Vector Calculus

Vector-values functions
To describe the location of the airplane following the path
shown in the following figure

you might consider using a point ( x,y,z) in three

dimensions. However, it turns out to be more convenient to
describe its location at any given time by the endpoint of a
vector whose initial point is located at the origin ( position
vector)( see the figure below) for vectors indicating the
location of the plane at a number of times .

Definition
A vector valued function r(t) is a mapping from its domain
D R to its range R V , so that for each t in D, r(t)=v for
exactly one vector v V . We can always write a vectorvalued function as :
3

r (t ) f (t )i g (t ) j h(t )k

. . . . . (1)

for some scalar functions f, g, and h ( called the component

functions of ( r ).
For each t, we regard r(t) as a position vector . The
endpoint of r(t) then can be viewed as tracing out a curve,
as we have shown in the last figure.

Observe that for r(t) as defined in equation (1) , this curve

is the same as that described by the parametric equations
x=f(t), y=g(t), and z=h(t). In three dimensions, such a
curve is referred to as a space curve. Likewise, we can
define a vector-valued function r(t) in V2 by
r (t ) f (t )i g (t ) j , for some scalar functions f and g.
Remark
We routinely use the variable t to represent the independent
variable for vector-valued functions, since in many
applications t represents time

Example
Sketch a graph of the curve traced out by the endpoint of
the two-dimensional vector-valued function:
r (t ) (t 1)i (t 2 2) j

Solution
Substituting some values for t, we have :
r (0) i 2 j 1,2
r ( 2) 3i 2 j 3,2 ,
and
r (2) i 2 j 1,2

We plot these in the following figure

The endpoints of all position vectors r(t) lie on the curve C,

described parametrically by
C : x t 1, y t 2 2

tR

we can eliminate the parameter by solving for t in terms of

x : t x 1

the curve is then given by : y t 2 ( x 1) 2

Notice that the graph of this is a parabola opening upward,
with vertex at the point (1,2) ( see the figure)
2

The small arrows marked on the graph indicate the

orientation, that is, the direction of increasing values of t. If
the curve describes the path of an object, then the
orientation indicated the direction in which the object
traverses the path. In this case, we can easily determine the
orientation from the parametric representation of the
curve. Since x t 1 , observe that x increases as t increases.
Eliminating the parameter from the parametric
representation of a curve is not always so easy as in the
following example:
Example
Sketch a graph of the curve traced out by the endpoint of
the vector-valued function :
r (t ) 4 cos ti 3 sin tj ,

tR

Solution
In this case, the curve can be written parametrically as :
x 4 cos t , y 3 sin t ,

tR

Instead of solving for the parameter t, it often helps to look

for some relationship between the variables. Here,
x

4

x

4

3
y

cos 2 t sin 2 t 1
1

which is the equation of an ellipse ( see the figure

To determine the orientation of the curve here, we will need

to look carefully at both parametric equations. First, fix a
starting place on the curve, for convenience, say (4,0) . This
corresponds to t 0, 2 , 4 , ... . As t increases , notice that
cost ( and hence, x) decreases initially, while sint
increases , we get the clockwise orientation indicated in the
last figure.
6

Vector Fields
To analyze the flight of an airplane, engineers use wind
tunnel tests to provide information about the flow of air
over the wings and around the fuselage. As you can
imagine, to model such a test mathematically, we need to
be able to describe the velocity of the air at various points
throughout the tunnel. So, we need to define a function
that assigns a vector to each point in space. Such a
function would have both a multidimensional domain and a
multidimensional range( like the vector- valued functions)
Although vector fields in higher dimensions can be very
useful, we will focus here on vector fields in two and three
dimensions.
Definition
A vector field in the plane is a function F(x,y) mapping
points in R2 into the set of two-dimensional vectors V2, we
write :
F ( x, y )

f 1 ( x, y ), f 2 ( x, y ) f 1 ( x, y )i f 2 ( x, y ) j

for scalar functions f ( x, y ) and f ( x, y) .

In space, a vector field is a function F ( x, y, z ) mapping
points in R into the set of three- dimensional vectors V .
In this case, we write;
1

F ( x, y, z ) f 1 ( x, y, z ), f 2 ( x, y, z ), f 3 ( x, y, z )
f 1 ( x, y, z )i f 2 ( x, y, z ) j f 3 ( x, y, z )k

for scalar functions f ( x, y, z ) , f

1

( x, y , z )

and

f 3 ( x, y , z )

To describe a two-dimensional vector field graphically, we

draw a collection of the vectors F ( x, y ) for various points
( x, y ) in the domain, in each case drawing the vector so
that its initial points is located at ( x, y ) .

Example
For the vector field F ( x, y) x y,3 y x , evaluate
(a) F(1,0)
(b) F(0,1) and
(c) F(-2,1)
Plot each vector using the point (x,y) as the initial
point.
Solution
(a) Taking x=1 and y=0 , we have
F (1,0) 1 0,0 1 1,1

In the following figure ,

we have plotted the vector 1,1 with its initial point

located at the point (1,0), so that its terminal point is
located at (2,-1).
(b) Taking x=0 and y=1, we have
F (0,1) 0 1,3 2 1,3 . In the same figure above , we
have also indicated the vector 1,3 , taking the point
(0,1) as its initial point , so that its terminal point is
located at (1,4)
(c ) with x=-2 and y=1, we have F (2,1) 2 1,3 2
In the figure above , the vector 1,5 is plotted by
placing its initial point at
(-2,1) and its terminal point at (-3,6).

1,5

Graphing vector fields poses something of a problem.

Notice that the graph of a two-dimensional vector field
would be four-dimensional ( i.e., two independent
variables plus two dimensions for the vectors). Likewise,
the graph of a three-dimensional vector field would be
six-dimensional . Despite this , we can visualize many of
the important properties of a vector field by plotting a
number of values of the vector field as we had started to
do in the previous figure. In general, by the graph of
vector field F(x,y), we mean a two-dimensional graph
with vectors F(x,y), plotted with their initial point
located at (x,y), for a variety of points (x,y).

Example
Graph the vector fields
F ( x, y ) x , y ,

G ( x, y )

x, y
x2 y2

, and

H ( x, y ) y , x

, and

identify any patterns.

Solution
First choose a variety of points (x,y), evaluate the vector
field at these points and plot the vectors using (x,y) as
the initial point. Notice that in the following table we
10

have chosen points on the axes and in each of the four

quadrants.

The vectors indicated in the table are plotted in the

following figure

A computer- generated plot of the vector field is shown

in the figure below
11

Notice that the vectors all points away from the origin
and increase in length as the initial points get farther
from the origin. In fact, the initial point (x,y) lies a
distance x y from the origin and the vector x, y
has length x y , so , the length of each vector
corresponds to the distance from its initial point to the
origin. This gives us an important clue about the graph
of G(x,y). Although the formula may look messy, notice
that G(x,y) is the same as F(x,y) except for the division
by x y . Which is the magnitude of the vector x, y .
Recall that dividing a vector by its magnitude yields a
unit vector in the same direction. Thus, for each x, y ,
G(x,y) has the same direction as F(x,y), but is a unit
vector. A computer-generated plot of G(x,y) is shown in
the figure below
2

12

the following table shows some sample vectors for

H(x,y)

The plot of these vectors are shown in the following

figure
13

A computer-generated plot of H(x,y) is shown in the

figure below

If we think of H(x,y) as representing the velocity field for

a fluid in motion , the vectors suggest a circular rotation
14

of the fluid. Recall that tangent lines to a circle are

perpendicular to radius lines. The radius vector from the
origin to the point (x,y) is x, y , which is perpendicular
to the vector y, x , since x, y . y, x =0. Also, notice
that the vectors are not of constant size. As for F(x,y) ,
the length of the vector y, x is x y , which is the
distance from the origin to the initial point (x,y).
2

Gradient Fields
An important type of vector field with which we already
have some experience is the gradient field. Where the
vector field is the gradient of some scalar function.
Because of the importance of gradient fields, there are a
number of terms associated with them. In the following
definition we do not specify the number of independent
variables , since the terms can be applied to functions of
two, three or more variables.
Definition
For any scalar function f, the vector field F f is called
the gradient field for the function f. We call f a potential
function for F. Whenever F f , for some scalar function
f, we refer to F as a conservative vector field.

Example
15

Find the gradient fields corresponding to the functions:

(a)
(b)

f ( x, y ) x 2 y e y
1
g ( x, y , z ) 2
x y2 z2

and use a CAS to sketch the fields.

Solution
(a) we first compute the partial derivatives
and
f ( x, y )

f
x2 e y
y

f f
,
x y

f
2 xy
x

, so that

2 xy , x 2 e y

A computer-generated graph of
figure below

f ( x , y )

is shown in the

16

(b) for g(x,y,z)=(x2+y2+z2)-1 , we have

g
2x
( x 2 y 2 z 2 ) 2 ( 2 x )
2
x
x y2 z2

and by symmetry, we conclude that

g
2y
( x 2 y 2 z 2 ) 2 (2 y )
2
y
x y2 z2

and z ( x y
This gives us
g ( x , y , z )

z 2 ) 2 (2 z )

f f f
,
,
x y z

2z
2

y2 z2

.
2

2 x, y , z
(x y 2 z 2 )2
2

A computer-generated graph of
following figure

g ( x, y , z )

is shown in the

17

Many calculations involving vector field simplify

dramatically if the vector field is a gradient field( i.e., if the
vector field is conservative). To take full advantage of these
simplifications, you will need to be able to construct a
potential function that generates a given conservative field.
Example
Determine whether each of the following vector fields is
conservative. If it is, find a corresponding potential
function f(x,y):
(a) F ( x, y) 2 xy 3, x cos y
(b) G ( x, y) 3x y 2 y, x y 2 x
2

Solution
The idea here is to try to construct a potential function. In
the process of trying to do so, we may instead recognize
that there is no potential function for the given vector field.
For (a), if f(x,y) is a potential function for F(x,y), we have
that
f ( x, y ) F ( x, y ) 2 xy 3, x 2 cos y

so that

f
2 xy 3, and
x

f
x 2 cos y
y

. . . . .

( 2)

integrating the first of these two equations with respect to x

and treating y as a constant, we get
f ( x, y )

(2 xy 3)dx x

y 3x g ( y )

(3)

18

Here, we have added an arbitrary function of y alone, g(y),

rather than a constant of integration, because any function
of y is treated as a constant when integrating with respect
to x. Differentiating the expression for f(x,y) with respect to
y gives us
f
( x, y ) x 2 g ( y ) x 2 cos y
y

from equation (2). This gives us g ( y) cos y , so that

g ( y ) cos ydy sin y c , and from equation (3) we have
f ( x, y ) x 2 y 3 x sin y c

where c is an arbitrary constant. Since we have been able

to construct a potential function, the vector field F(x,y) is
conservative.

b) Again, we assume that there is a potential function f

for G(x,y) and try to construct it. In this case, we have
g ( x, y ) G ( x, y ) 3 x 2 y 2 2 y , x 2 y 2 x

so that

g
3x 2 y 2 2 y
x

g
x 2 y 2x
y

. . .

(4)

Integrating the first equation in (4) with respect to x , we

have
g ( x, y )

3x

y 2 2 y dx x 3 y 2 2 yx h( y )

where h is an arbitrary function of y. Differentiating this

with respect to y, we have

19

f
( x, y ) 2 x 3 y 2 x h ( y ) x 2 y 2 x
y

from equation (4), solving for

h ( y ) ,

we get

h ( y ) x y 2 x 2 x y 2 x x y 2 x y
2

which is impossible, since h(y) is a function of y alone.

We then conclude that there is no potential function for
G(x,y) and so, the vector field is not conservative

Line Integrals
In section (5) we used integration to find the mass of a thin
rod with variable mass density . There, we had observed
that if the rod extends from x=a to x=b and has mass
density function (x), then the mass of the rod is given by
b

( x)dx . This definition is fine for objects that are

a

20

essentially one-dimensional, but what if we wanted to find

the mass of a helical spring ( see the figure)

By using the same technique we can solve a wide variety of

problems. As we have done before, we will derive a
solution by first approximating the curve with line
segments and then taking a limit.
In this three-dimensional setting, the density function has
the form (x,y,z) ( where is measured in units of mass per
unit length). We assume that the object is in the shape of a
curve C in three dimensions with endpoints (a,b,c) and
(d,e,f) . Further , we assume that the curve is oriented,
which means that there is a direction to the curve. For
example , the curve C might start at (a,b,c) and end at
(d,e,f). We first partition the curve into n pieces with
endpoint
(a,b,c)=(xo,yo,zo), (x1,y1,z1),(x2,y2,z2),,(xn,yn,zn)=(d,e,f) as
indicated in the following figure

21

we will use the shorthand i to denote the point ( xi,yi,zi),

and Ci for the section of the curve C extending from i-1 to
i , for each i=1,2,,n. Our initial objective is to
approximate the mass of the portion of the object along Ci.
Note that if the segment Ci is small enough, we can
consider the density to be constant on Ci . In this case , the
mass of this segment would simply be the product of the
density and the length of Ci . For some point (xi*,yi*,zi*) on
Ci , we approximate the density on Ci by (xi*,yi*,zi*) . The
mass of the section Ci is then approximately
(xi*,yi*,zi*)si
where si represent the arc length of Ci . The mass m of the
entire object is then approximately the sum of the masses of
the n segments
n

m ( xi* , y i* , z i* ) s i
i 1

22

This approximation will be improved as we divide the curve

into more and more segments that are shorter and shorter
in length. Finally, taking the norm of the partition p to be
the maximum of the arc lengths si ( i=1,2,,n) , we have
m lim

p 0

(x , y , z
i 1

*
i

*
i

*
i

)si

. . . . . .(1)

provided the limit exists and is the same for every choice of
the evaluation points (xi*,yi*,zi*) , i=1,2,,n
You might recognize that equation (1) looks like the limit of
a Riemann sum ( an integral). As it turns out, this limit
arises naturally in numerous applications. We pause now to
give this limit a name and identify some useful properties

Definition
The line integral of f(x,y,z) with respect to arc length along
the oriented curve C in three-dimensional space, written
f ( x, y, z )ds , is defined by
C

f ( x, y , z ) ds lim

p 0

f (x
i 1

*
i

, yi* , zi* ) si

provided that this limit exists and is the same for all
choices of evaluation points.
Furthermore, we can define line integrals of functions
f(x,y) of two variables along an oriented curve C in the

23

xy-plane in a similar way. Often, the curve C is specified

by parametric equations . The following theorem allows us
to evaluate the line integral as a definite integral of a
function of one variable.

Theorem ( Evaluation Theorem)

Suppose that f(x,y,z) is continuous in a region D containing
the curve C and that C is described parametrically by
( x(t),y(t),z(t)), for a t b, where x(t),y(t),and z(t) have
continuous first derivatives. Then,

f ( x, y, z )ds

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

y (t ) z (t ) dt
2

Suppose that f(x,y) is continuous in a region D containing

the curve C and that C is described parametrically by
(x(t),y(t)) , for a t b , where x(t) and y(t) have
continuous first derivatives. Then

x(t ) 2 y(t ) 2 dt

f ( x, y )ds f ( x(t ), y (t ))
a

Proof
We prove the result for the case of a curve in two
dimensional and can prove the three-dimensional case in
the same manner. From the definition ( adjusted for the
two-dimensional case), we have
f ( x, y )ds lim f ( x , y )s . . . . .(2)
n

p 0

i 1

*
i

*
i

24

where si represents the arc length of the section of the

curve C between ( xi-1,yi-1) and (xi,yi). Choose to,t1,,tn so
that x(ti)=xi and y(ti)=yi, for i=1,2,,n. We approximate the
arc length of such a small section of the curve by the
straight-line distance:
xi xi 1 2 y i yi 1 2

si

Further, since x(t) and y(t) have continuous first

derivatives, we have by the mean value theorem ( as in the
derivation of the arc length formula in chapter 9 ) , that
s x x y y x(t y(t t
2

for some

i 1

t i* (t i 1 , t i )
n

i 1

* 2
i

. Together with equation (2), this gives us

f ( x, y )ds lim f ( x(t

p 0

* 2
i

i 1

*
i

), y (ti* )

x(t ) y(t )
*
i

f ( x(t ), y (t )) x(t ) y(t )

2

*
i

t i

dt

A curve C is called smooth if it satisfies the hypotheses of

the last theorem and one additional condition. Specifically,
C is smooth if it can be described parametrically by x=x(t),
y=y(t) and z=z(t) , for a t b, where x(t) , y(t) and z(t)
all have continuous first derivatives and
x(t ) y (t ) z (t ) 0 on the interval [a,b].
2

25

Example
Find the mass of a spring in the shape of the helix defined
parametrically by :
x=2cost , y=t , z=2sint , for 0 t 6 , with density
(x,y,z)=2y.
Solution
A graph of the helix is shown in the figure below:

The density is
( x, y , z ) 2 y 2t

and from equation (3) , the arc length element ds is given

by
ds

x(t ) 2 y(t ) 2 z(t ) 2 dt

2 sin t 2 1 2 2 cos t 2 dt

5dt

where we have used the identity

4 sin t 4 cos t 4. from the evaluation theorem , we have :
2

26

mass ( x, y, z ) ds
C

2 5

6
2

2
t

( x, y , z )

5dt
ds

tdt
0

36 2 5

We should point out that this example is unusual in at least

one respect: we were able to compute the integral exactly.
Most line integrals of the type defined in the first definition
are too complicated to evaluate exactly and will need to be
approximated with some numerical method .

Example
Evaluate
(2 x
c

y ) ds

, where C is the upper half of the unit circle

x +y =1.
Solution
We first need the parametric equation to represent C .
Recall that the unit circle can be parameterized by means
of the equations
x= cost , and y=sint
27

and the upper half of the circle is described by the

parameter interval 0 t ( see the figure)

therefore we have

dx

dt

2
2
c(2 x y )ds (2 cos t sin t )
0

dy
dt
dt

(2 cos 2 t sin t ) sin 2 t cos 2 t dt

0

cos 3 t
(2 cos t sin t )dt 2t

0
2

2
0

2
3

28

Suppose now that C is a piecewise-smooth curve; that is ,C

is a union of finite number of smooth curves C1,C2,,Cn .
( as in figure) ,

where the initial point of Ci+1 is the terminal of Ci. Then we

define the integral of f along C as the sum of the integrals
of f along each of the smooth pieces of C:
f ( x, y )ds
c

c1

f ( x, y ) ds f ( x, y )ds ... f ( x, y ) ds
c2

cn

29

Example
Evaluate
2 xds , where c consists of the arc c1 of the parabola y=x2
from (0,0) to (1,1) followed by the vertical line segment c2
from (1,1) to (1,2).
c

Solution
The curve C is shown in the figure below

C1 is the graph of a function of x , so we can choose x as

the parameter and the equation for c1 become:
x=x

y=x2

0x 1

30

dx

dx

c2 xds 2 x
1

dy
dx
dx

therefore

2 x 1 4 x 2 dx
0

1 2
. (1 4 x) 2
4 3

5 5 1
6

on c2 we choose y as the parameter , so the equations of c2

are
x=1 , y=y , 1 y 2
and

dx
c 2 xds 1 2(1) dy
2
2

dy

dy

Thus 2 xds 2 xds 2 xds 5

c

c1

c2

dy 2dy 2
1

5 1
2
6

It frequently happens that line integrals with respect to x

occur together. When this happens, its customary to
abbreviate by writing:
P( x, y )dx Q( x, y )dy P ( x, y )dx Q( X , y )dy
c

when we are setting up a line integral, sometimes the most

difficult thing is to think of a parametric representation for
a curve whose geometric description is given. In particular
we often need to parameterize a line segment, so its useful
to remember that a vector representation of the line
segment that starts at r0 and ends at r1 is given by:
r (t ) (1 t )ro tr1

0 t 1

31

Example
Evaluate y dx xdy , where
(a) C=C is the line segment from (-5,-3) to (0,2) and
(b) C=c is the arc of the parabola x=4-y2 from ( -5,-3)
to (0,2)
( see the figure)
2

Solution
(a) A parametric representation for the line segment is
x=5t-5
y=5t-3
0t1
We use r 5,3 , and r 0,2
Then dx= 5dt , dy=5dt ,
o

32

c1

y dx xdy (5t 3) 2 (5dt ) (5t 5)(5dt )

2

5 ( 25t 2 25t 4) dt
0

25t 3 25t 2

4t
2
3

5
6

(b) Since the parabola is given as a function of y, lets take

y as the parameter and write c2 as
x=4-y2 y=y
-3 y 2
then dx=-2y , we have

c2

y 2 dx xdy

( 2 y )dy (4 y 2 )dy

3
2

(2 y

y 2 4)dy

y4
y3

4 y
4
3

245
6

Note that we got different answers in parts (a) and (b)

of the example above even though the two curves had
the same endpoints. Thus, in general, the value of a
line integral depends not just on the endpoints of the
curve but also on the path. The answers also depend
on the direction , or orientation, of the curve. If (-c1)
denotes the line segment from (0,2) to (-5,-3), you can
verify, using parameterization
x=-5t , y=2-5t
5
That : y dx xdy 6

0t1

33

In general, a given parameterization x=x(t) , y=y(t) ,

a t b determines an orientation of a curve C , with the
positive direction corresponding to increasing values of the
parameter t ( see the figure below , where the initial point
A corresponds to the parameter value a and the terminal
point B corresponds to t=b)

If c denotes the curve consisting of the same point as C

but with the opposite orientation ( from initial point B to
terminal point A as in the same figure above), then we have

f ( x, y ) dx f ( x, y ) dx
c

f ( x, y ) dy f ( x, y ) dy
c

34

but if we integrate with respect to arc length, the value of

the line integral does not change when we reverse the
orientation of the curve :

f ( x, y ) ds

f ( x, y ) ds

this is because si is always positive, whereas xi , and

yi change sign when we reverse the orientation of C.

Example
Evaluate the line integral (2 x 3 yz )ds , where C is the
curve defined parametrically by x=cost , y=sint , and
z=cost with 0 t 2.
2

Solution
A graph of C is shown in the figure below:

35

The integrand is given by :

f ( x, y, z ) 2 x 2 3 yz 2 cos 2 t 3 sin t cos t

The arc length element is given by

ds

x (t ) 2 y (t ) 2 z (t ) 2 dt

( sin t ) 2 (cos t ) 2 ( sin t ) 2 dt 1 sin 2 t dt

where we have used the identity sin2t+cos2t=1. We now

have

(2 x 2 3 yz )ds

(2 cos

0

t 3 sin t cos t )

2
1 sin
tdt 6.9922
ds

2 x 2 3 yz

The last integral was approximated numerically.

Example
Evaluate the line integral 2 x yds , where C is :
(a) the portion of the parabola y=x2 from (-1,1) to (2,4)
2

36

(b) the portion of the parabola from (2,4) to (-1,1).

Solution
(a) A sketch of the curve is shown in the figure below

Taking x=t as the parameter, we can write parametric

equations for the curve as x=t and y=t2, for -1 t 2. using
this, the integrand becomes 2x2y=2t2t2=2t4. The arc length
element is
ds

x (t ) 2 y (t ) 2 dt

1 4t 2 dt

The integral is now written as:

2

2
4
C 2 x yds 2 t

1 2 x 2 y

2
1 4t dt 45.391
ds

where we have evaluated the integral numerically.

37

(b) The curve is the same as in part (a) , except that the
orientation is backward ( see the figure )

In this case , we represent the curve with the parametric

equations x=-t and y=t2 , for -2 t 1. Observe that every
thing in the integral remains the same and we have

2 xyds

2t

1 4t 2 dt 45.391

as before.

Theorem (*)
Suppose that f(x,y,z) is a continuous is a region D
containing the curve C and that C is described
parametrically by x=x(t) , y=y(t), and z=z(t), where t
ranges from t=a to t=b and x(t),y(t) and z(t) have
continuous first derivatives. Then:

38

f ( x, y, z ) dx f ( x(t ), y (t ), z (t )) x (t )dt

f ( x, y, z ) dy f ( x(t ), y (t ), z (t )) y (t )dt

f ( x, y, z ) dz f ( x(t ), y (t ), z (t )) z (t ) dt

a
b

and

a
b

we will examine some simple examples. Recall that the line

integral along a given curve with respect to arc length will
not traverse the curve in the opposite direction. On the
other hand you can see from the next example that the line
integral with respect to x,y,or z change sign when the
orientation of the curve changes.

Example
Compute the line integral (4 xz 2 y )dx , where C is the
line segment :
(a) from (2,1,0) to (4,0,2)
(b) from (4,0,2) to (2,1,0)
C

Solution
First, parametric equations for C for part (a) are
x 2 (4 2)t 2 2t
y 1 (0 1)t 1 t

and

for

0 t 1

z 0 (2 0)t 2t

39

The integrand is then

4xz+2y=4(2+2t)(2t)+2(1-t)=16t2+14t+2
and the element dx is given by:
dx x (t ) dt 2dt

From the theorem above , the line integrals is now given

by
1

2
C (4 xz 2 y)dx (16t 14t 2)
0

4 xz 2 y

86
(2)dt
3
dx

For part (b) , we can observe that the line segment

connects the same two points as in part (a) , but in the
opposite direction. The same parametric equations will
then work, with the single change that t will run from t=1
to t=0. This gives us

(4 xz 2 y )dx (16t 2 14t 2)(2)dt

1

86
3

where you should recall that reversing the order of

integration changes the sign of the integral.

Theorem
Suppose that f(x,y,z) is a continuous function in some
region D containing the oriented curve C. Then , the
following hold.
(i)
If C is piecewise-smooth , then

f ( x, y , z ) dx f ( x, y , z ) dx
C

40

(ii)

If C C C ... C where C1, C2, Cn are all

smooth and the terminal point of C is the same as
the initial point of C, for i=1,2,,n-1, then
1

f ( x, y , z ) dx

C1

f ( x, y , z ) dx

C2

f ( x, y, z ) dx ...

Cn

f ( x, y , z )

Example
Compute 4 xdy 2 ydz , where C consists of the line
segment from (0,1,0) to (0,1,1), followed by the line
segment from (0,1,1) to (2,1,1) and followed by the line
segment from (2,1,1) to (2,4,1).
C

Solution
The sketch of the curves is in the following figure

41

Parametric equations for the first segment C1 are x=0,

y=1 and z=t with 0 t 1. On this segment, we have
dy=0dt and dz=1dt. On the second segment C2,
parametric equations are x=2t, y=1 and z=1 with
0 t 1. On this segment , we have dy=dz=0dt. On the
third segment, parametric equations are x=2,y=3t+1
and z=1 with 0 t 1 . On this segment , we have
dy=3dt and dz=0dt. Putting this all together, we have

4 xdy 2 ydz 4 xdy 2 ydz 4 xdy 2 ydz ... 4 xdy 2 ydz

C1

C2

Cn

1
1

4
(
0
)
(
0
)

2
(
1
)
(
1
)
dt

4(2t ) (0)
0
0
4x
y (t )
2y
z ( t )
4x
y ( t )

4
(2t ) (0)
0
4x
y (t )
1

2
(1) (0)dt
2y

z ( t )

2 (1) (0)dt 26dt 26

0
2y
z ( t )
1

42

Notice that the line integral will be zero if the integrand

simplifies to 0 or if the variable of integration is constant
along the curve. For instance, if z is constant on some
curve, then the change in z ( given by dz) will be 0 on
that curve.
Recall that our motivation for introducing line integrals
with respect to the three coordinate variables was to
compute the work done by the force field while moving
an object along a curve. The work performed by the
force field : F ( x, y, z ) F ( x, y, z ), F ( x, y, z ), F ( x, y, z )
along the curve defined parametrically by x=x(t),y=y(t)
and z=z(t) , for a t b , is given by
1

F1 ( x(t ), y (t ), z (t )) x (t )dt F2 ( x(t ), y (t ), z (t )) y (t )dt

a

F3 ( x(t ), y (t ), z (t )) z (t ) dt

we should

now recognize that we can rewrite each of the three

terms in this expression for work using theorem (*) , to
obtain
W

F1 ( x, y , z ) dx

F2 ( x, y , z ) dy

F3 ( x, y , z ) dz

we can write such a combination of the line integrals in

a simpler form by introducing some notation . Suppose
that a vector field F ( x, y, z ) F ( x, y, z ), F ( x, y, z ), F ( x, y, z ) .
Since we use r=xi+yj+zk, we define dr=dxi+dyj+dzk or
simply dr dx, dy, dz
we now define the line integral
1

F ( x, y, z ).dr

F1 ( x, y, z ) dx F2 ( x, y , z ) dy F3 ( x, y , z ) dz

F1 ( x, y , z ) dx F2 ( x, y , z ) dy F3 ( x, y, z )dz
C

43

In this case where F(x,y,z) is the force field, the work

done by F in moving a particle along the curve C can be
written simply as :
W

F ( x, y , z ).dr

Notice how different parts of F ( x, y, z ).dr correspond

to our knowledge of work. The only way in which the xcomponent of force affects the work done is when the
object moves in the x-direction ( i.e,. when dx0).
Similarly, the y-component of force contributes to the
work only when dy0 and the z-component of force
contributes to the work only when dz0.
C

Example
Compute the work done by the force field
F ( x, y , z ) 4 y ,2 xz ,3 y acting on an object as it moves along
the helix defined parametrically by x=2cost , y=2sint
and z=3t, from the point (2,0,0) to the point (-2,0,3).
Solution
The work is given by
W

F ( x, y , z ).dr

4 ydx 2 xzdy 3 ydz

We have already provided parametric equations for the

curve, but not the range of t-values. Notice that from
z=3t, you can determine that (2,0,0) corresponds to t=0
and (-2,0,3) corresponds to t= . Substituting in for
x,y,z and dx=-2sintdt , dy=2costdt and dz=3dt, we have

44

4 ydx 2 xzdy 3 ydz

0 4 (2sint )
4y

(2 sin t 2(2 cos t )(3t )

x ( t )

(2 cos t )

y ( t )

2 xz

3(2 sin t )

3y

(3)dt

z( t )

( 16 sin 2 t 24t cos 2 t 18 sin t )dt 36 8 6 2

0

this result is yielded by using a computer algebra system

to evaluate the final integral
Example
Compute the work done by the force field F ( x, y ) y, x
acting on an object as it moves along the parabola y=x21 from (1,0) to (-2,3).
Solution
The work is given by
W

F ( x, y ).dr

ydx xdy

here, we use x=t and y=t2-1 as parametric equations for

the curve, with t ranging from t=1 to t=-2. In this case,
dx=1dt and dy=2tdt and the work is
W

ydx xdy

(t
1

1)(1) (t )(2t ) dt

(t

1) dt 6

Independence of path and conservative vector field

From the previous discussion and examples , there are a
lot of steps needed to evaluate a line integral . First we
must parameterize the curve , rewrite the line integral as
45

a definite integral and then evaluate the resulting

definite integral. While this process is unavoidable for
many line integrals, we will now consider a group of line
integrals that are the same along every curve connecting
the given endpoints. We will determine the circumstances
under which this occurs and see that when this does
happen, there is a simple way to evaluate the integral.
We begin with a simple observation. Consider the line
integral F .dr , where F ( x, y) 2 x,3 y and C1 is the
straight line segment joining the two points (0,0) and
(1,2) ( see the figure )
2

C1

To parameterize the curve, we take x=t and y=2t, for

0 t 1. We then have

46

C1

F .dr

C1

2 x,3 y 2 . dx, dy

2t 12t

C1

2 xdx 3 y 2 dy

( 2) dt 9

For the same vector field F(x,y), consider now F .dr ,

where C2 is made up of the horizontal line segment from
(0,0) to (1,0) followed by the vertical line segment from
(1,0) to (1,2) ( see the figure)
C2

In this case we have

F .dr
C2

C2

2 x,3 y 2 . dx, dy

2
2 xdx 3 y dy 9

Look carefully at these two line integrals. Although the

integrands are the same and the endpoints of the two
47

curves are the same, the curves followed are quite

different. You should try to compute this line integral
over several additional curves from (0,0) to (1,2). You
will find that each line integral has the same value : 9 .
This integral is an example of one that is the same along
any curve from (0,0) to (1,2).

Let C be any piecewise-smooth curve, traces out by the

endpoint of the vector-valued function r(t), for a t b.
In this context, we usually refer to a curve connecting
two given points as a path. We say that the line integral
F .dr is independent of path in the domain D if the
integral is the same for every path contained in D that
has the same beginning and ending points.
C

Definition
A region D R ( for n 2 ) is called connected if every
pair of points in D can be connected by a piecewisesmooth curve lying entirely in D.
n

In the figure below you can see a region in R2 that is

connected

48

And in the following figure we indicate a region that is

not connected.

We are now able to prove a result concerning integrals

that are independent of path.
Theorem (1)

49

Suppose that the vector field F ( x, y) M ( x, y ), N ( x, y ) is

continuous on the open, connected region D R . Then,
the line integral F ( x, y).dr is independent of path in D
if and only if F is conservative on D.
2

Proof
Recall that a vector field F is conservative whenever
F f , for some scalar function f ( called a potential
function for F). There are several things to prove here.
First, suppose that F is conservative, with F ( x, y ) f ( x, y )
.Then
F ( x, y ) M ( x, y ), N ( x, y ) f ( x, y )

f x ( x, y ), f y ( x, y )

and so , we must have

M ( x, y ) f x ( x, y )

and

N ( x, y ) f y ( x , y )

Let A(x,y) and B(x,y) be any two points in D and let C be

any smooth path from A to B, lying in D and defined
parametrically by :
C: x=g(t), y=h(t) , where t1 t t2 . Then we
have

F ( x, y ).dr M ( x, y )dx N ( x, y )dy

C

f x ( x, y )dx f y ( x, y )dy
C

t2

f x ( g (t ), h(t )) g (t ) f y ( g (t ), h(t ))h(t ) dt

(1)

t1

Notice that since fx and fy were assumed to be

continuous, we have by the chain rule that
d
f ( g (t ), h(t )) f x ( g (t ), h(t ) g (t ) f y ( g (t ), h(t )h (t )
dt

which is the integrand in equation (1)

50

By the fundamental theorem in calculus, we now have

F ( x, y ).dr

t2

( g (t ), h(t )) g (t ) f y ( g (t ), h(t ))h (t ) dt

t1

d
f ( g (t ), h(t )) dt
dt
f ( g (t 2 ), h(t 2 )) f ( g (t1 ), h(t1 ))

t2

t1

f ( x 2 , y 2 ) f ( x1 , y1 )

In particular this tells us that the value of the integral

depends only on the value of the potential function at the
two endpoints of the curve and not on the particular path
followed. That is, the line integral is independent of
path, as desired.
Next we need to prove that if the integral is independent
of path, that the vector field must be conservative. So,
suppose that f ( x, y ).dr is independent of path in D.
For any points (u,v) and ( x , y ) D , define the function
C

( u ,v )

f (u , v )

F ( x, y ).dr

( xo , y o )

( we are using the variables u and v , since the variables

x and y inside the integral are dummy variables and
cannot be used both inside and outside the line integral).
Notice that since the line integral is independent of path
in D, we need not specify a path over which to
integrate ; its the same over every path in D connecting
these points. ( since D is connected, there is always a
path lying in D that connects the points.) Further , since
D is open , there is a disk centered at (u,v) and lying
completely inside D. Pick any point (x1,v) in the disk with
x1<u and let C1 be any path from (xo,yo) to (x1,v) lying in
51

D. So, in particular, if we integrate over the path

consisting of C1 followed by the horizontal path C2
indicated in the following figure

we must have
f (u , v)

( x1 ,v )

( u ,v )

F ( x, y ).dr

( xo , y o )

F ( x, y).dr

(2)

( x1 , v )

Observe that the first integral in (2) is independent of u.

So, taking the partial derivative of both sides of (2) with
respect to u , we get
( x ,v )

( u ,v )

f u (u , v)
F ( x, y ).dr
F ( x, y ).dr

u ( xo , yo )
u ( x1,v )
( u ,v )

F ( x, y ).dr
u ( x1,v )
( u ,v )

M ( x, y )dx N ( x, y )dy
u ( x1,v )

52

Notice that on the second portion of the indicated path, y

is a constant and so, dy=0. This gives us
( u ,v )

f u (u , v)

( u ,v )

M ( x, y )dx N ( x, y )dy
M ( x, y )dx

u ( x1 ,v )
u ( x1,v )

Finally, from the second form of the fundamental

theorem of calculus, we have
( u ,v )

f u (u , v)
M ( x, y )dx M (u , v)
u ( x1,v )

(3)

Similarly, pick any point (u,y1) in the disk centered

at(u,v) with y1<v and let C1 be any path from ( xo,yo) to
( u,y1) lying in D. Then, integrating over the path
consisting of C1 followed by the vertical path C2
indicated in the figure below

we find that

53

f (u , v)

( u , y1 )

(u ,v )

F ( x, y).dr

F ( x, y ).dr

( xo , y o )

(4) In

( u , y1 )

this case , the

first integral is independent of v. So, differentiating both

sides of (4) with respect to v , we have
(u , y )

f v (u , v)

( u ,v )

F ( x, y ).dr
F ( x, y ).dr

v ( xo , yo )
v ( u , y1 )

( u ,v )

F ( x, y ).dr
v ( u , y1 )
( u ,v )

M ( x, y ) dx N ( x, y )dy
v ( u , y1 )

N ( x, y )dy N (u , v)
v ( u, y1

( u ,v )

(5)

by the second form of the fundamental theorem of

calculus, where we have used the fact that on the second
part of the indicated path, x is a constant, so that dx=0.
Replacing u and v by x and y, respectively, in (3) and (5)
establishes that
F ( x, y ) M ( x, y ), N ( x, y )

f x ( x, y ), f y ( x, y ) f ( x, y )

so that F is conservative in D
Theorem (2)( Fundamental theorem for line integrals)
Suppose that F ( x, y) M ( x, y ), N ( x, y ) is continuous in the
open, connected region D R and that C is any
piecewise-smooth curve lying in D, with initial point
( x1,y1) and terminal point (x2,y2). Then , if F is
conservative on D, with F ( x, y ) f ( x, y) , we have
2

x ,y )

is clear that the

advantage of the last theorem for a conservative vector
C

F ( x, y ).dr f ( x, y ) ( x2 , y2 ) f ( x 2 , y 2 ) f ( x1 , y1 ) It
1

54

field, that we dont need to parameterize the path to

compute a line integral; we need only find a potential
function and then simply evaluate the potential function
between the endpoints of the curve.
Example
Show that for F ( x, y) 2 xy 3, x 4 y 5 , the line integral
F ( x, y ).dr is independent of path. Then, evaluate the
line integral for any curve C with initial point at (-1,2)
and terminal point at (2,3).
2

Solution
From theorem (1), the line integral is independent of
path if and only if the vector field F(x,y) is conservative.
So, we look for a potential function for F, that is, a
function f(x,y) of which
F ( x, y ) 2 xy 3, x 2 4 y 3 5 f ( x, y )

f x ( x, y ), f y ( x, y )

This occurs when

fx=2xy-3 and fy=x2+4y3+5
(6)
Integrating the first of these two equations with respect
to x ( note that we might just as easily integrate the
second one with respect to y), we get
f ( x, y ) ( 2 xy 3) dx x 2 y 3 x g ( y )

(7 )

where g(y) is some arbitrary function of y alone. ( Recall

that we get an arbitrary function of y instead of a
constant of integration, since we are integrating a
function of x and y with respect to x.) Differentiating
with respect to y , we get
f y ( x, y ) x 2 g ( y )

Notice that from (6) , we already have an expression for

fy. Setting these two expressions equal, we get
55

x 2 g ( y ) x 2 4 y 3 5

and substituting x2 from both sides , we get

g ( y ) 4 y 3 5

Finally, integrating this last expression with respect to y

gives us
g(y)=y4+5y+c
We now have from (7) that
f(x,y)=x2y-3x+y4+5y+c
Is a potential function for F(x,y), for any constant c.
Now that we have found a potential function, we have by
theorem (2) that for any path from (-1,2) to (2,3),

F ( x, y ).dr f ( x, y )

( 2 , 3)
( 1, 2 )

2 (3) 3(2) 3 4 5(3) c 2 3 2 4 5(2) c

71
2

Notice that

when we evaluated the line integral in the last example,

the constant c in the expression for the potential function
dropped out. For this reason, we usually take the
constant to be zero when we write down a potential
function.
We consider a curve C to be closed if its two endpoints
are the same. That is, for a plane curve C defined
parametrically by
C ( x, y ) / x g (t ), y h(t ), a t b

C is closed if (g(a),h(a))=(g(b),h(b)). Theorem (3)

provides us with an important connection between
conservative vector fields and line integrals along closed
curves.
Theorem (3)
Suppose that F(x,y) is continuous in the open, connected
region D R . Then F is conservative on D if and only if
2

56

F ( x, y ).dr 0

for every piecewise-smooth closed curve C

lying in D.
Proof
Suppose that F ( x, y ).dr 0 for every piecewise-smooth
closed curve C lying in D. Take any two points P and Q
lying in D and let C1 and C2 be any two piecewisesmooth closed curves from P to Q that lie in D, as
indicated in the figure below
C

Notice that since D is connected , there always exist such

curves .) Then , the curve C consisting of C1 followed by
C2 is a piecewise-smooth closed curve lying in D as
indicated in the following figure

57

It now follows that

0

F ( x, y ).dr

C1

C1

F ( x, y ).dr

C2

F ( x, y ).dr

F ( x, y ).dr F ( x, y ).dr
C2

F ( x, y ).dr F ( x, y ).dr
So that
Since C1 and C2 were any two curves from P to Q, we
have that F ( x, y ).dr is independent of path and so, F is
conservative by theorem (1) . The second half of the
theorem ( that F conservative implies F ( x, y).dr 0 for
every piecewise-smooth closed curve C lying in D) is a
simple consequence of theorem (2) .
The line integrals are not always independent of path.
Said , differently, not all vector fields are conservative.
In view of this, it would be helpful to have a simple way
of deciding whether or not a line integral is independent
of path before going through the process of trying to
construct a potential function.
Note that by theorem (1) , if F ( x, y) M ( x, y ), N ( x, y ) is any
continuous on the open , connected region D and the line
integral F ( x, y ).dr is independent of path, then F must
C1

C2

58

be conservative. That is, there is a function f(x,y) in

which F ( x, y ) f ( x, y ) , so that
M ( x, y ) f x ( x, y )

and

N ( x, y ) f y ( x, y )

Differentiating the first equation with respect to y and

the second with respect to x , we have
M y ( x, y ) f xy ( x, y )

and

N x ( x, y ) f yx ( x, y )

Notice that if My and Nx are continuous in D, then the

mixed second partial derivatives fxy(x,y) and fyx(x,y)
must be equal in D, then we have ;
M ( x, y ) N ( x, y ) for all (x,y) in D. As it turns out, if we
further assume that D is simply-connected ( that is ,
every closed curve in D encloses only points in D), then
the converse of this result is also true [ i.e., F ( x, y ).dr
is independent of path whenever My=Nx in D].the
simply- connected region is shown in the following
figure
y

The following figure showing a region that is not simplyconnected .

59

You can think about simply connected regions as

regions that have no holes. The following results can be
found.
Theorem (4)
Suppose that M(x,y) and N(x,y) have continuous first
partial derivatives on a simply-connected region D. Then
M ( x, y )dx N ( x, y )dy is independent of path in D if and
only if My(x,y)=Nx(x,y) for all (x,y) in D.
C

We have already proved that independence of path

implies that My(x,y)=Nx(x,y) for all (x,y) in D.
Example
Determine whether or not the line integral
(e x sin y )dx ( x cos y )dy is independent of path.
2x

Solution
In this case, we have

60

(e 2 x x sin y ) x cos y
y

Nx
( x 2 cos y ) 2 x cos y
x

My

and

so that My Nx . By theorem (4) the line integral is thus

not independent of path.
Greens Theorem
In this section , we develop a connection between certain
line integrals around a closed curve in the plane and
double integrals over the region enclosed by the curve.
Greens theorem is a significant result with far-reaching
implications. It is of fundamental importance in the
analysis of fluid flows and in the theories of electricity
and magnetism.
Recall that for a plane curve C defined parametrically
by
C ( x, y ) / x g (t ), y h(t ), a t b

C is closed if its two endpoints are the same, i.e.,

(g(a),h(a))=(g(b),h(b)). A curve C is simple if it does not
intersect itself, except at the endpoints. We illustrate a
simple closed curve and a closed curve that is not simple
in the following figure

61

We say that a simple closed curve C has positive

orientation if the region R enclosed by C stays to the left
of C, as the curve is traversed; a curve has negative
orientation if the region R stays to the right of C. In the
following figures we illustrate a simple closed curve with
positive orientation and one with negative orientation ,
respectively

We use the notation F ( x, y ).dr to denote a line integral

along a simple closed curve C oriented in the positive
direction. We can now state the main result of the
section
C

62

Theorem (1) ( Greens Theorem)

Let C be a piecewise-smooth , simple closed curve in the
plane with positive orientation and let R be the region
enclosed by C, together with C. Suppose that M(x,y) and
N(x,y) are continuous and have continuous first partial
derivatives in some open region D, with R D . Then
N
M

x
y

M ( x, y )dx N ( x, y )dy

dA

Proof
Here, we assume that the region R can be written in the
form
R ( x, y ) / a x b, and g1 ( x) y g 2 ( x)
where g1(x) g2(x) , for all x in [a,b], g1(a)=g2(a) and
g1(b)=g2(b), as illustrated in the figure below

63

Notice that we can divide C into two pieces as

indicated in the last figure
C C1 C 2

where C1 is the bottom portion of the curve, defined by

C1 ( x, y ) / a x b, y g1 ( x)

and C2 is the top portion of the curve, defined by

C 2 ( x, y ) / a x b, y g 2 ( x)

where the orientation is as indicated in the figure. From

the evaluation theorem for line integrals , we then have
M ( x, y)dx M ( x, y )dx M ( x, y )dx

C1

C2

M ( x, g1 ( x))dx M ( x, g 2 ( x))dx

Where the minus

M ( x, g1 ( x)) M ( x, g 2 ( x)) dx

(1)

sign in front of the second integral comes from our

traversing C2 backward ( i.e., from right to left). On
the other hand, notice that we can write
b g ( x)

2
M
M
dA

R y
a g ( x) y dydx
1

y g 2 ( x)

M ( x, y ) y g
a

1 ( x)

dx

the

fundamnetal

theorem

of

calculus

M ( x, g 2 ( x )) M ) x, g1 ( x)) dx
a

Together with equation (1) , this gives us

M ( x, y)dx

M
dA
y

( 2)

64

We now assume that we can also write the region R in

form
R ( x, y ) / c y d , and h1 ( y ) x h2 ( y )
Where h1(y) h2(y) for all y in [ c,d], h1( c)=h2( c) and
h1(d)=h2(d)
Here, we write C C C , as illustrated in the following
figure
3

In this case , notice that we can write

N ( x, y)dy N ( x, y)dy N ( x, y)dy

C3

C4

N (h1 ( y ), y )dy N (h2 ( y ), y )dy

d

N (h2 ( y ), y )dy N (h1 ( y ), y )dy

Where the
(3)

minus sign in front of the first integral accounts for our

traversing C3 backward ( in this case from top to
bottom). Further, notice that

65

d h ( y)

2
N
N
dA

R x
c h ( y ) dx dxdy
1

N ( h2 ( y ), y ) N (h1 ( y ), y )dy
c

Together with (3) , this will give us

N

N ( x, y )dy x dA

( 4)

Adding together equations (2) and (4) , we have

N
M

x
y

M ( x, y )dx N ( x, y )dy

dA

Example
Use Greens theorem to rewrite and evaluate
( x y )dx 3xy dy , where C consists of the portion of y=x2
from (2,4) to (0,0), followed by the line segments from (0,0)
to (2,0) and from (2,0) to (2,4).
2

Solution
We indicate the curve C and the enclosed region R in the
figure below

66

Notice that C is a piecewise-smooth , simple closed curve

with positive orientation. Further, for M(x,y)=x 2+y3 and
N(x,y)=3xy2, M and N are continuous and have continuous
first partial derivatives everywhere. Greens theorem then
says that
(x

N M
dA

y
x

y 3 ) dx 3 xy 2 dy
R

(3 y 2 3 y 2 ) dA 0
R

Notice , in this example , since the integrand of the double

integral was zero, evaluating the double integral was far
easier than evaluating the line integral directly. There is
another simple way of thinking of the line integral in the
same example. Notice that we can write this as F ( x, y ).dr
where F ( x, y) x y ,3xy . Notice further that F is
1
conservative [ with potential function f ( x, y) 3 x xy and so,
by theorem (3) in line integral section states that the line
integral of F over any piecewise-smooth, closed curve must
be zero.
C

Example
Evaluate the line integral (7 y e )dx [15 x sin( y 8 y )]dy ,
where C is the circle of radius 3 centered at the point (5,7), as shown in the figure below
sin x

67

Solution
First, notice that it will be virtually impossible to evaluate
the line integral directly. However , taking M(x,y)=7y-esinx
and N(x,y)=15x-sin(y3+8y), notice that M and N are
continuous and have continuous first partial derivatives
everywhere. So, we may apply Greens theorem, which
gives us:
(7 y e

sin x

)dx [15 x sin( y 3 8 y )]dy

N M

x
y

dA

(15 7)dA
R

8 dA 72
R

Where

dA
R

dA (3)

is simply the area inside the region R,

The example above may make you suspect that the

integrand was chosen carefully so that the line integral was
impossible to evaluate directly, but so that the integrand of
the double integral was trivial. The problem was cooked up
68

simply to illustrate the power of Greens theorem. More

significantly, Greens theorem provides us with a wealth of
interesting observations. One of these is as follows.
Suppose that C is a piecewise-smooth, simple closed curve
enclosing the region R . Then, taking M(x,y)=0 and
N(x,y)=x, we have
N M

x
y

xdy

dA

dA
R

Which is simply the area of the region R. Alternatively,

notice that if we take M(x,y) =-y and N(x,y)=0, we have
N M

x
y

ydx

dA

dA
R

Which again the area of R. Putting these two results

together, we also have
1

dA 2
C

xdy ydx

(5)

Example ( finding area )

Find the area enclosed by the ellipse
x2 y2

1
a2 b2

Solution
First, observe that the ellipse corresponds to the simple
closed curve C defined parametrically by:
C ( x, y ) / x cos t ,

y sin t , 0 t 2

Where a,b>0. you should also observe that C is smooth

and positively oriented ( see the figure )
69

From equation (5) , we have that the area A of the ellipse

is given by
A

1
1
xdy ydx [(a cos t )(b cos t ) (b sin t )( a sin t )]dt

C
2
2 0

1
2

(ab cos

t ab sin 2 t )dt ab

Example ( evaluation of line integral)

Evaluate the line integral (e 6 xy)dx (8 x sin y )dy , where
C is the positively-oriented boundary of the region bounded
by the circles of radius 1 and 3 , centered at the origin and
lying in the first quadrant, as indicated in the figure below
x

70

Solution
Since C consists of four distinct pieces, evaluating the line
integral directly by parameterizing the curve is probably
not a good choice. On the other hand, since C is a
piecewise-smooth, simple closed curve, we have by Greens
theorem

(e x 6 xy ) dx (8 x 2 sin y 2 )dy

[ x (8 x

sin y 2 )

(e x 6 xy )]dA
y

(16 x 6 x)dA 10 xdA

R

Where R is the region between two circles and lying in the

first quadrant. Notice that this is easy to compute using
polar coordinate, as follows:

71

(e
C

6 xy ) dx (8 x 2 sin y 2 ) dy

10 x
R

r cos

dA

rdrd

/2 3

(10r cos )rdrd

0 1

/2

10r 3
cos
3

r 3

d
r 1

10 3 3
260
/2
(3 1 )sin 0
3
3

For simplicity, we often will use the notation R to refer to

the boundary of the region R, oriented in the positive
direction. Using this notation, the conclusion of Greens
theorem is written as

M ( x, y ) dx N ( x, y ) dy

N M

x
y

dA

Greens theorem can be extended to the case where a

region is not simply-connected ( i.e., where the region has
one or more holes). We must emphasize that when dealing
with such regions, the integration is taken over the entire
boundary of the region ( not just the outer-most portion of
the boundary) and that the boundary curve is traversed is
the positive direction, always keeping the region to the left.
For instance, for the region R illustrated in the following
figure with a single hole

72

Notice that the boundary of R, R , consists of two separate

curves , C1 and C2 where C2 is traversed clockwise, in
order to keep the orientation positive on all of the
boundary. Since the region is not simply-connected, we
may not apply Greens theorem directly. Rather, we first
make two horizontal slits in the region, as indicated in the
following figure

This operation will divide the region R into two simplyconnected regions R1 and R2 . We can then apply Greens
theorem in each of R1 and R2 separately. Adding the double
integrals over R1 and R2 gives us the double integral over
all R. We have
73

N M

x
y

dA

N M

x
y

R1

R1

N M

x
y
R2

dA

M ( x, y ) dx N ( x, y ) dy

R2

dA

M ( x, y ) dx N ( x, y ) dy

Further, since the line integral over the common portions

of R and R ( i.e., the slits) are traversed in the opposite
direction( one way on R and the other on R , the line
integrals over these portions will cancel out, leaving only
the line integrals over C1 and C2 .
This gives us :
1

N M

x
y

dA

R1

C1
C

M ( x, y ) dx N ( x, y ) dy

R2

M ( x, y ) dx N ( x, y ) dy

M ( x, y ) dx N ( x, y ) dy M ( x, y ) dx N ( x, y ) dy
C2

M ( x, y ) dx N ( x, y ) dy

This says that Greens theorem also holds for regions with
a single hole. This case can be extended to regions with
any finite number of holes.
Curl and Divergence
Greens theorem relates the line integral of a function over
the boundary of a plane region R to the double integral of a
related function over the region R. In some cases, the line
integral is easier to evaluate, while in other cases, the
double integral is easier. More significantly, Greens
theorem provides us with a connection between physical
quantities measured on the boundary of a plane region with
related quantities in the interior of the region. The goal of
this section is to extend Greens theorem to results that
relate triple integrals, double integrals , and line integrals.
The first step is to understand the vector operations of curl
and divergence . Both curl and divergence are
74

generalization of the notion of derivative that are applied

to vector fields. Both directly measure important physical
quantities related to vector field F(x,y,z).
Definition(1)
The curl of the vector field
F ( x, y , z ) F1 ( x, y , z ), F2 ( x, y , z ), F3 ( x, y , z )

Curl

is the vector field

F3 F2
F2 F1
F1 F3
i
k
F

j
z
x
y
z
y
x

Defined at all points at which the indicated partial

derivatives exist.
An easy way to remember curl F is to use cross product
notation, as follows. Notice that using a determinant, we
can write
i

xF
x
F1

y
F2

z
F3

whenever

F3 F2
F 2 F1
F3 F1
i
k

j
z
z
y
x
y
x

F3 F2 F3 F1 F 2 F1

curl
y
z x
z x
y

(1)

all of the identical partial derivatives are defined.

Example
Compute curl F for :
(a) F ( x, y, z ) x y,3x yz, z
(b) F ( x, y, z ) x y, y , e
2

Solution
Using the cross product notation in (1) , we have that for
(a):
75

Curl

F xF
x
F1

y
F2

k
i

z
x
F3
x2 y

y
3 x yz

z
z3

( z 3 ) (3 x yz )
( z 3 ) ( x 2 y)
(3 x yz ) ( x 2 y )
i
j
k

z
z
x
dy
y

(0 y )i (0 0) j (3 x 2 )k y ,0,3 x 2

Similarly, for part (b), we have

Curl

F xF
x
F1

y
F2

k
i

z
x
F3
x3 y

y
y5

z
ez

(e z ) ( y 5 )
(e z ) ( x 3 y )
( y 5 ) ( x 3 y)
i
j
k

z
z
y
y
x

(0 0)i (0 0) j (0 1) k 0,0,1

Notice that in part (b) , that the only term that contributes to
the curl is the term (-y) in the i-component of F(x,y,z). This
illustrates an important property of the curl. Terms in icomponent of the vector field involving only x will not
contribute to the curl, nor will terms in the j-component
involving only y nor terms in k-component involving only
z. These observations are useful to simplify some
calculations. For instance, notice that
curl x 3 , sin 2 y ,

z 2 1 x2

curl 0,0, x 2
x 0,0, x 2 0,2 x,0

The simplification discussed above gives an important hint

about what the curl measures, since the variables must get
mixed up to produce a nonzero curl.

76

The following example provides a clue as to the meaning of

the curl of a vector field.
Example
Compute the curl of :
(a) F ( x, y, z ) xi yj
(b) G ( x, y, z ) yi xj
Solution
(a) we have
i

xF
x
F1

y
F2

k
i

z
x
F3
x

y
y

z
0

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

(b) we have
i

xG
x
F1

y
F2

k
i

z
x
F3
y

y
x

z
0

0 0,(0 0),1 1 0,0,2

Graph of the vector fields F and G in two dimensions are

shown in the following figures

77

78

It is helpful to think of each of these vector fields as the

velocity field for a fluid in motion across the xy-plane .
In this case, the vectors indicated in the graph of the
velocity field indicate the direction of flow of the fluid.
For the vector field x, y,0 , observe that the fluid flows
directly away from the origin, so that the fluid has no
rotation and in (a) we found that curl F=0 . By contrast,
the vector field y, x,0 indicates a clockwise rotation of
the fluid, while in (b) we computed a nonzero curl. In
particular, notice that if you curl the fingers of your right
hand so that your fingertips point in the direction of the
flow, your thumb will point into the page , in the
direction of -k, which has the same direction as
Curl y , x,0 x y , x,0 2k

We noted earlier that there is no contribution to the curl

of a vector field F(x,y,z) from terms in the i-component of
F that involve only x, nor from terms in the j-component
of F involving only y nor terms in the k-component of F
involving only z. By contrast, these terms make
important contributes to the divergence of a vector field,
the other major vector operation introduced in this
section.
Definition (2)
The divergence of the vector field
F ( x, y , z ) F1 ( x, y , z ), F2 ( x, y , z ), F3 ( x, y , z )

is the scalar function

:
F1 F2 F3

x
y
z

,
Defined at all points at which all the indicated partial
derivatives exist.
divF ( x, y, z )

79

While we wrote the curl using cross product notation,

note that we can write the divergence of a vector field
using dot product notation, as follows:
.F

F1 F2 F3

,
,
. F1 , F2 , F3

divF ( x, y , z )
x y z
x
y
z

Example
Compute divF for
(a) F ( x, y, z ) x y,3x yz, z
(b) F ( x, y, z ) x y, z , e
2

( 2)

Solution
(a) we have from equation (2) that
divF .F

( x 2 y ) (3 x yz ) ( z 3 )

2 xy z 3 z 2
x
y
z

(b) again from equation (2) we have that

divF .F

( x 3 y ) ( z 5 ) (e y )

3x 2 0 0 3x 2
x
y
z

In part (b) the only term contributing to the divergence is x3

term in the i-component of F. Further, observe that in
general , the divergence F(x,y,z) is not affected by terms in
the i-component of F that do not involve x, terms in the jcomponent of F that do not involve y terms in the kcomponent that do not involve z. Returning to the two
-dimensional vector field in the previous example , we can
develop a graphical interpretation of the divergence.
Example
Compute the divergence of
(a) F ( x, y) xi yj
(b) F ( x, y ) yj xi
80

Solution
(a) we have
(b) we have

( x) ( y )

2
x
y
y ( x)
.F

0
x
y
.F

Graphs of vector fields in (a) and (b) are shown in the

following figures respectively

Notice the boxes that we have superimposed on the graph

of each vector field. If F(x,y) represents the velocity field of
a fluid in motion, try to use the graph to estimate the net
flow of fluid into or out of each box. For y, x , the fluid
is rotating in circular paths, so that the velocity of any
particle on a given circle centered at the origin is a
constant. This suggests that the flow into the box should
equal the flow out of the box and the net flow is zero. For
the vector field x, y notice that the arrows coming into the
box are shorter than those exiting the box . This says that
the net flow out of the box is positive(i.e., there is more

81

fluid exiting the box than entering the box) in this case the
divergence is positive.
Example
If f(x,y,z) is a scalar function and F(x,y,z) is a vector field,
determine whether each operation is a scalar function, a
vector field or undefined :
(a) x(f )
(b) x(.F )
(c ) .(f )
Solution
Examine each of these expressions one step at a time,
working from the inside out.
For (a) , f is a vector field , so the curl of f is defined
and gives a vector field . In (b) .F is a scalar function, so
the curl of .F is undefined. In ( c) , f is a vector field, so
the divergence of f is defined and gives f a scalar
function.
Theorem (1)
Suppose that F ( x, y, z ) F ( x, y, z ), F ( x, y, z ), F ( x, y, z ) is a vector
field whose components F1,F2,and F3 have continuous firstorder partial derivatives throughout an open region D R .
If F is conservative, then xF 0
1

Example
Use the theorem above to determine whether the following
fields are conservative:
82

(a) F
(b) F

cos x z , y 2 , xz
2 xz ,3 z 2 , x 2 6 yz

Solution
(a) we have
i

xF
x
cos x z

y
y2

0 0,1 z ,0 0 0
z
xz

So by theorem (1) above F is not conservative.

(b) we have
i

xF
x
2 xz

y
3z 2

6 z 6 z ,2 x 2 x,0 0,1 z ,0 0 0
z
x 2 6 yz

Notice that in this case theorem (1) does not tell us

whether or not F is conservative . However , we can
notice that
F ( x, y , z ) 2 xz ,3 z , x 6 yz
since we have found a potential
function for F , we now see that it is indeed a
conservative field.
2

Theorem (2)
Suppose that F ( x, y, z ) F ( x, y, z ), F ( x, y, z ), F ( x, y, z ) is a vector
field whose components F1,F2,and F3 have continuous firstorder partial derivatives throughout all of R3. Then F is
conservative, if and only if xF 0 .
1

Notice that half of this theorem is already known from

theorem(1). Also , notice that we required that the
83

components of F have continuous first partial derivatives

throughout all of R . The other half of the theorem requires
the additional theorem ( Stokes theorem).
3

Summary
We can summarize a number of equivalent properties for
three-dimensional vector fields. Suppose that
F ( x, y , z ) F ( x, y , z ), F ( x, y , z ), F ( x, y , z ) is a vector field whose
components F , F , and F have continuous first partial
derivatives throughout all R . Then the following are
equivalent:
1. F ( x, y, z ) is conservative
2. F. dr is independent of path.
3. F . dr 0 for every piecewise-smooth closed curve C .
4. xF 0
5. F ( x, y, z ) is a gradient field ( F f for some potential function f )
1

We end this section by rewriting Greens theorem in

terms of the curl and divergence.
First suppose that F ( x, y ) M ( x, y ), N ( x, y ) is a vector
field, for some functions M ( x, y ) and N ( x, y ) . Suppose that
R is a region in the xy plane whose boundary curve C is
piecewise-smooth, positively oriented , simple and
closed and that M and N are continuous and have
continuous first partial derivatives in some open region
D , where R D . Then, from Greens theorem, we have
N M

x
y

dA

dx N dy

84

Notice that the integrand of the double integral, x y

, is the k component of F .
Further, since dz 0 on any curve lying in the xy plane ,
we have M dx N dy F . dr
Thus, we can write Greens theorem in the form
C

F . dr ( F ). k dA

To take Greens theorem in another direction, suppose

that F and R are as just defined and suppose that C is
traced out by the endpoint of the vector-valued function
r (t ) x (t ), y (t ) , for a t b , where x (t ) and y (t ) have
continuous first partial derivatives for a t b . Recall that
the unit tangent vector to the curve is given by:
T (t )

x (t )
y (t )
,

r (t )
r (t )

Its then easy to verify that the exterior unit normal

vector to C at any point ( i.e., the unit normal vector
y (t ) x(t )
that points out of R ) is given by n(t ) r (t ) , r (t )
See the figure below:

85

Now, from theorem (2), we have

b

F .nds ( F .n)(t ) r (t ) dt

M ( x(t ), y (t )) y (t ) N ( x (t ), y (t )) x (t )

r (t ) dt
r (t )
r (t )
a

M ( x(t ), y (t )) y (t )dt N ( x(t ), y(t )) x(t )dt

a

M ( x, y)dy N ( x, y)dx

M N

x
y

dA

From Greens theorem. Finally, recognize that the

integrand of the double integral is the divergence of
and this gives us another vector form of Greens
theorem: F . n ds .F ( x, y ) dA
C

This form of Greens theorem is generalized to the

divergence.

86