Professional Documents
Culture Documents
Integration Theorums PDF
Integration Theorums PDF
Integration Theorums PDF
W W L CHEN
c
This chapter is available free to all individuals, on the understanding that it is not to be used for financial gain,
and may be downloaded and/or photocopied, with or without permission from the author.
However, this document may not be kept on any information storage and retrieval system without permission
from the author, unless such system is not accessible to any individuals other than its owners.
Chapter 12
INTEGRATION THEOREMS
(1)
where the functions 1 : [A1 , B1 ] R and 2 : [A1 , B1 ] R are continuous in the interval [A1 , B1 ] and
where 1 (x) 2 (x) for every x [A1 , B1 ], is called an elementary region of type 1. A region of the
type
R = {(x, y) R2 : y [A2 , B2 ] and 1 (y) x 2 (y)},
(2)
where the functions 1 : [A2 , B2 ] R and 2 : [A2 , B2 ] R are continuous in the interval [A2 , B2 ] and
where 1 (y) 2 (y) for every y [A2 , B2 ], is called an elementary region of type 2. Furthermore, an
elementary region of type 3 is one which is of both type 1 and type 2; in other words, one that can be
described by both (1) and (2).
Greens theorem relates a line integral along a simple closed curve C in R2 to a double integral over
the region R enclosed by the curve. We say that C has positive orientation if the region R is on the left
when we follow the curve C, and has negative orientation otherwise. For example, a circle followed in
the anticlockwise direction has positive orientation with respect to the region it encloses.
THEOREM 12A. (GREENS THEOREM) Suppose that R R2 is an elementary region of type 3,
with boundary curve C followed with positive orientation. Suppose further that the functions P : R R
and Q : R R are both continuously differentiable. Then
Z
ZZ
Q P
dxdy.
(3)
(P dx + Q dy) =
x
y
C
R
Chapter 12 : Integration Theorems
page 1 of 19
Remarks. (1) Consider the vector field F = (P, Q) in R2 . Then (3) can be written as
Z
ZZ
Q P
F ds =
dxdy.
x
y
C
R
(2) Consider a vector field F (x, y, z) = (P (x, y), Q(x, y), 0) in R3 , and imagine the region R to be a
surface S on the xy-plane, with boundary curve C. We have
Z
Z
Z
F ds =
(P, Q, 0) (dx, dy, dz) =
(P dx + Q dy).
(4)
C
On the other hand, we can parametrize the surface S by the function : R R3 : (x, y) 7 (x, y, 0).
Then tx = (1, 0, 0) and ty = (0, 1, 0), so that tx ty = (0, 0, 1). Hence
ZZ
Z
ZZ
Q P
Q P
(0, 0, 1) dxdy =
dxdy.
(5)
(curl F ) dS =
0, 0,
x
y
x
y
R
R
If we take the oriented surface S to have normal in the positive z-direction, then
Z
Z
(curl F ) dS = (curl F ) dS.
(6)
F ds =
(curl F ) dS.
(7)
(8)
Next, suppose that is an orientation preserving parametrization of C. Then a tangent vector at a point
(x(t), y(t)) is given by (x0 (t), y 0 (t)). Rotating this vector in the clockwise direction by an angle /2 gives
an outward normal vector to C at the point (x(t), y(t)). This outward normal vector is (y 0 (t), x0 (t)),
with unit vector
n=
(y 0 (t), x0 (t))
.
k(x0 (t), y 0 (t))k
It follows that
Z
(P dy Q dx) =
F n ds.
(9)
(div F ) dxdy =
R
F n ds.
This is the 2-dimensional version of Gausss divergence theorem which we shall study in Section 12.3.
Chapter 12 : Integration Theorems
page 2 of 19
xxxxx
(4) Greens theorem can be extended to regions R which are finite unions of essentially disjoint
elementary regions of type 3. For example, consider the annulus
R = {(x, y) R2 : 1 x2 + y 2 4}.
We can cut R into four subregions of type 3 by the lines x = 0 and y = 0.
C2
C1
1
and
C2 = {(x, y) R2 : x2 + y 2 = 4},
with C1 followed in the clockwise direction and C2 followed in the anticlockwise direction.
Proof
of Theorem 12A. Consider first of all the integral
xxxxx
Z
P dx.
(10)
Since R is an elementary region of type 3, it is an elementary region of type 1, and so can be described
in the form (1).
C2
R
C1
A1
B1
The boundary curve C of this region can be split into four parts. There are two straight line segments,
from (A1 , 2 (A1 )) to (A1 , 1 (A1 )) and from (B1 , 1 (B1 )) to (B1 , 2 (B1 )). There are also two curves
C1 = {(x, 1 (x)) : x [A1 , B1 ]}
and
followed from (A1 , 1 (A1 )) to (B1 , 1 (B1 )) and from (B1 , 2 (B1 )) to (A1 , 2 (A1 )) respectively. The
contribution from the two straight line segments to the integral (10) is zero, since dx = 0 on these two
line segments. It follows that
Z
C
P dx =
P dx +
C1
Z
C2
B1
P dx =
B1
P (x, 1 (x)) dx
A1
B1
P (x, 2 (x)) dx
A1
A1
page 3 of 19
P
dxdy =
y
B1
2 (x)
1 (x)
A1
!
Z B1
P
dy dx =
(P (x, 2 (x)) P (x, 1 (x))) dx,
y
A1
P dx =
ZZ
P
dxdy.
y
(11)
Q dy =
ZZ
Q
dxdy.
x
(12)
(x dy y dx) =
ZZ
dxdy.
R
This is equal to the area of R. Suppose now that R is the region bounded by the hypocycloid C of four
cusps, given by the equation x2/3 + y 2/3 = 1 and parametrized by
: [0, 2] R2 : t 7 (cos3 t, sin3 t).
Then dx = 3 cos2 t sin t dt and dy = 3 sin2 t cos t dt. Hence the area of the region bounded by the
hypocycloid is given by
1
2
Z
0
3
(3 sin t cos t + 3 cos t sin t) dt =
2
2
Z
0
sin2 t cos2 t dt =
3
.
8
Example 12.1.2. Let P (x, y) = x2 y cos x + 2xy sin x y 2 ex and Q(x, y) = x2 sin x 2yex . Then
Q
P
= x2 cos x + 2x sin x 2yex =
.
x
y
It follows from Greens theorem that
Z
(P dx + Q dy) = 0
(13)
for the boundary curve C of any elementary region of type 3. Note that (13) holds if C is the boundary
curve of any elementary region of type 3 in which the equality
P
Q
=
x
y
holds.
Chapter 12 : Integration Theorems
page 4 of 19
Example 12.1.3. Let P (x, y) = x2 xy 3 and Q(x, y) = y 2 2xy, and let R denote the square with
vertices (0, 0), (2, 0), (2, 2) and (0, 2).
C3
C4
C2
C1
The boundary curve is then C = C1 C2 C3 C4 , where C1 , C2 , C3 , C4 are the four sides of R followed
in the anticlockwise direction with initial point (0, 0), and can be parametrized respectively by
1 : [0, 2] R2 : t 7 (t, 0),
2 : [0, 2] R2 : t 7 (2, t),
3 : [0, 2] R2 : t 7 (2 t, 2),
4 : [0, 2] R2 : t 7 (0, 2 t).
We have
Z
(P dx + Q dy) =
(P dx + Q dy) =
C1
(P, Q) ds =
1
2
0
(t2 , 0) (1, 0) dt =
8
3
and
Z
C2
(P, Q) ds =
2
Z 2
0
Z
0
16
,
3
as well as
Z
(P dx + Q dy) =
C3
Z
3
Z 2
0
(P, Q) ds =
40
3
and
Z
(P dx + Q dy) =
C4
=
Hence
Z
Z
(P dx + Q dy) =
C
Z
4
Z 2
0
(P, Q) ds =
8
(0, (2 t)2 ) (0, 1) dt = .
3
(P dx + Q dy) +
C1
Z
C2
(P dx +1Q dy) +
Z
C3
(P dx + Q dy) +
(P dx + Q dy) = 8.
C4
page 5 of 19
(P dx + Q dy) =
P (x, 0) dx +
x2 dx +
Z
0
Z
0
Q(2, y) dy
(y 2 4y) dy
Z
0
P (x, 2) dx
(x2 8x) dx
0
2
Q(0, y) dy
y 2 dy = 8.
xxxxx
x
y
R
ZZ
dxdy. =
ZZ
Z
0
(3xy 2y) dx dy =
(6y 2 4y) dy = 8.
y=x
1
C2
R
C1
1
The boundary curve is then C = C1 C2 , where C1 is the part of the parabola from (0, 0) to (1, 1) and
C2 is the part of the line from (1, 1) and (0, 0). The curves C1 and C2 can be parametrized respectively
by
1 : [0, 1] R2 : t 7 (t, t2 )
2 : [0, 1] R2 : t 7 (1 t, 1 t).
and
We have
Z
(P dx + Q dy) =
C1
Z
1
Z 1
0
(P, Q) ds =
19
20
and
Z
(P dx + Q dy) =
C2
Z
2
Z 1
0
(P, Q) ds =
Z
0
Hence
Z
(P dx + Q dy) =
Z
C1
(P dx + Q dy) +
Z
C2
(P dx + Q dy) =
1
.
20
page 6 of 19
Q P
x
y
dxdy. =
ZZ
(x 2y) dxdy =
Z
(x 2y) dy dx =
x2
Z
0
(x4 x3 ) dy =
1
.
20
We shall not give a rigorous proof here. Instead, we only very roughly give an outline of the main
ideas, and show that the result may be deduced from Greens theorem. In the sketch below, we often
make extra assumptions which are not normally necessary.
Heuristics of Theorem 12B. Write F = (F1 , F2 , F3 ). Then
Z
C
F ds =
(F1 dx + F2 dy + F3 dz)
(14)
and
Z
(curl F ) dS =
(curl(F1 , 0, 0)) dS +
(curl(0, F2 , 0)) dS +
(curl(0, 0, F3 )) dS.
(15)
Suppose that
: R R3 : (u, v) 7 (x, y, z).
Let C 0 denote the boundary curve of R, and consider the integral
Z
F1 dx.
C
Since
dx =
Chapter 12 : Integration Theorems
x
x
du +
dv,
u
v
page 7 of 19
F1 dx =
F1
du + F1
dv =
F1
F1
dudv
u
v
u
v
v
u
C
C0
R
ZZ
ZZ
F1 x
2x
F1 x
2x
F1 x F1 x
=
+ F1
F1
dudv =
dudv.
u v
uv
v u
vu
u v
v u
R
R
Next, note that
F1 x F1 y
F1 z
F1
=
+
+
u
x u
y u
z u
F1
F1 x F1 y F1 z
=
+
+
.
v
x v
y v
z v
and
Hence
Z
C
F1 x F1 y
F1 z x
+
+
x u
y u
z u v
R
F1 x F1 y F1 z x
+
+
dudv
x v
y v
z v u
ZZ
F1 z x
F1 y F1 z x
F1 y
=
+
+
dudv
y u
z u v
y v
z v u
R
ZZ
F1 z x z x
F1 x y x y
=
dudv
z u v
v u
y u v
v u
R
ZZ
F1 (z, x) F1 (x, y)
=
dudv.
z (u, v)
y (u, v)
R
F1 dx =
ZZ
,
,
dudv
z
y
(u, v) (u, v) (u, v)
S
R
ZZ
F1 (z, x) F1 (x, y)
=
dudv.
z (u, v)
y (u, v)
R
Combining (16) and (17), we obtain
Z
F1 dx =
(17)
(18)
Similarly, we have
Z
Z
xxxxx
F2 dy = (curl(0, F2 , 0)) dS
C
(16)
and
F3 dz =
(curl(0, 0, F3 )) dS.
(19)
3
S
x
Chapter 12 : Integration Theorems
page 8 of 19
Then the boundary curve C is given by x2 + y 2 = 9, followed in the anticlockwise direction. Consider
the vector field F (x, y, z) = (2y, 3x, z 2 ). Let us first of all evaluate the integral
Z
F ds.
C
F ds =
F ((t)) 0 (t) dt =
Z
0
(curl F ) dS.
xxxxx
ZZ
(curl F ) dS =
z
2
n
x
Chapter 12 : Integration Theorems
page 9 of 19
Then the boundary curve C is given by x2 + y 2 = 4, followed in the clockwise direction. Consider the
vector field F (x, y, z) = (x z, x3 + yz, 3xy 2 ). Let us first of all evaluate the integral
Z
F ds.
C
F ds =
0
2
F ((t)) 0 (t) dt =
Z
0
(curl F ) dS.
= 3
Z
u cos v dv du = 12.
In other words, the value of the integral depends only on the endpoints of the path .
Chapter 12 : Integration Theorems
page 10 of 19
With the help of Stokess theorem, we can characterize gradient vector fields.
THEOREM 12C. Suppose that F : R3 R3 is a continuously differentiable vector field. Then the
following statements are equivalent:
(a) For any oriented simple closed curve C, we have
Z
F ds = 0.
C
(b) For any two oriented simple curves C1 and C2 with the same initial point and the same terminal
point, we have
Z
Z
F ds =
F ds.
C1
C2
C2
If C is not simple, then an elaboration of this argument will give the same result. To show that (b)(c),
let C be any oriented simple curve with initial point (0, 0, 0) and terminal point (x, y, z), and write
Z
xxxxx
f (x, y, z) =
F ds.
C
Since (b) holds, F (x, y, z) is independent of the choice of C. In particular, we can take C to be the line
segment from (0, 0, 0) to (x, 0, 0), followed by the line segment from (x, 0, 0) to (x, y, 0), followed by the
line segment from (x, y, 0) to (x, y, z).
z
(x, y, z)
(x, 0, 0)
(x, y, 0)
x
Assume first of all that x, y, z are all positive. Then the three line segments can be parametrized
respectively by
1 : [0, x] R3 : t 7 (t, 0, 0),
2 : [0, y] R3 : t 7 (x, t, 0),
3 : [0, z] R3 : t 7 (x, y, t),
so that writing F = (F1 , F2 , F3 ), we have
Z x
Z
f (x, y, z) =
F1 (t, 0, 0) dt +
0
F2 (x, t, 0) dt +
Z
0
F3 (x, y, t) dt.
page 11 of 19
With a little modification in argument, this last formula can be shown to hold even if x, y, z are not all
positive. By the Fundamental theorem of calculus, we clearly have
f
= F3 .
z
By using different paths, it can be shown that
f
= F1
x
and
f
= F2 ,
y
so that f = F . That (c)(d) is proved in Theorem 8G. Finally, to prove that (d)(a), we simply
apply Stokess theorem with any surface S whose boundary is C.
Remarks. (1) In the statement of Theorem 12C, it is possible to assume that the vector field F is
continuously differentiable in R3 , except possibly at a finite number of points. The proof only needs
minor modification.
(2) There is a 2-dimensional version of Theorem 12C. Recall that
Q P
x
y
is the scalar curl of a vector field F = (P, Q) in R2 . Thus there exists a function f : R2 R such that
F = f everywhere in R2 if and only if
Q P
=0
x
y
everywhere in R2 . Here Greens theorem plays the role of Stokess theorem in establishing the result.
However, we cannot allow exceptions to the condition that F is continuously differentiable in R2 .
(3) Theorem 12C is in some sense the converse of Theorem 8G. Recall now Theorem 8F, that for any
twice continuously differentiable vector field F in R3 , we have div(curl F ) = 0. One can prove that if G
is a continuously differentiable vector field in R3 with div G = 0, then there exists a vector field F in R3
such that G = curl F .
(20)
page 12 of 19
F dS =
(div F ) dV.
(2) Gausss theorem is in fact valid for any region V which can be expressed as a union of finitely
many essentially disjoint symmetric elementary regions.
(3) We shall see that the proof of Gausss theorem is very similar to that of Greens theorem.
Sketch of Proof of Theorem 12D. Write F = (F1 , F2 , F3 ). Then
Z
Z
F dS = ((F1 , 0, 0) + (0, F2 , 0) + (0, 0, F3 )) dS
S
ZS
Z
Z
= (F1 , 0, 0) dS + (0, F2 , 0) dS + (0, 0, F3 ) dS
S
(21)
and
F2
F3
F1
+
+
dxdydz
(div F ) dxdydz =
x
y
z
V
ZZZ
ZZZ
ZZZV
F2
F3
F1
dxdydz +
dxdydz +
dxdydz.
=
x
y
V
V z
V
ZZZ
ZZZ
(22)
(0, 0, F3 ) dS =
ZZZ
F3
dxdydz.
z
(23)
Since V is a symmetric elementary region, it can be described in the form (20), so that
!
ZZZ
ZZ
Z 2 (x,y)
F3
F3
dxdydz =
dz dxdy
z
V z
R
1 (x,y)
ZZ
=
(F3 (x, y, 2 (x, y)) F3 (x, y, 1 (x, y))) dxdy.
(24)
On the other hand, the boundary surface S can be partitioned into six surfaces, with bottom surface
S1 = {(x, y, z) R3 : (x, y) R and z = 1 (x, y)},
top surface
S2 = {(x, y, z) R3 : (x, y) R and z = 2 (x, y)},
and four side surfaces S3 , S4 , S5 , S6 corresponding to the four edges of the elementary region R. The
normal vectors to the surfaces S3 , S4 , S5 , S6 are all horizontal, with no component in the z-direction.
Hence
Z
Z
Z
Z
(0, 0, F3 ) dS =
(0, 0, F3 ) dS =
(0, 0, F3 ) dS =
(0, 0, F3 ) dS = 0,
S3
Chapter 12 : Integration Theorems
S4
S5
S6
page 13 of 19
and so
Z
(0, 0, F3 ) dS =
(0, 0, F3 ) dS +
S1
(0, 0, F3 ) dS.
(25)
S2
1, 0,
1
x
1
1
1
0, 1,
=
,
,1 .
y
x
y
(0, 0, F3 ) dS =
ZZ
S1
ZZ
1
1
(0, 0, F3 )
,
, 1 dxdy =
F3 (x, y, 1 (x, y)) dxdy.
x
y
R
R
(26)
1, 0,
2
x
2
2
2
0, 1,
=
,
,1 .
y
x
y
(0, 0, F3 ) dS =
ZZ
S2
ZZ
2
2
(0, 0, F3 )
,
, 1 dxdy =
F3 (x, y, 2 (x, y)) dxdy.
x
y
R
R
(27)
(F1 , 0, 0) dS =
ZZZ
F1
dxdydz
x
and
(0, F2 , 0) dS =
ZZZ
V
F2
dxdydz.
y
(28)
F dS.
page 14 of 19
+
=
Next, note that
ZZZ
Z
8
.
3
Z
sin u cos3 u du
(div F ) dxdydz =
ZZZ
dv
(2 + 2y + 2z) dxdydz = 2
ZZZ
(1 + y + z) dxdydz.
We can write
V = {(x, y, z) R3 : x2 + y 2 1 and
1 x2 y 2 z
p
1 x2 y 2 },
so that
ZZZ
z dxdydz =
Z 1x2 y2
ZZ
x2 +y 2 1
1x2 y 2
!
z dz
dxdy = 0.
Similarly
ZZZ
y dxdydz = 0.
Hence
ZZZ
(div F ) dxdydz = 2
ZZZ
dxdydz =
8
,
3
since the volume of the unit sphere is equal to 4/3. This verifies Gausss theorem.
Example 12.3.2. Let V be the cube with vertices (1, 1, 1), with boundary surface S. Consider the
vector field F (x, y, z) = (x, y, z). We have shown in Example 11.4.6 that
Z
F dS = 24.
S
Now
ZZZ
(div F ) dxdydz = 3
ZZZ
dxdydz = 24.
This verifies Gausss theorem. In fact, we can generalize this observation. Suppose that S is the boundary
surface of any region V in R3 for which Gausss theorem holds. Then
Z
ZZZ
r dS = 3
dxdydz,
S
page 15 of 19
dS
=
3
0
if (0, 0, 0) 6 V ,
S r
where r = (x, y, z) denotes any point on S, and r = krk = (x2 + y 2 + z 2 )1/2 .
Sketch of Proof. Suppose first of all that (0, 0, 0) 6 V . Then the vector field
r
r3
is continuously differentiable on V , and so it follows from Gausss theorem that
Z
ZZZ
r
r
dS
=
div
dxdydz.
3
r3
S r
V
It is easy to check that
div
r
=0
r3
whenever r 6= 0. The desired conclusion therefore holds in this case. Suppose now that (0, 0, 0) V .
Since (0, 0, 0) 6 S, it follows that there exists > 0 such that the open ball B(), with centre (0, 0, 0)
and radius , satisfies B() V . Now let = V \ B(), the region V with the open ball B() removed.
Clearly this region has boundary surface S T , where T is the boundary surface of B() with normal
pointing towards (0, 0, 0). Applying Gausss theorem to this region (note that is not an elementary
region), we have
Z
Z
ZZZ
r
r
r
dS
+
dS
=
div
dxdydz = 0,
3
3
r3
S r
T r
so that
Z
S
r
dS =
r3
Z
T
r
dS.
r3
dS =
(sin2 u cos v, sin2 u sin v, sin u cos u) dudv
3
r
3
T
R
ZZ
=
(sin u cos v, sin u sin v, cos u) (sin2 u cos v, sin2 u sin v, sin u cos u) dudv
R
Z
Z 2
ZZ
=
sin u dudv =
sin u du
dv = 4.
R
page 16 of 19
x2
y
+ y2
and
Q(x, y) =
x2
x
,
+ y2
and let R be the disc in R2 with centre (0, 0) and radius 1, with boundary C followed in the
anticlockwise direction. Evaluate the integrals
Z
ZZ
Q P
(P dx + Q dy)
and
dxdy,
x
y
C
R
and explain why they have different values.
5. Consider the paraboloid S = {(x, y, z) R3 : 2z = x2 + y 2 and z 2}, with boundary curve C.
Consider also the vector field F (x, y, z) = (3y, xz, yz 2 ).
a) Show that curl F = (z 2 + x, 0, z 3).
b) Let R = {(u, v) R2 : u2 + v 2 4}. Show that
3
: R R : (u, v) 7
u2 + v 2
u, v,
2
(curl F ) dS =
ZZ
u2 + v 2
2
2
!
u2 + v 2
+u +
+ 3 dudv.
2
2
Using the substitution u = r cos and v = r sin , show that the integral above has value 20.
c) Suppose that the surface S is oriented with normal vector pointing towards the positive z-axis.
Determine whether the parametrization is orientation preserving or orientation reversing, and
write down the value of the integral
Z
(curl F ) dS.
S
d) Find a suitable orientation of the path C which you must specify carefully, and verify Stokess
theorem.
Chapter 12 : Integration Theorems
page 17 of 19
a) F (x, y, z) = (x2 + y 4, 3xy, 2xz + z 2 ) and S is the part of the sphere x2 + y 2 + z 2 = 16 above
the xy-plane.
b) F (x, y, z) = (x2 + y 4, 3xy, 2xz + z 2 ) and S is the part of the paraboloid z = 4 (x2 + y 2 ) above
the xy-plane..
c) F (x, y, z) = (2yz, 2 x 3y, x2 + z) and S is the surface of the intersection of the cylinders
x2 + y 2 = a2 and x2 + z 2 = a2 in the first octant x, y, z 0.
9. Suppose that F (x, y, z) = (2xy + 3, x2 4z, 4y).
a) Show that curl F = 0 everywhere in R3 .
b) Choose any curve C from (3, 1, 2) to (2, 1, 1), and evaluate the integral
F ds
c) Suppose that C 0 is a curve from (3, 1, 2) to (2, 1, 1) different from your choice in part (b).
Z
Comment on the value of the integral
F ds, and justify your assertion by quoting the relevant
C0
results.
10. Consider two surfaces S1 and S2 with the same boundary. Describe with sketches how S1 and S2
must be oriented to ensure that
Z
Z
(curl F ) dS =
(curl F ) dS.
S1
S2
12. Suppose that b > a > 0, and that C is any path joining any point on the sphere x2 + y 2 + z 2 = a2 to
any point on the sphere x2 + y 2 + z 2 = b2 . Suppose also that F (x, y, z) = 5r3 r, where r = (x, y, z)
and r = krk = (x2 + y 2 + z 2 )1/2 . Use Theorem 12C to show that
Z
F ds = b5 a5 .
C
13. Suppose that F (x, y, z) = (4x, 2y 2 , z 2 ). Suppose further that V is the region bounded by the
cylinder x2 + y 2 = 4 and the planes z = 0 and z = 3, and that S is the boundary surface of V
oriented with outward normal. Verify that
Z
ZZZ
F dS =
(div F ) dxdydz.
S
Chapter 12 : Integration Theorems
V
page 18 of 19
r
dS =
r2
ZZZ
V
1
dxdydz.
r2
17. Suppose that n is the outward unit normal at any point on the surface S of a region V . Prove that
the surface area of S is equal to
ZZZ
(div n) dxdydz.
V
(f g) dS =
ZZZ
(f 2 g + (f ) (g)) dxdydz.
(f g gf ) dS =
ZZZ
(f 2 g g2 f ) dxdydz.
page 19 of 19