Part 1: Double Integrals and Boundary Curves
When a vector field F( x,y) =
á M(x,y) ,N( x,y)
ñ is conservative, then it has
a potential and line integrals of the form òABF·dr are independent of the path C which connects A to B. When
a vector field is not conservative, then another form of the
fundamental theorem called Green's theorem is used to reveal
properties of the field.
To begin with, a region R is simply connected if its boundary is a
simple closed curve. Moreover, we denote its boundary curve by ¶R
and we define the orientation of ¶R to be such that the region R
is always to the left of a tangent vector to ¶R.
For example, if f( x) > g( x) for all x in (a,b) and if f( a) = g( a) and f(b) = g( b) , then y = f( x) and y = g(x) enclose a simply connected region R whose boundary ¶R
is made up of two curves, C1 and C2.
The curve C1 is parametrized
by r( t) =
á -t,f( -t)
ñ for t in [ -b,-a] so that it starts at ( b,f( b) ) and ends at ( a,f( a)).The
double integral of the derivative My with respect to y
of a function M(x,y) over R is given by

R |
MydA = |
ó õ
|
b
a
|
|
ó õ
|
f( x)
g( x)
|
My( x,y) dydx |
|
Since My is a derivative with respect to y, we must have

R |
MydA |
|
|
|
|
ó õ
|
b
a
|
M( x,y) |
|
y = f( x)
y = g( x) |
dx |
| |
|
|
|
ó õ
|
b
a
|
M( x,f( x) ) -M( x,g(x) ) dx |
| |
|
|
ó õ
|
b
a
|
M( x,f( x) ) dx- |
ó õ
|
b
a
|
M(x,g( x) ) dx |
|
|
If we let x = -t, dx = -dt, t( a) = -a, t(b) = -b in the first integral, then

R |
MydA = - |
ó õ
|
-b
-a
|
M( -t,f( -t) )dt - |
ó õ
|
b
a
|
M( x,g( x) ) dx |
|
However, if x = -t, then dx/dt= -1. Likewise, in the second
integral we let x = t, so that dx/dt = 1 and

R |
MydA |
|
|
|
- |
ó õ
|
-a
-b
|
M( -t,f( -t) ) |
dx
dt
|
dt - |
ó õ
|
b
a
|
M( t,g( t) ) |
dx
dt
|
dt |
| |
|
|
- |
ó õ
|
C1
|
M( x,y) dx- |
ó õ
|
C2
|
M( x,y) dx |
| |
|
|
- |
é ë
|
|
ó õ
|
C1
|
M( x,y) dx+ |
ó õ
|
C2
|
M( x,y)dx |
ù û
|
|
| |
|
| - |  |
¶R
|
M(x,y) dx |
|
|
where
denotes a line integral around a closed curve with the
natural orientation. That is, we have shown that

R |
MydA = - | ¶R | M( x,y) dx |
| (1) |
Thus, a line integral on the boundary of R can be used to evaluate a
double integral over R.
EXAMPLE 1 Use (1) to evaluate
Solution: First we convert into a double integral over the unit
disk D:
|
|
ó õ
|
1
-1
|
|
ó õ
|
|
3xy2dydx = |
óó
õõ |
D |
3xy2 dA | |
|
|
Identifying My = 3xy2 implies that M = òMydy = ò3xy2dy = xy3. Thus, (1) says that
óó
õõ |
D |
3xy2dA = - |  |
¶D
|
xy3dx |
|
Moreover, r( t) =
á cos( t),sin( t)
ñ , t in [ 0,2p] , is a
parametrization of ¶D, so that
|
|  |
¶D
|
xy3dx = |
ó õ
|
2p
0
|
xy3 |
dx
dt
|
dt = |
ó õ
|
2p
0
|
cos( t) sin3( t) [-sin( t) ] dt |
|
If u = sin( t) , then du = cos( t) dt and u(0) = u( 2p) = 0. Thus,
óó
õõ |
D |
3xy2dA = - |  |
¶D
|
xy3dx = |
ó õ
|
2p
0
|
sin4( t) cos( t) dt = - |
ó õ
|
0
0
|
u4du = 0 |
|
Check your Reading: Why is u( 0) = u(2p) = 0 in example 1?