Green’s Theorem

1. Let C be the boundary of the unit square 0 ≤ x ≤Z 1, 0 ≤ y ≤ 1, oriented counterclockwise, and let F~ be
the vector field F~ (x, y) = hey + x, x2 − yi. Find F~ · d~r.
C

Solution. Let’s write P (x, y) = e + x and Q(x, y) = x2 − y, so that F~ = hP, Qi. Let R be the
y

region 0 ≤ x ≤ 1, 0 ≤ y ≤ 1. The boundary of R, oriented “correctly” (so that a penguin Z walking

along it keeps R on his left), is the given curve C. So, Green’s Theorem says that F~ · d~r =
ZZ ZZ C

(Qx − Py ) dA = (2x − ey ) dA. We compute this by converting it to an iterated integral:

R R
Z ZZ
F~ · d~r = (2x − ey ) dA
C R
Z 1Z 1
= (2x − ey ) dx dy
0 0
Z 1 x=1 
= x2 − xey  dy

0 x=0
Z 1
= (1 − ey ) dy
0
y=1
= y − ey 

y=0

= 2−e

2. Let C be the oriented curve Z consisting of line segments from (0, 0) to (2, 3) to (2, 0) back to (0, 0), and
let F~ (x, y) = hy , x i. Find
2 2
F~ · d~r.
C

Solution. Here is a picture of the curve C, along with the interior of the triangle, which we’ll call R:
3

1 2

The boundary of R, oriented “correctly” (so that a penguin walking along it keeps R on Zhis left side),
is −C (that is, it’s C with the opposite orientation). So, Green’s Theorem says that F~ · d~r =
ZZ Z −C

(Qx − Py ) dA, where F = hP, Qi. We are looking for F~ · d~r, which we know is the negative of
R C

1
 Z
F~ · d~r. Therefore,
−C
Z ZZ
F~ · d~r = − (Qx − Py ) dA
C R
ZZ
= − (2x − 2y) dA
R
Z 2 Z 3x/2
= − (2x − 2y) dy dx
0 0
Z 2 y=3x/2 
= − 2xy − y 2  dx

0 y=0
Z 2
3 2
= − x dx
0 4
x=2 !
1 3 
= − x
4 x=0
= −2

3. Find the area of the region enclosed by the parameterized curve ~r(t) = ht − t2 , t − t3 i, 0 ≤ t ≤ 1.
0.4

0.2

0
0 0.2

Solution. Let R be theZ Zregion in question. We know from #2(a) on the worksheet “Double Integrals”
that the area of R is 1 dA. Normally, we would evaluate this by converting it to an iterated
R
integral; in this case, that’s quite challenging, so we’ll instead use Green’s Theorem to evaluate this
integral. If we can come up with a vector
ZZ field F~ (x,
Z y) = hP (x, y), Q(x, y)i such that Qx − Py = 1,
then Green’s Theorem will say that 1 dA = F~ · d~r, where C is the boundary of the region,
R C
traveled counterclockwise (so that a penguin walking along C keeps R on his left). One such vector
field is F~ (x, y) = h0, xi.
We are given a parameterization ~r(t) of the curve, and this parameterization does in fact travel the

2
 curve counterclockwise.(1) So,
ZZ Z
1 dA = F~ · d~r
R C
Z 1
= F~ (~r(t)) · ~r0 (t) dt
0
Z 1
= h0, t − t2 i · h1 − 2t, 1 − 3t2 i dt
0
Z 1
= (t − t2 )(1 − 3t2 ) dt
0
Z 1
= (t − t2 − 3t3 + 3t4 ) dt
0
t=1
1 2 1 3 3 4 3 5 
= t − t − t + t 
2 3 4 5 t=0
1
=
60

4. Let F~ (x, y) = hP (x, y), Q(x, y)i be any vector field defined on the region R (in R2 ) shown in the picture,
and let C1 and
Z Z C2 be the oriented
ZZ curves shown in the picture. What does Green’s Theorem say about
F~ · d~r, F~ · d~r, and (Qx − Py ) dA?
C1 C2 R

C1

C2 R

Solution. The boundary of R consists of two curves, C1 and C2 . A penguin walking along C1 in the
indicated direction would indeed keep R on his left, but a penguin walking along C2 in the indicated
direction would have R on his right. So, the boundary of R is really C1 together with −C2 , which
ZZ Z Z
means (Qx − Py ) dA = F~ · d~r − F~ · d~r .
R C1 C2

* +
x y
5. Let F~ (x, y) = hP (x, y), Q(x, y)i = p ,p . You can check that Py = Qx .
x2 + y 2 x2 + y 2

(a) What is wrong with the following reasoning? “Py = Qx , so F~ is conservative.”

Solution. F~ is not defined at the origin, so its domain is R2 except the point (0, 0). This domain
is not simply connected, so we cannot conclude anything from the fact that Py = Qx .

(b) Let C be any simple closed curve in R2 that does not enclose the origin, oriented counterclockwise.
(1) This is not completely obvious, but there’s an easy way to tell at the end whether the parameterization went the right way

— we are looking for an area, so our final answer must be positive.

3
 (A
Z simple curve is a curve that does not cross itself.) Use Green’s Theorem to explain why
F~ · d~r = 0.
C

Solution. Since C does not go around the origin, F~ is defined on the interior R of C. (The
only point where F~ is not defined
Z is the Zorigin,
Z but that’s not in R.) Therefore, we can use
Green’s Theorem, which says F~ · d~r = (Qx − Py ) dA. Since Qx − Py = 0, this says that
Z C R

F~ · d~r = 0.
C

(c) Let a be a positive constant, and let C be the circle x2 + y 2 = a2 , oriented counterclockwise.
Parameterize C (check your parameterization by plugging
Z it into the equation x2 + y 2 = a2 ), and
use the definition of the line integral to show that F~ · d~r = 0. (Why doesn’t the reasoning from
C
(b) work in this case?)
Solution. One possible parameterization of C is ~r(t) = ha cos t, a sin ti, 0 ≤ t ≤ 2π. Then,
Z Z 2π
F~ · d~r = F~ (~r(t)) · r~0 (t) dt
C 0
Z 2π * +
a cos t a sin t
= p ,p · h−a sin t, a cos ti dt
0 (a cos t)2 + (a sin t)2 (a cos t)2 + (a sin t)2
Z 2π
= 0 dt
0
= 0,

as we wanted.
We cannot use the reasoning from (b) since F~ is not defined in the whole interior of C (in
particular, it’s not defined at the origin, which is inside C).

(d) Let C be anyZsimple closed curve in R2 that does enclose the origin, oriented counterclockwise.
Explain why F~ · d~r = 0. (Hint: Use (c) and #4.)
C

Solution. No matter what C looks like, we can draw a giant circle x2 + y 2 = a2 around the origin
that encloses all of C. Let’s orient this giant circle counterclockwise and call it C 0 , and let’s have
R be the region between C and C 0 R:

C¢ R

Notice that F~ is defined on all of R (because it is defined everywhere except the origin, and R

4
 doesn’t include the origin). So, #4 tells us that
ZZ Z Z
(Qx − Py ) dA = F~ · d~r − F~ · d~r.
R C0 C
Z
We showed in (c) that F~ · d~r = 0, so this simplifies to
C0
ZZ Z
(Qx − Py ) dA = − F~ · d~r.
R C

Since Qx = Py inside of R, the

Z double integral is really a double integral of 0, so it’s equal to 0.
Therefore, we conclude that F~ · d~r = 0 as well.
C

(e) Is it valid to conclude from the above reasoning that, if F~ (x, y) = hP (x, y), Q(x, y)i is a vector
field defined everywhere except the origin and Py = Qx , then F~ is conservative?

Solution. No! The ~

* calculation in (c) only
+ applied to this particular vector field F (x, y) =
x y
hP (x, y), Q(x, y)i = p ,p .
2
x +y 2 x + y2
2

There are vector fields that are defined everywhere except the origin and satisfy Py = Qx but are
still not conservative; the vector field in #4(b) of the worksheet “The Fundamental Theorem for
Line Integrals; Gradient Vector Fields” is an example.

6. In this problem, you’ll prove Green’s Theorem in the case where the region is a rectangle. Let F~ (x, y) =
hP (x, y), Q(x, y)i be a vector field on the rectangle R = [a, b] × [c, d] in R2 .
ZZ Z d Z b
(a) Show that [Qx (x, y) − Py (x, y)] dA = [Q(b, y) − Q(a, y)] dy − [P (x, d) − P (x, c)] dx.
R c a

Solution. Let’s first break the given double integral into a difference of two double integrals:
ZZ ZZ ZZ
[Qx (x, y) − Py (x, y)] dA = Qx (x, y) dA − Py (x, y) dA.
R R R

Now, we’ll convert the double integrals on the right side to iterated integrals. This is easy, since
the region R is just a rectangle. However, we’re going to do the two iterated integrals in different
orders: it makes sense to first integrate Qx with respect to x (since it’s a derivative with respect
to x) and to first integrate Py with respect to y:
ZZ Z d Z b Z b Z d
[Qx (x, y) − Py (x, y)] dA = Qx (x, y) dx dy − Py (x, y) dy dx.
R c a a c

When we integrate Qx with respect to x, we just get Q; similarly, when we integrate Py with
respect to y, we just get P :
ZZ Z d x=b  Z b y=d 
[Qx (x, y) − Py (x, y)] dA = Q(x, y) dy − P (x, y) dx
 
R c x=a a y=c
Z d Z b
= [Q(b, y) − Q(a, y)] dy − [P (x, d) − P (x, c)] dx,
c a

which is exactly what we were asked to show.

5
 Z
(b) Let C be the boundary of R, traversed counterclockwise. Show that F~ · d~r is also equal to
Z d Z b C

[Q(b, y) − Q(a, y)] dy − [P (x, d) − P (x, c)] dx.

c a

Solution. Here is a picture of C:

y

c
x
a b

As we can see, it’s composed of 4 pieces, and we’ll parameterize each separately. The bottom piece
has y = c, so only x varies, and we can parameterize it using ~r1 (t) = ht, ci with a ≤ t ≤ b. The
right piece has x = b, so only y varies, and we can parameterize it using ~r2 (t) = hb, ti, c ≤ t ≤ d.
The top piece has y = d, so only x varies, and we’d like to parameterize it using ~r3 (t) = ht, di.
The slight problem with this is that it goes the wrong direction: as t increases, ht, di goes to the
right. This is actually not a problem, as long as we account for it later. So, we’ll go ahead and
use ~r3 (t) = ht, di with a ≤ t ≤ b. Similarly, for the left piece, we’ll use ~r4 (t) = ha, ti, c ≤ t ≤ d.
Here’s a diagram showing the various things we’ve parameterized:
y
Ó
r3 HtL
d
Ó Ó
r4 HtL r2 HtL

c Ó
r1 HtL
x
a b

As we can see from the two diagrams,

Z Z Z Z Z
~
F · d~r = ~
F · d~r + F~ · d~r − F~ · d~r − F~ · d~r.
C ~
r1 (t) ~
r2 (t) ~
r3 (t) ~
r4 (t)
Z
Plugging the four parameterizations into this, F~ · d~r is equal to
C
Z b Z d Z b Z d
F~ (t, c) · h1, 0i dt + F~ (b, t) · h0, 1i dt − F~ (t, d) · h1, 0i dt − F~ (a, t) · h0, 1i dt.
a c a c

Writing F~ (x, y) = hP (x, y), Q(x, y)i, we can simplify this to

Z Z b Z d Z b Z d
F~ · d~r = P (t, c) dt + Q(b, t) dt − P (t, d) dt − Q(a, t) dt.
C a c a c

This is exactly what we were supposed to show, which is more obvious if we rename t to be x in
the first and third integrals, rename t to be y in the second and fourth integrals, and rearrange

6
the terms:
Z Z b Z d Z b Z d
F~ · d~r = P (x, c) dx + Q(b, y) dy − P (x, d) dx − Q(a, y) dy
C a c a c
Z d Z d Z b Z b
= Q(b, y) dy − Q(a, y) dy − P (x, d) dx + P (x, c) dx
c c a a
Z d Z b
= [Q(b, y) − Q(a, y)] dy − [P (x, d) − P (x, c)] dx
c a