1 Green’s Theorem

Green’s theorem states that a line integral around the boundary of a plane region D can be computed
as a double integral over D. More precisely, if D is a “nice” region in the plane and C is the boundary
of D with C oriented so that D is always on the left-hand side as one goes around C (this is the
positive orientation of C), then
Z ZZ
€ ∂Q ∂P 
P dx + Qdy = − dxdy
C D ∂x ∂y

if the partial derivatives of P and Q are continuous on D.

Regions that are simultaneously of type I and II are “nice” regions, i.e., Green’s theorem is true
for such regions. The next two propositions prove this.

Theorem 1. If D is a region of type I then

Z ZZ
∂P
P dx = − dxdy.
C D ∂y

Proof. If D = {(x, y) | a ≤ x ≤ b, f (x) ≤ y ≤ g(x)} with f (x), g(x) continuous on a ≤ x ≤ b, we

have C = C1 + C2 + C3 + C4 , where C1 , C2 , C3 , C4 are as shown below.

Since the region is of type I, we have

ZZ Z b‚
∂P ƒ
dxdy = P (x, g(x)) − P (x, f (x)) dx.
D ∂y a
Using the standard parametrizations of C1 and C3 , we have
Z b Z Z b Z
P (x, f (x)) dx = P dx, P (x, g(x)) dx = − P dx
a C1 a C3

We thus obtain ZZ Z Z Z
∂P
dxdy = P dx + P dx = P dx
D ∂y C1 C3 C

since the line integral of P dx is zero on C2 and C4 as x is constant there.

Theorem 2. If D is a region of type II then

Z ZZ
∂Q
Q dy = dxdy.
C D ∂x

Proof. If D = {(x, y) | h(y) ≤ x ≤ k(y), c ≤ y ≤ d} with h(y), k(y) continuous on c ≤ y ≤ d, we

have C = C1 + C2 + C3 + C4 , where C1 , C2 , C3 , C4 are as shown below.

Since the region is of type II, we have

ZZ Z d‚
∂Q ƒ
dxdy = Q(k(y), y) − Q(h(y), y) dy.
D ∂y c

Using the standard parametrizations of C2 and C4 , we have

Z d Z Z d Z
Q(k(y), y) dy = Q dy, Q(h(y), y) dx = − Q dy
c C2 c C4

We thus obtain ZZ Z Z Z
∂Q
dxdy = Q dy + Q dy = Q dy
D ∂x C2 C4 C

since the line integral of Q dx is zero on C1 and C3 as y is constant there.

Putting these two theorems together, we obtain

Theorem 3. If D is a region of the plane that is simultaneously type I and type II then
Z ZZ
€ ∂Q ∂P 
P dx + Qdy = − dxdy.
C D ∂x ∂y

Green’ Theorem can easily be extended to any region that can be decomposed into a finite
number of regions with are both type I and type II. Such regions we call ”nice”. Fortunately, most
regions are nice. For example, consider the region below.
Since D is the union of D1 ,D2 and D3 , we have
ZZ ZZ ZZ ZZ
= + + .
D D1 D2 D3

Since the regions D1 , D2 , D3 are all type I and type II and the positively oriented boundaries of
D1 , D2 D3 are respectively C1 − E2 − E1 , E1 + C2 , E2 + C2 , we have
ZZ Z Z Z
€ ∂Q ∂P 
− dxdy = P dx + Q dy − P dx + Q dy − P dx + Q dy,
D1 ∂x ∂y C1 E1 E2
ZZ Z Z
€ ∂Q ∂P 
− dxdy = P dx + Q dy + P dx + Q dy,
D2 ∂x ∂y E1 C2
ZZ Z Z
€ ∂Q ∂P 
− dxdy = P dx + Q dy + P dx + Q dy.
D3 ∂x ∂y E2 C2

Adding these equations, we get

ZZ Z Z Z Z
€ ∂Q ∂P 
− dxdy = CP dx + Q dy + P dx + Q dy + P dx + Q dy = P dx + Q dy,
D ∂x ∂y 1 C2 C3 C

where C = C1 + C2 + C3 is the positively oriented boundary of D. This yields Green’s Theorem for
D.
The reader is invited to prove Green’s Theorem for the region below using the given decomposition
into regions which are type I and type II.

2 The flux form of Green’s Theorem

Let R be a region for which Green’s Theorem holds and let C be the positively oriented boundary
−
→ −
→ → −
of R. For each point P on C let T be the unit tangent vector at P and let N = T × ~k, where ~k is
the unit normal to the x, y-plane.
−
→
Theorem 4. If F = P~i + Q~j is a twice continuously differentiable vector field on R, then
Z ZZ
−
→ − → € ∂P ∂Q 
F · N ds = + dxdy.
C R ∂x ∂y
 u · (−
Proof. Since −
→ →
v ×w ~ = (→
−
w ×− u)·−
→ →v , we have
Z Z Z ZZ
−
→ − → ~ → −
− → € ∂P ∂Q 
F · ( T × k) ds = ~
(k × F ) · T ds = −Q dx + P dy = + dxdy.
C C C R ∂x ∂y

This theorem is called the flux form of Green’s Theorem since

Z
−
→ − →
F · N ds
C
−
→ ∂Q
is the flux of F across C. The function ∂P ∂x + ∂y is called the divergence of the vector field
−
→ −
→
F = P~i + Q~j and is denoted by div( F ). For this reason, Theorem 4 is also called the 2-dimensional
∂~ ∂ ~
Divergence Theorem. Note that,if ∇ = ∂x i + ∂y j, we have
→
− −
→
div( F ) = ∇ · F .
→
− € 
∂P ~ →
−
The vector field ∇ × F = ∂Q ∂x − ∂y k is the called the curl of the vector field F and is also denoted
→
−
by curl( F ). The first form of Green’s Theorem can be stated as
Z ZZ
−
→ → − −
→
F · T ds = curl( F ) · ~k dxdy.
C R
These two equivalent forms of Green’s Theorem in the plane give rise to two distinct theorems
in three dimensions. The usual form of Green’s Theorem corresponds to Stokes’ Theorem and the
flux form of Green’s Theorem to Gauss’ Theorem, also called the Divergence Theorem. In Adams’
textbook, in Chapter 9 of the third edition, he first derives the Gauss theorem in §9.3, followed, in
Example 6 of §9.3, by the two dimensional version of it that has here been referred to as the flux
form of Green’s Theorem. He then uses this two dimensional version in §9.4 to derive the usual
form of Green’s Theorem which he uses to prove Stokes’ Theorem, the three dimensional version of
Green’s Theorem. [In this COURSE PAK (see Stokes’ theorem) it is also shown how to deduce
Stokes’ Theorem from Green’s Theorem.]
→
−
Green’s Theorem can be used to give a physical interpretation of the curl in the case F is the
velocity field →
−
v of a flow. If Cr is a circle of radius r with center P , then the average value of the
−
→
angular velocity ωr = − →
v · T /r on Cr is
Z
1 →
− −
→
ωr = 2
v · T ds.
2πr Cr
If Dr is the closed disk with boundary Cr , the average value of curl(−→v ) · ~k on Dr is
ZZ
1 2v T
2
curl(−
→
v ) · ~k dxdy = = 2ωr .
πr Dr r
Taking the limit as r → 0, we find that that the angular velocity of the flow around P is
1
ω = lim ωr = curl(− →v )(P ) · ~k
r→0 2
and hence that curl(→ −v )(P ) = 2ω~k. For this reason, we sometimes denote curl(~v ) by rot(−
→v ). The
−
→
vector field ~v is said to be irrotational if curl( v ) = 0.
Using the flux form of Green’s Theorem we can, in the same way, give a physical interpretation
of div(−
→
v )(P ). The flux of → −v across Cr per unit area of Dr is
Z ZZ
1 −
→ →
− 1
v · T ds = div(−
→v ) dxdy = div(−
→
v )(Qr )
πr2 Cr πr2 Dr

for some point Q in D . Taking the limit as r → 0, we find that div(−

r r
→
v )(P ) measures the rate
of change of the quantity of fluid or gas flowing from P per unit area. For this reason, P since
called a source if div(−
→ v )(P ) < 0. The vector field →
v )(P ) > 0 and a sink if div(→
− −
v is said to be
−
→
incompressible if div( v ) = 0.