Ch18 LT Instructor Solution
Ch18 LT Instructor Solution
Ch18 LT Instructor Solution
18 THEOREMS OF
VE C TOR ANALYSIS
18.1 Greens Theorem (ET Section 17.1)
Preliminary Questions
1. Which vector field F is being integrated in the line integral x 2 d y e y d x?
SOLUTION The line integral can be rewritten as e y d x + x 2 d y. This is the line integral of F = e y , x 2 along
the curve.
2. Draw a domain in the shape of an ellipse and indicate with an arrow the boundary orientation of the boundary curve.
Do the same for the annulus (the region between two concentric circles).
SOLUTION The orientation on C is counterclockwise, meaning that the region enclosed by C lies to the left in traversing
C.
C
For the annulus, the inner boundary is oriented clockwise and the outer boundary is oriented counterclockwise. The
region between the circles lies to the left while traversing each circle.
3. The circulation of a gradient vector field around a closed curve is zero. Is this fact consistent with Greens Theorem?
Explain.
SOLUTION Greens Theorem asserts that
Q P
F ds = P dx + Q dy = dA (1)
C C D x y
If F is a gradient vector field, the cross partials are equal, that is,
P Q Q P
= =0 (2)
y x x y
Combining (1) and (2) we obtain C F ds = 0. That is, Greens Theorem implies that the integral of a gradient vector
field around a simple closed curve is zero.
4. Which of the following vector fields possess the following property: For every simple closed curve C, F ds is
C
equal to the area enclosed by C?
(a) F = y, 0
(b) F = x, y
2
(c) F = sin(x 2 ), x + e y
SOLUTION By Greens Theorem,
Q P
F ds = dx dy (1)
C D x y
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1195
Exercises
1. Verify Greens Theorem for the line integral x y d x + y d y, where C is the unit circle, oriented counterclockwise.
C
SOLUTION
Step 1. Evaluate the line integral. We use the parametrization ( ) = cos , sin , 0 2 of the unit circle. Then
d x = sin d , d y = cos d
and
x y d x + y d y = cos sin ( sin d ) + sin cos d = cos sin2 + sin cos d
D
x
The integral of x over the disk D is zero, since by symmetry the positive and negative values of x cancel each other.
Therefore,
Q P
dx dy = 0 (2)
D x y
Step 3. Compare. The line integral in (1) is equal to the double integral in (2), as stated in Greens Theorem.
2
2. Let I = F ds, where F = y + sin x 2 , x 2 + e y and C is the circle of radius 4 centered at the origin.
C
(a) Which is easier: evaluating I directly or using Greens Theorem?
(b) Evaluate I using the easier method.
SOLUTION
(a) Using the parametrization ( ) = 4 cos , 4 sin for the circle, we have
d x = 4 sin d , d y = 4 cos d
and
(y + sin x 2 )d x + (x 2 + e y )d y = 4 sin + sin(16 cos2 ) (4 sin ) d + (16 cos2 + e16 sin ) 4 cos d
2 2
= 16 sin2 4 sin sin(16 cos2 ) + 64 cos3 + 4 cos e16 sin d
2
Clearly, the double integral in (2) is much easier to evaluate than the line integral in (1).
(b) To evaluate the double integral in (2), we notice that by symmetry the integral of 2x over D is zero, since the positive
and negative values of x cancel each other. Hence,
Q P
dA = (2x 1) d x d y = 2x d x d y 1 dx dy
D x y D D D
= 0 Area(D) = 42 = 16
In Exercises 311, use Greens Theorem to evaluate the line integral. Orient the curve counterclockwise unless otherwise
indicated.
3. y 2 d x + x 2 d y, where C is the boundary of the unit square 0 x 1, 0 y 1
C
SOLUTION
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1197
1 C
x
1
D
(1, 1)
C
D
x
(0, 0) (1, 0)
0yx
y=x
D
x
0 1
5. x 2 y d x, where C is the unit circle centered at the origin
C
SOLUTION
1198 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
D
x
y= x
C
D
y = x2
x
0 1
y=x
D
y = x2
x
0 1
We have P = Q = x 2 , therefore
Q P
= 2x 0 = 2x
x y
Hence,
1 x 1 x 1 1
I = 2x d A = 2x d y d x = 2x y dx = 2x(x x 2 ) d x = (2x 2 2x 3 ) d x
D 0 x2 0 y=x 2 0 0
2 3 1 4 1 2 1 1
= x x = =
3 2 0 3 2 6
8. (ln x + y) d x x 2 d y, where C is the rectangle with vertices (1, 1), (3, 1), (1, 4), and (3, 4)
C
SOLUTION By Greens Theorem,
Q P
I = F ds = dA
C D x y
(1, 4) (3, 4)
(1, 1) (3, 1)
9. The line integral of F = x 3 , 4x around the boundary of the parallelogram in Figure 15 (note the orientation)
(2, 2) (4, 2)
x
(2, 0)
FIGURE 15
1200 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
Q P
=40=4
x y
Hence, Greens Theorem implies
Q P
x 3 d x + 4x d y = dA = 4dA = 4 d A = 4 Area(D) = 4 4 = 16
C D x y D D
10. The line integral of F = e x+y , e xy around the boundary of the parallelogram in Figure 15
SOLUTION We compute the curl of F = e x+y , e xy . Here, P = e x+y and Q = e xy , hence
Q P
= e xy e x+y = e x (ey e y ).
x y
y
x=y
x=y+2
11. x y d x + (x 2 + x) d y, where C is the path in Figure 16
C
(0, 1)
x
(1, 0) (1, 0)
FIGURE 16
SOLUTION
(0, 1)
C
D
x
(1, 0) (1, 0)
By symmetry, the positive and negative values of x cancel each other, causing the first integral to be zero. Thus,
21
x y d x + (x 2 + x) d y = 0 + d A = Area(D) = = 1.
C D 2
12. Let C R be the circle of radius R centered at the origin. Use Greens Theorem to determine F ds, where F is a
C2
vector field such that F ds = 9 and curlz (F) = x 2 + y 2 for 1 x 2 + y 2 4.
C1
SOLUTION We use Greens Theorem for the annulus D between the circles C(1) and C(2) oriented as shown in the
figure.
y
C(8)
D C(3)
x
3 8
That is,
F ds F ds = curl(F) d x d y
C (2) C (1) D
In Exercises 1316, use Eq. (5) to calculate the area of the given region.
Therefore,
1 1 2 9
A= x dy y dx = 9 d = 2 = 9 .
2 C 2 0 2
14. The triangle with vertices (0, 0), (1, 0), and (1, 1)
SOLUTION We parametrize the segments by
O A : t, 0 , t = 0 to t = 1
AB : 1, t , t = 0 to t = 1
B O : t, t , t = 1 to t = 0
1202 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
B = (1, 1)
x
O = (0, 0) A = (1, 0)
15. The region between the x-axis and the cycloid parametrized by c(t) = (t sin t, 1 cos t) for 0 t 2 (Figure
17)
x
2
FIGURE 17 Cycloid.
x
O A = (2 , 0)
Therefore,
1 1
A= x dy y dx + x dy y dx (1)
2 OA 2
We compute the two integrals. The segment O A is parametrized by t, 0, t = 0 to t = 2 . Hence, x = t and y = 0.
Therefore,
x d y y d x = t 0 dt 0 dt = 0
x dy y dx = 0 (2)
OA
Hence,
0 2
x dy y dx = (t sin t + 2 cos t 2) dt = (2 2 cos t t sin t) dt
2 0
2 2
= 2t 2 sin t + t cos t sin t = 2t 3 sin t + t cos t = 6 (3)
0 0
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1203
16. The region between the graph of y = x 2 and the x-axis for 0 x 2
SOLUTION The boundary of the region consists of the curve and the segments O A and AB shown in the figure.
y
B = (2, 4)
4
x
0 A = (2, 0)
y
2
x
2 2
(a) Show that the folium has a parametrization in terms of t = y/x given by
3t 3t 2
x= , y= ( < t < ) (t = 1)
1 + t3 1 + t3
1204 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
9t 2
x dy y dx = dt
(1 + t 3 )2
Hint: By the Quotient Rule,
y
x2 d = x dy y dx
x
(c) Find the area of the loop of the folium.
SOLUTION
2
(a) We show that x = 3t 3 , y = 3t 3 satisfy the equation x 3 + y 3 3x y = 0 of the folium:
1+t 1+t
3 3
3t 3t 2 3t 3t 2
x 3 + y 3 3x y = + 3
1 + t3 1 + t3 1 + t3 1 + t3
27t 3 + 27t 6 27t 3 (1 + t 3 ) 27t 3 1 + t 3 (1 + t 3 ) 0
= 3
3
= 3
= 3
=0
(1 + t 3 ) (1 + t 3 ) (1 + t 3 ) (1 + t 3 )
2
This proves that the curve parametrized by x = 3t 3 , y = 3t 3 lies on the folium of Descartes. This parametrization
1+t 1+t
parametrizes the whole folium since the two equations can be solved for t in terms of x and y. That is,
3t
x=
1 + t3 y
t=
3t 2 x
y=
1 + t3
A glance at the graph of the folium shows that any line y = t x, with slope t, intersects the folium exactly once. Thus,
there is a one-to-one relationship between the values of t and the points on the graph.
(b) We differentiate the two sides of t = xy with respect to t. Using the Quotient Rule gives
x ddty y ddtx
1=
x2
or
2
dy dx 3t
x y = x2 =
dt dt 1 + t3
This equality can be written in the form
9t 2
x dy y dx = 2
dt
(1 + t 3 )
(c) We use the formula for the area enclosed by a closed curve and the result of part (b) to find the required area. That is,
1 1 9t 2
A= x dy y dx = dt
2 C 2 0 (1 + t 3 )2
From our earlier discussion on the parametrization of the folium, we see that the loop is traced when the parameter t
is increasing along the interval 0 t < . We compute the improper integral using the substitution u = 1 + t 3 ,
du = 3t 2 dt. This gives
R 1+R 3 3
1 9t 2 1 3 du 3 1 1+R
A= lim dt = lim = lim
2 R 0 (1 + t 3 )2 2 R 1 u2 2 R u u=1
3 1 3 3
= lim 1 = (1 0) =
2 R 1 + R3 2 2
y
2
C
t=
x
2 t=0 2
2
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1205
18. Follow the procedure of Exercise 17 to find the area of the loop of the lemniscate curve with equation (x 2 + y 2 )2 =
x y (Figure 19).
0.5
0.5
x
0.5
0.5
FIGURE 19 Lemniscate.
SOLUTION We first find the parametrization of the lemniscate, determined by the parameter t = xy . We substitute
y = t x in the equation of the lemniscate and solve for x in terms of t. This gives
2 2 2
(x 2 + y 2 ) = (x 2 + t 2 x 2 ) = (1 + t 2 ) x 4
xy = x t x = t x2
or (taking x positive)
t 1/2
x=
1 + t2
hence
t 3/2
y = tx = .
1 + t2
We obtain the parametrization
t 1/2 t 3/2
x= , y= .
1 + t2 1 + t2
One loop is traced as 0 t < .
y
t=
x
t=0
The area enclosed by one loop is given by the following line integral:
1
A= x dy y dx (1)
2 C
x d y y d x = x 2 dt
1/2
Now x = t 2 , hence
1+t
t
x dy y dx = 2
dt
(1 + t 2 )
We substitute in (1) and compute the resulting improper integral using the substitution u = 1 + t 2 , du = 2t dt. We get
R 1+R 2 1
1 1 t 1 t dt 1 2 du du
A= x dy y dx = 2
dt = lim 2
= lim
2 C 2 0 (1 + t 2 ) 2 R 0 (1 + t 2 ) 2 R 1 u2
1206 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
2
1 1 1+R 1 1 1 1
= lim = lim 1 = (1 0) =
4 R u u=1 4 R 1 + R2 4 4
20. For the vector fields (A)(D) in Figure 20, state whether the curl at the origin appears to be positive, negative, or
zero.
y y
x x
(A) (B)
y y
x x
(C) (D)
FIGURE 20
SOLUTION The vector field (A) does not have spirals nor it is a shear flow. Therefore, the curl appears to be zero. The
vector field (B) rotates in the clockwise direction, hence we expect the curl to be negative. The vector field (C) rotates
counterclockwise around the origin, indicating a positive curl. Finally, in the vector field (D) the fluid flows straight
toward the origin without spiraling. We expect the curl to be zero.
21. Let F = 2xe y , x + x 2 e y and let C be the quarter-circle path from A to B in Figure 21. Evaluate I = F ds as
C
follows:
(a) Find a function (x, y) such that F = G + , where G = 0, x.
(b) Show that the line integrals of G along the segments O A and O B are zero.
(c) Use Greens Theorem to show that
I = (B) ( A) + 4
and evaluate I .
y
B = (0, 4)
x
O A = (4, 0)
FIGURE 21
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1207
SOLUTION
(a) We need to find a potential function (x, y) for the difference
F G = 2xe y , x + x 2 e y 0, x = 2xe y , x 2 e y
We let (x, y) = x 2 e y .
(b) We use the parametrizations AO : t, 0, 0 t 4 and O B : 0, t, 0 t 4 to evaluate the integrals of
G = 0, x. We get
4 4
G ds = 0, t 1, 0 dt = 0 dt = 0
OA 0 0
4 4
G ds = 0, 0 0, 1 dt = 0 dt = 0
OB 0 0
B = (0, 4)
x
O A = (4, 0)
We now use Greens Theorem. Since G = 0, x, we have P = 0 and Q = x, hence Q P
x y = 1 0 = 1. Thus,
42
G ds = 1 d A = Area(D) = = 4 (3)
B O+O A+C D 4
Combining (1), (2), and (3), we obtain
I = (B) ( A) + 4
B = (0, 4)
C
x
O A = (4, 0)
22. Compute the line integral of F = x 3 , 4x along the path from A to B in Figure 22. Hint: To save work, use Greens
Theorem to relate this line integral to the line integral along the vertical path from B to A.
1208 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
A = (1, 0)
x
B = (1, 1)
FIGURE 22
SOLUTION We denote by C the path from A to B, and D is the region enclosed by C and the segment B A.
y
C 1
D
A = (1, 0)
x
2
B = (1, 1)
We compute the integrals on the right-hand side. We parametrize the segment AB by 1, t, with t from 0 to 1. We
get
1 1 0
F ds = 1, 4 0, 1 dt = 4 dt = 4 dt = 4 (2)
AB 0 0 1
D = (0, 6)
(2, 4)
(2, 2)
A = (0, 0)
x
FIGURE 23
SOLUTION
D = (0, 6)
C
C = (2, 4)
D
B = (2, 2)
A = (0, 0)
x
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1209
Let F = sin x + y, 3x + y, hence P = sin x + y and Q = 3x + y. We denote by C1 the closed path determined by C
and the segment D A. Then by Greens Theorem,
Q P
P dx + Q dy = dA = (3 1) d A = 2 d A = 2 Area(D) (1)
C1 D x y D D
D = (0, 6)
C = (2, 4)
h
B = (2, 2)
A = (0, 0)
x
D A : x = 0, y = t, t varies from 6 to 0.
We get
0 0
d
P dx + Q dy = F(0, t) 0, t dt = sin 0 + t, 3 0 + t 0, 1 dt
DA 6 dt 6
0 0
t 2 0
= t, t 0, 1 dt = t dt = = 18
6 6 2 t=6
24. Estimate the circulation of a vector field F around a circle of radius R = 0.1, assuming that curlz (F) takes the value
4 at the center of the circle.
SOLUTION We estimate the circulation by
F ds curl(F)(P)Area(D) (1)
C
C
1
P
We are given that curl(F)(P) = 4. The area of the disk of radius R = 0.1 is 0.12 = /100. Substituting in (1) gives
the estimation
F ds 4 =
C 100 25
1210 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
25. Let F be the velocity field. Estimate the circulation of F around a circle of radius R = 0.05 with center P, assuming
that curlz (F)(P) = 3. In which direction would a small paddle placed at P rotate? How fast would it rotate (in radians
per second) if F is expressed in meters per second?
SOLUTION We use the following estimation:
F ds curl(F)(P)Area(D) (1)
C
C
D
0.05
P
We are given that curl(F)(P) = 3. Also, the area of the disk of radius R = 0.05 is 0.052 = 0.0025 . Therefore, we
obtain the following estimation:
F ds 3 0.0025 0.024.
C
Since the curl is negative, the paddle would rotate in the clockwise direction. Using the formula |curl(F)| = 2 , we see
that the angular speed is = 1.5 radians per second.
26. Referring to Figure 24, suppose that F ds = 12. Use Greens Theorem to determine F ds, assuming that
C2 C1
curlz (F) = 3 in D.
C1 3
C2 D
x
2 5
FIGURE 24
Hence,
F ds 12 = 3 Area(D) (1)
C1
We compute the area of D as the difference between the area of the rectangle and the area of the inner disk. That is,
Area(D) = 6 10 22 = 60 4
Use Greens Theorem to determine the circulation of F around C1 , assuming that curlz (F) = 9 on the shaded region.
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1211
C1
D
C3 C2
1 1
FIGURE 25
SOLUTION We must calculate C1 F ds. We use Greens Theorem for the region D between the three circles C1 , C2 ,
and C3 . Because of orientation, the line integrals C2 F ds = C2 F ds and C3 F ds = C3 F ds must be used
in applying Greens Theorem. That is,
F ds F ds F ds = curl(F) d A
C1 C2 C3 D
Area(D) = 52 12 12 = 23
28. Area of a Polygon Greens Theorem leads to a convenient formula for the area of a polygon.
(a) Let C be the line segment joining (x 1 , y1 ) to (x 2 , y2 ). Show that
1 1
y d x + x d y = (x 1 y2 x 2 y1 )
2 C 2
(b) Prove that the area of the polygon with vertices (x 1 , y1 ), (x 2 , y2 ), . . . , (x n , yn ) is equal [where we set
(x n+1 , yn+1 ) = (x 1 , y1 )] to
1 n
(xi yi+1 xi+1 yi )
2 i=1
SOLUTION
(a) We parametrize the segment from (x 1 , y1 ) to (x 2 , y2 ) by
x = t x 2 + (1 t)x 1 , y = t y2 + (1 t)y1 , 0t 1
An = (xn , yn)
An 1 = (xn 1, yn 1)
A1 = (x1, y1)
A2 = (x2, y2)
1212 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
(b) Let Ai = (xi , yi ), i = 1, 2, . . . , n, and let C be the closed curve determined by the polygon. By the formula for the
area enclosed by a simple closed curve, the area of the polygon is
1
A= y d x + x d y
2 C
We use additivity of line integrals and the result in part (a) to write the integral as follows:
1 n1
A= y d x + x d y + y d x + x d y
2 i=1 Ai Ai+1 An A1
1 n1
= (xi yi+1 xi+1 yi ) + (x n y1 x 1 yn )
2 i=1
1 n1 1
= (xi yi+1 xi+1 yi ) + (x n y1 x 1 yn )
2 i=1 2
1 n
A= (xi yi+1 xi+1 yi ).
2 i=1
29. Use the result of Exercise 28 to compute the areas of the polygons in Figure 26. Check your result for the area of the
triangle in (A) using geometry.
y y
(3, 5)
5 5
4 4
(2, 3) (1, 3) (5, 3)
3 3
2 2
(3, 2)
1 (1, 1) 1
(2, 1) (5, 1)
x x
1 2 3 4 5 3 2 1 1 2 3 4 5
(A) (B)
FIGURE 26
SOLUTION
(a) The vertices of the triangle are
5
4
(2, 3)
3
2
1
(2, 1) (5, 1)
x
1 2 3 4 5
Using the formula obtained in Exercise 28, the area of the triangle is the following sum:
1
A= (x y x 2 y1 ) + (x 2 y3 x 3 y2 ) + (x 3 y1 x 1 y3 )
2 1 2
1 1
= (2 1 5 1) + (5 3 2 1) + (2 1 2 3) = (3 + 13 4) = 3
2 2
We verify our result using the formula for the area of a triangle:
1 1
A= bh = (5 2) (3 1) = 3
2 2
(b) The vertices of the polygon are
(x 1 , y1 ) = (x 6 , y6 ) = (1, 1)
(x 2 , y2 ) = (1, 3)
(x 3 , y3 ) = (3, 2)
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1213
(x 4 , y4 ) = (5, 3)
(x 5 , y5 ) = (3, 5)
y
(3, 5)
5
4
(1, 3) (5, 3)
3
2
(3, 2)
(1, 1) 1
x
3 2 1 1 2 3 4 5
Using the formula in part (a), the area of the polygon is the following sum:
1
A= (x 1 y2 x 2 y1 ) + (x 2 y3 x 3 y2 ) + (x 3 y4 x 4 y) + (x 4 y5 x 5 y4 ) + (x 5 y1 x 1 y5 )
2
1
= (1 3 1 1) + (1 2 3 3) + (3 3 5 2) + 5 5 (3) 3 + 3 1 (1) 5
2
1
= (4 7 1 + 34 + 2) = 12
2
P y 1 (x 2 + y 2 ) + y 2y y2 x 2
= = =
y y x + y 2 2
(x 2 + y 2 )
2
(x 2 + y 2 )
2
Since Q P
x = y , the curl of F is identically zero for (x, y) = (0, 0):
Q P
curl(F) = =0
x y
(b) Let C R be the circle of radius R centered at the origin. We parametrize C R by
Hence,
R sin R cos sin cos
F (C( )) = , = ,
R 2 cos2 + R 2 sin2 R 2 cos2 + R 2 sin2 R R
sin cos
F (C( )) C ( ) = , R sin , R cos = sin2 + cos2 = 1
R R
The line integral is thus
2 2 2
F ds = F (C( )) C ( ) d = 1 d = = 2 .
C 0 0 0
31. Prove that if C is a simple closed curve whose interior contains the origin, then F ds = 2 (Figure 27). Hint:
C
Apply Greens Theorem to the domain between C and C R where R is so small that C R is contained in C.
1214 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
C
CR
x
FIGURE 27
C
CR D
x
Let D denote the region between C R and C. We apply Greens Theorem to the region D. The curve C is oriented
counterclockwise and C R is oriented clockwise. We have
F ds + F ds = curl(F) d A (1)
C CR D
From Exercise 30(b) we know that C R F ds = 2 . Since D does not contain the origin, we have by part (a) of
Exercise 30, curl (F) = 0 on D. Substituting in (1) we obtain
F ds 2 = 0dA = 0
C D
or
F ds = 2 .
C
In Exercises 3234, the conjugate of a vector field F = P, Q is the vector field F = Q, P.
32. Explain the following statement: F is the vector field obtained by rotating the vectors of F counterclockwise through
an angle of 2 .
SOLUTION We show that the vector b = y, x is obtained by rotating the vector a = x, y counterclockwise. First,
notice that a and b are on the same circle centered at the origin, since a = b = x 2 + y 2 . Notice also that since
a b = yx + x y = 0, the angle between the two vectors is 2 , and the two vectors are in adjacent quadrants. In addition,
the x-coordinate of b and the y-coordinate of a have opposite signs. By these reasonings we conclude that b is obtained
by rotating a counterclockwise by 2 .
33. The normal component of F at a point P on a simple closed C is the quantity F(P) n(P), where n(P) is the
C is defined as the line integral of the normal component
outward-pointing unit normal vector. The flux of F across curve
around C (Figure 28). Show that the flux across C is equal to F ds.
C
P
n(P)
FIGURE 28 The flux of F is the integral of the normal component F n around the curve.
By the definition of the vector line integral, the line integral of the vector field F over C is
F ds = (F T) ds (1)
C C
In Exercise 32 we showed that F is a rotation of F by 2 counterclockwise. Therefore, the angle between F and the
tangent T is equal to the angle between F and the normal n. Also, F = F. Hence,
F T = F T cos = F cos
F n = Fn cos = F cos
P Q
34. Define div(F) = + . Use Greens Theorem to prove that for any simple closed curve C,
x y
Flux across C = div(F) d A
D
where D is the region enclosed by C. This is a two-dimensional version of the Divergence Theorem discussed in Section
18.3.
SOLUTION Since F = P, Q and F = Q, P, we have
P Q
div(F) = +
x y
P P Q
curl(F ) = (Q) = +
x y x y
Therefore,
2 2
= + 2 9
x 2 y
35. Let F = . Show that curlz (F ) = , where F is the conjugate vector field (defined in Exercises 3234).
1216 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
SOLUTION For a vector field F = P, Q, the conjugate vector field is F = Q, P. By the given information,
F = = , F = ,
x y y x
36. Let n denote the outward-pointing unit normal vector to a simple closed C. The normal derivative of a function ,
denoted , is the directional derivative Dn ( ) = n. Prove the formula
n
ds = d A
C n D
where D is the domain enclosed by C. Hint: Let F = . Show that = F T, where T is the unit tangent vector
n
pointing in the counterclockwise direction along C, and apply Greens Theorem.
SOLUTION In Exercise 32 we showed that for any vector field F, F is a rotation of F by 2 counterclockwise. The unit
tangent en is a rotation of n by 2 counterclockwise.
F*
n
en
q q
F
These properties imply that the angle between F and n is equal to the angle between F and en , and F = F .
Therefore,
Now, if F = , then
= n = F n = F en
n
By the definition of the vector line integral C F ds = C (F en ) ds. Therefore,
ds = (F en ) ds = F ds
C n C C
Using Greens Theorem and the equality curl(F ) = obtained in Exercise 35, we get
ds = F ds = curl(F ) d A = d A.
C n C D D
37. Let P = (a, b) and let C(r ) be the circle of radius r centered at P. The average value of a continuous function on
C(r ) is defined as the integral
2
1
I (r ) = (a + r cos , b + r sin ) d
2 0
(a) Show that
(a + r cos , b + r sin ) = (a + r cos , b + r sin )
n r
(b) Use differentiation under the integral sign to prove that
d 1
I (r ) = ds
dr 2 r C (r ) n
S E C T I O N 18.1 Greens Theorem (ET Section 17.1) 1217
n = cos , sin .
We now differentiate I (r ) under the integral sign, and use (3) and (2) to obtain
2 2
d 1 1
I (r ) = (a + r cos , b + r sin ) d = r (a + r cos , b + r sin ) d
dr 2 0 r 2 0
2
1
= (a + r cos , b + r sin ) d (4)
2 0 n
2
=r (a + r cos , b + r sin ) d (5)
0 n
Combining (4) and (5) we get
d 1
I (r ) = ds.
dr 2 r C (r ) n
38. Prove that m(r ) I (r ) M(r ), where m(r ) and M(r ) are the minimum and maximum values of on C(r ). Then
use the continuity of to prove that lim I (r ) = (P).
r 0
SOLUTION I (r ) is defined by
2
1
I (r ) = (a + r cos , b + r sin ) d
2 0
The points on C(r ) have the form (a + r cos , b + r sin ). Therefore, since m(r ) and M(r ) are the minimum and
maximum values of on C(r ), we have for all 0 2 ,
Now, since is continuous and the functions sin and cos are bounded for all 0 2 , the following holds:
lim (a + r cos , b + r sin ) = lim (a + r cos , b + r sin ) = (a, b)
r 0 r 0
Combining (1), (2), and the Squeeze Theorem, we obtain the following conclusion:
In Exercises 3940, let D be the region bounded by a simple closed curve C. A function (x, y) on D (whose second-
order partial derivatives exist and are continuous) is called harmonic if = 0, where is the Laplace operator
defined in Eq. (9).
39. Use the results of Exercises 37 and 38 to prove the mean-value property of harmonic functions: If is harmonic,
then I (r ) = (P) for all r .
SOLUTION In Exercise 37 we showed that
d 1
I (r ) = d A
dr 2 r D
If is harmonic, = 0. Therefore the right-hand side of the equality is zero, and we get
d
I (r ) = 0
dr
We conclude that I (r ) is constant, that is, I (r ) has the same value for all r . The constant value is determined by the
limit limr 0 I (r ) = (P) obtained in Exercise 38. That is, I (r ) = (P) for all r .
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1219
40. Show that f (x, y) = x 2 y 2 is harmonic. Verify the mean-value property for f (x, y) directly [expand f (a +
r cos , b + r sin ) as a function of and compute I (r )]. Show that x 2 + y 2 is not harmonic and does not satisfy the
mean-value property.
We show that the function f (x, y) = x 2 y 2 is harmonic by showing that f = 2f + 2f = 0. We
2 2
SOLUTION
x y
have
f f
= 2x, = 2y
x y
2 f 2 f
= 2, = 2
x2 y 2
Hence,
2 f 2 f
f = + =22=0
x2 y 2
We now verify the mean-value property for f . That is, we show that for all r ,
2
1
I f (r ) = f (a + r cos , b + r sin ) d = f (a, b)
2 0
We compute the integrand:
f (a + r cos , b + r sin ) = x 2 y 2 = (a + r cos )2 (b + r sin )2
= a 2 + 2ar cos + r 2 cos2 b2 + 2br sin + r 2 sin2
However, we have f (a, b) = a 2 b2 . Hence, for all r , I f (r ) = f (a, b), which proves the mean-value property for f .
For g(x, y) = x 2 + y 2 we have
gx x = 2, g yy = 2, and g = 2 + 2 = 4 = 0.
We check the mean value property:
2 2
1 1
I g (r ) = g(a + r cos , b + r sin ) d = (a + r cos )2 + (b + r sin )2 d
2 0 2 0
2
1
= a 2 + b2 + 2r (a cos + b sin ) + r 2 d = a 2 + b2 + r 2 = a 2 + b2 = (a, b)
2 0
The mean value property does not hold for g.
n n
(A) (B)
FIGURE 13
1220 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
SOLUTION The indicated orientation is defined so that if the normal vector is moving along the boundary curve, the
surface lies to the left. Since the surfaces are oriented by the outward-pointing normal vectors, the induced orientation is
as shown in the figure:
n n
(A) (B)
Hence,
F ds = F ds
S1 S2
Exercises
In Exercises 14, calculate curl(F).
1. F = z y 2 , x + z 3 , y + x 2
SOLUTION We have
i j k
curl(F) = = (1 3z 2 )i (2x 1)j + (1 + 2y)k = 1 3z 2 , 1 2x, 1 + 2y
x y z
z y2 x + z3 y+x
2
y y z
2. F = , ,
x z x
SOLUTION The curl is the following vector:
i j k
curl(F) = = 0 y i z 0 j + 0 1 k = y , z , 1
x y z z2 x2 x z2 x 2 x
y
y z
x z x
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1221
3. F = e y , sin x, cos x
SOLUTION We have
i j k
curl(F) = = 0i ( sin x)j + (cos x e y )k = 0, sin x, cos x e y
x y z
ey sin x cos x
x y
4. F = , , 0
x 2 + y2 x 2 + y2
SOLUTION
i j k
2x y
2x y
curl(F) = x y z = 0i 0j + + k=0
(x 2 + y 2 )
2
(x 2 + y 2 )
2
x y
0
x 2 + y2 x 2 + y2
In Exercises 58, verify Stokes Theorem for the given vector field and surface, oriented with an upward-pointing normal.
Step 1. Compute the line integral around the boundary curve. The boundary curve C is the unit circle oriented in the
counterclockwise direction. We parametrize C by
Then,
F (t) = 2 cos t sin t, cos t, sin t
(t) = sin t, cos t, 0
F (t) (t) = 2 cos t sin t, cos t, sin t sin t, cos t, 0 = 2 cos t sin2 t + cos2 t
Step 2. Compute the flux of the curl through the surface. We parametrize the surface by
( , t) = t cos , t sin , 1 t 2 , 0 t 1, 0 2
Since the normal is always supposed to be pointing upward, the z-coordinate of the normal vector must be positive.
Therefore, the normal vector is
n = 2t 2 cos , 2t 2 sin , t
Hence,
curl(F) n = 1, 0, 1 2t cos 2t 2 cos , 2t 2 sin , t = 2t 2 cos + t 2t 2 cos = t
C3 C2
y
2 C1 3
x
Hence,
F 1 (t) 1 (t) = 0, 0, 2 2t 2, 3, 0 = 0
F 2 (t) 2 (t) = 3t 3t 2 , 0, 0 0, 3, 1 = 0
F 3 (t) 3 (t) = 0, 0, 2t 2, 0, 1 = 2t
Step 3. Compute the flux of the curl through the surface. We parametrize the portion of the plane by
x y
(x, y) = x, y, 1 for (x, y) D
2 3
y
3
3x
y= +3
2
x
0 2
The integrals in (1) and (2) are equal as stated in Stokes Theorem.
7. F = e yz , 0, 0 , the square with vertices (1, 0, 1), (1, 1, 1), (0, 1, 1), and (0, 0, 1)
SOLUTION
Step 1. Compute the integral around the boundary curve. The boundary consists of four segments C1 , C2 , C3 , and C4
shown in the figure:
1224 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
(0, 0, 1)
C4
C1 (0, 1, 1)
S
(1, 0, 1)
C3
C2
(1, 1, 1)
Hence,
F 1 (t) 1 (t) = e1 , 0, 0 1, 0, 0 = e1
F 2 (t) 2 (t) = et1 , 0, 0 0, 1, 0 = 0
F 3 (t) 3 (t) = 1, 0, 0 1, 0, 0 = 1
F 4 (t) 4 (t) = et1 , 0, 0 0, 1, 0 = 0
Step 3. Compute the flux of the curl through the surface. We parametrize the surface by
The upward pointing normal is n = 0, 0, 1. We express curl(F) in terms of the parameters x and y:
curl(F) ((x, y)) = 0, e y1 , e y1
Hence,
curl(F) n = 0, e y1 , e y1 0, 0, 1 = e y1
= 1 + e1 = e1 1 (1)
We parametrize the upper hemisphere by ( , ) = (cos sin , sin sin , cos ), where 0 2 , 0 2 . We
compute the normal to the surface:
We obtain:
/2 2
curl(F) ds = sin sin2 + 3 cos sin d d
S 0 0
/2
2 /2
= 2
sin d sin d + 6 cos sin d
0 0 0
sin2 /2
= 0 + 6 = 3 (1)
2 0
The integrals in (1) and (2) are equal. This verifies Stokes Theorem.
In Exercises 910, use Stokes Theorem to compute the flux of curl(F) through the given surface.
Then also
curl(F) dS = 0
S
10. F = x 2 + y 2 , x + z 2 , 0 , the part of the cone z 2 = x 2 + y 2 such that 2 z 4
SOLUTION
C1
4
2
C2
y
x
The boundary C of the surface consists of the circle C1 : x 2 + y 2 = 16 in the plane z = 4 oriented clockwise, and the
circle C2 : x 2 + y 2 = 4 in the plane z = 2 oriented counterclockwise. We use the parametrizations
11. Let S be the surface of the cylinder (not including the top and bottom) of radius 2 for 1 z 6, oriented with
outward-pointing normal (Figure 14).
(a) Indicate with an arrow the orientation of S (the top and bottom circles).
(b) Verify Stokes Theorem for S and F = yz 2 , 0, 0 .
x y
FIGURE 14
SOLUTION
(a) The induced orientation is defined so that as the normal vector travels along the boundary curve, the surface lies to
its left. Therefore, the boundary circles on top and bottom have opposite orientations, which are shown in the figure.
z
C1
6
n
C2
1
x y
(b) We verify Stokes Theorem for S and F = yz 2 , 0, 0 .
Step 1. Compute the integral around the boundary circles. We use the following parametrizations:
Step 3. Compute the flux of the curl through the surface. We parametrize S by
n = 2 cos , 2 sin , 0
We obtain the following integral (and use the integral we computed before):
6
6 2 6 2
curl(F) dS = 8z sin2 d dz = 8z dz sin2 d = 4z 2 = 140
S 1 0 1 0 1
The line integral and the flux have the same value. This verifies Stokes Theorem.
12. Let S be the portion of the plane z = x contained in the half-cylinder of radius R depicted in Figure 15. Use
Stokes Theorem to calculate the circulation of F = z, x, y + 2z around the boundary of S (a half-ellipse) in the
counterclockwise direction when viewed from above. Hint: Show that curl(F) is orthogonal to the normal vector to the
plane.
z
y
x
FIGURE 15
We conclude that
curl(F) ds = 0dA = 0
S D
This equation, along with Stokes Theorem, allows us to conclude that the circulation of F around C is zero.
2
13. Let I be the flux of F = e y , 2xe x , z 2 through the upper hemisphere S of the unit sphere.
2
(a) Let G = e y , 2xe x , 0 . Find a vector field A such that curl(A) = G.
(b) Use Stokes Theorem to show that the flux of G through S is zero.
Hint: Calculate the circulation of A around S.
(c) Calculate I . Hint: Use (b) to show that I is equal to the flux of 0, 0, z 2 through S.
SOLUTION
(a) We search for a vector field A so that G = curl(A). That is,
A3 A2 A1 A3 A2 A1 2
, , = e y , 2xe x , 0
y z z x x y
We note that the third coordinate ofthis curl vector must be zero; this canbe satisfied if A1 = 0 and A2 = 0. With this
2 2
in mind, we let A = 0, 0, e y e x . The vector field A = 0, 0, e y e x satisfies this equality. Indeed,
A3 A2 A1 A3 2 A2 A1
= ey , = 2xe x , =0
y z z x x y
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1229
2
(b) We found that G = curl(A), where A = 0, 0, e y e x . We compute the flux of G through S. By Stokes Theorem,
G dS = curl(A) dS = A ds
S S C
y
C
x
Therefore,
2
A ds = 0 dt = 0
C 0
2
(c) We rewrite the vector field F = e y , 2xe x , z 2 as
2
2
F = e y , 2xe x , z 2 = e y , 2xe x , 0 + 0, 0, z 2 = curl(A) + 0, 0, z 2
Therefore,
F dS = curl(A) dS + 0, 0, z 2 dS (1)
S S S
In part (b) we showed that the first integral on the right-hand side is zero. Therefore,
F dS = 0, 0, z 2 dS (2)
S S
The upper hemisphere is parametrized by
( , ) = (cos sin , sin sin , cos ), 0 2 , 0 .
2
with the outward pointing normal
n = sin cos sin , sin sin , cos
Therefore,
2 /2 /2
0, 0, z 2 dS = sin cos3 d d = 2 sin cos3 d
S 0 0 0
cos4 /2
= 2 = (0 1) =
4 0 2 2
Combining with (2) we obtain the solution
F dS = .
S 2
1230 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
14. Let F = 0, z, 1 and let S be the spherical cap x 2 + y 2 + z 2 1, where z 12 . Evaluate F dS directly
S
as a surface integral. Then verify that F = curl(A), where A = (0, x, x z) and evaluate the surface integral again using
Stokes Theorem.
SOLUTION
z
1
1
2
C
f0
x y
We first compute the surface integral directly. The spherical cap is parametrized by
( , ) = (cos sin , sin sin , cos ), 0 2 , 0
3
1
2 1
f0
cos f 0 = 1
2
f 0 = p3
We now evaluate the flux using Stokes Theorem. We first notice that F = curl(A), where A = 0, x, x z. We verify it:
A3 A2
=00=0
y z
A1 A3
= 0 z = z
z x
A2 A1
=10=1
x y
Indeed, we see that curl(A) = F. Applying Stokes Theorem, we have
F dS = curl(A) dS = A ds (2)
S S C
3
2
C 1 f0 1
2
x y
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1231
To compute the line integral, we notice that the boundary curve is the circle x 2 + y 2 = 34 in the plane z = 12 . We
parametrize C by
3 3 1
(t) = cos t, sin t, , 0 t 2
2 2 2
S
x
r
(b) Outside the solenoid B is the zero field, hence B = 0 on every domain lying outside the solenoid. Therefore, Stokes
Theorem implies that
A dS = curl(A) dS = B dS = 0 dS = 0.
S S S S
16. The magnetic field B due to a small current loop (which we place at the origin) is called a magnetic dipole (Figure
16). Let = (x 2 + y 2 + z 2 )1/2 . For large, B = curl(A), where
y x
A = 3, 3,0
(a) Let C be a horizontal circle of radius R located far from the origin with center on z-axis. Show that A is tangent to C.
(b) Use Stokes Theorem to calculate the flux of B through C.
y
x Current loop
FIGURE 16
SOLUTION
(a) We parametrize C by
A
z0
R
y
x Current loop
We write A = y3 , x3 , 0 in terms of the parameter t:
R sin t R cos t
A (c(t)) = 3 , , 0
3
We compute the line integral. Since A and c (t) are parallel, we have
R 2 sin2 t R 2 cos2 t
A (c(t)) c (t) = A (c(t)) c (t) = + R 2 sin2 t + R 2 cos2 t
6 6
R R2 R2
= 3 R= 3 =
3/2
R 2 cos2 t + R 2 sin2 t + z 02
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1233
R2 1
= =
z 0 2
3/2 z 0 2
3/2
R 2 cos2 t + sin2 t + R R 1+ R
Hence,
2
dt 2
B dS = A ds = =
S C 0 z 0 2
3/2 z 0 2
3/2
R 1+ R R 1+ R
17. A uniform magnetic field B has constant strength b teslas in the z-direction [i.e., B = 0, 0, b].
(a) Verify that A = 12 B r is a vector potential for B, where r = x, y, 0.
(b) Calculate the flux of B through the rectangle with vertices A, B, C, and D in Figure 17.
FIGURE 17
SOLUTION
SOLUTION Let S1 be the rectangle whose boundary is C, and let S2 denote the wedge-shaped box. Then by Stokes
Theorem,
F ds = curl(F) dS (1)
C S1
and:
F ds = curl(F) dS (2)
C S2
We find the curl of F = x 2 y, x, 0 :
i j k
curl(F) = = 0, 0, 1 + x 2
x y z
x 2 y x 0
We first compute the line integral via the surface integral in (1). To find a parametrization for S1 , we compute the equation
of the plane through A = (6, 0, 4), C = (0, 3, 0), D = (0, 0, 4). A normal to the plane is
AD AC = 6, 0, 0 6, 3, 4 = 6i (6i + 3j 4k) = 18k 24j
= 0, 24, 18 = 6 0, 4, 3
x
0 6
Then
4 4 4 4
= 1, 0, 0 0, 1, = i (j k) = k + j = 0, , 1
x y 3 3 3 3
The upward pointing normal is
4
n = 0, , 1
3
Also,
curl(F) ((x, y)) = 0, 0, 1 + x 2
Hence,
4
curl(F) n = 0, 0, 1 + x 2 0, , 1 = 1 + x 2
3
The integral in (1) is thus
3 6 6
F ds = curl(F) dS = (1 + x 2 ) d A = (1 + x 2 ) d x d y = 3 (1 + x 2 ) d x
C S1 D 0 0 0
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1235
x 3 6
=3 x+ = 234 (3)
3 0
We now compute the line integral via the surface integral (2). The surface S2 consists of two rectangles R1 and R2 and
two triangles T1 and T2 , parametrized by
D3
y
0 3
D4
y
0 3
x+y+z=4
4
4 y
4
x
FIGURE 18
(x, y) = x, y, 4 x y
Hence,
= 1, 0, 1 0, 1, 1 = (i k)(j k) = k + j + i = 1, 1, 1
x y
The normal determined by the induced orientation is
n = 1, 1, 1
Let D be the parameter domain in the parametrization (x, y) = (x, y, 4 x y) of S; that is, D will be the base
triangle in the x y plane that lies underneath the pyramid in the picture. To compute the surface integral in (1) we compute
the values
curl(F) ((x, y)) = 4 2x, 2x 2(4 x y), 2(4 x y) = 4 2x, 8 + 4x + 2y, 8 2x 2y
curl(F) n = 4 2x, 8 + 4x + 2y, 8 2x 2y 1, 1, 1 = 4 2x 8 + 4x + 2y + 8 2x 2y = 4
for any two closed curves lying on a cylinder whose central axis is the z-axis (Figure 19).
C2
C1
y
x
FIGURE 19
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1237
SOLUTION We denote by S the part of the cylinder for which C1 and C2 are boundary curves. Using Stokes Theorem
(notice that C1 and C2 have the same orientations), we have
F ds F ds = curl(F) dS (1)
C1 C2 S
( , z) = R cos , R sin , z
n = R cos , R sin , 0
x y z
21. Let C be the triangular boundary of the portion of the plane + + = 1 lying in the octant x, y, z 0. Use
a b c
Stokes Theorem to find positive constants a, b, c such that the line integral of F = y 2 , 2z + x, 2y 2 around C is zero.
Hint: Choose constants so that curl(F) is orthogonal to the normal vector.
SOLUTION
b y
a
x
origin is
v0 = 3, 1, 4. Estimate the circulation around the small parallelogram
22. The curl of a vector field
F at the
spanned by the vectors A = 0, 12 , 12 and B = 0, 0, 13 .
SOLUTION We use the following approximation, relying on Stokes Theorem:
F ds (v0 en ) A(P) (1)
C
z
(
A = 0, 12 , 1
2 )
(
B = 0, 0, 1
3 )
en
0 y
x
en = 1, 0, 0 (2)
We compute the area of the parallelogram spanned by the vectors O A = 0, 12 , 12 and O B = 0, 0, 13 :
1 1 1 1
O A O B = (j + k) k = (j k + k k) = i
2 3 6 6
1 1
A(P) =
6 i = 6 (3)
We now substitute v0 = 3, 1, 4, (2), and (3) in (1) to obtain the approximation
1 1 1
F ds 3, 1, 4 1, 0, 0 = 3 =
C 6 6 2
1
y
Unit circle
x
FIGURE 20
SOLUTION Since F has a vector potentialthat is, F is the curl of a vector fieldthe flux of F through a surface
depends only on the boundary curve C. Now, the surface S and the unit disc S1 in the x y-plane share the same boundary
C. Therefore,
F dS = F dS (1)
S S1
x
1
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1239
24. Find the flux of F through the surface S in Figure 20, assuming that F has a vector potential and F(x, y, 0) =
cos x, 0, 0.
SOLUTION Since F has a vector potentialthat is, F is the curl of a vector fieldthe flux of F through a surface
depends only on the boundary curve C. Now, the surface S and the unit disc S1 in the x y-plane share the same boundary
C. Therefore,
F dS = F dS (1)
S S1
z
S1
y
C
x
25. Use Eq. (8) to prove that if a is a constant vector, then curl( a) = a.
SOLUTION By Eq. (8) we have
curl( a) = curl(a) + a
Since a is a constant vector, all the partial derivatives of the components of a are zero, hence the curl of a is the zero
vector:
curl(a) = 0
Thus we obtain
curl( a) = a
1240 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
26. A vector field F is called radial if it is of the form F = () x, y, z for some function (), where =
x 2 + y 2 + z 2 . Show that the curl of a radial vector field is zero. Hint: It is enough to show that one component of
the curl is zero, since the calculation for the other two components is similar by symmetry.
SOLUTION Let v = x, y, z. We must show that curl( v) = 0. By Eq. (8) we have
curl( v) = curl(v) + v
Therefore,
curl( v) = v (1)
We now must prove that the cross product is zero. To compute the gradient of , we use the derivatives
2x x
= =
x 2 x +y +z
2 2 2
2y y
= =
y 2 x +y +z
2 2 2
2z z
= =
z 2 x 2 + y2 + z2
v = 0
curl( v) = 0
curl(( )) = 0 7
SOLUTION We have
= , ,
x y z
We compute each component of curl ( ) . The first component is
2 2
= =0
y z z y yz zy
The second component of curl(( )) is
2 2
= =0
z x x z z x xz
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1241
curl(( )) = 0
curl( F) = curl(F) + F 8
SOLUTION We evaluate the curl of F. Since F = F1 , F2 , F3 , using the Product Rule for scalar functions we
have
curl( F) = ( F3 ) ( F2 ), ( F1 ) ( F3 ), ( F2 ) ( F1 )
y z z x x y
F F F F
= F + 3 F 2, F + 1 F 3,
y 3 y z 2 z z 1 z x 3 x
F F
F + 2 F 1
x 2 x y 1 y
F3 F2 F1 F3 F2 F1
= , ,
y z z x x y
+ F3 F2 , F1 F3 , F2 F1 (1)
y z z x x y
The vector in the first term is curl(F). We show that the second term is the cross product F. We compute the cross
product:
i j k
F = = F3 F2 i F3 F1 j + F2 F1 k
x y z y z x z x y
F1 F2 F3
= F3 F2 , F1 F3 , F2 F1
y z z x x y
Therefore,(1) gives
curl( F) = curl(F) + F
29. Assume that the second partial derivatives of and exist and are continuous. Use(7) and (8) to prove that
( ) ds = ( ) ( ) ds
S S
where S is a smooth surface with boundary S.
SOLUTION By Stokes Theorem, we have
( ) ds = curl( ) dS
S S
We now use Eq.(8) to evaluate the curl of . That is,
( ) ds = curl( ) + dS
S S
= curl( ) dS + ( ) ( ) dS (1)
S S
Now, since the gradient field is conservative, this field satisfies the cross-partials condition. In other words,
curl( ) = 0
30. Explain carefully why Greens Theorem is a special case of Stokes Theorem.
SOLUTION Let C be a simple closed curve enclosing a region D oriented counterclockwise in the x y-plane. We must
show, using Stokes Theorem, that for F = F1 , F2 ,
F2 F1
F ds = dA
C D x y
We consider D as a surface in three-space with parametrization
n = 0, 0, 1
z
D y
x C
We compute the curl of F = F1 , F2 , 0. Since F1 = F1 (x, y), F2 = F2 (x, y), and F3 = 0, the curl of F is the vector
F3 F2 F1 F3 F2 F1 F F1
curl(F) = , , = 0, 0, 2
y z z x x y x y
Hence,
F F1 F2 F1
curl(F) n = 0, 0, 2 0, 0, 1 =
x y x y
By Stokes Theorem we have
F2 F1
F ds = curl(F) dS = curl(F) n d A = dA
C D D D x y
We thus showed that Greens Theorem is a special case of Stokes Theorem for two dimensions.
(x, y) y
D C0
x
S E C T I O N 18.2 Stokes Theorem (ET Section 17.2) 1243
Let
Hence,
F3 F3
curl(F) ((x, y)) n = (x, y, f (x, y)) (x, y, f (x, y)) , 0 f x (x, y), f y (x, y), 1
y x
F3 (x, y, f (x, y)) F3 (x, y, f (x, y))
= f x (x, y) + f y (x, y)
y x
The surface integral is thus
F3 (x, y, f (x, y)) F3 (x, y, f (x, y))
curl(F) dS = f x (x, y) + f y (x, y) d x d y (2)
S D y x
We now evaluate the line integral in (1). We have
d
F (t) (t) = 0, 0, F3 x(t), y(t), f x(t), y(t) x (t), y (t), f x(t), y(t)
dt
d
= F3 x(t), y(t), f x(t), y(t) f x(t), y(t) (3)
dt
Using the Chain Rule gives
d
f x(t), y(t) = f x x(t), y(t) x (t) + f y x(t), y(t) y (t)
dt
Substituting in (3), we conclude that the line integral is
b
F ds = F3 x(t), y(t), f x(t), y(t) f x x(t), y(t) x (t) + f y x(t), y(t) y (t) dt (4)
C a
Then the integral in (4) is the line integral of the planar vector field G over C0 . That is,
F ds = G ds
C C0
32. Let F be a continuously differentiable vector field in R3 , Q a point, and S a plane containing Q with unit normal
vector e. Let Cr be a circle of radius r centered at Q in S and let Sr be the disk enclosed by Cr . Assume Sr is oriented
with unit normal vector e.
(a) Let m(r ) and M(r ) be the minimum and maximum values of curl(F(P)) e for P Sr . Prove that
1
m(r ) 2 curl(F) dS M(r )
r Sr
(b) Prove that
1
curl(F(Q)) e = lim F ds
r 0 r 2 Cr
This proves that curl(F(Q)) e is the circulation per unit area in the plane S.
SOLUTION
(a) We may assume that the circle lies on the x y-plane, and parametrize Sr by
Q
en(Q) = e
Sr
By part(a) we have
1
m(r ) F ds M(r ) (2)
r 2 Cr
We take the limit over the circles of radius r centered at Q, as r 0. As r 0, the regions Sr are approaching the
center Q. The continuity of the curl implies that
Therefore,
1
lim m(r ) lim F ds lim M(r )
r 0 r 0 r 2 Cr r 0
1
curl(F)(Q) e lim F ds curl(F)(Q) e
r 0 r 2 Cr
Hence,
1
lim F ds = curl(F)(Q) e
r 0 r 2 Cr
div(F) = 3x 2 + 3y 2 + 3z 2
(b) The curl acts on vector valued functions, and is such a function. Therefore, curl( ) is defined. Since the
gradient field is conservative, the cross partials of are equal, or equivalently, curl( ) is the zero vector.
(c) The curl is defined on vector fields rather than on scalar functions. Therefore, curl( ) is undefined. Obviously,
curl( ) is also undefined.
1246 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
(d) The curl is defined on the vector field F and the divergence is defined on the vector field curl(F). Therefore the
expression div (curl(F)) is meaningful. We show that this vector is automatically zero:
F3 F2 F1 F3 F2 F1
div (curl (F)) = div , ,
y z z x x y
F3 F2 F1 F3 F2 F1
= + +
x y z y z x z x y
2 F3 2 F2 2 F1 2 F3 2 F2 2 F1
= + +
xy xz yz y x z x zy
2
F3 2
F3 2
F2 2
F2 2 F1 2 F1
= + +
xy y x z x xz yz zy
=0+0+0=0
(e) The curl acts on vector valued functions, whereas div(F) is a scalar function. Therefore the expression curl (div(F))
has no meaning.
(f) div(F) is a scalar function, hence (divF) is meaningful. It is not necessarily the zero vector as shown in the follow-
ing example:
F = x 2, y2, z2
div (F) = 2x + 2y + 2z
(divF) = 2, 2, 2 = 0, 0, 0
4. Which of the following statements is correct (where F is a continuously differentiable vector field defined every-
where)?
(a) The flux of curl(F) through all surfaces is zero.
(b) If F = , then the flux of F through all surfaces is zero.
(c) The flux of curl(F) through all closed surfaces is zero.
SOLUTION
(a) This statement holds only for conservative
fields. If F is not conservative, there exist closed curves such that C F
ds = 0, hence by Stokes Theorem S curl(F) dS = 0.
(b) This statement is false. Consider the unit sphere S in the three-dimensional space and the function (x, y, z) =
x 2 + y 2 + z 2 . Then F = = 2x, 2y, 2z and div (F) = 2 + 2 + 2 = 6. Using the Divergence Theorem, we have (W
is the unit ball in R 3 )
F dS = div(F) d V = 6 dV = 6 d V = 6 Vol(W)
S W W W
(c) This statement is correct, as stated in the corollary of Stokes Theorem in section 18.2.
5. How does the Divergence Theorem imply that the flux of F = x 2 , y e z , y 2zx through a closed surface is equal
to the enclosed volume?
SOLUTION By the Divergence Theorem, the flux is
F dS = div(F) d V = (2x + 1 2x) d V = 1 d V = Volume(W)
S W W W
Therefore the statement is true.
Exercises
In Exercises 14, compute the divergence of the vector field.
1. F = x y, yz, y 2 x 3
SOLUTION The divergence of F is
div(F) = (x y) + (yz) + (y 2 x 3 ) = y + z + 0 = y + z
x y z
2. xi + yj + zk
S E C T I O N 18.3 Divergence Theorem (ET Section 17.3) 1247
SOLUTION
div(F) = (x) + (y) + (z) = 1 + 1 + 1 = 3
x y z
3. F = x 2zx 2 , z x y, z 2 x 2
SOLUTION
div(F) = (x 2zx 2 ) + (z x y) + (z 2 x 2 ) = (1 4zx) + (x) + (2zx 2 ) = 1 4zx x + 2zx 2
x y z
4. sin(x + z)i ye x z k
SOLUTION
div(F) = sin(x + z) + (ye x z ) = cos(x + z) yxe x z
x z
In Exercises 58, verify the Divergence Theorem for the vector field and region.
2 y
Then,
3 4 3 4
F dS = F (1 (x, z)) 0, 1, 0 d x dz = z, x, 0 0, 1, 0 d x dz
S1 0 0 0 0
1248 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
3 4
x 2 4
= x d x dz = 3 = 24
0 0 2 0
3 2 3 2
F dS = F (2 (y, z)) 1, 0, 0 d y dz = z, 0, y 1, 0, 0 d y dz
S2 0 0 0 0
3 2
z 2 3
= z d y dz = 2 = 9
0 0 2 0
3 4 3 4
F dS = F (3 (x, z)) 0, 1, 0 d x dz = z, x, 2 0, 1, 0 d x dz
S3 0 0 0 0
3 4
x 2 4
= x d x dz = 3 = 24
0 0 2 0
3 2 3 2
F dS = F (4 (y, z)) 1, 0, 0 d y dz = z, 4, y 1, 0, 0 d y dz
S4 0 0 0 0
3 2
z 2 3
= z d y dz = 2 =9
0 0 2 0
2 4 2 4
F dS = F (5 (x, y)) 0, 0, 1 d x d y = 0, x, y 0, 0, 1 d x d y
S5 0 0 0 0
2 4
y 2 2
= y d x d y = 4 = 8
0 0 2 0
2 4 2 4
F dS = F (6 (x, y)) n d x d y = 3, x, y 0, 0, 1 d x d y
S6 0 0 0 0
2 4
y 2 2
= y dx dy = 4 =8
0 0 2 0
We now evaluate the triple integral in (1). We compute the divergence of F = z, x, y:
div(F) = (z) + (x) + (y) = 0
x y z
Hence,
div(F) d V = 0 dV = 0 (3)
R R
The equality of the integrals in (2) and (3) verifies the Divergence Theorem.
6. F = y, x, z and the region x 2 + y 2 + z 2 4
SOLUTION Let S be the surface of the sphere and R the ball enclosed by S. We compute both sides of the Divergence
Theorem:
F dS = div(F) d V (1)
S R
Step 1. Integral over sphere. We use the parametrization
Then,
F ( , ) n = 2 sin sin , cos sin , cos 4 sin cos sin , sin sin , cos
= 8 sin cos sin3 + cos sin sin3 + cos2 sin
y
1
Let S be the surface of the cylinder and R the region enclosed by S. We compute the two sides of the Divergence
Theorem:
F dS = div(F) d V (1)
S R
We first calculate the surface integral.
Step 1. Integral over the side of the cylinder. The side of the cylinder is parametrized by
Then,
Step 2. Integral over the top of the cylinder. The top of the cylinder is parametrized by
(x, y) = (x, y, 2)
!
with parameter domain D = (x, y) : x 2 + y 2 1 . The upward pointing normal is
Also,
Hence,
F dS = 3y d A = 0
top D
D
x
1
Step 3. Integral over the bottom of the cylinder. We parametrize the bottom by
Let S be the surface enclosing the given region R. We must verify the equality
F dS = div(F) d V (1)
S R
We first compute the surface integral on the left-hand side.
Step 1. Integral over the side of the surface. The side of the surface is parametrized by
( , t) = t cos , t sin , t 2 , 0 t 2, 0 2
S E C T I O N 18.3 Divergence Theorem (ET Section 17.3) 1251
Also,
F (( , t)) n = t cos , 0, 0 2t 2 cos , 2t 2 sin , t = 2t 3 cos2
Also,
Hence,
F dS = 0dA = 0 (3)
top D
x
2
D
div(F) = (x) + (y) + (z) = 1 + 1 + 1 = 3
x y z
Using the Divergence Theorem and the volume of the sphere, we obtain
F dS = div(F) d V = 3 dV = 3 1 d V = 3 Volume (W)
S W W W
4
= 3 13 = 4
3
div(F) = (y) + (z) + (x) = 0 + 0 + 0 = 0
x y z
Hence, by the Divergence Theorem (W is the unit ball),
F dS = div(F) d V = 0 dV = 0
S W W
11. F = x 3 , 0, z 3 , S is the sphere x 2 + y 2 + z 2 = 4.
SOLUTION We compute the divergence of F = x 3 , 0, z 3 :
3
div(F) = (x ) + (0) + (z 3 ) = 3x 2 + 3z 2 = 3(x 2 + z 2 )
x y z
Using the Divergence Theorem we obtain (W is the region inside the sphere)
F dS = div(F) d V = 3(x 2 + z 2 ) d V
S W W
We convert the integral to spherical coordinates. We have
x y
We compute the divergence of F = x, y 2 , z + y :
2
div(F) = (x) + (y ) + (z + y) = 1 + 2y + 1 = 2 + 2y.
x y z
By the Divergence Theorem we have
F dS = div(F) d V = (2 + 2y) d V
S W W
We compute the triple integral. Denoting by D the disk x 2 + y 2 4 in the x y-plane, we have
8 8
F dS = (2 + 2y) dz d x d y = (2 + 2y)z dx dy = (2 + 2y)(8 x) d x d y
S D x D z=x D
2
14. F = x 2 z 2 , e z cos x, y 3 , S is the boundary of the region bounded by x + 2y + 4z = 12 and the coordinate
planes in the first octant.
2
SOLUTION We compute the divergence of F = x 2 z 2 , e z cos x, y 3 :
2 z2
div(F) = (x z 2 ) + e cos x + (y 3 ) = 2x.
x y z
By the Divergence Theorem,
F dS = div(F) d V = 2x d V .
S W W
1254 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
6
y
12
6
y = 12 x + 6
x
12
Thus,
12 x/2+6 3 y x 12 x/2+6 3 y x
2 4 2 4
F dS = 2x dz d y d x = 2x z dy dx
S 0 0 0 0 0 z=0
12 x/2+6
12
y x
y2 x y x/2+6
= 2x 3 dy dx = 2x 3y dx
0 0 2 4 0 4 4 y=0
12 2
x
x
6 x2
= 2x 3 6 dx
0 4 2 4
We let u = 6 x2 and du = 12 d x:
12 2 6
6 x2 2 3 1 4 6
2x dx = 2(6 u)u du = 4u u = 216.
0 4 0 2 0
15. F = x + y, z, z x, S is the boundary of the region between the paraboloid z = 9 x 2 y 2 and the x y-plane.
SOLUTION We compute the divergence of F = x + y, z, z x,
div(F) = (x + y) + (z) + (z x) = 1 + 0 + 1 = 2.
x y z
z
9
y
3
x 2 + y2 = 9
x
3
x = r cos , y = r sin , 0 r 3, 0 2
2 3 3
9r 2 r 4 3
F dS = 2 9 r 2 r dr d = 4 (9r r 3 ) dr = 4 = 81
S 0 0 0 2 4 0
2
16. F = e z , sin(x 2 z), x 2 + 9y 2 , S is the region x 2 + y 2 z 8 x 2 y 2 .
SOLUTION We compute the divergence of F:
z2
div(F) = (e ) + sin(x 2 z) + x 2 + y2 = 0
x y z
The Divergence Theorem implies that
F dS = div(F) d V = 0
S W
17. Let W be the region in Figure 16 bounded by the cylinder x 2 + y 2 = 9, the plane z = x + 1, and the x y-plane. Use
the Divergence Theorem to compute the flux of F = z, x, y + 2z through the boundary of W.
y
FIGURE 16
div(F) = (z) + (x) + (y + 2z) = 2
x y z
By the Divergence Theorem we have
F dS = div(F) d V = 2 dV
S W W
To compute the triple integral, we identify the projection D of the region on the x y-plane. D is the region in the x y plane
enclosed by the circle x 2 + y 2 = 9 and the line 0 = x + 1 or x = 1. We obtain the following integral:
x+1 x+1
F dS = 2 dV = 2 dz d x d y = 2z dx dy = (2x + 2) d x d y
S W D 0 D z=0 D
We compute the double integral as the difference of two integrals: the integral over the disk D2 of radius 3, and the
integral over the part D1 of the disk, shown in the figure.
1256 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
D1
x
3
We obtain
F dS = (2x + 2) d x d y (2x + 2) d x d y
S D2 D1
= 2x d x d y + 2 dx dy (2x + 2) d x d y
D2 D2 D1
D2
x
3
The first integral is zero due to symmetry. The second integral is twice the area of D2 , that is, 2 32 = 18 . Therefore,
F dS = 18 (2x + 2) d x d y
S D1
We compute the double integral over the upper part of D1 . Due to symmetry, this integral is equal to half of the integral
over D1 .
y
q
x
1 3
div(F) = (cx y) + (y z) + (3x + 4cz) = c + 1 + 4c = 5c + 1
x y z
Therefore, div(F) = 0 if 5c + 1 = 0 or c = 15 .
19. Volume as a Surface Integral Let F = x, y, z. Prove that if W is a region R3 with a smooth boundary S, then
1
Volume(W) = F dS 9
3 S
We compute the surface integral of F over S, using the Divergence Theorem. Since div(F) = x (x) + y (y) + z
(z) =
3, we get
F dS = div(F) d V = 3 dV = 3 1 dV (2)
S W W W
We combine (1) and (2) to obtain
F dS = 3 volume(W)
S
or
1
Volume(W) = F dS
3 S
20. Use Eq. (9) to calculate the volume of the unit ball as a surface integral over the unit sphere.
SOLUTION Let S be the unit sphere and W is the unit ball. By Eq. (9) we have
1
Volume(W) = F dS, F = x, y, z
3 S
To compute the surface integral, we parametrize S by
Then
F ( , ) n = cos sin , sin sin , cos cos sin2 , sin sin2 , cos sin
= cos2 sin3 + sin2 sin3 + cos2 sin = sin3 (cos2 + sin2 ) + cos2 sin
21. Show that (a cos sin , b sin sin , c cos ) is a parametrization of the ellipsoid
x
2 y
2 z
2
+ + =1
a b c
Then use Eq. (9) to calculate the volume of the ellipsoid as a surface integral over its boundary.
SOLUTION For the given parametrization,
= sin2 + cos2 = 1
Conversely, for each (x, y, z) on the ellipsoid, there exists and so that (1) holds. Therefore ( , ) parametrizes the
whole ellipsoid. Let W be the interior of the ellipsoid S. Then by Eq. (9):
1
Volume(W) = F dS, F = x, y, z
3 S
We compute the surface integral, using the given parametrization. We first compute the normal vector:
= a sin sin , b cos sin , 0
= a cos cos , b sin cos , c sin
= ab sin2 sin cos k ac sin sin2 j ab cos2 sin cos k bc cos sin2 i
= bc cos sin2 , ac sin sin2 , ab sin cos
= abc sin3 + abc sin cos2 = abc sin3 + abc sin (1 sin2 )
= abc sin
div(curl(F)) = 0
SOLUTION Let F = P(x, y, z), Q(x, y, z), R(x, y, z). We compute the curl of F:
i j k
R Q P R Q P
curlF = = , ,
x y z y z z x x y
P Q R
S E C T I O N 18.3 Divergence Theorem (ET Section 17.3) 1259
Since the second-order partials are continuous, the mixed partials are equal. Therefore,
div (curl(F)) = 0
23. Find and prove a Product Rule expressing div( f F) in terms of div(F) and f .
SOLUTION Let F = P, Q, R. We compute div( f F):
div( f F) = div f P, f Q, f R = ( f P) + ( f Q) + ( f R)
x y z
Applying the product rule for scalar functions we obtain
P f Q f R f
div( f F) = f + P + f + Q + f + R
x x y y z z
P Q R f f f
= f + + + P+ Q+ R = f div(F) + F f
x y z x y z
We thus proved the following identity:
div( f F) = f div(F) + F f
Then prove that the cross product of two irrotational vector fields is incompressible [F is called irrotational if curl(F) = 0
and incompressible if div(F) = 0].
SOLUTION We compute the left-hand side of the identity. For F = P, Q, R and G = S, T, U we have
i j k
F G = P Q R = (QU RT )i (PU RS)j + (P T Q S)k
S T U
div(F G) = (QU RT ) (PU RS) + (P T Q S)
x y z
= (Q x U + QUx Rx T RTx ) Py U + PU y R y S RS y + (Pz T + P Tz Q z S Q Sz )
= S(R y Q z ) + T (Pz Rx ) + U (Q x Py ) P(U y Tz ) Q(Sz Ux ) R(Tx S y )
Thus,
G curl(F) F curl(G) = S, T, U R y Q z , Pz Rx , Q x Py
P, Q, R U y Tz , Sz Ux , Tx S y
= S(R y Q z ) + T (Pz Rx ) + U (Q x Py )
P(U y Tz ) Q(Sz Ux ) R(Tx S y )
Thus, if both F and G are irrotational (that is, with curl zero), then their cross product is source-free (that is, with
divergence zero), as div F G = G 0 F 0 = 0.
1260 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
We now compute the divergence of this vector. Using the Product Rule for scalar functions and the equality of the mixed
partials, we obtain
div( f g) = ( f y gz f z g y ) + ( f z g x f x gz ) + ( f x g y f y g x )
x y z
= f yx gz + f y gzx f zx g y f z g yx + f zy gx + f z gx y f x y gz f x gzy + f x z g y + f x g yz
f yz gx f y gx z
= ( f yx f x y )gz + (gzx gx z ) f y + ( f x z f zx )g y + (gx y g yx ) f z
+ ( f zy f yz )gx + (g yz gzy ) f x = 0
2 2 2
= + 2 + 2
x 2 y z
curl(curl(F)) = (div(F)) F
Hence,
F2 F1 F1 F3 F3 F2 F2 F1
curl (curl(F)) = , ,
y x y z z x z y z x x y
F1 F3 F3 F2
x z x y y z
2 F2 2 F1 2 F1 2 F3 2 F3 2 F2 2 F2 2 F1
= + , + ,
y x y 2 z 2 z x zy z 2 x 2 xy
2 F1 2 F3 2 F3 2 F2
+ (1)
xz x2 y 2 yz
Therefore,
2 F1 2 F1 2 F1 2 F2 2 F2 2 F2 2 F3 2 F3 2 F3
(div(F)) F = (div(F)) + + , + + , + +
x2 y 2 z 2 x 2 y 2 z 2 x 2 y 2 z 2
S E C T I O N 18.3 Divergence Theorem (ET Section 17.3) 1261
2 F2 2 F3 2 F1 2 F1 2 F1 2 F3 2 F2 2 F2
= + , + ,
xy xz y 2 z 2 y x yz x 2 z 2
2 F1 2 F2 2 F3 2 F3
+ (2)
z x zy x2 y 2
Since the mixed partials are equal, the expressions obtained in (1) and (2) are the same. This proves the desired identity.
27. A function satisfying = 0 is called harmonic.
(a) Show that = div( ) for any function .
(b) Show that is harmonic if and only if div( ) = 0.
(c) Show that if F is the gradient of a harmonic function, then curl(F) = 0 and div(F) = 0.
1
(d) Show F = x z, yz, (x 2 y 2 ) is the gradient of a harmonic function. What is the flux of F through a closed
2
surface?
SOLUTION
(a) We compute the divergence of :
2 2 2
div( ) = div , , = + 2 + 2 =
x y z x 2 y z
(b) In part (a) we showed that = div( ). Therefore = 0 if and only if div( ) = 0. That is, is harmonic if
and only if is divergence free.
(c) We are given that F = , where = 0. In part (b) we showed that
div(F) = div( ) = 0
2
= xz =z
x x2
2
= yz = z
y y 2
x 2 y2 2
= =0
z 2 z 2
Therefore,
2 2 2
= + 2 + 2 = zz+0=0
x 2 y z
Since F is the gradient of a harmonic function, we know by part (c) that div(F) = 0. Therefore, by the Divergence
Theorem, the flux of F through a closed surface is zero:
F dS = div(F) d V = 0 dV = 0
S W W
28. Let F = n er , where n is any number, = (x 2 + y 2 + z 2 )1/2 , and er = 1 x, y, z is the unit radial vector.
1262 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
Hence,
F1 2 (n1)/2
n 1 (n3)/2 (n1)/2
= (x + y 2 + z 2 ) x = (x 2 + y 2 + z 2 ) 2x x + (x 2 + y 2 + z 2 )
x x 2
(n3)/2
(n3)/2
= (x 2 + y 2 + z 2 ) (n 1)x 2 + x 2 + y 2 + z 2 = (x 2 + y 2 + z 2 ) (nx 2 + y 2 + z 2 )
Similarly,
F2 2 (n1)/2
(n3)/2 2
= (x + y 2 + z 2 ) y = (x 2 + y 2 + z 2 ) (x + ny 2 + z 2 )
y y
and
F3 2 (n1)/2
(n3)/2 2
= (x + y 2 + z 2 ) z = (x 2 + y 2 + z 2 ) (x + y 2 + nz 2 )
z z
The divergence of F is the sum
(n3)/2
div(F) = (x 2 + y 2 + z 2 ) = (nx 2 + y 2 + z 2 + x 2 + ny 2 + z 2 + x 2 + y 2 + nz 2 )
(n3)/2 (n1)/2
= (x 2 + y 2 + z 2 ) (n + 2)(x 2 + y 2 + z 2 ) = (n + 2)(x 2 + y 2 + z 2 ) = (n + 2)n1
(b) Let W denote the ball inside the sphere of radius R. We may apply the Divergence Theorem when the components
of F in (1) are defined and have continuous derivatives in W, that is, when n1 n3
2 0 and 2 0 or when n 3. In this
case, we have
F dS = div(F) d V = (n + 2)n1 d V
S W W
We compute the triple integral by converting it to spherical coordinates. We obtain
2 R R
F dS = (n + 2)n1 2 sin d d d = 2 (n + 2) n+1 sin d d
S 0 0 0 0 0
R R
= 2 (n + 2) n+1 d sin d = 2 (n + 2) cos n+1 d
0 0 =0 0
R
n+2 R R n+2
= 4 (n + 2) n+1 d = 4 (n + 2) = 4 (n + 2) = 4 R n+2
0 n + 2 =0 n+2
That is,
F dS = 4 R n+2 , n 3.
S
We now consider the case n < 3. We evaluate the surface integral directly, using the parametrization, ( , ) =
(R cos sin , R sin sin , R cos ), 0 2 , 0 .
n = R 2 sin er
Then,
F ( , ) n = n er R 2 sin er = n R 2 sin
S E C T I O N 18.3 Divergence Theorem (ET Section 17.3) 1263
n/2
= R 2 sin R 2 cos2 sin2 + R 2 sin2 sin2 + R 2 cos2 = R n+2 sin
Hence,
2
F dS = R n+2 sin d d = 2 R n+2 cos = 4 R n+2
S 0 0 0
Similarly,
y z
= and =
y z
Therefore,
n x
( ) = n n1 = n n1 = n n2 x
x x
n y
( ) = n n1 = n n1 = n n2 y
y y
n z
( ) = n n1 = n n1 = n n2 z
z z
Therefore, if n = 1, then
n+1
F=
n+1
We now show that F = 1 er is also a gradient vector field. We compute the gradient of ln . Using the Chain Rule and
the partial derivatives (1), we have
1 1x x
(ln ) = = = 2
x x
1 1y y
(ln ) = = = 2
y y
1 1z z
(ln ) = = = 2
z z
Therefore, the gradient of ln is
x y z
(ln ) = , , = 1 1 x, y, z = 1 er = F.
2 2 2
We conclude that F is a gradient vector for all values of n.
(e) Since F is a gradient vector field, it is conservative. Therefore, the line integral of F over a closed curve is zero.
(f) Using Exercise 27 part (b), = n is harmonic if and only if
div (n ) = 0
1264 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
or
div(n1 er ) = 0 or n=0
div(n1 er ) = (n + 1)n2 = 0 n = 1 or n = 0.
SOLUTION
(a) We first compute the partial derivatives of :
1 1/2 x
= (x 2 + y 2 + z 2 ) 2x =
x 2
1 1/2 y
= (x 2 + y 2 + z 2 ) 2y =
y 2
1 1/2 z
= (x 2 + y 2 + z 2 ) 2z = (1)
z 2
We compute the partial derivatives of z3 , using the Chain Rule and the partial derivatives in (1):
z 3 x 3zx
=z ( ) = z (3)4 = 3z 4 = 5 = 3z 5 x
x 3 x x
z 3 y
=z ( ) = z (3)4 = 3z 4 = 3z 5 y
y 3 y y
z z
= (z 3 ) = 1 3 + z (3)4 = 3 3z 4 = 3 3z 2 5
z 3 z z
Therefore,
z
E= = 3z 5 xi 3z 5 yj + (3 3z 2 5 )k
3
= 3 k 3z 4 1 (xi + yj + zk) = 3 k 3z 4 er
(b) To compute the flux S E dS we use the parametrization ( , ) = (R cos sin , R sin sin , R cos ), 0
2 , 0 :
n = R 2 sin er
S E C T I O N 18.3 Divergence Theorem (ET Section 17.3) 1265
We compute E ( , ) n. Since = R on S, we get
E ( , ) n = R 3 k 3z R 4 er R 2 sin er = R 1 sin k er 3z R 2 sin
Hence,
2
2
E dS = R 1 sin 2 d d = sin 2 d = cos 2 =0
S 0 0 R 0 R =0
To find div(E) we compute the following derivatives, using (1) and the laws of differentiation. This gives
(3z 5 x) = 3z (5 x) = 3z 56 x + 5 1
x x x
x
= 3z 56 x + 5 = 3z 7 (5x 2 2 )
Similarly,
(3z 5 y) = 3z 7 (5y 2 2 )
y
and
(3z 2 5 + 3 ) = 6z 5 3z 2 (5)6 34
z z z
z z
= 6z 5 + 15z 2 6 34 = 3z 7 (5z 2 32 )
Hence,
= 15z 7 (2 2 ) = 0
(d) Since E is not defined at the origin, which is inside the ball W, we cannot use the Divergence Theorem to compute
the flux of E through the sphere.
30. Let E be the electric field due to a long, uniformly charged rod of radius R with charge density per unit length
(Figure 18). By symmetry, we may assume that E is everywhere perpendicular to the rod and its magnitude E(r ) depends
only on the distance r to the rod (strictly speaking, this would hold only if the rod were infinite, but it is nearly true if the
rod is long enough). Show that E(r ) = for r > R. Hint: Apply Gausss Law to a cylinder of radius R and of unit
2 0 r
length with its axis along the rod.
r
R
y
x Charged rod
FIGURE 18
SOLUTION Gausss Theorem asserts that if S is a closed surface, then the total charge Q enclosed by S is given by
Q= 0 E dS
S
1266 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
where 0 is the dielectric coefficient in vacuum. Here, S is a closed cylinder of radius r and unit length that encloses a
charge Q = 1 = . Therefore, by Gausss Theorem, for r > R,
= 0 E dS = 0 E dS + 0 E dS
S integral cylinder bases
E is perpendicular to the normal vectors of the two bases, hence the second integral is zero, and we have
2 1
= 0 E dS = 0 E n d y d (1)
S 0 0
r
R
1
Since E and n point in the same direction and E = E(r ), n = r , we have
E n = En = E(r ) r
Hence, by (1),
2 1 1
= 0 r E(r ) d y d = 2 0 r E(r ) d y = 2 0 r E(r )
0 0 0
or
E(r ) = .
2 0 r
31. Let I = F dS, where
S
2yz xz xy
F= , 2, 2
2
( = x 2 + y 2 + z 2 ) and S is the boundary of a region W.
(a) Check that F is divergence-free.
(b) Show that I = 0 if S is a sphere centered at the origin. Note that the Divergence Theorem cannot be used. Why not?
(c) Give an argument showing that I = 0 for all S.
SOLUTION
(a) To find div(F), we first compute the partial derivatives of = x 2 + y2 + z2:
2x x 2y y 2z z
= = , = = , = =
x 2 x +y +z
2 2 2 y 2 x +y +z
2 2 2 z 2 x +y +z
2 2 2
div(F) = 4x yz 4 + 2x yz 4 + 2x yz 4 = 0
The Divergence Theorem cannot be used since F is not defined at the origin, which is inside the ball with the boundary
S.
(c) In part (b) we showed that F n = 0 for all values of R. Therefore, I = 0 over all spheres centered at the origin.
SOLUTION
(a) By the theorem on evaluating directional derivatives, Den = en , hence,
Den d S = en d S (1)
S S
By the definition of the vector surface integral, we have
dS = ( en ) d S
S S
Combining with (1) gives
Den d S = dS
S S
We now apply the Divergence Theorem and the identity div( ) = shown in part (a) of Exercise 27, to write
Den d S = dS = div( ) d V = d V
S S W W
(b) If is harmonic, then = 0; therefore, by the equality of part (a) we have
Den d S = d V = 0 d V = 0.
S W W
33. Assume that is harmonic. Show that div( ) = 2 and conclude that
Den d S = 2 d V
S W
div( ) = (2)
1268 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
= 0 (3)
div( ) = 2 + 0 = 2 (4)
Den = en
Hence,
Den d S = ( en ) d S (5)
S S
By the definition of the vector surface integral we have
dS = ( en ) d S (6)
S S
Combining (5) and (6) and using the Divergence Theorem and equality (4), we get
Den d S = dS = div( ) d V = 2 d V
S S W W
34. Let F = P, Q, R be a vector field defined on R3 such that div(F) = 0. Use the following steps to show that F has
a vector potential.
(a) Let A = f, 0, g. Show that
g f g f
curl(A) = , ,
y z x y
(b) Fix any value y0 and show that if we define
y
f (x, y, z) = R(x, t, z) dt + (x, z)
y0
y
g(x, y, z) = P(x, t, z) dt + (x, z)
y0
div(F) = Px + Q y + Rz = 0 Px + Rz = Q y (3)
potential A, which is
A = f, 0, g (1)
y z
f (x, y, z) = R(x, t, z) dt + Q(x, y0 , t) dt
y0 z0
y
g(x, y, z) = P(x, t, z) dt
y0
36. Show that F = 2ye z x y, y, yz z has a vector potential and find one.
SOLUTION As shown in Exercise 34, if F is divergence free, then F has a vector potential. We show that div(F) = 0:
div(F) = (2ye z x y) + (y) + (yz z) = y + 1 + y 1 = 0
x y z
We find a vector potential A, using the result in Exercise 34:
A = f, 0, g (1)
Using z 0 = 0, we have
y z
f (x, y, z) = R(x, t, z) dt + Q(x, y0 , t) dt
y0 0
1270 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
y
g(x, y, z) = P(x, t, z) dt
y0
Hence, P(x, y, z) = 2ye z x y, Q(x, y, z) = y, and R(x, y, z) = yz z. We choose y0 = 0 and compute the functions
f and g:
y z
t2z y y 2z y 2
f (x, y, z) = (t z z) dt + 0 dt = zt = zy =z y
0 0 2 t=0 2 2
y
xt 2 y x y2 x
g(x, y, z) = (2te z xt) dt = t 2 e z = y 2 ez = y 2 ez
0 2 t=0 2 2
37. A vector field with a vector potential has zero flux through every closed surface in its domain. In the text, we observed
er
that although the inverse-square radial vector field F = 2 satisfies div(F) = 0, F cannot have a vector potential on its
domain {(x, y, z) = (0, 0, 0)} because the flux of F through a sphere containing the origin is nonzero.
(a) Show that the method of Exercise 34 produces a vector potential A such that F = curl(A) on the restricted domain
D consisting of R3 with the y-axis removed.
(b) Show that F also has a vector potential on the domains obtained by removing either the x-axis or the z-axis from R3 .
(c) Does the existence of a vector potential on these restricted domains contradict the fact that the flux of F through a
sphere containing the origin is nonzero?
SOLUTION
<x,y,z>
(a) We have F(x, y, z) = er2 = 3/2 , hence
(x 2 +y 2 +z 2 )
x
P(x, y, z) = 3/2
(x 2 + y 2 + z 2 )
y
Q(x, y, z) = 3/2
(x 2 + y 2 + z 2 )
z
R(x, y, z) = 3/2
(x 2 + y 2 + z 2 )
In Exercise 34, we defined the functions (taking y0 = z 0 = 0)
y z y
z z
f (x, y, z) = dt + Q(x, 0, t) dt = dt
0 (x 2 + t 2 + z 2 )3/2 0 0 (x 2 + t 2 + z 2 )3/2
y
x
g(x, y, z) = dt
0 (x + t + z 2 )3/2
2 2
These functions are defined for (x, z) = (0, 0), since the points with x = 0 and z = 0 are on the y-axis. (Notice that for
any fixed (x, z) = (0, 0) the interval of integration do not intersect the y-axis, therefore they are contained in the domain
D.) For (x, z) = (0, 0) we have by the Fundamental Theorem of Calculus
y
g x x
= dt = = P(x, y, z)
y y 0 (x 2 + t 2 + z 2 )3/2 (x 2 + y 2 + z 2 )
3/2
y 2 3/2 1/2
f g (x + t 2 + z 2 ) z 32 (x 2 + t 2 + z 2 ) 2z
= dt
z x 0 (x 2 + t 2 + z 2 )
3
y 2 3/2 1/2
(x + t 2 + z 2 ) x 32 (x 2 + t 2 + z 2 ) 2x
3
dt
0 (x 2 + t 2 + z 2 )
y 2 1/2
(x + t 2 + z 2 ) (x 2 + t 2 + z 2 3z 2 )
= 3
dt
0 (x 2 + t 2 + z 2 )
y 2 1/2
(x + t 2 + z 2 ) (x 2 + t 2 + z 2 3x 2 )
3
dt
0 (x 2 + t 2 + z 2 )
Chapter Review Exercises 1271
y 2
x + t 2 2z 2 + t 2 + z 2 2x 2
= 5/2
dt
0 (x 2 + t 2 + z 2 )
y
x 2 2t 2 + z 2 y
= 5/2
dt = 3/2
= Q(x, y, z)
0 (x 2 + t 2 + z 2 ) (x 2 + y 2 + z 2 )
The last integral can be verified by showing that
y x 2 2y 2 + z 2
=
y (x 2 + y 2 + z 2 )3/2 (x 2 + y 2 + z 2 )
5/2
and
y
f z z
= dt = = R(x, y, z)
y y 0 (x 2 + t 2 + z 2 )3/2 (x + y + z 2 )
2 2 3/2
A = 0, f, g
x y
g(x, y, z) = Q(t, y, z)dt + P(x 0 , t, z) dt
x0 y0
x
f (x, y, z) = R(t, y, z) dt
x0
Using similar procedure to that in Exercise 34, one can show that
F = curl(A).
In removing the z-axis the proof is similar, with corresponding modifications of the functions in Exercise 34.
(c) The ball inside any sphere containing the origin must intersect the x, y, and z axes; therefore, F does not have a
vector potential in the ball, and the flux of F through the sphere may differ from zero, as in our example.
We now compute the integral using Greens Theorem. We compute the curl of F. Since P = x + y 2 and Q = x 2 y,
we have
Q P
= 2x 2y
x y
1272 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
Thus,
F ds = (2x 2y) d x d y
C D
y
D
x
1
2. Let R be the boundary of the rectangle in Figure 1 and let R1 and R2 be the boundaries of the two triangles, all
oriented counterclockwise.
(a) Determine F ds if F ds = 4 and F ds = 2.
R1 R R2
y
Rectangle R
R1
R2
x
FIGURE 1
SOLUTION
(a) Since all boundaries are oriented counterclockwise, the segment D B is oriented in opposite directions as part of the
boundaries R1 and R2 .
y
R
D C
R1
R2
A B
x
Therefore, the contributions of this segment to the sum of the line integrals over R1 and R2 cancel each other and the
following equality holds:
F ds = F ds + F ds
R R1 R2
(b) Reversing the orientation of the curve gives the opposite integral. Therefore if R is oriented clockwise, the line
integral is the opposite of 4; that is,
F ds = 4.
R
Chapter Review Exercises 1273
In Exercises 36, use Greens Theorem to evaluate the line integral around the given closed curve.
3. x y 3 d x + x 3 y d y, where C is the rectangle 1 x 2, 2 y 3, oriented counterclockwise.
C
SOLUTION
3 C
x
1 2
4. (3x + 5y cos y) d x + x sin y d y, where C is any closed curve enclosing a region with area 4, oriented counter-
C
clockwise.
SOLUTION The components of F are P = 3x + 5y cos y and Q = x sin y. Therefore the curl of F is
Q P
= sin y (5 + sin y) = 5
x y
Using Greens Theorem we obtain
Q P
P dx + Q dy = dA = (5) d A
C D x y D
= 5 1 d A = 5Area( A) = 5 4 = 20
D
5. y 2 d x x 2 d y, where C consists of the arcs y = x 2 and y = x, 0 x 1, oriented clockwise.
C
SOLUTION We compute the curl of F.
C
y= x
y = x2
x
0 1
1274 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
(0, 4)
D
(0, 1)
x
(1, 0)
The components of the vector field are P = ye x and Q = xe y , hence the flux is
Q P
= e y ex .
x y
Greens Theorem implies that
Q P
ye x d x + xe y d y = dA = (e y e x ) d A
C0 D x y D
0 4x+4 0 4x+4
= (e y e x ) d y d x = e y ye x dx
1 x+1 1 y=x+1
0
= e4x+4 (4x + 4)e x e x+1 (x + 1)e x dx
1
0
= e4x+4 e x+1 (3x + 3)e x d x
1
0 0
= (e4x+4 e x+1 3e x ) d x 3xe x d x
1 1
y = 4x + 4
D
1
x
1 y=x+1
e4 3
= e + 3e1 10.58
4 4
7. Let c(t) = t 2 (1 t), t(t 1)2 .
(a) Plot the path c(t) for 0 t 1.
(b) Calculate the area of the region enclosed by c(t) for 0 t 1.
SOLUTION
(a) The path c(t) for 0 t 1 is shown in the figure:
y
0.1
x
0 0.1
Therefore,
We obtain the following integral (note that the path must be counterclockwise):
1 0 2 2 1 1 4 3 2 1 t5 t4 t 3 1 1
A= t (t 1) dt = (t 2t + t ) dt = + =
2 1 2 0 2 5 2 3 0 60
8. In (a)(d), state whether the equation is an identity (valid for all F or ). If it is not, provide an example in which the
equation does not hold.
(a) curl( ) = 0 (b) div( ) = 0
(c) div(curl(F)) = 0 (d) (div(F)) = 0
SOLUTION
(a) This equality is valid for all since
= x , y , z
i j k
curl( ) = = (zy yz )i (zx x z )j ( yx x y )k
x y z
x y z
By the equality of the mixed partials, we conclude that curl( ) is the zero vector.
(b) This equation is not an identity. Take (x, y, z) = x 2 + y + z. Then = 2x, 1, 1 and
div( ) = (2x) + (1) + (1) = 2 + 0 + 0 = 2 = 0.
x y z
1276 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
(c) The equality div (curl(F)) = 0 is an identity. To prove it, we let F = P, Q, R. Then
i j k
curlF = F = = (R y Q z )i (Rx Pz )j + (Q x Py )k
x y z
P Q R
div (curl(F)) = (R y Q z ) (Rx Pz ) + (Q x Py )
x y z
= R yx Q zx (Rx y Pzy ) + Q x z Pyz
= (R yx Rx y ) + (Pzy Pyz ) + (Q x z Q zx ) = 0
2
div(F) = (x ) + (y) + (z) = 2x + 2
x y z
(div(F)) = (2x + 2), (2x + 2), (2x + 2) = 2, 0, 0 = 0
x y z
In Exercises 912, calculate the curl and divergence of the vector field.
9. F = yi zk
SOLUTION We compute the curl of the vector field,
i j k
curl(F) =
x y z
y 0 z
(0) (y)
= (z) (0) i (z) (y) j + k
y z x z x y
= 0i + 0j 1k = k
The divergence of F is
div(F) = (y) + (0) + (z) = 0 + 0 1 = 1.
x y z
10. F = e x+y , e y+z , x yz
SOLUTION The curl of F = e x+y , e y+z , x yz is the following vector:
i j k
curl(F) = F =
x y z
x+y
e e y+z x yz
y+z x+y
= (x yz) e y+z i (x yz) e x+y j + e e k
y z x z x y
= (x z e y+z )i (yz)j e x+y k = x z e y+z , yz, e x+y
The divergence of F is
x+y y+z
divF = (e )+ (e ) + (x yz) = e x+y + e y+z + x y.
x y z
11. F = (ex y z )
2 2 2
In Exercise 8 we proved the identity curl( ) = 0. Here, = ex y z , and we have curl ex y z
2 2 2 2 2 2
SOLUTION =
0. To compute div F, we first write F explicitly:
F = ex y z = 2xex y z , 2yex y z , 2zex y z = P, Q, R
2 2 2 2 2 2 2 2 2 2 2 2
Chapter Review Exercises 1277
P Q R
div(F) = + +
x y z
= 2ex y z + 4x 2 ex y z + 2ex y z + 4y 2 ex y z
2 2 2 2 2 2 2 2 2 2 2 2
+ 2ex y z + 4z 2 ex y z
2 2 2 2 2 2
= 2ex y z 2(x 2 + y 2 + z 2 ) 3
2 2 2
12. e = 1 x, y, z = x 2 + y 2 + z 2
SOLUTION It can be easily verified that e = for (x, y, z) = x 2 + y 2 + z 2 = . Therefore, by the identity
curl( ) = 0 (provided in Exercise 8), we have
curl(e ) = curl( ) = 0
div(e ) = (x 1 ) + (y 1 ) + (z 1 ) = (1 x 2 x ) + (1 y 2 y ) + (1 z 2 z )
x y z
x2 y2 z2 2
= 31 2 (x x + y y + z z ) = 31 2 + + = 31 2 = 21
13. Recall that if F1 , F2 , and F3 are differentiable functions of one variable, then
SOLUTION We use the linearity of the curl and the property mentioned in the exercise to compute the curl of F:
3
3
curl F = curl x 2 + y 2 , ln y + z 2 , z 3 sin z 2 e z = curl x 2 , ln y, z 3 sin(z 2 )e z + curl y 2 , z 2 , 0
2 2
= 0 + curl y 2 , z 2 , 0 = (0) z 2 , y 2 (0), z y = 2z, 0, 2y
y z z x x y
14. Give an example of a nonzero vector field F such that curl(F) = 0 and div(F) = 0.
SOLUTION Let F = x, y, 0. Then
i j k
curl(F) = = (0) + (y) i (0) (x) j + (y) (x) k = 0
x y z y z x z x y
x y 0
div(F) = (x) + (y) + (0) = 1 1 + 0 = 0
x y z
15. Verify the identities of Exercises 22 and 24 in Section 18.3 for the vector fields F = x z, ye x , yz and G =
2
z , x y3, x 2 y .
SOLUTION We first show div(curl(F)) = 0. Let F = P, Q, R = x z, ye x , yz . We compute the curl of F:
i j k
R Q P R Q P
curlF = = , ,
x y z y z z x x y
P Q R
Thus,
Thus,
F G = x 2 y 2 e x x y 4 z, yz 3 x 3 yz, x 2 y 3 z yz 2 e x
Thus,
So, we calculate
16. Suppose that S1 and S2 are surfaces with the same oriented boundary curve C. Which of the following conditions
guarantees that the flux of F through S1 is equal to the flux of F through S2 ?
(a) F = for some function
(b) F = curl(G) for some vector field G
SOLUTION If F = curl(G), then by the Theorem on Surface Independence for Curl Vector Fields, the flux of F through
a surface S depends only the oriented boundary S. Since S1 and S2 have the same oriented boundary curve, we conclude
that the flux of F through S1 is equal to the flux of F through S2 . The condition in (a) that F is conservative does not
guarantee that the flux of F through S1 is equal to the flux through S2 .
17. Prove that if F is a gradient vector field, then the flux of curl(F) through a smooth surface S (whether closed or not)
is equal to zero.
SOLUTION If F is a gradient vector field, then F is conservative; therefore the line integral of F over any closed curve
is zero. Combining with Stokes Theorem yields
curl(F) dS = F ds = 0
S S
18. Verify Stokes Theorem for F = y, z x, 0 and the surface z = 4 x 2 y 2 , z 0, oriented by outward-pointing
normals.
SOLUTION We begin by computing the line integral C F ds. The boundary curve is the circle x 2 + y 2 = 4 (in the
x y plane) oriented in the counterclockwise direction. We use the parametrization
Then,
Then,
Tu = = cos v, sin v, 2u
u
Tv = = u sin v, u cos v, 0
v
i j k
Tu Tv = cos v sin v 2u = (2u 2 cos v)i + (2u 2 sin v)j + (u cos2 v + u sin2 v)k
u sin v u cos v 0
= 2u 2 cos v, 2u 2 sin v, u
The surface is oriented outwards, hence the z-component of the normal vector is nonnegative. Therefore, the normal
vector is (recall that u 0)
n = 2u 2 cos v, 2u 2 sin v, u
By (1) and (2), both the line integral and the flux of the curl are equal to 8 . Thus, this example verifies Stokes
Theorem.
x2
In Exercises 1920, let F = z 2 , x + z, y 2 and let S be the upper half of the ellipsoid + y 2 + z 2 = 1, oriented by
4
outward-pointing normals.
19. Use Stokes Theorem to compute curl(F) dS.
S
SOLUTION We compute the curl of F = z 2 , x + z, y 2 :
i j k
curl(F) = = (2y 1)i (0 2z)j + (1 0)k = 2y 1, 2z, 1
x y z
z2 x +z y2
1280 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
2
Let C denote the boundary of S, that is, the ellipse x4 + y 2 = 1 in the x y-plane, oriented counterclockwise. Then by
Stokes Theorem we have
curl(F) dS = F ds (1)
S C
We parametrize C by
Then
F (r (t)) r (t) = 0, 2 cos t, sin2 t 2 sin t, cos t, 0 = 2 cos2 t
Then
sin3 t 4 cos2 t 2 4
A (r (t)) r (t) = , 0, 2 sin t, cos t, 0 = sin t
3 2 3
(0, 0, 1)
y2 + z2 = 1
x (0, 1, 0) y
FIGURE 2
Chapter Review Exercises 1281
en = 1, 0, 0 .
Hence,
22. Verify the Divergence Theorem for F = 0, 0, z and the region x 2 + y 2 + z 2 = 1.
SOLUTION Let S be the boundary of the unit sphere W. We calculate both sides of the equation:
F dS = div (F) d V (1)
S W
We start with the surface integral. We parametrize S by
( , ) = cos sin , sin sin , cos , 0 < 2 , 0
Hence,
F ( , ) n = 0, 0, cos sin cos sin , sin sin , cos = cos2 sin
z
5
1 2
x 4
y
We compute the divergence of F = x y, yz, x 2 z + z 2 :
div(F) = xy + yz + (x 2 z + z 2 ) = y + z + x 2 + 2z = x 2 + y + 3z
x y z
The Divergence Theorem gives
5 4 1 5 4 3 1
x
x y, yz, x 2 z + z 2 dS = (x 2 + y + 3z) d x d y dz = + (y + 3z)x d y dz
S 1 2 0 1 2 3 x=0
5 4 5 4
1 1 1
= + y + 3z d y dz = y + y 2 + 3zy dz
1 1 3 1 3 2 y=1
5 5
4 16 1 1 17
= + + 12z + + 3z dz = + 9z dz
1 3 2 3 2 1 2
17z 9z 2 5 (17 5 + 9 25) (17 + 9)
= + = = 142
2 2 1 2
24. F = x y, yz, x 2 z + z 2 , S is the boundary of the unit sphere.
SOLUTION We use spherical coordinates:
with
0 2 , 0 , 01
We obtain
x y, yz, x 2 z + z 2 dS = div(F) d V = (x 2 + y + 3z) d V
S V V
2 1
= (2 cos2 sin2 + sin sin + 3 cos ) 2 sin d d d
0 0 0
1
2
= cos2 d sin3 d 4 d
0 0 0
1
2
+ sin d sin2 d 3 d
0 0 0
1
+ 6 cos sin d 3 d
0 0
4 1 4
= +0+0=
3 5 15
2
25. F = x yz + x y, 12 y 2 (1 z) + e x , e x +y , S is the boundary of the solid bounded by the cylinder x 2 + y 2 = 16
2
Let S denote the surface of the solid W. The Divergence Theorem gives
0
F dS = div(F) d V = 2y d V = 2y dz d x d y
S W W D y4
0
= 2yz dx dy = 2y (0 (y 4)) d x d y = (8y 2y 2 ) d x d y
D z=y4 D D
26. F = sin(yz), x 2 + z 4 , x cos(x y) , S is any smooth closed surface that is the boundary of a region in R3 .
SOLUTION We compute the divergence of F:
2
div(F) = (sin(yz)) + x + z4 + (x cos x(x y)) = 0
x y z
Let W denote the solid inside S. The Divergence Theorem gives
F dS = div(F) d V = 0 dV = 0
S W W
16 = 8 Volume (W)
or
Volume (W) = 2
28. Show that the circulation of F = x 2 , y 2 , z(x 2 + y 2 ) around any curve C on the surface of the cone z 2 = x 2 + y 2
is equal to zero (Figure 3).
z2 = x2 + y2
y
x
FIGURE 3
1284 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
SOLUTION Let S be the part of the cone that is inside C. Then by Stokes Theorem, the circulation of F around C is
F ds = curl(F) dS (1)
C S
We parametrize the cone by
Then,
Tr = = cos , sin , 1
r
T = = r sin , r cos , 0
i j k
n = T Tr = r sin r cos 0 = r cos , r sin , r
cos sin 1
We see that curl(F) is tangent to the cone at all points on the cone, hence the surface integral in (1) is zero. We conclude
that the circulation of F around any curve C on the cone is zero.
In Exercises 2932, let F be a vector field whose curl and divergence at the origin are
en
y
C
x
The unit normal vector to the disk R is en = k = 0, 0, 1. The area of the disk is
C
R
en y
en = i = 1, 0, 0 .
The area of the square is Area(R) = 0.032 = 0.0009, and by the given information the curl at the origin is 2, 1, 4.
Therefore (1) gives the estimation
F ds 2, 1, 4 1, 0, 0 0.0009 = 2 0.0009 = 0.0018
C
The estimate depend on curl(F), en and the area of the square. Hence, if we flip the square over (such that en points along
the negative x-axis), then we will get a different answer.
31. Suppose that F is the velocity field of a fluid and imagine placing a small paddle wheel at the origin. Find the equation
of the plane in which the paddle wheel should be placed to make it rotate as quickly as possible.
SOLUTION The paddle wheel has the maximum spin when the circulation of the velocity field F around the wheel is
maximum. The maximum circulation occurs when en , and the curl of F at the origin (i.e., the vector 2, 1, 4) point in
the same direction. Therefore, the plane in which the paddle wheel should be placed is the plane through the origin with
the normal 2, 1, 4. This plane has the equation, 2x y + 4z = 0.
32. Estimate the flux of F through the box of side 0.5 in Figure 4. Does the result depend on how the box is oriented
relative to the coordinate axes?
z
0.5
FIGURE 4
The volume of the box W is 0.53 , and we are given that div(F)(0) = 2. This gives the estimation
F dS 2 0.53 = 0.25.
S
The negative sign shows that there is a net inflow across the box. Our estimation of the flux does not depend on the
orientation of the box; rather, it depends on the magnitude of the divergence of F.
33. The velocity field of a fluid (in meters per second) is
F = x 2 + y 2 , 0, z 2
1286 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
in the x y-plane.
(a) Show that no fluid flows across D.
(b) Use (a) to show that the rate of fluid flow across S is equal to div(F) d V . Compute this triple integral using
W
spherical coordinates.
SOLUTION
(a) To show that no fluid flows across D, we show that the normal component of F at each point on D is zero. At each
point P = (x, y, 0) on the x y-plane,
F(P) = x 2 + y 2 , 0, 02 = x 2 + y 2 , 0, 0 .
Moreover, the unit normal vector to the x y-plane is en = (0, 0, 1). Therefore,
F(P) en = x 2 + y 2 , 0, 0 0, 0, 1 = 0.
Since D is contained in the x y-plane, we conclude that the normal component of F at each point on D is zero. Therefore,
no fluid flows across D.
(b) By the Divergence Theorem and the linearity of the flux we have
F dS + F dS = div(F) d V
S D W
Since the flux through the disk D is zero, we have
F dS = div(F) d V (1)
S W
To compute the triple integral, we first compute div(F):
2
div(F) = (x + y 2 ) + (0) + (z 2 ) = 2x + 2z = 2(x + z).
x y z
z
(a) Estimate the flow rate (in cubic meters per second) through a small surface S around the origin if S encloses a region
of volume 0.01 m3 .
(b) Estimate the circulation of F about a circle in the x y-plane of radius r = 0.1 m centered at the origin (oriented
counterclockwise when viewed from above).
(c) Estimate the circulation of F about a circle in the yz-plane of radius r = 0.1 m centered at the origin (oriented
counterclockwise when viewed from the positive x-axis).
SOLUTION
(a) We use the approximation
F dS div(F)(0)Vol(W) (1)
S
en = k
R P = (0, 0.1, 0)
y
The unit normal vector to R is en = 0, 0, 1 = k and the area of the disc is 0.12 = 0.01 . We compute the curl at P:
i j k
curl(F) = = 2y ey(z+1) (y) i (2x 0)j + (0 3)k
x y z
(3y 4) ey(z+1) x 2 + y 2
= 2y + yey(z+1) i 2xj 3k
Hence,
C
R
P = (0, 0.1, 0)
y
x en = i
1288 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
We get
F ds (0.2 + 0.1e0.1 )i 3k i 0.01 = (0.2 + 0.1e0.1 ) 0.01 = 0.009
C
x
35. Let (x, y) = x + 2 . The vector field F = (Figure 5) provides a model in the plane of the velocity field
x + y2
of an incompressible, irrotational fluid flowing past a cylindrical obstacle (in this case, the obstacle is the unit circle
x 2 + y 2 = 1).
(a) Verify that F is irrotational [by definition, F is irrotational if curl(F) = 0].
y
3
2
1
x
3 2 1 1 2 3
1
2
3
x
FIGURE 5 The vector field for (x, y) = x + .
x 2 + y2
(b) Verify that F is tangent to the unit circle at each point along the unit circle except (1, 0) and (1, 0) (where F = 0).
(c) What is the circulation of F around the unit circle?
(d) Calculate the line integral of F along the upper and lower halves of the unit circle separately.
SOLUTION
(a) In Exercise 8, we proved the identity curl( ) = 0. Since F is a gradient vector field, it is irrotational; that is,
curl(F) = 0 for (x, y) = (0, 0), where F is defined.
(b) We compute F explicitly:
y2 x 2 2x y
F = = , = 1+ ,
x y (x 2 + y 2 )
2
(x 2 + y 2 )
2
Now, using x = cos t and y = sin t as a parametrization of the circle, we see that
F = 1 + sin2 t cos2 t, 2 cos t sin t = 2 sin2 t, 2 cos t sin t ,
and so
which is clearly perpendicular to the radial vector x, y for the circle.
(c) We use our expression of F from Part (b):
y2 x 2 2x y
F = = 1 + 2
, 2
(x 2 + y 2 ) (x 2 + y 2 )
Now, F is not defined at the origin and therefore we cannot use Greens Theorem to compute the line integral along the
unit circle. We thus compute the integral directly, using the parametrization
x
1
Chapter Review Exercises 1289
Then,
sin2 t cos2 t 2 cos t sin t
F (c(t)) c (t) = 1 + 2
, 2
sin t, cos t
(cos2 t + sin2 t) (cos2 t + sin2 t)
= 1 + sin2 t cos2 t, 2 cos t, sin t sin t, cos t = 2 sin2 t, 2 cos t sin t sin t, cos t
Hence,
2
F ds = 2 sin t dt = 0
C 0
(d) We denote by C1 and C2 the upper and lower halves of the unit circle. Using part (c) we have
F ds + F ds = 0 F ds = F ds (1)
C1 C2 C2 C1
C1
x
1
C2
To compute the circulation along C1 , we compute the integral as in part (c), only that the limits of integration are now
t = 0 and t = . Using the computations in part (c) we obtain
F ds = 2 sin2 t dt = 4
C1 0
Therefore, by (1),
F ds = 4.
C2
36. Figure 6 shows the vector field F = , where (x, y) = ln(x 2 + (y 1)2 ) + ln(x 2 + (y + 1)2 ), which is the
velocity field for the flow of a fluid with sources of equal strength at (0, 1) (note that is undefined at these two points).
Show that F is both irrotational and incompressible, that is, curlz (F) = 0 and div(F) = 0 [in computing div(F), treat F
as a vector field in R3 with a zero z-component]. Is it necessary to compute curlz (F) to conclude that it is zero?
y (0, 1)
(0, 1)
FIGURE 6
SOLUTION Since F is a gradient field it is irrotational. This property was proved in Exercise 8, where we showed that
curl( ) = 0 for all . To show that F is incompressible, we first find F explicitly.
2x 2x 2(y 1) 2(y + 1)
F(x, y, z) = , , = + , + ,0
x y z x 2 + (y 1)2 x 2 + (y + 1)2 x 2 + (y 1)2 x 2 + (y + 1)2
= F1 , F2 , F3
Hence,
F1 F2 F3
div F(x, y, z) = + +
x y z
1290 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
2 x 2 + (y 1)2 2x 2x 2 x 2 + (y + 1)2 2x 2x
=
+
x 2 + (y 1)2 x 2 + (y + 1)2
2 x 2 + (y 1)2 2(y 1) 2(y 1) 2 x 2 + (y + 1)2 2 2(y + 1)2
+
+
x 2 + (y 1)2 x 2 + (y + 1)2
=0+0=0
Note that, again by Exercise 8, the divergence of is zero, and hence so also is the divergence of F.
37. In Section 18.1, we showed that if C is a simple closed curve, oriented counterclockwise, then the line integral is
1
Area enclosed by C = x dy y dx 1
2 C
Suppose that C is a path from P to Q that is not closed but has the property that every line through the origin intersects
C in at most one point, as in Figure 7. Let R be the region enclosed by C and the two radial segments joining P and Q to
the origin. Show that the line integral in Eq. (1) is equal to the area of R. Hint: Show that the line integral of F = y, x
along the two radial segments is zero and apply Greens Theorem.
y
C
Q
R
P
x
FIGURE 7
SOLUTION
C
Q
R
P
x
Q
Then,
F (c(t)) c (t) = t tan , t 1, tan = t tan + t tan = 0
F (d(t)) d (t) = t tan , t 1, tan = t tan + t tan = 0
Therefore,
F ds = F ds = 0.
QO OP
38. Suppose that the curve C in Figure 7 has the polar equation r = f ( ).
(a) Show that c( ) = ( f ( ) cos , f ( ) sin ) is a counterclockwise parametrization of C.
(b) In Section 12.4, we showed that the area of the region R is given by the formula
1
Area of R = f ( )2 d
2
Use the result of Exercise 37 to give a new proof of this formula. Hint: Evaluate the line integral in (1) using c( ).
SOLUTION
(a) The curve r = f ( ) in polar coordinates can be parametrized using as a parameter. Since x = r cos and
y = r sin , we have
c( ) = ( f ( ) cos , f ( ) sin ) , .
We have
dy
= f ( ) sin + f ( ) cos d y = f ( ) sin + f ( ) cos d
d
dx
= f ( ) cos f ( ) sin d x = f ( ) cos f ( ) sin d
d
Hence
x dy y dx = f ( ) cos f ( ) sin + f ( ) cos f ( ) sin f ( ) cos f ( ) sin d
C
= f ( ) f ( ) cos sin + f 2 ( ) cos2 f ( ) f ( ) sin cos + f 2 ( ) sin2 d
= f 2 ( ) cos2 + sin2 d = f 2 ( ) d
Substituting in (1) we obtain
1 2
area of R = f ( ) d
2
39. Prove the following generalization of Eq. (1). Let C be a simple closed curve in the plane (Figure 8)
S: ax + by + cz + d = 0
Plane S
n = a, b, c
C
R
y
x
FIGURE 8
1292 C H A P T E R 18 F U N D A M E N TA L TH E O R E M S O F V E C T O R A N A LY S I S (ET CHAPTER 17)
Therefore,
1
curl(F) en = 2 a, b, c a, b, c
a + b2 + c2
2
2
= (a 2 + b2 + c2 ) = 2 a 2 + b2 + c2
a 2 + b2 + c2
Hence,
curl(F) dS = curl(F) en d S = 2 a 2 + b2 + c2 d S = 2 a 2 + b2 + c2 1 dS (2)
S S S S
The sign of S 1 dS is determined by the orientation of S. Since the area is a positive value, we have
1 ds = Area (S)
S
40. Use the result of Exercise 39 to calculate the area of the triangle with vertices (1, 0, 0), (0, 1, 0), and (0, 0, 1) as a
line integral. Verify your result using geometry.
SOLUTION In Exercise 39 we showed that if C is a simple closed curve in the plane ax + by + cz + d = 0, then the
area of the region R enclosed by C is equal to
1
(bz cy) d x + (cx az) d y + (ay bx) dz, n = a, b, c (1)
2n C
z
C = (0, 0, 1)
B = (0, 1, 0)
y
A = (1, 0, 0)
x
Chapter Review Exercises 1293
We use this formula where C is the triangle ABC parametrized counterclockwise. We compute the upward-pointing
normal to the plane of the triangle:
n = AB AC = (j i) (k i) = i + k + j = i + j + k.
We substitute n = 12 + 12 + 12 = 3 and a = b = c = 1 in (1) to obtain:
1
area of R = (z y) d x + (x z) d y + (y x) dz (2)
2 3 C
We parametrized the oriented segments by
The integral along C is the sum of these three integrals. That is,
(z c) d x + (x z) d y + (y x) dz = 1 + 1 + 1 = 3
C
We combine with (2) to obtain the following area of the triangle:
1 3
area of R = 3 =
2 3 2
We verify this solution geometrically. The triangle spanned by the vectors AB = j i and AC = k i has area
1 1
1
AB AC = i + j + k = 3.
2 2 2
The two answers match.