Chapter 1 Calc Sinlge Multi

APPENDIX A SOLUTIONS 269
APPENDIX
Solutions for Section A
1. The graph is
-3 -2 -1 1 2 3
-40
-20
20
40
(a) The range appears to be y 30.
(b) The function has two zeros.
5. The largest root is at about 2.5.
9. Using a graphing calculator, we see that when x is around 0.45, the graphs intersect.
13. (a) Only one real zero, at about x = 1.15.
(b) Three real zeros: at x = 1, and at about x = 1.41 and x = 1.41.
17. (a) Since f is continuous, there must be one zero between = 1.4 and = 1.6, and another between = 1.6 and
= 1.8. These are the only clear cases. We might also want to investigate the interval 0.6 0.8 since f()
takes on values close to zero on at least part of this interval. Now, = 0.7 is in this interval, and f(0.7) = 0.01 < 0,
so f changes sign twice between = 0.6 and = 0.8 and hence has two zeros on this interval (assuming f is not
really wiggly here, which its not). There are a total of 4 zeros.
(b) As an example, we nd the zero of f between = 0.6 and = 0.7. f(0.65) is positive; f(0.66) is negative. So
this zero is contained in [0.65, 0.66]. The other zeros are contained in the intervals [0.72, 0.73], [1.43, 1.44], and
[1.7, 1.71].
(c) Youve found all the zeros. A picture will conrm this; see Figure A.1.
0.2 0.4 0.6 0.8 1 1.2 1.4
1.6
1.8
f() = sin 3 cos 4 + 0.8
Figure A.1
270 APPENDIX /SOLUTIONS
21.
1
2 4
4
5
x
f(x) = 4x x
2
Bounded and 5 f(x) 4.
Solutions for Section B
1. 2e
i/2
5. 0e
i
, for any .
9. 3 4i
13.
1
4

9i
8
17. 5
3
(cos
3
2
+ i sin
3
2
) = 125i
21. One value of
3
i is
3
e
i
2 = (e
i
2
)
1
3
= e
i
6
= cos

6
+ i sin

6
=
3
2
+
i
2
25. One value of (4 + 4i)
2/3
is [
32e
(i3/4)
]
(2/3)
= (
32)
2/3
e
(i/2)
= 2
5/3
cos

2
+ i2
5/3
sin

2
= 2i
3
4
29. We have
i
1
=
1
i
=
1
i

i
i
= i,
i
2
=
1
i
2
= 1,
i
3
=
1
i
3
=
1
i

i
i
= i,
i
4
=
1
i
4
= 1.
The pattern is
i
n
=
i n = 1, 5, 9,
1 n = 2, 6, 10,
i n = 3, 7, 11,
1 n = 4, 8, 12, .
Since 36 is a multiple of 4, we know i
36
= 1.
Since 41 = 4 10 + 1, we know i
41
= i.
33. To conrm that z =
a + bi
c + di
, we calculate the product
z(c + di) =
ac + bd
c
2
+ d
2
=
bc ad
c
2
+ d
2
i
(c + di)
=
ac
2
+ bcd bcd + ad
2
+ (bc
2
acd + acd + bd
2
)i
c
2
+ d
2
=
a(c
2
+ d
2
) + b(c
2
+ d
2
)i
c
2
+ d
2
= a + bi.
37. True, since
a is real for all a 0.

APPENDIX C SOLUTIONS 271
41. True. We can write any nonzero complex number z as re
i
, where r and are real numbers with r > 0. Since r > 0, we
can write r = e
c
for some real number c. Therefore, z = re
i
= e
c
e
i
= e
c+i
= e
w
where w = c + i is a complex
number.
45. Using Eulers formula, we have:
e
i(2)
= cos 2 + i sin 2
On the other hand,
e
i(2)
= (e
i
)
2
= (cos + i sin )
2
= (cos
2
sin
2
) + i(2 cos sin )
Equating real parts, we nd
cos 2 = cos
2
sin
2
.
49. Replacing by (x + y) in the formula for sin :
sin(x + y) =
1
2i
e
i(x+y)
e
i(x+y)
=
1
2i
e
ix
e
iy
e
ix
e
iy
=
1
2i
((cos x + i sin x) (cos y + i sin y) (cos (x) + i sin (x)) (cos (y) + i sin (y)))
=
1
2i
((cos x + i sin x) (cos y + i sin y) (cos x i sin x) (cos y i sin y))
= sin xcos y + cos xsin y.
Solutions for Section C
1. (a) f
(x) = 3x
2
+ 6x + 3 = 3(x + 1)
2
. Thus f
(x) > 0 everywhere except at x = 1, so it is increasing everywhere

except perhaps at x = 1. The function is in fact increasing at x = 1 since f(x) > f(1) for x > 1, and
f(x) < f(1) for x < 1.
(b) The original equation can have at most one root, since it can only pass through the x-axis once if it never decreases.
It must have one root, since f(0) = 6 and f(1) = 1.
(c) The root is in the interval [0, 1], since f(0) < 0 < f(1).
(d) Let x0 = 1.
x0 = 1
x1 = 1
f(1)
f
(1)
= 1
1
12
=
11
12
0.917
x2 =
11
12

f
11
12
11
12
0.913
x3 = 0.913
f(0.913)
f
(0.913)
0.913.
Since the digits repeat, they should be accurate. Thus x 0.913.
5. Let f(x) = sin x 1 + x; we want to nd all zeros of f, because f(x) = 0 implies sin x = 1 x.
Graphing sin x and 1 x in Figure C.2, we see that f(x) has one solution at x
1
2
.
/2
1
1
y = sin x
y = 1 x
Figure C.2
Letting x0 = 0.5, and using Newtons method, we have f
(x) = cos x + 1, so that

x1 = 0.5
sin(0.5) 1 + 0.5
cos(0.5) + 1
0.511,
x2 = 0.511
sin(0.511) 1 + 0.511
cos(0.511) + 1
0.511.
Thus sin x = 1 x has one solution at x 0.511.
9. Let f(x) = ln x
1
x
, so f
(x) =
1
x
+
1
x
2
.
Now use Newtons method with an initial guess of x0 = 2.
x1 = 2
ln 2
1
2
1
2
+
1
4
1.7425,
x2 1.763,
x3 1.763.
Thus x 1.763 is a solution. Since f
(x) > 0 for positive x, f is increasing: it must be the only solution.

Solutions for Section D
Exercises
1. The magnitude is 3
i =
3
2
+ 0
2
= 3.
The angle of 3
i is 0 because the vector lies along the positive x-axis.

5. 2v + w = (2 2)
i + (4 + 3)
j = 7
j .
9. Two vectors have opposite direction if one is a negative scalar multiple of the other. Since
5
j =
5
6
(6
j )
the vectors 5
j and 6
j have opposite direction. Similarly, 6
j and
j have opposite direction.

13. Scalar multiplication by 2 doubles the magnitude of a vector without changing its direction. Thus, the vector is 2(4
i
3
j ) = 8
i 6
j .
17. In components, the vector from (7, 7) to (9, 11) is (9 7)
i + (11 9)
j = 2
i + 2
j .
In components, the vector from (8, 10) to (10, 12) is (10 8)
i + (12 10)
j = 2
i + 2
j .
The two vectors are equal.
21. The velocity is v (t) = e
t
i + (1/(1 + t))
j . When t = 0, the velocity vector is v =
i +
j .
The speed is v =
1
2
+ 1
2
=
2.
The acceleration is a (t) = e
t
i 1/((1 + t)
2
)
j . When t = 0, the acceleration vector is a =
j .
10.1 SOLUTIONS 217
CHAPTER TEN
Solutions for Section 10.1
Exercises
1. Let f(x) =
1
1 x
= (1 x)
1
. Then f(0) = 1.
f
(x) = 1!(1 x)
2
f
(0) = 1!,
f
(x) = 2!(1 x)
3
f
(0) = 2!,
f
(x) = 3!(1 x)
4
f
(0) = 3!,
f
(4)
(x) = 4!(1 x)
5
f
(4)
(0) = 4!,
f
(5)
(x) = 5!(1 x)
6
f
(5)
(0) = 5!,
f
(6)
(x) = 6!(1 x)
7
f
(6)
(0) = 6!,
f
(7)
(x) = 7!(1 x)
8
f
(7)
(0) = 7!.
P3(x) = 1 +x +x
2
+x
3
,
P5(x) = 1 +x +x
2
+x
3
+x
4
+x
5
,
P7(x) = 1 +x +x
2
+x
3
+x
4
+x
5
+x
6
+x
7
.
5. Let f(x) = cos x. Then f(0) = cos(0) = 1, and
f
(x) = sin x f
(0) = 0,
f
(x) = cos x f
(0) = 1,
f
(x) = sin x f
(0) = 0,
f
(4)
(x) = cos x f
(4)
(0) = 1,
f
(5)
(x) = sin x f
(5)
(0) = 0,
f
(6)
(x) = cos x f
(6)
(0) = 1.
Thus,
P2(x) = 1
x
2
2!
,
P4(x) = 1
x
2
2!
+
x
4
4!
,
P6(x) = 1
x
2
2!
+
x
4
4!

x
6
6!
.
9. Let f(x) =
1
1 +x
= (1 +x)
1/2
. Then f(0) = 1.
f
(x) =
1
2
(1 +x)
3/2
f
(0) =
1
2
,
f
(x) =
3
2
2
(1 +x)
5/2
f
(0) =
3
2
2
,
f
(x) =
35
2
3
(1 +x)
7/2
f
(0) =
35
2
3
,
f
(4)
(x) =
357
2
4
(1 +x)
9/2
f
(4)
(0) =
357
2
4
218 Chapter Ten /SOLUTIONS
Then,
P2(x) = 1
1
2
x +
1
2!
3
2
2
x
2
= 1
1
2
x +
3
8
x
2
,
P3(x) = P2(x)
1
3!
3 5
2
3
x
3
= 1
1
2
x +
3
8
x
2
5
16
x
3
,
P4(x) = P3(x) +
1
4!
3 5 7
2
4
x
4
= 1
1
2
x +
3
8
x
2
5
16
x
3
+
35
128
x
4
.
13. Let f(x) =
1 +x = (1 +x)
1/2
.
Then f
(x) =
1
2
(1 +x)
1/2
, f
(x) =
1
4
(1 +x)
3/2
, and f
(x) =
3
8
(1 +x)
5/2
. The Taylor polynomial of degree
three about x = 1 is thus
P3(x) = (1 + 1)
1/2
+
1
2
(1 + 1)
1/2
(x 1) +
1
4
(1 + 1)
3/2
2!
(x 1)
2
+
3
8
(1 + 1)
5/2
3!
(x 1)
3
=
2
_
1 +
x 1
4

(x 1)
2
32
+
(x 1)
3
128
_
.
Problems
17. Using the fact that
f(x) P3(x) = f(0) +f
(0)x +
f
(0)
2!
x
2
+
f
(0)
3!
x
3
and identifying coefcients with those given for P3(x), we obtain the following:
(a) f(0) = constant term which equals 2, so f(0) = 2.
(b) f
(0) = coefcient of x which equals 1, so f
(0) = 1.
(c)
f
(0)
2!
= coefcient of x
2
which equals 1/3, so f
(0) = 2/3.
(d)
f
(0)
3!
= coefcient of x
3
which equals 2, so f
(0) = 12.
21. Since P2(x) is the second degree Taylor polynomial for f(x) about x = 0, P2(0) = f(0), which says a = f(0). Since
d
dx
P2(x)
x=0
= f
(0),
b = f
(0); and since

d
2
dx
2
P2(x)
x=0
= f
(0),
2c = f
(0). In other words, a is the y-intercept of f(x), b is the slope of the tangent line to f(x) at x = 0 and c tells us the
concavity of f(x) near x = 0. So c < 0 since f is concave down; b > 0 since f is increasing; a > 0 since f(0) > 0.
25.
lim
x0
sin x
x
= lim
x0
x
x
3
3!
x
= lim
x0
_
1
x
2
3!
_
= 1.
29. (a) Since the coefcient of the x-term of each f is 1, we know f
1
(0) = f
2
(0) = f
3
(0) = 1. Thus, each of the fs slopes
upward near 0, and are in the second gure.
The coefcient of the x-term in g1 and in g2 is 1, so g
1
(0) = g
2
(0) = 1. For g3 however, g
3
(0) = 1. Thus,
g1 and g2 slope up near 0, but g3 slopes down. The gs are in the rst gure.
(b) Since g1(0) = g2(0) = g3(0) = 1, the point A is (0, 1).
Since f1(0) = f2(0) = f3(0) = 2, the point B is (0, 2).
(c) Since g3 slopes down, g3 is I. Since the coefcient of x
2
for g1 is 2, we know
g
1
(0)
2!
= 2 so g
1
(0) = 4.
10.2 SOLUTIONS 219
By similar reasoning g
2
(0) = 2. Since g1 and g2 are concave up, and g1 has a larger second derivative, g1 is III and
g2 is II.
Calculating the second derivatives of the fs from the coefcients x
2
, we nd
f
1
(0) = 4 f
2
(0) = 2 f
3
(0) = 2.
Thus, f1 and f3 are concave up, with f1 having the larger second derivative, so f1 is III and f3 is II. Then f2 is
concave down and is I.
33. (a) The equation sin x = 0.2 has one solution near x = 0 and innitely many others, one near each multiple of . See
Figure 10.1. The equation x
x
3
3!
= 0.2 has three solutions, one near x = 0 and two others. See Figure 10.2.
y = 0.2
x
y
Figure 10.1: Graph of y = sin x and y = 0.2
y = 0.2
x
y
Figure 10.2: Graph of y = x
x
3
3!
and y = 0.2
(b) Near x = 0, the cubic Taylor polynomial x x
3
/3! sin x. Thus, the solutions to the two equations near x = 0 are
approximately equal. The other solutions are not close. The reason is that x x
3
/3! only approximates sin x near
x = 0 but not further away. See Figure 10.3.
sin x
y = 0.2
x x
3
/3!
x
y
Figure 10.3
Exercises
1. Differentiating (1 +x)
3/2
:
f(x) = (1 +x)
3/2
f(0) = 1,
f
(x) = (3/2)(1 +x)

1/2
f
(0) =
3
2
,
f
(x) = (1/2)(3/2)(1 +x)

1/2
= (3/4)(1 +x)
1/2
f
(0) =
3
4
,
f
(x) = (1/2)(3/4)(1 +x)

3/2
= (3/8)(1 +x)
3/2
f
(0) =
3
8
.
f(x) = (1 +x)
3/2
= 1 +
3
2
x +
(3/4)x
2
2!
+
(3/8)x
3
3!
+
= 1 +
3x
2
+
3x
2
8

x
3
16
+
5.
f(x) =
1
1x
= (1 x)
1
f(0) = 1,
f
(x) = (1 x)
2
(1) = (1 x)
2
f
(0) = 1,
f
(x) = 2(1 x)
3
(1) = 2(1 x)
3
f
(0) = 2,
f
(x) = 6(1 x)
4
(1) = 6(1 x)
4
f
(0) = 6.
f(x) =
1
1 x
= 1 + 1 x +
2x
2
2!
+
6x
3
3!
+
= 1 +x +x
2
+x
3
+
9.
f(x) = sin x f(
4
) =
2
2
,
f
(x) = cos x f
4
) =
2
2
,
f
(x) = sin x f
4
) =
2
2
,
f
(x) = cos x f
4
) =
2
2
.
sin x =
2
2
+
2
2
_
x

4
_
2
2
(x

4
)
2
2!

2
2
(x

4
)
3
3!

=
2
2
+
2
2
_
x

4
_
2
4
_
x

4
_
2
2
12
_
x

4
_
3

13.
f(x) = tan x f(
4
) = 1,
f
(x) =
1
cos
2
x
f
4
) = 2,
f
(x) =
2(sin x)
cos
3
x
=
2 sin x
cos
3
x
f
4
) = 4,
f
(x) =
6 sin x(sin x)
cos
4
x
+
2
cos
2
x
f
4
) = 16.
tan x = 1 + 2
_
x

4
_
+ 4
(x

4
)
2
2!
+ 16
_
x

4
_
3
3!
+
= 1 + 2
_
x

4
_
+ 2
_
x

4
_
2
+
8
3
_
x

4
_
3
+
17. The general term can be written as x
n
for n 0.
21. The general term can be written as (1)
k
x
2k+1
/(2k + 1)! for k 0.
Problems
25. By looking at Figure 10.4 we can that the Taylor polynomials are reasonable approximations for the function f(x) =
1
1+x
between x = 1 and x = 1. Thus a good guess is that the interval of convergence is 1 < x < 1.
10.2 SOLUTIONS 221
1 1 2 3
1
2
3
4
f(x) =
1
1+x
P
2
(x)
P
4
(x) P
10
(x)
x
Figure 10.4
29. (a) We have shown that the series is
1 +px +
p(p 1)
2!
x
2
+
p(p 1)(p 2)
3!
x
3
+
so the general term is
p(p 1) . . . (p (n 1))
n!
x
n
.
(b) We use the ratio test
lim
n
|an+1|
|an|
= |x| lim
n
p(p 1) . . . (p (n 1))(p n) n!
(n + 1)!p(p 1) . . . (p (n 1))
= |x| lim
n
p n
n + 1
.
Since p is xed, we have
lim
n
p n
n + 1
= 1, so R = 1.
33. We dene e
i
to be
e
i
= 1 +i +
(i)
2
2!
+
(i)
3
3!
+
(i)
4
4!
+
(i)
5
5!
+
(i)
6
6!
+
Suppose we consider the expression cos +i sin , with cos and sin replaced by their Taylor series:
cos +i sin =
_
1

2
2!
+

4
4!

6
6!
+
_
+i
_

3
3!
+

5
5!

_
Reordering terms, we have
cos +i sin = 1 +i

2
2!

i
3
3!
+

4
4!
+
i
5
5!

6
6!

Using the fact that i
2
= 1, i
3
= i, i
4
= 1, i
5
= i, , we can rewrite the series as
cos +i sin = 1 +i +
(i)
2
2!
+
(i)
3
3!
+
(i)
4
4!
+
(i)
5
5!
+
(i)
6
6!
+
Amazingly enough, this series is the Taylor series for e
x
with i substituted for x. Therefore, we have shown that
cos +i sin = e
i
.
37. This is the series for cos x with x replaced by 10, so the series converges to cos 10.
41. This is the series for cos x with x = 1 substituted. Thus
1
1
2!
+
1
4!

1
6!
+ = cos 1.
Exercises
1. Substitute y = x into e
y
= 1 +y +
y
2
2!
+
y
3
3!
+ . We get
e
x
= 1 + (x) +
(x)
2
2!
+
(x)
3
3!
+
= 1 x +
x
2
2!

x
3
3!
+ .
5. Since
d
dx
(arcsin x) =
1
1x
2
= 1 +
1
2
x
2
+
3
8
x
4
+
5
16
x
6
+ , integrating gives
arcsin x = c +x +
1
6
x
3
+
3
40
x
5
+
5
112
x
7
+ .
Since arcsin 0 = 0, c = 0.
9.
3
cos(
2
) =
3
_
1
(
2
)
2
2!
+
(
2
)
4
4!

(
2
)
6
6!
+
_
=
3

7
2!
+

11
4!

15
6!
+
13. Multiplying out gives (1 +x)
3
= 1 +3x+3x
2
+x
3
. Since this polynomial equals the original function for all x, it must
be the Taylor series. The general term is 0 x
n
for n 4.
17. Using the binomial expansion for (1 +x)
1/2
with x = h/T:
T +h =
_
T +
T
T
h
_
1/2
=
_
T
_
1 +
h
T
__
1/2
=
T
_
1 +
h
T
_
1/2
=
T
_
1 + (1/2)
_
h
T
_
+
(1/2)(1/2)
2!
_
h
T
_
2
+
(1/2)(1/2)(3/2)
3!
_
h
T
_
3

_
=
T
_
1 +
1
2
_
h
T
_
1
8
_
h
T
_
2
+
1
16
_
h
T
_
3

_
.
21.
a
a
2
+x
2
=
a
a(1 +
x
2
a
2
)
1
2
=
_
1 +
x
2
a
2
_
1
2
= 1 +
_
1
2
_
x
2
a
2
+
1
2!
_
1
2
__
3
2
_
_
x
2
a
2
_
2
+
1
3!
_
1
2
__
3
2
__
5
2
_
_
x
2
a
2
_
3
+
= 1
1
2
_
x
a
_
2
+
3
8
_
x
a
_
4
5
16
_
x
a
_
6
+
Problems
25. From the series for ln(1 +y),
ln(1 +y) = y
y
2
2
+
y
3
3

y
4
4
+ ,
10.3 SOLUTIONS 223
we get
ln(1 +y
2
) = y
2
y
4
2
+
y
6
3

y
8
4
+
The Taylor series for sin y is
sin y = y
y
3
3!
+
y
5
5!

y
7
7!
+
So
sin y
2
= y
2
y
6
3!
+
y
10
5!

y
14
7!
+
The Taylor series for cos y is
cos y = 1
y
2
2!
+
y
4
4!

y
6
6!
+
So
1 cos y =
y
2
2!

y
4
4!
+
y
6
6!
+
Near y = 0, we can drop terms beyond the fourth degree in each expression:
ln(1 +y
2
) y
2
y
4
2
sin y
2
y
2
1 cos y
y
2
2!

y
4
4!
.
(Note: These functions are all even, so what holds for negative y will hold for positive y.)
Clearly 1 cos y is smallest, because the y
2
term has a factor of
1
2
. Thus, for small y,
y
2
2!

y
4
4!
< y
2
y
4
2
< y
2
so
1 cos y < ln(1 +y
2
) < sin(y
2
).
29. Since e
x
=
n=0
x
n
n!
and sinh 2x = (e
2x
e
2x
)/2, the Taylor expansion for sinh 2x is
sinh 2x =
1
2
_

n=0
(2x)
n
n!

n=0
(2x)
n
n!
_
=
1
2
_

n=0
(1 (1)
n
)
(2x)
n
n!
_
=
m=0
(2x)
2m+1
(2m+ 1)!
.
Since cosh 2x = (e
2x
+e
2x
)/2, we have
cosh 2x =
1
2
_

n=0
(2x)
n
n!
+
n=0
(2x)
n
n!
_
=
1
2
_

n=0
(1 + (1)
n
)
(2x)
n
n!
_
=
m=0
(2x)
2m
(2m)!
.
33. (a) =
mM
m+M
.
If M >> m, then the denominator m+M M , so
mM
M
= m.
(b)
= m
_
M
m+M
_
= m
_
1
M
M
m
M
+
M
M
_
= m
_
1
1 +
m
M
_
We can use the binomial expansion since
m
M
< 1.
= m
_
1
m
M
+
_
m
M
_
2
_
m
M
_
3
+
_
(c) If m
1
1836
M, then
m
M

1
1836
0.000545.
So a rst order approximation to would give = m(1 0.000545). The percentage difference from = m is
0.0545%.
37. (a) To nd when V takes on its minimum values, set
dV
dr
= 0. So
V0
d
dr
_
2
_
r0
r
_
6
_
r0
r
_
12
_
= 0
V0
_
12r
6
0
r
7
+ 12r
12
0
r
13
_
= 0
12r
6
0
r
7
= 12r
12
0
r
13
r
6
0
= r
6
r = r0.
Rewriting V
(r) as
12r
6
0
V0
r7
_
1
_
r0
r
_
6
_
, we see that V
(r) > 0 for r > r0 and V
(r) < 0 for r < r0. Thus,

V = V0(2(1)
6
(1)
12
) = V0 is a minimum.
(Note: We discard the negative root r0 since the distance r must be positive.)
(b)
V (r) = V0
_
2
_
r0
r
_
6
_
r0
r
_
12
_
V
(r) = V0(12r
6
0
r
7
+ 12r
12
0
r
13
)
V
(r) = V0(84r
6
0
r
8
156r
12
0
r
14
)
V (r0) = V0
V
(r0) = 0
V
(r0) = 72V0r
2
0
The Taylor series is thus:
V (r) = V0 + 72V0r
2
0
(r r0)
2
1
2
+
(c) The difference between V and its minimum value V0 is
V (V0) = 36V0
(r r0)
2
r
2
0
+
which is approximately proportional to (r r0)
2
since terms containing higher powers of (r r0) have relatively
small values for r near r0.
(d) From part (a) we know that dV/dr = 0 when r = r0, hence F = 0 when r = r0. Since, if we discard powers of
(r r0) higher than the second,
V (r) V0
_
1 36
(r r0)
2
r
2
0
_
giving
F =
dV
dr
72
r r0
r
2
0
(V0) = 72V0
r r0
r
2
o
.
So F is approximately proportional to (r r0).
Exercises
1. The error bound in approximating e
01
using the Taylor polynomial of degree 3 for f(x) = e
x
about x = 0 is:
|E3| = |f(0.1) P3(0.1)|
M |0.1 0|
4
4!
=
M(0.1)
4
24
,
where |f
(4)
(x)| M for 0 x 0.1. Now, f
(4)
(x) = e
x
. Since e
x
is increasing for all x, we see that |f
(4)
(x)| is
maximized for x between 0 and 0.1 when x = 0.1. Thus,
|f
(4)
| e
0.1
,
10.4 SOLUTIONS 225
so
|E3|
e
0.1
(0.1)
4
24
= 0.00000460.
The Taylor polynomial of degree 3 is
P3(x) = 1 +x +
1
2!
x
2
+
1
3!
x
3
.
The approximation is P3(0.1), so the actual error is
E3 = e
0.1
P3(0.1) = 1.10517092 1.10516667 = 0.00000425,
which is slightly less than the bound.
5. The error bound in approximating ln(1.5) using the Taylor polynomial of degree 3 for f(x) = ln(1 +x) about x = 0 is:
|E4| = |f(0.5) P3(0.5)|
M |0.5 0|
4
4!
=
M(0.5)
4
24
,
where |f
(4)
(x)| M for 0 x 0.5. Since
f
(4)
(x) =
3!
(1 +x)
4
and the denominator attains its minimum when x = 0, we have |f
(4)
(x)| 3!, so
|E4|
3! (0.5)
4
24
= 0.0156.
The Taylor polynomial of degree 3 is
P3(x) = 0 +x + (1)
x
2
2!
+ (1)(2)
x
3
3!
= x
1
2
x
2
+
1
3
x
3
.
The approximation is P3(0.5), so the actual error is
E3 = ln(1.5) P3(0.5) = 0.4055 0.4167 = 0.0112
which is slightly less, in absolute value, than the bound.
Problems
9. Let f(x) =
1 +x. We use a Taylor polynomial with x = 1 to approximate

2. The error bound for the Taylor
approximation of degree three for f(x) =
2 about x = 0 is:
|E3| = |f(1) P3(1)|
M |1 0|
4
4!
=
M
24
,
where |f
(4)
(x)| M for 0 x 1.
Now,
f
(4)
(x) =
15
16
(1 +x)
7/2
=
15
16(1 +x)
7/2
.
Since 1 (1 +x)
7/2
for x between 0 and 1, we see that
|f
(4)
(x)| =
15
16(1 +x)
7
2
15
16
for x between 0 and 1. Thus,
|E3|
15
16 24
< 0.039
13. (a) The second-degree Taylor polynomial for f(t) = e
t
is P2(t) = 1 + t + t
2
/2. Since the full expansion of e
t
=
1 +t +t
2
/2 +t
3
/6 +t
4
/24 + is clearly larger than P2(t) for t > 0, P2(t) is an underestimate on [0, 0.5].
(b) Using the second-degree error bound, if |f
(3)
(t)| M for 0 t 0.5, then
|E2|
M
3!
|t|
3
M(0.5)
3
6
.
Since |f
(3)
(t)| = e
t
, and e
t
is increasing on [0, 0.5],
f
(3)
(t) e
0.5
<
4 = 2.
So
|E2|
(2)(0.5)
3
6
< 0.047.
17. (a)
Table 10.1
E1 = sin x x
x sin x E
0.5 0.4794 0.0206
0.4 0.3894 0.0106
0.3 0.2955 0.0045
0.2 0.1987 0.0013
0.1 0.0998 0.0002
Table 10.2
E1 = sin x x
x sin x E
0 0 0
0.1 0.0998 0.0002
0.2 0.1987 0.0013
0.3 0.2955 0.0045
0.4 0.3894 0.0106
0.5 0.4794 0.0206
(b) See answer to part (a) above.
(c)
0.5 0.5
0.03
0.03
x
E
1
The fact that the graph of E1 lies between the horizontal lines at 0.03 shows that |E1| < 0.03 for 0.5
x 0.5.
21.
sin x = x
x
3
3!
+
x
5
5!

Write the error in approximating sin x by the Taylor polynomial of degree n = 2k + 1 as En so that
sin x = x
x
3
3!
+
x
5
5!
(1)
k
x
2k+1
(2k + 1)!
+En.
(Notice that (1)
k
= 1 if k is even and (1)
k
= 1 if k is odd.) We want to show that if x is xed, En 0 as k .
Since f(x) = sin x, all the derivatives of f(x) are sin x or cos x, so we have for all n and all x
|f
(n+1)
(x)| 1.
Using the bound on the error given in the text on page 528, we see that
|En|
1
(2k + 2)!
|x|
2k+2
.
By the argument in the text on page 528, we know that for all x,
|x|
2k+2
(2k + 2)!
=
|x|
n+1
(n + 1)!
0 as n = 2k + 1 .
Thus the Taylor series for sin x does converge to sin x for every x.
10.5 SOLUTIONS 227
Exercises
1. No, a Fourier series has terms of the form cos nx, not cos
n
x.
5.
a0 =
1
2
_

f(x) dx =
1
2
__
0
1 dx +
_

0
1 dx
_
= 0
a1 =
1
f(x) cos xdx =

1
__
0
cos xdx +
_

0
cos xdx
_
=
1
_
sin x
+ sin x
0
_
= 0.
Similarly, a2 and a3 are both 0.
(In fact, notice f(x) cos nx is an odd function, so
_
f(x) cos nx = 0.)

b1 =
1
f(x) sin xdx =

1
__
0
sin xdx +
_

0
sin xdx
_
=
1
_
cos x
+ (cos x)
0
_
=
4
b2 =
1
f(x) sin 2xdx =

1
__
0
sin 2xdx +
_

0
sin 2xdx
_
=
1
_
1
2
cos 2x
+ (
1
2
cos 2x)
0
_
= 0.
b3 =
1
f(x) sin 3xdx =

1
__
0
sin 3xdx +
_

0
sin 3xdx
_
=
1
_
1
3
cos 3x
+ (
1
3
cos 3x)
0
_
=
4
3
.
Thus, F1(x) = F2(x) =
4
sin x and F3(x) =

4
sin x +
4
3
sin 3x.
1
1
x
F
1
(x) = F
2
(x) =
4
sin x
1
1
x
F
3
(x) =
4
sin x +
4
3
sin 3x
9.
a0 =
1
2
_

h(x) dx =
1
2
_

0
xdx =

4
As in Problem 10, we use the integral table (III-15 and III-16) to nd formulas for an and bn.
an =
1
h(x) cos(nx) dx =
1
_

0
xcos nxdx =
1
_
x
n
sin(nx) +
1
n
2
cos(nx)
_
0
=
1
_
1
n
2
cos(n)
1
n
2
_
=
1
n
2
_
cos(n) 1
_
.
Note that since cos(n) = (1)
n
, an = 0 if n is even and an =
2
n
2
if n is odd.
bn =
1
h(x) cos(nx) dx =
1
_

0
xsin xdx
=
1
x
n
cos(nx) +
1
n
2
sin(nx)
_
0
=
1

n
cos(n)
_
=
1
n
cos(n)
=
1
n
(1)
n+1
if n 1
We have that the n
th
Fourier polynomial for h (for n 1) is
Hn(x) =

4
+
n
i=1
_
1
i
2
_
cos(i) 1
_
cos(ix) +
(1)
i+1
sin(ix)
i
_
.
This can also be written as
Hn(x) =

4
+
n
i=1
(1)
i+1
sin(ix)
i
+
[
n
2
]
i=1
2
(2i 1)
2
cos((2i 1)x)
where
_
n
2
denotes the biggest integer smaller than or equal to

n
2
. In particular, we have the graphs in Figure 10.5.
'
H
1
(x)
' H
2
(x)
'
H
3
(x)
'
h(x)
x
Figure 10.5
Problems
13. We have f(x) = x, 0 x < 1. Let t = 2x . Notice that as x varies from 0 to 1, t varies from to . Thus
if we rewrite the function in terms of t, we can nd the Fourier series in terms of t in the usual way. To do this, let
g(t) = f(x) = x =
t+
2
on t < . We now nd the fourth degree Fourier polynomial for g.
ao =
1
2
_

g(t) dt =
1
2
_

t +
2
dt =
1
(2)
2
_
t
2
2
+t
_
=
1
2
10.5 SOLUTIONS 229
Notice, a0 is the average value of both f and g. For n 1,
an =
1
t +
2
cos(nt)dt =
1
2
2
_

(t cos(nt) + cos(nt))dt
=
1
2
2
_
t
n
sin(nt) +
1
n
2
cos(nt) +

n
sin(nt)
_
= 0.
bn =
1
t +
2
sin(nt) dt =
1
2
2
_

(t sin(nt) + sin(nt)) dt
=
1
2
2
_
t
n
cos(nt) +
1
n
2
sin(nt)

n
cos(nt)
_
=
1
2
2
(
4
n
cos(n)) =
2
n
cos(n) =
2
n
(1)
n+1
.
We get the integrals for an and bn using the integral table (formulas III-15 and III-16).
Thus, the Fourier polynomial of degree 4 for g is:
G4(t) =
1
2
+
2
sin t
1
sin 2t +
2
3
sin 3t
1
2
sin 4t.
Now, since g(t) = f(x), the Fourier polynomial of degree 4 for f can be found by replacing t in terms of x again. Thus,
F4(x) =
1
2
+
2
sin(2x )
1
sin(4x 2) +
2
3
sin(6x 3)
1
2
sin(8x 4).
Now, using the fact that sin(x ) = sin x and sin(x 2) = sin x, etc., we have:
F4(x) =
1
2

2
sin(2x)
1
sin(4x)
2
3
sin(6x)
1
2
sin(8x).
1
1 f(x) = x
F
4
(x)
x
17. Let f(x) = a
k
cos kx +b
k
sin kx. Then the energy of f is given by
1
(f(x))
2
dx =
1
(a
k
cos kx +b
k
sin kx)
2
dx
=
1
(a
2
k
cos
2
kx 2a
k
b
k
cos kxsin kx +b
2
k
sin
2
kx) dx
=
1
_
a
2
k
_

cos
2
kxdx 2a
k
b
k
_

cos kxsin kxdx +b

2
k
_

sin
2
kxdx
_
=
1
_
a
2
k
2a
k
b
k
0 +b
2
k
= a
2
k
+b
2
k
.
21. (a)
3 2
1
5
1
5
2 3
1
x
f(x)
The energy of the pulse train f is
E =
1
(f(x))
2
dx =
1
_
1/5
1/5
1
2
dx =
1
_
1
5

_
1
5
__
=
2
5
.
Next, nd the Fourier coefcients:
a0 = average value of f on [, ] =
1
2
( Area) =
1
2
_
2
5
_
=
1
5
,
a
k
=
1
f(x) cos kxdx =

1
_
1/5
1/5
cos kxdx =
1
k
sin kx
1/5
1/5
=
1
k
_
sin
_
k
5
_
sin
_
k
5
__
=
1
k
_
2 sin
_
k
5
__
,
b
k
=
1
f(x) sin kxdx =

1
_
1/5
1/5
sin kxdx =
1
k
cos kx
1/5
1/5
=
1
k
_
cos
_
k
5
_
cos
_
k
5
__
=
1
k
(0) = 0.
The energy of f contained in the constant term is
A
2
0
= 2a
2
0
= 2
_
1
5
_
2
=
2
25
2
which is
A
2
0
E
=
2/25
2
2/5
=
1
5
0.063662 = 6.3662% of the total.
The fraction of energy contained in the rst harmonic is
A
2
1
E
=
a
2
1
E
=
_
2 sin
1
5
_
2
2
5
0.12563.
The fraction of energy contained in both the constant term and the rst harmonic together is
A
2
0
E
+
A
2
1
E
0.06366 + 0.12563 = 0.18929 = 18.929%.
(b) The formula for the energy of the k
th
harmonic is
A
2
k
= a
2
k
+b
2
k
=
_
2 sin
k
5
k
_
2
+ 0
2
=
4 sin
2 k
5
k
2
2
.
By graphing this formula as a continuous function for k 1, we see its overall behavior as k gets larger in Figure 10.6.
The energy spectrum for the rst ve terms is shown in Figure 10.7.
10.5 SOLUTIONS 231
1 10 20 30 40
0.005
0.01
0.015
0.02
k
y
y =
4 sin
2 k
5
k
2
2
Figure 10.6
1 2 3 4 5
2
5
2
0.005
0.01
0.015
0.02
k
A
2
k
0
Figure 10.7
(c) The constant term and the rst ve harmonics contain
A
2
0
E
+
A
2
1
E
+
A
2
2
E
+
A
2
3
E
+
A
2
4
E
+
A
2
5
E
61.5255%
of the total energy of f.
(d) The fth Fourier approximation to f is
F5(x) =
1
5
+
2 sin(
1
5
)
cos x +
sin(
2
5
)
cos 2x +
2 sin(
3
5
)
3
cos 3x +
sin(
4
5
)
2
cos 4x +
2 sin 1
5
cos 5x.
3 2
1
5
1
5
2 3
1
x
f(x)
F
5
(x)
For comparison, below is the thirteenth Fourier approximation to f.
3 2
1
5
1
5
2 3
1
x
'
F
13
(x)
25. We make the substitution u = mx, dx =
1
m
du. Then
_

cos
2
mxdx =
1
m
_
u=m
u=m
cos
2
udu.
By Formula IV-18 of the integral table, this equals
1
m
_
1
2
cos usin u
_
m
m
+
1
m
1
2
_
m
m
1 du = 0 +
1
2m
u
m
m
=
1
2m
u
m
m
=
1
2m
(2m) = .
29. (a) To show that g(t) is periodic with period 2, we calculate
g(t + 2) = f
_
b(t + 2)
2
_
= f
_
bt
2
+b
_
= f
_
bt
2
_
= g(t).
Since g(t + 2) = g(t) for all t, we know that g(t) is periodic with period 2. In addition
g
_
2x
b
_
= f
_
b(2x/b)
2
_
= f(x).
(b) We make the change of variable t = 2x/b, dt = (2/b)dx in the usual formulas for the Fourier coefcients of
g(t), as follows:
a0 =
1
2
_

t=
g(t)dt =
1
2
_
b/2
x=b/2
g
_
2x
b
_
2
b
dx =
1
b
_ b
2
b
2
f(x) dx
a
k
=
1
_

t=
g(t) cos(kt) dt =
1
_
b/2
x=b/2
g
_
2x
b
_
cos
_
2kx
b
_
2
b
dx
=
2
b
_
b/2
b/2
f(x) cos
_
2kx
b
_
dx
b
k
=
1
_

t=
g(t) sin(kt) dt =
1
_
b/2
x=b/2
g
_
2x
b
_
sin
_
2kx
b
_
2
b
dx
=
2
b
_
b/2
b/2
f(x) sin
_
2kx
b
_
dx
(c) By part (a), the Fourier series for f(x) can be obtained by substituting t = 2x/b into the Fourier series for g(t)
which was found in part (b).
Solutions for Chapter 10 Review
Exercises
1. e
x
1 +e(x 1) +
e
2
(x 1)
2
5. f
(x) = 3x
2
+ 14x 5, f
(x) = 6x + 14, f
(x) = 6. The Taylor polynomial about x = 1 is

P3(x) = 4 +
12
1!
(x 1) +
20
2!
(x 1)
2
+
6
3!
(x 1)
3
= 4 + 12(x 1) + 10(x 1)
2
+ (x 1)
3
.
Notice that if you multiply out and collect terms in P3(x), you will get f(x) back.
9. Substituting y = t
2
in sin y = y
y
3
3!
+
y
5
5!

y
7
7!
+ gives
sin t
2
= t
2
t
6
3!
+
t
10
5!

t
14
7!
+
13. We use the binomial series to expand 1/
1 z
2
and multiply by z
2
. Since
1
1 +x
= (1 +x)
1/2
= 1
1
2
x +
(1/2)(3/2)
2!
x
2
+
(1/2)(3/2)(5/2)
3!
x
3
+
= 1
1
2
x +
3
8
x
2
5
16
x
3
+ .
Substituting x = z
2
gives
z
2
1 z
2
= 1
1
2
(z
2
) +
3
8
(z
2
)
2
5
16
(z
2
)
3
+
= 1 +
1
2
z
2
+
3
8
z
4
+
5
16
z
6
+ .
Multiplying by z
2
, we have
z
2
1 z
2
= z
2
+
1
2
z
4
+
3
8
z
6
+
5
16
z
8
+ .
SOLUTIONS to Review Problems for Chapter Ten 233
17.
a
a +b
=
a
a(1 +
b
a
)
=
_
1 +
b
a
_
1
= 1
b
a
+
_
b
a
_
2
_
b
a
_
3
+
Problems
21. Factoring out a 3, we see
3
_
1 + 1 +
1
2!
+
1
3!
+
1
4!
+
1
5!
+
_
= 3e
1
= 3e.
25. This is the series for sin x with x = 2 substituted. Thus
2
8
3!
+
32
5!

128
7!
+ = 2
2
3
3!
+
2
5
5!

2
7
7!
+ = sin 2.
29. The second degree Taylor polynomial for f(x) around x = 3 is
f(x) f(3) +f
(3)(x 3) +
f
(3)
2!
(x 3)
2
= 1 + 5(x 3)
10
2!
(x 3)
2
= 1 + 5(x 3) 5(x 3)
2
.
Substituting x = 3.1, we get
f(3.1) 1 + 5(3.1 3) 5(3.1 3)
2
= 1 + 5(0.1) 5(0.01) = 1.45.
33. (a) The series for
sin 2
is
sin 2
=
1
_
2
(2)
3
3!
+
(2)
5
5!

_
= 2
4
2
3
+
4
4
15

so lim
0
sin 2
= 2.
(b) Near = 0, we make the approximation
sin 2
2
4
3
2
so the parabola is y = 2
4
3
2
.
37.
1 1
1
y = e
x
2
y =
1
1+x
2
x
y
(a)
e
x
2
= 1 x
2
+
x
4
2!

x
6
3!
+
1
1 +x
2
= 1 x
2
+x
4
x
6
+
Notice that the rst two terms are the same in both series.
(b)
1
1 +x
2
is greater.
(c) Even, because the only terms involved are of even degree.
(d) The coefcients for e
x
2
become extremely small for higher powers of x, and we can counteract the effect of these
powers for large values of x. The series for
1
1+x
2
has no such coefcients.
41. (a) If h is much smaller than R, we can say that (R +h) R, giving the approximation
F =
mgR
2
(R +h)
2

mgR
2
R
2
= mg.
(b)
F =
mgR
2
(R +h)
2
=
mg
(1 +h/R)
2
= mg(1 +h/R)
2
= mg
_
1 +
(2)
1!
_
h
R
_
+
(2)(3)
2!
_
h
R
_
2
+
(2)(3)(4)
3!
_
h
R
_
3
+
_
= mg
_
1
2h
R
+
3h
2
R
2

4h
3
R
3
+
_
(c) The rst order correction comes from term 2h/R. The approximation for F is then given by
F mg
_
1
2h
R
_
.
If the rst order correction alters the estimate for F by 10%, we have
2h
R
= 0.10 so h = 0.05R 0.05(6400) = 320 km.
The approximation F mg is good to within 10% that is, up to about 300 km.
45. The situation is more complicated. Lets rst consider the case when g
(0) = 0. To be specic let g
(0) > 0. Then

g(x) P3(x) = g(0) +
g
(0)
3!
x
3
.
So, g(x) g(0)
g
(0)
3!
x
3
. (Notice that
g
(0)
3!
> 0 is a constant.) Now, no matter how small an open interval I
around x = 0 is, there are always some x1 and x2 in I such that x1 < 0 and x2 > 0, which means that
g
(0)
3!
x
3
1
< 0
and
g
(0)
3!
x
3
2
> 0, i.e. g(x1) g(0) < 0 and g(x2) g(0) > 0. Thus, g(0) is neither a local minimum nor a local
maximum. (If g
(0) < 0, the same conclusion still holds. Try it! The reasoning is similar.)
Now lets consider the case when g
(0) = 0. If g
(4)
(0) > 0, then by the fourth degree Taylor polynomial approxi-
mation to g at x = 0, we have
g(x) g(0)
g
(4)
(0)
4!
x
4
> 0
for x in a small open interval around x = 0. So g(0) is a local minimum. (If g
(4)
(0) < 0, then g(0) is a local maximum.)
In general, suppose that g
(k)
(0) = 0, k 2, and all the derivatives of g with order less than k are 0. In this case
g looks like cx
k
near x = 0, which determines its behavior there. Then g(0) is neither a local minimum nor a local
maximum if k is odd. For k even, g(0) is a local minimum if g
(k)
(0) > 0, and g(0) is a local maximum if g
(k)
(0) < 0.
49. Since g(x) = f(x +c), we have that [g(x)]
2
= [f(x +c)]
2
, so g
2
is f
2
shifted horizontally by c. Since f has period 2,
so does f
2
and g
2
. If you think of the denite integral as an area, then because of the periodicity, integrals of f
2
over any
interval of length 2 have the same value. So
Energy of f =
_

(f(x))
2
dx =
_
+c
+c
(f(x))
2
dx.
Now we know that
Energy of g =
1
(g(x))
2
dx
=
1
(f(x +c))
2
dx.
Using the substitution t = x +c, we see that the two energies are equal.
CHECK YOUR UNDERSTANDING 235
CAS Challenge Problems
53. (a) The Taylor polynomial is
P10(x) = 1 +
x
2
12

x
4
720
+
x
6
30240

x
8
1209600
+
x
10
47900160
(b) All the terms have even degree. A polynomial with only terms of even degree is an even function. This suggests that
f might be an even function.
(c) To show that f is even, we must show that f(x) = f(x).
f(x) =
x
e
x
1
+
x
2
=
x
1
1
e
x
x
2
=
xe
x
e
x
1

x
2
=
xe
x
1
2
x(e
x
1)
e
x
1
=
xe
x
1
2
xe
x
+
1
2
x
e
x
1
=
1
2
xe
x
+
1
2
x
e
x
1
=
1
2
x(e
x
1) +x
e
x
1
=
1
2
x +
x
e
x
1
=
x
e
x
1
+
x
2
= f(x)
CHECK YOUR UNDERSTANDING
1. False. For example, both f(x) = x
2
and g(x) = x
2
+x
3
have P2(x) = x
2
.
5. False. The Taylor series for sin x about x = is calculated by taking derivatives and using the formula
f(a) +f
(a)(x a) +
f
(a)
2!
(x a)
2
+ .
The series for sin x about x = turns out to be
(x ) +
(x )
3
3!

(x )
5
5!
+ .
9. False. The derivative of f(x)g(x) is not f
(x)g
(x). If this statement were true, the Taylor series for (cos x)(sin x) would
have all zero terms.
13. True. For large x, the graph of P10(x) looks like the graph of its highest powered term, x
10
/10!. But e
x
grows faster than
any power, so e
x
gets further and further away from x
10
/10! P10(x).
17. True. Since f is even, f(x) sin(mx) is odd for any m, so
bm =
1
f(x) sin x(mx) dx = 0.

21. False. The quadratic approximation to f1(x)f2(x) near x = 0 is
f1(0)f2(0) + (f
1
(0)f2(0) +f1(0)f
2
(0))x +
f
1
(0)f2(0) + 2f
1
(0)f
2
(0) +f1(0)f
2
(0)
2
x
2
.
On the other hand, we have
L1(x) = f1(0) +f
1
(0)x, L2(x) = f2(0) +f
2
(0)x,
so
L1(x)L2(x) = (f1(0) +f
1
(0)x)(f2(0) +f
2
(0)x) = f1(0)f2(0) + (f
1
(0)f2(0) +f
2
(0)f1(0))x +f
1
(0)f
2
(0)x
2
.
The rst two terms of the right side agree with the quadratic approximation to f1(x)f2(x) near x = 0, but the term of
degree 2 does not.
For example, the linear approximation to e
x
is 1+x, but the quadratic approximation to (e
x
)
2
= e
2x
is 1+2x+2x
2
,
not (1 +x)
2
= 1 + 2x +x
2
.
11.1 SOLUTIONS 237
CHAPTER ELEVEN
Exercises
1. (a) = (III), (b) = (IV), (c) = (I), (d) = (II).
5. In order to prove that y = A+Ce
kt
is a solution to the differential equation
dy
dt
= k(y A),
we must show that the derivative of y with respect to t is in fact equal to k(y A):
y = A+Ce
kt
dy
dt
= 0 + (Ce
kt
)(k)
= kCe
kt
= k(Ce
kt
+AA)
= k
_
(Ce
kt
+A) A
_
= k(y A).
9. Differentiating and using the fact that
d
dt
(cosh t) = sinh t and
d
dt
(sinh t) = cosh t,
we see that
dx
dt
= C1 sinh t +C2 cosh t
d
2
x
dt
2
=
2
C1 cosh t +
2
C2 sinh t
=
2
(C1 cosh t +C2 sinh t) .
Therefore, we see that
d
2
x
dt
2
=
2
x.
Problems
13. Since y = x
2
+k, we know that y
= 2x. Substituting y = x
2
+k and y
= 2x into the differential equation, we get

10 = 2y xy
= 2(x
2
+k) x(2x)
= 2x
2
+ 2k 2x
2
= 2k.
Thus, k = 5 is the only solution.
238 Chapter Eleven /SOLUTIONS
17. (a) If y = Cx
n
is a solution to the given differential equation, then we must have
x
d (Cx
n
)
dx
3(Cx
n
) = 0
x(Cnx
n1
) 3(Cx
n
) = 0
Cnx
n
3Cx
n
= 0
C(n 3)x
n
= 0.
Thus, if C = 0, we get y = 0 is a solution, for every n. If C = 0, then n = 3, and so y = Cx
3
is a solution.
(b) Because y = 40 for x = 2, we cannot have C = 0. Thus, by part (a), we get n = 3. The solution to the differential
equation is
y = Cx
3
.
To determine C if y = 40 when x = 2, we substitute these values into the equation.
40 = C 2
3
40 = C 8
C = 5.
So, now both C and n are xed at specic values.
21.
(I) y = e
x
, y
= e
x
, y
= e
x
(II) y = x
3
, y
= 3x
2
, y
= 6x
(III) y = e
x
, y
= e
x
, y
= e
x
(IV) y = x
2
, y
= 2x
3
, y
= 6x
4
and so:
(a) (I),(III) because y
= y in each case.
(b) (IV) because x
2
y
+ 2xy
2y = x
2
(6x
4
) + 2x(2x
3
) 2x
2
= 6x
2
4x
2
2x
2
= 0.
(c) (II),(IV) because x
2
y
= 6y in each case.
Exercises
1. See Figure 11.1. Other choices of solution curves are, of course, possible.
y
x
y
x
Figure 11.1
11.3 SOLUTIONS 239
5. (a) See Figure 11.2.
4 4
4
4
x
y
i
iii ii
Figure 11.2
(b) The solution through (1, 0) appears to be linear, so its equation is y = x 1.
(c) If y = x 1, then y
= 1 and x +y = x + (x 1) = 1, so this checks as a solution.

Problems
9. (a) and (b) See Figure 11.3
y
x
(i)
(ii)
(iii)
(iv)
(v)
(vi)
Figure 11.3
(c) Figure 11.3 shows that a solution will be increasing if its y-values fall in the range 1 < y < 2. This makes sense
since if we examine the equation y
= 0.5(1 +y)(2 y), we will nd that y
> 0 if 1 < y < 2. Notice that if the

y-value ever gets to 2, then y
= 0 and the function becomes constant, following the line y = 2. (The same is true if
ever y = 1.)
From the graph, the solution is decreasing if y > 2 or y < 1. Again, this also follows from the equation, since
in either case y
< 0.
The curve has a horizontal tangent if y
= 0, which only happens if y = 2 or y = 1. This also can be seen on

the graph in Figure 11.3.
13. (a) II (b) VI (c) IV (d) I (e) III (f) V
Exercises
1. (a) In Table 11.1, we see that y(0.4) 1.5282.
(b) In Table 11.2, we see that y(0.4) = 1.4. (This answer is exact.)
Table 11.1 Eulers method for
y
= x +y with y(0) = 1
x y y =(slope)x
0 1 0.1 = (1)(0.1)
0.1 1.1 0.12 = (1.2)(0.1)
0.2 1.22 0.142 = (1.42)(0.1)
0.3 1.362 0.1662 = (1.662)(0.1)
0.4 1.5282
y
= x +y with y(1) = 0
x y y =(slope)x
1 0 0.1 = (1)(0.1)
0.9 0.1 0.1 = (1)(0.1)
0.8 0.2 0.1 = (1)(0.1)
0.7 0.3
.
.
.
.
.
. Notice that y
0 1 decreases by 0.1
.
.
.
.
.
. for every step
0.4 1.4
Problems
5. (a) x = 0.5
y
= 2x, with y(0) = 1

x y y =(slope)x
0 1 0 = (2 0)(0.5)
0.5 1 0.5 = (2 0.5)(0.5)
1 1.5
x = 0.25
Table 11.4 Eulers method for y
= 2x,
with y(0) = 1
x y y =(slope)x
0 1 0 = (2 0)(0.25)
0.25 1 0.125 = (2 0.25)(0.25)
0.50 1.125 0.25 = (2 0.5)(0.25)
0.75 1.375 0.375 = (2 0.75)(0.25)
1 1.75
(b) General solution is y = x
2
+C, and y(0) = 1 gives C = 1. Thus, the solution is y = x
2
+ 1. So the true value of y
when x = 1 is y = 1
2
+ 1 = 2.
(c) When x = 0.5, error = 0.5.
When x = 0.25, error = 0.25.
Thus, decreasing x by a factor of 2 has decreased the error by a factor of 2, as expected.
9. By looking at the slope elds, or by computing the second derivative
d
2
y
dx
2
= 2x 2y
dy
dx
= 2x 2x
2
y + 2y
3
,
we see that the solution curve is concave up, so Eulers method gives an underestimate.
11.4 SOLUTIONS 241
13. Assume that x > 0 and that we use n steps in Eulers method. Label the x-coordinates we use in the process
x0, x1, . . . , xn, where x0 = 0 and xn = x. Then using Eulers method to nd y(x), we get
Table 11.5
x y y = (slope)x
P
0
0 = x
0
0 f(x
0
)x
P
1
x
1
f(x
0
)x f(x
1
)x
P
2
x
2
f(x
0
)x + f(x
1
)x f(x
2
)x
.
.
.
.
.
.
.
.
.
.
.
.
Pn x = xn
n1
i=0
f(x
i
)x
Thus the result from Eulers method is
n1
i=0
f(xi)x. We recognize this as the left-hand Riemann sum that approxi-
mates
_
x
0
f(t) dt.
Exercises
1. Separating variables gives
_
1
P
dP =
_
2dt,
so
ln |P| = 2t +C.
Therefore
P = e
2t+C
= Ae
2t
.
The initial value P(0) = 1 gives 1 = A, so
P = e
2t
.
_
dy
y
=
_
1
3
dx
ln |y| =
1
3
x +C.
Solving for y, we have
y = Ae
1
3
x
, where A = e
C
.
Since y(0) = A = 10, we have
y = 10e
1
3
x
.
_
dz
z
=
_
5 dt
ln |z| = 5t +C.
Solving for z, we have
z = Ae
5t
, where A = e
C
.
Using the fact that z(1) = 5, we have z(1) = Ae
5
= 5, so A = 5/e
5
. Therefore,
z =
5
e
5
e
5t
= 5e
5t5
.
_
dy
y 200
=
_
0.5dt
ln |y 200| = 0.5t +C
y = 200 +Ae
0.5t
, where A = e
C
.
The initial condition, y(0) = 50, gives
50 = 200 +A, so A = 150.
Thus,
y = 200 150e
0.5t
.
17. Factoring out the 0.1 gives
dm
dt
= 0.1m+ 200 = 0.1(m+ 2000)
_
dm
m+ 2000
=
_
0.1 dt,
so
ln |m+ 2000| = 0.1t +C,
and
m = Ae
0.1t
2000, where A = e
C
.
Using the initial condition, m(0) = Ae
(0.1)0
2000 = 1000, gives A = 3000. Thus
m = 3000e
0.1t
2000.
dz
dt
= te
z
e
z
dz = tdt
_
e
z
dz =
_
t dt,
so
e
z
=
t
2
2
+C.
Since the solution passes through the origin, z = 0 when t = 0, we must have
e
0
=
0
2
+C, so C = 1.
Thus
e
z
=
t
2
2
1,
or
z = ln
_
1
t
2
2
_
.
11.4 SOLUTIONS 243
dw
d
= w
2
sin
2
_
dw
w
2
=
_
sin
2
d,
so
1
w
=
1
2
cos
2
+C.
According to the initial conditions, w(0) = 1, so 1 =
1
2
+C and C =
1
2
. Thus,
1
w
=
1
2
cos
2
1
2
1
w
=
cos
2
+ 1
2
w =
2
cos
2
+ 1
.
Problems
_
dR
R
=
_
k dt.
Integrating gives
ln |R| = kt +C,
so
|R| = e
kt+C
= e
kt
e
C
R = Ae
kt
, where A = e
C
or A = 0.
_
dP
P a
=
_
k dt.
Integrating yields
ln |P a| = kt +C,
so
P = a +Ae
kt
where A = e
C
or A = 0.
37. Separating variables and integrating gives
_
1
R
2
+ 1
dR =
_
adx
or
arctan R = ax +C
so that
R = tan(ax +C).
dx
dt
=
xln x
t
,
so
_
dx
xln x
=
_
dt
t
,
and thus
ln | ln x| = ln t +C,
so
| ln x| = e
C
e
ln t
= e
C
t.
Therefore
ln x = At, where A = e
C
or A = 0, so x = e
At
.
45. (a), (b) See Figure 11.4.
x
y
Figure 11.4
(c) Since
dy
dx
=
y
x
,
we have
_
dy
y
=
_
dx
x
,
so
ln |y| = ln |x| +C,
giving
|y| = e
ln |x|+C
= (|x|)
1
e
C
.
Thus,
y =
A
x
, where A = e
C
or A = 0.
Exercises
1. (a) (I)
(b) (IV)
(c) (II) and (IV)
(d) (II) and (III)
11.5 SOLUTIONS 245
5. (a) The equilibriumsolutions occur where the slope y
= 0, which occurs on the slope eld where the lines are horizontal,
or (looking at the equation) at y = 2 and y = 1. Looking at the slope eld, we can see that y = 2 is stable, since
the slopes at nearby values of y point toward it, whereas y = 1 is unstable.
(b) Draw solution curves passing through the given points by starting at these points and following the ow of the slopes,
as shown in Figure 11.5.
x
y
Figure 11.5
Problems
9. (a) Separating variables
_
dy
100 y
=
_
dt
ln |100 y| = t +C
|100 y| = e
tC
100 y = (e
C
)e
t
= Ae
t
whereA = e
C
y = 100 Ae
t
.
(b) See Figure 11.6.
1
100
s
C = 25
s
C = 50
s
C = 100
t
y
Figure 11.6
(c) Substituting y = 0 when t = 0 gives
0 = 100 +Ce
0
so C = 100. Thus solution is
y = 100 100e
t
.
13. (a) Since we are told that the rate at which the quantity of the drug decreases is proportional to the amount of the drug
left in the body, we know the differential equation modeling this situation is
dQ
dt
= kQ.
Since we are told that the quantity of the drug is decreasing, we know that k < 0.
(b) We know that the general solution to the differential equation
dQ
dt
= kQ
is
Q = Ce
kt
.
(c) We are told that the half life of the drug is 3.8 hours. This means that at t = 3.8, the amount of the drug in the body
is half the amount that was in the body at t = 0, or, in other words,
0.5Q(0) = Q(3.8).
Solving this equation gives
0.5Q(0) = Q(3.8)
0.5Ce
k(0)
= Ce
k(3.8)
0.5C = Ce
k(3.8)
0.5 = e
k(3.8)
ln(0.5) = k(3.8)
ln(0.5)
3.8
= k
k 0.182.
(d) From part (c) we know that the formula for Q is
Q = Ce
0.182t
.
We are told that initially there are 10 mg of the drug in the body. Thus at t = 0, we get
10 = Ce
0.182(0)
so
C = 10.
Thus our equation becomes
Q(t) = 10e
0.182t
.
Substituting t = 12, we get
Q(t) = 10e
0.182t
Q(12) = 10e
0.182(12)
= 10e
2.184
Q(12) 1.126 mg.
17. According to Newtons Law of Cooling, the temperature, T, of the roast as a function of time, t, satises
T
(t) = k(350 T)
T(0) = 40.
Solving this differential equation, we get that T = 350 310e
kt
for some k > 0. To nd k, we note that at t = 1 we
have T = 90, so
90 = 350 310e
k(1)
260
310
= e
k
k = ln
_
260
310
_
0.17589.
11.5 SOLUTIONS 247
Thus, T = 350 310e
0.17589t
. Solving for t when T = 140, we have
140 = 350 310e
0.17589t
210
310
= e
0.17589t
t =
ln(210/310)
0.17589
t 2.21 hours.
21. Lake Superior will take the longest, because the lake is largest (V is largest) and water is moving through it most slowly
(r is smallest). Lake Erie looks as though it will take the least time because V is smallest and r is close to the largest. For
Erie, k = r/V = 175/460 0.38. The lake with the largest value of r is Ontario, where k = r/V = 209/1600 0.13.
Since e
kt
decreases faster for larger k, Lake Erie will take the shortest time for any xed fraction of the pollution to be
removed.
For Lake Superior,
dQ
dt
=
r
V
Q =
65.2
12,200
Q 0.0053Q
so
Q = Q0e
0.0053t
.
When 80% of the pollution has been removed, 20% remains so Q = 0.2Q0. Substituting gives us
0.2Q0 = Q0e
0.0053t
so
t =
ln(0.2)
0.0053
301 years.
(Note: The 301 is obtained by using the exact value of
r
V
=
65.2
12,200
, rather than 0.0053. Using 0.0053 gives 304 years.)
For Lake Erie, as in the text
dQ
dt
=
r
V
Q =
175
460
Q 0.38Q
so
Q = Q0e
0.38t
.
When 80% of the pollution has been removed
0.2Q0 = Q0e
0.38t
t =
ln(0.2)
0.38
4 years.
So the ratio is
Time for Lake Superior
Time for Lake Erie

301
4
75.
In other words it will take about 75 times as long to clean Lake Superior as Lake Erie.
25. (a) The differential equation is
dT
dt
= k(T A),
where A = 10
F is the outside temperature.

(b) Integrating both sides yields
_
dT
T A
=
_
k dt.
Then ln |T A| = kt +C, so T = A+Be
kt
. Thus
T = 10 + 58e
kt
.
Since 10:00 pm corresponds to t = 9,
57 = 10 + 58e
9k
47
58
= e
9k
ln
47
58
= 9k
k =
1
9
ln
47
58
0.0234.
At 7:00 the next morning (t = 18) we have
T 10 + 58e
18(0.0234)
= 10 + 58(0.66)
48
F,
so the pipes wont freeze.
(c) We assumed that the temperature outside the house stayed constant at 10
F. This is probably incorrect because the

temperature was most likely warmer during the day (between 1 pm and 10 pm) and colder after (between 10 pm and
7 am). Thus, when the temperature in the house dropped from 68
F to 57
F between 1 pm and 10 pm, the outside

temperature was probably higher than 10
F, which changes our calculation of the value of the constant k. The house
temperature will most certainly be lower than 48
F at 7 am, but not by muchnot enough to freeze.

Exercises
1. (a) If B = f(t), where t is in years,
dB
dt
= Rate of money earned from interest + Rate of money deposited
dB
dt
= 0.10B + 1000.
(b) We use separation of variables to solve the differential equation
dB
dt
= 0.1B + 1000.
_
1
0.1B + 1000
dB =
_
dt
1
0.1
ln |0.1B + 1000| = t +C1
0.1B + 1000 = C2e
0.1t
B = Ce
0.1t
10,000
For t = 0, B = 0, hence C = 10,000. Therefore, B = 10,000e
0.1t
10,000.
5. Using (Rate balance changes) = (Rate interest is added) (Rate payments are made), when the interest rate is i, we have
dB
dt
= iB 100.
Solving this equation, we nd:
dB
dt
= i
_
B
100
i
_
_
dB
B
100
i
=
_
i dt
ln
B
100
i
= it +C
B
100
i
= Ae
it
, where A = e
C
.
At time t = 0 we start with a balance of $1000. Thus
1000
100
i
= Ae
0
, so A = 1000
100
i
.
11.6 SOLUTIONS 249
Thus B =
100
i
+ (1000
100
i
)e
it
.
When i = 0.05, B = 2000 1000e
0.05t
.
When i = 0.1, B = 1000.
When i = 0.15, B = 666.67 + 333.33e
0.15t
.
We now look at the graph in Figure 11.7 when i = 0.05, i = 0.1, and i = 0.15.
t 13.86
1000
i = 0.1
i = 0.15
i = 0.05
years
B
Figure 11.7
Problems
9. Let C(t) be the current owing in the circuit at time t, then
dC
dt
= C
where > 0 is the constant of proportionality between the rate at which the current decays and the current itself.
The general solution of this differential equation is C(t) = Ae
t
but since C(0) = 30, we have that A = 30, and
so we get the particular solution C(t) = 30e
t
.
When t = 0.01, the current has decayed to 11 amps so that 11 = 30e
0.01
which gives = 100 ln(11/30) =
100.33 so that,
C(t) = 30e
100.33t
.
13. Let the depth of the water at time t be y. Then
dy
dt
= k
y, where k is a positive constant. Separating variables,

_
dy
y
=
_
k dt,
so
2
y = kt +C .
When t = 0, y = 36; 2
36 = k 0 +C, so C = 12.
When t = 1, y = 35; 2
35 = k + 12, so k 0.17.
Thus, 2
y 0.17t + 12. We are looking for t such that y = 0; this happens when t
12
0.17
71 hours, or about 3
days.
17. Let V (t) be the volume of water in the tank at time t, then
dV
dt
= k
V
This is a separable equation which has the solution
V (t) = (
kt
2
+C)
2
Since V (0) = 200 this gives 200 = C
2
so
V (t) = (
kt
2
+
200)
2
.
However, V (1) = 180 therefore
180 = (
k
2
+
200)
2
,
so that k = 2
_
180
200
_
= 1.45146. Therefore,
V (t) = (0.726t +
200)
2
.
The tank will be half-empty when V (t) = 100, so we solve
100 = (0.726t +
200)
2
to obtain t = 5.7 days. The tank will be half empty in 5.7 days.
The volume after 4 days is V (4) which is approximately 126.32 liters.
21. (a) The balance in the account at the beginning of the month is given by the following sum
_
balance in
account
_
=
_
previous months
balance
_
+
_
interest on
previous months balance
_
+
_
monthly deposit
of $100
_
Denote month is balance by Bi. Assuming the interest is compounded continuously, we have
_
previous months
balance
_
+
_
interest on previous
months balance
_
= Bi1e
0.1/12
.
Since the interest rate is 10% = 0.1 per year, interest is
0.1
12
per month. So at month i, the balance is
Bi = Bi1e
0.1
12
+ 100
Explicitly, we have for the ve years (60 months) the equations:
B0 = 0
B1 = B0e
0.1
12
+ 100
B2 = B1e
0.1
12
+ 100
B3 = B2e
0.1
12
+ 100
.
.
.
.
.
.
B60 = B59e
0.1
12
+ 100
In other words,
B1 = 100
B2 = 100e
0.1
12
+ 100
B3 = (100e
0.1
12
+ 100)e
0.1
12
+ 100
= 100e
(0.1)2
12
+ 100e
0.1
12
+ 100
B4 = 100e
(0.1)3
12
+ 100e
(0.1)2
12
+ 100e
(0.1)
12
+ 100
.
.
.
.
.
.
B60 = 100e
(0.1)59
12
+ 100e
(0.1)58
12
+ + 100e
(0.1)1
12
+ 100
B60 =
59
k=0
100e
(0.1)k
12
(b) The sum B60 =
59
k=0
100e
(0.1)k
12
can be written as B60 =
59
k=0
1200e
(0.1)k
12
(
1
12
) which is the left Riemann sum for
_
5
0
1200e
0.1t
dt, with t =
1
12
and N = 60. Evaluating the sum on a calculator gives B60 = 7752.26.
11.6 SOLUTIONS 251
(c) The situation described by this problem is almost the same as that in Problem 20, except that here the money is being
deposited once a month rather than continuously; however the nominal yearly rates are the same. Thus we would
expect the balance after 5 years to be approximately the same in each case. This means that the answer to part (b)
of this problem should be approximately the same as the answer to part (c) to Problem 20. Since the deposits in this
problem start at the end of the rst month, as opposed to right away, we would expect the balance after 5 years to be
slightly smaller than in Problem 20, as is the case.
Alternatively, we can use the Fundamental Theorem of Calculus to show that the integral can be computed
exactly
_
5
0
1200e
0.1t
dt = 12000(e
(0.1)5
1) = 7784.66
Thus
_
5
0
1200e
0.1t
dt represents the exact solution to Problem 20. Since 1200e
0.1t
is an increasing function, the left
hand sum we calculated in part (b) of this problem underestimates the integral. Thus the answer to part (b) of this
problem should be less than the answer to part (c) of Problem 20.
25. (a)
dQ
dt
= r Q = (Q
r
)
_
dQ
Qr/
= =
_
dt
ln
Q
r
= t +C
Q
r
= Ae
t
When t = 0, Q = 0, so A =
r
and
Q =
r
(1 e
t
)
So,
Q = lim
t
Q =
r
.
r
Q
t
Q =
r
(1 e
t
)
(b) Doubling r doubles Q. Since Q = r/, the time to reach
1
2
Q is obtained by solving
r
2
=
r
(1 e
t
)
1
2
= 1 e
t
e
t
=
1
2
t =
ln(1/2)
=
ln 2
.
So altering r does not alter the time it takes to reach
1
2
Q. See Figure 11.8.
ln 2
2r
Q
t
Q =
r
(1 e
t
)
Q =
2r
(1 e
t
)
Figure 11.8
(c) Q is halved by doubling , and so is the time, t =
ln 2
, to reach
1
2
Q.
29. (a) Newtons Law of Motion says that
Force = (mass) (acceleration).
Since acceleration, dv/dt, is measured upward and the force due to gravity acts downward,
mgR
2
(R +h)
2
= m
dv
dt
so
dv
dt
=
gR
2
(R +h)
2
.
(b) Since v =
dh
dt
, the chain rule gives
dv
dt
=
dv
dh

dh
dt
=
dv
dh
v.
Substituting into the differential equation in part (a) gives
v
dv
dh
=
gR
2
(R +h)
2
.
(c) Separating variables gives
_
v dv =
_
gR
2
(R +h)
2
dh
v
2
2
=
gR
2
(R +h)
+C
Since v = v0 when h = 0,
v0
2
2
=
gR
2
(R + 0)
+C gives C =
v0
2
2
gR,
so the solution is
v
2
2
=
gR
2
(R +h)
+
v0
2
2
gR
v
2
= v0
2
+
2gR
2
(R +h)
2gR
(d) The escape velocity v0 ensures that v
2
0 for all h 0. Since the positive quantity
2gR
2
(R +h)
0 as h , to
ensure that v
2
0 for all h, we must have
v0
2
2gR.
When v0
2
= 2gR so v0 =
2gR, we say that v0 is the escape velocity.

11.7 SOLUTIONS 253
Exercises
1. We see from the differential equation that k = 0.05 and L = 2800, so the general solution is
P =
2800
1 +Ae
0.05t
.
5. We rewrite
10P 5P
2
= 10P
_
1
P
2
_
,
so k = 10 and L = 2. Since P0 = L/4, we have A = (L P0)/P0 = 3. Thus
P =
2
1 + 3e
10t
.
The time to peak dP/dt is
t =
1
k
ln A = ln(3)/10.
9. (a) We see that k = 0.035 which tells us that the quantity P grows by about 3.5% per unit time when P is very small
relative to L. We also see that L = 6000 which tells us the upper limit on the value of P if P is initially below 6000.
(b) The largest rate of change occurs when P = L/2 = 3000.
Problems
13. A continuous growth rate of 0.2% means that
1
P
dP
dt
= 0.2% = 0.002.
Separating variables and integrating gives
_
dP
P
=
_
0.002 dt
P = P0e
0.002t
= (6.6 10
6
)e
0.002t
.
17. (a) For 1800 we have
1
P
dP
dt
=
1
P(1800)
P(1810) P(1790)
20
=
1
5.3
7.2 3.9
20
= 0.0311.
For 1930 we have
1
P
dP
dt
=
1
P(1930)
P(1940) P(1920)
20
=
1
122.8
131.7 105.7
20
= 0.0106.
(b) The slope between these points is
Slope =
0.0311 0.0106
5.3 122.8
= 0.000174.
Thus the tted line has an equation of the form
1
P
dP
dt
= k 0.000174P.
Using the point P = 5.3, (1/P)dP/dt = 0.0311, we solve for the vertical intercept k:
0.0311 = k 0.000174(5.3)
k = 0.0311 + 0.000174(5.3) = 0.032.
Thus k/L = 0.000174, so L = 0.0320/.000174 = 184.
(c) These are quite close to the values given in the text.
21. By rewriting the equation, we see that it is logistic:
1
P
dP
dt
=
(100 P)
1000
.
Before looking at its solution, we explain why there must always be at least 100 individuals. Since the population begins
at 200, the quantity dP/dt is initially negative, so the population initially decreases. It continues to do so while P > 100.
If the population ever reached 100, then dP/dt would be 0. This would mean the population stopped changingso if the
population ever decreased to 100, thats where it would stay. The fact that dP/dt is always negative for P > 100 also
shows that the population is always under 200, as shown in Figure 11.9.
100
200
t
P
Figure 11.9
The solution, as given by the formula derived in the chapter, is
P =
100
1 0.5e
0.1t
25. (a) Let I be the number of informed people at time t, and I0 the number who know initially. Then this model predicts
that dI/dt = k(M I) for some positive constant k. Solving this, we nd the solution is
I = M (M I0)e
kt
.
We sketch the solution with I0 = 0. Notice that dI/dt is largest when I is smallest, so the information spreads fastest
in the beginning, at t = 0. In addition, Figure 11.10 shows that I M as t , meaning that everyone gets the
information eventually.
(b) In this case, the model suggests that dI/dt = kI(M I) for some positive constant k. This is a logistic model with
carrying capacity M. We sketch the solutions for three different values of I0 in Figure 11.11.
M
t
I
Figure 11.10
I
0
= 0
I
0
= 0.05M
I
0
= 0.75M
0.5M
M
t
I
Figure 11.11
(i) If I0 = 0 then I = 0 for all t. In other words, if nobody knows something, it does not spread by word of mouth!
(ii) If I0 = 0.05M, then dI/dt is increasing up to I = M/2. Thus, the information is spreading fastest at I = M/2.
(iii) If I0 = 0.75M, then dI/dt is always decreasing for I > M/2, so dI/dt is largest when t = 0.
11.8 SOLUTIONS 255
b
2a
b
a
0 P
dP
dt
Figure 11.12
b
a
b
2a
P
t
'
inflection point
threshold
Figure 11.13
Figure 11.12 shows that dP/dt is negative for P <
b
a
, making P a decreasing function when P(0) <
b
a
. When
P >
b
a
, the sign of dP/dt is positive, so P is an increasing function. Thus solution curves starting above
b
a
are
increasing, and those starting below
b
a
are decreasing. See Figure 11.13.
For P >
b
a
, the slope,
dP
dt
, increases with P, so the graph of P against t is concave up. For 0 < P <
b
a
, the
value of P decreases with time. As P decreases, the slope
dP
dt
decreases for
b
2a
< P <
b
a
, and increases toward 0 for
0 < P <
b
2a
. Thus solution curves starting just below the threshold value of
b
a
are concave down for
b
2a
< P <
b
a
and concave up and asymptotic to the t-axis for 0 < P <
b
2a
. See Figure 11.13.
(c) P =
b
a
is called the threshold population because for populations greater than
b
a
, the population will increase without
bound. For populations less than
b
a
, the population will go to zero, i.e. to extinction.
33. The population dies out if H is large enough that dP/dt < 0 for all P. The largest value for dP/dt = kP(1 P/L)
occurs when P = L/2; then
dP
dt
= kP
_
1
P
L
_
= k
L
2
_
1
L/2
L
_
=
kL
4
.
Thus if H > kL/4, we have dP/dt < 0 for all P and the population dies out if the quota is met.
Exercises
1. Since
dS
dt
= aSI,
dI
dt
= aSI bI,
dR
dt
= bI
we have
dS
dt
+
dI
dt
+
dR
dt
= aSI +aSI bI +bI = 0.
Thus
d
dt
(S +I +R) = 0, so S +I +R = constant.
5. If w = 2 and r = 2, then
dw
dt
= 2 and
dr
dt
= 2, so initially the number of worms decreases and the number of robins
increases. In the long run, however, the populations will oscillate; they will even go back to w = 2 and r = 2. See
Figure 11.14.
1 2 3
1
2
3
w (worms in millions)
r (robins in thousands)
(2500 robins)
Figure 11.14
1 2 3
1
2
3
w
r
(3, 1)
Figure 11.15
9. The numbers of robins begins to increase while the number of worms remains approximately constant. See Figure 11.15.
The numbers of robins and worms oscillate periodically between 0.2 and 3, with the robin population lagging behind
the worm population.
13. x decreases quickly while y increases more slowly.
Problems
17. (a) The x population is unaffected by the y populationit grows exponentially no matter what the y population is, even
if y = 0. If alone, the y population decreases to zero exponentially, because its equation becomes dy/dt = 0.1y.
(b) Here, interaction between the two populations helps the y population but does not effect the x population. This is not
a predator-prey relationship; instead, this is a one-way relationship, where the y population is helped by the existence
of xs. These equations could, for instance, model the interaction of rhinoceroses (x) and dung beetles (y).
21. (a) Thinking of y as a function of x and x as a function of t, then by the chain rule:
dy
dt
=
dy
dx
dx
dt
, so:
dy
dx
=
dy/dt
dx/dt
=
0.01x
0.05y
=
x
5y
10 20 30 40 50 60
10
20
30
x (thousand US troops)
y (thousand Japanese troops)
(b) The gure above shows the slope eld for this differential equation and the trajectory starting at x0 = 54, y0 = 21.5.
The trajectory goes to the x-axis, where y = 0, meaning that the Japanese troops were all killed or wounded before
the US troops were, and thus predicts the US victory (which did occur). Since the trajectory meets the x-axis at
x 25, the differential equation predicts that about 25,000 US troops would survive the battle.
(c) The fact that the US got reinforcements, while the Japanese did not, does not alter the predicted outcome (a US
victory). The US reinforcements have the effect of changing the trajectory, altering the number of troops surviving
the battle. See the graph below.
11.9 SOLUTIONS 257
10 20 30 40 50 60
10
20
30
x (thousand US troops)
y (thousand Japanese troops)
25. (a) Taking the constants of proportionality to be a and b, with a > 0 and b > 0, the equations are
dx
dt
= axy
dy
dt
= bxy
(b)
dy
dx
=
dy/dt
dx/dt
=
bxy
axy
=
b
a
. Solving the differential equation gives y =
b
a
x + C, where C depends on the initial
sizes of the two armies.
(c) The sign of C determines which side wins the battle. Looking at the general solution y =
b
a
x + C, we see that if
C > 0 the y-intercept is at C, so y wins the battle by virtue of the fact that it still has troops when x = 0. If C < 0
then the curve intersects the axes at x =
a
b
C, so x wins the battle because it has troops when y = 0. If C = 0, then
the solution goes to the point (0, 0), which represents the case of mutual annihilation.
(d) We assume that an army wins if the opposing force goes to 0 rst.
1 2 3 4
1
2
3
4
x (guerrilla)
y (guerrilla)
C > 0
y wins
C < 0
x wins
y = bx/a
(i.e. C = 0)
Exercises
1. (a) To nd the equilibrium points we set
20x 10xy = 0
25y 5xy = 0.
So, x = 0, y = 0 is an equilibrium point. Another one is given by
10y = 20
5x = 25.
Therefore, x = 5, y = 2 is the other equilibrium point.
(b) At x = 2, y = 4,
dx
dt
= 20x 10xy = 40 80 = 40
dy
dt
= 25y 5xy = 100 40 = 60.
Since these are not both zero, this point is not an equilibrium point.
Problems
5. We rst nd the nullclines. Vertical nullclines occur where
dx
dt
= 0, which happens when x = 0 or y =
1
3
(2 x).
Horizontal nullclines occur where
dy
dt
= y(1 2x) = 0, which happens when y = 0 or x =
1
2
. These nullclines are
shown in Figure 11.16.
Equilibrium points (also shown in Figure 11.16) occur at the intersections of vertical and horizontal nullclines. There
are three such points for this system of equations; (0, 0), (
1
2
,
1
2
) and (2, 0).
The nullclines divide the positive quadrant into four regions as shown in Figure 11.16. Trajectory directions for these
regions are shown in Figure 11.17.
2 1/2
2/3
y
x
I II
III IV
Figure 11.16: Nullclines and
equilibrium points (dots)
2 1/2
2/3
y
x

E E
T
T
Figure 11.17: General directions of
trajectories and equilibrium points
(dots)
9. We assume that x, y 0 and then nd the nullclines.
dx
dt
= x(1
x
2
y) = 0 when x = 0 or y +
x
2
= 1.
dy
dt
= y(1
y
3
x) = 0 when y = 0 or x +
y
3
= 1.
We nd the equilibrium points. They are (2, 0), (0, 3), (0, 0), and (
4
5
,
3
5
). The nullclines and equilibrium points are shown
in Figure 11.18.
1 2
1
3
x
y
'
x + y/3 = 1
dy/dt = 0
y = 1 x/2
dx/dt = 0
(
4
5
,
3
5
)
Figure 11.18: Nullclines and equilibrium points
(dots)
1 2
1
3
x
y
(I)
(IV)
(III)
(II)
(
4
5
,
3
5
)
Figure 11.19: General directions of trajectories
and equilibrium points (dots)
Figure 11.19 shows that if the initial point is in sector (I), the trajectory heads toward the equilibrium point (0, 3).
Similarly, if the trajectory begins in sector (III), then it heads toward the equilibrium (2, 0) over time. If the trajectory
begins in sector (II) or (IV), it can go to any of the three equilibrium points (2, 0), (0, 3), or (
4
5
,
3
5
).
dx
dt
= 0 when x =
10.5
0.45
= 23.3
dy
dt
= 0 when 8.2x 0.8y 142 = 0
11.10 SOLUTIONS 259
There is an equilibrium point where the trajectories cross at x = 23.3, y = 61.7
In region I,
dx
dt
> 0,
dy
dt
< 0.
In region II,
dx
dt
< 0,
dy
dt
< 0.
In region III,
dx
dt
< 0,
dy
dt
> 0.
In region IV,
dx
dt
> 0,
dy
dt
> 0.
23.3
61.7
x (Soviet)
y (US)
Region I
E
c
Region II
'
c
Region III
s
'
T
Region IV
E
T
'
dx
dt
= 0
E
dy
dt
= 0
'
Figure 11.20: Nullclines and equilibrium point (dot) for
US-Soviet arms race
23.3
61.7
x (Soviet)
y (US)
Figure 11.21: Trajectories for US-Soviet arms race.
(c) All the trajectories tend toward the equilibrium point x = 23.3, y = 61.7. Thus the model predicts that in the long
run the arms race will level off with the Soviet Union spending 23.3 billion dollars a year on arms and the US 61.7
billion dollars.
(d) As the model predicts, yearly arms expenditure did tend toward 23 billion for the Soviet Union and 62 billion for the
US.
Exercises
1. If y = 2 cos t + 3 sin t, then y
= 2 sin t + 3 cos t and y
= 2 cos t 3 sin t. Thus, y
+y = 0.
5. If y(t) = Asin(t) +Bcos(t) then
y
= Acos(t) Bsin(t)
y
=
2
Asin(t)
2
Bcos(t)
therefore
y
+
2
y =
2
Asin(t)
2
Bcos(2t) +
2
(Asin(t) +Bcos(t)) = 0
for all values of A and B, so the given function is a solution.
9. The amplitude is
3
2
+ 7
2
=
58.
Problems
13. First, we note that the solutions of:
(a) x
+x = 0 are x = Acos t +Bsin t;

(b) x
+ 4x = 0 are x = Acos 2t +Bsin 2t;

(c) x
+ 16x = 0 are x = Acos 4t +Bsin 4t.

This follows from what we know about the general solution to x
+
2
x = 0.
The period of the solutions to (a) is 2, the period of the solutions to (b) is , and the period of the solutions of (c) is

2
.
Since the t-scales are the same on all of the graphs, we see that graphs (I) and (IV) have the same period, which is twice
the period of graph (III). Graph (II) has twice the period of graphs (I) and (IV). Since each graph represents a solution, we
have the following:
equation (a) goes with graph (II)
equation (b) goes with graphs (I) and (IV)
equation (c) goes with graph (III)
The graph of (I) passes through (0, 0), so 0 = Acos 0 +Bsin 0 = A. Thus, the equation is x = Bsin 2t. Since the
amplitude is 2, we see that x = 2 sin 2t is the equation of the graph. Similarly, the equation for (IV) is x = 3 sin 2t.
The graph of (II) also passes through (0, 0), so, similarly, the equation must be x = Bsin t. In this case, we see that
B = 1, so x = sin t.
Finally, the graph of (III) passes through (0, 1), and 1 is the maximum value. Thus, 1 = Acos 0+Bsin 0, so A = 1.
Since it reaches a local maximum at (0, 1), x
(0) = 0 = 4Asin 0 + 4Bcos 0, so B = 0. Thus, the solution is

x = cos 4t.
17. At t = 0, we nd that y = 1, which is clearly the lowest point on the path. Since y
= 3 sin 3t, we see that y
= 0
when t = 0. Thus, at t = 0 the object is at rest, although it will move up after t = 0.
21. (a) Since a mass of 3 kg stretches the spring by 2 cm, the spring constant k is given by
3g = 2k so k =
3g
2
.
See Figure 11.22.
T
c
x cm
equilibrium
3 kg
Figure 11.22
Suppose we measure the displacement x from the equilibrium; then, using
Mass Acceleration = Force
gives
3x
= kx =
3gx
2
x
+
g
2
x = 0
Since at time t = 0, the brick is 5 cm below the equilibrium and not moving, the initial conditions are x(0) = 5 and
x
(0) = 0.
(b) The solution to the differential equation is
x = Acos
__
g
2
t
_
+Bsin
__
g
2
t
_
.
Since x(0) = 5, we have
x = Acos(0) +Bsin(0) = 5 so A = 5.
In addition,
x
(t) = 5
_
g
2
sin
__
g
2
t
_
+B
_
g
2
cos
__
g
2
t
_
11.11 SOLUTIONS 261
so
x
(0) = 5
_
g
2
sin(0) +B
_
g
2
cos(0) = 0 so B = 0.
Thus,
x = 5 cos
_
g
2
t.
25. (a) 36
d
2
Q
dt
2
+
Q
9
= 0 so
d
2
Q
dt
2
=
Q
324
.
Thus,
Q = C1 cos
1
18
t +C2 sin
1
18
t .
Q(0) = 0 = C1 cos 0 +C2 sin 0 = C1,
so C1 = 0.
So, Q = C2 sin
1
18
t, and
Q
= I =
1
18
C2 cos
1
18
t.
Q
(0) = I(0) = 2 =
1
18
C2 cos
_
1
18
0
_
=
1
18
C2,
so C2 = 36.
Therefore, Q = 36 sin
1
18
t.
(b) As in part (a), Q = C1 cos
1
18
t +C2 sin
1
18
t.
According to the initial conditions:
Q(0) = 6 = C1 cos 0 +C2 sin 0 = C1,
so C1 = 6.
So Q = 6 cos
1
18
t +C2 sin
1
18
t.
Thus,
Q
= I =
1
3
sin
1
18
t +
1
18
C2 cos
1
18
t.
Q
(0) = I(0) = 0 =
1
3
sin
_
1
18
0
_
+
1
18
C2 cos
_
1
18
0
_
=
1
18
C2,
so C2 = 0.
Therefore, Q = 6 cos
1
18
t.
Exercises
1. The characteristic equation is r
2
+ 4r + 3 = 0, so r = 1 or 3.
Therefore y(t) = C1e
t
+C2e
3t
.
2
+ 7 = 0, so r =
7i.
Therefore s(t) = C1 cos
7t +C2 sin
7t.
2
+r + 1 = 0, so r =
1
2

3
2
i.
Therefore p(t) = C1e
t/2
cos
3
2
t +C2e
t/2
sin
3
2
t.
13. The characteristic equation is
r
2
+ 5r + 6 = 0
which has the solutions r = 2 and r = 3 so that
y(t) = Ae
3t
+Be
2t
The initial condition y(0) = 1 gives
A+B = 1
and y
(0) = 0 gives
3A2B = 0
so that A = 2 and B = 3 and
y(t) = 2e
3t
+ 3e
2t
2
+ 6r + 5 = 0, so r = 1 or 5.
Therefore y(t) = C1e
t
+C2e
5t
.
y
(t) = C1e
t
5C2e
5t
y
(0) = 0 = C1 5C2
y(0) = 1 = C1 +C2
Therefore C2 = 1/4, C1 = 5/4 and y(t) =
5
4
e
t
1
4
e
5t
.
21. The characteristic equation is
r
2
+ 5r + 6 = 0
which has the solutions r = 2 and r = 3 so that
y(t) = Ae
2t
+Be
3t
The initial condition y(0) = 1 gives
A+B = 1
and y(1) = 0 gives
Ae
2
+Be
3
= 0
so that A =
1
1 e
and B =
e
1 e
and
y(t) =
1
1 e
e
2t
+
e
1 e
e
3t
Problems
25. (a) x
+ 4x = 0 represents an undamped oscillator, and so goes with (IV).

(b) x
4x = 0 has characteristic equation r

2
4 = 0 and so r = 2. The solution is C1e
2t
+C2e
2t
. This represents
non-oscillating motion, so it goes with (II).
(c) x
0.2x
+ 1.01x = 0 has characteristic equation r

2
0.2 + 1.01 = 0 so b
2
4ac = 0.04 4.04 = 4, and
r = 0.1 i. So the solution is
C1e
(0.1+i)t
+C2e
(0.1i)t
= e
0.1t
(Asin t +Bcos t).
The negative coefcient in the x
term represents an amplifying force. This is reected in the solution by e

0.1t
, which
increases as t increases, so this goes with (I).
(d) x
+ 0.2x
+ 1.01x has characteristic equation r

2
+ 0.2r + 1.01 = 0 so b
2
4ac = 4. This represents a damped
oscillator. We have r = 0.1 i and so the solution is x = e
0.1t
(Asin t +Bcos t), which goes with (III).
29. Recall that s
+ bs
+ cs = 0 is overdamped if the discriminant b

2
4c > 0, critically damped if b
2
4c = 0, and
underdamped if b
2
4c < 0. Since b
2
4c = b
2
20, the solution is overdamped if b > 2
5 or b < 2
5, critically
damped if b = 2
5, and underdamped if 2
5 < b < 2
5.
33. The frictional force is F
drag
= c
ds
dt
. Thus spring (iv) has the smallest frictional force.
SOLUTIONS to Review Problems for Chapter Eleven 263
37. (a) If r1 =
b
b
2
4c
2
then r1 < 0 since both b and
b
2
4c are positive.
If r2 =
b+
b
2
4c
2
, then r2 < 0 because
b =
b
2
>
_
b
2
4c.
(b) The general solution to the differential equation is of the form
y = C1e
r
1
t
+C2e
r
2
t
and since r1 and r2 are both negative, y must go to 0 as t .
41. In this case, the differential equation describing the charge is Q
+ Q
+
1
4
Q = 0, so the characteristic equation is
r
2
+r +
1
4
= 0. This equation has one root, r =
1
2
, so the equation for charge is
Q(t) = (C1 +C2t)e
1
2
t
,
Q
(t) =
1
2
(C1 +C2t)e
1
2
t
+C2e
1
2
t
=
_
C2
C1
2

C2t
2
_
e
1
2
t
.
(a) We have
Q(0) = C1 = 0 ,
Q
(0) = C2
C1
2
= 2.
Thus, C1 = 0, C2 = 2, and
Q(t) = 2te
1
2
t
.
(b) We have
Q(0) = C1 = 2 ,
Q
(0) = C2
C1
2
= 0.
Thus, C1 = 2, C2 = 1, and
Q(t) = (2 +t)e
1
2
t
.
(c) The resistance was decreased by exactly the amount to switch the circuit from the overdamped case to the critically
damped case. Comparing the solutions of parts (a) and (b) in Problems 40, we nd that in the critically damped case
the net charge goes to 0 much faster as t .
Exercises
1. The slope elds in (I) and (II) appear periodic. (I) has zero slope at x = 0, so (I) matches y
= sin x, whereas (II) matches

y
= cos x. The slope in (V) tends to zero as x , so this must match y
= e
x
2
. Of the remaining slope elds,
only (III) shows negative slopes, matching y
= xe
x
. The slope in (IV) is zero at x = 0, so it matches y
= x
2
e
x
. This
leaves eld (VI) to match y
= e
x
.
5. This equation is separable, so we integrate, giving
_
dP =
_
t dt
so
P(t) =
t
2
2
+C.
9. This equation is separable, so we integrate, using the table of integrals or partial fractions, to get
_
1
R 3R
2
dR = 2
_
dt
_
1
R
dR +
_
3
1 3R
dR = 2
_
dt
so
ln |R| ln |1 3R| = 2t +C
ln
R
1 3R
= 2t +C
R
1 3R
= Ae
2t
R =
Ae
2t
1 + 3Ae
2t
.
13. 1 +y
2
dy
dx
= 0 gives
dy
dx
= y
2
+1, so
_
dy
1+y
2
=
_
dx and arctan y = x+C. Since y(0) = 0 we have C = 0, giving
y = tan x.
17.
dy
dx
=
0.2y(18+0.1x)
x(100+0.5y)
giving
_
(100+0.5y)
0.2y
dy =
_
18+0.1x
x
dx, so
_ _
500
y
+
5
2
_
dy =
_
_
18
x
+
1
10
_
dx.
Therefore, 500 ln |y| +
5
2
y = 18 ln |x| +
1
10
x +C. Since the curve passes through (10,10), 500 ln 10 +25 = 18 ln 10 +
1 +C, so C = 482 ln 10 + 24. Thus, the solution is
500 ln |y| +
5
2
y = 18 ln |x| +
1
10
x + 482 ln 10 + 24.
We cannot solve for y in terms of x, so we leave the answer in this form.
21.
df
dx
=
_
xf(x) gives
_
df
f(x)
=
_

xdx, so 2
_
f(x) =
2
3
x
3
2
+ C. Since f(1) = 1, we have 2 =
2
3
+ C so C =
4
3
.
Thus, 2
_
f(x) =
2
3
x
3
2
+
4
3
, so f(x) = (
1
3
x
3
2
+
2
3
)
2
.
(Note: this is only dened for x 0.)
25. (1+t
2
)y
dy
dt
= 1y implies that
_
y dy
1y
=
_
dt
1+t
2
implies that
_ _
1 +
1
1y
_
dy =
_
dt
1+t
2
. Therefore yln |1y| =
arctan t +C. y(1) = 0, so 0 = arctan 1 +C, and C =
4
, so y ln |1 y| = arctan t

4
. We cannot solve for
y in terms of t.
29. The characteristic equation of 9z
+z = 0 is
9r
2
+ 1 = 0
If we write this in the form r
2
+br +c = 0, we have that r
2
+ 1/9 = 0 and
b
2
4c = 0 (4)(1/9) = 4/9 < 0
This indicates underdamped motion and since the roots of the characteristic equation are r =
1
3
i, the general equation
is
y(t) = C1 cos
_
1
3
t
_
+C2 sin
_
1
3
t
_
Problems
33. (a) The slope eld for dy/dx = xy is in Figure 11.23.
x
y
Figure 11.23
x
y
Figure 11.24
(b) Some solution curves are shown in Figure 11.24.
(c) Separating variables gives
_
1
y
dy =
_
xdx
or
ln |y| =
1
2
x
2
+C.
Solving for y gives
y(x) = Ae
x
2
/2
where A = e
C
. In addition, y(x) = 0 is a solution. So y(x) = Ae
x
2
/2
is a solution for any A.
37. (a) x =
1
5
= 0.2.
At x = 0:
y0 = 1, y
= 4; so y = 4(0.2) = 0.8. Thus, y1 = 1 + 0.8 = 1.8.

At x = 0.2:
y1 = 1.8, y
= 3.2; so y = 3.2(0.2) = 0.64. Thus, y2 = 1.8 + 0.64 = 2.44.

At x = 0.4:
y2 = 2.44, y
= 2.56; so y = 2.56(0.2) = 0.512. Thus, y3 = 2.44 + 0.512 = 2.952.

At x = 0.6:
y3 = 2.952, y
= 2.048; so y = 2.048(0.2) = 0.4096. Thus, y4 = 3.3616.

At x = 0.8:
y4 = 3.3616, y
= 1.6384; so y = 1.6384(0.2) = 0.32768. Thus, y5 = 3.68928. So y(1) 3.689.

(b)
5
5
x
y
Since solution curves are concave down for 0 y 5, and y(0) = 1 < 5, the estimate from Eulers method
will be an overestimate.
(c) Solving by separation:
_
dy
5 y
=
_
dx, so ln |5 y| = x +C.
Then 5 y = Ae
x
where A = e
C
. Since y(0) = 1, we have 5 1 = Ae
0
, so A = 4.
Therefore, y = 5 4e
x
, and y(1) = 5 4e
1
3.528.
(Note: as predicted, the estimate in (a) is too large.)
(d) Doubling the value of n will probably halve the error and, therefore, give a value half way between 3.528 and 3.689,
which is approximately 3.61.
41. (a) If A is surface area, we know that for some constant K
dV
dt
= KA.
If r is the radius of the sphere, V = 4r
3
/3 and A = 4r
2
. Solving for r in terms of V gives r = (3V/4)
1/3
, so
dV
dt
= K(4r
2
) = K4
_
3V
4
_
2/3
so
dV
dt
= kV
2/3
where k is another constant, k = K(4)
1/3
3
2/3
.
(b) Separating variables gives
_
dV
V
2/3
=
_
k dt
3V
1/3
= kt +C.
Since V = V0 when t = 0, we have 3V
1/3
0
= C, so
3V
1/3
= kt + 3V
1/3
0
.
Solving for V gives
V =
_
k
3
t +V
1/3
0
_
3
.
This function is graphed in Figure 11.25.
3V
1/3
0
/k
V
0
t
V
Figure 11.25
(c) The snowball disappears when V = 0, that is when
k
3
t +V
1/3
0
= 0
giving
t =
3V
1/3
0
k
.
45. Recall that s
+ bs
+ c = 0 is overdamped if the discriminant b

2
4c > 0, critically damped if b
2
4c = 0, and
underdamped if b
2
4c < 0. Since b
2
4c = 16 4c, the circuit is overdamped if c < 4, critically damped if c = 4,
and underdamped if c > 4.
49. Let I be the number of infected people. Then, the number of healthy people in the population is M I. The rate of
infection is
Infection rate =
0.01
M
(M I)I.
and the rate of recovery is
Recovery rate = 0.009I.
Therefore,
dI
dt
=
0.01
M
(M I)I 0.009I
or
dI
dt
= 0.001I(1 10
I
M
).
This is a logistic differential equation, and so the solution will look like the following graph:
M
10
I
t
The limiting value for I is
1
10
M, so 1/10 of the population is infected in the long run.
53. (a) Using the integral equation with n + 1 replaced by n, we have
yn(a) = b +
_
a
a
(yn1(t)
2
+t
2
) dt = b + 0 = b.
(b) We have a = 1 and b = 0, so the integral equation tells us that
yn+1(s) =
_
s
1
(yn(t)
2
+t
2
) dt.
With n = 0, since y0(s) = 0, the CAS gives
y1(s) =
_
s
1
0 +t
2
dt =
1
3
+
s
3
3
.
Then
y2(s) =
_
s
1
(y1(t)
2
+t
2
) dt =
17
42
+
s
9
+
s
3
3

s
4
18
+
s
7
63
,
and
y3(s) =
_
s
1
(y2(t)
2
+t
2
) dt
=
157847
374220
+
289 s
1764

17 s
2
378
+
82 s
3
243

17 s
4
252
+
s
5
42

s
6
486
+
s
7
63

11 s
8
1764
+
5 s
9
6804
+
2 s
11
2079

s
12
6804
+
s
15
59535
.
(c) The solution y, and the approximations y1, y2, y3 are graphed in Figure 11.26. The approximations appear to be
accurate on the range 0.5 s 1.5.
1 2
0.5
1
2
3
y(s)
y
1
(s)
y
2
(s)
y
3
(s)
x
y
Figure 11.26
1. True. The general solution to y
= ky is y = Ce
kt
.
5. True. No matter what initial value you pick, the solution curve has the x-axis as an asymptote.
9. True. Rewrite the equation as dy/dx = xy +x = x(y + 1). Since the equation now has the form dy/dx = f(x)g(y), it
can be solved by separation of variables.
13. False. This is a logistic equation with equilibrium values P = 0 and P = 2. Solution curves do not cross the line P = 2
and do not go from (0, 1) to (1, 3).
17. True. Since f
(x) = g(x), we have f
(x) = g
(x). Since g(x) is increasing, g
(x) > 0 for all x, so f
(x) > 0 for all x.

Thus the graph of f is concave up for all x.
21. False. Let g(x) = 0 for all x and let f(x) = 17. Then f
(x) = g(x) and limxg(x) = 0, but limxf(x) = 17.

25. True. The slope of the graph of f is dy/dx = 2x y. Thus when x = a and y = b, the slope is 2a b.
29. False. Since f
(1) = 2(1) 5 = 3, the point (1, 5) could not be a critical point of f.

33. True. We will use the hint. Let w = g(x) f(x). Then:
dw
dx
= g
(x) f
(x) = (2x g(x)) (2x f(x)) = f(x) g(x) = w.

Thus dw/dx = w. This equation is the equation for exponential decay and has the general solution w = Ce
x
. Thus,
lim
x
(g(x) f(x)) = lim
x
Ce
x
= 0.
37. If we differentiate implicitly the equation for the family, we get 2x2ydy/dx = 0. When we solve, we get the differential
equation we want dy/dx = x/y.
12.1 SOLUTIONS 269
CHAPTER TWELVE
Exercises
1. The distance of a point P = (x, y, z) from the yz-plane is |x|, from the xz-plane is |y|, and from the xy-plane is |z|.
So A is closest to the yz-plane, since it has the smallest x-coordinate in absolute value. B lies on the xz-plane, since its
y-coordinate is 0. C is farthest from the xy-plane, since it has the largest z-coordinate in absolute value.
5. Your nal position is (1, 1, 1). This places you in front of the yz-plane, to the left of the xz-plane, and above the
xy-plane.
9. The graph is all points with y = 4 and z = 2, i.e., a line parallel to the x-axis and passing through the points
(0, 4, 2); (2, 4, 2); (4, 4, 2) etc. See Figure 12.1.
z
x
y
2
4
(0, 4, 2)
(2, 4, 2)
(4, 4, 2)
Figure 12.1
13. (a) 80-90
F
(b) 60-72
F
(c) 60-100
F
17. Table 12.1 gives the amount M spent on beef per household per week. Thus, the amount the household spent on beef in a
year is 52M. Since the households annual income is I thousand dollars, the proportion of income spent on beef is
P =
52M
1000I
= 0.052
M
I
.
Thus, we need to take each entry in Table 12.1, divide it by the income at the left, and multiply by 0.052. Table 12.2 shows
the results.
Table 12.1 Money spent on beef
($/household/week)
Income
($1,000)
Price of Beef ($)
3.00 3.50 4.00 4.50
20 7.95 9.07 10.04 10.94
40 12.42 14.18 15.76 17.46
60 15.33 17.50 19.88 21.78
80 16.05 18.52 20.76 22.82
100 17.37 20.20 22.40 24.89
Table 12.2 Proportion of annual income spent on beef
Income
($1,000)
Price of Beef ($)
3.00 3.50 4.00 4.50
20 0.021 0.024 0.026 0.028
40 0.016 0.018 0.020 0.023
60 0.013 0.015 0.017 0.019
80 0.010 0.012 0.013 0.015
100 0.009 0.011 0.012 0.013
270 Chapter Twelve /SOLUTIONS
Problems
21.
Table 12.3 Temperature adjusted for wind-chill at 5 mph
Temperature (
F) 35 30 25 20 15 10 5 0
Adjusted temperature (
F) 31 25 19 13 7 1 5 11
Table 12.4 Temperature adjusted for wind-chill at 20 mph
Temperature (
F) 35 30 25 20 15 10 5 0
Adjusted temperature (
F) 24 17 11 4 2 9 15 22
25. The gravitational force on a 100 kg object which is 7, 000, 000 meters from the center of the earth (or about 600 km above
the earths surface) is about 820 newtons.
29. The distance of any point with coordinates (x, y, z) from the x-axis is
y
2
+ z
2
. The distance of the point from the
xy-plane is |x|. Since the condition states that these distances are equal, the equation for the condition is
y
2
+ z
2
= |x| i.e. y
2
+ z
2
= x
2
.
This is the equation of a cone whose tip is at the origin and which opens along the x-axis with a slope of 1 as shown in
Figure 12.2.
x
y
z
Figure 12.2
33. Using the distance formula, we nd that
Distance from P1 to P =
206
152
170
113
So P4 = (4, 2, 7) is closest to P = (6, 0, 4).
Exercises
1. (a) is (IV), since z = 2 + x
2
+ y
2
is a paraboloid opening upward with a positive z-intercept.
(b) is (II), since z = 2 x
2
y
2
is a paraboloid opening downward.
(c) is (I), since z = 2(x
2
+ y
2
) is a paraboloid opening upward and going through the origin.
(d) is (V), since z = 2 + 2x y is a slanted plane.
(e) is (III), since z = 2 is a horizontal plane.
12.2 SOLUTIONS 271
5. The graph is a bowl opening up, with vertex at the point (0, 0, 4). See Figure 12.3.
4
x
y
z
Figure 12.3
x y
z
Figure 12.4
9. In the xy-plane, the graph is a circle of radius 2. Since there are no restrictions on z, we extend this circle along the z-axis.
The graph is a circular cylinder extended in the z-direction. See Figure 12.4.
Problems
13. The one-variable function f(a, t) represents the effect of an injection of a mg at time t. Figure 12.5 shows the graphs of
the four functions f(1, t) = te
4t
, f(2, t) = te
3t
, f(3, t) = te
2t
, and f(4, t) = te
t
corresponding to injections of
1, 2, 3, and 4 mg of the drug. The general shape of the graph is the same in every case: The concentration in the blood is
zero at the time of injection t = 0, then increases to a maximum value, and then decreases toward zero again. We see that
if a larger dose of the drug is administered, the peak of the graph is later and higher. This makes sense, since a larger dose
will take longer to diffuse fully into the bloodstream and will produce a higher concentration when it does.
1 2 3 4 5
0.1
0.2
0.3
'
a = 1
'
a = 2
a = 3
a = 4
t (hours)
C (mg per liter)
Figure 12.5: Concentration C = f(a, t) of the drug resulting from an a mg injection
17. (a) This is a bowl; z increases as the distance from the origin increases, from a minimum of 0 at x = y = 0.
(b) Neither. This is an upside-down bowl. This function will decrease from 1, at x = y = 0, to arbitrarily large negative
values as x and y increase due to the negative squared terms of x and y. It will look like the bowl in part (a) except
ipped over and raised up slightly.
(c) This is a plate. Solving the equation for z gives z = 1 xy which describes a plane whose x and y slopes are 1.
It is perfectly at, but not horizontal.
(d) Within its domain, this function is a bowl. It is undened at points at which x
2
+ y
2
> 5, but within those limits it
describes the bottom half of a sphere of radius
5 centered at the origin.

(e) This function is a plate. It is perfectly at and horizontal.
21. One possible equation: z = (x y)
2
. See Figure 12.6.
z
y
x
Figure 12.6
25. (a) If we have iron stomachs and can consume cola and pizza endlessly without ill effects, then we expect our happiness
to increase without bound as we get more cola and pizza. Graph (IV) shows this since it increases along both the
pizza and cola axes throughout.
(b) If we get sick upon eating too many pizzas or drinking too much cola, then we expect our happiness to decrease once
either or both of those quantities grows past some optimum value. This is depicted in graph (I) which increases along
both axes until a peak is reached, and then decreases along both axes.
(c) If we do get sick after too much cola, but are always able to eat more pizza, then we expect our happiness to decrease
after we drink some optimum amount of cola, but continue to increase as we get more pizza. This is shown by graph
(III) which increases continuously along the pizza axis but, after reaching a maximum, begins to decrease along the
cola axis.
29. The paraboloid is z = x
2
+ y
2
+ 5, so it is represented by
z = f(x, y) = x
2
+ y
2
+ 5
and
g(x, y, z) = x
2
+ y
2
+ 5 z = 0.
Other answers are possible.
Exercises
1. Well set z = 4 at the peak. See Figure 12.7.
x
y
4
3
2
1
Figure 12.7
12.3 SOLUTIONS 273
5. The contour where f(x, y) = x + y = c, or y = x + c, is the graph of the straight line with slope 1 as shown in
Figure 12.8. Note that we have plotted the contours for c = 3, 2, 1, 0, 1, 2, 3. The contours are evenly spaced.
x
y
2 1 1 2
2
1
1
2
c
=
0
c
=
1
c
=
2 c
=
3
c
=
1
c
=
2
c
=
3
Figure 12.8
9. The contour where f(x, y) = xy = c, is the graph of the hyperbola y = c/x if c = 0 and the coordinate axes if c = 0,
as shown in Figure 12.9. Note that we have plotted contours for c = 5, 4, 3, 2, 1, 0, 1, 2, 3, 4, 5. The contours
become more closely packed as we move further from the origin.
x
y
c
=
1 c
=
2
c
=
3
c
=
4
c
=
1
c
=
2
c
=
3
c
=
4
c
=
1
c
=
2
c
=
3
c
=
4
c
=
1
c
=
2
c
=
3
c
=
4
c
=
0
Figure 12.9
13. The contour where f(x, y) = cos(
x
2
+ y
2
) = c, where 1 c 1, is a set of circles centered at (0, 0), with radius
cos
1
c +2k with k = 0, 1, 2, .. and cos
1
c +2k, with k = 1, 2, 3, ... as shown in Figure 12.10. Note that we have
plotted contours for c = 0, 0.2, 0.4, 0.6, 0.8, 1.
5
2
5
2
5
2
5
2
x
y
c = 1
'
'
c
=
0
Figure 12.10
17. The values in Table 12.5 are not constant along rows or columns and therefore cannot be the lines shown in (I) or (IV).
Also observe that as you move away from the origin, whose contour value is 0, the z-values on the contours increase.
Thus, this table corresponds to diagram (II).
The values in Table 12.6 are also not constant along rows or columns. Since the contour values are decreasing as you
move away from the origin, this table corresponds to diagram (III).
Table 12.7 shows that for each xed value of x, we have constant contour value, suggesting a straight vertical line at
each x-value, as in diagram (IV).
Table 12.8 also shows lines, however these are horizontal since for each xed value of y we have constant contour
values. Thus, this table matches diagram (I).
Problems
21. (a) The contour lines are much closer together on path A, so path A is steeper.
(b) If you are on path A and turn around to look at the countryside, you nd hills to your left and right, obscuring the
view. But the ground falls away on either side of path B, so you are likely to get a much better view of the countryside
from path B.
(c) There is more likely to be a stream alongside path A, because water follows the direction of steepest descent.
25. Figure 12.11 shows an east-west cross-section along the line N = 50 kilometers.
Figure 12.12 shows an east-west cross-section along the line N = 100 kilometers.
60 120 180
0.5
1
1.5
2
2.5
East
Density of the
fox population P
Figure 12.11
30 60 90 120 150 180
0.5
1
1.5
East
Density of the
fox population P
Figure 12.12
Figure 12.13 shows a north-south cross-section along the line E = 60 kilometers.
Figure 12.14 shows a north-south cross-section along the line E = 120 kilometers.
35 70
0.5
1
North
Density of the
fox population P
Figure 12.13
50 100 150
0.5
1
1.5
2
2.5
North
Density of the
fox population P
Figure 12.14
12.3 SOLUTIONS 275
29. (a) Multiply the values on each contour of the original contour diagram by 3. See Figure 12.15.
6
3
0
3
6
x
y
Figure 12.15: 3f(x, y)
12
11
10
9
8
x
y
Figure 12.16: f(x, y) 10
(b) Subtract 10 from the values on each contour. See Figure 12.16.
(c) Shift the diagram 2 units to the right and 2 units up. See Figure 12.17.
4
3
2
1
0
x
y
Figure 12.17: f(x 2, y 2)
2
1
0
1
2
x
y
Figure 12.18: f(x, y)
(d) Reect the diagram about the y-axis. See Figure 12.18.
33. Since f(x, y) = x
2
y
2
= (x y)(x + y) = 0 gives x y = 0 or x + y = 0, the contours f(x, y) = 0 are the
lines y = x or y = x. In the regions between them, f(x, y) > 0 or f(x, y) < 0 as shown in Figure 12.19. The surface
z = f(x, y) is above the xy-plane where f > 0 (that is on the shaded regions containing the x-axis) and is below the
xy-plane where f < 0. This means that a person could sit on the surface facing along the positive or negative x-axis, and
with his/her legs hanging down the sides below the y-axis. Thus, the graph of the function is saddle-shaped at the origin.
f > 0
f < 0
f > 0
f < 0
f = 0 f = 0
y = x
y = x
x
y
Figure 12.19
Exercises
1.
Table 12.5
x\y 0.0 1.0
0.0 1.0 1.0
2.0 3.0 5.0
5. A table of values is linear if the rows are all linear and have the same slope and the columns are all linear and have the
same slope. The table might represent a linear function since the slope in each row is 5 and the slope in each column is 2.
9. Since
0 = c + m 0 + n 0 c = 0
1 = c + m 0 + n 2 c + 2n = 1
4 = c + m (3) + n 0 c 3m = 4
we get:
c = 0, m =
4
3
, n =
1
2
.
Thus, z =
4
3
x
1
2
y.
13. (a) Since z is a linear function of x and y with slope 2 in the x-direction, and slope 3 in the y-direction, we have:
z = 2x + 3y + c
We can write an equation for changes in z in terms of changes in x and y:
z = (2(x + x) + 3(y + y) + c) (2x + 3y + c)
= 2x + 3y
Since x = 0.5 and y = 0.2, we have
z = 2(0.5) + 3(0.2) = 0.4
So a 0.5 change in x and a 0.2 change in y produces a 0.4 change in z.
(b) As we know that z = 2 when x = 5 and y = 7, the value of z when x = 4.9 and y = 7.2 will be
z = 2 + z = 2 + 2x + 3y
where z is the change in z when x changes from 4.9 to 5 and y changes from 7.2 to 7. We have x = 4.9 5 =
0.1 and y = 7.2 7 = 0.2. Therefore, when x = 4.9 and y = 7.2, we have
z = 2 + 2 (0.1) + 5 0.2 = 2.4
Problems
17. (a) Expenditure, E, is given by the equation:
E = (price of raw material 1)m1 + (price of raw material 2)m2 + C
where C denotes all the other expenses (assumed to be constant). Since the prices of the raw materials are constant,
but m1 and m2 are variables, we have a linear function.
(b) Revenue, R, is given by the equation:
R = (p1)q1 + (p2)q2.
Since p1 and p2 are constant, while q1 and q2 are variables, we again have a linear function.
(c) Revenue is again given by the equation,
R = (p1)q1 + (p2)q2.
Since p2 and q2 are now constant, the term (p2)q2 is also constant. However, since p1 and q1 are variables, the (p1)q1
term means that the function is not linear.
12.5 SOLUTIONS 277
21. The function, g, has a slope of 3 in the x direction and a slope of 1 in the y direction, so g(x, y) = c + 3x + y. Since
g(0, 0) = 0, the formula is g(x, y) = 3x + y.
25. See Figure 12.20.
x
y
z
1
2
2
Figure 12.20
29. (a) The contours of f have equation
k = c + mx + ny, where k is a constant.
Solving for y gives:
y =
m
n
x +
k c
n
Since c, m, n and k are constants, this is the equation of a line. The coefcient of x is the slope and is equal to m/n.
(b) Substituting x + n for x and y m for y into f(x, y) gives
f(x + n, y m) = c + m(x + n) + n(y m)
Multiplying out and simplifying gives
f(x + n, y m) = c + mx + mn + ny nm
f(x + n, y m) = c + mx + ny = f(x, y)
(c) Part (b) tells us that if we move n units in the x direction and m units in the y direction, the value of the function
f(x, y) remains constant. Since contours are lines where the function has a constant value, this implies that we remain
on the same contour. This agrees with part (a) which tells us that the slope of any contour line will be m/n. Since
the slope is y/x, it follows that changing y by m and x by n will keep us on the same contour.
Exercises
1. (a) Observe that setting f(x, y, z) = c gives a cylinder about the x-axis, with radius
c. These surfaces are in graph (I).

(b) By the same reasoning the level curves for h(x, y, z) are cylinders about the y-axis, so they are represented in graph
(II).
5. If we solve for z, we get z = (1 x
2
y)
2
, so the level surface is the graph of f(x, y) = (1 x
2
y)
2
.
9. A hyperboloid of two sheets.
13. Yes,
z = f(x, y) =
2
5
x +
3
5
y 2.
Problems
17. The top half of the sphere is represented by
z = f(x, y) =
10 x
2
y
2
and
g(x, y, z) = x
2
+ y
2
+ z
2
= 10, z 0.
21. The equation of any plane parallel to the plane z = 2x+3y5 has x-slope 2 and y-slope 3, so has equation z = 2x+3yc
for any constant c, or 2x + 3y z = c. Thus we could take g(x, y, z) = 2x + 3y z. Other answers are possible.
25. In the xz-plane, the equation x
2
/4 + z
2
= 1 is an ellipse, with widest points at x = 2 on the x-axis and crossing the
z-axis at z = 1. Since the equation has no y term, the level surface is a cylinder of elliptical cross-section, centered
along the y-axis.
29. Lets consider the function y = 2 + sin z drawn in the yz-plane in Figure 12.21.
x
y
z
y = 2 + sin z
2
Figure 12.21
Now rotate this graph around the z-axis. Then, a point (x, y, z) is on the surface if and only if x
2
+y
2
= (2+sin z)
2
.
Thus, the surface generated is a surface of rotation with the prole shown in Figure 12.21.
Similarly, the surface with equation x
2
+y
2
= (f(z))
2
is the surface obtained rotating the graph of y = f(z) around
the z-axis.
33. For values of f < 4, the level surfaces are spheres, with larger f giving smaller radii. See Figure 12.22.
y
z
x
f = 2
E
f = 1
f = 0
Figure 12.22
12.6 SOLUTIONS 279
Exercises
1. No, 1/(x
2
+ y
2
) is not dened at the origin, so is not continuous at all points in the square 1 x 1, 1 y 1.
5. The function tan() is undened when = /2 1.57. Since there are points in the square 2 x 2, 2 y 2
with x y = /2 (e.g. x = 1, y = /2) the function tan(xy) is not dened inside the square, hence not continuous.
9. Since f does not depend on y we have:
lim
(x,y)(0,0)
f(x, y) = lim
x0
x
x
2
+ 1
=
0
0 + 1
= 0.
Problems
13. We want to show that f does not have a limit as (x, y) approaches (0, 0). Let us suppose that (x, y) tends to (0, 0) along
the line y = mx. Then
f(x, y) = f(x, mx) =
x
2
m
2
x
2
x
2
+ m
2
x
2
=
1 m
2
1 + m
2
.
Therefore
lim
x0
f(x, mx) =
1 m
2
1 + m
2
and so for m = 1 we get
lim
(x,y)(0,0)
y=x
f(x, y) =
1 1
1 + 1
=
0
2
= 0
and for m = 0
lim
(x,y)(0,0)
y=0
f(x, y) =
1 0
1 + 0
= 1.
Thus no matter how close they are to the origin, there will be points (x, y) such that f(x, y) is close to 0 and points (x, y)
where f(x, y) is close to 1. So the limit:
lim
(x,y)(0,0)
f(x, y) does not exist.
17. We will study the continuity of f at (a, 0). Now f(a, 0) = 1 a. In addition:
lim
(x,y)(a,0)
y>0
f(x, y) = lim
xa
(1 x) = 1 a
lim
(x,y)(a,0)
y<0
f(x, y) = lim
xa
2 = 2.
If a = 3, then
lim
(x,y)(3,0)
y>0
f(x, y) = 1 3 = 2 = lim
(x,y)(3,0)
y<0
f(x, y)
and so lim
(x,y)(3,0)
f(x, y) = 2 = f(3, 0). Therefore f is continuous at (3, 0).
On the other hand, if a = 3, then
lim
(x,y)(a,0)
y>0
f(x, y) = 1 a = 2 = lim
(x,y)(a,0)
y<0
f(x, y)
so lim
(x,y)(a,0)
f(x, y) does not exist. Thus f is not continuous at (a, 0) if a = 3.
Thus, f is not continuous along the line y = 0. (In fact the only point on this line where f is continuous is the point
(3, 0).)
21. The function f(x, y) = x
2
+ y
2
+ 1 gets closer and closer to 1 as (x, y) gets closer to the origin. To make f continuous
at the origin, we need to have f(0, 0) = 1. Thus c = 1 will make the function continuous at the origin.
Exercises
1. An example is the line z = x in the xz-plane. See Figure 12.23.
x
y
z
Figure 12.23
5. Given (x, y) we can solve uniquely for z, namely z = 2 +
x
5
+
y
5

3x
2
5
+ y
2
. Thus, z is a function of x and y:
z = f(x, y) = 2 +
x
5
+
y
5

3x
2
5
+ y
2
.
9. Contours are lines of the form 3x 5y + 1 = c as shown in Figure 12.24. Note that for the regions of x and y given, the
c values range from 12 < c < 12 and are evenly spaced.
2 1 1 2
2
1
1
2
x
y
1
2
4
0
4
8
1
2
Figure 12.24
13. These conditions describe a line parallel to the z-axis which passes through the xy-plane at (2, 1, 0).
17. A contour diagram is linear if the contours are parallel straight lines, equally spaced for equally spaced values of z. This
contour diagram does not represent a linear function.
21. The level surfaces appear to be circular cylinders centered on the z-axis. Since they dont change with z, there is no z in
the formula, and we can use the formula for a circle in the xy-plane, x
2
+ y
2
= r
2
. Thus the level surfaces are of the
form f(x, y, z) = x
2
+ y
2
= c for c > 0.
SOLUTIONS to Review Problems for Chapter Twelve 281
25. (a) The value of z decreases as x increases. See Figure 12.25.
(b) The value of z increases as y increases. See Figure 12.26.
x
z
Figure 12.25
y
z
Figure 12.26
Problems
29. Points along the positive x-axis are of the form (x, 0); at these points the function looks like 2x/2x = 1 everywhere
(except at the origin, where it is undened). On the other hand, along the y-axis, the function looks like y
2
/y
2
= 1.
Since approaching the origin along two different paths yields numbers that are not the same, the limit does not exist.
33. When h is xed, say h = 1, then
V = f(r, 1) = r
2
1 = r
2
Similarly,
f(r,
2
3
) =
4
9
r
2
and f(r,
1
3
) =

9
r
2
When r is xed, say r = 1, then
f(1, h) = (1)
2
h = h
Similarly,
f(2, h) = 4 and f(3, h) = 9h.
1 2 3 4 5
0
5
10
15
20
25
r
Volume
h =
1
3
f(r,
1
3
)
h =
2
3
f(r,
2
3
)
f(r, 1)
h = 1
Figure 12.27
1 2 3 4 5 0
10
20
30
40
50
r = 1
f(1, h)
r = 2
f(2, h)
r = 3
f(3, h)
h
Volume
Figure 12.28
37. The function f(0, t) = cos t sin 0 = 0 gives the displacement of the left end of the string as time varies. Since that point
remains stationary, the displacement is zero. The function f(1, t) = cos t sin 1 = 0.84 cos t gives the displacement of the
point at x = 1 as time varies. Since cos t oscillates back and forth between 1 and 1, this point moves back and forth with
maximum displacement of 0.84 in either direction. Notice the maximum displacements are greatest at x = /2 where
sin x = 1.
41. (a)
f(x, f(x, y)) = 3 + x + 2(3 + x + 2y) = (3 + 2 3) + (1 + 2)x + 2
2
y = 9 + 3x + 4y
f(x, f(x, f(x, y))) = 3 + x + 2(3 + x + 2(3 + x + 2y))
= (3 + 2 3 + 2
2
3) + (1 + 2 + 2
2
)x + 2
3
y = 21 + 7x + 8y
(b) From part (a) we guess that the general pattern for k nested fs is
(3 + 2 3 + 2
2
3 + + 2
k1
3) + (1 + 2 + 2
2
+ + 2
k1
)x + 2
k
y
Thus
f(x, f(x, f(x, f(x, f(x, f(x, y)))))) =
(3 + 2 3 + 2
2
3 + + 2
5
3) + (1 + 2 + 2
2
+ + 2
5
)x + 2
6
y = 189 + 63x + 64y.
1. Could not be true. If the origin is on the level curve z = 1, then z = f(0, 0) = 1 = 1. So (0, 0) cannot be on both
z = 1 and z = 1.
5. True. For every point (x, y), compute the value z = e
(x
2
+y
2
)
at that point. The level curve obtained by getting z equal
to that value goes through the point (x, y).
9. False. If, for example, d = 2 meters and H = 57
C, there could be many times t at which the water temperature is 57

C
at 2 meters depth.
13. True. If there were such an intersection point, that point would have two different temperatures simultaneously.
17. False. The point (0, 0, 0) does not satisfy the equation.
21. True. The x-axis is where y = z = 0.
25. False. If x = 10, substituting gives 10
2
+ y
2
+ z
2
= 10, so y
2
+ z
2
= 90. Since y
2
+ z
2
cannot be negative, a point
with x = 10 cannot satisfy the equation.
29. True. The cross-sections with y = c are of the form z = 1 c
2
, which are horizontal lines.
33. False. Wherever f(x, y) = 0 the graphs of f(x, y) and f(x, y) will intersect.
37. True. If f = c then the contours are of the form c = y
2
+ (x 2)
2
, which are circles centered at (2, 0) if c > 0. But if
c = 0 the contour is the single point (2, 0).
41. False. As a counterexample, consider any function with one variable missing, e.g. f(x, y) = x
2
. The graph of this is not
a plane (it is a parabolic cylinder) but has contours which are lines of the form x = c.
45. True. The graph of g is the same as the graph of f translated down by 5 units, so the horizontal slice of f at height 5 is
the same as the horizontal slice of g at height 0.
49. True. f(0, 0) = 0, f(0, 1) = 4 give a y slope of 4, but f(0, 0) = 0, f(0, 3) = 5 give a y slope of 5/3. Since linearity
means the y slope must be the same between any two points, this function cannot be linear.
53. True. Functions can have only one value for a given input, so their graphs can intersect a vertical line at most once. A
vertical plane would not satisfy this property, so cannot be the graph of a function.
57. True. Both are the set of all points (x, y, z) in 3-space satisfying z = x
2
+ y
2
.
61. True. The level surfaces are of the form x + 2y +z = k, or z = k x 2y. These are the graphs of the linear functions
f(x, y) = k x 2y, each of which has x-slope of 1 and y-slope equal to 2. Thus they form parallel planes.
65. False. For example, the function g(x, y, z) = sin(x + y + z) has level surfaces of the form x + y + z = k, where
k = arcsin(c) + n, for n = 0, 1, 2, . . .. These surfaces are planes (for 1 c 1).
13.1 SOLUTIONS 283
CHAPTER THIRTEEN
Exercises
1. The vectors are a =
i + 3
j ,
b = 3
i + 2
j , v = 2
i 2
j , and w =
i + 2
j .
5. a =
b =c = 3
k ,

d = 2
i + 3
k , e =
j ,

f = 2
i
9. 4
i + 8
j 0.5
i + 0.5
k = 4.5
i + 8
j + 0.5
k
13. 0.6
i + 0.2
k + 0.3
i + 0.3
k = 0.9
i + 0.2
j 0.7
k
17. v =
_
1
2
+ (1)
2
+ 3
2
=
11.
21.
5a + 2
b = 5(2
j +
k ) + 2(3
i + 5
j + 4
k )
= (5(2)
j + 5(1)
k ) + (2(3)
i + 2(5)
j + 2(4)
k )
= (10
j + 5
k ) + (6
i + 10
j + 8
k ) = (0 6)
i + (10 + 10)
j + (5 + 8)
k
= 6
i + 20
j + 13
k .
25.
y x = (4
i 7
j ) (2
i + 9
j ) = (4 (2))
i + (7 9)
j = 6
i 16
j
=
_
6
2
+ (16)
2
=
36 + 256 =
292 = 2
73.
Problems
29. If two vectors are parallel, they are scalar multiples of one another. Thus
a
2
5a
=
6
3
.
Solving for a gives
a
2
= 2 5a so a = 0, 10.
33. (a) The components are v1 = 2 cos /4 =
2, v2 = 2 sin /4 =
2. See Figure 13.1. Thus v =
i +
j .
(b) Since the vector lies in the xz-plane, its y-component is 0. Its x-component is 1 cos(
6
)
i =
3
2
i and its z-component

is 1 sin(
6
)
j =
1
2
j . See Figure 13.2. So the vector is
3
2
i +
1
2
k .
x
y
45
=

4
Figure 13.1
x
y
z
6
= 30
Figure 13.2
284 Chapter Thirteen /SOLUTIONS
37.
v
y
z
w
x
v w
u
Figure 13.3
u
v
w
E
w = v + (u )
Figure 13.4
Break the hexagon up into 6 equilateral triangles, as shown in Figure 13.3.
Then u v + w =
0 , so w = v u
Similarly, x = u , y = v , z = w = u v .
41.
30 km/hr
truck
T
r
40 km/hr
Police car
'
x East
North
y
C Police car
P
Truck
O
Figure 13.5
Since both vehicles reach the crossroad in exactly one hour, at the present the truck is at O in Figure 13.5; the police
car is at P and the crossroads is at C. If r is the vector representing the line of sight of the truck with respect to the police
car.
r = 40
i 30
j
Exercises
1. Scalar
5. Writing

P = (P1, P2, , P50) where Pi is the population of the i-th state, shows that

P can be thought of as a vector
with 50 components.
9. We need to calculate the length of each vector.
21
i + 35
j =
_
21
2
+ 35
2
=
1666 40.8,
40
i =
40
2
= 40.
So the rst car is faster.
13.2 SOLUTIONS 285
Problems
13. The velocity vector of the plane with respect to the air has the form
v = a
i + 80
k where v = 480.
(See Figure 13.6.) Therefore
a
2
+ 80
2
= 480 so a =
480
2
80
2
473.3 km/hr. We conclude that v 473.3
i +
80
k .
The wind vector is
w = 100(cos 45
i + 100(sin 45
j
70.7
i + 70.7
j
The velocity vector of the plane with respect to the ground is then
v + w = (473.3
i + 80
k ) + (70.7
i + 70.7
j )
= 544
i + 70.7
j + 80
k
From Figure 13.7, we see that the velocity relative to the ground is
544
i + 70.7
j .
The ground speed is therefore
544
2
+ 70.7
2
548.6 km/hr.
a
i
50
k
v
Figure 13.6: Side view
a
i
Velocity relative
to ground
w
Figure 13.7: Top view
17. Let

R be the resultant force, and let

F 1 and

F 2 be the forces exerted by the larger and smaller tugs. See Figure 13.8.
Then
F1 =
5
4
F2 . The y components of the vectors

F 1 and

F 2 must cancel each other in order to ensure that the
ship travels due east, hence
F1 sin 30
F2 sin ,
so
5
4
F2 sin 30
F2 sin ,
giving sin =
5
8
, and hence = sin
1 5
8
= 38.7
.
30
F
1
F
2
T
E
T
x
y
N
Figure 13.8
21. We want the total force on the object to be zero. We must choose the third force

F 3 so that

F 1 +

F 2 +

F 3 = 0. Since
F 1 +

F 2 = 11
i 4
j , we need

F 3 = 11
i + 4
j .
25. (a) Let x-axis be the East direction and y-axis be the North direction. From Figure 13.9,
= sin
1
(4/5) = 53.1
.
That is, he should steer at 53.1
east of south.
*
*
*
E
T T
x
y
N
5 km/hr
Steering direction
4 km/hr
Figure 13.9
(b)
E
T T
x
y
N
4 km/hr
4
10 km/hr
R
5 km/hr
R
Figure 13.10
Let

R be the resultant of the wind and river velocities, that is
R = 4
i + (10 cos(
4
)
i + 10 cos(
4
)
j )
= (4 + 5
2)
i + 5
j .
From Figure 13.10, we see that to get the the x-component of his rowing velocity and the x-component of

R to
cancel each other, we must have
5 sin = 4 + 5
2
= sin
1
_
4 + 5
2
5
_
= 37.9
.
However for this value of , the y-component of the velocity is
5
2 5 cos(37.9
) = 3.1.
Since the y-component is positive, the man will not move across the river in a southward direction.
13.3 SOLUTIONS 287
29.
v
v
v
v
Figure 13.11
The vectors v , v and v are all parallel. Figure 13.11 shows them with , > 0, so all the vectors are in the same
direction. Notice that v is a vector times as long as v and v is times as long as v . Therefore v +v is a vector
( +) times as long as v , and in the same direction. Thus,
v +v = ( +)v .
33. According to the denition of scalar multiplication, 1 v has the same direction and magnitude as v , so it is the same as
v .
Exercises
1. a y = (2
j +
k ) (4
i 7
j ) = 14.
5. c a +a y = (
i + 6
j ) (2
j +
k ) + (2
j +
k ) (4
i 7
j ) = 12 14 = 2.
9. Since c c is a scalar and (c c )a is a vector, the answer to this equation is another scalar. We could calculate c c , then
(c c )a , and then take the dot product ((c c )a ) a . Alternatively, we can use the fact that
((c c )a ) a = (c c )(a a ).
Since
c c = (
i + 6
j ) (
i + 6
j ) = 1
2
+ 6
2
= 37
a a = (2
j +
k ) (2
j +
k ) = 2
2
+ 1
2
= 5,
we have,
(c c )(a a ) = 37(5) = 185
13. The equation can be rewritten as
z 5x + 10 = 15 3y
5x + 3y +z = 5
so n = 5
i + 3
j +
k .
17. (a) Writing the plane in the form 2x + 3y z = 0 shows that a normal vector is
n = 2
i + 3
k .
Any multiple of this vector is also a correct answer.
(b) Any vector perpendicular to n is parallel to the plane, so one possible answer is
v = 3
i 2
j .
Many other answers are possible.
21. (a) We rst nd the unit vector in direction v . Since ||v || =
3
2
+ 4
2
= 5, the unit vector in direction of v is
u = 0.6
i + 0.8
j . Then
F
parallel
= (
F u )u
= (0.4 0.6 + 0.3 0.8)u
=

0 .
Notice that the component of

F in direction v is equal to
0 . This makes sense (and could have been predicted) since
F is perpendicular to v .
(b) We have
F perp =

F

F
parallel
=

F .
(c) Since work is the dot product of the force and displacement vectors, we have
W =

F v = 0.4 3 + 0.3 4 = 0.
Notice that since the force is perpendicular to the displacement, the work done is zero.
25. (a) We rst nd the unit vector in direction v . Since ||v || =
_
5
2
+ (1)
2
=
26, the unit vector in direction of v is

u = v /
26. Then
F
parallel
= (
F u )u
= (20/
26)u
=
100
26
i +
20
26
j
= 3.846
i 0.769
j .
(b) We have
F perp =

F

F
parallel
= (20
j ) (3.846
i 0.769
j ) = 3.846
i 19.231
j .
(c) Since work is the dot product of the force and displacement vectors, we have
W =

F v = 20.
Problems
29. (a) Perpendicular vectors have a dot product of 0. Since a c = 1(2) 3(1) 1 1 = 0, and
b

d = 1(1) +
1(1) + 2 1 = 0, the pairs we want are a , c and
b ,

d .
(b) Parallel vectors are multiples of one another, so there are no parallel vectors in this set.
(c) Since v w = ||v |||| w || cos , the dot product of the vectors we want is positive. We have
a
b = 1 1 3 1 1 2 = 4
a

d = 1(1) 3(1) 1 1 = 1
b c = 1(2) + 3(1) + 2 1 = 1
c

d = 2(1) 1(1) + 1 1 = 4,
and we already knowa c =
b

d = 0. Thus, the pairs of vectors with an angle of less than /2 between them are
a ,

d and c ,

d .
(d) Vectors with an angle of more than /2 between them have a negative dot product, so pairs are a ,
b and
b , c .
33. If the vectors are perpendicular, we need
v w = (2a
i a
j + 16
k ) (5
i +a
k ) = 10a a
2
16 = 0.
Solving 10a a
2
16 = (a 2)(a 8) = 0 gives a = 2, 8.
37. The plane is
3x y + 4z = 3 1 1 5 + 4 2
3x y + 4z = 6.
13.3 SOLUTIONS 289
41. Two planes are parallel if their normal vectors are parallel. Since the plane 3x + y + z = 4 has normal vector n =
3
i +
j +
k , the plane we are looking for has the same normal vector and passes through the point (2, 3, 2). Thus, it
has the equation
3x +y +z = 3 (2) + 3 + 2 = 1.
45. (a) The points A, B and C are shown in Figure 13.12.
x
y
z
B
C
A
Figure 13.12
First, we calculate the vectors which form the sides of this triangle:
AB = (4
i + 2
j +
k ) (2
i + 2
j + 2
k ) = 2
BC = (2
i + 3
j +
k ) (4
i + 2
j +
k ) = 2
i +
AC = (2
i + 3
j +
k ) (2
i + 2
j + 2
k ) =
k
Now we calculate the lengths of each of the sides of the triangles:
AB =
_
2
2
+ (1)
2
=
BC =
_
(2)
2
+ 1
2
=
AC =
_
1
2
+ (1)
2
=
2
Thus the length of the shortest side of S is
2.
(b) cos
BAC =
AB
AC
AB
AC
=
20+01+(1)(1)
2
0.32
49. (a) The speed of the current is c =
5 = 2.24 m/sec.
(b) The speed of the current in the direction of the canoes motion is the component of c in the direction of v . This is
given by:
Speed of current in direction of canoes motion =
c v
v
=
(1)(5) + (2)(3)
5
2
+ 3
2
=
11
34
= 1.89 m/sec.
Notice that the speed of the current in the direction of the canoe is less than the speed of the current in the direction
in which the current is moving.
53. Let u = 3
i + 4
j and v = 5
i 12
j . We seek a vector w = x
i +y
j such that the cosine of the angle between u and

w equals the cosine of the angle between v and w . Thus
u w
u w
=
v w
v w
or
3x + 4y
5
_
x
2
+y
2
=
5x 12y
13
_
x
2
+y
2
.
Simplifying, we have x = 8y. The vector we want is of the form w = 8y
i +y
j , but should we take y > 0 or y < 0?

The smaller of the two angles formed by u and v is between 0
and 180
0
, and so w must make an acute angle with u
and v . If y > 0 then u w = 20y < 0 indicating an obtuse angle and if y < 0 then u w = 20y > 0 indicating an
acute angle. We have w = 8y
i +y
j with y < 0. Thus w can be any positive multiple of the vector 8
j .
57. Since u w = v w , (u v ) w = 0. This equality holds for any w , so we can take w = u v . This gives
u v
2
= (u v ) (u v ) = 0,
that is,
u v = 0.
This implies u v = 0, that is, u = v .
61. Let u and v be the displacement vectors from C to the other two vertices. Then
c
2
= u v
2
= (u v ) (u v )
= u u v u u v +v v
= u
2
2uv cos C +v
2
= a
2
2ab cos C +b
2
Exercises
1. v w =

k
j =
i (remember
i ,
j ,
k are unit vectors along the axes, and you must use the right hand rule.)
5. v = 2
i 3
j +
k , and w =
i + 2
k
v w =
i

j

k
2 3 1
1 2 1
i + 3
j + 7
k
9.
x
z
y
i +
j
Figure 13.13
By denition, (
i +
j ) (
j ) is in the direction of
k . The magnitude is
i +
j sin

2
=
2 1 = 2.
So (
i +
j ) (
j ) = 2
k . See Figure 13.13.

13. The displacement vector from (3, 4, 2) to (2, 1, 0) is:
a = 5
i 3
j 2
k .
The displacement vector from (3, 4, 2) to (0, 2, 1) is:
b = 3
i 2
k .
Therefore the vector normal to the plane is:
n =a
b =
i +
j +
k .
Using the rst point, the equation of the plane can be written as:
(x 3) + (y 4) + (z 2) = 0.
The equation of the plane is thus:
x +y +z = 3.
13.4 SOLUTIONS 291
Problems
17. We use the same normal n = 4
i + 26
j + 14
k and the point (0, 0, 0) to get 4(x 0) + 26(y 0) + 14(z 0) = 0, or

4x + 26y + 14z = 0.
21. The normal vectors to the planes are n1 = 2
i 3
j + 5
k and n2 = 4
i +

j 3
k . The line of intersection is

perpendicular to both normal vectors (picture the pages in a partially open book). Hence the vector we need is n1 n2 =
4
i + 26
j + 14
k .
25. (a) If we let
PQ in Figure 13.14 be the vector from point P to point Q and
PR be the vector from P to R, then
PQ =
i + 2
PR = 2
k ,
then the area of the parallelogram determined by
PQ and
PR is:
Area of
parallelogram
=
PQ
PR =
i

j

k
1 0 2
2 0 1
= 3
j = 3.
Thus, the area of the triangle PQR is
_
Area of
triangle
_
=
1
2
_
Area of
parallelogram
_
=
3
2
= 1.5.
x
y
z
P = (0, 1, 0)
Q = (1, 1, 2)
(2, 1, 1) = R
Figure 13.14
(b) Since n =

PQ
PR is perpendicular to the plane PQR, and from above, we have n = 3
j , the equation of the

plane has the form 3y = C. At the point (0, 1, 0) we get 3 = C, therefore 3y = 3, i.e., y = 1.
29. Since
v w = v w sin ,
and
v w = v w cos ,
so
v w
v w
=
v w sin
v w cos
= tan ,
so
tan =
2
i 3
j + 5
k
3
=
38
3
= 2.055.
33. The quantities
a (
b c )
and
(a
b ) c
both represent the volume of the same parallelepiped, namely that dened
by the three vectors a ,
b , and c , and therefore must be equal. Thus, the two triple products a (
b c ) and (a
b ) c
must be equal except perhaps for their sign. In fact, both are positive if a ,
b , c are right-handed and negative if a ,
b , c
are left-handed. This can be shown by drawing a picture:
a
c
b
right-handed
a
b
c
left-handed
Figure 13.15
37. Write v and w in components and expand using the distributive property of the cross product.
v w = (v1
i +v2
j +v3
k ) (w1
i +w2
j +w3
k )
= v1w1
i +v1w2
j +v1w3
k
+v2w1
i +v2w2
j +v2w3
k
+v3w1
i +v3w2
j +v3w3
k
Now we use the fact that
i =

0 ,
j =

k ,
k =
j ,
i =
k ,
j =

0 ,
k =
i ,
i =
j ,
j =
i ,
k =
0 . Thus we have
v w =

0 +v1w2
k +v1w3(
j ) +v2w1(
k ) +
0 +v2w3
i +v3w1
j +v3w2(
i ) +
0
= (v2w3 v3w2)
i + (v3w1 v1w3)
j + (v1w2 v2w1)
k .
41. The area vector for face OAB =
1
2
b a .
The area vector for face OBC =
1
2
a c .
The area vector for face OAC =
1
2
b c .
The area vector for face ABC =
1
2
(
b a ) (c a ).
1
2
b a +
1
2
c
b +
1
2
a c +
1
2
(
b a ) (c a ) =
1
2
b a +
1
2
c
b +
1
2
a c +
1
2
(
b c
b a a c a a ) = 0.
45. (a) Since
u v = (u2v3 u3v2)
i + (u3v1 u1v3)
j + (u1v2 u2v1)
k ,
we have
Area of S = u v =
_
(u2v3 u3v2)
2
+ (u3v1 u1v3)
2
+ (u1v2 u2v1)
2
_
1/2
.
(b) The two edges of R are given by the projections of u and v onto the xy-plane. These are the vectors

U and

V ,
obtained by omitting the
k -components of u and v : we have

U = u1
i +u2
j and

V = v1
i +v2
j , Thus
Area of R =
V = (u1v2 u2v1)
k = |u1v2 u2v1| .
(c) The vector m
i + n
k is normal to the plane z = mx + ny + c. Since the vectors u and v are in the plane
(theyre the sides of

S ), the vector u v is also normal to the plane. Thus, these two vectors are scalar multiples of
one another. Suppose
u v = (m
i +n
k )
SOLUTIONS to Review Problems for Chapter Thirteen 293
Since the
k component of u v is (u1v2 u2v1)
k , comparing the
k -components tells us that

= (u1v2 u2v1).
Thus,
(u1v2 u2v1)(m
i +n
k ) = u v = (u2v3 u3v2)
i + (u3v1 u1v3)
j + (u1v2 u2v1)
k ,
so
m =
u2v3 u3v2
u2v1 u1v2
n =
u3v1 u1v3
u2v1 u1v2
.
(d) We have
(1 +m
2
+n
2
) (Area of R)
2
=
_
1 +
_
u2v3 u3v2
u2v1 u1v2
_
2
+
_
u3v1 u1v3
u2v1 u1v2
_
2
_
(u1v2 u2v1)
2
= (u2v3 u3v2)
2
+ (u3v1 u1v3)
2
+ (u1v2 u2v1)
2
= (Area of S)
2
.
Exercises
1. Scalar. u v = (2
i 3
j 4
k ) (
j ) = 2 0 3(1) 4 1 = 1.
5. 5c = 5
i + 30
j
9. 3v w v = 2v w = 2(2
i + 3
k ) (
j + 2
k ) = 3
i + 7
j 4
k .
13. For any vector v , we have v v =
0 .
17. The cross product of two parallel vectors is
0 , so the cross product of any vector with itself is
0 .
21. (a) We have v w = 3 4 + 2 (3) + (2) 1 = 4.
(b) We have v w = 4
i 11
j 17
k .
(c) A vector of length 5 parallel to v is
5
v
v =
5
17
(3
i + 2
j 2
k ) = 3.64
i + 2.43
j 2.43
k .
(d) The angle between vectors v and w is found using
cos =
v w
v w
=
4
17
26
= 0.190,
so = 79.0
.
(e) The component of vector v in the direction of vector w is
v w
w
=
4
26
= 0.784.
(f) The answer is any vector a such that a v = 0. One possible answer is 2
i 2
j +
k .
(g) A vector perpendicular to both is the cross product:
v w = 4
i 11
j 17
k .
25. n = 4
i + 6
k (the coefcients of x, y, z are the same as the coefcients of
i ,
j , and
k .)
29. Since

F = 2
d , the two vectors are parallel in the same direction, so
F
parallel
=

F and

F perp =
0 .
The work done is
W =

F

d = 2 + 8 = 10.
Notice that this is the same as the magnitude of the force,
F =
20, times the distance traveled,
d =
5, since the
force is the same direction as the displacement.
33. The unit vector in the direction of

d is u = (1/
2)(
i +
j ). Thus
F
parallel
=
_
F u
_
u =
2
2
u =
i +
j ,
F perp =

F

F
parallel
=
j .
Notice that

F perp u = 0, as we expect. The work done is
W =

F

d = 2 0 = 2.
Problems
37. (a) The velocity vector for the boat is
b = 25
i and the velocity vector for the current is

c = 10 cos(45
i 10 sin(45
j = 7.07
i 7.07
j .
The actual velocity of the boat is
b +c = 17.93
i 7.07
j .
(b)
b +c = 19.27 km/hr.
(c) We see in Figure 13.16 that tan =
7.07
17.93
, so = 21.52
south of east.
T
c
7.07
E '
17.93
b
c
b +c
Figure 13.16
41. Vectors v 1, v 4, and v 8 are all parallel to each other. Vectors v 3, v 5, and v 7 are all parallel to each other, and are all
perpendicular to the vectors in the previous sentence. Vectors v 2 and v 9 are perpendicular.
45. Since the plane is normal to the vector 5
i +
j 2
k and passes through the point (0, 1, 1), an equation for the plane is
5x +y 2z = 5 0 + 1 1 + (2) (1) = 3
5x +y 2z = 3.
49. (a) Since
PQ = (3
i + 5
j + 7
k ) (
i + 2
j + 3
k ) = 2
i + 3
j + 4
k ,
and
PR = (2
i + 5
j + 3
k ) (
i + 2
j + 3
k ) =
i + 3
j ,
PQ
PR =
i

j

k
2 3 4
1 3 0
= 12
i + 4
j + 3
k ,
which is a vector perpendicular to the plane containing P, Q and R. Since
PQ
PR =
_
(12)
2
+ 4
2
+ 3
2
= 13,
the unit vectors which are perpendicular to a plane containing P, Q, and R are
12
13
i +
4
13
j +
3
13
k ,
or the unit vector pointing to the opposite direction,
12
13
i
4
13
j
3
13
k .
SOLUTIONS to Review Problems for Chapter Thirteen 295
(b) The angle between PQ and PR is for which
cos =
PQ
PR
PQ
PR
=
2 1 + 3 3 + 4 0
2
2
+ 3
2
+ 4
2
1
2
+ 3
2
+ 0
2
=
11
290
,
so
= cos
1
(
11
290
) 49.76
.
(c) The area of triangle PQR =
1
2
PQ
PR =
13
2
.
(d) Let d be the distance from R to the line through P and Q (see Figure 13.17), then
1
2
d
PQ = the area of PQR =

13
2
.
Therefore,
d =
13
PQ
=
13
2
2
+ 3
2
+ 4
2
=
13
29
.
P
R
Q
d
Figure 13.17
53. (a) On the x-axis, y = z = 0, so 5x = 21, giving x =
21
5
. So the only such point is (
21
5
, 0, 0).
(b) Other points are (0, 21, 0), and (0, 0, 3). There are many other possible answers.
(c) n = 5
j + 7
k . It is the normal vector.

(d) The vector between two points in the plane is parallel to the plane. Using the points from part (b), the vector 3
k
(21
j ) = 21
j + 3
k is parallel to the plane.

57. (a) Suppose v =
OP as in Figure 13.18. The
i component of
OP is the projection of
OP on the x-axis:
OT = v cos
i .
Similarly, the
j and
k components of
OP are the projections of
OP on the y-axis and the z-axis respectively. So:
OS = v cos
OQ = v cos
k
Since v =
OT +
OS +
OQ, we have
v = v cos
i +v cos
j +v cos
k .
(b) Since
v
2
= v v = (v cos
i +v cos
j +v cos
k )
(v cos
i +v cos
j +v cos
k )
= v
2
(cos
2
+ cos
2
+ cos
2
)
so
cos
2
+ cos
2
+ cos
2
= 1.
x
y
z
T
S
Q
Figure 13.18
61. (a) From the geometric denition of the dot product, we have
cos =
|a
b |
a
b
=
10
14
9
.
Using sin
2
= 1 cos
2
, we get
x + 2y + 3z = 0
2x +y + 2z = 0
x
2
+y
2
+z
2
= a
2
b
2
(1 cos
2
) = (14)(9)
_
1
100
(14)(9)
_
Solving these equations we get x = 1, y = 4, z = 3 or x = 1, y = 4, and z = 3. Thus c =
i 4
j + 3
k
or c =
i + 4
j 3
k .
(b) a
b =
i + 4
j 3
k . This is the same as one of the answers in part (a). The conditions in part (a) ensured that
c is perpendicular to a and
b and that it has magnitude a
b | sin |. The cross product is the solution that, in

addition, satises the right-hand rule.
1. False. There are exactly two unit vectors: one in the same direction as v and the other in the opposite direction. Explicitly,
the unit vectors parallel to v are
1
v
v .
5. False. If v and w are not parallel, the three vectors v , w and v w can be thought of as three sides of a triangle. (If
the tails of v and w are placed together, then v w is a vector from the head of w to the head of v .) The length of one
side of a triangle is less than the sum of the lengths of the other two sides. Alternatively, a counterexample is v =
i and
w =
j . Then
j =
2 but
j = 0.
9. False. To nd the displacement vector from (1, 1, 1) to (1, 2, 3) we subtract

i +
j +

k from
i + 2
j + 3
k to get
(1 1)
i + (2 1)
j + (3 1)
k =
j + 2
k .
13. True. The cosine of the angle between the vectors is negative when the angle is between /2 and .
17. False. If the vectors are nonzero and perpendicular, the dot product will be zero (e.g.
j = 0).
21. True. The cross product yields a vector.
25. False. If u and w are two different vectors both of which are parallel to v , then v u = v w =

0 , but u = w . A
counterexample is v =
i , u = 2
i and w = 3
i .
29. True. Any vector w that is parallel to v will give v w =

0 .
14.1 SOLUTIONS 297
CHAPTER FOURTEEN
Exercises
1. If h is small, then
fx(3, 2)
f(3 +h, 2) f(3, 2)
h
.
With h = 0.01, we nd
fx(3, 2)
f(3.01, 2) f(3, 2)
0.01
=
3.01
2
(2+1)

3
2
(2+1)
0.01
= 2.00333.
With h = 0.0001, we get
fx(3, 2)
f(3.0001, 2) f(3, 2)
0.0001
=
3.0001
2
(2+1)

3
2
(2+1)
0.0001
= 2.0000333.
Since the difference quotient seems to be approaching 2 as h gets smaller, we conclude
fx(3, 2) 2.
To estimate fy(3, 2), we use
fy(3, 2)
f(3, 2 +h) f(3, 2)
h
.
fy(3, 2)
f(3, 2.01) f(3, 2)
0.01
=
3
2
(2.01+1)

3
2
(2+1)
0.01
= 0.99668.
fy(3, 2)
f(3, 2.0001) f(3, 2)
0.0001
=
3
2
(2.0001+1)

3
2
(2+1)
0.0001
= 0.9999667.
Thus, it seems that the difference quotient is approaching 1, so we estimate
fy(3, 2) 1.
5. P/t: The unit is dollars per month. This is the rate at which payments change as the number of months it takes to pay
off the loan changes. The sign is negative because payments decrease as the pay-off time increases.
P/r: The unit is dollars per percentage point. This is the rate at which payments change as the interest rate
changes. The sign is positive because payments increase as the interest rate increases.
9. Moving right from P in the direction of increasing x increases f, so fx(P) > 0.
Moving up from P in the direction of increasing y increases f, so fy(P) > 0.
13. For fw(10, 25) we get
fw(10, 25)
f(10 +h, 25) f(10, 25)
h
.
Choosing h = 5 and reading values from Table 12.2 on page 644 of the text, we get
fw(10, 25)
f(15, 25) f(10, 25)
5
=
13 15
5
= 0.4
F/mph
This means that when the wind speed is 10 mph and the true temperature is 25
F, as the wind speed increases from

10 mph by 1 mph we feel an approximately 0.4
F drop in temperature. This rate is negative because the temperature you

feel drops as the wind speed increases.
298 Chapter Fourteen /SOLUTIONS
Problems
17. The values of z increase as we move in the direction of increasing x-values, so fx is positive. The values of z decrease as
we move in the direction of increasing y-values, so fy is negative. We see in the contour diagram that f(2, 1) = 10. We
estimate the partial derivatives:
fx(2, 1)
z
x
=
14 10
4 2
= 2,
fy(2, 1)
z
y
=
6 10
2 1
= 4.
21. (a) (i) Near A, the value of z increases as x increases, so fx(A) > 0.
(ii) Near A, the value of z decreases as y increases, so fy(A) < 0.
(b) fx(P) changes from positive to negative as P moves from A to B along a straight line, because after P crosses the
y-axis, z decreases as x increases near P.
fy(P) does not change sign as P moves from A to B along a straight line; it is negative along AB.
25. (a) Estimating T(x, t) from the gure in the text at x = 15, t = 20 gives
T
x
(15,20)
T(23, 20) T(15, 20)

23 15
=
20 23
8
=
3
8
C per m,
T
t
(15,20)
T(15, 25) T(15, 20)

25 20
=
25 23
5
=
2
5
C per min.
At 15 m from heater at time t = 20 min, the room temperature decreases by approximately 3/8
C per meter and

increases by approximately 2/5
C per minute.
(b) We have the estimates,
T
x
(5,12)
T(7, 12) T(5, 12)

7 5
=
25 27
2
= 1

C per m,
T
t
(5,12)
T(5, 40) T(5, 12)

40 12
=
30 27
28
=
3
28
C per min.
At x = 5, t = 12 the temperature decreases by approximately 1
C per meter and increases by approximately 3/28
C
per minute.
29. (a)
p
c
= fc(c, s) = rate of change in blood pressure as cardiac output increases while systemic vascular resistance
remains constant.
(b) Suppose that p = kcs. Note that c (cardiac output), a volume, s (SVR), a resistance, and p, a pressure, must all be
positive. Thus k must be positive, and our level curves should be conned to the rst quadrant. Several level curves
are shown in Figure 14.1. Each level curve represents a different blood pressure level. Each point on a given curve is
a combination of cardiac output and SVR that results in the blood pressure associated with that curve.
1 2 3
1
2
3
p = 9k
p = 4k
p = 1k
c (vol)
s (SVR)
Figure 14.1
A
B
c E
After
Nitroglycerine,
SVR drops
After Dopamine,
cardiac output increases
T
c (vol)
s (SVR)
Figure 14.2
14.2 SOLUTIONS 299
(c) Point B in Figure 14.2 shows that if the two doses are correct, the changes in pressure will cancel. The patients
cardiac output will have increased and his SVR will have decreased, but his blood pressure wont have changed.
(d) At point F in Figure 14.3, the patients blood pressure is normalized, but his/her cardiac output has dropped and his
SVR is up.
c (vol)
s (SVR)
c
2
c
1
D
E
F
'
T
E
resident
gives drug
increasing
SVR
during heart attack

cardiac output drops
Figure 14.3
Note: c1 and c2 are the cardiac outputs before and after the heart attack, respectively.
Exercises
1. (a) Make a difference quotient using the two points (3, 2) and (3, 2.01) that have the same x-coordinate 3 but whose
y-coordinates differ by 0.01. We have
fy(3, 2)
f(3, 2.01) f(3, 2)
2.01 2
=
28.0701 28
0.01
= 7.01.
(b) Differentiating gives fy(x, y) = x + 2y, so fy(3, 2) = 3 + 2 2 = 7.
5.
z
x
=

x
_
(x
2
+x y)
7
= 7(x
2
+x y)
6
(2x + 1) = (14x + 7)(x
2
+x y)
6
.
z
y
=

y
_
(x
2
+x y)
7
= 7(x
2
+x y)
6
.
9.

T
_
2r
T
_
=
2r
T
2
13. av =
2v
r
17.

r
_
2r
v
_
=
2
v
21. zx = sin x, zx(2, 3) = sin 2 0.9
25.

M
_
2r
3/2
GM
_
= 2r
3/2
(
1
2
)(GM)
3/2
(G) = r
3/2
G
GM
GM
=
r
3/2
M
GM
29.
V
r
=
8
3
rh and
V
h
=
4
3
r
2
.
33. zx = 7x
6
+yx
y1
, and zy = 2
y
ln 2 +x
y
ln x
Problems
37. (a) The difference quotient for evaluating fw(2, 2) is
fw(2, 2)
f(2 + 0.01, 2) f(2, 2)
h
=
e
(2.01) ln 2
e
2 ln 2
0.01
=
e
ln(2
2.01
)
e
ln(2
2
)
0.01
=
2
(2.01)
2
2
0.01
2.78
The difference quotient for evaluating fz(2, 2) is
fz(2, 2)
f(2, 2 + 0.01) f(2, 2)
h
=
e
2 ln(2.01)
e
2 ln 2
0.01
=
(2.01)
2
2
2
0.01
= 4.01
(b) Using the derivative formulas we get
fw =
f
w
= ln z e
wln z
= z
w
ln z
fz =
f
z
= e
wln z
w
z
= w z
w1
so
fw(2, 2) = 2
2
ln 2 2.773
fz(2, 2) = 2 2
21
= 4.
41. (a)
E
m
= c
2
_
1
_
1 v
2
/c
2
1
_
. We expect this to be positive because energy increases with mass.
(b)
E
v
= mc
2
1
2
_
(1 v
2
/c
2
)
3/2
_
2v
c
2
_
=
mv
(1 v
2
/c
2
)
3/2
. We expect this to be positive because energy
increases with velocity.
45. Since fx(x, y) = 4x
3
y
2
3y
4
, we could have
f(x, y) = x
4
y
2
3xy
4
.
In that case,
fy(x, y) =

y
(x
4
y
2
3xy
4
) = 2x
4
y 12xy
3
as expected. More generally, we could have f(x, y) = x
4
y
2
3xy
4
+C, where C is any constant.
Exercises
1. The partial derivatives are
zx = x and zy = 4y,
so
z(2, 1) = 4, zx(2, 1) = 2 and zy(2, 1) = 4.
The tangent plane to z =
1
2
(x
2
+ 4y
2
) at (x, y) = (2, 1) has equation
z = z(2, 1) +zx(2, 1)(x 2) +zy(2, 1)(y 1)
= 4 + 2(x 2) + 4(y 1)
= 4 + 2x + 4y.
14.3 SOLUTIONS 301
5. Since
zx = y cos xy, we have zx
_
2,
3
4
_
=
3
4
cos
_
2
3
4
_
= 0.
zy = xcos xy, we have zy
_
2,
3
4
_
= 2 cos
_
2
3
4
_
= 0.
Since
z
_
2,
3
4
_
= sin
_
2
3
4
_
= 1,
the tangent plane is
z = z
_
2,
3
4
_
+zx
_
2,
3
4
_
(x 2) +zy
_
2,
3
4
__
y
3
4
_
z = 1.
9. Since gu = 2u +v and gv = u, we have
dg = (2u +v) du +udv
13. We have df = fx dx+fy dy. Finding the partial derivatives, we have fx = e
y
so fx(1, 0) = e
0
= 1, and fy = xe
y
so fy(1, 0) = 1e
0
= 1. Thus, df = dx dy.
Problems
17. (a) The units are dollars/square foot.
(b) The price of land 300 feet from the beach and of area near 1000 square feet is greater for larger plots by about $3 per
square foot.
(c) The units are dollars/foot.
(d) The price of a 1000 square foot plot about 300 feet from the beach is less for plots farther from the beach by about
$2 per extra foot from the beach.
(e) Compared to the 998 ft
2
plot at 295 ft from the beach, the other plot costs about 7 3 = $21 more for the extra
7 square feet but about 10 2 = $20 less for the extra 10 feet you have to walk to the beach. The net difference is
about a dollar, and the smaller plot nearer the beach is cheaper.
21. Since fx(x, y) =
x
x
2
+y
3
and fy(x, y) =
3y
2
2
x
2
+y
3
,
fx(1, 2) =
1
1
2
+2
3
=
1
3
and fy(1, 2) =
32
2
2
1
2
+2
3
= 2.
Thus the differential at the point (1, 2) is
df = df(1, 2) = fx(1, 2)dx +fy(1, 2)dy =
1
3
dx + 2dy.
Using the differential at the point (1, 2), we can estimate f(1.04, 1.98). Since
f fx(1, 2)x +fy(1, 2)y
where f = f(1.04, 1.98) f(1, 2) and x = 1.04 1 and y = 1.98 2, we have
f(1.04, 1.98) f(1, 2) +fx(1, 2)(1.04 1) +fy(1, 2)(1.98 2)
=
_
1
2
+ 2
3
+
0.04
3
2(0.02) 2.973.
25. (a) The linear approximation gives
f(520, 24) 24.20, f(480, 24) 23.18,
f(500, 22) 25.52, f(500, 26) 21.86.
The approximations for f(520, 24) and f(500, 26) agree exactly with the values in the table; the other two do not.
The reason for this is that the partial derivatives were estimated using difference quotients with these values.
(b) We could get a more balanced estimate by using a difference quotient that uses the values on both sides. Thus, we
could estimate the partial derivatives as follows:
fT (500, 24)
f(520, 24) f(480, 24)
40
=
(24.20 23.19)
40
= 0.02525,
and
fp(500, 24)
f(500, 26) f(500, 22)
4
=
(21.86 25.86)
4
= 1.
This yields the linear approximation
V = f(T, p) 23.69 + 0.02525(T 500) (p 24) ft
3
.
This approximation yields values
f(520, 24) 24.195, f(480, 24) 23.185,
f(500, 22) 25.69, f(500, 26) 21.69.
Although none of these predictions are accurate, the error in the predictions that were wrong before has been reduced.
This new linearization is a better all-round approximation for values near (500, 24).
29. The error in is approximated by d, where
d =

r
dr +

p
dp.
We need to nd
r
=

8
p4r
3
v
and
p
=

8
r
4
v
.
For r = 0.005 and p = 10
5
we get
r
(0.005, 10
5
) = 3.14159 10
7
,

p
(0.005, 10
5
) = 0.39270,
so that
d =

r
dr +

p
dp.
is largest when we take all positive values to give
d = 3.14159 10
7
0.00025 + 0.39270 1000 = 8246.68.
This seems quite large but (0.005, 10
5
) = 39269.9 so the maximum error represents about 20% of any value
computed by the given formula. Notice also the relative error in r is 5%, which means the relative error in r
4
is 20%.
Exercises
1. Since the partial derivatives are
f
x
=
15
2
x
4
0 =
15
2
x
4
f
y
= 0
24
7
y
5
=
24
7
y
5
we have
grad f =
f
x
i +
f
y
j =
_
15
2
x
4
_
i
_
24
7
y
5
_
j .
14.4 SOLUTIONS 303
zx = e
y
and zy = xe
y
+e
y
+ye
y
,
we have
z = e
y
i +e
y
(1 +x +y)
j .
fr = sin and f
= r cos ,
we have
f = sin
i +r cos
j .
f
=
(2 3)(2 + 0) (2 0)(2 + 3)
(2 3)
2
=
4 6 (4 + 6)
(2 3)
2
=
12
(2 3)
2
f
=
(2 3)(0 + 3) (0 3)(2 + 3)
(2 3)
2
=
(6 9) + (6 + 9)
(2 3)
2
=
12
(2 3)
2
we have
grad f =
f
i +
f
j =
_
12
(2 3)
2
_
i +
_
12
(2 3)
2
_
j .
fr = 2(h +r) and f
h
= 2r,
we have
f(2, 3) = 10
i + 4
j .
f
x
=
1
2
(tan x +y)
1/2
_
1
cos
2
x
+ 0
_
=
1
2 cos
2
x
tan x +y
,
and
f
y
=
1
2
(tan x +y)
1/2
(0 + 1) =
1
2
tan x +y
,
then
grad f =
f
x
i +
f
y
j =
_
1
2 cos
2
x
tan x +y
_
i +
_
1
2
tan x +y
_
j .
Hence we have
grad f
(0,1)
=
_
1
2(cos(0))
2
_
tan(0) + 1
_
i +
_
1
2
_
tan(0) + 1)
_
j
=
_
1
2(1)
2
0 + 1
_
i +
_
1
2
0 + 1
_
j
=
1
2
i +
1
2
j
25. Since fx = 2 cos(2x y) and fy = cos(2x y), at (1, 2) we have grad f = 2 cos(0)
i cos(0)
j = 2
j . Thus
f
u
(1, 2) = grad f
_
3
5
i
4
5
j
_
=
2 3 1(4)
5
=
10
5
= 2.
29. Since df = fxdx +fydy and grad f = fx
i +fy
j we have
grad f = (x + 1)ye
x
i +xe
x
j .
33. Since the values of z increase as we move in direction
i +
j from the point (1, 1), the directional derivative is positive.

37. The approximate direction of the gradient vector at point (0, 2) is

j , since the gradient vector is perpendicular to the
contour and points in the direction of increasing z-values. Answers may vary since answers are approximate and any
positive multiple of the vector given is also correct.
41. The approximate direction of the gradient vector at point (2, 2) is
j , since the gradient vector is perpendicular to

the contour and points in the direction of increasing z-values. Answers may vary since answers are approximate and any
positive multiple of the vector given is also correct.
Problems
45. (a) The unit vector u in the same direction as v is
u =
1
v
v =
1
_
(1)
2
+ 3
2
v =
1
10
i +
3
10
j = 0.316228
i + 0.948683
j .
The vector w of length 0.1 in the direction of v is
w = 0.1u = 0.0316228
i + 0.0948683
j .
The displacement vector from P to Q is w . Hence
Q = (4 0.0316228, 5 + 0.0948683) = (3.96838, 5.09487).
(b) Since the distance from P to Q is 0.1, the directional derivative of f at P in the direction of Q is approximately
f
u

f(Q) f(P)
0.1
=
3.01052 3
0.1
= 0.1052.
(c) We have
grad f(x, y) =
1
2
x +y
i +
1
2
x +y
j
gradf(4, 5) =
1
6
i +
1
6
j .
The directional derivative at P = (4, 5) in the direction of u is
f
u
(4, 5) = grad f(4, 5) u =
1
6
1
10
+
1
6
3
10
=
1
3
10
= 0.1054.
49. At points (x, y) where the gradients are dened and are not the zero vector, the level curves of f and g intersect at right
angles if and only grad f grad g = 0.
We have grad f grad g = (
i +
j ) (
j ) = 0. The level curves of f and g are straight lines that cross at right
angles. See Figure 14.4.
14.4 SOLUTIONS 305
x
y
g = c
2
f = c
1
Figure 14.4
53. f
i
(3, 1) means the rate of change of f in the x direction at (3, 1). Thus,
f
i
(3, 1)
f(4, 1) f(3, 1)
1
=
2 1
1
= 1.
57. One way to do this is to estimate the gradient vector and then nd grad f(x, y) u . This is a useful approach since it is
easier to estimate grad f than to estimate f
u
directly. Since grad f(x, y) = (fx(x, y), fy(x, y)) we can simply estimate
the x- and y-derivatives of f at (3, 1) to nd grad f at that point. In the x-direction we see that the function is increasing.
This implies that the x-derivative is positive. To estimate its value we estimate the slope in the x-direction. Applying the
same reasoning to nd fy, we get:
fx(3, 1)
f(3 + 1, 1) f(3, 1)
1
= 2 1 = 1,
fy(3, 1)
f(3, 1) f(3, 1 0.6)
0.6

1 2
0.6
1.67.
This gives us our estimated value for grad f(3, 1)
i 1.67
j . Now, if u = (2
i +
j )/
5,
f
u
(3, 1) = grad f(3, 1) (2
i +
j )/
5
(
i 1.67
j ) (2
i +
j )/
5 1.64
61. First, we check that 3 = 2
2
1. Then let f(x, y) = y x
2
+ 1 so that the given curve is the contour f(x, y) = 0.
Since fx = 2x and fy = 1, we have grad f(2, 3) = 4
i +
j . Since gradients are perpendicular to contours, a vector

normal to the curve at (2, 3) is n = 4
i +1
j . Using the normal vector to a line the same way we use the normal vector
to a plane, we get that an equation of the tangent line is 4(x 2) + (y 3) = 0. Notice, if we had instead found the
slope of the tangent line using dy/dx = 2x, we get (y 3) = 4(x 2), which agrees with the equation we got using the
gradient.
65. (a) Negative. f is perpendicular to the level curve at the point P, so its x-component which is f
i is negative.
(b) Positive. The y-component of f is in the same direction as
j at P and hence the dot product will be positive.

(c) Positive. The partial derivative with respect to x at Q is positive because the value of f is increasing in the positive x
direction at Q. (Note that Q lies between the level curves with values 3 and 4 and that the one with value 4 is further
in the positive x direction from Q.)
(d) Negative. Again, Q lies between the level curves with values 3 and 4 and the one with value 3 is further from Q in
the positive y direction, so the partial derivative with respect to y at Q is negative.
69. We see that
gradf = (3y)
i + (3x + 2y)
j ,
so at the point (2, 3), we have
grad f = 9
i + 12
j .
(a) The directional derivative is f
v
v
=
(9)(3) + (12)(1)
10
=
15
10
4.74.
(b) The direction of maximum rate of change is f(2, 3) = 9
i + 12
j .
(c) The maximum rate of change is f =
225 = 15.
73. Directional derivative = f u , where u = unit vector. If we move from (4, 5) to (5, 6), we move in the direction
i +
j
so u =
1
i +
1
j . So,
f u = fx
_
1
2
_
+fy
_
1
2
_
= 2.
Similarly, if we move from (4, 5) to (6, 6), the direction is 2
i +
j so u =
2
i +
1
j . So
f u = fx
_
2
5
_
+fy
_
1
5
_
= 3.
Solving the system of equations for fx and fy
fx +fy = 2
2
2fx +fy = 3
5
gives
fx = 3
5 2
2
fy = 4
2 3
5.
Thus at (4, 5),
f = (3
5 2
2)
i + (4
2 3
5)
j .
77. (a) If
j points north and
i points east, then the direction the car is driving is
i . A unit vector in this direction is

u =
1
2
(
i ).
The gradient of the height function is
grad h =
h
E
i +
h
N
j = 50
i + 100
j .
So the directional derivative is
h
u
= grad g
h = (50
i + 100
j )
1
2
(
i ) =
100 50
2
= 35.355 ft/mi.
(b) The car is traveling at v mi/hr, so
Rate of change of with respect to time = v
miles
hour
35.355
ft
mile
= 35.355v ft/hour.
81. (a) To estimate the change in f, we use the gradient vector to estimate the change in f in moving from P to Q. Because
the contours are approximately parallel, moving from P to Q takes you to the same contour as moving from P to R.
(See Figure 14.5.) If is the angle between u and grad f(a, b), then
Change in f
between P and Q
= Change in f
=
_
Rate of change
in direction PR
__
Distance traveled
between P and R
_
grad f(hcos ).
14.5 SOLUTIONS 307
'
'
h
c
o
s
T
c
h
Q
u
grad f
P = (a, b)
R
Figure 14.5
(b) Since u is a unit vector, we use the denition of f
u
(a, b) to estimate
f
u
(a, b)
Change in f
h

grad f(a, b)hcos
h
= gradf(a, b) cos = grad fu cos = grad f(a, b) u .
This approximation gets better as we choose h smaller and smaller, and in the limit we get the formula:
f
u
(a, b) = grad f(a, b) u .
Exercises
1. Since fx = 2x, fy = 0 and fz = 0, we have
grad f = 2x
i .
5. Since f(x, y, z) =
1
xyz
, we have
fx =
1
x
2
yz
, fy =
1
xy
2
z
fz =
1
xyz
2
,
we have
gradf =
1
xyz
_
1
x
i +
1
y
j +
1
z
k
_
.
9. We have fx = e
y
sin z, fy = xe
y
sin z, and fz = xe
y
cos z. Thus
grad f = e
y
sin z
i +xe
y
sin z
j +xe
y
cos z
k .
13. We have fx = 0, fy = 2yz and fz = y
2
. Thus gradf = 2yz
j +y
2
k and grad f(1, 0, 1) =
0 .
17. We have
grad(xln(yz))
(2,1,e)
= ln(yz)
i +
x
y
j +
x
z
(2,1,e)
=
i + 2
j +
2
e
k .
21. We have gradf = y
i + x
j + 2z
k , so grad f(1, 1, 0) =

i +

j . A unit vector in the direction we want is u =
(1/
2)(
i +
k ). Therefore, the directional derivative is

gradf(1, 1, 0) u =
1(1) + 1 0 + 0 1
2
=
1
2
.
25. First, we check that (1)
2
(1)
2
+ 2
2
= 4. Then let f(x, y, z) = x
2
y
2
+ z
2
so that the given surface is the level
surface f(x, y, z) = 4. Since fx = 2x, fy = 2y, and fz = 2z, we have grad f(1, 1, 2) = 2
i 2
j + 4
k . Since
gradients are perpendicular to level surfaces, a vector normal to the surface at (1, 1, 2) is n = 2
i 2
j + 4
k . Thus
an equation for the tangent plane is
2(x + 1) 2(y 1) + 4(z 2) = 0.
29. First, we check that cos(1 + 1) = e
2+2
. Then we let f(x, y, z) = cos(x + y) e
xz+2
, so that the given surface
is the level surface f(x, y, z) = 0. Since fx = sin(x + y) ze
xz+2
, fy = sin(x + y), and fz = xe
xz+2
, we
have grad f(1, 1, 2) = 2
i +
k . Since gradients are perpendicular to level surfaces, a vector normal to the surface at
(1, 1, 2) is n = 2
i +
k . Thus an equation for the tangent plane is

2(x + 1) + (z 2) = 0.
33. (a) We have fx = 2x yz, fy = 2y xz and fz = xy so
grad f = (2x yz)
i + (2y xz)
j xy
k .
(b) At the point (2, 3, 1) we have
grad f(2, 3, 1) =
i + 4
j + 6
k .
Thus an equation of the tangent plane to the level surface at the point (2, 3, 1) is
(x 2) + 4(y 3) + 6(z 1) = 0
or
x + 4y + 6z = 20.
37. Let f(x, y, z) = x
2
+y
2
so that the surface is the level surface f(x, y, z) = 1. Since
grad f = 2x
i + 2y
j
we have
grad f(1, 0, 1) = 2
i .
Thus an equation of the tangent plane at the point (1, 0, 1) is
2(x 1) + 0(y 0) + 0(z 1) = 0
or
2(x 1) = 0.
The tangent plane is given by the equation
x = 1.
Problems
41. The gradient of (a) is 2x
i + 2y
j + 2z
k , which points radially outward from the origin, so (a) goes with (III).
The gradient of (c) is parallel to the gradient of (a) but pointing inward, so (c) goes with (IV).
The gradient of (b) is 2x
i + 2y
j , which points radially outward from the z-axis, so (b) goes with (I).
The gradient of (d) is parallel to the gradient of (b) but pointing inward, so (d) goes with (II).
14.5 SOLUTIONS 309
45. The surface is given by F(x, y, z) = 0 where F(x, y, z) = x y
3
z
7
. The normal direction is
grad F =
F
x
i +
F
y
j +
F
z
k =
i 3y
2
z
7
j 7y
3
z
6
k .
Thus, at (1, 1, 1) a normal vector is
i + 3
j + 7
k . The tangent plane has the equation

1(x 1) + 3(y (1)) + 7(z (1)) = 0
x + 3y + 7z = 9.
49. (a) A normal to the surface is given by the gradient of the function f(x, y, z) = x
2
+y
2
+ 3z
2
,
grad f = 2x
i + 2y
j + 6z
k .
At the point (0.6, 0.8, 1), a normal to the surface and to the tangent plane is
n = 1.2
i + 1.6
i + 6
k .
Since the plane goes through the point (0.6, 0.8, 1), its equation is
1.2(x 0.6) + 1.6(y 0.8) + 6(z 1) = 0
1.2x + 1.6y + 6z = 8.
(b) We want to know if there are x, y, z values such that a normal to the surface is parallel to the normal to the plane.
That is, is 2x
i + 2y
j + 6z
k parallel to 8
i + 6
j + 30
k ? Yes, if 2x = 8t, 2y = 6t, 6z = 30t for some value

of t. That is, if
x = 4t, y = 3t, z = 5t.
Substituting these equations into the equation for the surface and solving for t, we get
(4t)
2
+ (3t)
2
+ 3 (5t)
2
= 4
100t
2
= 4
t =
_
4
100
= 0.2.
So there are two points on the surface at which the tangent plane is parallel to the plane 8x + 6y + 30z = 1. They
are
(0.8, 0.6, 1).
53. (a) In the direction of grad F:
grad F
(1,1,1)
=
_
(2x + 2xz
2
)
i + (4y
3
)
j + (2x
2
z)
k
_
(1,1,1)
= 4
i + 4
j + 2
k .
(b) The rate of change in the direction of gradF with respect to distance = F =
16 + 16 + 4 = 6. Now we want
rate of change with respect to time. If we move at 4 units/sec:
Rate of change of
Conc
Time
= Rate of change of
Conc
Dist
Rate of change of
Dist
Time
= 6 4 = 24mg/cm
3
/sec.
57. (a) We have G = (2x 5y)
i + (5x + 2yz)
j + (y
2
)
k , so G(1, 2, 3) = 8
i + 7
j + 4
k . The rate of change is

given by the directional derivative in the direction v :
Rate of change in density = G
v
v
= (8
i + 7
j + 4
k )
(2
i +
j 4
k )
21
=
16 + 7 16
21
=
25
21
5.455.
(b) The direction of maximum rate of change is G(1, 2, 3) = 8
i + 7
j + 4
k .
(c) The maximum rate of change is G(1, 2, 3) =
_
(8)
2
+ 7
2
+ 4
2
=
129 11.36.
61. (a) is (V) since r +a is a vector not a scalar.
(b) is (IV) since grad(r a ) = grad(a1x +a2y +a3z) =a .
(c) is (V) since r a is a vector not a scalar.
65.
grad( r ) = grad(1x +2y +3z)
= 1
i +2
j +3
k = .
Exercises
1. Using the chain rule we see:
dz
dt
=
z
x
dx
dt
+
z
y
dy
dt
= y
2
e
t
+ 2xy cos t
= (sin t)
2
e
t
+ 2e
t
sin t cos t
= sin(t)e
t
(2 cos t sin t)
We can also solve the problem using one variable methods:
z = e
t
(sin t)
2
dz
dt
=
d
dt
(e
t
(sin t)
2
)
=
de
t
dt
(sin t)
2
+e
t
d(sin t)
2
dt
= e
t
(sin t)
2
+ 2e
t
sin t cos t
= e
t
sin t(2 cos t sin t)
5. Using the chain rule we see:
dz
dt
=
z
x
dx
dt
+
z
y
dy
dt
= 2t(sin y +y cos x) +
1
t
(xcos y + sin x)
= 2t sin(ln t) + 2t ln(t) cos(t
2
) +t cos(ln t) +
sin t
2
t
This problem can also be solved using one variable methods. Attempting to solve the problem that way will demonstrate
the advantage of using the chain rule.
9. Since z is a function of two variables x and y which are functions of two variables u and v, the two chain rule identities
which apply are:
z
u
=
z
x
x
u
+
z
y
y
u
= e
y
(2u) +xe
y
(2u)
= 2ue
y
(1 +x) = 2ue
(u
2
v
2
)
(1 +u
2
+v
2
).
z
v
=
z
x
x
v
+
z
y
y
v
= e
y
(2v) +xe
y
(2v)
= 2ve
y
(1 x) = 2ve
(u
2
v
2
)
(1 u
2
v
2
).
14.6 SOLUTIONS 311
13. Since z is a function of two variables x and y which are functions of two variables u and v, the two chain rule identities
which apply are:
z
u
=
z
x
x
u
+
z
y
y
u
z
v
=
z
x
x
v
+
z
y
y
v
First to nd z/u
z
u
= (e
y
ye
x
) sin v + (xe
y
+e
x
)(v sin u)
= (e
v cos u
v(cos u)e
usin v
) sin v (u(sin v)e
v cos u
+e
usin v
)v sin u
Now we nd z/v using the same method.
z
v
= (e
y
ye
x
)ucos v + (xe
y
+e
x
) cos u
= (e
v cos u
v(cos u)e
usin v
)ucos v + (u(sin v)e
v cos u
+e
usin v
) cos u
Problems
17.
2020 2040
13
14
15
t (years)
R(in)
T
c
2
E '
30
Figure 14.6: Global warming predictions:
Rainfall as a function of time
2020 2040
23
25
27
t (years)
T (
C)
E '
40
T
c
3
Figure 14.7: Global warming predictions:
Temperature as a function of time
We know that, as long as the temperature and rainfall stay close to their current values of R = 15 inches and
T = 30
C, a change, R, in rainfall and a change, T, in temperature produces a change, C, in corn production given
by
C 3.3R 5T.
Now both R and T are functions of time t (in years), and we want to nd the effect of a small change in time, t, on R
and T. Figure 14.6 shows that the slope of the graph for R versus t is about 2/30 0.07 in/year when t = 2020.
Similarly, Figure 14.7 shows the slope of the graph of T versus t is about 3/40 0.08
C/year when t = 2020. Thus,

around the year 2020,
R 0.07t and T 0.08t.
Substituting these into the equation for C, we get
C (3.3)(0.07)t (5)(0.08)t 0.6t.
Since at present C = 100, corn production will decline by about 0.6 % between the years 2020 and 2021. Now C
0.6t tells us that when t = 2020,
C
t
0.6, and therefore, that
dC
dt
0.6.
21.
w
x
w
y
w
z
w
x y z
u v
x
u
x
v
y
u
y
v
z
u
z
v
Figure 14.8
The tree diagram in Figure 14.8 tells us that
w
u
=
w
x
x
u
+
w
y
y
u
+
w
z
z
u
,
w
v
=
w
x
x
v
+
w
y
y
v
+
w
z
z
v
.
25. All are done using the chain rule.
(a) We have u = x, v = 3. Thus du/dx = 1 and dv/dx = 0 so
f
(x) = Fu(x, 3)(1) +Fv(x, 3)(0) = Fu(x, 3).

(b) We have u = 3, v = x. Thus du/dx = 0 and dv/dx = 1 so
f
(x) = Fu(3, x)(0) +Fv(3, x)(1) = Fv(3, x).

(c) We have u = x, v = x. Thus du/dx = dv/dx = 1 so
f
(x) = Fu(x, x)(1) +Fv(x, x)(1) = Fu(x, x) +Fv(x, x).

(d) We have u = 5x, v = x
2
. Thus du/dx = 5 and dv/dx = 2x so
f
(x) = Fu(5x, x
2
)(5) +Fv(5x, x
2
)(2x).
29. We have z = h(x, y) where h(x, y) = f(x)g(y), x = t, and y = t. The chain rule gives dz/dt = (h/x)dx/dt +
(h/y)dy/dt = h/x+h/y. Since h/x = f
(x)g(y) and h/y = f(x)g
(y) we have dz/dt = f
(x)g(y)+
f(x)g
(y) = f
(t)g(t) +f(t)g
(t).
33. To calculate
_
U
P
_
T
, we think of U as a function of P and T, as in U1(T, P). Thus
_
U
P
_
T
=
U1
P
.
37. (a) Thinking of V as a function of P and T gives
dV =
_
V
P
_
T
dP +
_
V
T
_
P
dT.
14.7 SOLUTIONS 313
(b) Substituting for dV in the following expression for dU,
dU =
_
U
P
_
V
dP +
_
U
V
_
P
dV,
we get
dU =
_
U
P
_
V
dP +
_
U
V
_
P
__
V
P
_
T
dP +
_
V
T
_
P
dT
_
.
Rearranging terms gives
dU =
__
U
P
_
V
+
_
U
V
_
P
_
V
P
_
T
_
dP +
_
U
V
_
P
_
V
T
_
P
dT.
(c) The formula for dU obtained by thinking of U as a function of P and T is
dU =
_
U
T
_
P
dT +
_
U
P
_
T
dP.
(d) Comparing coefcients of dP and dT in the two formulas gives
_
U
T
_
P
=
_
U
V
_
P
_
V
T
_
P
_
U
P
_
T
=
_
U
P
_
V
+
_
U
V
_
P
_
V
P
_
T
.
Exercises
1. We have fx = 6xy +5y
3
and fy = 3x
2
+15xy
2
, so fxx = 6y, fxy = 6x +15y
2
, fyx = 6x +15y
2
, and fyy = 30xy.
5. We have fx = 2ye
2xy
and fy = 2xe
2xy
, so fxx = 4y
2
e
2xy
, fxy = 4xye
2xy
+ 2e
2xy
, fyx = 4xye
2xy
+ 2e
2xy
, and
fyy = 4x
2
e
2xy
.
9. Since f(x, y) = sin(x
2
+y
2
), we have
fx = (cos (x
2
+y
2
))2x , fy = (cos (x
2
+y
2
))2y
fxx = (sin (x
2
+y
2
))4x
2
+ 2 cos (x
2
+y
2
)
fxy = (sin (x
2
+y
2
))4xy = fyx
fyy = (sin (x
2
+y
2
))4y
2
+ 2 cos (x
2
+y
2
).
13. The quadratic Taylor expansion about (0, 0) is given by
f(x, y) Q(x, y) = f(0, 0) +fx(0, 0)x +fy(0, 0)y +
1
2
fxx(0, 0)x
2
+fxy(0, 0)xy +
1
2
fyy(0, 0)y
2
.
First we nd all the relevant derivatives
f(x, y) = e
2x
2
y
2
fx(x, y) = 4xe
2x
2
y
2
fy(x, y) = 2ye
2x
2
y
2
fxx(x, y) = 4e
2x
2
y
2
+ 16x
2
e
2x
2
y
2
fyy(x, y) = 2e
2x
2
y
2
+ 4y
2
e
2x
2
y
2
fxy(x, y) = 8xye
2x
2
y
2
Now we evaluate each of these derivatives at (0, 0) and substitute into the formula to get as our nal answer:
Q(x, y) = 1 2x
2
y
2
f(x, y) Q(x, y) = f(0, 0) +fx(0, 0)x +fy(0, 0)y +
1
2
fxx(0, 0)x
2
+fxy(0, 0)xy +
1
2
fyy(0, 0)y
2
.
So rst we nd all the relevant derivatives:
f(x, y) = ln(1 +x
2
y)
fx(x, y) =
2x
1 +x
2
y
fy(x, y) =
1
1 +x
2
y
fxx(x, y) =
2(1 +x
2
y) 4x
2
(1 +x
2
y)
2
fyy(x, y) =
1
(1 +x
2
y)
2
fxy(x, y) =
2x
(1 +x
2
y)
2
Substituting into the formula we get as our answer:
Q(x, y) = y +x
2
y
2
2
21. (a) fx(P) < 0 because f decreases as you go to the right.
(b) fy(P) = 0 because f does not change as you go up.
(c) fxx(P) < 0 because fx decreases as you go to the right (fx changes from a small negative number to a large negative
number).
(d) fyy(P) = 0 because fy does not change as you go up.
(e) fxy(P) = 0 because fx does not change as you go up.
25. (a) fx(P) > 0 because f increases as you go to the right.
(b) fy(P) > 0 because f increases as you go up.
(c) fxx(P) = 0 because fx does not change as you go to the right. (Notice that the level curves are equidistant and
parallel, so the partial derivatives of f do not change if you move horizontally or vertically.)
(d) fyy(P) = 0 because fy does not change as you go up.
(e) fxy(P) = 0 because fx does not change as you go up.
29. Since f(x, y) =
1 + 2x y, the rst and second derivatives are

fx =
1
1 + 2x y
fy =
1
2
1 + 2x y
fxx =
1
(1 + 2x y)
3/2
fxy =
1
2(1 + 2x y)
3/2
fyy =
1
4(1 +x 2y)
3/2
,
so we nd that
f(0, 0) = 1
fx(0, 0) = 1
fy(0, 0) = 1/2
fxx(0, 0) = 1
fxy(0, 0) = 1/2
fyy(0, 0) = 1/4.
14.7 SOLUTIONS 315
The best quadratic approximation for f(x, y) for (x, y) near (0, 0) is
f(x, y) 1 +x
1
2
y
1
2
x
2
+
1
2
xy
1
8
y
2
.
Problems
33. We have f(1, 0) = 1 and the relevant derivatives are:
fx = e
y
so fx(1, 0) = 1
fy = xe
y
so fy(1, 0) = 1
fxx = 0 so fxx(1, 0) = 0
fxy = e
y
so fxy(1, 0) = 1
fyy = xe
y
so fyy(1, 0) = 1 .
Thus the linear approximation, L(x, y) to f(x, y) at (1, 0), is given by:
f(x, y) L(x, y) = f(1, 0) +fx(1, 0)(x 1) +fy(1, 0)(y 0)
= 1 + (x 1) y .
The quadratic approximation, Q(x, y) to f(x, y) near (1, 0), is given by:
f(x, y) Q(x, y) = f(1, 0) +fx(1, 0)(x 1) +fy(1, 0)(y 0) +
1
2
fxx(1, 0)(x 1)
2
+fxy(1, 0)(x 1)(y 0) +
1
2
fyy(1, 0)(y 0)
2
= 1 + (x 1) y (x 1)y +
1
2
y
2
.
The values of the approximations are
L(0.9, 0.2) = 1 0.1 0.2 = 0.7
Q(0.9, 0.2) = 1 0.1 0.2 + 0.02 + 0.02 = 0.74
and the exact value is
f(0.9, 0.2) = (0.9)e
0.2
0.737
Observe that the quadratic approximation is closer to the exact value.
37. Differentiating, we get
Fx = e
x
sin y +e
y
cos x Fy = e
x
cos y +e
y
sin x
Fxx = e
x
sin y e
y
sin x Fyy = e
x
sin y +e
y
sin x = Fxx
Thus, Fxx +Fyy = 0.
41. Since zy = g(x), zyy = 0, because g is a function of x only.
45. (a) Since P and Q lie on the same level curve, we have a = k.
(b) We have b = fx and c = fy. Since the gradient of f at P (respectively Q) points toward M or away from M, from
the gure, we see fx(P) and fy(P) have opposite signs, while fx(Q) and fy(Q) have the same signs. Thus Q is the
point (x1, y1), so P is (x2, y2).
(c) Since b = fx(Q) > 0 and c = fy(Q) > 0, the value of f must increase as we go away from M. Thus, M must be a
minimum (the surface is a valley).
(d) Since M is a minimum, m = fx(P) < 0 and n = fy(P) > 0.
49. If f and g have exactly the same contour diagrams inside a circle about the origin, then any partial derivative of f of
any order has the same value at the point (0, 0) as the corresponding partial derivative of g. Since the Taylor polynomials
of degree 2 for f and g near (0, 0) are constructed from the values of the partial derivatives at (0, 0), the two Taylor
polynomials are identical. It makes no difference what the contour diagrams look like outside the circle. See Figure 14.9
for one possible solution.
f(x, y)
x
y
g(x, y)
x
y
Figure 14.9
Exercises
1. Not differentiable at (0, 0).
5. Not differentiable where x = 0; that is, not differentiable on y-axis.
9. Not differentiable at (1, 2).
Problems
13. (a) The contour diagram for f(x, y) = xy/
_
x
2
+y
2
is shown in Figure 14.10.
2 2
2
2
x
y
1
.5
.25
1
.5
.25
1
.5
.25
1
.5
.25
Figure 14.10
(b) By the chain rule, f is differentiable at all points (x, y) where x
2
+y
2
= 0, and so at all points (x, y) = (0, 0).
(c) The partial derivatives of f are given by
fx(x, y) =
y
3
(x
2
+y
2
)
3/2
, for (x, y) = (0, 0),
and
fy(x, y) =
x
3
(x
2
+y
2
)
3/2
, for (x, y) = (0, 0).
Both fx and fy are continuous at (x, y) = (0, 0).
14.8 SOLUTIONS 317
(d) If f were differentiable at (0, 0), the chain rule would imply that the function
g(t) =
_
f(t, t), t = 0
0, t = 0
would be differentiable at t = 0. But
g(t) =
t
2
2t
2
=
1
t
2
|t|
=
1
2
|t|,
which is not differentiable at t = 0. Hence, f is not differentiable at (0, 0).
(e) The partial derivatives of f at (0, 0) are given by
fx(0, 0) = lim
x0
f(x, 0) f(0, 0)
x
= lim
x0
x0
x
2
+0
2
0
x
= lim
x0
0 0
x
= 0,
fy(0, 0) = lim
y0
f(0, y) f(0, 0)
y
= lim
y0
0y
0
2
+y
2
0
y
= lim
y0
0 0
y
= 0.
The limit lim
(x,y)(0,0)
fx(x, y) does not exist since if we choose x = y = t, t = 0, then
fx(x, y) = fx(t, t) =
t
3
(2t
2
)
3/2
=
t
3
2
2 |t|
3
=
_
1
2
2
, t > 0,
1
2
2
, t < 0.
Thus, fx is not continuous at (0, 0). Similarly, fy is not continuous at (0, 0).
17. (a) The contour diagram of f(x, y) =
_
|xy| is shown in Figure 14.11.
2 2
2
2
x
y
.5
1
1.5
.5
1
1.5
.5
1
1.5
.5
1
1.5
Figure 14.11
x
y
z
Figure 14.12
(b) The graph of f(x, y) =
_
|xy| is shown in Figure 14.12.
(c) f is clearly differentiable at (x, y) where x = 0 and y = 0. So we need to look at points (x0, 0), x0 = 0 and (0, y0),
y0 = 0. At (x0, 0):
fx(x0, 0) = lim
xx
0
f(x, 0) f(x0, 0)
x x0
= 0
fy(x0, 0) = lim
y0
f(x0, y) f(x0, 0)
y
= lim
y0
_
|x0y|
y
which does not exist. So f is not differentiable at the points (x0, 0), x0 = 0. Similarly, f is not differentiable at the
points (0, y0), y0 = 0.
(d)
fx(0, 0) = lim
x0
f(x, 0) f(0, 0)
x
= 0
fy(0, 0) = lim
y0
f(0, y) f(0, 0)
y
= 0
(e) Let u = (
i +
j )/
2:
f
u
(0, 0) = lim
t0
+
f(
t
2
,
t
2
) f(0, 0)
t
= lim
t0
+
_
t
2
2
t
=
1
2
.
We know that f(0, 0) =

0 because both partial derivatives are 0. But if f were differentiable, f
u
(0, 0) =
f(0, 0) u = fx(0, 0)
1
2
+ fy(0, 0)
1
2
= 0. But since, in fact, f
u
(0, 0) = 1/
2, we conclude that f
is not differentiable.
Exercises
1. Vector. Taking the gradient and substituting x = 1, y = 2 gives
grad(x
3
e
y/2
)
(1,2)
=
_
3x
2
e
y/2
i
x
3
2
e
y/2
j
_
(1,2)
= 3e
1
i
1
2
e
1
j .
5. Taking the derivative of f with respect to x and treating y as a constant we get fx = 2xy + 3x
2
7y
6
. To nd fy we
take the derivative of f with respect to y and treat x as a constant. Thus, fy = x
2
42xy
5
.
9. For both partial derivatives we use the quotient rule. Thus,
fx =
2xy(x
2
+y
2
) x
2
y2x
(x
2
+y
2
)
2
=
2x
3
y + 2xy
3
2x
3
y
(x
2
+y
2
)
2
=
2xy
3
(x
2
+y
2
)
2
,
and
fy =
x
2
(x
2
+y
2
) x
2
y2y
(x
2
+y
2
)
2
=
x
4
+x
2
y
2
2x
2
y
2
(x
2
+y
2
)
2
=
x
4
x
2
y
2
(x
2
+y
2
)
2
.
13. Since we take the derivative with respect to N, we use the power rule to obtain fN = cN
1
V
.
17. zx =
1
2ay
(2)
1
x
3
+
15x
4
abc
y
=
1
ax
3
y
+
15abcx
4
y
=
15a
2
bcx
7
1
ax
3
y
21. Using the chain rule and the quotient rule,
w
_
x
2
yw xy
3
w
7
w 1
_
7/2
=
7
2
_
x
2
yw xy
3
w
7
w 1
_
9/2
_
(w 1)(x
2
y 7xy
3
w
6
) (x
2
yw xy
3
w
7
)(1)
(w 1)
2
_
=
7
2
_
w 1
x
2
yw xy
3
w
7
_
9/2
_
(w 1)(x
2
y 7xy
3
w
6
) (x
2
yw xy
3
w
7
)
(w 1)
2
_
=
7
2
_
w 1
x
2
yw xy
3
w
7
_
9/2
x
2
y + 6xy
3
w
7
7xy
3
w
6
(w 1)
2
SOLUTIONS to Review Problems for Chapter Fourteen 319
25. The rst order partial derivatives are
ux = e
x
sin y, uy = e
x
cos y.
Thus the second order partials are
uxx = e
x
sin y, uyy = e
x
sin y.
29. Since fx = 2x, fy = 2y + 3y
2
and fz = 0, we have
grad f = 2x
i + (2y + 3y
2
)
j .
33. Since fx = 2xsin(x
2
+y
2
+z
2
), fy = 2y sin(x
2
+y
2
+z
2
) and fz = 2z sin(x
2
+y
2
+z
2
), we have
grad f = 2xsin(x
2
+y
2
+z
2
)
i + 2y sin(x
2
+y
2
+z
2
)
j + 2z sin(x
2
+y
2
+z
2
)
k .
f
x
= cos(xy) (y) sin(xy) (y) = y[cos(xy) sin(xy)]
f
y
= cos(xy) (x) sin(xy) (x) = x[cos(xy) sin(xy)]
we have
grad f =
f
x
i +
f
y
j
= y[cos(xy) sin(xy)]
i +x[cos(xy) sin(xy)]
j .
41. We have grad f = e
y
i +xe
y
j , so gradf(3, 0) =
i +3
j . A unit vector in the direction we want is u = (1/5)(4
i 3
j ).
Therefore, the directional derivative is
grad f(3, 0) u =
1 4 + 3(3)
5
= 1.
45. We have gradf = (e
x+z
cos y)
i (e
x+z
sin y)
j + (e
x+z
cos y)
k , so grad f(1, 0, 1) =
i +
k . A unit vector in the

direction we want is u = (1/
3)(
i +
j +
k ). Therefore, the directional derivative is

grad f(1, 0, 1) u =
1 1 + 1 1
3
=
2
3
.
49. Let f(x, y, z) = z
2
4x
2
3y
2
so that the surface (a hyperboloid) is the level surface f(x, y, x) = 9. Since
grad f = 8x
i 6y
j + 2z
k
we have
grad f(1, 1, 4) = 8
i 6
j + 8
k .
Thus an equation of the tangent plane at (1, 1, 4) is
8(x 1) 6(y + 6) + 8(z 4) = 0
so
4x 3y + 4z = 9.
f(x, y) Q(x, y) = f(0, 0) +fx(0, 0)x +fy(0, 0)y +
1
2
fxx(0, 0)x
2
+fxy(0, 0)xy +
1
2
fyy(0, 0)y
2
.
First we nd all the relevant derivatives
f(x, y) = (x + 1)
3
(y + 2)
fx(x, y) = 3(x + 1)
2
(y + 2)
fy(x, y) = (x + 1)
3
fxx(x, y) = 6(x + 1)(y + 2)
fyy(x, y) = 0
fxy(x, y) = 3(x + 1)
2
Now we evaluate each of these derivatives at (0, 0) and substitute into the formula to get as our nal answer:
Q(x, y) = 2 + 6x +y + 6x
2
+ 3xy
Notice this is the same as what you get if you expand (x + 1)
3
(y + 2) and keep only the terms of degree 2 or less.
Problems
57. (a) Let f(x, y, z) = 2x
2
2xy
2
+az so that the given surface is the level surface f(x, y, z) = a. Since
grad f = (4x 2y
2
)
i 4xy
j +a
k
we have
grad f(1, 1, 1) = 2
i 4
j +a
k .
Thus an equation of the tangent plane at (1, 1, 1) is
2(x 1) 4(y 1) +a(z 1) = 0
or
2x 4y +az = a 2.
(b) Substituting x = 0, y = 0, z = 0 into the equation for the tangent plane in part (a) we have 0 = a 2. The tangent
plane passes through the origin if a = 2.
61. The partial derivative, Q/b is the rate of change of the quantity of beef purchased with respect to the price of beef,
when the price of chicken stays constant. If the price of beef increases and the price of chicken stays the same, we expect
consumers to buy less beef and more chicken. Thus when b increases, we expect Q to decrease, so Q/b < 0.
On the other hand, Q/c is the rate of change of the quantity of beef purchased with respect to the price of chicken,
when the price of beef stays constant. An increase in the price of chicken is likely to cause consumers to buy less chicken
and more beef. Thus when c increases, we expect Q to increase, so Q/c > 0.
65. We need the partial derivatives, fx(1, 0) and fy(1, 0). We have
fx(x, y) = 2xe
xy
+x
2
ye
xy
, so fx(1, 0) = 2
fy(x, y) = x
3
e
xy
, so fy(1, 0) = 1.
(a) Since f(1, 0) = 1, the tangent plane is
z = f(1, 0) + 2(x 1) + 1(y 0) = 1 + 2(x 1) +y.
(b) The linear approximation can be obtained from the equation of the tangent plane:
f(x, y) 1 + 2(x 1) +y.
(c) At (1, 0), the differential is
df = fxdx +fydy = 2dx +dy.
69. The directional derivative is approximately the change in z (as we move in direction v ) divided by the horizontal change
in position. The directional derivative is f
i

2 1
4 1
=
1
3
0.3.
73. The directional derivative is approximately the change in z (as we move in direction v ) divided by the horizontal
change in position. In the direction of v we go from point (3, 3) to point (1, 4). We have the directional derivative
2 3
2
i +
j
=
1
5
0.4.
77. (a) Incorrect. grad H is not the rate of change of H. In fact, theres no such thing as the rate of change of H, although
directional derivatives can give its rate of change in a particular direction. For example, this expression would give
the wrong answer if the ant was crawling along a contour of H, since then the rate of change of the temperature it
experiences is zero even though grad H and v might not be zero.
(b) Correct. If u = v /v , then
grad H v = grad H
v
v
v = (grad H u )v = H
u
v
= (Rate of change of H in direction u in deg/cm)(Speed of ant in cm/sec)
= Rate of change of H in deg/sec.
(c) Incorrect, this is the directional derivative, which gives the rate of change with respect to distance, not time.
81. We have
P
x
= 5e
0.1
x
2
+y
2
+z
2
(0.1)(2x)
1
2
(x
2
+y
2
+z
2
)
1/2
= 0.5e
0.1
x
2
+y
2
+z
2 x
_
x
2
+y
2
+z
2
,
and similarly
P
y
= 0.5e
0.1
x
2
+y
2
+z
2 y
_
x
2
+y
2
+z
2
, and
P
z
= 0.5e
0.1
x
2
+y
2
+z
2 z
_
x
2
+y
2
+z
2
.
Thus the gradient of P at (0, 0, 1) is
grad P =
P
x
(x,y,z)=(0,0,1)
i +
P
y
(x,y,z)=(0,0,1)
j +
P
z
(x,y,z)=(0,0,1)
k = 0.5e
0.1
k .
Let u be a unit vector in the direction of v , so
u =
v
v
.
Then
Rate of change of pressure in atm/sec = (Directional derivative in direction u in atm/mi)(Speed of spacecraft in mi/sec)
= P
u
v = (grad P u )v = grad P
v
v
v = grad P v
= 0.5e
0.1
k (
i 2.5
k ) = 0.5 2.5e
0.1
= 1.131 atm/sec.
85. (a) The chain rule gives
z
x
=
z
u
u
x
+
z
v
v
x
= (2u e
v
)1 + (ue
v
)2.
At (x, y) = (1, 2), we have u = 1 + 2 2 = 5 and v = 2 1 2 = 0, so
z
y
(x,y)=(1,2)
= (2 5 e
0
)1 5e
0
2 = 1.
(b) The chain rule gives
z
y
=
z
u
u
y
+
z
v
v
y
= (2u e
v
)2 + (ue
v
)(1).
At (x, y) = (1, 2), we have (u, v) = (5, 0), so
z
y
(x,y)=(1,2)
= (2 5 e
0
)2 + 5e
0
= 23.
89. The error, dT, in the period T is given by
dT =
T
l
dl +
T
g
dg,
where
T
l
=
_
l
g

l
and
T
g
=
_
l
g

g
,
so that
T
l
(2, 9.8) = 0.7096, Tg(2, 9.8) = 0.1448.
We also have that
dl = 0.01, dg = 0.01.
The maximum discrepancy in the period is then given by
dT = 0.7096(0.01) 0.1448(0.01) = 0.008544.
93. The directional derivative, or slope, of f at (2, 1) in the direction perpendicular to these contours is the length of the
gradient. At the point (2, 1), we have grad f = 3
i + 4
j , so grad f = 5. Thus if d is the distance between the

contours, we have
7.3 7
d
= Slope in direction perpendicular to contours = 5.
Thus d = 0.3/5 = 0.06.
97. By the chain rule,
fr(2, 1) = fx(2, 1) cos +fy(2, 1) sin
f
(2, 1) = fx(2, 1)(r sin ) +fy(2, 1)(r cos ).

Since x
2
+y
2
= r
2
, we have r =
2
2
+ 1
2
=
5, and cos = x/r = 2/
5, and sin = y/r = 1/
5. Thus
fr(2, 1) = (3)
2
5
+ (4)
1
5
=
2
5
f
(2, 1) = (3)(1) + (4)(2) = 11.

For u = (1/
5)(2
i +
j ), we have f
u
(2, 1) = 2/
5. Since u is pointing radially out from the origin to the point

(2, 1), we expect that f
u
(2, 1) = fr(2, 1).
101. (a) The rst-order Taylor polynomial of a function f about a point (a, b) is equal to
f(a, b) +fx(a, b)(x a) +fy(a, b)(y b).
Computing the partial derivatives, we get:
fx = 2(x 1)e
(x1)
2
+(y3)
2
fy = 2(y 3)e
(x1)
2
+(y3)
2
fx(0, 0) = 2(1)e
(1)
2
+(3)
2
= 2e
10
fy(0, 0) = 2(3)e
(1)
2
+(3)
2
= 6e
10
Thus,
f(x, y) e
10
2e
10
x 6e
10
y
(b) The second-order Taylor polynomial of a function f about the point (1, 3) is given by
f(1, 3) + fx(1, 3)(x 1) +fy(1, 3)(y 3)
+
1
2
fxx(1, 3)(x 1)
2
+fxy(1, 3)(x 1)(y 3) +
1
2
fyy(1, 3)(y 3)
2
.
Computing the partial derivatives, we get:
fx = 2(x 1)e
(x1)
2
+(y3)
2
fy = 2(y 3)e
(x1)
2
+(y3)
2
fxx = (4(x 1)
2
+ 2)e
(x1)
2
+(y3)
2
fxy = 4(x 1)(y 3)e
(x1)
2
+(y3)
2
fyy = (4(y 3)
2
+ 2)e
(x1)
2
+(y3)
2
Substituting in the point (1, 3) to these partial derivatives, we get:
fx(1, 3) = 0
fy(1, 3) = 0
fxy(1, 3) = 0
fxx(1, 3) = (4(0)
2
+ 2)e
0
2
+0
2
= 2
fyy(1, 3) = (4(0)
2
+ 2)e
0
2
+0
2
= 2
Thus,
f(x, y) e
0
+ 0(x 1) + 0(y 3) +
2
2
(x 2)
2
+ 0(x 1)(y 3) +
2
2
(y 3)
2
f(x, y) 1 + (x 1)
2
+ (y 3)
2
.
(c) A vector perpendicular to the level curve is grad f. At the point (0, 0), we have
grad f = fx(0, 0)
i +fy(0, 0)
j
Computing partial derivatives, we have
fx = 2(x 1)e
(x1)
2
+(y3)
2
fy = 2(y 3)e
(x1)
2
+(y3)
2
fx(0, 0) = 2(1)e
(1)
2
+(3)
2
= 2e
10
fy(0, 0) = 2(3)e
(1)
2
+(3)
2
= 6e
10
Therefore, a perpendicular vector is grad f = 2e
10
i 6e
10
j . Any multiple of grad f, say 2
i 6
j , will
do.
(d) Since the surface can be represented by the level surface
F(x, y, z) = f(x, y) z = 0,
a vector perpendicular to the surface at (0, 0) is given by
grad F = fx(0, 0)
i +fy(0, 0)
k = 2e
10
i 6e
10
k
105. (a) We have f(1, 2) = A0 +A1 +2A2 +A3 +2A4 +4A5, fx(1, 2) = A1 +2A3 +2A4, and fy(1, 2) = A2 +A4 +4A5,
so the linear approximation is
L(x, y) = A0 +A1 + 2A2 +A3 + 2A4 + 4A5 + (A1 + 2A3 + 2A4)(x 1) + (A2 +A4 + 4A5)(y 2).
Also, m(t) = 1 +B1t and n(t) = 2 +C1t.
(b) Using a CAS to compute the derivatives, we nd that they are both the same:
d
dt
f(x(t), y(t))|t=0 =
d
dt
l(m(t), n(t))|t=0 = A1B1 + 2A3B1 + 2A4B1 +A2C1 +A4C1 + 4A5C1.
This can be explained using the chain rule and the fact that the derivative of a function at a point is the same as the
derivative of its linear approximation there:
d
dt
f(x(t), y(t))|t=0 =

x
f(x(0), y(0))x
(0) +

y
f(x(0), y(0))y
(0)
=

x
l(x(0), y(0))m
(0) +

y
l(x(0), y(0))n
(0) =
d
dt
l(m(t), n(t))|t=0
1. True. This is the instantaneous rate of change of f in the x-direction at the point (10, 20).
5. True. The property fx(a, b) > 0 means that f increases in the positive x-direction near (a, b), so f must decrease in the
negative x-direction near (a, b).
9. True. A function with constant fx(x, y) and fy(x, y) has constant x-slope and constant y-slope, and therefore has a graph
which is a plane.
13. True. Since g is a function of x only, it can be treated like a constant when taking the y partial derivative.
17. False. The equation z = 2 +2x(x 1) +3y
2
(y 1) is not linear. The correct equation is z = 2 +2(x 1) +3(y 1),
which is obtained by evaluating the partial derivatives at the point (1, 1).
21. False. The graph of f is a paraboloid, opening upward. The tangent plane to this surface at any point lies completely under
the surface (except at the point of tangency). So the local linearization underestimates the value of f at nearby points.
25. True. If f is linear, then f(x, y) = mx+ny +c for some m, n and c. So fx = mand fy = n giving df = mdx+ndy,
which is linear in the variables dx and dy.
29. False. The gradient is perpendicular to the contour of f at (a, b).
33. False. The gradient vector at (3, 4) has no relation to the direction of the vector 3
i +4
j . For example, if f(x, y) = x+2y,

then grad f =
i + 2
j , which is not perpendicular to 3
i + 4
j . The gradient vector grad f(3, 4) is perpendicular to the

contour of f passing through the point (3, 4).
37. True. The length of the gradient gives the maximal directional derivative in any direction. The gradient vector is grad f(0, 0) =
i +
j , which has length
2.
41. True. The denition of fx(x, y) is the limit of the difference quotient
fx(x, y) = lim
h0
f(x +h, y) f(x, y)
h
.
The symmetry of f gives
f(x +h, y) f(x, y)
h
=
f(y, x +h) f(y, x)
h
.
The denition of fy(y, x) is the limit of the difference quotient
fy(y, x) = lim
h0
f(y, x +h) f(y, x)
h
.
Thus fx(x, y) = fy(y, x).
45. True. Take the direction perpendicular to grad f at that point. If grad f = 0, any direction will do.
15.1 SOLUTIONS 325
CHAPTER FIFTEEN
Exercises
1. The point A is not a critical point and the contour lines look like parallel lines. The point B is a critical point and is a local
maximum; the point C is a saddle point.
5. At the origin, the second derivative test gives
D = kxxkyy (kxy)
2
=
_
(sin xsin y)(sin xsin y) (cos xcos y)
2
_
x=0,y=0
= sin
2
0 sin
2
0 cos
2
0 cos
2
0
= 1 < 0.
Thus k(0, 0) is a saddle point.
9. To nd the critical points, we solve fx = 0 and fy = 0 for x and y. Solving
fx = 3x
2
6x = 0
fy = 2y + 10 = 0
shows that x = 0 or x = 2 and y = 5. There are two critical points: (0, 5) and (2, 5).
We have
D = (fxx)(fyy) (fxy)
2
= (6x 6)(2) (0)
2
= 12x 12.
When x = 0, we have D = 12 < 0, so f has a saddle point at (0, 5). When x = 2, we have D = 12 > 0 and
fxx = 6 > 0, so f has a local minimum at (2, 5).
13. Find the critical point(s) by setting
fx = (xy + 1) + (x +y) y = y
2
+ 2xy + 1 = 0,
fy = (xy + 1) + (x +y) x = x
2
+ 2xy + 1 = 0,
then we get x
2
= y
2
, and so x = y or x = y.
If x = y, then x
2
+2x
2
+1 = 0, that is, 3x
2
= 1, and there is no real solution. If x = y, then x
2
2x
2
+1 = 0,
which gives x
2
= 1. Solving it we get x = 1 or x = 1, then y = 1 or y = 1, respectively. Hence, (1, 1) and (1, 1)
are critical points.
Since
fxx(x, y) = 2y,
fxy(x, y) = 2y + 2x and
fyy(x, y) = 2x,
the discriminant is
D(x, y) = fxxfyy f
2
xy
= 2y 2x (2y + 2x)
2
= 4(x
2
+xy +y
2
).
thus
D(1, 1) = 4(1
2
+ 1 (1) + (1)
2
) = 4 < 0,
D(1, 1) = 4((1)
2
+ (1) 1 + 1
2
) = 4 < 0.
Therefore (1, 1) and (1, 1) are saddle points.
326 Chapter Fifteen /SOLUTIONS
Problems
17. At the critical point x = 1, y = 0,
fx = 2x +A = 0, so 2 +A = 0 or A = 2
fy = 2y = 0.
Thus, f(x, y) = x
2
2x +y
2
+B has a critical point at (1, 0). Since fxx = 2 and fyy = 2 and fxy = 0 at (1, 0),
D = fxxfyy (fxy)
2
= 2 2 0
2
= 4,
so the second derivative test shows the critical point at (1, 0) is a local minimum. The value of the minimum is
f(1, 0) = 1
2
2 1 + 0
2
+B = 20, so B = 21.
21. (a) P is a local maximum.
(b) Q is a saddle point.
(c) R is a local minimum.
(d) S is none of these.
25. At a critical point,
fx = cos xsin y = 0 so cos x = 0 or sin y = 0;
and
fy = sin xcos y = 0 so sin x = 0 or cos y = 0.
Case 1: Assume cos x = 0. This gives
x =
3
2
,
2
,

2
,
3
2
,
(This can be written more compactly as: x = k +/2, for k = 0, 1, 2, .)
If cos x = 0, then sin x = 1 = 0. Thus in order to have fy = 0 we need cos y = 0, giving
y =
3
2
,
2
,

2
,
3
2
,
(More compactly, y = l +/2, for l = 0, 1, 2, )
Case 2: Assume sin y = 0. This gives
y = 2, , 0, , 2,
(More compactly, y = l, for l = 0, 1, 2, )
If sin y = 0, then cos y = 1 = 0, so to get fy = 0 we need sin x = 0, giving
x = , 2, , 0, , 2,
(More compactly, x = k for k = 0, 0, 1, 2, )
Hence we get all the critical points of f(x, y). Those from Case 1 are as follows:

_
2
,
2
_
,
_
2
,

2
_
,
_
2
,
3
2
_

_
2
,
2
_
,
_
2
,

2
_
,
_
2
,
3
2
_

_
3
2
,
2
_
,
_
3
2
,

2
_
,
_
3
2
,
3
2
_

Those from Case 2 are as follows:
(, ), (, 0), (, ), (, 2)
(0, ), (0, 0), (0, ), (0, 2)
(, ), (, 0), (, ), (, 2)
15.1 SOLUTIONS 327
More compactly these points can be written as,
(k, l), for k = 0, 1, 2, , l = 0, 1, 2,
and (k +

2
, l +

2
), for k = 0, 1, 2, , l = 0, 1, 2,
To classify the critical points, we nd the discriminant. We have
fxx = sin xsin y, fyy = sin xsin y, and fxy = cos xcos y.
Thus the discriminant is
D(x, y) = fxxfyy f
2
xy
= (sin xsin y)(sin xsin y) (cos xcos y)
2
= sin
2
xsin
2
y cos
2
xcos
2
y
= sin
2
y cos
2
x. (Use: sin
2
x = 1 cos
2
x and factor.)
At points of the form (k, l) where k = 0, 1, 2, ; l = 0, 1, 2, , we have
D(x, y) = 1 < 0 so (k, l) are saddle points.
At points of the form (k +

2
, l +

2
) where k = 0, 1, 2, ; l = 0, 1, 2,
D(k +

2
, l +

2
) = 1 > 0, so we have two cases:
If k and l are both even or k and l are both odd, then
fxx = sin xsin y = 1 < 0, so (k +

2
, l +

2
) are local maximum points.
If k is even but l is odd or k is odd but l is even, then
fxx = 1 > 0 so (k +

2
, l +

2
) are local minimum points.
29. (a) (a, b) is a critical point. Since the discriminant D = fxxfyy f
2
xy
= f
2
xy
< 0, (a, b) is a saddle point.
1
1
x
y
0
0
0
0
1
3
5
7
9
1
3
5
7
9
7
Figure 15.1
33. The rst order partial derivatives are
fx(x, y) = 2kx 2y and fy(x, y) = 2ky 2x.
And the second order partial derivatives are
fxx(x, y) = 2k fxy(x, y) = 2 fyy(x, y) = 2k
Since fx(0, 0) = fy(0, 0) = 0, the point (0, 0) is a critical point. The discriminant is
D = (2k)(2k) 4 = 4(k
2
1).
For k = 2, the discriminant is positive, D = 12. When k = 2, fxx(0, 0) = 4 which is positive so we have a local
minimum at the origin. When k = 2, fxx(0, 0) = 4 so we have a local maximum at the origin. In the case k = 0,
D = 4 so the origin is a saddle point.
Lastly, when k = 1 the discriminant is zero, so the second derivative test can tell us nothing. Luckily, we can factor
f(x, y) when k = 1. When k = 1,
f(x, y) = x
2
2xy +y
2
= (x y)
2
.
This is always greater than or equal to zero. So f(0, 0) = 0 is a minimum and the surface is a trough-shaped parabolic
cylinder with its base along the line x = y.
When k = 1,
f(x, y) = x
2
2xy y
2
= (x +y)
2
.
This is always less than or equal to zero. So f(0, 0) = 0 is a maximum. The surface is a parabolic cylinder, with its top
ridge along the line x = y.
1
4
8
12
16
y
x
k = 2
1
1
5
5
10
10
20
20
30
30
k = 1
x
y
k = 0
1
1
4
4
8
8
12
12
16
16
4
4
8
8
12
12
16
16
1
1
x
y
1
1
5
5
10
10
20
20
30
30
k = 1
x
y
1
4
8
12
16
y
x
k = 2
Figure 15.2
Exercises
1. Mississippi lies entirely within a region designated as 80s so we expect both the maximum and minimum daily high
temperatures within the state to be in the 80s. The southwestern-most corner of the state is close to a region designated as
90s, so we would expect the temperature here to be in the high 80s, say 87-88. The northern-most portion of the state is
located near the center of the 80s region. We might expect the high temperature there to be between 83-87.
Alabama also lies completely within a region designated as 80s so both the high and low daily high temperatures
within the state are in the 80s. The southeastern tip of the state is close to a 90s region so we would expect the temperature
here to be about 88-89 degrees. The northern-most part of the state is near the center of the 80s region so the temperature
there is 83-87 degrees.
15.2 SOLUTIONS 329
Pennsylvania is also in the 80s region, but it is touched by the boundary line between the 80s and a 70s region. Thus
we expect the low daily high temperature to occur there and be about 80 degrees. The state is also touched by a boundary
line of a 90s region so the high will occur there and be 89-90 degrees.
New York is split by a boundary between an 80s and a 70s region, so the northern portion of the state is likely to be
about 74-76 while the southern portion is likely to be in the low 80s, maybe 81-84 or so.
California contains many different zones. The northern coastal areas will probably have the daily high as low as
65-68, although without another contour on that side, it is difcult to judge how quickly the temperature is dropping off
to the west. The tip of Southern California is in a 100s region, so there we expect the daily high to be 100-101.
Arizona will have a low daily high around 85-87 in the northwest corner and a high in the 100s, perhaps 102-107 in
its southern regions.
Massachusetts will probably have a high daily high around 81-84 and a low daily high of 70.
5. To maximize z = x
2
+ y
2
, it sufces to maximize x
2
and y
2
. We can maximize both of these at the same time by
taking the point (1, 1), where z = 2. It occurs on the boundary of the square. (Note: We also have maxima at the points
(1, 1), (1, 1) and (1, 1) which are on the boundary of the square.)
To minimize z = x
2
+y
2
, we choose the point (0, 0), where z = 0. It does not occur on the boundary of the square.
9. The function f has no global maximum or global minimum.
13. Suppose x is xed. Then for large values of y the sign of f is determined by the highest power of y, namely y
3
. Thus,
f(x, y) as y
f(x, y) as y .
So f does not have a global maximum or minimum.
Problems
17. Let the sides be x, y, z cm. Then the volume is given by V = xyz = 32.
The surface area S is given by
S = 2xy + 2xz + 2yz.
Substituting z = 32/(xy) gives
S = 2xy +
64
y
+
64
x
.
At a critical point,
S
x
= 2y
64
x
2
= 0
S
y
= 2x
64
y
2
= 0,
The symmetry of the equations (or by dividing the equations) tells us that x = y and
2x
64
x
2
= 0
x
3
= 32
x = 32
1/3
= 3.17 cm.
Thus the only critical point is x = y = (32)
1/3
cm and z = 32/
_
(32)
1/3
(32)
1/3
_
= (32)
1/3
cm. At the critical point
SxxSyy (Sxy)
2
=
128
x
3

128
y
3
2
2
=
(128)
2
x
3
y
3
4.
Since D > 0 and Sxx > 0 at this critical point, the critical point x = y = z = (32)
1/3
is a local minimum. Since
S as x, y , the local minimum is a global minimum.
21. The box is shown in Figure 15.3. Cost of four sides = (2hl +2wh)(1)c/. Cost of two bottoms = (2wl)(2)c/. Thus the total
cost C (in cents) of the box is
C = 2(hl +wh) + 4wl.
But volume wlh = 512, so l = 512/(wh), thus
C =
1024
w
+ 2wh +
2048
h
.
To minimize C, nd the critical points of C by solving
C
h
= 2w
2048
h
2
= 0,
Cw = 2h
1024
w
2
= 0.
We get
2wh
2
= 2048
2hw
2
= 1024.
Since w, h = 0, we can divide the rst equation by the second giving
2wh
2
2hw
2
=
2048
1024
,
so
h
w
= 2,
thus
h = 2w.
Substituting this in C
h
= 0, we obtain h
3
= 2048, so h = 12.7 cm. Thus w = h/2 = 6.35 cm, and l = 512/(wh) =
6.35 cm. Now we check that these dimensions minimize the cost C. We nd that
D = C
hh
Cww C
2
hw
= (
4096
h
3
)(
2048
w
3
) 2
2
,
and at h = 12.7, w = 6.35, C
hh
> 0 and D = 16 4 > 0, thus C has a local minimum at h = 12.7 and w = 6.35.
Since C increases without bound as w, h 0 or , this local minimum must be a global minimum.
Therefore, the dimensions of the box that minimize the cost are w = 6.35 cm, l = 6.35 cm and h = 12.7 cm.
w
l
h
Figure 15.3
25. The total revenue is
R = pq = (60 0.04q)q = 60q 0.04q
2
,
and as q = q1 +q2, this gives
R = 60q1 + 60q2 0.04q
2
1
0.08q1q2 0.04q
2
2
.
Therefore, the prot is
P(q1, q2) = R C1 C2
= 13.7 + 60q1 + 60q2 0.07q
2
1
0.08q
2
2
0.08q1q2.
At a local maximum point, we have grad P =
0 :
P
q1
= 60 0.14q1 0.08q2 = 0,
P
q2
= 60 0.16q2 0.08q1 = 0.
15.3 SOLUTIONS 331
Solving these equations, we nd that
q1 = 300 and q2 = 225.
To see whether or not we have found a local maximum, we compute the second-order partial derivatives:
2
P
q
2
1
= 0.14,

2
P
q
2
2
= 0.16,

2
P
q1q2
= 0.08.
Therefore,
D =

2
P
q
2
1
2
P
q
2
2

2
P
q1q2
= (0.14)(0.16) (0.08)
2
= 0.016,
and so we have found a local maximum point. The graph of P(q1, q2) has the shape of an upside down paraboloid since
P is quadratic in q1 and q2, hence (300, 225) is a global maximum point.
29. (a) We have f(2, 1) = 120.
(i) If x > 20 then f(x, y) > 10x > 200 > f(2, 1).
(ii) If y > 20 then f(x, y) > 20y > 400 > f(2, 1).
(iii) If x < 0.01 and y 20 then f(x, y) > 80/(xy) > 80/((0.01)(20)) = 400 > f(2, 1).
(iv) If y < 0.01 and x 20 then f(x, y) > 80/(xy) > 80/((20)(0.01)) = 400 > f(2, 1).
(b) The continuous function f must achieve a minimum at some point (x0, y0) in the closed and bounded region R
:
0.01 x 20, 0.01 y 20. Since (2, 1) is in R
, we must have f(x0, y0) f(2, 1). By part (a), f(x0, y0)
is less than all values of f in the part of R that is outside R
, so f(x0, y0) is a minimum for f on all of R. Since

(x0, y0) is not on the boundary of R, it must be a critical point of f.
(c) The only critical point of f in R is the point (2, 1), so by part (b) f has a global minimum there.
Exercises
1. Our objective function is f(x, y) = x +y and our equation of constraint is g(x, y) = x
2
+y
2
= 1. To optimize f(x, y)
with Lagrange multipliers, we solve f(x, y) = g(x, y) subject to g(x, y) = 1. The gradients of f and g are
f(x, y) =
i +
j ,
g(x, y) = 2x
i + 2y
j .
So the equation f = g becomes
i +
j = (2x
i + 2y
j )
Solving for gives
=
1
2x
=
1
2y
,
which tells us that x = y. Going back to our equation of constraint, we use the substitution x = y to solve for y:
g(y, y) = y
2
+y
2
= 1
2y
2
= 1
y
2
=
1
2
y =
_
1
2
=
2
2
.
Since x = y, our critical points are (
2
2
,
2
2
) and (
2
2
,
2
2
). Since the constraint is closed and bounded, maximum
and minimum values of f subject to the constraint exist. Evaluating f at the critical points we nd that the maximum
value is f(
2
2
,
2
2
) =
2 and the minimum value is f(
2
2
,
2
2
) =
2.
5. Our objective function is f(x, y) = 3x 2y and our equation of constraint is g(x, y) = x
2
+ 2y
2
= 44. Their gradients
are
f(x, y) = 3
i 2
j ,
g(x, y) = 2x
i + 4y
j .
So the equation f = g becomes 3
i 2
j = (2x
i + 4y
j ). Solving for gives us

=
3
2x
=
2
4y
,
which we can use to nd x in terms of y:
3
2x
=
2
4y
4x = 12y
x = 3y.
Using this relation in our equation of constraint, we can solve for y:
x
2
+ 2y
2
= 44
(3y)
2
+ 2y
2
= 44
9y
2
+ 2y
2
= 44
11y
2
= 44
y
2
= 4
y = 2.
Thus, the critical points are (6, 2) and (6, 2). Since the constraint is closed and bounded, maximum and minimum
values of f subject to the constraint exist. Evaluating f at the critical points, we nd that the maximum is f(6, 2) =
18 + 4 = 22 and the minimum value is f(6, 2) = 18 4 = 22.
9. The objective function is f(x, y, z) = x + 3y + 5z and the equation of constraint is g(x, y, z) = x
2
+ y
2
+ z
2
= 1.
Their gradients are
f(x, y, z) =
i + 3
j + 5
k ,
g(x, y, z) = 2x
i + 2y
j + 2z
k .
So the equation f = g becomes
i + 3
j + 5
k = (2x
i + 2y
j + 2z
k ). Solving for we nd
=
1
2x
=
3
2y
=
5
2z
.
Which provides us with the equations
2y = 6x
10x = 2z.
Solving the rst equation for y gives us y = 3x. Solving the second equation for z gives us z = 5x. Substituting these
into the equation of constraint, we can nd x:
x
2
+ (3x)
2
+ (5x)
2
= 1
x
2
+ 9x
2
+ 25x
2
= 1
35x
2
= 1
x
2
=
1
35
x =
_
1
35
=
35
35
.
Since y = 3x and z = 5x, the critical points are at (
35
35
, 3
35
35
,
35
7
). Since the constraint is closed and bounded,
maximum and minimum values of f subject to the constraint exist. Evaluating f at the critical points, we nd the max-
imum is f(
35
35
, 3
35
35
,
35
7
) =
35
35
35
=
35, and the minimum value is f(
35
35
, 3
35
35
,
35
7
) =
35.
15.3 SOLUTIONS 333
13. The region x
2
+y
2
2 is the shaded disk of radius
2 centered at the origin (including the circle x

2
+y
2
= 2) as shown
in Figure 15.4.
2
1
3
5
1
3
5
x
y
Figure 15.4
We rst nd the local maxima and minima of f in the interior of our disk. So we need to nd the extrema of
f(x, y) = x + 3y, in the region x
2
+y
2
< 2.
As
fx = 1
fy = 3
f does not have critical points. Now lets nd the local extrema of f on the boundary of the disk. We want to nd the
extrema of f(x, y) = x + 3y subject to the constraint g(x, y) = x
2
+y
2
2 = 0. We use Lagrange multipliers
grad f = grad g and x
2
+y
2
= 2,
which give
1 = 2x
3 = 2y
x
2
+y
2
= 2.
As cannot be zero, we solve for x and y in the rst two equations and get x =
1
2
and y =
3
2
. Plugging into the third
equation gives
8
2
= 10
so =
5
2
and we get the solutions (
1
5
,
3
5
) and (
1
5
,
3
5
). Evaluating f at these points gives
f(
1
5
,
3
5
) = 2
5 and
f(
1
5
,
3
5
) = 2
5
The region x
2
+ y
2
2 is closed and bounded, so maximum and minimum values of f in the region exist. Therefore
(
1
5
,
3
5
) is a global maximum of f and (
1
5
,
3
5
) is a global minimum of f on the whole region x
2
+ y
2
2.
17. We rst nd the critical points of f:
fx = 2xy = 0, fy = x
2
+ 6y 1 = 0.
From the rst equation, we get either x = 0 or y = 0. If x = 0, from the second equation we get 6y 1 = 0 so y = 1/6.
If instead y = 0, then from the second equation x = 1. We conclude that the critical points are (0, 1/6), (1, 0), and
(1, 0). All three critical points satisfy the constraint x
2
+y
2
10.
The Lagrange conditions, grad f = grad g, are:
2xy = 2x, x
2
+ 6y 1 = 2y
From the rst equation, when x = 0, we divide by x to get = y. Substituting into the second equation, we get
x
2
+ 6y 1 = 2y
2
.
Then using x
2
= 10 y
2
from the constraint, we have
10 y
2
+ 6y 1 = 2y
2
,
so 3y
2
6y 9 = 0. Factoring, we get 3(y 3)(y + 1) = 0. From the constraint, we get x = 1 when y = 3
and x = 3 for y = 1. If instead x = 0, so that we cannot divide by x in the rst Lagrange equation, then from the
constraint, y =
10. Summarizing, the following points are either critical points or satisfy the Lagrange conditions:
(1, 0), (1, 0), (0, 1/6), (1, 3), (3, 1), (0,
10).
These are the candidates for global maximum or minimum points. The corresponding values for f(x, y) = x
2
y +3y
2
y
are:
0, 0, 1/12, 27, 5, 30
10.
The largest value is 30 +
10 at the point (0,
10) and the smallest value is 5 at (3, 1).

Problems
21. (a) C. The vectors v and w may, or may not, be parallel.
(b) F. Since v w = 2 = ||v || || w || cos , we must have cos > 0, so 0 < /2. Thus, the statement /2 < <
is false.
(c) T. Since ||v || || w || cos = 2 and cos 1, it is true that either ||v || or || w || or both must be greater than 1.
(d) F. Since P is the hottest point on R, we know grad g = v is perpendicular to R at P.
(e) F. Since v is perpendicular to R and w is not perpendicular to v (because v w = 0), we know that w is not
tangent to R at P.
25. The maximum and minimum values change by approximately c. The Lagrange conditions give:
3 = 2x, 2 = 4y.
Solving for and setting the expressions equal, we get x = 3y. Substituting into the constraint, we get y = 2, so the
points satisfying the Lagrange conditions are (6, 2) and (6, 2). The corresponding values of f(x, y) = 3x 2y are
22 and 22. From the rst equation, we have = 3/(2x). Thus the minimum value changes by 3/(12)c = c/4
and the maximum changes by 3/(12)c = c/4.
29. (a) We want to minimize C subject to g = x +y = 39. Solving C = g gives
10x + 2y =
2x + 6y =
so y = 2x. Solving with x +y = 39 gives x = 13, y = 26, = 182. Therefore C = $4349.
(b) Since = 182, increasing production by 1 will cause costs to increase by approximately $182. (because =
C
g
= rate of change of C with g). Similarly, decreasing production by 1 will save approximately $182.
33. (a) Let g(x, y) = x +y. We are minimizing f(x, y) = x
2
+ 2y
2
subject to the constraint g(x, y) = c.
The method of Lagrange multipliers is to solve the equations
fx = gx fy = gy g = c,
which are
2x = 4y = x +y = c.
We have
x =
2c
3
y =
c
3
=
4c
3
,
15.3 SOLUTIONS 335
so there is a critical point at (2c/3, c/3). Since moving away from the origin increases values of f in Figure 15.5, we
see that f has a minimum on the constraint. The minimum value is
m(c) = f
_
2c
3
,
c
3
_
=
2c
2
3
.
Constraint
x + y = c
(
2c
3
,
c
3
)
x
y
Figure 15.5
(b) Calculations in part (a) showed that = 4c/3.
(c) The multiplier is the rate of change of m(c) as c increases and the constraint moves: that is, = m
(c).
37. (a) The objective function is the complementary energy,
f
2
1
2k1
+
f
2
2
2k2
, and the constraint is f1 +f2 = mg. The Lagrangian
function is
L(f1, f2, ) =
f
2
1
2k1
+
f
2
2
2k2
(f1 +f2 mg).
We look for solutions to the system of equations we get from grad L =

0 :
L
f1
=
f1
k1
= 0
L
f2
=
f2
k2
= 0
L
= (f1 +f2 mg) = 0.

Combining
L
f1
L
f2
=
f1
k1
f2
k2
= 0 with
L
= 0 gives the two equation system

f1
k1
f2
k2
= 0
f1 +f2 = mg.
Substituting f2 = mg f1 into the rst equation leads to
f1 =
k1
k1 +k2
mg
f2 =
k2
k1 +k2
mg.
(b) Hookes Law states that for a spring
Force of spring = Spring constant Distance stretched or compressed from equilibrium.
Since f1 = k1 and f2 = k2 , the Lagrange multiplier equals the distance the mass stretches the top spring
and compresses the lower spring.
41. (a) The objective function f(x, y) = px +qy gives the cost to buy x units of input 1 at unit price p and y units of input
2 at unit price q.
The constraint g(x, y) = u tells us that we are only considering the cost of inputs x and y that can be used to
produce quantity u of the product.
Thus the number C(p, q, u) gives the minimum cost to the company of producing quantity u if the inputs it
needs have unit prices p and q.
(b) The Lagrangian function is
L(x, y, ) = px +qy (xy u).
We look for solutions to the system of equations we get from grad L =

0 :
L
x
= p y = 0
L
y
= q x = 0
L
= (xy u) = 0.
We see that = p/y = q/x so y = px/q. Substituting for y in the constraint xy = u leads to x =
_
qu/p,
y =
_
pu/q and =
_
pq/u. The minimum cost is thus
C(p, q, u) = p
_
qu
p
+q
_
pu
q
= 2
pqu.
45. (a) If the prices are p1 and p2 and the budget is b, the quantities consumed are constrained by
p1x1 +p2x2 = b.
We want to maximize
u(x1, x2) = a ln x1 + (1 a) ln x2
subject to the constraint
p1x1 +p2x2 = b.
Using Lagrange multipliers, we solve
u
x1
=
a
x1
= p1
u
x2
=
1 a
x2
= p2,
giving x1 = a/(p1) and x2 = (1 a)/(p2). Substituting into the constraint, we get
a
+
(1 a)
= b
so
=
1
b
.
Thus
x1 =
ab
p1
x2 =
(1 a)b
p2
so the maximum satisfaction is given by
S = u(x1, x2) = u
_
ab
p1
,
(1 a)b
p2
_
= a ln
_
ab
p1
_
+ (1 a) ln
_
(1 a)b
p2
_
= a ln a +a ln b a ln p1 + (1 a) ln(1 a) + (1 a) ln b (1 a) ln p2
= a ln a + (1 a) ln(1 a) + ln b a ln p1 (1 a) ln p2.
(b) We want to calculate the value of b needed to achieve u(x1, x2) = c. Thus, we solve for b in the equation
c = a ln a + (1 a) ln(1 a) + ln b a ln p1 (1 a) ln p2.
SOLUTIONS to Review Problems for Chapter Fifteen 337
Since
ln b = c a ln a (1 a) ln(1 a) +a ln p1 + (1 a) ln p2,
we have
b =
e
c
e
a ln p
1
e
(1a) ln p
2
e
a ln a
e
(1a) ln(1a)
=
e
c
p
a
1
p
1a
2
a
a
(1 a)
(1a)
.
Exercises
1. The critical points of f are obtained by solving fx = fy = 0, that is
fx(x, y) = 2y
2
2x = 0 and fy(x, y) = 4xy 4y = 0,
so
2(y
2
x) = 0 and 4y(x 1) = 0
The second equation gives either y = 0 or x = 1. If y = 0 then x = 0 by the rst equation, so (0, 0) is a critical point. If
x = 1 then y
2
= 1 from which y = 1 or y = 1, so two further critical points are (1, 1), and (1, 1).
Since
D = fxxfyy (fxy)
2
= (2)(4x 4) (4y)
2
= 8 8x 16y
2
,
we have
D(0, 0) = 8 > 0, D(1, 1) = D(1, 1) = 16 < 0,
and fxx = 2 < 0. Thus, (0, 0) is a local maximum; (1, 1) and (1, 1) are saddle points.
fx = cos x + cos (x +y).
fy = cos y + cos (x +y).
Setting fx = 0 and fy = 0 gives
cos x = cos y
For 0 < x < and 0 < y < , cos x = cos y only if x = y. Then, setting fx = fy = 0:
cos x + cos 2x = 0,
cos x + 2 cos
2
x 1 = 0,
(2 cos x 1)(cos x + 1) = 0.
So cos x = 1/2 or cos x = 1, that is x = /3 or x = . For the given domain 0 < x < , 0 < y < , we only
consider the solution when x = /3 then y = x = /3. Therefore, the critical point is (
3
,

3
).
Since
fxx(x, y) = sin x sin (x +y) fxx(
3
,

3
) = sin

3
sin
2
3
=
3
fxy(x, y) = sin (x +y) fxy(
3
,

3
) = sin
2
3
=
3
2
fyy(x, y) = sin y sin (x +y) fyy(
3
,

3
) = sin

3
sin
2
3
=
3
the discriminant is
D(x, y) = fxxfyy f
2
xy
= (
3)(
3) (
3
2
)
2
=
9
4
> 0.
Since fxx(
3
,

3
) =
3 < 0, (
3
,

3
) is a local maximum.
fx = y +
1
x
, fy = x + 2y.
For critical points, solve fx = 0 and fy = 0 simultaneously. From fy = x + 2y = 0 we get that x = 2y. Substituting
into fx = 0, we have
y +
1
x
= y
1
2y
=
1
y
(y
2
1
2
) = 0
Since
1
y
= 0, y
2
1
2
= 0, therefore
y =
1
2
=
2
2
,
and x =
2. So the critical points are

_
2,
2
2
_
and
_
2,
2
2
_
. But x must be greater than 0, so
_
2,
2
2
_
is
not in the domain.
The contour diagram for f in Figure 15.6 (drawn by computer), shows that
_
2,
2
2
_
is a saddle point of f(x, y).
4 5 6 7 8
4
3
2
1
1
2
3
4
x
y
0
c
10
T
10
E
f(
2,
2
2
) =
21
2
+ ln
2
20
0
10
20
30
Figure 15.6: Contour map of f(x, y) = xy + ln x +y
2
10
We can also see that
_
2,
2
2
_
is a saddle point analytically.
Since fxx =
1
x
2
, fyy = 2, fxy = 1, the discriminant is:
D(x, y) = fxxfyy f
2
xy
=
2
x
2
1.
D
_
2,
2
2
_
= 2 < 0, so
_
2,
2
2
_
is a saddle point.
13. The objective function is f(x, y) = x
2
xy +y
2
and the equation of constraint is g(x, y) = x
2
y
2
= 1. The gradients
of f and g are
f(x, y) = (2x y)
i + (x + 2y)
j ,
g(x, y) = 2x
i 2y
j .
Therefore the equation f(x, y) = g(x, y) gives
2x y = 2x
x + 2y = 2y
x
2
y
2
= 1.
Let us suppose that = 0. Then 2x = y and 2y = x give x = y = 0. But (0, 0) is not a solution of the third equation,
so we conclude that = 0. Now lets multiply the rst two equations
2y(2x y) = 2x(x + 2y).
As = 0, we can cancel it in the equation above and after doing the algebra we get
x
2
4xy +y
2
= 0
which gives x = (2 +
3)y or x = (2
3)y.
If x = (2 +
3)y, the third equation gives

(2 +
3)
2
y
2
y
2
= 1
so y 0.278 and x 1.038. These give the critical points (1.038, 0.278), (1.038, 0.278).
If x = (2
3)y, from the third equation we get

(2
3)
2
y
2
y
2
= 1.
But (2
3)
2
1 0.928 < 0 so the equation has no solution. Evaluating f gives
f(1.038, 0.278) = f(1.038, 0.278) 0.866
Since y on the constraint, rewriting f as
f(x, y) =
_
x
y
2
_
2
+
3
4
y
2
shows that f has no maximum on the constraint. The minimum value of f is 0.866. See Figure 15.7.
2
2
(1.038, 0.278)
(1.038, 0.278)
0.2
0.866
2
x
x
2
y
2
= 1
y
Figure 15.7
17. Our objective function is f(x, y, z) = x
2
2y +2z
2
and our equation of constraint is g(x, y, z) = x
2
+y
2
+z
2
1 = 0.
To optimize f(x, y, z) with Lagrange multipliers, we solve f(x, y, z) = g(x, y, z) subject to g(x, y, z) = 0. The
gradients of f and g are
f(x, y, z) = 2x
i 2
j + 4z
k ,
g(x, y) = 2x
i + 2y
j + 2z
k .
We get,
x = x
1 = y
2z = z
x
2
+y
2
+z
2
= 1.
From the rst equation we get x = 0 or = 1.
If x = 0 we have
1 = y
2z = z
y
2
+z
2
= 1.
From the second equation z = 0 or = 2. So if z = 0, we have y = 1 and we get the solutions (0, 1, 0),(0, 1, 0). If
z = 0 then = 2 and y =
1
2
. So z
2
=
3
4
which gives the solutions (0,
1
2
,
3
2
), (0,
1
2
,
3
2
).
If x = 0, then = 1, so y = 1, which implies, from the equation x
2
+y
2
+z
2
= 1, that x = 0, which contradicts
the assumption.
Since the constraint is closed and bounded, maximum and minimum values of f subject to the constraint exist. There-
fore, evaluating f at the critical points, we get f(0, 1, 0) = 2, f(0, 1, 0) = 2 and f(0,
1
2
,
3
2
) = f(0,
1
2
,
3
2
) =
4. So the maximum value of f is 4 and the minimum is 2.
21. The region x +y 1 is the shaded half plane (including the line x +y = 1) shown in Figure 15.8.
1 2 3
1
2
3
1
3
5
1
3
5
x
y
Figure 15.8
Lets look for the critical points of f in the interior of the region. As
fx = 3x
2
fy = 1
there are no critical points inside the shaded region. Now lets nd the extrema of f on the boundary of our region. We
want the extrema of f(x, y) = x
3
+y subject to the constraint g(x, y) = x +y 1 = 0. We use Lagrange multipliers
grad f = grad g and x +y = 1,
which give
3x
2
=
1 =
x +y = 1.
From the rst two equations we get 3x
2
= 1, so the solutions are
(
1
3
, 1
1
3
) and (
1
3
, 1 +
1
3
).
Evaluating f at these points we get
f(
1
3
, 1
1
3
) = 1
2
3
3
f(
1
3
, 1 +
1
3
) = 1 +
2
3
3
.
From the contour diagram in Figure 15.8, we see that (
1
3
, 1
1
3
) is a local minimum and (
1
3
, 1 +
1
3
) is a local
maximum of f on x +y = 1. Are they global extrema as well?
If we take x very big and y = 1 x then f(x, y) = x
3
+y = x
3
x + 1 which can be made as big as we want (if
we choose x big enough). So there will be no global maximum.
Similarly, taking x negative with big absolute value and y = 1 x, f(x, y) = x
3
+y = x
3
x +1 can be made as
small as we want (if we choose x small enough). So there is no global minimum. This can also be seen from Figure 15.8.
25. If xy = 10, then f(x, y) = x
2
y
2
= x
2
100/x
2
and x can take any nonzero value. Since
lim
x
_
x
2
100
x
2
_
= ,
we see f has no maximum on the constraint. Since
lim
x0
_
x
2
100
x
2
_
= ,
we see f has no minimum on the constraint.
Problems
29. (a) (i) Suppose N = kA
p
. Then the rule of thumb tells us that if A is multiplied by 10, the value of N doubles. Thus
2N = k(10A)
p
= k10
p
A
p
.
Thus, dividing by N = kA
p
, we have
2 = 10
p
so taking logs to base 10 we have
p = log 2 = 0.3010.
(where log 2 means log
10
2). Thus,
N = kA
0.3010
.
(ii) Taking natural logs gives
ln N = ln(kA
p
)
ln N = ln k +p ln A
ln N ln k + 0.301 ln A
Thus, ln N is a linear function of ln A.
(b) Table 15.1 contains the natural logarithms of the data:
Table 15.1 ln N and ln A
Island ln A ln N
Redonda 1.1 1.6
Saba 3.0 2.2
Montserrat 2.3 2.7
Puerto Rico 9.1 4.3
Jamaica 9.3 4.2
Hispaniola 11.2 4.8
Cuba 11.6 4.8
Using a least squares t we nd the line:
ln N = 1.20 + 0.32 ln A
This yields the power function:
N = e
1.20
A
0.32
= 3.32A
0.32
Since 0.32 is pretty close to log 2 0.301, the answer does agree with the biological rule.
33. (a) The problem is to maximize
V = 1000D
0.6
N
0.3
subject to the budget constraint in dollars
40000D + 10000N 600000
or (in thousand dollars)
40D + 10N 600
(b) Let B = 40D + 10N = 600 (thousand dollars) be the budget constraint. At the optimum
V = B,
so
V
D
=
B
D
= 40
V
N
=
B
N
= 10.
Thus
V/D
V/N
= 4.
Therefore, at the optimum point, the rate of increase in the number of visits with respect to an increase in the number
of doctors is four times the corresponding rate for nurses. This factor of four is the same as the ratio of the salaries.
(c) Differentiating and setting V = B yields
600D
0.4
N
0.3
= 40
300D
0.6
N
0.7
= 10
Thus, we get
600D
0.4
N
0.3
40
= =
300D
0.6
N
0.7
10
So
N = 2D.
To solve for D and N, substitute in the budget constraint:
600 40D 10N = 0
600 40D 10 (2D) = 0
So D = 10 and N = 20.
=
600(10
0.4
)(20
0.3
)
40
14.67
Thus the clinic should hire 10 doctors and 20 nurses. With that staff, the clinic can provide
V = 1000(10
0.6
)(20
0.3
) 9,779 visits per year.
(d) From part c), the Lagrange multiplier is = 14.67. At the optimum, the Lagrange multiplier tells us that about 14.67
extra visits can be generated through an increase of $1,000 in the budget. (If we had written out the constraint in
dollars instead of thousands of dollars, the Lagrange multiplier would tell us the number of extra visits per dollar.)
(e) The marginal cost, MC, is the cost of an additional visit. Thus, at the optimum point, we need the reciprocal of the
Lagrange multiplier:
MC =
1

1
14.67
0.068 (thousand dollars),
that is, at the optimum point, an extra visit costs the clinic 0.068 thousand dollars, or $68.
This production function exhibits declining returns to scale (e.g. doubling both inputs less than doubles output,
because the two exponents add up to less than one). This means that for large V , increasing V will require increasing
D and N by more than when V is small. Thus the cost of an additional visit is greater for large V than for small. In
other words, the marginal cost will rise with the number of visits.
37. The objective function is
f(x, y, z) =
_
(x a)
2
+ (y b)
2
+ (z c)
2
,
and the constraint is
g(x, y, z) = Ax +By +Cz +D = 0.
Partial derivatives of f and g are
fx =
1
2
2 (x a)
f(x, y, z)
=
x a
f(x, y, z)
,
fy =
1
2
2 (y b)
f(x, y, z)
=
y b
f(x, y, z)
,
fz =
1
2
2 (z c)
f(x, y, z)
=
z c
f(x, y, z)
,
gx = A, gy = B, and gz = C.
Using Lagrange multipliers, we need to solve the equations
grad f = grad g
where grad f = fx
i +fy
j +fz
k and grad g = gx
i +gy
j +gz
k . This gives a system of equations:

x a
f(x, y, z)
= A
y b
f(x, y, z)
= B
z c
f(x, y, z)
= C
Ax +By +Cz +D = 0.
Now
xa
A
=
yb
B
=
zc
C
= f(x, y, z) gives
x =
A
B
(y b) +a,
z =
C
B
(y b) +c,
Substitute into the constraint,
A
_
A
B
(y b) +a
_
+By +C
_
C
B
(y b) +c
_
+D = 0,
_
A
2
B
+B +
C
2
B
_
y =
A
2
B
b Aa +
C
2
B
b Cc D.
Hence
y =
(A
2
+C
2
)b B(Aa +Cc +D)
A
2
+B
2
+C
2
,
y b =
B(Aa +Bb +Cc +D)
A
2
+B
2
+C
2
x a =
A
B
(y b)
=
A(Aa +Bb +Cc +D)
A
2
+B
2
+C
2
z c =
C
B
(y b)
=
C(Aa +Bb +Cc +D)
A
2
+B
2
+C
2
Thus the minimum f(x, y, z) is
f(x, y, z) =
_
(x a)
2
+ (y b)
2
+ (z c)
2
=
_
_
A(Aa +Bb +Cc +D)
A
2
+B
2
+C
2
_
2
+
_
B(Aa +Bb +Cc +D)
A
2
+B
2
+C
2
_
2
+
_
C(Aa +Bb +Cc +D)
A
2
+B
2
+C
2
_
2
1/2
=
|Aa +Bb +Cc +D|
A
2
+B
2
+C
2
.
The geometric meaning is nding the shortest distance from a point (a, b, c) to the plane Ax +By +Cz +D = 0.
41. Cost of production, C, is given by C = p1W +p2K = b. At the optimal point, q = C.
Since q =
_
c(1 a)W
a
K
a
_
i +
_
caW
1a
K
a1
_
j and C = p1
i +p2
j , we get
c(1 a)W
a
K
a
= p1 and caW
1a
K
a1
= p2.
Now, marginal productivity of labor is given by
q
W
= c(1 a)W
a
K
a
and marginal productivity of capital is given by
q
K
= caW
1a
K
a1
, so their ratio is given by
q
W
q
K
=
c(1 a)W
a
K
a
caW
1a
K
a1
=
p1
p2
=
p1
p2
which is the ratio of the cost of one unit of labor to the cost of one unit of capital.
45. You should try to anticipate your opponents choice. After you choose a value , your opponent will use calculus to
nd the point (x, y) that maximizes the function f(x, y) = 10 x
2
y
2
2x (2x + 2y). At that point, we have
fx = 2x 2 2 = 0 and fy = 2y 2 = 0, so your opponent will choose x = 1 and y = . This gives a
value L(1, , ) = 10(1)
2
()
2
2(1)(2(1)+2()) = 11+2+2
2
which you
want to make as small as possible. You should choose to minimize the function h() = 11 + 2 + 2
2
. You choose
so that h
() = 2 + 4 = 0, or = 1/2. Your opponent then chooses (x, y) = (1 , ) = (1/2, 1/2), giving

a nal score of L(1/2, 1/2, 1/2) = 10.5. No choice of that you can make can force the value of L below 10.5. But
your choice of = 1/2 makes it impossible for your opponent to force the value of L above 10.5.
49. (a) We have grad f = 3
i + 2
j and grad g = (4x 4y)
i + (4x + 10y)
j , so the Lagrange multiplier equations are

3 = (4x 4y)
2 = (4x + 10y)
2x
2
4xy + 5y
2
= 20
Solving these with a CAS we get = 0.4005, x = 3.9532, y = 2.0806 and = 0.4005, x = 3.9532, y =
2.0806. We have f(3.9532, 2.0806) = 11.0208, and f(3, 9532, 2.0806) = 21.0208. The constraint equation
is 2x
2
4xy +5y
2
= 20, or, completing the square, 2(x y)
2
+3y
2
= 20. This has the shape of a skewed ellipse,
so the constraint curve is bounded, and therefore the local maximum is a global maximum. Thus the maximum value
is 21.0208.
(b) The maximum value on g = 20.5 is 21.0208 + 0.5(0.4005) = 21.2211. The maximum value on g = 20.2 is
21.0208 + 0.2(0.4005) = 21.1008.
(c) We use the same commands in the CAS from part (a), with 20 replaced by 20.5 and 20.2, and get the maximum
values 21.2198 for g = 20.5 and 21.1007 for g = 20.2. These agree with the approximations we found in part (b) to
2 decimal places.
1. True. By denition, a critical point is either where the gradient of f is zero or does not exist.
5. True. The graph of this function is a cone that opens upward with its vertex at the origin.
9. False. For example, the linear function f(x, y) = x +y has no local extrema at all.
13. False. For example, the linear function f(x, y) = x + y has neither a global minimum or global maximum on all of
2-space.
17. False. On the given region the function f is always less than one. By picking points closer and closer to the circle
x
2
+ y
2
= 1 we can make f larger and larger (although never larger than one). There is no point in the open disk that
gives f its largest value.
21. True. The point (a, b) must lie on the constraint g(x, y) = c, so g(a, b) = c.
25. False. Since grad f and grad g point in opposite directions, they are parallel. Therefore (a, b) could be a local maximum
or local minimum of f constrained to g = c. However the information given is not enough to determine that it is a
minimum. If the contours of g near (a, b) increase in the opposite direction as the contours of f, then at a point with
grad f(a, b) = grad g(a, b) we have 0, but this can be a local maximum or minimum.
For example, f(x, y) = 4 x
2
y
2
has a local maximum at (1, 1) on the constraint g(x, y) = x + y = 2. Yet
at this point, grad f = 2
i 2
j and grad g =

i +
j , so grad f and grad g point in opposite directions.

29. True. Since f(a, b) = M, we must satisfy the Lagrange conditions that fx(a, b) = g(a, b) and fy(a, b) = gy(a, b),
for some . Thus fx(a, b)/fy(a, b) = gx(a, b)/gy(a, b).
33. False. The value of at a minimum point gives the proportional change in m for a change in c. If > 0 and the change
in c is positive, the change in m will also be positive.
16.1 SOLUTIONS 345
CHAPTER SIXTEEN
Exercises
1. Mark the values of the function on the plane, as shown in Figure 16.1, so that you can guess respectively at the smallest
and largest values the function takes on each small rectangle.
Lower sum =
f(xi, yi)xy
= 4xy + 6xy + 3xy + 4xy
= 17xy
= 17(0.1)(0.2) = 0.34.
Upper sum =
f(xi, yi)xy
= 7xy + 10xy + 6xy + 8xy
= 31xy
= 31(0.1)(0.2) = 0.62.
1.0 1.1 1.2
2.0
2.2
2.4
x
y
5
4
3
7
6
5
10
8
4
E '
x = 0.1
T
c
y = 0.2
Figure 16.1
(0, 0)
(0, 4)
(4, 0)
(4, 4)
Figure 16.2
5. (a) If we take the partition of R consisting of just R itself, we get
Lower bound for integral = minRf AR = 0 (4 0)(4 0) = 0.
Similarly, we get
Upper bound for integral = maxRf AR = 4 (4 0)(4 0) = 64.
(b) The estimates asked for are just the upper and lower sums. We partition R into subrectangles R
(a,b)
of width 2 and
height 2, where (a, b) is the lower-left corner of R
(a,b)
. The subrectangles are then R
(0,0)
, R
(2,0)
, R
(0,2)
, and R
(2,2)
,
as in Figure 16.2. Then we nd the lower sum
Lower sum =
(a,b)
AR
(a,b)
min
R
(a,b)
f =
(a,b)
4 (Min of f on R
(a,b)
)
= 4
(a,b)
(Min of f on R
(a,b)
)
= 4(f(0, 0) + f(2, 0) + f(0, 2) + f(2, 2))
= 4(
0 0 +
2 0 +
0 2 +
2 2)
= 8.
346 Chapter Sixteen /SOLUTIONS
Similarly, the upper sum is
Upper sum = 4
(a,b)
(Max of f on R
(a,b)
)
= 4(f(2, 2) + f(4, 2) + f(2, 4) + f(4, 4))
= 4(
2 2 +
4 2 +
2 4 +
4 4)
= 24 + 16
2 46.63.
The upper sum is an overestimate and the lower sum is an underestimate, so we can get a better estimate by averaging
them to get 16 + 8
2 27.3.
Problems
9. The function being integrated is f(x, y) = 1, which is positive everywhere. Thus, its integral over any region is positive.
13. The region D is symmetric both with respect to x and y axes. The function being integrated is f(x, y) = 5x, which is an
odd function in x. Since D is symmetric with respect to x, the contributions to the integral cancel out. Thus, the integral
of the function over the region D is zero.
17. The function being integrated, f(x, y) = y y
3
is always negative in the region B since in that region 1 < y < 0 and
|y
3
| < |y|. Thus, the integral is negative.
21. The function f(x, y) = e
x
is positive for any value of x. Thus, its integral is always positive for any region, such as D,
with nonzero area.
25. The question is asking which graph has more volume under it, and from inspection, it appears that it would be the graph
for the mosquitos.
29. Let R be the region 0 x 60, 0 y 8. Then
Volume =
_
R
w(x, y) dA
Lower estimate: 102(1+4+8+10+10+8+0+3+4+6+6+4+0+1+2+3+3+2+0+0+1+1+1+1) = 1580.
Upper estimate:
10 2(8+13+16+17+17+16+4+8+10+11+11+10+3+4+6+7+7+6+1+2+3+4+4+3) = 3820.
The average of the two estimates is 2700 cubic feet.
Exercises
1. See Figure 16.3.
x
y
Figure 16.3
16.2 SOLUTIONS 347
5. We evaluate the inside integral rst:
_
4
0
(4x + 3y) dx = (2x
2
+ 3yx)
4
0
= 32 + 12y.
Therefore, we have
_
3
0
_
4
0
(4x + 3y) dxdy =
_
3
0
(32 + 12y) dy = (32y + 6y
2
)
3
0
= 150.
9. Calculating the inner integral rst, we have
_
1
0
_
1
0
ye
xy
dxdy =
_
1
0
_
e
xy
1
0
_
dy =
_
1
0
_
e
y
e
0
_
dy =
_
1
0
(e
y
1) dy = (e
y
y)
1
0
= e
1
1(e
0
0) = e2.
13. This region lies between x = 0 and x = 4 and between the lines y = 3x and y = 12, and so the iterated integral is
_
4
0
_
12
3x
f(x, y) dydx.
Alternatively, we could have set up the integral as follows:
_
12
0
_
y/3
0
f(x, y) dxdy.
17. The line connecting (1, 0) and (4, 1) is
y =
1
3
(x 1)
So the integral is
_
4
1
_
2
(x1)/3
f dy dx
21.
_
4
1
_
y
y
x
2
y
3
dxdy =
_
4
1
y
3
x
3
3
y
dy
=
1
3
_
4
1
(y
6
y
9
2
) dy
=
1
3
_
y
7
7

y
11/2
11/2
_
4
1
=
1
3
__
4
7
7

4
11/2
2
11
_
_
1
7

2
11
_
_
656.082
See Figure 16.4.
1 2 4
1
4
x
y
y = 4
y = x
x =
y
Figure 16.4
25. In the other order, the integral is
_
1
0
_
2
0
x + y dy dx.
First we keep x xed and calculate the inside integral with respect to y:
_
2
0
x + y dy =
2
3
(x + y)
3/2
y=2
y=0
=
2
3
_
(x + 2)
3/2
x
3/2
.
Then the outside integral becomes
_
1
0
2
3
_
(x + 2)
3/2
x
3/2
dx =
2
3
_
2
5
(x + 2)
5/2
2
5
x
5/2
_
1
0
=
2
3

2
5
_
3
5/2
1 2
5/2
= 2.38176
Note that the answer is the same as the one we got in Exercise 24.
Problems
29. The region is bounded by x = 1, x = 4, y = 2, and y = 2x. Thus
Volume =
_
4
1
_
2x
2
(6x
2
y) dydx.
To evaluate this integral, we evaluate the inside integral rst:
_
2x
2
(6x
2
y) dy = (3x
2
y
2
)
2x
2
= 3x
2
(2x)
2
3x
2
(2
2
) = 12x
4
12x
2
.
Therefore, we have
_
4
1
_
2x
2
(6x
2
y) dydx =
_
4
1
(12x
4
12x
2
) dx =
_
12
5
x
5
4x
3
_
4
1
= 2203.2.
The volume of this object is 2203.2.
33. (a) The line x = y/2 is the line y = 2x, and y = x and y = 2x intersect at x = 0. Thus, R is the shaded region in
Figure 16.5. One expression for the integral is
_
R
f dA =
_
3
0
_
2x
x
x
2
e
x
2
dy dx.
Another expression is obtained by reversing the order of integration. When we do this, it is necessary to split R into
two regions on a line parallel to the x-axis along the point of intersection of y = x and x = 3; this line is y = 3.
Then we obtain
_
R
f dA =
_
3
0
_
y
y/2
x
2
e
x
2
dxdy +
_
6
3
_
3
y/2
x
2
e
x
2
dxdy.
(b) We evaluate the rst integral. Integrating with respect to y rst:
_
3
0
_
2x
x
x
2
e
x
2
dy dx =
_
3
0
(x
2
e
x
2
y)
2x
x
dx =
_
3
0
x
3
e
x
2
dx.
We use integration by parts with u = x
2
, v
= xe
x
2
. Then u
= 2x and v =
1
2
e
x
2
, so
_
3
0
x
3
e
x
2
dx =
1
2
x
2
e
x
2
3
0
_
3
0
xe
x
2
dx =
1
2
x
2
e
x
2
3
0
1
2
e
x
2
3
0
=
_
1
2
(9)e
9
0
_
_
1
2
e
9
1
2
_
=
1
2
+ 4e
9
.
16.2 SOLUTIONS 349
1 2 3
1
2
3
4
5
6
7
(3, 3)
(3, 6)
y = 2x
y = x
x
y
Figure 16.5
9
x
y
3
x = 9
x = y
2
Figure 16.6
37. As given, the region of integration is as shown in Figure 16.6.
Reversing the limits gives
_
9
0
_

x
0
y sin (x
2
) dydx =
_
9
0
_
y
2
sin (x
2
)
2
x
0
_
dx
=
1
2
_
9
0
xsin (x
2
) dx
=
cos (x
2
)
4
9
0
=
1
4

cos (81)
4
= 0.056.
41. (a) The contour f(x, y) = 1 lies in the xy-plane and has equation
2e
(x1)
2
y
2
= 1,
so
(x 1)
2
y
2
= ln(1/2)
(x 1)
2
+ y
2
= ln 2 = 0.69.
This is the equation of a circle centered at (1, 0) in the xy-plane.
Other contours are of the form
2e
(x1)
2
y
2
= c
(x 1)
2
y
2
= ln(c/2).
Thus, all the contours are circles centered at the point (1, 0).
(b) The cross-section has equation z = f(1, y) = e
y
2
. If x = 1, the base region in the xy-plane extends from
y =
3 to y =
3. See Figure 16.7, which shows the circular region below W in the xy-plane. So
Area =
_

3
3
e
y
2
dy.
(c) Slicing parallel to the y-axis, we get
Volume =
_
2
2
_

4x
2
4x
2
e
(x1)
2
y
2
dy dx.
2 2
2
2
(1,
3) y =
4 x
2
y =
4 x
2
(1,
3)
x
y
Figure 16.7: Region beneath W in the
xy-plane
45.
Volume =
_
2
0
_
2
0
xy dy dx =
_
2
0
1
2
xy
2
2
0
dx
=
_
2
0
2xdx
= x
2
2
0
= 4
49. Let R be the triangle with vertices (1, 0), (2, 2) and (0, 1). Note that (3x + 2y + 1) (x + y) = 2x + y + 1 > 0 for
x, y > 0, so z = 3x + 2y + 1 is above z = x + y on R. We want to nd
Volume =
_
R
((3x + 2y + 1) (x + y)) dA =
_
R
(2x + y + 1) dA.
We need to express this in terms of double integrals.
1 2
1
2
x
y
R
1
R
2
y = 1 + 0.5x
y = 1 x
y = 2x 2
(2, 2)
O
Figure 16.8
16.2 SOLUTIONS 351
To do this, divide R into two regions with the line x = 1 to make regions R1 for x 1 and R2 for x 1. See
Figure 16.8. We want to nd
_
R
(2x + y + 1) dA =
_
R
1
(2x + y + 1) dA +
_
R
2
(2x + y + 1) dA.
Note that the line connecting (0, 1) and (1, 0) is y = 1 x, and the line connecting (0, 1) and (2, 2) is y = 1 +0.5x. So
_
R
1
(2x + y + 1) dA =
_
1
0
_
1+0.5x
1x
(2x + y + 1) dy dx.
The line between (1, 0) and (2, 2) is y = 2x 2, so
_
R
2
(2x + y + 1) dA =
_
2
1
_
1+0.5x
2x2
(2x + y + 1) dy dx.
We can now compute the double integral for R1:
_
1
0
_
1+0.5x
1x
(2x + y + 1) dy dx =
_
1
0
_
2xy +
y
2
2
+ y
_
1+0.5x
1x
dx
=
_
1
0
_
21
8
x
2
+ 3x
_
dx
=
_
7
8
x
3
+
3
2
x
2
_
1
0
dx
=
19
8
,
and the double integral for R2:
_
2
1
_
1+0.5x
2x2
(2x + y + 1) dy dx =
_
2
1
(2xy + y
2
/2 + y)
1+0.5x
2x2
dx
=
_
1
0
_
39
8
x
2
+ 9x +
3
2
_
dx
=
_
13
8
x
3
+
9
2
x
2
+
3
2
x
_
2
1
=
29
8
.
So, Volume =
19
8
+
29
8
=
48
8
= 6.
53. Assume the length of the two legs of the right triangle are a and b, respectively. See Figure 16.9. The line through (a, 0)
and (0, b) is given by
y
b
+
x
a
= 1. So the area of this triangle is
A =
1
2
ab.
Thus the average distance from the points in the triangle to the y-axis (one of the legs) is
Average distance =
1
A
_
a
0
_

b
a
x+b
0
xdy dx
=
2
ab
_
a
0
_
b
a
x
2
+ bx
_
dx
=
2
ab
_
b
3a
x
3
+
b
2
x
2
_
a
0
=
2
ab
_
a
2
b
6
_
=
a
3
.
Similarly, the average distance from the points in the triangle to the x-axis (the other leg) is
Average distance =
1
A
_
b
0
_

a
b
y+a
0
y dxdy
=
2
ab
_
b
0
_
a
b
y
2
+ ay
_
dy
=
2
ab
_
ab
2
6
_
=
b
3
.
x
y
b
a
Figure 16.9
a
b
(a, b)
x
y
Figure 16.10
57. The force, F, acting on A, a small piece of area, is given by
F pA,
where p is the pressure at that point. Thus, if R is the rectangle, the total force is given by
F =
_
R
p dA.
We choose coordinates with the origin at one corner of the plate. See Figure 16.10. Suppose p is proportional to the
square of the distance from the corner represented by the origin. Then we have
p = k(x
2
+ y
2
), for some positive constant k.
Thus, we want to compute
_
R
k(x
2
+ y
2
)dA. Rewriting as an iterated integral, we have
F =
_
R
k(x
2
+ y
2
) dA =
_
b
0
_
a
0
k(x
2
+ y
2
) dxdy = k
_
b
0
_
x
3
3
+ xy
2
a
0
_
dy
= k
_
b
0
_
a
3
3
+ ay
2
_
dy = k
_
a
3
y
3
+ a
y
3
3
b
0
_
=
k
3
(a
3
b + ab
3
).
16.3 SOLUTIONS 353
Exercises
1.
_
W
f dV =
_
2
0
_
1
1
_
3
2
(x
2
+ 5y
2
z) dz dy dx
=
_
2
0
_
1
1
(x
2
z + 5y
2
z
1
2
z
2
)
3
2
dy dx
=
_
2
0
_
1
1
(x
2
+ 5y
2
5
2
) dy dx
=
_
2
0
(x
2
y +
5
3
y
3
5
2
y)
1
1
dx
=
_
2
0
(2x
2
+
10
3
5) dx
= (
2
3
x
3
5
3
x)
2
0
=
16
3

10
3
= 2
5. The region is the half cylinder in Figure 16.11.
9. The region is the cylinder in Figure 16.12.
13. The region is the quarter sphere in Figure 16.13.
1
1
1
x
y
z
Figure 16.11
x
y
z
1
1
1
Figure 16.12
1
1
1
x
y
z
Figure 16.13
Problems
17. The sphere x
2
+ y
2
+ z
2
= 9 intersects the plane z = 2 in the circle
x
2
+ y
2
+ 2
2
= 9
x
2
+ y
2
= 5.
The upper half of the sphere is given by z =
_
9 x
2
y
2
. Thus, using the limits from Figure 16.14 gives
V =
_

5
5
_

5x
2
5x
2
_

9x
2
y
2
2
1 dz dy dx.
The order of integration of x and y can be reversed.
5
y =
5 x
2
y =
5 x
2
x
y
Figure 16.14
21. A slice through W for a xed value of x is a semi-circle the boundary of which is z
2
= r
2
x
2
y
2
, for z 0, so the
inner integral is
_

r
2
x
2
y
2
0
f(x, y, z) dz.
Lining up these stacks parallel to y-axis gives a slice from y =
r
2
x
2
to y =
r
2
x
2
giving
_

r
2
x
2
r
2
x
2
_

r
2
x
2
y
2
0
f(x, y, z) dz dy.
Finally, there is a slice for each x between 0 and r, so the integral we want is
_
r
0
_

r
2
x
2
r
2
x
2
_

r
2
x
2
y
2
0
f(x, y, z) dz dy dx.
25. The pyramid is shown in Figure 16.15. The planes y = 0, and y x = 4, and 2x+y +z = 4 intersect the plane z = 6
in the lines y = 0, y x = 4, 2x + y = 10 on the z = 6 plane as shown in Figure 16.16.
These three lines intersect at the points (4, 0, 6), (5, 0, 6), and (2, 6, 6). Let R be the triangle in the planes
z = 6 with the above three points as vertices. Then, the volume of the solid is
V =
_
6
0
_
(10y)/2
y4
_
42xy
6
dz dxdy
=
_
6
0
_
(10y)/2
y4
(10 2x y) dxdy = 162
=
_
6
0
(10x x
2
xy)
(10y)/2
y4
dy
=
_
6
0
(
9y
2
4
27y + 81) dy
= 162
16.3 SOLUTIONS 355
x
y
z
2
4
4
'
2x + y + z = 4
E
y = 0
'
z = 6 bottom
'
y x = 4 backside
(4, 0, 6)
(5, 0, 6)
(2, 6, 6)
Figure 16.15
(4, 0, 6) (5, 0, 6)
(2, 6, 6)
z = 6 plane
x
y
y x = 4 2x + y 6 = 4
Figure 16.16
29. The required volume, V , is given by
V =
_
5
0
_
3
0
_
x
2
0
dz dy dx
=
_
5
0
_
3
0
x
2
dy dx
=
_
5
0
x
2
y
y=3
y=0
dx
=
_
5
0
3x
2
dx
= 125.
33. (a) The vectors u =

i
j and v =

i
k lie in the required plane so p = u v =

i +
j +
k is perpendicular
to this plane. Let (x, y, z) be a point in the plane, then (x 1)
i + y
j + z
k is perpendicular to p , so ((x 1)
i +
y
j + z
k ) (
i +
j +
j ) = 0 and so
(x 1) + y + z = 0.
Therefore, the equation of the required plane is x + y + z = 1.
(b) The required volume, V , is given by
V =
_
1
0
_
1x
0
_
1xy
0
dz dy dx
=
_
1
0
_
1x
0
(1 x y) dy dx
=
_
1
0
y xy
1
2
y
2
1x
0
dx
=
_
1
0
_
1 x x(1 x)
1
2
(1 x)
2
_
dx
=
_
1
0
1
2
(1 x)
2
dx
=
1
6
.
37. From the problem, we know that (x, y, z) is in the cube which is bounded by the three coordinate planes, x = 0, y = 0,
z = 0 and the planes x = 2, y = 2, z = 2. We can regard the value x
2
+y
2
+z
2
as the density of the cube. The average
value of x
2
+ y
2
+ z
2
is given by
average value =
_
V
(x
2
+ y
2
+ z
2
) dV
volume(V )
=
_
2
0
_
2
0
_
2
0
(x
2
+ y
2
+ z
2
) dxdydz
8
=
_
2
0
_
2
0
_
x
3
3
+ (y
2
+ z
2
)x
_
2
0
dydz
8
=
_
2
0
_
2
0
_
8
3
+ 2y
2
+ 2z
2
_
dydz
8
=
_
2
0
_
8
3
y +
2
3
y
3
+ 2z
2
y
_
2
0
dz
8
=
_
2
0
_
16
3
+
16
3
+ 4z
2
_
dz
8
=
_
32
3
z +
4
3
z
3
_
2
0
8
=
_
64
3
+
32
3
_
8
= 4.
41. Zero. The value of y is positive on the half of the cone above the second and third quadrants and negative (of equal
absolute value) on the half of the cone above the third and fourth quadrants. The integral of y over the entire solid cone is
zero because the integrals over the four quadrants cancel.
45. Negative. If (x, y, z) is any point inside the cone then z < 2. Hence the function z 2 is negative on W and so is its
integral.
49. Positive. The value of x is positive on the half-cone, so its integral is positive.
53. Zero. Write the triple integral as an iterated integral, say integrating rst with respect to y. For xed x and z, the y-integral
is over an interval symmetric about 0. The integral of y over such an interval is zero. If any of the inner integrals in an
iterated integral is zero, then the triple integral is zero.
57. The intersection of two cylinders x
2
+z
2
= 1 and y
2
+z
2
= 1 is shown in Figure 16.17. This region is bounded by four
surfaces:
z =
_
1 x
2
, z =
_
1 x
2
, y =
_
1 z
2
, and y =
_
1 z
2
So the volume of the given solid is
V =
_
1
1
_

1x
2
1x
2
_

1z
2
1z
2
dy dz dx
x
y
z
Figure 16.17
16.4 SOLUTIONS 357
61. The volume V of the solid is 1 2 3 = 6. We need to compute
m
6
_
W
x
2
+ y
2
dV =
m
6
_
1
0
_
2
0
_
3
0
x
2
+ y
2
dz dy dx
=
m
6
_
1
0
_
2
0
3(x
2
+ y
2
) dy dx
=
m
2
_
1
0
(x
2
y + y
3
/3)
2
0
dx
=
m
2
_
1
0
(2x
2
+ 8/3) dx = 5m/3
Exercises
1.
_
2
0
_

2
0
f rdr d
5. A circle is best described in polar coordinates. The radius is 5, so r goes from 0 to 5. To include the entire circle, we need
to go from 0 to 2. The integral is
_
2
0
_
5
0
f(r cos , r sin ) r dr d.
x
y
= /2
= /2
r = 4
Figure 16.18
x
y
= 3/4
= 3/2
r = 4

r = 3
Figure 16.19
1 2
1
2
x
y
Figure 16.20
17. The region is pictured in Figure 16.20. Using polar coordinates, we get
_
R
(x
2
y
2
)dA =
_
/2
0
_
2
1
r
2
(cos
2
sin
2
)rdr d =
_
/2
0
(cos
2
sin
2
)
1
4
r
4
2
1
d
=
15
4
_
/2
0
(cos
2
sin
2
) d
=
15
4
_
/2
0
cos 2 d
=
15
4

1
2
sin 2
/2
0
= 0.
Problems
21. From the given limits, the region of integration is in Figure 16.21.
6
x
y
y = x
6
y = x
Figure 16.21
In polar coordinates, /4 /4. Also,
6 = x = r cos . Hence, 0 r
6/ cos . The integral becomes

_

6
0
_
x
x
dy dx =
_
/4
/4
_

6/cos
0
r dr d
=
_
/4
/4
_
r
2
2
6/cos
0
_
d =
_
/4
/4
6
2 cos
2
d
= 3 tan
/4
/4
= 3 (1 (1)) = 6.
Notice that we can check this answer because the integral gives the area of the shaded triangular region which is
1
2

6
(2
6) = 6.
25. The average value of the function r on the disc R of radius a is
Average of r =
1
Area of R
_
R
rdA =
1
a
2
_
2
0
_
a
0
rrdrd =
1
a
2
_
2
0
a
3
3
d =
1
a
2
2
a
3
3
=
2a
3
.
29. A rough graph of the base of the spring is in Figure 16.22, where the coil is roughly of width 0.01 inches. The volume is
equal to the product of the base area and the height. To calculate the area we use polar coordinates, taking the following
integral:
Area =
_
4
0
_
0.26+0.04
0.25+0.04
rdrd
=
1
2
_
4
0
(0.26 0.04)
2
(0.25 0.04)
2
d
=
1
2
_
4
0
0.01 (0.51 + 0.08)d
= 0.0051 2 +
1
4
(0.0008
2
)
4
0
= 0.0636
Therefore, the volume= 0.0636 0.2 = 0.0127 in
3
.
Figure 16.22
16.5 SOLUTIONS 359
33. (a) The curve r = 1/(2 cos ) or r cos = 1/2 is the line x = 1/2. The curve r = 1 is the circle of radius 1 centered at
the origin. See Figure 16.23.
1
2
1
x = 1/2
= /3
s
= /3
x
y
Figure 16.23
(b) The line intersects the circle where 2 cos = 1, so = /3. From Figure 16.23 we see that
Area =
_
/3
/3
_
1
1/(2 cos )
r dr d.
Evaluating gives
Area =
_
/3
/3
_
r
2
2
1
1/(2 cos )
_
d =
1
2
_
/3
/3
_
1
1
4 cos
2
_
d
=
1
2
_

tan
4
_
/3
/3
=
1
2
_
2
3
2
3
4
_
=
4 3
3
12
.
Exercises
1. (a) is (IV); (b) is (II); (c) is (VII); (d) is (VI); (e) is (III); (f) is (V).
5. The plane has equation = /4.
9.
_
W
f dV =
_
3
1
_
2
0
_
1
0
(sin (r
2
)) rdr d dz
=
_
3
1
_
2
0
(
1
2
cos r
2
)
1
0
d dz
=
1
2
_
3
1
_
2
0
(cos 1 cos 0) d dz
=
_
3
1
(cos 1 1) dz = 4(cos 1 1) = 4(1 cos 1)
13. Using cylindrical coordinates, we get:
_
1
0
_
2
0
_
4
0
f rdr d dz
17. We use Cartesian coordinates, oriented as shown in Figure 16.24. The slanted top has equation z = mx, where m is the
slope in the x-direction, so m = 1/5. Then if f is an arbitrary function, the triple integral is
_
5
0
_
2
0
_
x/5
0
f dzdydx.
x
y
z
1
2
5
(5, 2, 0)
(5, 2, 1)
c
(5, 0, 1)
Figure 16.24
Problems
21. We want the volume of the region above the cone = /3 and below the sphere = 3:
V =
_
2
0
_
/3
0
_
3
0
2
sin d dd.
The order of integration can be altered and other coordinates can be used.
25. We use cylindrical coordinates since the sphere x
2
+ y
2
+ z
2
= 10, or r
2
+ z
2
= 10, and the plane z = 1 can both be
simply expressed. The plane cuts the sphere in the circle r
2
+ 1
2
= 10, or r = 3. Thus
V =
_
2
0
_
3
0
_

10r
2
1
r dz dr d,
or
V =
_
2
0
_

10
1
_

10z
2
0
r dr dz d.
Order of integration can be altered and other coordinates can be used.
29. In rectangular coordinates, a cone has equation z = k
_
x
2
+ y
2
for some constant k. Since z = 4 when
_
x
2
+ y
2
=
2
2
= 2, we have k = 2. Thus, the integral is
_
2
2
_

4x
2
4x
2
_
4
2
x
2
+y
2
h(x, y, z) dz dy dx.
33. The region is a solid cylinder of height 1, radius 1 with base on the xy-plane and axis on the z-axis. We have:
_
1
0
_
1
1
_

1x
2
1x
2
1
(x
2
+ y
2
)
1/2
dy dxdz =
_
1
0
_
2
0
_
1
0
1
r
r dr d dz
=
_
1
0
_
2
0
r
1
0
d dz
=
_
1
0
_
2
0
d dz = 2.
Note that the integral is improper, but it can be shown that the result is correct.
16.5 SOLUTIONS 361
37. The cone is centered along the positive x-axis and intersects the sphere in the circle
(y
2
+ z
2
) + y
2
+ z
2
= 4
y
2
+ z
2
= 2.
We use spherical coordinates with measured from the x-axis and measured in the yz-plane. (Alternatively, the volume
we want is equal to the volume between the cone z =
_
x
2
+ y
2
and the sphere x
2
+ y
2
+ z
2
= 4.) The cone is given
by = /4. The sphere has equation = 2. Thus
Volume =
_
2
0
_
/4
0
_
2
0
2
sin d dd
=
_
2
0
_
/4
0
3
3
sin
2
0
dd
=
_
2
0
_
/4
0
8
3
sin dd
=
_
2
0
8
3
cos
/4
0
d =
_
2
0
8
3
_
1
1
2
_
d
=
16
3
_
1
1
2
_
.
41. The region whose volume we want is shown in Figure 16.25:
x
y
z
=

6
=

3
5
T
c
2
Figure 16.25
Using cylindrical coordinates, the volume is given by the integral:
V =
_
2
0
_
/3
/6
_
5
0
r dr d dz
=
_
2
0
_
/3
/6
r
2
2
5
0
d dz
=
25
2
_
2
0
_
/3
/6
d dz
=
25
2
_
2
0
_
3

6
_
dz
=
25
2

6
2 =
25
6
.
45. The cylinder has radius 2. Using cylindrical coordinates to nd the mass and integrating with respect to r rst, we have
Mass =
_
2
0
_
3
0
_
2
0
(1 + r)r dr dz d =
_
2
0
_
3
0
_
r
2
2
+
r
3
3
_
2
0
dz d = 2 3
_
4
2
+
8
3
_
= 28 gm.
49. (a) We use spherical coordinates. Since = 9 where = 6 and = 11 where = 7, the density increases at a rate
of 2 gm/cm
3
for each cm increase in radius. Thus, since density is a linear function of radius, the slope of the linear
function is 2. Its equation is
11 = 2( 7) so = 2 3.
(b) Thus,
Mass =
_
2
0
_

0
_
7
6
(2 3)
2
sin d dd.
(c) Evaluating the integral, we have
Mass = 2
_
cos
0
_
_
2
4
4

3
3
3
7
6
_
= 2 2(425.5) = 1702 gm = 5346.991 gm.
53. We must rst decide on coordinates. We pick spherical coordinates with the common center of the two spheres as the
origin. We imagine the half-melon with the at side horizontal and the positive z-axis going through the curved surface.
See Figure 16.26. The volume is given by the integral
Volume =
_
2
0
_
/2
0
_
b
a
2
sin ddd.
Evaluating gives
Volume =
_
2
0
_
/2
0
sin

3
3
p=b
p=a
dd = 2(cos )
/2
0
_
b
3
3

a
3
3
_
=
2
3
(b
3
a
3
).
To check our answer, notice that the volume is the difference between the volumes of two half spheres of radius a and b.
These half spheres have volumes 2b
3
/3 and 2a
3
/3, respectively.
x
y
z
a
b
Figure 16.26
57. We rst need to nd the mass of the solid, using cylindrical coordinates:
m =
_
2
0
_
1
0
_

z/a
0
r dr dz d
=
_
2
0
_
1
0
z
2a
dz d
=
_
2
0
1
4a
d =

2a
It makes sense that the mass would vary inversely with a, since increasing a makes the paraboloid skinnier. Now for
the z-coordinate of the center of mass, again using cylindrical coordinates:
16.5 SOLUTIONS 363
z =
2a
_
2
0
_
1
0
_

z/a
0
zr dr dz d
=
2a
_
2
0
_
1
0
z
2
2a
dz d
=
2a
_
2
0
1
6a
d =
2
3
61. Assume the base of the cylinder sits on the xy-plane with center at the origin. Because the cylinder is symmetric about the
z-axis, the force in the horizontal x or y direction is 0. Thus we need only compute the vertical z component of the force.
We are going to use cylindrical coordinates; since the force is G mass/(distance)
2
, a piece of the cylinder of volume
dV located at (r, , z) exerts on the unit mass a force with magnitude G( dV )/(r
2
+ z
2
). See Figure 16.27.
Vertical component
of force
=
G( dV )
r
2
+ z
2
cos =
G dV
r
2
+ z
2

z
r
2
+ z
2
=
Gz dV
(r
2
+ z
2
)
3/2
.
Adding up all the contributions of all the dV s, we obtain
Vertical force =
_
H
0
_
2
0
_
R
0
Gzr
(r
2
+ z
2
)
3/2
drddz
=
_
H
0
_
2
0
(Gz)
_
r
2
+ z
2
_
R
0
ddz
=
_
H
0
_
2
0
(Gz)
_
R
2
+ z
2
+
1
z
_
ddz
=
_
H
0
2G
_
1
z
R
2
+ z
2
_
dz
= 2G(z
_
R
2
+ z
2
)
H
0
= 2G(H
_
R
2
+ H
2
+ R) = 2G(H + R
_
R
2
+ H
2
)
r
2
+ z
2
z
Figure 16.27
65. Use cylindrical coordinates, with the z-axis being the axis of the cable. Consider a piece of cable of length 1. Then
Stored energy =
1
2
_
b
a
_
1
0
_
2
0
E
2
r d dz dr =
q
2
8
2
_
b
a
_
1
0
_
2
0
1
r
d dz dr
=
q
2
4
_
b
a
1
r
dr =
q
2
4
(ln b ln a) =
q
2
4
ln
b
a
.
So the stored energy is proportional to ln(b/a) with constant of proportionality q
2
/4.
Exercises
1. We have p(x, y) = 0 for all points (x, y) satisfying x 3, since all such points lie outside the region R. Therefore the
fraction of the population satisfying x 3 is 0.
5. Since p(x, y) = 0 for all (x, y) outside the rectangle R, the population is given by the volume under the graph of p over
the region inside the rectangle R and to the right of the line x = y. Therefore the fraction of the population is given by
the double integral:
_
1
0
_
2
y
xy dxdy =
_
1
0
x
2
y
2
2
y
dy =
_
1
0
_
2y
y
3
2
_
dy =
_
y
2
y
4
8
_
1
0
=
7
8
.
9. No, p is not a joint density function. Since p(x, y) = 0 outside the region R, the volume under the graph of p is the same
as the volume under the graph of p over the region R, which is 2 not 1.
13. Yes, p is a joint density function. In the region R we have 1 x
2
+ y
2
, so p(x, y) = (2/)(1 x
2
y
2
) 0 for all x
and y in R, and p(x, y) = 0 for all other (x, y). To check that p is a joint density function, we check that the total volume
under the graph of p over the region R is 1. Using polar coordinates, we get:
_
R
p(x, y)dA =
2
_
2
0
_
1
0
(1 r
2
)r dr d =
2
_
2
0
_
r
2
2

r
4
4
_
1
0
d =
2
_
2
0
1
4
d = 1.
Problems
17. (a) For a density function,
1 =
_

f(x, y) dy dx =
_
2
0
_
1
0
kx
2
dy dx
=
_
2
0
kx
2
dx
=
kx
3
3
2
0
=
8k
3
.
So k = 3/8.
(b)
_
1
0
_
2y
0
3
8
x
2
dxdy =
_
1
0
1
8
(2 y)
3
dy =
1
32
(2 y)
4
1
0
=
15
32
(c)
_
1/2
0
_
1
0
3
8
x
2
dxdy =
_
1/2
0
1
8
x
3
1
0
dy =
_
1/2
0
1
8
dy =
1
16
.
21. (a) If t 0, then F(t) = 0 because the average of two positive numbers can not be negative. If 1 < t then F(t) = 1
because the average of two numbers each at most 1 is certain to be less than or equal to 1. For any t, we have
F(t) =
_
R
p(x, y)dA where R is the region of the plane dened by (x + y)/2 t. Since p(x, y) = 0 outside the
unit square, we need integrate only over the part of R that lies inside the square, and since p(x, y) = 1 inside the
square, the integral equals the area of that part of the square. Thus, we can calculate the area using area formulas.
For 0 t 1, we draw the line (x + y)/2 = t, which has x- and y-intercepts of 2t. Figure 16.28 shows that for
0 < t 1/2,
F(t) = Area of triangle =
1
2
2t 2t = 2t
2
.
In Figure 16.29, when x = 1, we have y = 2t 1. Thus, the vertical side of the unshaded triangle is 1 (2t 1) =
2 2t. The horizontal side is the same length, so for 1/2 < t 1,
F(t) = Area of Square Area of triangle = 1
2
1
2
(2 2t)
2
1 2(1 t)
2
.
16.7 SOLUTIONS 365
The nal result is:
F(t) =
_
_
_
0 if t 0
2t
2
if 0 < t 1/2
1 2(1 t)
2
if 1/2 < t 1
1 if 1 < t
.
1
2t
1
2t
x+y
2
= t
x
y
Figure 16.28: For 0 < t
1
2
1 2t
2t
1
x+y
2
= t
E '
2 2t
E '
2t 1
T
c
2 2t
T
c
2t 1
x
y
Figure 16.29: For
1
2
< t 1
(b) The probability density function p(t) of z is the derivative of its cumulative distribution function. We have
p(t) =
_
_
_
0 if t 0
4t if 0 < t 1/2
4 4t if 1/2 < t 1
0 if 1 < t
.
See Figure 16.30.
(c) The values of x and y and equally likely to be near 0, 1/2, and 1. Notice from the graph of the density function in
Figure 16.30 that even though x and y separately are equally likely to be anywhere between 0 and 1, their average
z = (x + y)/2 is more likely to be near 1/2 than to be near 0 or 1.
0.5 1
2
0
x
p(t)
Figure 16.30
Exercises
1. We have
(x, y)
(s, t)
=
xs xt
ys yt
5 2
3 1
= 1.
Therefore,

(x, y)
(s, t)
= 1.
5. We have
(x, y, z)
(s, t, u)
=
xs xt xu
ys yt yu
zs zt zu
3 1 2
1 5 1
2 1 1
.
This 3 3 determinant is computed the same way as for the cross product, with the entries 3, 1, 2 in the rst row playing
the same role as
i ,
j ,
k . We get
(x, y, z)
(s, t, u)
= ((5)(1) (1)(1))3 + ((1)(2) (1)(1))1 + ((1)(1) (2)(5))2 = 13.
9. The square T is dened by the inequalities
0 s = ax 1 0 t = by 1
that correspond to the inequalities
0 x 1/a = 50 0 y 1/b = 10
that dene R. Thus a = 1/50 and b = 1/10.
13. Inverting the change of variables gives x = s at, y = t.
The four edges of R are
y = 0, y = 5, y =
1
3
x, y =
1
3
(x 10).
The change of variables transforms the edges to
t = 0, t = 5, t =
1
3
s +
1
3
at, t =
1
3
s +
1
3
at +
10
3
.
These are equations for the edges of a rectangle in the st-plane if the last two are of the form: s = (Constant). This
happens when the t terms drop out, or a = 3. With a = 3 the change of variables gives
_ _
T
(x, y)
(s, t)
ds dt
over the rectangle
T : 0 t 5, 0 s 10.
Problems
17. Given
_
_
x = sin cos
y = sin sin
z = cos ,
(x, y, z)
(, , )
=
sin cos cos cos sin sin

sin sin cos sin sin cos
cos sin 0
= cos
cos cos sin sin

cos sin sin cos
+ sin
sin cos sin sin

sin sin sin cos
= cos (
2
cos
2
cos sin +
2
sin
2
cos sin )
+ sin ( sin
2
cos
2
+ sin
2
sin
2
)
=
2
cos
2
sin +
2
sin
3
=
2
sin .
SOLUTIONS to Review Problems for Chapter Sixteen 367
21. The area of the ellipse is
_ _
R
dxdy where R is the region x
2
+ 2xy + 2y
2
1. We must change variables in both the
area element dA = dxdy and the region R.
Inverting the variable change gives x = s t, y = t. Thus
(x, y)
(s, t)
=
x
s
x
t
y
s
y
t
1 1
0 1
= 1.
Therefore
dxdy =
(x, y)
(s, t)
ds dt = ds dt.
The region of integration is
x
2
+ 2xy + 2y
2
= (s t)
2
+ 2(s t)t + 2t
2
= s
2
+ t
2
1.
Let T be the unit disc s
2
+ t
2
1. We have
_ _
R
dxdy =
_ _
T
ds dt = Area of T = .
25. Given
_
s = xy
t = xy
2
,
we have
(s, t)
(x, y)
=
s
x
s
y
t
x
t
y
y x
y
2
2xy
= xy
2
= t.
Since
(s, t)
(x, y)

(x, y)
(s, t)
= 1,
(x, y)
(s, t)
= t so
(x, y)
(s, t)
=
1
t
So
_
R
xy
2
dA =
_
T
t
(x, y)
(s, t)
ds dt =
_
T
t(
1
t
) ds dt =
_
T
ds dt,
where T is the region bounded by s = 1, s = 4, t = 1, t = 4.
Then
_
R
xy
2
dA =
_
4
1
ds
_
4
1
dt = 9.
Exercises
1. We use Cartesian coordinates, oriented so that the cube is in the rst quadrant. See Figure 16.31. Then, if f is an arbitrary
function, the integral is
_
2
0
_
3
0
_
5
0
f dxdydz.
Other answers are possible. In particular, the order of integration can be changed.
x
y
z
5
(5, 3, 0)
(5, 3, 2)
2
Figure 16.31
9. The region is the half cylinder in Figure 16.33.
1 1
1
1
x
y
x
2
+ y
2
= 1
Figure 16.32
1
1
1
x
y
z
Figure 16.33
2
4
x
y
x = (y 4)/2 or y = 2x + 4
Figure 16.34
(b)
_
2
0
_
2x+4
0
g(x, y) dy dx.
17. First use integration by parts, with y as the variable, u = x
2
y, u
= x
2
, v =
sin (xy)
x
, v
= cos (xy). Then,

_
4
3
_
1
0
x
2
y cos (xy) dy dx =
_
4
3
_
[xy sin (xy)]
1
0
_
1
0
xsin (xy) dy
_
dx
=
_
4
3
_
xsin x + [cos (xy)]
1
0
_
dx
=
_
4
3
(xsin x + cos x 1) dx.
Now use integration by parts again, with u = x, u
= 1, v = cos x, v
= sin x. Then,
_
4
3
(xsin x + cos x 1) dx = [xcos x]
4
3
+
_
4
3
cos xdx +
_
4
3
(cos x 1) dx
= (xcos x + 2 sin x x)|
4
3
= 4 cos 4 + 2 sin 4 + 3 cos 3 2 sin 3 1.
Thus,
_
4
3
_
1
0
x
2
y cos (xy) dy dx = 4 cos 4 + 2 sin 4 + 3 cos 3 2 sin 3 1.
21. (a) A vertical plane perpendicular to the x-axis: x = 2.
(b) A cylinder: r = 3.
(c) A sphere: =
3.
(d) A cone: = /4.
(e) A horizontal plane: z = 5.
(f) A vertical half-plane: = /4.
25. From Figure 16.35, we have the following iterated integrals:
x
y
z
x
2
+ y
2
+ z
2
= 1
Figure 16.35
(a)
_
R
f dV =
_
1
1
_

1x
2
1x
2
_

1x
2
y
2
0
f(x, y, z) dzdydx
(b)
_
R
f dV =
_
1
1
_

1y
2
1y
2
_

1x
2
y
2
0
f(x, y, z) dzdxdy
(c)
_
R
f dV =
_
1
1
_

1y
2
0
_

1y
2
z
2
1y
2
z
2
f(x, y, z) dxdzdy
(d)
_
R
f dV =
_
1
1
_

1x
2
0
_

1x
2
z
2
1x
2
z
2
f(x, y, z) dydzdx
(e)
_
R
f dV =
_
1
0
_

1z
2
1z
2
_

1x
2
z
2
1x
2
z
2
f(x, y, z) dydxdz
(f)
_
R
f dV =
_
1
0
_

1z
2
1z
2
_

1y
2
z
2
1y
2
z
2
f(x, y, z) dxdydz
Problems
29. The region is a hollow half-sphere, with inner radius

3 and outer radius

4 = 2. See Figure 16.36. In spherical
coordinates, the integral is
_

0
_

0
_
2
2
sin d dd.
The order of integration can be altered.
x
y
z
c
2
Figure 16.36
33. W is a cylindrical shell, so cylindrical coordinates should be used. See Figure 16.37.
T
c
4
E '
1
E '
1
x
y
z
Figure 16.37
_
W
z
(x
2
+ y
2
)
3/2
dV =
_
4
0
_
2
0
_
2
1
z
r
3
rdr d dz
=
_
4
0
_
2
0
_
2
1
z
r
2
dr d dz
=
_
4
0
_
2
0
(
z
r
)
2
1
d dz
=
_
4
0
_
2
0
z
2
d dz
=
_
4
0
z
2
2 dz =
1
2
z
2
4
0
= 8
37. Cant tell, since y
3
is both positive and negative for x < 0.
41. Zero. You can see this in several ways. One way is to observe that xy is positive on the part of the sphere above and below
the rst and third quadrants (where x and y are of the same sign) and negative (of equal absolute value) on the part of the
sphere above and below the second and fourth quadrants (where x and y have opposite signs). These add up to zero in the
integral of xy over all of W.
Another way to see that the integral is zero is to write the triple integral as an iterated integral, say integrating rst
with respect to x. For xed y and z, the x-integral is over an interval symmetric about 0. The integral of x over such an
interval is zero. If any of the inner integrals in an iterated integral is zero, then the triple integral is zero.
45. Negative. Since z
2
1 0 in the sphere, its integral is negative.
49. The depth of the lake is given in meters and the diameter in kilometers. We should work with a single unit of length. In
this solution we work with kilometers, but meters would work just as well.
The shape of the lake suggests integration in polar coordinates, with r km measured from the center of the island.
Thus t = r 1 is the distance in kilometers from the island when r varies between 1 and 5. The depth of the lake r km
from the center of the island is
Depth =
100(r 1)(4 (r 1))
1000
=
1
2
+
3r
5

r
2
10
km.
Volume of the lake =
_
2
0
_
5
1
_
1
2
+
3r
5

r
2
10
_
r dr d =
32
5
= 20.1 km
3
.
53. (a)
_
2
0
_

0
_
2
1
2
sin ddd.
(b)
_
2
0
_
2
0
_

4r
2
4r
2
r dzdrd
_
2
0
_
1
0
_

1r
2
1r
2
r dzdrd.
57. The region of integration is shown in Figure 16.38, and the mass of the given solid is given by
x
y
z
3
4
12
z = 12 4x 3y
'
4x + 3y = 12
or y =
1
3
(12 4x)
Figure 16.38
Mass =
_
R
dV
=
_
3
0
_ 1
3
(124x)
0
_
124x3y
0
x
2
dzdydx
=
_
3
0
_ 1
3
(124x)
0
x
2
z
z=124x3y
z=0
dydx
=
_
3
0
_ 1
3
(124x)
0
x
2
(12 4x 3y) dydx
=
_
3
0
x
2
(12y 4xy
3
2
y
2
)
y=
1
3
(124x)
0
dx
=
_
8x
3
4x
4
+
8
15
x
5
_
3
0
=
108
5
.
61. The plane (x/p) + (y/q) + (z/r) = 1 cuts the axes at the points (p, 0, 0); (0, q, 0); (0, 0, r). Since p, q, r are positive,
the region between this plane and the coordinate planes is a pyramid in the rst octant. Solving for z gives
z = r
_
1
x
p

y
q
_
= r
rx
p

ry
q
.
The volume, V , is given by the double integral
V =
_
R
_
r
rx
p

ry
q
_
dA,
where R is the region shown in Figure 16.39. Thus
V =
_
p
0
_
qqx/p
0
_
r
rx
p

ry
q
_
dydx
=
_
p
0
_
_
ry
rxy
p

ry
2
2q
_
y=qqx/p
y=0
_
dx
=
_
p
0
_
r
_
q
qx
p
_
r
p
x
_
q
qx
p
_
r
2q
_
q
qx
p
_
2
_
dx
=
_
p
0
rq
2rqx
p
+
rqx
2
p
2

rq
2
2q
_
1
2x
p
+
x
2
p
2
_
dx
=
_
rqx
rqx
2
p
+
rqx
3
p
2
3

rqx
2
+
rqx
2
p2

rqx
3
2p
2
3
_
p
0
= pqr pqr +
pqr
3

pqr
2
+
pqr
2

pqr
6
=
pqr
6
.
p
q
R
_
_
_
x
p
+
y
q
= 1
y = q
qx
p
x
y
Figure 16.39
65. We must rst decide on coordinates. We pick Cartesian coordinates with the smaller sphere centered at the origin, the
larger one centered at (0, 0, 1). A vertical cross-section of the region in the xz-plane is shown in Figure 16.40. The
smaller sphere has equation x
2
+ y
2
+ z
2
= 1. The larger sphere has equation x
2
+ y
2
+ (z + 1)
2
= 2.
1
x
2
+ y
2
+ z
2
= 1
x
2
+ y
2
+ (z + 1)
2
= 2
x
z
Figure 16.40
Let R represent the region in the xy-plane which lies directly underneath (or above) the region whose volume we
want. The curve bounding this region is a circle, and we nd its equation by solving the system:
x
2
+ y
2
+ z
2
= 1
x
2
+ y
2
+ (z + 1)
2
= 2
Subtracting the equations gives
(z + 1)
2
z
2
= 1
2z + 1 = 1
z = 0.
Since z = 0, the two surfaces intersect in the xy-plane in the circle x
2
+ y
2
= 1. Thus R is x
2
+ y
2
1.
The top half of the small sphere is represented by z =
_
1 x
2
y
2
; the top half of the large sphere is represented
by z = 1 +
_
2 x
2
y
2
. Thus the volume is given by
Volume =
_
1
1
_

1x
2
1x
2
_

1x
2
y
2
1+
2x
2
y
2
dzdydx.
Starting to evaluate the integral, we get
Volume =
_
1
1
_

1x
2
1x
2
(
_
1 x
2
y
2
+ 1
_
2 x
2
y
2
) dydx.
We simplify the integral by converting to polar coordinates
Volume =
_
2
0
_
1
0
_
_
1 r
2
+ 1
_
2 r
2
_
r drd
=
_
2
0
_
(1 r
2
)
3/2
3
+
r
2
2
+
(2 r
2
)
3/2
3
_
1
0
d
= 2
_
1
2
+
1
3

_
1
3
+
2
3/2
3
__
= 2
_
7
6

2
2
3
_
= 1.41.
69. The outer circle is a semicircle of radius 4. This is shown in Figure 16.41, with center at D. Thus, CE = 2 and DC = 2,
while AD = 4. Notice that angle ADO is a right angle.
O
r 2
D
2
C
2
E
A
4
r
B
Figure 16.41
Suppose the large circle has center O and radius r. Then OA = r and OD = OC DC = r 2. Applying
Pythagoras Theorem to triangle OAD gives
r
2
= 4
2
+ (r 2)
2
r
2
= 16 + r
2
4r + 4
r = 5.
If we put the origin at O, the equation of the large circle is x
2
+ y
2
= 25. In the same coordinates, the equation of the
small circle, which has center at D = (3, 0), is (x 3)
2
+ y
2
= 16. The right hand side of the two circles are given by
x =
_
25 y
2
and x = 3 +
_
16 y
2
.
Since the y-coordinate of A is 4 and the y-coordinate of B is 4, we have
Area =
_
4
4
_
3+
16y
2
25y
2
1 dxdy
=
_
4
4
(3 +
_
16 y
2
_
25 y
2
) dy
= 13.95.
73. In Cartesian coordinates the integral is
_
D
3
_
x
2
+ y
2
dA =
_
1
1
_

1x
2
1x
2
3
_
x
2
+ y
2
dydx.
In polar coordinates it is
_
D
3
_
x
2
+ y
2
dA =
_
2
0
_
1
0
3
r
2
rdrd =
_
2
0
_
1
0
r
5/3
drd
=
_
2
0
3
8
d =
6
7
The Cartesian coordinate version requires the use of a computer algebra system. Some CASs may be able to handle it
and may give the answer in terms of functions called hypergeometric functions. To compare the answers are the same you
may need to ask the CAS to give a numerical value for the answer. Its possible your CAS will not be able to handle the
integral at all.
1. False. For example, if f(x, y) < 0 for all (x, y) in the region R, then
_
R
f dA is negative.
5. True. The double integral is the limit of the sum
A0
(x, y)A. Each of the terms (x, y)A is an approximation of
the total population inside a small rectangle of area A. Thus the limit of the sum of all of these numbers as A
gives the total population of the region R.
9. False. There is no reason to expect this to be true, since the behavior of f on one half of R can be completely unrelated
to the behavior of f on the other half. As a counterexample, suppose that f is dened so that f(x, y) = 0 for points
(x, y) lying in S, and f(x, y) = 1 for points (x, y) lying in the part of R that is not in S. Then
_
S
f dA = 0, since
f = 0 on all of S. To evaluate
_
R
f dA, note that f = 1 on the square S1 which is 0 x 1, 1 y 2. Then
_
R
f dA =
_
S
1
f dA = Area(S1) = 1, since f = 0 on S.
13. True. For any point in the region of integration we have 1 x 2, and so y is between the positive numbers 1 and 8.
17. False. The given limits describe only the upper half disk where y 0. The correct limits are
_
a
a
_
a
2
x
2
a
2
x
2
fdydx.
21. False. The integral gives the total mass of the material contained in W.
25. True. Both sets of limits describe the solid region lying above the rectangle 1 x 1, 0 y 1, z = 0 and below
the parabolic cylinder z = 1 x
2
.
29. False. As a counterexample, let W1 be the solid cube 0 x 1, 0 y 1, 0 z 1, and let W2 be the solid cube
1
2
x 0,
1
2
y 0,
1
2
z 0. Then volume(W1) = 1 and volume(W2) =
1
8
. Now if f(x, y, z) = 1, then
_
W
1
f dV = 1 1 which is less than
_
W
2
f dV =
1
8
1.
17.1 SOLUTIONS 375
CHAPTER SEVENTEEN
Exercises
1. We want the bottom half of a semicircle of radius 1 centered at (0, 1). The equations x = cos t, y = 1 + sin t describe
clockwise motion in this circle, passing (1, 1) when t = and (1, 1) when t = 2. So a possible parameterization is
x = cos t, y = 1 + sin t, t 2.
5. Since we are moving on the y-axis, x = 0, and y goes from 2 to 1. Thus a possible parameterization is
x = 0, y = t, 2 t 1.
9. One possible parameterization is
x = 3 + 2t, y = 4 + 2t, z = 2 3t.
13. The displacement vector from the rst point to the second is v = 4
i 5
j 3
k . The line through point (1, 5, 2) and

with direction vector v = 4
i 5
j 3
k is given by parametric equations

x = 1 + 4t,
y = 5 5t,
z = 2 3t.
Other parameterizations of the same line are also possible.
17. The line passes through (3, 0, 0) and (0, 0, 5). The displacement vector from the rst of these points to the second is
v = 3
i 5
k . The line through point (3, 0, 0) and with direction vector v = 3
i 5
k is given by parametric
equations
x = 3 3t,
y = 0,
z = 5t.
Other parameterizations of the same line are also possible.
21. The circle lies in the plane z = 2, so one possible answer is
x = 3 cos t, y = 3 sin t, z = 2.
25. The xy-plane is z = 0, so a possible answer is
x = t
2
, y = t, z = 0.
29. Since its diameters are parallel to the y and z-axes and its center is in the yz-plane, the ellipse must lie in the yz-plane,
x = 0. The ellipse with the same diameters centered at the origin would have its y-coordinate range between 5/2 and
5/2 and its z-coordinate range between 1 and 1. Thus this ellipse has equation
x = 0, y =
5
2
cos t, z = sin t.
376 Chapter Seventeen /SOLUTIONS
To move the center to (0, 1, 2), we add 1 to the equation for y and 2 to the equation for z, so one possible answer for
our ellipse is
x = 0, y = 1 +
5
2
cos t, z = 2 + sin t.
33. The displacement vector between the points is u = 3
i + 5
k , so a possible parameterization of the line is

x = 1 + 3t, y = 2, z = 3 + 5t.
37. The line segment PQhas length 10, so it must be a diameter of the circle. The center of the circle is therefore the midpoint
of PQ, which is the point (5, 0). The upper arc of the circle between P and Q can be parameterized as follows:
r (t) = 5
i + 5(cos t
i + sin t
j ), 0 t .
The lower arc can be parameterized as follows:
r (t) = 5
i + 5(cos t
i + sin t
j ), t 2.
Problems
41. We nd the parameterization in terms of the displacement vector
OP = 2
i + 5
j from the origin to the point P and the

displacement vector
PQ = 10
i + 4
j from P to Q.
r (t) =
OP + (t 10)
PQ or r (t) = (2 + (t 10)10)
i + (5 + (t 10)4)
j
45. These equations parameterize a line. Since (3 + t) + (2t) + 3(1 t) = 6, we have x + y + 3z = 6. Similarly,
x y z = (3 + t) 2t (1 t) = 2. That is, the curve lies entirely in the plane x + y + 3z = 6 and in the plane
x y z = 2. Since the normals to the two planes, n1 =
i +
j +3
k and n2 =
k are not parallel, the line is

the intersection of two nonparallel planes, which is a straight line in 3-dimensional space.
49. The coefcients of t in the parameterizations show that line r 1 is parallel to the vector 3
i + 2
j +

k and line r 2 is
parallel to 6
i + 4
j + 3
k . Since these vectors are not parallel, the lines are not parallel, so the lines are different.
53. Add the two equations to get 2x+3z = 5, or x =
3
2
z+
5
2
. Subtract the two equations to get 2yz = 1, or y =
1
2
z+
1
2
.
So a possible parameterization is
x =
3
2
t +
5
2
, y =
1
2
t +
1
2
, z = t.
57. (a) Both paths are straight lines, the rst passes through the point (1, 4, 1) in the direction of the vector
j + 2
k
and the second passes through (7, 6, 1) in the direction of the vector 2
i + 2
j +

k . The two paths are not
parallel.
(b) Is there a time t when the two particles are at the same place at the same time? If so, then their coordinates will be
the same, so equating coordinates we get
1 + t = 7 + 2t
4 t = 6 + 2t
1 + 2t = 1 + t.
Since the rst equation is solved by t = 6, the second by t = 10/3, and the third by t = 0, no value of t solves all
three equations. The two particles never arrive at the same place at the same time, and so they do not collide.
(c) Are there any times t1 and t2 such that the position of the rst particle at time t1 is the same as the position of the
second particle at time t2? If so then
1 + t1 = 7 + 2t2
4 t1 = 6 + 2t2
1 + 2t1 = 1 + t2.
We solve the rst two equations and get t1 = 2 and t2 = 4. This is a solution for the third equation as well, so the
three equations are satised by t1 = 2 and t2 = 4. At time t = 2 the rst particle is at the point (1, 2, 3), and at time
t = 4 the second is at the same point. The paths cross at the point (1, 2, 3), and the rst particle gets there rst.
17.1 SOLUTIONS 377
61. The helices wind around a cylinder of radius , which explains the signicance of . As t increases from 0 to 2, the
helix winds once around the cylinder, climbing upward a distance of 2. Thus controls how stretched out the helix is
in the vertical direction. See Figure 17.1 and Figure 17.2.
x y
z
x y
z
x y
z
Figure 17.1: Three values of with the same
x y
z
x y
z
x y
z
Figure 17.2: Three values of with the same
65. The line r = a + t
b is parallel to the vector
b and through the point with position vector a .

(a) is (vii). The equation
b r = 0 is a plane perpendicular to
b and satised by (0, 0, 0).

(b) is (ii). For any constant k, the equation
b r = k is a plane perpendicular to
b . If k = ||a || = 0, the plane does not

contain the origin.
(c) is (iv). The equation (a
b ) (r a ) = 0 is the equation of a plane which is satised by r = a , so the point with

position vector a lies on the plane. Since a
b is perpendicular to
b , the plane is parallel to the line, and therefore

it contains the line.
69. (a) If n v = 0, then n and v are perpendicular. Since P1 is perpendicular to n and L is parallel to v , we see that P1
and L are parallel. In fact, L may lie in the plane.
(b) Since n v is perpendicular to n and to v , the vector n v is parallel to P1 and perpendicular to L. Thus, P2,
which is perpendicular to n v , is
(i) Perpendicular to P1.
(ii) Parallel to L.
Exercises
1. The velocity vector v is given by:
v =
d(t)
dt
i +
_
d
dt
(t
3
t)
_
j =
i + (3t
2
1)
j .
The acceleration vector a is given by:
a =
dv
dt
=
d(1)
dt
i +
_
d
dt
(3t
2
1)
_
j = 6t
j .
v =
d
dt
(3 cos t)
i +
d
dt
(4 sin t)
j = 3 sin t
i + 4 cos t
j .
The acceleration vector a is given by:
a =
dv
dt
=
d
dt
(3 sin t)
i +
d
dt
(4 cos t)
j = 3 cos t
i 4 sin t
j .
v =
d
dt
(t)
i +
d
dt
(t
2
)
j +
d
dt
(t
3
)
k
=
i + 2t
j + 3t
2
k .
The speed is given by:
v =
_
1 + 4t
2
+ 9t
4
.
Now v is never zero since 1 + 4t
2
+ 9t
4
1 for all t. Thus, the particle never stops.
13. We have
Length =
_
2
1
_
(x
(t))
2
+ (y
(t))
2
+ (z
(t))
2
dt =
_
2
1
_
5
2
+ 4
2
+ (1)
2
dt =
42.
This is the length of a straight line from the point (8, 5, 2) to (13, 9, 1).
17. The velocity vector v is
v =
dx
dt
i +
dy
dt
j +
dz
dt
k = 0
i + 2(3) cos(3t)
j + 2(3)(sin(3t))
k
= 6 cos(3t)
j 6 sin(3t)
k .
The acceleration vector a is
a =
d
2
x
dt
2
i +
d
2
y
dt
2
j +
d
2
z
dt
2
k = 6(3)(sin(3t))
j 6(3) cos(3t)
k
= 18 sin(3t)
j 18 cos(3t)
k .
To check that v and a are perpendicular, we check that the dot product is zero:
v a = (6 cos(3t)
j 6 sin(3t)
k ) (18 sin(3t)
j 18 cos(3t)
k )
= 108 cos(3t) sin(3t) + 108 sin(3t) cos(3t) = 0.
The speed is
v = 6 cos(3t)
j 6 sin(3t)
k = 6
_
sin
2
(3t) + cos
2
(3t) = 6,
and so is constant. The magnitude of the acceleration is
a = 18 sin(3t)
j 18 cos(3t)
k = 18
_
sin
2
(3t) + cos
2
(3t) = 18,
which is also constant.
17.2 SOLUTIONS 379
Problems
21. Plotting the positions on the xy plane and noting their times gives the graph shown in Figure 17.3.
2 4 6 8
0
2
4
6
8
10
x
y
t = 0
t = 0.5
t = 1
t = 1.5
t = 2
t = 2.5, 3.5
t = 3
t = 4
Figure 17.3
(a) We approximate dx/dt by x/t calculated between t = 1.5 and t = 2.5:
dx
dt

x
t
=
3 7
2.5 1.5
=
4
1
= 4.
Similarly,
dy
dt

y
t
=
10 5
2.5 1.5
=
5
1
= 5.
So,
v (2) 4
i + 5
j and Speed = v =
41.
(b) The particle is moving vertically at about time t = 1.5. Note that the particle is momentarily stopped at about t = 3;
however it is not moving parallel to the y-axis at this instant.
(c) The particle stops at about time t = 3 and reverses course.
25. (a) The vector
PQ between the points is given by
PQ = 2
i + 5
j + 3
k .
Since ||
PQ|| =
2
2
+ 5
2
+ 3
2
=
38, the velocity vector of the motion is

v =
5
38
(2
i + 5
j + 3
k ).
(b) The motion is along a line starting at the point (3, 2, 5) and with the velocity vector from part (a). The equation of
the line is
r = 3
i + 2
j 5
k + tv = 3
i + 2
j 5
k +
5
38
(2
i + 5
j + 3
k )t,
so
x = 3 +
10
38
t, y = 2 +
25
38
t, z = 5 +
15
38
t.
29. (a) Since z = 90 feet when t = 0, the tower is 90 feet high.
(b) The child reaches the bottom when z = 0, so t = 90/5 = 18 minutes.
(c) Her velocity is given by
v =
dr
dt
= (10 sin t)
i + (10 cos t)
j 5
k ,
so
Speed = ||v || =
_
(10 sin t)
2
+ (10 cos t)
2
+ (5)
2
=
_
10
2
+ 5
2
=
125 ft/min.
(d) Her acceleration is given by
a =
dv
dt
= (10 cos t)
i (10 sin t)
j ft/min
2
.
33. (a) For any positive constant k, the parameterization
x = 5 sin(kt) y = 5 cos(kt)
moves counterclockwise on a circle of radius 5 starting at the point (0, 5). We choose k to make the period 8 seconds.
If k 8 = 2, then k = /4 and the parameterization is
x = 5 sin
_
t
4
_
y = 5 cos
_
t
4
_
.
(b) Since it takes 8 seconds for the particle to go around the circle
Speed =
Circumference of circle
8
=
2(5)
8
=
5
4
cm/sec.
37. (a) The center of the wheel moves horizontally, so its y-coordinate will never change; it will equal 1 at all times. In one
second, the wheel rotates 1 radian, which corresponds to 1 meter on the rim of a wheel of radius 1 meter, and so the
rolling wheel advances at a rate of 1 meter/sec. Thus the x-coordinate of the center, which equals 0 at t = 0, will
equal t at time t. At time t the center will be at the point (x, y) = (t, 1).
(b) By time t the spot on the rim will have rotated t radians clockwise, putting it at angle t as in Figure 17.4. The
coordinates of the spot with respect to the center of the wheel are (cos(t), sin(t)). Adding these to the coordinates
(t, 1) of the center gives the location of the spot as (x, y) = (t + cos t, 1 sin t). See Figure 17.5.
(t, 1)
t
x
y
Figure 17.4
x
y
Figure 17.5
41. (a) Let x represent horizontal displacement (in cm) from some starting point and y the distance (in cm) above the ground.
Since
25 km/hr =
25 10
5
60
2
= 694.444 cm/sec,
if t is in seconds, the motion of the center of the pedal is given by
x
i + y
j = 694.444t
i + 30
j .
The circular motion of your foot relative to the center is described by
h
i + k
j = 20 cos(2t)
i + 20 sin(2t)
j ,
so the motion of the light on your foot relative to the ground is described by
x
i + y
j = (694.444t + 20 cos(2t))
i + (30 + 20 sin(2t))
j .
600 1200
20
40
x
y
Figure 17.6
(c) Suppose your pedal is rotating with angular velocity radians/sec, so that the motion is described by
x
i + y
j = (694.444t + 20 cos t)
i + (30 + 20 sin t)
j .
The light moves backward if dx/dt is negative. Since
dx
dt
= 694.444 20 sin t,
the minimum value of dx/dt occurs when t = /2, and then
dx
dt
= 694.444 20 < 0
giving
34.722 radians/sec.
Since there are 2 radians in a complete revolution, an angular velocity of 34.722 radians/sec means 34.722/2
5.526 revolutions/sec.
45. (a) Using the product rule for differentiation we get
d
dt
(r r ) = r
dr
dt
+
dr
dt
r = 2r
dr
dt
.
(b) Since a is a constant, da /dt = 0 so the product rule gives
d
dt
(a r ) = a
dr
dt
.
(c) The product rule gives
d
dt
(r
3
r ) = r
3
dr
dt
+
d
dt
(r
3
)r = r
3
dr
dt
+ 3r
2
r .
Exercises
1. (a) Parallel to y-axis.
(b) Length increasing in x-direction.
(c) Length not dependent on y.
5.

V = x
i
9.

V = y
i + x
j
x
y
Figure 17.7:

F (x, y) = y
j
x
y
Figure 17.8:

F (x, y) = y
i + x
j
Problems
21.

F (x, y) = a
i + b
j for any real numbers a and b is a constant vector eld. For example,

F (x, y) = 3
i 4
j .
17.4 SOLUTIONS 381
25. If

F (x, y) = f(x, y)((1 + y
2
)
i (x + y)
j ) where f(x, y) is any function, then

F

G = 0, which shows that

F is
perpendicular to

G .
For example

F (x, y) = (y + cos x)((1 + y
2
)
i (x + y)
j ).
29. One possible solution is

F (x, y) = x
i . See Figure 17.9.

3 2 1 1 2 3
3
2
1
1
2
3
x
y
Figure 17.9
33. (a) The line l is parallel to the vector v =
i 2
j 3
k . The vector eld

F is parallel to the line when

F is a multiple
of v . Taking the multiple to be 1 and solving for x, y, z we nd a point at which this occurs:
x = 1
x + y = 2
x y + z = 3
gives x = 1, y = 3, z = 7, so a point is (1, 3, 7). Other answers are possible.
(b) The line and vector eld are perpendicular if

F v = 0, that is
(x
i + (x + y)
j + (x y + z)
k ) (
i 2
j 3
k ) = 0
x 2x 2y 3x + 3y 3z = 0
4x + y 3z = 0.
One point which satises this equation is (0, 0, 0). There are many others.
(c) The equation for this set of points is 4x + y 3z = 0. This is a plane through the origin.
37. (a) Dividing a vector

F by its magnitude always produces the unit vector in the same direction as

F .
(b) Since
N = (1/F)(v
i + u
j ) = (1/F)
_
v
2
+ u
2
= (1/F)F = 1,
then

N is a unit vector. We check that

N is perpendicular to

F using the dot product of

N and

F :
N

F = (1/F)(v
i + u
j ) (u
i + v
j ) = 0.
Which side of

F does

N point? The vector

k is pointing out of the diagram. Since the cross product

k

F is
perpendicular to both
k and

F , then

N lies in the xy-plane and points at a right angle to the direction of

F . By the
right-hand rule,

N points to the left as shown in the gure.
Exercises
1. Since x
(t) = 3 and y
(t) = 0, we have x = 3t + x0 and y = y0. Thus, the solution curves are y = constant.
18 18
18
18
x
y
Figure 17.10: The eld v = 3
i
18 18
18
18
x
y
Figure 17.11: The ow y =constant
5. As
v (t) =
dx
dt
i +
dy
dt
j ,
the system of differential equations is
_
dx
dt
= x
dy
dt
= 0.
Since
d
dt
(x(t)) =
d
dt
(ae
t
) = x
and
d
dt
(y(t)) =
d
dt
(b) = 0,
the given ow satises the system. The solution curves are the horizontal lines y = b. See Figures 17.12 and 17.13.
x
y
Figure 17.12: v (t) = x
i
18 18
18
18
x
y
Figure 17.13: The ow x(t) = ae
t
, y(t) = b
9. As
v (t) =
dx
dt
i +
dy
dt
j ,
the system of differential equations is
_
dx
dt
= y
dy
dt
= x.
17.4 SOLUTIONS 383
Since
dx(t)
dt
=
d
dt
[a(e
t
+ e
t
)] = a(e
t
e
t
) = y(t)
and
dy(t)
dt
=
d
dt
[a(e
t
e
t
)] = a(e
t
+ e
t
) = x(t),
the given ow satises the system. By eliminating the parameter t in x(t) and y(t), the solution curves obtained are
x
2
y
2
= 4a
2
. See Figures 17.14 and 17.15.
x
y
Figure 17.14: v (t) = y
i + x
j
x
y
Figure 17.15: The ow
x(t) = a(e
t
+ e
t
), y(t) = a(e
t
e
t
)
Problems
13. This corresponds to area C in Figure 17.16.
0 100 200 300 400 500 600
0
100
200
300
400
500
600

km
km
E
A
D
'
C
E
B
Figure 17.16
17. (a) Perpendicularity is indicated by zero dot product. We have v grad H = (Hy
i + Hx
j ) (Hx
i + Hy
j ) = 0.
(b) If r (t) = x(t)
i + y(t)
j is a ow line we have, using the chain rule,

d
dt
H(x(t), y(t)) = Hx
dx
dt
+ Hy
dy
dt
= Hx(Hy) + Hy(Hx) = 0.
Thus H(x(t), y(t)) is constant which shows that a ow line stays on a single level curve of H.
For a different solution, use geometric reasoning. The vector eld v is tangent to the level curves of H because,
by part (a), v and the level curves are both perpendicular to the same vector eld grad H. Thus the level curves of
H and the owlines of v run in the same direction.
21. Let r (t) = x(t)
i + y(t)
j be a ow line of v . If f(x, y) has the same value at all points (x(t), y(t)) then the ow line
lies on a level curve of f. We can check whether
g(t) = f(x(t), y(t)) = x(t)
2
y(t)
2
is constant by computing the derivative g
(t). Since v = y
i + x
j , we have dx/dt = y and dy/dt = x. Thus,

g
(t) = 2x
dx
dt
2y
dy
dt
= 2xy 2yx = 0
and g is constant. This means that the ow line lies on a level curve of f. The ow lines are parameterized hyperbolas
with equation x
2
y
2
= c.
Exercises
1. A horizontal disk of radius 5 in the plane z = 7.
5. Since z = r =
_
x
2
+ y
2
, we have a cone around the z-axis. Since 0 r 5, we have 0 z 5, so the cone has
height and maximum radius of 5.
9. The top half of the sphere (z 0).
17.5 SOLUTIONS 385
Problems
13. Two vectors in the plane containing P = (0, 0, 0), Q = (1, 2, 3), and R = (2, 1, 0) are the displacement vectors
v 1 =

PQ =
i + 2
j + 3
k
v 2 =

PR = 2
i +
j .
Letting r 0 = 0
i + 0
j + 0
k =
0 we have the parameterization

r (s, t) = r 0 + sv 1 + tv 2
= (s + 2t)
i + (2s + t)
j + 3s
k .
17. To parameterize the plane we need two nonparallel vectors v 1 and v 2 that are parallel to the plane. Such vectors are
perpendicular to the normal vector to the plane, n =

i + 2
j + 3
k . We can choose any vectors v 1 and v 2 such that

v 1 n = v 2 n = 0.
One choice is
v 1 = 2
i j v 2 = 3
k .
Letting r 0 = 5
i +
j + 4
k we have the parameterization

r (s, t) = r 0 + sv 1 + tv 2
= (5 + 2s + 3t)
i + (1 s)
j + (4 t)
k .
21. Since you walk 5 blocks east and 1 block west, you walk 5 blocks in the direction of v 1, and 1 block in the opposite
direction. Thus,
s = 5 1 = 4,
Similarly,
t = 4 2 = 2.
Hence
x
i + y
j + z
k = (x0
i + y0
j + z0
k ) + 4 v1 + 2 v2
= (x0
i + y0
j + z0
k ) + 4(2
i 3
j + 2
k ) + 2(
i + 4
j + 5
k )
= (x0 + 10)
i + (y0 4)
j + (z0 + 18)
k .
Thus the coordinates are:
x = x0 + 10, y = y0 4, z = z0 + 18.
25. Set up the coordinates as in Figure 17.17. The surface is the revolution surface obtained by revolving the curve shown in
Figure 17.18 about the z axis. From the measurements given, we obtain the equation of the curve in Figure 17.18:
x = cos
_
3
z
_
+ 3, 0 z 48
(a) Rotating this around the z-axis, and taking z = t as the parameter, we get the parametric equations
x = (cos
_
3
t
_
+ 3) cos
y = (cos
_
3
t
_
+ 3) sin
z = t 0 2, 0 t 48
(b) We know that the points in the curve consists of cross-sections of circles parallel to the xy plane and of radius
cos((/3)z + 3). Thus,
Area of cross-section =
_
cos
_
3
z + 3
__
2
Integrating over z, we get
Volume =
_
48
0
_
cos

3
z + 3
_
2
dz
=
_
48
0
_
cos
2

3
z + 6 cos

3
z + 9
_
dz
= 456 in
3
.
x y
z
Figure 17.17
T
c
6
E '
2
E '
4
x
z
3
Figure 17.18
29. Let (, /2) be the original coordinates. If < , then the new coordinates will be ( + , /4). If , then the new
coordinates will be ( , /4).
33. The vase obtained by rotating the curve z = 10
x 1, 1 x 2, around the z-axis is shown in Figure 17.19. At

height z, the cross-section is a horizontal circle of radius a. Thus, a point on this horizontal circle is given by
r = a cos
i + a sin
j + z
k .
However, the radius a varies, so we need to express it in terms of the other parameters and z. If you look at the xz-plane,
the radius of this circle is given by x, so solving for x in z = 10
x 1 gives
a = x =
_
z
10
_
2
+ 1.
Thus, a parameterization is
r =
_
_
z
10
_
2
+ 1
_
cos
i +
_
_
z
10
_
2
+ 1
_
sin
j + z
k
so
x =
_
_
z
10
_
2
+ 1
_
cos , y =
_
_
z
10
_
2
+ 1
_
sin , z = z,
where 0 2, 0 z 10.
2
2
x
y
z
10
z = 10
x 1
Figure 17.19
x
y
z
Figure 17.20: The surface x = s + t,
y = s t, z = s
2
+ t
2
for 0 s 1,
0 t 1
SOLUTIONS to Review Problems for Chapter Seventeen 387
37. (a) From the rst two equations we get:
s =
x + y
2
, t =
x y
2
.
Hence the equation of our surface is:
z =
_
x + y
2
_
2
+
_
x y
2
_
2
=
x
2
2
+
y
2
2
,
which is the equation of a paraboloid.
The conditions: 0 s 1, 0 t 1 are equivalent to: 0 x + y 2, 0 x y 2. So our surface is
dened by:
z =
x
2
2
+
y
2
2
, 0 x + y 2 0 x y 2
(b) The surface is shown in Figure 17.20.
Exercises
1. The line has equation
r = 2
j + 3
k + t(5
i + 4
k ),
or, equivalently
x = 2 + 5t
y = 1 + 4t
z = 3 t.
5. The parameterization x
i + y
j = (4 + 4 cos t)
i + (4 + 4 sin t
) j gives the correct circle, but starts at (8, 4). To start on

the x-axis we need
x
i + y
j = (4 + 4 cos(t

2
))
i + (4 + 4 sin(t

2
))
j = (4 + 4 sin t)
i + (4 4 cos t)
j .
9. Since the vector n = grad(2x 3y + 5z) = 2
i 3
j + 5
k is perpendicular to the plane, this vector is parallel to the

line. Thus the equation of the line is
x = 1 + 2t, y = 1 3t, z = 1 + 5t.
13. See Figure 17.21. The parameterization is
r = 10 cos
_
2t
30
_
i 10 sin
_
2t
30
_
j + 7
k .
x
y
z
(0, 0, 7)
*
Figure 17.21
v =
d
dt
(2 + 3t
2
)
i +
d
dt
(4 + t
2
)
j +
d
dt
(1 t
2
)
k = 6t
i + 2t
j 2t
k .
21. Vector. Differentiating using the chain rule gives
Velocity =
_
cos t
2
3 + sin t
_
i +
_
sin t
2
3 + cos t
_
j .
25. The direction vectors of the lines,
i +4
j 2
k and 2
i 8
j +4
k , are multiplies of each other (the second is 2 times

the rst). Thus the lines are parallel. To see if they are the same line, we take the point corresponding to t = 0 on the rst
line, which has position vector 3
i + 3
k , and see if it is on the second line. So we solve

(1 + 2t)
i + (11 8t)
j + (4t 5)
k = 3
i + 3
k .
This has solution t = 1, so the two lines have a point in common and must be the same line, parameterized in two different
ways.
29. The vector eld points in a clockwise direction around the origin. Since
_
y
_
x
2
+ y
2
_
i
_
x
_
x
2
+ y
2
_
j =
_
x
2
+ y
2
_
x
2
+ y
2
= 1,
the length of the vectors is constant everywhere.
x
y
Figure 17.22
Problems
33. (a) Avector eld associates a vector to every point in a region of the space. In other words, a vector eld is a vector-valued
function of position given by v =

f (r ) =

f (x, y, z)
(b) (i) Yes, r +a = (x + a1)
i + (y + a2)
j + (z + a3)
k is a vector-valued function of position.

(ii) No, r a is a scalar.
(iii) Yes.
(iv) x
2
+ y
2
+ z
2
is a scalar.
37. The displacement vector from (1, 1, 1) to (2, 1, 3) is

d = (2
j +3
k ) (
i +
j +
k ) =
i 2
j +2
k meters. The
velocity vector has the same direction as

d and is given by
v =
d
5
= 0.2
i 0.4
j + 0.4
k meters/sec.
Since v is constant, the acceleration a =

0 .
SOLUTIONS to Review Problems for Chapter Seventeen 389
41. (a) In order for the particle to stop, its velocity v = (dx/dt)
i + (dy/dt)
j must be zero, so we solve for t such that

dx/dt = 0 and dy/dt = 0, that is
dx
dt
= 3t
2
3 = 3(t 1)(t + 1) = 0,
dy
dt
= 2t 2 = 2(t 1) = 0.
The value t = 1 is the only solution. Therefore, the particle stops when t = 1 at the point (t
3
3t, t
2
2t)|t=1 =
(2, 1).
(b) In order for the particle to be traveling straight up or down, the x-component of the velocity vector must be 0. Thus,
we solve dx/dt = 3t
2
3 = 0 and obtain t = 1. However, at t = 1 the particle has no vertical motion, as we saw
in part (a). Thus, the particle is moving straight up or down only when t = 1. Since the velocity at time t = 1 is
v (1) =
dx
dt
t=1
i +
dy
dt
t=1
j = 4
j ,
the motion is straight down. The position at that time is (t
3
3t, t
2
2t)|t=1 = (2, 3).
(c) For horizontal motion we need dy/dt = 0. That happens when dy/dt = 2t 2 = 0, and so t = 1. But from part (a)
we also have dx/dt = 0 also at t = 1, so the particle is not moving at all when t = 1. Thus, there is no time when
the motion is horizontal.
45. (a) fx =
_
2x(x
2
+ y
2
) 2x(x
2
y
2
)
(x
2
+ y
2
)
2
=
4xy
2
(x
2
+ y
2
)
2
.
fy =
_
2y(x
2
+ y
2
) 2y(x
2
y
2
)
(x
2
+ y
2
)
2
=
4yx
2
(x
2
+ y
2
)
2
.
f(1, 1) =
j , i.e., south-east.
(b) We need a vector u such that f(1, 1) u = 0, i.e., such that (
j ) u = 0. The vector u =
i +
j clearly works;
so does u =
j . Dividing by the length to get a unit vector, we have u =

1
i +
1
j or u =
1
i
1
j .
(c) f is a function of x and y, which in turn are functions of t. Thus, the chain rule can be used to show how f changed
with t.
df
dt
=
f
x

dx
dt
+
f
y

dy
dt
=
4xy
2
(x
2
+ y
2
)
2
2e
2t
4x
2
y
(x
2
+ y
2
)
2
(6t
2
+ 6).
At t = 0, x = 1, y = 1; so,
df
dt
=
4
4
2
4
4
6 = 4.
49. Note that uniform circular motion is possible with the given conditions, since v a = 0 shows that the velocity and
acceleration vectors are perpendicular.
For uniform circular motion in an orbit of radius R, we have a = v
2
/R. Thus R = v
2
/a = 52/
13
for both parts (a) and (b).
The center of the orbit is at distance R in the direction of the acceleration vector from the point P on the orbit. The
vector
R
a
a
=
52
13
2
i + 3
13
= 8
i + 12
j
thus extends from the point P to the center of the orbit.
(a) The center of the orbit is at the point (0 + 8, 0 + 12) = (8, 12)
(b) The center of the orbit is at the point (10 + 8, 50 + 12) = (18, 62)
53. At time t the particle has polar coordinates r = r (t) = at and = t. At time t, the ray from the origin to the
particle is at angle t radians from the positive x-axis. The ray is therefore rotating at a rate of radians per unit time.
The parameter is the rate of change of the polar angle of the particle measured in radians per unit time. The larger
is, the quicker the particle completes a complete revolution (a 360
trip) around the origin. At time t, the particle is at

distance at from the origin. Thus a equals the rate of change of the particles distance from the origin. The larger a is, the
faster the particle moves away from the origin.
57. (a) The cone of height h, maximum radius a, vertex at the origin and opening upward is shown in Figure 17.23.
y
x
z
h
a
Figure 17.23
By similar triangles, we have
r
z
=
a
h
,
so
z =
hr
a
.
Therefore, one parameterization is
x = r cos , 0 r a,
y = r sin , 0 < 2,
z =
hr
a
.
(b) Since r = az/h, we can write the parameterization in part (a) as
x =
az
h
cos , 0 z h,
y =
az
h
sin , 0 < 2,
z = z.
61. The parameterization for a sphere of radius a using spherical coordinates is
x = a sin cos , y = a sin sin , z = a cos .
Think of the ellipsoid as a sphere whose radius is different along each axis and you get the parameterization:
_
_
x = a sin cos , 0 ,
y = b sin sin , 0 2,
z = c cos .
To check this parameterization, substitute into the equation for the ellipsoid:
x
2
a
2
+
y
2
b
2
+
z
2
c
2
=
a
2
sin
2
cos
2
a
2
+
b
2
sin
2
sin
2
b
2
+
c
2
cos
2
c
2
= sin
2
(cos
2
+ sin
2
) + cos
2
= 1.
65. (a)
r

F = (x
i + y
j ) (y(1 y
2
)
i + x(1 y
2
)
j )
= xy(1 y
2
) + yx(1 y
2
) = 0
This means that the tangent line to the ow line at a point is always perpendicular to the vector from the origin to that
point. Hence the ow lines are circles centered at the origin.
(b) The circle r (t) = cos t
i + sin t
j has velocity vector v (t) = sin t
i + cos t
j = y
i + x
j = (1 y
2
)
F . Thus
the velocity vector is a scalar multiple of

F , and hence parallel to

F . However, since v (t) is not equal to

F (r (t)),
it is not a ow line.
(c) Using a CAS, we nd
v (t) =
t
(1 + t
2
)
3/2
i +
_
t
2
(1 + t
2
)
3/2
+
1
1 + t
2
_
j =
t
(1 + t
2
)
3/2
i +
1
(1 + t
2
)
3/2
j
and
F (r (t)) =
_
_
t
_
1
t
2
1+t
2
_
1 + t
2
_
_
i +
1
t
2
1+t
2
1 + t
2
j =
t
(1 + t
2
)
3/2
i +
1
(1 + t
2
)
3/2
j = v (t).
Although the circle parameterized in part (b) has velocity vectors parallel to

F at each point of the circle, its speed
is not equal to the magnitude of the vector eld. The circle in part (c) is parameterized at the correct speed to be the
ow line.
1. False. The y coordinate is zero when t = 0, but when t = 0 we have x = 2 so the curve never passes through (0, 0).
5. False. When t = 0, we have (x, y) = (0, 1). When t = /2, we have (x, y) = (1, 0). Thus the circle is being traced
out clockwise.
9. True. To nd an intersection point, we look for values of s and t that make the coordinates in the rst line the same as the
coordinates in the second. Setting x = t and x = 2s equal, we see that t = 2s. Setting y = 2 +t equal to y = 1 s, we
see that t = 1s. Solving both t = 2s and t = 1s yields t =
2
3
, s =
1
3
. These values of s and t will give equal
x and y coordinates on both lines. We need to check if the z coordinates are equal also. In the rst line, setting t =
2
3
gives z =
7
3
. In the second line, setting s =
1
3
gives z =
1
3
. As these are not the same, the lines do not intersect.
13. False. The velocity vector is v (t) = r

(t) = 2t
j . Then v (1) = 2
j and v (1) = 2
j , which are not

equal.
17. False. As a counterexample, consider the curve r (t) = t
2
i + t
2
j for 0 t 1. In this case, when t is replaced by t,

the parameterization is the same, and is not reversed.
21. True, since the vectors x
j are parallel to the y-axis.

25. True. Any ow line which stays in the rst quadrant has x, y .
29. True. If (x, y) were a point where the y-coordinate along a ow line reached a relative maximum, then the tangent vector
to the ow line, namely

F (x, y), there would have to be horizontal (or

0 ), that is its
j component would have to be 0.

But the
j component of

F is always 2.
33. False. There is only one parameter, s. The equations parameterize a line.
37. True. If the surface is parameterized by r (s, t) and the point has parameters (s0, t0) then the parameter curves r (s0, t)
and r (s, t0).
41. False. Suppose r (t) = t
i + t
j . Then r

(t) =
i +
j and
r

(t) r (t) = (t
i + t
j ) (
i +
j ) = 2t.
So r

(t) r (t) = 0 for t = 0.
18.1 SOLUTIONS 393
CHAPTER EIGHTEEN
Exercises
1. Positive, because the vectors are longer on the portion of the path that goes in the same direction as the vector eld.
5. Negative, because the vector eld points in the opposite direction to the path.
9. Since

F is a constant vector eld and the curve is a line,
_
C
F dr =

F r , where r = 7
j . Therefore,
_
C
F dr = (3
i + 4
j ) 7
j = 28
13. At every point, the vector eld is parallel to segments r = x
i of the curve. Thus,

_
C
F dr =
_
6
2
x
i dx
i =
_
6
2
xdx =
x
2
2
6
2
= 16.
17. Since the curve is along the y-axis, only the
j component of the vector eld contributes to the integral:

_
C
(2
j + 3
k ) dr =
_
C
2
j dr = 2 Length of C = 2 10 = 20.
Problems
21. Since it appears that C1 is everywhere perpendicular to the vector eld, all of the dot products in the line integral are zero,
hence
_
C
1
F dr 0. Along the path C2 the dot products of

F with ri are all positive, so their sum is positive and
we have
_
C
1
F dr <
_
C
2
F dr . For C3 the vectors ri are in the opposite direction to the vectors of

F , so the dot
products

F ri are all negative; so,
_
C
3
F dr < 0. Thus, we have

_
C
3
F dr <
_
C
1
F dr <
_
C
2
F dr
25. The line integral along C1 is negative, the line integral along C2 is negative, and the line integral along C3 appears to be
zero.
29. The vector eld

F is in the same direction as C if b > 0, so we want b < 0. No restriction is needed on c.
33. This vector eld is illustrated in Figure 18.1. It is perpendicular to C2 and C4 at every point, since

F (x, y) r (x, y) = 0
and C2 and C4 are radial line segments, then
_
C
2
F dr =
_
C
4
F dr = 0.
Since C3 is longer than C1, and the vector eld is larger in magnitude along C3, the line integral along C3 has greater
absolute value than that along C1. The line integral along C3 is positive and the line integral along C1 is negative, so
_
C
F dr =
_
C
3
F dr +
_
C
1
F dr > 0.
See Figure 18.1.
394 Chapter Eighteen /SOLUTIONS
1 2 3
2
1
1
2
x
y
C
4
C
2
C
3
C
1
Figure 18.1
x
y
(i)
x
y
(ii)
x
y
(iii)
x
y
(iv)
Figure 18.2
(b) For (i) and (iii) a closed curve can be drawn; not for the others.
41. See Figure 18.3. The example chosen is the vector eld

F (x, y) = y
j and the path C is the line from (0, 1) to (0, 1).

Since the vectors are symmetric about the x-axis, the dot products

F r cancel out along C to give 0 for the line
integral. Many other answers are possible.
18.2 SOLUTIONS 395
(0, 1)
(0, 1)
x
y
C
Figure 18.3
45. Let r = r . Since r points outward, in the opposite direction to

F , we expect the answer to be negative.
_
C
F dr =
_
C
GMmr
r
3
dr =
_
10000
8000
GMm
r
2
dr
=
GMm
r
10000
8000
= GMm
_
1
10000

1
8000
_
= 2.5 10
5
GMm.
49. In Problem 48 we saw that the surface where the potential is zero is a sphere of radius a. Let S be any sphere centered at
the origin, and let P1 be a point on S, and C1 a path from P0 to P1. If P is any point on S, then P can be reached from
P0 by a path, C, consisting of C1 followed by C2, where C2 is a path from P1 to P lying entirely on the sphere, S. Then
_
C
2
E dr = 0, since

E is perpendicular to the sphere. So
(P) =
_
C
E dr =
_
C
1
E dr
_
C
2
E dr =
_
C
1
E dr = (P1).
Thus, is constant on S. The equipotential surfaces are spheres centered at the origin.
Exercises
1. Only the
i -component contributes to the line integral, so dr =
i dx and
_
C
(2x
i + 3y
j ) dr =
_
(5,0,0)
(1,0,0)
(2x
i + 3y
j )
i dx =
_
5
1
2xdx = x
2
5
1
= 24.
5. Since

F = (x
2
+y)
i +y
3
j , the line integral along the third segment, which is parallel to the z-axis, is zero. On the rst
segment, which is parallel to the y-axis, only the
j -component contributes. On the second segment, which is parallel to

the x-axis, only the
i -component contributes. On the rst segment x = 4 and y varies from 0 to 3; on the second segment
y = 3 and x varies from 4 to 0. Thus, we have
_
C
F dr =
_
3
0
((4
2
+ y)
i + y
3
j )
j dy +
_
0
4
((x
2
+ 3)
i + 3
3
j )
i dx
=
_
3
0
y
3
dy +
_
0
4
(x
2
+ 3) dx =
y
4
4
3
0
_
x
3
3
+ 3x
_
4
0
=
81
4

64
3
12 =
157
12
.
9. The line can be parameterized by (1 + 2t, 2 + 2t), for 0 t 1, so the integral looks like
_
C
F dr =
_
1
0
F (1 + 2t, 2 + 2t) (2
i + 2
j ) dt
=
_
1
0
[(1 + 2t)
2
i + (2 + 2t)
2
j ] (2
i + 2
j ) dt
=
_
1
0
2(1 + 4t + 4t
2
) + 2(4 + 8t + 4t
2
) dt
=
_
1
0
(10 + 24t + 16t
2
) dt
= (10t + 12t
2
+ 16t
3
/3)
1
0
= 10 + 12 + 16/3 (0 + 0 + 0) = 82/3
13. The curve C is parameterized by
r = cos t
i + sin t
j , for 0 t 2,
so,
r

(t) = sin t
i + cos t
j .
Thus,
_
C
F dr =
_
2
0
(2 sin t
i sin (sin t)
j ) (sin t
i + cos t
j )dt
=
_
2
0
(2 sin
2
t sin (sin t) cos t)dt
= sin t cos t t + cos (sin t)
2
0
= 2.
17. Since r = x(t)
i + y(t)
j + z(t)
k = t
i + t
2
j + t
3
k , for 1 t 2,
we have r

(t) = x
(t)
i + y
(t)
j + z
(t)
k =
i + 2t
j + 3t
2
k . Then
_
C
F dr =
_
2
1
(t
i + 2t
3
t
2
j + t
k ) (
i + 2t
j + 3t
2
k ) dt
=
_
2
1
(t + 4t
6
+ 3t
3
) dt
=
t
2
2
+
4t
7
7
+
3t
4
4
2
1
=
2389
28
85.32
21.
_
C
3xdx y sin xdy
25. From x = t
2
and y = t
3
we get dx = 2tdt and dy = 3t
2
dt. Hence
_
C
ydx + xdy =
_
5
1
t
3
(2t)dt + t
2
(3t
2
)dt =
_
5
1
5t
4
dt = 5
5
1 = 3124.
18.2 SOLUTIONS 397
Problems
29. (a) Figure 18.4 shows the curves.
1
1
C
2
C
1
x
y
Figure 18.4
(b) On C1, only the
j -component of

F contributes to the integral. There dr =
j dy, so
_
C
1
F dr =
_
1
1
y
j dy =
_
1
1
y dy =
y
2
2
1
1
= 0.
On C2, we have r

(t) = sin t
i + cos t
j , so
_
C
2
F dr =
_
3/2
/2
((cos t + 3 sin t)
i + sin t
j ) (sin t
i + cos t
j ) dt
=
_
3/2
/2
cos t sin t 3 sin
2
t + cos t sin t dt =
_
3/2
/2
3 sin
2
t dt
= 3
_
t
2

sin t cos t
2
_
3/2
/2
=
3
2
.
33. (a) The line integral
_
C
(xy
i + x
j ) dr is positive. This follows from the fact that all of the vectors of xy
i + x
j
at points along C point approximately in the same direction as C (meaning the angles between the vectors and the
direction of C are less than /2).
(b) Using the parameterization x(t) = t, y(t) = 3t, with x
(t) = 1, y
(t) = 3, we have
_
C
F dr =
_
4
0
F (t, 3t) (
i + 3
j ) dt
=
_
4
0
(3t
2
i + t
j ) (
i + 3
j ) dt
=
_
4
0
(3t
2
+ 3t) dt
=
_
t
3
+
3
2
t
2
_
4
0
= 88.
(c) Figure 18.5 shows the oriented path C
, with the turn around points P and Q. The particle rst travels from the
origin to the point P (call this path C1), then backs up from P to Q (call this path C2), then goes from Q to the point
(4, 12) in the original direction (call this path C3). See Figure 18.6. Thus, C
= C1 + C2 + C3. Along the parts of

C1 and C2 that overlap, the line integrals cancel, so we are left with the line integral over the part of C1 that does not
overlap with C2, followed by the line integral over C3. Thus, the line integral over C
is the same as the line integral

over the direct route from the point (0, 0) to the point (4, 12).
Q
P
(4, 12)
1 2 3 4
0
1
2
3
4
5
6
7
8
9
10
11
12
x
y
Figure 18.5
C
1
C
2
C
3
Figure 18.6
(d) The parameterization
(x(t), y(t)) =
_
1
3
(t
3
6t
2
+ 11t), (t
3
6t
2
+ 11t)
_
has (x(0), y(0)) = (0, 0) and (x(4), y(4)) = (4, 12). The form of the parameterization we were given shows that
the second coordinate is always three times the rst. Thus all points on the parameterized curve lie on the line y = 3x.
We have to do a bit more work to guarantee that all points on the curve lie on the line between the point (0, 0)
and the point (4, 12); it is possible that they might shoot off to, say, (100, 300) before returning to (4, 12). Lets
investigate the maximum and minimum values of f(t) = t
3
6t
2
+11t on the interval 0 t 4. We can do this on
a graphing calculator or computer, or use single-variable calculus. We already know the values of f at the endpoints,
namely 0 and 12. Well look for local extrema:
0 = f
(t) = 3t
2
12t + 11
which has roots at t = 2
1
3
. These are the values of t where the particle changes direction: t = 2
1
3
corresponds to point P and t = 2+
1
3
corresponds to point Q of C
. At these values of t we have f(2

1
3
) 6.4,
and f(2 +
1
3
) 5.6. The fact that these values are between 0 and 12 shows that f takes on its maximum and
minimum values at the endpoints of the interval and not in between.
(e) Using the parameterization given in part (d), we have
r

(t) = x
(t)
i + y
(t)
j =
1
3
(3t
2
12t + 11)
i + (3t
2
12t + 11)
j .
Thus,
_
C
F dr
=
_
4
0
F (
1
3
(t
3
6t
2
+ 11t), t
3
6t
2
+ 11t) (
1
3
(3t
2
12t + 11)
i + (3t
2
12t + 11)
j ) dt
=
_
4
0
(
1
3
(t
3
6t
2
+ 11t)
2
i +
1
3
(t
3
6t
2
+ 11t)
j ) (
1
3
(3t
2
12t + 11)
i + (3t
2
12t + 11)
j ) dt
=
_
4
0
1
3
(t
3
6t
2
+ 11t)(3t
2
12t + 11)
_
((t
3
6t
2
+ 11t)
i +
j ) (
1
3
i +
j )
_
dt
=
_
4
0
1
3
(t
3
6t
2
+ 11t)(3t
2
12t + 11)
_
1
3
(t
3
6t
2
+ 11t) + 1
_
dt
=
1
9
_
4
0
(t
3
6t
2
+ 11t)(3t
2
12t + 11)(t
3
6t
2
+ 11t + 3) dt
Numerical integration yields an answer of 88, which agrees with the answer found in part b).
18.3 SOLUTIONS 399
37. The integral corresponding to A(t) = (t, t) is
_
1
0
3t dt.
The integral corresponding to D(t) = (e
t
1, e
t
1) is
3
_
ln 2
0
(e
2t
e
t
) dt.
The substitution s = e
t
1 has ds = e
t
dt. Also s = 0 when t = 0 and s = 1 when t = ln 2. Thus, substituting into the
integral corresponding to D(t) and using the fact that e
2t
= e
t
e
t
gives
3
_
ln 2
0
(e
2t
e
t
) dt = 3
_
ln 2
0
(e
t
1)e
t
dt =
_
1
0
3s ds.
The integral on the right-hand side is the same as the integral corresponding to A(t). Therefore we have
3
_
ln 2
0
(e
2t
e
t
) dt =
_
1
0
3s ds =
_
1
0
3t dt.
Alternatively, the substitution t = e
w
1 converts the integral corresponding to A(t) into the integral corresponding to
B(t).
Exercises
1. Since

F is a gradient eld, with

F = gradf where f(x, y) = x
2
+ y
4
, we use the Fundamental Theorem of Line
Integrals. The starting point of the path C is (2, 0) and the end is (0, 2). Thus,
_
C
F dr = f(0, 2) f(2, 0) = 16 4 = 12.

5. Path-independent, because the vector eld appears constant.
9. Since

F = 3x
2
i + 4y
3
j = grad(x
3
+ y
4
), we take f(x, y) = x
3
+ y
4
. Then by the Fundamental Theorem of Line
Integrals,
_
C
F dr = f(1, 0) f(1, 0) = (1)

3
1
3
= 2.
13. Since

F = y sin(xy)
i + xsin(xy)
j = grad(cos(xy)), the Fundamental Theorem of Line Integrals gives

_
C
F dr = cos(xy)
(3,18)
(1,2)
= cos(54) + cos(2) = cos(2) cos(54).
17. Since

F = 2xy
2
ze
x
2
y
2
z
i +2x
2
yze
x
2
y
2
z
j +x
2
y
2
e
x
2
y
2
z
k = grad(e
x
2
y
2
z
) and the curve C is closed, the Fundamental
Theorem of Line Integrals tells us that
_
C
F dr = 0, since
_
C
F dr = e
x
2
y
2
z
(1,0,1)
(1,0,1)
= e
0
e
0
= 0.
Problems
21. Yes. If f(x, y) =
1
2
x
2
, then grad f = x
i .
25. (a) To nd the change in f by computing a line integral, we rst choose a path C between the points; the simplest is a
line. We parameterize the line by (x(t), y(t)) = (t, t/2), with 0 t 1. Then (x
(t), y
(t)) = (1, /2), so the

Fundamental Theorem of Line Integrals tells us that
f(1,

2
) f(0, 0) =
_
C
grad f dr
=
_
1
0
grad f
_
t,
t
2
_
i +

2
j
_
dt
=
_
1
0
_
2te
t
2
sin
_
t
2
_
i + e
t
2
cos
_
t
2
_
j
_
i +

2
j
_
dt
=
_
1
0
_
2te
t
2
sin
_
t
2
_
+
e
t
2
2
cos
_
t
2
_
_
dt
=
_
1
0
d
dt
_
e
t
2
sin
_
t
2
__
dt
= e
t
2
sin
_
t
2
_
1
0
= e = 2.718.
This integral can also be approximated numerically.
(b) The other way to nd the change in f between these two points is to rst nd f. To do this, observe that
2xe
x
2
sin y
i + e
x
2
cos y
j =

x
_
e
x
2
sin y
_
i +

y
_
e
x
2
sin y
_
j = grad
_
e
x
2
sin y
_
.
So one possibility for f is f(x, y) = e
x
2
sin y. Thus,
Change in f
(1,/2)
(0,0)
= e
x
2
sin y
(1,/2)
(0,0)
= e
1
sin
_
2
_
e
0
sin 0 = e.
The exact answer conrms our calculations in part (a) which show that the answer is e.
29. This vector eld is not a gradient eld, so we evaluate the line integral directly. Let C1 be the path along the x-axis from
(0, 0) to (3, 0) and let C2 be the path from (3, 0) to (3/
2, 3/
2) along x
2
+ y
2
= 9. Then
_
C
H dr =
_
C
1
H dr +
_
C
2
H dr .
On C1, the vector eld has only a
j component (since y = 0), and

H is therefore perpendicular to the path. Thus,
_
C
1
H dr = 0.
On C2, the vector eld is tangent to the path. The path is one eighth of a circle of radius 3 and so has length 2(3/8) =
3/4.
_
C
2
H dr =
H Length of path = 3
_
3
4
_
=
9
4
.
Thus,
_
C
H dr =
9
4
.
33. Although this curve is complicated, the vector eld is a gradient eld since
F = sin
_
x
2
_
sin
_
y
2
_
i cos
_
x
2
_
cos
_
y
2
_
j = grad
_
2 cos
_
x
2
_
sin
_
y
2
__
.
18.3 SOLUTIONS 401
Thus, only the endpoints of the curve, P and Q, are needed. Since P = (3/2, 3/2) and Q = (3/2, 3/2) and
F = grad(2 cos(x/2) sin(y/2)), we have

_
C
F dr = 2 cos
_
x
2
_
sin
_
y
2
_
Q=(3/2,3/2)
P=(3/2,3/2)
= 2 cos
_
3
4
_
sin
_
3
4
_
+ 2 cos
_
3
4
_
sin
_
3
4
_
= 2 cos
_
3
4
_
sin
_
3
4
_
+ 2 cos
_
3
4
_
sin
_
3
4
_
= 2
1
2
2
1
2
= 2.
37. (a) By the Fundamental Theorem of Line Integrals
_
(3,4)
(0,2)
grad f dr = f(3, 4) f(0, 2) = 66 57 = 9.
(b) By the Fundamental Theorem of Line Integrals, since C is a closed path,
_
C
grad f dr = 0.
41. (a) Work done by the force is the line integral, so
Work done against force =
_
C
F dr =
_
C
(mg
k ) dr .
Since r = (cos t)
i + (sin t)
j + t
k , we have r

= (sin t)
i + (cos t)
j +
k ,
Work done against force =
_
2
0
mg
k (sin t
i + cos t
j +
k )dt
=
_
2
0
mg dt = 2 mg.
(b) We know from physical principles that the force is conservative. (Because the work done depends only on the vertical
distance moved, not on the path taken.) Alternatively, we see that
F = mg
k = grad(mgz),
so

F is a gradient eld and therefore path independent, or conservative.
45. (a) We have
grad h =
i
grad = 2y
i + 2x
F grad = y
i = y grad h.
Thus,

F grad is a multiple of grad h.
(b) By part (a) the vector elds

F and grad have the same components perpendicular to grad h, which is to say the
same components in the direction of the level curve C of h. Thus, the line integrals of

F and grad along C are
equal. Using the Fundamental Theorem of Calculus for Line Integrals, we have
_
C
F dr =
_
C
grad dr = (Q) (P) = 60 30 = 30.
49. (a) By the chain rule
dh
dt
=
f
x
dx
dt
+
f
y
dy
dt
= fxx
(t) + fyy
(t),
which is the result we want.
(b) Using the parameterization of C that we were given,
_
C
grad f dr =
_
b
a
(fx(x(t), y(t))
i + fy(x(t), y(t))
j ) (x
(t)
i + y
(t)
j )dt
=
_
b
a
(fx(x(t), y(t))x
(t) + fy(x(t), y(t))y
(t))dt.
Using the result of part (a), this gives us
_
C
grad f dr =
_
b
a
h
(t)dt
= h(b) h(a) = f(Q) f(P).
Exercises
1. We know that
f
x
= 2xy and
f
y
= x
2
,
so, integrating with respect to x, thinking of y as a constant gives
f(x, y) = x
2
y + C(y).
Differentiating with respect to y gives
f
y
= x
2
+ C
(y),
so we take C(y) = k for some constant K
2
. Thus
f(x, y) = x
2
y + K.
5. Yes, since

F = 2xy
i + x
2
j = grad(x
2
y).
9. The domain of the vector eld

F = (2xy
3
+ y)
i + (3x
2
y
2
+ x)
j is the whole xy-plane. We apply the curl test:

F1
y
= 6xy
2
+ 1 =
F2
x
so

F is the gradient of a function f. In order to compute f we rst integrate
f
x
= 2xy
3
+ y
with respect to x thinking of y as a constant. We get
f(x, y) = x
2
y
3
+ xy + C(y)
Differentiating with respect to y and using the fact that f/y = 3x
2
y
2
+ x gives
f
y
= 3x
2
y
2
+ x + C
(y) = 3x
2
y
2
+ x
Thus C
(y) = 0 so C is constant and

f(x, y) = x
2
y
3
+ xy + C.
18.4 SOLUTIONS 403
13. We have
F1
y
=
(x
2
+ y
2
)1 y(2y)
(x
2
+ y
2
)
2
=
x
2
y
2
(x
2
+ y
2
)
2
F2
x
=
(x
2
+ y
2
)1 x(2x)
(x
2
+ y
2
)
2
=
y
2
x
2
(x
2
+ y
2
)
2
=
x
2
y
2
(x
2
+ y
2
)
2
.
Thus
F1
y
=
F2
x
. However, the domain of the vector eld contains a hole at the origin, so the curl test does not apply.
This is not a gradient eld. See Example 7 on page 957 of the text.
17. By Greens Theorem, with R representing the interior of the circle,
_
C
F dr =
_
R
_

x
(xy)

y
(3y)
_
dA =
_
R
(y 3) dA.
The integral of y over the interior of the circle is 0, by symmetry, because positive contributions of y from the top half of
the circle cancel those from the bottom half. Thus
_
R
y dA = 0.
So
_
C
F dr =
_
R
(y 3) dA =
_
R
3 dA = 3 Area of circle = 3 (1)
2
= 3.
Problems
21. The curve is closed, so we can use Greens Theorem. If R represents the interior of the region
_
C
F dr =
_
R
_
F2
x

F1
y
_
dA =
_
R
_
(x)
x

(x y)
y
_
dA
=
_
R
(1 (1)) dA =
_
R
2 dA = 2 Area of sector.
Since R is 1/8 of a circle, R has area (3
2
)/8. Thus,
_
C
F dr = 2
9
8
=
9
4
.
25. Since

F = x
j , we have F2/x = 1 and F1/y = 0. Thus, using Greens Theorem if R is the region enclosed by the
closed curve C, we have
_
C
F dr =
_
R
_
F2
x

F1
y
_
dxdy =
_
R
1 dxdy = Area of R
29. (a) The curve, C, is closed and oriented in the correct direction for Greens Theorem. See Figure 18.7. Writing R for the
interior of the circle, we have
_
C
_
(x
2
y)
i + (y
2
+ x)
j
_
dr =
_
R
_
(y
2
+ x)
x

(x
2
y)
y
_
dxdy
=
_
R
(1 (1)) dxdy = 2
_
R
dxdy
= 2 Area of circle = 2( 3
2
) = 18.
(b) The circle given has radius R and center (a, b). The argument in part (a) works for any circle of radius R, oriented
counterclockwise. So the line integral has the value 2R
2
.
5
4 R C
(x 5)
2
+ (y 4)
2
= 9
x
y
Figure 18.7
33. (a) We see that

F ,

G ,

H are all gradient vector elds, since
grad(xy) =

F for all x, y
grad(arctan(x/y)) =

G except where y = 0
grad
_
(x
2
+ y
2
)
1/2
_
=

H except at (0, 0).
Other answer are possible. For example grad(arctan(y/x)) =

G for x = 0.
(b) Parameterizing the unit circle, C, by x = cos t, y = sin t, 0 t , we have r

(t) = sin t
i + cos t
j , so
_
C
F dr =
_
2
0
((sin t)
i + (cos t)
j ) ((sin t)
i + (cos t)
j ) dt =
_
2
0
cos(2t) dt = 0.
The vector eld

G is tangent to the circle, pointing in the opposite direction to the parameterization, and of length 1
everywhere. Thus
_
C
G dr = 1 Length of circle = 2.
The vector eld

H points radially outward, so it is perpendicular to the circle everywhere. Thus
_
C
H dr = 0.
(c) Greens Theorem does not apply to the computation of the line integrals for

G and

H because their domains do not
include the origin, which is in the interior, R, of the circle. Greens Theorem does apply to

F = y
i + x
j .
_
C
F dr =
_
R
_
F2
x

F1
y
_
dxdy =
_
0 dxdy = 0.
37. (a) We use Greens Theorem. Let R be the region enclosed by the circle C. Then
_
C
F dr =
_
R
_
F2
x

F1
y
_
dA =
_
R
_

x
(e
y
2
+ 12x)

y
(3x
2
y + y
3
+ e
x
)
_
dA
=
_
R
(12 (3x
2
+ 3y
2
) dA =
_
R
(12 3(x
2
+ y
2
)) dA.
Converting to polar coordinates, we have
_
C
F dr =
_
2
0
_
1
0
(12 3r
2
)r dr d = 2
_
6r
2
3
4
r
4
_
1
0
= 2
_
6
3
4
_
=
21
2
.
(b) The integrand of the integral over the disk R is 12 3(x
2
+y
2
). Since the integrand is positive for x
2
+y
2
< 4 and
negative for x
2
+y
2
> 4, the integrand is positive inside the circle of radius 2 and negative outside that circle. Thus,
the integral over R increases with a until a = 2 and then decreases. The maximum value of the line integral occurs
when a = 2.
SOLUTIONS to Review Problems for Chapter Eighteen 405
Exercises
1. The angle between the vector eld and the curve is more than 90
at all points on C, so the line integral is negative.

5. Since

F = 6
i 7
j , consider the function f

f(x, y) = 6x 7y.
Then we see that grad f = 6
i 7
j , so we use the Fundamental Theorem of Calculus for Line Integrals:

_
C
F dr =
_
C
grad f dr
= f(4, 4) f(2, 6) = (4) (54) = 58.
9. Only the
j component contributes to the integral. On the y-axis, x = 0, so

_
C
F dr =
_
5
3
y
2
j dy =
y
3
3
5
3
=
98
3
.
13. The triangle C consists of the three paths shown in Figure 18.8.
(0, 0) (3, 0)
(3, 2)
C
1
C
2
C
3
Figure 18.8
Write C = C1 + C2 + C3 where C1, C2, and C3 are parameterized by
C1 : (t, 0) for 0 t 3; C2 : (3, t) for 0 t 2; C3 : (3 3t, 2 2t) for 0 t 1.
Then
_
C
F dr =
_
C
1
F dr +
_
C
2
F dr +
_
C
3
F dr
where
_
C
1
F dr =
_
3
0
F (t, 0)
i dt =
_
3
0
(2t + 4)dt = (t
2
+ 4t)
3
0
= 21
_
C
2
F dr =
_
2
0
F (3, t)
j dt =
_
2
0
(5t + 3)dt = (5t
2
/2 + 3t)
2
0
= 16
_
C
3
F dr =
_
1
0
F (3 3t, 2 2t) (3
i 2
j )dt
=
_
1
0
((4t + 8)
i + (19t + 13)
j ) (3
i 2
j )dt
= 50
_
1
0
(t 1)dt = 25.
So
_
C
F dr = 21 + 16 25 = 12.
17. The domain is all 3-space. Since F1 = y,
curl y
i =
_
F3
y

F2
z
_
i +
_
F1
z

F3
x
_
j +
_
F2
x

F1
y
_
k =
k =
0 ,
so

F is not path-independent.
21. The domain is all 3-space. Since F1 = y, F2 = x,
curl y
i + x
j =
_
F3
y

F2
z
_
i +
_
F1
z

F3
x
_
j +
_
F2
x

F1
y
_
k =
0 ,
so

F is path-independent
25. Since the line is parallel to the x-axis, only the
i -component contributes to the line integral. On C, we have dr =
i dx,
so
_
C
F dr =
_
12
2
5x
i dx =
5
2
x
2
12
2
= 350.
29. We can calculate this line integral either by calculating a separate line integral for each side, or by adding a line segment,
C1, from(1, 4) to (1, 1) to form the closed curve C+C1. Since we now have a closed curve, we can use Greens Theorem:
_
C+C
1
F dr =
_
C+C
1
(5x
i + 3x
j ) dr =
_
R
_

x
(3x)

y
(5y)
_
dxdy
=
_
R
3 dxdy = 3 Area of region = 3
_
2 3 +
1
2
3 4
_
= 36.
Since dr =
j dy on C1, we have
_
C
1
F dr =
_
1
4
3 1
j (
j dy) = 3 3 = 9.
Since
_
C+C
1
F dr =
_
C
F dr +
_
C
1
F dr =
_
C
F dr 9 = 36
we have
_
C
F dr = 45.
Problems
33. (a) The vector eld is everywhere perpendicular to the radial line from the origin to (2, 3), so the line integral is 0.
(b) Since the path is parallel to the x-axis, only the
i component of the vector eld contributes to the line integral. The
i
component is 3
i on this line, and the displacement along this line is 2
i , so
Line integral = (3
i ) (2
i ) = 6.
(c) The circle of radius 5 has equation x
2
+ y
2
= 25. On this curve, ||
F || =
_
(y
2
) + x
2
=
25 = 5. In addition,
F is everywhere tangent to the circle, and the path is 3/4 of the circle. Thus
Line integral = ||
F || Length of curve = 5
3
4
2(5) =
75
2
.
(d) Use Greens Theorem. Writing C for the curve around the boundary of the triangle, we have
F2
x

F1
y
= 1 (1) = 2,
so
_
C
F dr =
_
Triangle
2 dA = 2 Area of triangle = 2 7 = 14.
SOLUTIONS to Review Problems for Chapter Eighteen 407
37. (a) Since

F = (6x + y
2
)
i + 2xy
j = grad(3x
2
+ xy
2
), the vector eld

F is path independent, so
_
C
1
F dr = 0.
(b) Since C1 is closed, we use Greens Theorem, so
_
C
1
G dr =
_
Interior of C
1
_

x
(x + y)

y
(x y)
_
dA
= 2
_
C
1
dA = 2 Area inside C1 = 2
1
2
2 2 = 4.
(c) Since

F = grad(3x
2
+ xy
2
), using the Fundamental Theorem of Line Integrals gives
_
C
2
F dr = (3x
2
+ xy
2
)
(0,2)
(2,0)
= 0 3 2
2
= 12.
(d) Parameterizing the circle by
x = 2 cos t y = 2 sin t 0 t
3
2
,
gives
x
= 2 sin t y
= 2 cos t,
so the integral is
_
C
2
G dr =
_
3/2
0
_
(2 cos t 2 sin t)
i + (2 cos t + 2 sin t)
j
_
(2 sin t
i + 2 cos t
j ) dt
=
_
3/2
0
4(cos t sin t + sin
2
t + cos
2
t + sin t cos t) dt
= 4
_
3/2
0
dt = 4
3
2
= 6.
41. (a) Since

x
(y
5
+x)

y
(x
3
y) = 1+1 = 2, any closed curve oriented counterclockwise will do. See Figure 18.9.
x
y
Figure 18.9
x
y
Figure 18.10
(b) Since

x
(y
5
xy)

y
(x
3
) = y, any closed curve in the lower half-plane oriented counterclockwise or any
closed curve in the upper half-plane oriented clockwise will do. See Figure 18.10. Other answers are possible.
45. Since ||
F || 7, the line integral cannot be larger than 7 times the length of the curve. Thus
_
C
F dr 7 Circumference of circle = 7 2 = 14.

The line integral is equal to 14 if

F is everywhere of magnitude 7, tangent to the curve, and pointing in the direction in
which the curve is traversed.
The smallest possible value occurs if the vector eld is everywhere of magnitude 7, tangent to the curve and pointing
opposite to the direction in which the curve is transversed. Thus
_
C
F dr 14.
49. The free vortex appears to starts at about r = 200 meters (thats where the graph changes its behavior) and the tangential
velocity at this point is about 200 km/hr = 2 10
5
meters/hr.
Since v = (y
i + x
j ) for
_
x
2
+ y
2
200, at r = 200 we have
v =
_
(y)
2
+ x
2
= (200) = 2 10
5
meters/hr,
so
= 10
3
rad/hr.
Since v = K(x
2
+ y
2
)
1
(y
i + x
j ) for
_
x
2
+ y
2
200, at r = 200 we have
v = K(200
2
)
1
(200) =
K
200
= 2 10
5
meters/hr
so
K = 4 10
7
m
2
rad/hr.
53. We have
_
C
1
F dr =
_
3
0
F (r (t)) r

(t)dt
=
_
3
0
(2
_
2at + bt
2
_
+ 2t
_
2ct + dt
2
_
) dt = 18a + 18b + 36c + (81d/2)
and
_
C
2
F dr =
_
3
0
F (r (t)) r

(t)dt
=
_
3
0
_
2
_
2a (3 t) + b(3 t)
2
_
2
_
2c (3 t) + d(3 t)
2
_
(3 t)
_
dt
= 18a 18b 36c (81d/2)
The second integral is the negative of the rst. This is because C2 is the same curve as C1 but traveling in the opposite
direction.
1. A path-independent vector eld must have zero circulation around all closed paths. Consider a vector eld like

F (x, y) =
|x|
j , shown in Figure 18.11.

y
x
Figure 18.11
A rectangular path that is symmetric about the y-axis will have zero circulation: on the horizontal sides, the eld is
perpendicular, so the line integral is zero. The line integrals on the vertical sides are equal in magnitude and opposite in
sign, so they cancel out, giving a line integral of zero. However, this eld is not path-independent, because it is possible to
nd two paths with the same endpoints but different values of the line integral of

F . For example, consider the two points
(0, 0) and (0, 1). The path C1 in Figure 18.12 along the y axis gives zero for the line integral, because the eld is 0 along
the y axis, whereas a path like C2 will have a nonzero line integral. Thus the line integral depends on the path between
the points, so

F is not path-independent.
C
1
C
2
x
y
(0, 0)
(0, 1)
Figure 18.12
5. False. Because

F r is a scalar quantity,
_
C
F dr is also a scalar quantity.

9. False. We can calculate a line integral of any vector eld.
13. False. The relative sizes of the line integrals along C1 and C2 depend on the behavior of the vector eld

F along the
curves. As a counterexample, take the vector eld

F =

i , and C1 to be the line from the origin to (0, 2), while C2
is the line from the origin to (1, 0). Then the length of C1 is 2, which is greater than the length of C2, which is 1.
However
_
C
1
F dr = 0 (since

F is perpendicular to C1) while
_
C
2
F dr > 0 (since

F points along C2).
17. False. The relation between these two line integrals depends on the behavior of the vector eld along each of the curves,
so there is no reason to expect one to be the negative of the other. As an example, if

F (x, y) = y
i , then, by symmetry,
both line integrals are equal to the same negative number.
21. True. The dot product of the integrand 4
i with r

(t) =
i + 2t
j is 4, so the integral has value

_
2
0
4 dt = 8.
25. False. As a counterexample, consider the vector eld

F = x
i . Then if we parameterize C1 by r (t) = t
i , with 0 t 1,
we get
_
C
1
x
i dr =
_
1
0
t
i dt =
_
1
0
t dt =
t
2
2
1
0
=
1
2
.
A similar computation for C2 gives a line integral with value 2.
29. True. By the Fundamental Theorem for Line Integrals, if C is a path from P to Q, then
_
C
grad f dr = f(Q) f(P),
so the value of the line integral
_
C
grad f dr depends only on the endpoints and not the path.
33. True. Since a gradient eld is path-independent, and C1 and C2 have the same initial and nal points, the two line integrals
are equal.
37. True. The value of
F2
x

F1
y
is 0 0 = 0, so the eld is path-independent.
41. True. This vector eld has components F1 = x, F2 = y, and F3 = z. Using the 3-space curl test gives zero for all of the
components of curl
F , so the eld is path-independent.

19.1 SOLUTIONS 411
CHAPTER NINETEEN
Exercises
1. Scalar. Only the
j -component of the vector eld contributes to the ux and d

A =
j dA, so
_
S
(3
i + 4
j ) d
A = 4 Area of disk = 4 5
2
= 100.
5. On the surface, d
A =

k dA, so only the
k component of v contributes to the ux:

Flux =
_
S
v d
A =
_
S
(
j + 3
k )
k dA = 3 Area of disk = 3 2
2
= 12.
9. v

A = (2
i + 3
j + 5
k )
k = 5.
13. The disk has area 25, so its area vector is 25
j . Thus
Flux = (2
i + 3
j ) 25
j = 75.
17. Since the square, S, is in the plane y = 0 and oriented in the negative y-direction, d

A =
j dxdz and
_
S
F d
A =
_
S
(0 + 3)
j (
j dxdz) = 3
_
S
dxdz = 3 Area of square = 3(2
2
) = 12.
21. The only contribution to the ux is from the
j -component, and since d
A =
j dxdz on the square, S, we have

Flux =
_
S
(6
i + x
2
k ) d
A =
_
2
2
_
2
2
x
2
j dxdz =
_
2
2
x
3
3
2
2
dz =
16
3
4 =
64
3
.
25. We have d
A =
i dA, so
_
S
F d
A =
_
S
(2z
i + x
j + x
k )
i dA =
_
S
2z dA
=
_
2
0
_
3
0
2z dzdy = 18.
29. Since the disk is oriented in the positive x-direction, d

A =
i dydz, so we have
Flux =
_
Disk
F d
A =
_
Disk
e
y
2
+z
2
i dydz =
_
Disk
e
y
2
+z
2
dydz.
To calculate this integral, we use polar coordinates with y = r cos and z = r sin . Then r
2
= y
2
+ z
2
and
_
Disk
F d
A =
_
2
0
_
2
0
e
r
2
rdrd = 2
e
r
2
2
2
0
= (e
4
1).
33. Since r is perpendicular to S and ||r || = 3 on S, we have
_
S
r d
A = 3 Area of surface = 3 43
2
= 108.
412 Chapter Nineteen /SOLUTIONS
Problems
37. Since this vector eld points radially out from the origin, it is everywhere parallel to the vector representing the surface
area, d
A . Thus since
F (r ) = 1/R
2
on the surface, S,
F (r ) d
A =
1
R
2
dA,
so
_
S
F (r ) d
A =
1
R
2
Surface area of sphere =
1
R
2
(4R
2
) = 4.
41. By the symmetry of the sphere, the

i and

j components of

F do not contribute to the ux; only the

k component
contributes. The vector normal to S has a negative

k component, so we need c > 0. There are no conditions on a and b.
45. The square of side 2 in the plane x = 5, oriented in the positive x-direction, has area vector

A = 4
i . Since the vector

eld is constant
Flux = (a
i + b
j + c
k ) 4
i = 4a = 24.
Thus, a = 6 and we cannot say anything about the values of b and c.
49. (a) (i) The integral
_
W
dV represents the total charge in the volume W.
(ii) The integral
_
S
J d
A represents the total current owing out of the surface S.

(b) The total current owing out of the surface S is the rate at which the total charge inside the surface S (i.e., in the
volume W) is decreasing. In other words,
Rate current owing out of S =

t
(charge in W),
so
_
S
J d
A =

t
__
W
dV
_
.
Exercises
1. Only the z-component of the vector eld contributes to the ux. Since d

A =

k dxdy on the surface, we have
_
S
(3
i + 4
j + xy
k ) d
A =
_
7
0
_
5
0
xy dxdy =
_
7
0
x
2
y
2
5
0
dy =
_
7
0
25
2
y dy =
25y
2
4
7
0
=
1225
4
.
5. Writing the surface S as z = f(x, y) = y + 1, we have
d
A = (fx
i fy
j +
k )dxdy.
Thus,
_
S
F d
A =
_
R
F (x, y, f(x, y)) (fx
i fy
j +
k ) dxdy
=
_
1
0
_
1
0
(2x
j + y
k ) (
j +
k ) dxdy
=
_
1
0
_
1
0
(2x + y) dxdy =
_
1
0
(x
2
+ xy)
1
0
dy
=
_
1
0
(1 + y) dy = (y +
y
2
2
)
1
0
=
3
2
.
19.2 SOLUTIONS 413
9. On the curved side of the cylinder, only the components x
i + z
k contribute to the ux. Since x
i + z
k is perpendicular
to the curved surface and ||x
i + z
k || = 2 there (because the cylinder has radius 2), we have

Flux through sides = 2 Area of curved surface = 2 2 2 6 = 48.
On the at ends, only y
j contributes to the ux. On one end, y = 3 and d

A =

j dA; on the other end, y = 3 and
d
A =
j dA. Thus
Flux through ends = Flux through top + Flux through bottom
= 3
j (2
2
) + (3
j ) (
j (2
2
)) = 24.
So,
Total ux = 48 + 24 = 72.
13. Using z = 1 x y, the upward pointing area element is d

A = (
i +

j +

k ) dxdy, so the downward one is
d
A = (
k ) dxdy. Since S is oriented downward, we have

_
S
F d
A =
_
S
(x
i + y
j + z
k ) d
A
=
_
3
0
_
2
0
(x
i + y
j + (1 x y)
k ) (
k ) dxdy
=
_
3
0
_
2
0
(x y 1 + x + y) dxdy = 6.
17. Since y = f(x, z) = x
2
+ z
2
, we have
d
A = (fx
i +
j fz
k ) dxdz = (2x
i +
j 2z
k ) dxdz.
Thus, substituting y = x
2
+ z
2
into

F , we have
_
S
F d
A =
_
x
2
+z
2
1
((x
2
+ z
2
)
i +
j xz
k ) (2x
i +
j 2z
k ) dxdz
=
_
x
2
+z
2
1
(2x
3
2xz
2
+ 1 + 2xz
2
) dxdz
=
_
1
1
_

1z
2
1z
2
(1 2x
3
) dxdz
=
_
1
1
_

1z
2
1z
2
dxdz
_
1
1
_

1z
2
1z
2
2x
3
dxdz
= Area of disk
_
1
1
_
_
x
4
2
1z
2
1z
2
_
_
dz = 0 =
21. The ux of

F through S is given by
_
S
F d
A =
_
2
0
_
/2
0
(2 cos
k ) (sin cos
i + sin sin
j + cos
k )2
2
sin dd
=
_
2
=0
_
/2
=0
8 sin cos
2
dd = 16(
cos
3
3
)
/2
=0
=
16
3
.
Problems
25. The

k -component of

F does not contribute to the ux as it is perpendicular to the surface. The vector eld x
i + y
j is
everywhere perpendicular to S and has constant magnitude ||x
2
+ y
2
|| = 1 on the surface S. Thus
_
S
F d
A =
_
S
(x
i + y
j ) d
A = 1 Area of S = 1
2
=

2
.
Alternatively, the ux can be computed by integrating with respect to x and z, treating y as a function of x and z. A
parameterization of S is given by y =
1 x
2
, 0 x 1,0 y 1, 0 z 1. Thus,
_
S
F d
A =
_
1
0
_
1
0
(x
i +
_
1 x
2
j + z
k ) (yx
i +
j yz
k ) dxdz
=
_
1
0
_
1
0
(x
i +
_
1 x
2
j + z
k )
_
x
1 x
2
i +
j + 0
k
_
dxdz
=
_
1
0
_
1
0
_
x
2
1 x
2
+
_
1 x
2
_
dxdz
=
_
1
0
_
1
0
1
1 x
2
dxdz
= 1 arcsin x
1
0
=

2
.
29. The plane is x z = 0 over region 0 x
2, 0 y 2. See Figure 19.1.

x
z
2
2
2
y
Figure 19.1
Flux =
_
2
0
_

2
0
_
(e
xy
+ 3z + 5)
i + (e
xy
+ 5z + 3)
j + (3z + e
xy
)
k
_
(
k ) dxdy
=
_
2
0
_

2
0
(e
xy
+ 3z + 5 3z e
xy
) dxdy = 5(2)(
2) = 10
2
Alternatively, since a unit normal to the surface is n /
2 = (
j )/
2, writing dA = ||d
A||, we have
Flux =
_
S
H d
A =
_

H
2
dA =
_
5
2
dA
=
5
2
(Area of slanted square) =
5
2
4 = 10
2.
19.3 SOLUTIONS 415
Exercises
1. Since S is given by
r (s, t) = (s + t)
i + (s t)
j + (s
2
+ t
2
)
k ,
we have
r
s
=
i +
j + 2s
k and
r
t
=
j + 2t
k ,
and
r
s

r
t
=
i

j

k
1 1 2s
1 1 2t
= (2s + 2t)
i + (2s 2t)
j 2
k .
Since the
i component of this vector is positive for 0 < s < 1, 0 < t < 1, it points away from the z-axis, and so has the
opposite orientation to the one specied. Thus, we use
d
A =
r
s

r
t
ds dt,
and so we have
_
S
F d
A =
_
1
0
_
1
0
(s
2
+ t
2
)
k
_
(2s + 2t)
i + (2s 2t)
j 2
k
_
ds dt
= 2
_
1
0
_
1
0
(s
2
+ t
2
) ds dt = 2
_
1
0
_
s
3
3
+ st
2
_
s=1
s=0
dt
= 2
_
1
0
(
1
3
+ t
2
) dt = 2(
1
3
t +
t
3
3
)
1
0
= 2(
1
3
+
1
3
) =
4
3
.
5. The cross product r /s r /t is given by
r
s

r
t
=
i

j

k
2s 2 0
0 2t 5
= 10
i 10s
j + 4st
k .
Since the z-component, 4st, of the vector r /s r /t is positive for 0 < s 1, 1 t 3, we see that r /s
r /t points upward, in the direction of the orientation of S we were given. Thus, we use
d
A =
_
r
s

r
t
_
ds dt,
and so we have
_
S
F d
A =
_
1
0
_
3
1
(5t
i + s
2
j ) (10
i 10s
j + 4st
k ) dt ds
=
_
1
0
_
3
1
(50t 10s
3
) dt ds =
_
1
0
(25t
2
10s
3
t)
t=3
t=1
ds
=
_
1
0
(200 20s
3
) ds = (200s 5s
4
)
1
0
= 200 5 = 195.
9. A parameterization of S is
x = s, y = t, z = 3s + 2t, for 0 s 10, 0 t 20.
We compute
r
s

r
t
= (
i + 3
k ) (
j + 2
k ) = 3
i 2
j +
k
_
_
_
_
r

r
t
_
_
_
_
=
14
Surface area =
_
S
dA =
_
R
_
_
_
_
r

r
t
_
_
_
_
dA =
_
20
t=0
_
10
s=0
14dsdt = 200
14.
Problems
13. The elliptic cylindrical surface is parameterized by
r = x
i + y
j + z
k = a cos
i + b sin
j + z
k where 0 2, c z c.
We have
r

r
z
=
i

j

k
a sin b cos 0
0 0 1
= b cos
i + a sin
j .
This vector points away from the z-axis, so we use d

A = (b cos
i + a sin
j ) ddz, giving
_
S
F d
A =
_
c
c
_
2
0
(
b
a
(a cos )
i +
a
b
(b sin
j )) (b cos
i + a sin
j ) d dz
=
_
c
c
_
2
0
(b
2
cos
2
+ a
2
sin
2
) ddz
= 2c(a
2
+ b
2
).
17. The surface of S is parameterized by
r (, ) = x
i + y
j + z
k ,
where
_
_
x = a + d sin cos ,
y = b + d sin sin ,
z = c + d cos ,
for 0 , 0 2.
The vector r / r / points outward by the right-hand rule, so
d
A =
_
r
_
dd.
Thus,
F d
A =

F
_
r
_
dd
=
x
2
y
2
z
2
x
dd
=
(a + d sin cos )
2
(b + d sin sin )
2
(c + d cos )
2
d cos cos d cos sin d sin
d sin sin d sin cos 0
dd.
19.3 SOLUTIONS 417
Hence,
_
S
F d
A = d
2
_
2
0
_

0
( a
2
sin
2
cos + 2ad sin
3
cos
2
+ d
2
sin
4
cos
3
+ b
2
sin
2
sin + 2bd sin
3
sin
2
+ d
2
sin
4
sin
3
+ c
2
sin cos + 2cd sin cos
2
+ d
2
sin cos
3
) dd.
Since
_
2
0
cos d =
_
2
0
sin d =
_
2
0
cos
3
d =
_
2
0
sin
3
d = 0,
and
_

0
sin cos d =
_

0
sin cos
3
d = 0,
we have
_
S
F

A d
2
_
2
0
_

0
(2ad sin
3
cos
2
+ 2bd sin
3
sin
2
+ 2cd sin cos
2
) dd
= 2d
3
_

0
(a sin
3
+ b sin
3
+ 2c sin cos
2
) d
= 4d
3
_
/2
0
(a sin
3
+ b sin
3
+ 2c sin cos
2
) d
=
8
3
d
3
(a + b + c).
21. In terms of the st-parameterization,
d
A =
r
s

r
t
ds dt.
By the chain rule, we have
r
s
=
r
u
u
s
+
r
v
v
s
r
t
=
r
u
u
t
+
r
v
v
t
.
So taking the cross product, we get
r
s

r
t
=
_
r
u
u
s
+
r
v
v
s
_
_
r
u
u
t
+
r
v
v
t
_
=
_
u
s
v
t

u
t
v
s
_
r
u

r
v
.
Now suppose we are going to change variables in a double integral from uv-coordinates to st-coordinates. The
Jacobian is
(u, v)
(s, t)
=
u
s
v
s
u
t
v
t
=
u
s
v
t

u
t
v
s
.
Since the Jacobian is assumed to be positive, converting from a uv-integral to an st-integral gives:
_
T
F
r
u

r
v
dudv =
_
R
F
r
u

r
v
(u, v)
(s, t)
dsdt
=
_
R
F
r
u

r
v
_
u
s
v
t

u
t
v
s
_
dsdt.
However, we know that this gives us
_
T
F
r
u

r
v
dudv =
_
R
F
r
u

r
v
_
u
s
v
t

u
t
v
s
_
dsdt =
_
R
F
r
s

r
t
dsdt.
Thus, the ux integral in uv-coordinates equals the ux integral in st-coordinates.
Exercises
1. Scalar. Since the surface is closed and the vector eld is constant, the ux in one side equals the ux out on the other side,
so the net ux through the surface is 0.
5. The only contribution to the ux is from the face z = 3, since the vector eld is zero or parallel to the other faces. On this
face,

H = 3x
k . The vector eld is everywhere perpendicular to the face z = 3 but varies in magnitude from point to
point. On this surface, d
A =

k dxdy. Thus
Flux =
_
2
0
_
1
0
3x
k dxdy =
_
2
0
_
1
0
3xdxdy =
3x
2
2
1
0
y
2
0
= 3.
9. Only the
i -component contributes to the ux, so

Flux = 2 Area of surface = 2(3
2
) = 18
2
.
13. Since the vector eld is everywhere perpendicular to the surface of the sphere, and ||
F || = on the surface, we have

_
S
F d
A = ||
F || Area of sphere = 4()

2
= 4
4
.
17. The area vector of the surface is 49
j , so
_
S
G d
A = (2
j + 3
k ) (49
j ) = 98.
21. Only the
k component contributes to the ux. In the plane z = 4, we have

F = 2
i +3
j +4
k . On the square d
A =

k dA,
so we have
Flux =
_

F d
A = 4
k (
k Area of square) = 4(5

2
) = 100.
25. See Figure 19.2. The vector eld is a vortex going around the z-axis, and the square is centered on the x-axis, so the ux
going across one half of the square is balanced by the ux coming back across the other half. Thus, the net ux is zero, so
_
S
F

dA = 0.
x
y
z
n
Figure 19.2
SOLUTIONS to Review Problems for Chapter Nineteen 419
29. On the sphere of radius 2, the vector eld has ||
F || = 10 and points inward everywhere (opposite to the orientation of

the surface). So
Flux =
_
S
F d
A = ||
F || Area of sphere = 10 42
2
= 160 .
33. On the surface S, y is constant, y = 1, and d

A =
j dA, so,
_
S
F d
A =
_
S
(x
2
i + (x + e
1
)
k ) (
j ) dA =
_
S
(x + e
1
) dA
=
_
4
0
_
2
0
(x + e
1
) dxdz = 4(2 + 2e
1
) = 8(1 + e
1
).
37. First we have
zx =
x
_
x
2
+ y
2
zy =
y
_
x
2
+ y
2
.
Although z is not a smooth function of x and y at (0, 0), the improper integral that we get converges:
_
S
F d
A =
_
S
(x
2
i + y
2
j +
_
x
2
+ y
2
k ) (
x
_
x
2
+ y
2
i
y
_
x
2
+ y
2
j +
k ) dA
=
_
S
_
x
3
+ y
3
_
x
2
+ y
2
+
_
x
2
+ y
2
_
dA
Changing to polar coordinates we have
_
S
F d
A =
_
/2
0
_
1
0
(r
2
cos
3
r
2
sin
3
+ r)r drd
=
_
/2
0
_
r
4
4
(cos
3
+ sin
3
) +
1
3
r
3
r=1
r=0
_
d
=
_
/2
0
_
1
4
(cos
3
+ sin
3
) +
1
3
_
d
=
_
/2
0
_
1
4
(cos cos sin
2
+ sin sin cos
2
) +
1
3
_
d
=
1
4
(sin
1
3
sin
3
cos +
1
3
cos
3
) +

3
/2
0
=

6

1
3
.
Problems
41. (a) We have grad f = (y + yze
xyz
)
i + (x + xze
xyz
)
j + xye
xyz
k .
(b) By the Fundamental Theorem of Line Integrals, we have
_
C
grad f dr = (xy + e
xyz
)
(2,3,4)
(1,1,1)
= (2 3 + e
234
(1 1 + e
111
) = 5 + e
24
e
1
.
(c) Only the k-component of grad f contributes to the ux integral. On the xy-plane, d

A =

k dxdy and z = 0, so
_
S
grad f d
A =
_
2
0
_

4x
2
0
xye
0
dy dx =
_
2
0
xy
2
2
4x
2
0
=
_
2
0
x
2
(4 x
2
) dx =
_
x
2
x
4
8
_
2
0
= 2.
45. The vector eld

D has constant magnitude on S, equal to Q/4R
2
, and points radially outward, so
_
S
D d
A =
Q
4R
2
4R
2
= Q.
49. (a) Consider two opposite faces of the cube, S1 and S2. The corresponding area vectors are

A 1 = 4 and

A 2 = 4
(since the side of the cube has length 2). Since

E is constant, we nd the ux by taking the dot product, giving
Flux through S1 =

E

A 1 = (a
i + b
j + c
k ) 4
i = 4a.
Flux through S2 =

E

A 2 = (a
i + b
j + c
k ) (4
i ) = 4a.
Thus the uxes through S1 and S2 cancel. Arguing similarly, we conclude that, for any pair of opposite faces, the
sum of the uxes of

E through these faces is zero. Hence, by addition,
_
S
E d
A = 0.
(b) The basic idea is the same as in part (a), except that we now need to use Riemann sums. First divide S into two
hemispheres H1 and H2 by the equator C located in a plane perpendicular to

E . For a tiny patch S1 in the hemisphere
H1, consider the patch S2 in the opposite hemisphere which is symmetric to S1 with respect to the center O of the
sphere. The area vectors
A 1 and
A 2 satisfy
A 2 =
A 1, so if we consider S1 and S2 to be approximately

at, then

E
A 1 =
A 2. By decomposing H1 and H2 into small patches as above and using Riemann

sums, we get
_
H
1
E d
A =
_
H
2
E d
A , so
_
S
E d
A = 0.
(c) The reasoning in part (b) can be used to prove that the ux of

E through any surface with a center of symmetry is
zero. For instance, in the case of the cylinder, cut it in half with a plane z = 1 and denote the two halves by H1 and
H2. Just as before, take patches in H1 and H2 with A1 = A2, so that

E A1 =
A 2. Thus, we get
_
H
1
E d
A =
_
H
2
E d
A ,
which shows that
_
S
E d
A = 0.
53. (a) When x > 0, the vector x
i points in the positive x-direction, and when x < 0 it points in the negative x-direction.
Thus it always points from the inside of the ellipsoid to the outside, so we expect the ux integral to be positive. The
upper half of the ellipsoid is the graph of z = f(x, y) =
1
2
(1 x
2
y
2
), so the ux integral is
_
S
F d
A =
_
1/2
1/2
_
1/2
1/2
x
i (fx
i fy
j +
k ) dxdy
=
_
1/2
1/2
_
1/2
1/2
(xfx) dxdy =
_
1/2
1/2
_
1/2
1/2
x
2
_
1 x
2
y
2
dxdy
=
2 + 11 arcsin(
1
3
) + 10 arctan(
1
2
) 8 arctan(
5
2
)
12
= 0.0958.
Different CASs may give the answer in different forms. Note that we could have predicted the integral was positive
without evaluating it, since the integrand is positive everywhere in the region of integration.
(b) For x > 1, the quantity x + 1 is positive, so the vector eld (x + 1)
i always points in the direction of the positive

x-axis. It is pointing into the ellipsoid when x < 0 and out of it when x > 0. However, its magnitude is smaller when
1/2 < x < 0 than it is when 0 < x < 1/2, so the net ux out of the ellipsoid should be positive. The ux integral
is
_
S
F d
A =
_
1/2
1/2
_
1/2
1/2
(x + 1)
i (fx
i fy
j +
k ) dxdy
=
_
1/2
1/2
_
1/2
1/2
(x + 1)fx dxdy =
_
1/2
1/2
_
1/2
1/2
x(1 + x)
_
1 x
2
y
2
dxdy
=
2 11 arcsin(
1
3
) 10 arctan(
1
2
) + 8 arctan(
5
2
)
12
= 0.0958
The answer is the same as in part (a). This makes sense because the difference between the integrals in parts (a) and
(b) is the integral of
_
1/2
1/2
_
1/2
1/2
(x/
_
1 x
2
y
2
) dxdy, which is zero because the integrand is odd with respect
to x.
(c) This integral should be positive for the same reason as in part (a). The vector eld y
j points in the positive y-direction

when y > 0 and in the negative y-direction when y < 0, thus it always points out of the ellipsoid. Evaluating the
integral we get
_
S
F d
A =
_
1/2
1/2
_
1/2
1/2
y
j (fx
i fy
j +
k ) dxdy
=
_
1/2
1/2
_
1/2
1/2
(yfy) dxdy =
_
1/2
1/2
_
1/2
1/2
y
2
_
1 x
2
y
2
dxdy
=
2 2 arcsin(
1
3
) 19 arctan(
1
2
) + 8 arctan(
5
2
)
12
= 0.0958.
The symbolic answer appears different but has the same numerical value as in parts (a) and (b). In fact the answer is
the same because the integral here is the same as in part (a) except that the roles of x and y have been exchanged.
Different CASs may give different symbolic forms.
1. True. By denition, the ux integral is the limit of a sum of dot products, hence is a scalar.
5. True. The ow of this eld is in the same direction as the orientation of the surface everywhere on the surface, so the ux
is positive.
9. True. In the sum dening the ux integral for

F , we have terms like

F
A = (2
G )
A = 2(
A ). So each
term in the sum approximating the ux of

F is twice the corresponding term in the sum approximating the ux of

G ,
making the sum for

F twice that of the sum for

G . Thus the ux of

F is twice the ux of

G .
13. False. Both surfaces are oriented upward, so

A (x, y) and

B (x, y) both point upward. But they could point in different
directions, since the graph of z = f(x, y) is the graph of z = f(x, y) turned upside down.
1.1 SOLUTIONS 1
CHAPTER ONE
Exercises
1. Since t represents the number of years since 1970, we see that f(35) represents the population of the city in 2005. In
2005, the citys population was 12 million.
5. The slope is (3 2)/(2 0) = 1/2. So the equation of the line is y = (1/2)x + 2.
9. Rewriting the equation as
y =
12
7
x +
2
7
shows that the line has slope 12/7 and vertical intercept 2/7.
13. (a) is (V), because slope is negative, vertical intercept is 0
(b) is (VI), because slope and vertical intercept are both positive
(c) is (I), because slope is negative, vertical intercept is positive
(d) is (IV), because slope is positive, vertical intercept is negative
(e) is (III), because slope and vertical intercept are both negative
(f) is (II), because slope is positive, vertical intercept is 0
17. y = 5x 3. Since the slope of this line is 5, we want a line with slope
1
5
passing through the point (2, 1). The equation
is (y 1) =
1
5
(x 2), or y =
1
5
x +
7
5
.
21. Since x goes from 1 to 5 and y goes from 1 to 6, the domain is 1 x 5 and the range is 1 y 6.
25. The domain is all x-values, as the denominator is never zero. The range is 0 < y
1
2
.
29. If distance is d, then v =
d
t
.
Problems
33. (a) When the car is 5 years old, it is worth $6000.
(b) Since the value of the car decreases as the car gets older, this is a decreasing function. Apossible graph is in Figure 1.1:
(5, 6)
a (years
V (thousand dollars)
Figure 1.1
(c) The vertical intercept is the value of V when a = 0, or the value of the car when it is new. The horizontal intercept
is the value of a when V = 0, or the age of the car when it is worth nothing.
37. (a) f(30) = 10 means that the value of f at t = 30 was 10. In other words, the temperature at time t = 30 minutes was
10
C. So, 30 minutes after the object was placed outside, it had cooled to 10

C.
(b) The intercept a measures the value of f(t) when t = 0. In other words, when the object was initially put outside,
it had a temperature of a
C. The intercept b measures the value of t when f(t) = 0. In other words, at time b the
objects temperature is 0

C.
2 Chapter One /SOLUTIONS
41. (a) We have the following functions.
(i) Since a change in p of $5 results in a decrease in q of 2, the slope of q = D(p) is 2/5 items per dollar. So
q = b
2
5
p.
Now we know that when p = 550 we have q = 100, so
100 = b
2
5
550
100 = b 220
b = 320.
Thus a formula is
q = 320
2
5
p.
(ii) We can solve q = 320
2
5
p for p in terms of q:
5q = 1600 2p
2p = 1600 5q
p = 800
5
2
q.
The slope of this function is 5/2 dollars per item, as we would expect.
(b) A graph of p = 800
5
2
q is given in Figure 1.2.
100 320
550
800
q (items)
p (dollars)
Figure 1.2
45. (a) The line given by (0, 2) and (1, 1) has slope m =
21
1
= 1 and y-intercept 2, so its equation is
y = x + 2.
The points of intersection of this line with the parabola y = x
2
are given by
x
2
= x + 2
x
2
+ x 2 = 0
(x + 2)(x 1) = 0.
The solution x = 1 corresponds to the point we are already given, so the other solution, x = 2, gives the x-
coordinate of C. When we substitute back into either equation to get y, we get the coordinates for C, (2, 4).
1.2 SOLUTIONS 3
(b) The line given by (0, b) and (1, 1) has slope m =
b1
1
= 1b, and y-intercept at (0, b), so we can write the equation
for the line as we did in part (a):
y = (1 b)x + b.
We then solve for the points of intersection with y = x
2
the same way:
x
2
= (1 b)x + b
x
2
(1 b)x b = 0
x
2
+ (b 1)x b = 0
(x + b)(x 1) = 0
Again, we have the solution at the given point (1, 1), and a new solution at x = b, corresponding to the other point
of intersection C. Substituting back into either equation, we can nd the y-coordinate for C is b
2
, and thus C is given
by (b, b
2
). This result agrees with the particular case of part (a) where b = 2.
Exercises
1. The graph shows a concave up function.
5. Initial quantity = 5; growth rate = 0.07 = 7%.
9. Since e
0.25t
=
e
0.25
t
(1.2840)
t
, we have P = 15(1.2840)
t
. This is exponential growth since 0.25 is positive. We
can also see that this is growth because 1.2840 > 1.
13. (a) Let Q = Q0a
t
. Then Q0a
5
= 75.94 and Q0a
7
= 170.86. So
Q0a
7
Q0a
5
=
170.86
75.94
= 2.25 = a
2
.
So a = 1.5.
(b) Since a = 1.5, the growth rate is r = 0.5 = 50%.
Problems
17. (a) This is a linear function with slope 2 grams per day and intercept 30 grams. The function is Q = 30 2t, and the
graph is shown in Figure 1.3.
30
15
Q(grams)
t (days)
Q = 30 2t
Figure 1.3
30
15
Q(grams)
t (days)
Q = 30(0.88)
t
Figure 1.4
(b) Since the quantity is decreasing by a constant percent change, this is an exponential function with base 1 0.12 =
0.88. The function is Q = 30(0.88)
t
, and the graph is shown in Figure 1.4.
21. (a) Advertising is generally cheaper in bulk; spending more money will give better and better marginal results initially,
(Spending $5,000 could give you a big newspaper ad reaching 200,000 people; spending $100,000 could give you a
series of TV spots reaching 50,000,000 people.) See Figure 1.5.
(b) The temperature of a hot object decreases at a rate proportional to the difference between its temperature and the
temperature of the air around it. Thus, the temperature of a very hot object decreases more quickly than a cooler
object. The graph is decreasing and concave up. See Figure 1.6 (We are assuming that the coffee is all at the same
temperature.)
advertising
revenue
Figure 1.5
time
temperature
Figure 1.6
25. The difference, D, between the horizontal asymptote and the graph appears to decrease exponentially, so we look for an
equation of the form
D = D0a
x
where D0 = 4 = difference when x = 0. Since D = 4 y, we have
4 y = 4a
x
or y = 4 4a
x
= 4(1 a
x
)
The point (1, 2) is on the graph, so 2 = 4(1 a
1
), giving a =
1
2
.
Therefore y = 4(1 (
1
2
)
x
) = 4(1 2
x
).
29. (a) The formula is Q = Q0
1
2
(t/1620)
.
(b) The percentage left after 500 years is
Q0(
1
2
)
(500/1620)
Q0
.
The Q0s cancel giving
1
2
(500/1620)
0.807,
so 80.7% is left.
33. (a) This is the graph of a linear function, which increases at a constant rate, and thus corresponds to k(t), which increases
by 0.3 over each interval of 1.
(b) This graph is concave down, so it corresponds to a function whose increases are getting smaller, as is the case with
h(t), whose increases are 10, 9, 8, 7, and 6.
(c) This graph is concave up, so it corresponds to a function whose increases are getting bigger, as is the case with g(t),
whose increases are 1, 2, 3, 4, and 5.
37. Because the population is growing exponentially, the time it takes to double is the same, regardless of the population
levels we are considering. For example, the population is 20,000 at time 3.7, and 40,000 at time 6.0. This represents a
doubling of the population in a span of 6.0 3.7 = 2.3 years.
How long does it take the population to double a second time, from 40,000 to 80,000? Looking at the graph once
again, we see that the population reaches 80,000 at time t = 8.3. This second doubling has taken 8.3 6.0 = 2.3 years,
the same amount of time as the rst doubling.
Further comparison of any two populations on this graph that differ by a factor of two will show that the time that
separates them is 2.3 years. Similarly, during any 2.3 year period, the population will double. Thus, the doubling time is
2.3 years.
Suppose P = P0a
t
doubles from time t to time t + d. We now have P0a
t+d
= 2P0a
t
, so P0a
t
a
d
= 2P0a
t
. Thus,
canceling P0 and a
t
, d must be the number such that a
d
= 2, no matter what t is.
1.3 SOLUTIONS 5
Exercises
1.
2 2
4
4
x
y
(a)
2 2
4
4
x
y
(b)
2 2
4
4
x
y
(c)
2 2
4
4
x
y
(d)
2 2
4
4
x
y
(e)
2 2
4
4
x
y
(f)
5. This graph is the graph of m(t) shifted to the right by one unit. See Figure 1.7.
4 4 6
4
4
p(t)
t
y
Figure 1.7
9. (a) f(g(1)) = f(1
2
) = f(1) =
1 + 4 =
5
(b) g(f(1)) = g(
1 + 4) = g(
5) = (
5)
2
= 5
(c) f(g(x)) = f(x
2
) =
x
2
+ 4
(d) g(f(x)) = g(
x + 4) = (
x + 4)
2
= x + 4
(e) f(t)g(t) = (
t + 4)t
2
= t
2
t + 4
13. (a) f(t + 1) = (t + 1)
2
+ 1 = t
2
+ 2t + 1 + 1 = t
2
+ 2t + 2.
(b) f(t
2
+ 1) = (t
2
+ 1)
2
+ 1 = t
4
+ 2t
2
+ 1 + 1 = t
4
+ 2t
2
+ 2.
(c) f(2) = 2
2
+ 1 = 5.
(d) 2f(t) = 2(t
2
+ 1) = 2t
2
+ 2.
(e) [f(t)]
2
+ 1 =
t
2
+ 1
2
+ 1 = t
4
+ 2t
2
+ 1 + 1 = t
4
+ 2t
2
+ 2.
17. m(z) m(z h) = z
2
(z h)
2
= 2zh h
2
.
21. f
1
(75) is the length of the column of mercury in the thermometer when the temperature is 75
F.
25. Since a horizontal line cuts the graph of f(x) = x
2
+ 3x + 2 two times, f is not invertible. See Figure 1.8.
3
2
3
4
y = 2.5
f(x) = x
2
+ 3x + 2
x
y
Figure 1.8
29.
f(x) = (x)
3
+ (x)
2
+ (x) = x
3
+ x
2
x.
Since f(x) = f(x) and f(x) = f(x), this function is neither even nor odd.
33. Since
f(x) = e
(x)
2
1
= e
x
2
1
= f(x),
we see f is even.
Problems
37. This looks like a shift of the graph y = x
3
. The graph is shifted to the right 2 units and down 1 unit, so a possible formula
is y = (x 2)
3
1.
41. f is an increasing function since the amount of fuel used increases as ight time increases. Therefore f is invertible.
45. f(g(1)) = f(2) 0.4.
49. Using the same way to compute g(f(x)) as in Problem 46, we get Table 1.1. Then we can plot the graph of g(f(x)) in
Figure 1.9.
Table 1.1
x f(x) g(f(x))
3 3 2.6
2.5 0.1 0.8
2 1 1.4
1.5 1.3 1.8
1 1.2 1.7
0.5 1 1.4
0 0.8 1
0.5 0.6 0.6
1 0.4 0.3
1.5 0.1 0.3
2 0.3 1.1
2.5 0.9 2
3 1.6 2.2
3 3
3
3
x
g(f(x))
Figure 1.9
1.4 SOLUTIONS 7
53. We have approximately u(10) = 13 and v(13) = 60 so v(u(10)) = 60.
57. f(x) =
x, g(x) = x
2
+ 4
61. (a) The function f tells us C in terms of q. To get its inverse, we want q in terms of C, which we nd by solving for q:
C = 100 + 2q,
C 100 = 2q,
q =
C 100
2
= f
1
(C).
(b) The inverse function tells us the number of articles that can be produced for a given cost.
Exercises
1. Using the identity e
ln x
= x, we have e
ln(1/2)
=
1
2
.
5. Using the rules for ln, we have
ln
1
e
+ ln AB = ln 1 ln e + ln A + ln B
= 0 1 + ln A + ln B
= 1 + ln A + ln B.
9. Isolating the exponential term
20 = 50(1.04)
x
20
50
= (1.04)
x
.
Taking logs of both sides
log
2
5
= log(1.04)
x
log
2
5
= xlog(1.04)
x =
log(2/5)
log(1.04)
= 23.4.
13. To solve for x, we rst divide both sides by 600 and then take the natural logarithm of both sides.
50
600
= e
0.4x
ln(50/600) = 0.4x
x =
ln(50/600)
0.4
6.212.
17. Using the rules for ln, we have
2x 1 = x
2
x
2
2x + 1 = 0
(x 1)
2
= 0
x = 1.
21. Taking logs of both sides yields
nt =
log
Q
Q
0
log a
.
Hence
t =
log
Q
Q
0
nlog a
=
log Qlog Q0
nlog a
.
25. Since we want (1.5)
t
= e
kt
= (e
k
)
t
, so 1.5 = e
k
, and k = ln 1.5 = 0.4055. Thus, P = 15e
0.4055t
. Since 0.4055 is
positive, this is exponential growth.
29. If p(t) = (1.04)
t
, then, for p
1
the inverse of p, we should have
(1.04)
p
1
(t)
= t,
p
1
(t) log(1.04) = log t,
p
1
(t) =
log t
log(1.04)
58.708 log t.
Problems
33. We know that the y-intercept of the line is at (0,1), so we need one other point to determine the equation of the line. We
observe that it intersects the graph of f(x) = 10
x
at the point x = log 2. The y-coordinate of this point is then
y = 10
x
= 10
log 2
= 2,
so (log 2, 2) is the point of intersection. We can now nd the slope of the line:
m =
2 1
log 2 0
=
1
log 2
.
Plugging this into the point-slope formula for a line, we have
y y1 = m(x x1)
y 1 =
1
log 2
(x 0)
y =
1
log 2
x + 1 3.3219x + 1.
37. The vertical asymptote is where x + 2 = 0, or x = 2, so increasing a does not effect the vertical asymptote.
41. Since the factor by which the prices have increased after time t is given by (1.05)
t
, the time after which the prices have
doubled solves
2 = (1.05)
t
log 2 = log(1.05
t
) = t log(1.05)
t =
log 2
log 1.05
14.21 years.
45. Let B represent the sales (in millions of dollars) at Borders bookstores t years since 2000. Since B = 2108 when t = 0
and we want the continuous growth rate, we write B = 2108e
kt
. We use the information from 2005, that B = 3880
when t = 5, to nd k:
3880 = 2108e
k5
1.841 = e
5k
ln(1.841) = 5k
k = 0.122.
We have B = 2108e
0.122t
, which represents a continuous growth rate of 12.2% per year.
1.5 SOLUTIONS 9
49. Let t = number of years since 2000. Then the number of vehicles, V, in millions, at time t is given by
V = 213(1.04)
t
and the number of people, P, in millions, at time t is given by
P = 281(1.01)
t
.
There is an average of one vehicle per person when
V
P
= 1, or V = P. Thus, we must solve for t in the equation:
213(1.04)
t
= 281(1.01)
t
,
which leads to
1.04
1.01
t
=
(1.04)
t
(1.01)
t
=
281
213
Taking logs on both sides, we get
t log
1.04
1.01
= log
281
213
,
so
t =
log (281/213)
log (1.04/1.01)
= 9.5 years.
This model predicts one vehicle per person in 2009.
53. We assume exponential decay and solve for k using the half-life:
e
k(5730)
= 0.5 so k = 1.21 10
4
.
Now nd t, the age of the painting:
e
1.2110
4
t
= 0.995, so t =
ln 0.995
1.21 10
4
= 41.43 years.
Since Vermeer died in 1675, the painting is a fake.
Exercises
1. See Figure 1.10.
sin
3
2
= 1 is negative.
cos
3
2
= 0
tan
3
2
is undened.
5. See Figure 1.11.
sin
is positive.
cos
is positive.
tan
is positive.
9. 1 radian
180
radians
=
180
60
. See Figure 1.12.

sin (1) is negative
cos (1) is positive
tan (1) is negative.
E
Figure 1.10
w
Figure 1.11
%
Figure 1.12
13. (a) We determine the amplitude of y by looking at the coefcient of the cosine term. Here, the coefcient is 1, so the
amplitude of y is 1. Note that the constant term does not affect the amplitude.
(b) We know that the cosine function cos x repeats itself at x = 2, so the function cos(3x) must repeat itself when
3x = 2, or at x = 2/3. So the period of y is 2/3. Here as well the constant term has no effect.
(c) The graph of y is shown in the gure below.
2
3
4
5
6
x
y
17. The period is 2/ = 2, since when t increases from 0 to 2, the value of t increases from 0 to 2. The amplitude is 0.1,
since the function oscillates between 1.9 and 2.1.
21. This graph is an inverted cosine curve with amplitude 8 and period 20, so it is given by f(x) = 8 cos
x
10
.
25. This can be represented by a sine function of amplitude 3 and period 18. Thus,
f(x) = 3 sin
9
x
.
29. We rst isolate cos(2x + 1) and then use inverse cosine:
1 = 8 cos(2x + 1) 3
4 = 8 cos(2x + 1)
0.5 = cos(2x + 1)
cos
1
(0.5) = 2x + 1
x =
cos
1
(0.5) 1
2
0.0236.
There are innitely many other possible solutions since the cosine is periodic.
1.6 SOLUTIONS 11
Problems
33. sin x
2
is by convention sin(x
2
), which means you square the x rst and then take the sine.
sin
2
x = (sin x)
2
means nd sin x and then square it.
sin(sin x) means nd sin x and then take the sine of that.
Expressing each as a composition: If f(x) = sin x and g(x) = x
2
, then
sin x
2
= f(g(x))
sin
2
x = g(f(x))
sin(sin x) = f(f(x)).
37. 200 revolutions per minute is
1
200
minutes per revolution, so the period is
1
200
minutes, or 0.3 seconds.
41. (a) D = the average depth of the water.
(b) A = the amplitude = 15/2 = 7.5.
(c) Period = 12.4 hours. Thus (B)(12.4) = 2 so B = 2/12.4 0.507.
(d) C is the time of a high tide.
(b) Average value of population =
700+900
2
= 800, amplitude =
900700
2
= 100, and period = 12 months, so B =
2/12 = /6. Since the population is at its minimum when t = 0, we use a negative cosine:
P = 800 100 cos
t
6
.
Jan. 1 Jul. 1 Jan. 1
100
700
800
900
t (months)
P
Figure 1.13
w
E '
x
E '
x
T
c
h
T
c
h
Figure 1.14
49. Figure 1.14 shows that the cross-sectional area is one rectangle of area hw and two triangles. Each triangle has height h
and base x, where
h
x
= tan so x =
h
tan
.
Area of triangle =
1
2
xh =
h
2
2 tan
Total area = Area of rectangle + 2(Area of triangle)
= hw + 2
h
2
2 tan
= hw +
h
2
tan
.
Exercises
1. As x , y .
As x , y .
5. As x , 0.25x
1/2
is larger than 25,000x
3
.
9. f(x) = k(x+3)(x1)(x4) = k(x
3
2x
2
11x+12), where k < 0. (k
1
6
if the horizontal and vertical scales
are equal; otherwise one cant tell how large k is.)
Problems
13. (a) II and III because in both cases, the numerator and denominator each have x
2
as the highest power, with coefcient
= 1. Therefore,
y
x
2
x
2
= 1 as x .
(b) I, since
y
x
x
2
= 0 as x .
(c) II and III, since replacing x by x leaves the graph of the function unchanged.
(d) None
(e) III, since the denominator is zero and f(x) tends to when x = 1.
17. (a) (i) The water that has owed out of the pipe in 1 second is a cylinder of radius r and length 3 cm. Its volume is
V = r
2
(3) = 3r
2
.
(ii) If the rate of ow is k cm/sec instead of 3 cm/sec, the volume is given by
V = r
2
(k) = r
2
k.
(b) (i) The graph of V as a function of r is a quadratic. See Figure 1.15.
r
V
Figure 1.15
k
V
Figure 1.16
(ii) The graph of V as a function of k is a line. See Figure 1.16.
21. (a) (i) If (1, 1) is on the graph, we know that
1 = a(1)
2
+ b(1) + c = a + b + c.
(ii) If (1, 1) is the vertex, then the axis of symmetry is x = 1, so
b
2a
= 1,
and thus
a =
b
2
, so b = 2a.
But to be the vertex, (1, 1) must also be on the graph, so we know that a + b + c = 1. Substituting b = 2a,
we get a + c = 1, which we can rewrite as a = c 1, or c = 1 + a.
(iii) For (0, 6) to be on the graph, we must have f(0) = 6. But f(0) = a(0
2
) + b(0) + c = c, so c = 6.
(b) To satisfy all the conditions, we must rst, from (a)(iii), have c = 6. From (a)(ii), a = c 1 so a = 5. Also from
(a)(ii), b = 2a, so b = 10. Thus the completed equation is
y = f(x) = 5x
2
10x + 6,
which satises all the given conditions.
1.7 SOLUTIONS 13
25. We use the fact that at a constant speed, Time = Distance/Speed. Thus,
Total time = Time running + Time walking
=
3
x
+
6
x 2
.
Horizontal asymptote: x-axis.
Vertical asymptote: x = 0 and x = 2.
29. (a) R(P) = kP(L P), where k is a positive constant.
(b) A possible graph is in Figure 1.17.
L
P
R
Figure 1.17
Exercises
1. Yes, because 2x + x
2/3
is dened for all x.
5. Yes, because 2x 5 is positive for 3 x 4.
9. No, because e
x
1 = 0 at x = 0.
13. We have that f(0) = 1 < 0 and f(1) = 1 cos 1 > 0 and that f is continuous. Thus, by the Intermediate Value
Theorem applied to k = 0, there is a number c in [0, 1] such that f(c) = k = 0.
Problems
17. The voltage f(t) is graphed in Figure 1.18.
7
6
12
time (seconds)
Figure 1.18: Voltage change from 6V to 12V
Using formulas, the voltage, f(t), is represented by
f(t) =
6, 0 < t 7
12, 7 < t
Although a real physical voltage is continuous, the voltage in this circuit is well-approximated by the function f(t), which
is not continuous on any interval around 7 seconds.
21. For x > 0, we have |x| = x, so f(x) = 1. For x < 0, we have |x| = x, so f(x) = 1. Thus, the function is given by
f(x) =
1 x > 0
0 x = 0
1 x < 1 ,
so f is not continuous on any interval containing x = 0.
25. (a) The graphs of y = e
x
and y = 4 x
2
cross twice in Figure 1.19. This tells us that the equation e
x
= 4 x
2
has two
solutions.
Since y = e
x
increases for all x and y = 4 x
2
increases for x < 0 and decreases for x > 0, these are only
the two crossing points.
2 2
y = e
x
y = x
2
4
x
y
Figure 1.19
(b) Values of f(x) are in Table 1.2. One solution is between x = 2 and x = 1; the second solution is between x = 1
and x = 2.
Table 1.2
x 4 3 2 1 0 1 2 3 4
f(x) 12.0 5.0 0.1 2.6 3 0.3 7.4 25.1 66.6
29. The drug rst increases linearly for half a second, at the end of which time there is 0.6 ml in the body. Thus, for 0 t
0.5, the function is linear with slope 0.6/0.5 = 1.2:
Q = 1.2t for 0 t 0.5.
At t = 0.5, we have Q = 0.6. For t > 0.5, the quantity decays exponentially at a continuous rate of 0.002, so Q has the
form
Q = Ae
0.002t
0.5 < t.
We choose A so that Q = 0.6 when t = 0.5:
0.6 = Ae
0.002(0.5)
= Ae
0.001
A = 0.6e
0.001
.
Thus
Q =
1.2t 0 t 0.5
0.6e
0.001
e
.002t
0.5 < t.
Exercises
1. (a) As x approaches 2 from either side, the values of f(x) get closer and closer to 3, so the limit appears to be about 3.
(b) As x approaches 0 from either side, the values of f(x) get closer and closer to 7. (Recall that to nd a limit, we are
interested in what happens to the function near x but not at x.) The limit appears to be about 7.
(c) As x approaches 2 from either side, the values of f(x) get closer and closer to 3 on one side of x = 2 and get closer
and closer to 2 on the other side of x = 2. Thus the limit does not exist.
(d) As x approaches 4 from either side, the values of f(x) get closer and closer to 8. (Again, recall that we dont care
what happens right at x = 4.) The limit appears to be about 8.
1.8 SOLUTIONS 15
5. For 1 x 1, 1 y 1, the graph of y = xln |x| is in Figure 1.20. The graph suggests that
lim
x0
xln |x| = 0.
9. For 90
90
, 0 y 0.02, the graph of y =

sin
is shown in Figure 1.21. Therefore, by tracing along the

curve, we see that in degrees, lim
0
sin
= 0.01745 . . ..
1 1
1
1
x
y
y = xln |x|
Figure 1.20
90 0 90
0.01
0.02
(degree)
sin
Figure 1.21
1 2 3 4 5
2
0
2
4
6
8
10
x
f(x)
Figure 1.22
13. f(x) =
x
2
2 0 < x < 3
2 x = 3
2x + 1 3 < x
Figure 1.22 conrms that lim
x3
f(x) = lim
x3
(x
2
2) = 7 and that lim
x3+
f(x) = lim
x3+
(2x + 1) = 7, so lim
x3
f(x) =
7. Note, however, that f(x) is not continuous at x = 3 since f(3) = 2.
Problems
17. Since lim
x0
+ f(x) = 1 and lim
x0
f(x) = 1, we see that limx0 f(x) does not exist. Thus, f(x) is not continu-
ous at x = 0
21. When x = 0.1, we nd xe
1/x
2203. When x = 0.01, we nd xe
1/x
3 10
41
. When x = 0.001, the value of
xe
1/x
is too big for a calculator to compute. This suggests that lim
x0
+
xe
1/x
does not exist (and in fact it does not).
25. If x > 1 and x approaches 1, then p(x) = 55. If x < 1 and x approaches 1, then p(x) = 34. There is not a single number
that p(x) approaches as x approaches 1, so we say that lim
x1
p(x) does not exist.
29. From Table 1.3, it appears the limit is 1. This is conrmed by Figure 1.23. An appropriate window is 0.099 < x <
0.099, 1.01 < y < 0.99.
Table 1.3
x f(x)
0.1 0.99
0.01 0.9999
0.001 0.999999
0.0001 0.99999999
x f(x)
0.0001 0.99999999
0.001 0.999999
0.01 0.9999
0.1 0.99
0.099 0.099
1.01
0.99
Figure 1.23
0.047, 2.99 < y < 3.01.
Table 1.4
x f(x)
0.1 2.9552
0.01 2.9996
0.001 3.0000
0.0001 3.0000
x f(x)
0.0001 3.0000
0.001 3.0000
0.01 2.9996
0.1 2.9552
0.047 0.047
2.99
3.01
Figure 1.24
37. Divide numerator and denominator by x:
f(x) =
+ 3x
x 3
=
( + 3x)/x
(x 3)/x
,
so
lim
x
f(x) = lim
x
/x + 3
3/x
=
limx(/x + 3)
limx( 3/x)
=
3
.
41. Divide numerator and denominator by x
3
, giving
f(x) =
2x
3
16x
2
4x
2
+ 3x
3
=
2 16/x
4/x + 3
,
so
lim
x
f(x) = lim
x
2 16/x
4/x + 3
=
limx(2 16/x)
limx(4/x + 3)
=
2
3
.
45. f(x) =
2e
x
+ 3
3e
x
+ 2
, so lim
x
f(x) =
limx(2e
x
+ 3)
limx(3e
x
+ 2)
=
3
2
.
49. Division of numerator and denominator by x
2
yields
x
2
+ 3x + 5
4x + 1 + x
k
=
1 + 3/x + 5/x
2
4/x + 1/x
2
+ x
k2
.
As x , the limit of the numerator is 1. The limit of the denominator depends upon k. If k > 2, the denominator
approaches as x , so the limit of the quotient is 0. If k = 2, the denominator approaches 1 as x , so the
limit of the quotient is 1. If k < 2 the denominator approaches 0
+
as x , so the limit of the quotient is . Therefore
the values of k we are looking for are k 2.
53. In the denominator, we have lim
x
3
2x
+ 4 = 4. In the numerator, if k < 0, we have lim
x
3
kx
+ 6 = , so the
quotient has a limit of . If k = 0, we have lim
x
3
kx
+ 6 = 7, so the quotient has a limit of 7/4. If k > 0, we have
lim
x
3
kx
+ 6 = 6, so the quotient has a limit of 6/4.
57. (a) Since sin(n) = 0 for n = 1, 2, 3, . . . the sequence of x-values
1
,
1
2
,
1
3
, . . .
works. These x-values 0 and are zeroes of f(x).
(b) Since sin(n/2) = 1 for n = 1, 5, 9 . . . the sequence of x-values
2
,
2
5
,
2
9
, . . .
works.
(c) Since sin(n)/2 = 1 for n = 3, 7, 11, . . . the sequence of x-values
2
3
,
2
7
,
2
11
. . .
works.
(d) Any two of these sequences of x-values show that if the limit were to exist, then it would have to have two (different)
values: 0 and 1, or 0 and 1, or 1 and 1. Hence, the limit can not exist.
SOLUTIONS to Review Problems for Chapter One 17
61. We will show f(x) = x is continuous at x = c. Since f(c) = c, we need to show that
lim
xc
f(x) = c
that is, since f(x) = x, we need to show
lim
xc
x = c.
Pick any > 0, then take = . Thus,
|f(x) c| = |x c| < for all |x c| < = .
Exercises
1. The line of slope m through the point (x0, y0) has equation
y y0 = m(x x0),
so the line we want is
y 0 = 2(x 5)
y = 2x 10.
5. A circle with center (h, k) and radius r has equation (x h)
2
+(y k)
2
= r
2
. Thus h = 1, k = 2, and r = 3, giving
(x + 1)
2
+ (y 2)
2
= 9.
Solving for y, and taking the positive square root gives the top half, so
(y 2)
2
= 9 (x + 1)
2
y = 2 +
9 (x + 1)
2
.
See Figure 1.25.
(1, 2)
E '
3
x
y
Figure 1.25: Graph of y = 2 +
9 (x + 1)
2
9. Since T = f(P), we see that f(200) is the value of T when P = 200; that is, the thickness of pelican eggs when the
concentration of PCBs is 200 ppm.
13. (a) The equation is y = 2x
2
+ 1. Note that its graph is narrower than the graph of y = x
2
which appears in gray. See
Figure 1.26.
2
4
6
8 y = 2x
2
+ 1
y = x
2
Figure 1.26
1
2
3
4
5
6
7
8 y = 2(x
2
+ 1)
y = x
2
Figure 1.27
(b) y = 2(x
2
+ 1) moves the graph up one unit and then stretches it by a factor of two. See Figure 1.27.
(c) No, the graphs are not the same. Since 2(x
2
+ 1) = (2x
2
+ 1) + 1, the second graph is always one unit higher than
the rst.
17. (a) It was decreasing from March 2 to March 5 and increasing from March 5 to March 9.
(b) From March 5 to 8, the average temperature increased, but the rate of increase went down, from 12
between March
5 and 6 to 4
between March 6 and 7 to 2
between March 7 and 8.

From March 7 to 9, the average temperature increased, and the rate of increase went up, from 2
between March
7 and 8 to 9
between March 8 and 9.

21. (a) A polynomial has the same end behavior as its leading term, so this polynomial behaves as 5x
4
globally. Thus we
have:
f(x) as x , and f(x) as x +.
(b) Polynomials behave globally as their leading term, so this rational function behaves globally as (3x
2
)/(2x
2
), or 3/2.
Thus we have:
f(x) 3/2 as x , and f(x) 3/2 as x +.
(c) We see from a graph of y = e
x
that
f(x) 0 as x , and f(x) +as x +.
25. Collecting similar factors yields
1.04
1.03
t
=
12.01
5.02
. Solving for t yields
t =
log
12.01
5.02
log
1.04
1.03
= 90.283.
29. The amplitude is 2. The period is 2/5. See Figure 1.28.
4
5

2
5
2
5
4
5
2
4
6
y = 4 2 cos(5x)
x
y
Figure 1.28
33. y = kx(x + 5) = k(x
2
+ 5x), where k > 0 is any constant.
37. x = ky(y 4) = k(y
2
4y), where k > 0 is any constant.
41. There are many solutions for a graph like this one. The simplest is y = 1 e
x
, which gives the graph of y = e
x
, ipped
over the x-axis and moved up by 1. The resulting graph passes through the origin and approaches y = 1 as an upper
bound, the two features of the given graph.
45. f(x) = x
3
, g(x) = ln x.
SOLUTIONS to Review Problems for Chapter One 19
49. f(x) =
e
x
1 < x < 0
1 x = 0
cos x 0 < x < 1
Figure 1.29 conrms that lim
x0
f(x) = lim
x0
e
x
= e
0
= 1, and that lim
x0+
f(x) = lim
x0+
cos x = cos 0 = 1, so
lim
x0
f(x) = 1.
1 0 1
1
x
f(x)
Figure 1.29
Problems
53. (a) The amplitude of the sine curve is |A|. Thus, increasing |A| stretches the curve vertically. See Figure 1.30.
(b) The period of the wave is 2/|B|. Thus, increasing |B| makes the curve oscillate more rapidlyin other words, the
function executes one complete oscillation in a smaller interval. See Figure 1.31.
2
2
3
1
2
3
E
A = 1
E
A = 2
E
A = 3
x
y
Figure 1.30
2
2
B = 2
c
B = 1
B = 3
1
1
x
y
Figure 1.31
57. (a) This could be a linear function because w increases by 5 as h increases by 1.
(b) We nd the slope m and the intercept b in the linear equation w = b + mh. We rst nd the slope m using the rst
two points in the table. Since we want w to be a function of h, we take
m =
w
h
=
171 166
69 68
= 5.
Substituting the rst point and the slope m = 5 into the linear equation w = b + mh, we have 166 = b + (5)(68),
so b = 174. The linear function is
w = 5h 174.
The slope, m = 5, is in units of pounds per inch.
(c) We nd the slope and intercept in the linear function h = b + mw using m = h/w to obtain the linear function
h = 0.2w + 34.8.
Alternatively, we could solve the linear equation found in part (b) for h. The slope, m = 0.2, has units inches per
pound.
61. We can solve for the growth rate k of the bacteria using the formula P = P0e
kt
:
1500 = 500e
k(2)
k =
ln(1500/500)
2
.
Knowing the growth rate, we can nd the population P at time t = 6:
P = 500e
(
ln 3
2
)6
13,500 bacteria.
65. We will let
T = amount of fuel for take-off,
L = amount of fuel for landing,
P = amount of fuel per mile in the air,
m = the length of the trip in miles.
Then Q, the total amount of fuel needed, is given by
Q(m) = T + L + Pm.
69. (a) Beginning at time t = 0, the voltage will have oscillated through a complete cycle when cos(120t) = cos(2),
hence when t =
1
60
second. The period is
1
60
second.
(b) V0 represents the amplitude of the oscillation.
(c) See Figure 1.32.
1
120
1
60
V
0
t
V
Figure 1.32
0.015, 0.01 < y < 0.01.
Table 1.5
x f(x)
0.1 0.0666
0.01 0.0067
0.001 0.0007
0.0001 0.0001
x f(x)
0.0001 0.0001
0.001 0.0007
0.01 0.0067
0.1 0.0666
0.015 0.015
0.01
0.01
Figure 1.33
77. (a) A CAS gives f(x) = (x a)(x + a)(x + b)(x c).
(b) The graph of f(x) crosses the x-axis at x = a, x = a, x = b, x = c; it crosses the y-axis at a
2
bc. Since the
coefcient of x
4
(namely 1) is positive, the graph of f looks like that shown in Figure 1.34.
b a
a c
a
2
bc
x
y
Figure 1.34: Graph of
f(x) =
(xa)(x+a)(x+b)(xc)
81. (a) A CAS or division gives
f(x) =
x
3
30
x 3
= x
2
+ 3x + 9
3
x 3
,
so p(x) = x
2
+ 3x + 9, and r(x) = 3, and q(x) = x 3.
(b) The vertical asymptote is x = 3. Near x = 3, the values of p(x) are much smaller than the values of r(x)/q(x).
Thus
f(x)
3
x 3
for x near 3.
(c) For large x, the values of p(x) are much larger than the value of r(x)/q(x). Thus
f(x) x
2
+ 3x + 9 as x , x .
(d) Figure 1.35 shows f(x) and y = 3/(x 3) for x near 3. Figure 1.36 shows f(x) and y = x
2
+ 3x + 9 for
20 x 20. Note that in each case the graphs of f and the approximating function are close.
5
50
50
f(x)
f(x)
y =
3
x3
c
y =
3
x3
x
y
Figure 1.35: Close-up view of f(x) and
y = 3/(x 3)
20 20
500
500
f(x)
y = x
2
+ 3x + 9
x
y
Figure 1.36: Far-away view of f(x) and
y = x
2
+ 3x + 9
1. False. A line can be put through any two points in the plane. However, if the line is vertical, it is not the graph of a
function.
5. True. The highest degree term in a polynomial determines how the polynomial behaves when x is very large in the positive
or negative direction. When n is odd, x
n
is positive when x is large and positive but negative when x is large and negative.
Thus if a polynomial p(x) has odd degree, it will be positive for some values of x and negative for other values of x. Since
every polynomial is continuous, the Intermediate Value Theorem then guarantees that p(x) = 0 for some value of x.
9. False. Suppose y = 5
x
. Then increasing x by 1 increases y by a factor of 5. However increasing x by 2 increases y by a
factor of 25, not 10, since
y = 5
x+2
= 5
x
5
2
= 25 5
x
.
(Other examples are possible.)
13. True. The period is 2/(200) = 1/100 seconds. Thus, the function executes 100 cycles in 1 second.
17. False: When /2 < x < 3/2, we have cos |x| = cos x < 0 but | cos x| > 0.
21. False. A counterexample is given by f(x) = sin x, which has period 2, and g(x) = x
2
. The graph of f(g(x)) = sin(x
2
)
in Figure 1.37 is not periodic with period 2.
2

2
1
1
f(x)
x
y
Figure 1.37
25. True. If f is increasing then its reection about the line y = x is also increasing. An example is shown in Figure 1.38.
The statement is true.
4 6
4
4
f(x)
x = y
f
1
(x)
x
y
Figure 1.38
29. True. If b > 1, then ab
x
0 as x . If 0 < b < 1, then ab
x
0 as x . In either case, the function
y = a + ab
x
has y = a as the horizontal asymptote.
33. False. A counterexample is given by f(x) = x
2
and g(x) = x+1. The function f(g(x)) = (x+1)
2
is not even because
f(g(1)) = 4 and f(g(1)) = 0 = 4.
37. Let f(x) =
1
x + 7
. Other answers are possible.
41. Let f(x) = x and g(x) = 2x. Then f(x) + g(x) = x, which is decreasing. Note f is increasing since it has positive
slope, and g is decreasing since it has negative slope.
45. False. For example, f(x) = x/(x
2
+ 1) has no vertical asymptote since the denominator is never 0.
49. False. For example, if y = 4x+1 (so m = 4) and x = 1, then y = 5. Increasing x by 2 units gives 3, so y = 4(3) +1 =
13. Thus, y has increased by 8 units, not 4 + 2 = 6. (Other examples are possible.)
53. True. The constant function f(x) = 0 is the only function that is both even and odd. This follows, since if f is both even
and odd, then, for all x, f(x) = f(x) (if f is even) and f(x) = f(x) (if f is odd). Thus, for all x, f(x) = f(x)
i.e. f(x) = 0, for all x. So f(x) = 0 is both even and odd and is the only such function.
57. True, by Property 2 of limits in Theorem 1.2.
61. True. Suppose instead that limx3 g(x) does not exist but limx3(f(x)g(x)) did exist. Since limx3 f(x) exists and is
not zero, then limx3((f(x)g(x))/f(x)) exists, by Property 4 of limits in Theorem 1.2. Furthermore, f(x) = 0 for all x
in some interval about 3, so (f(x)g(x))/f(x) = g(x) for all x in that interval. Thus limx3 g(x) exists. This contradicts
our assumption that limx3 g(x) does not exist.
65. False. Although x may be far from c, the value of f(x) could be close to L. For example, suppose f(x) = L, the constant
function.
20.1 SOLUTIONS 423
CHAPTER TWENTY
Exercises
1. Scalar. Since
div
_
y
i x
j
x
2
+ y
2
_
=
y 2x
(x
2
+ y
2
)
2
+
x 2y
(x
2
+ y
2
)
2
= 0.
5. div

F =

x
(x + y) +

y
(y + z) +

z
(z + x) = 1 + 1 1 = 1
9. Using the formula for a r in Cartesian coordinates, we get
div

F =

x
(a2z a3y) +

y
(a3x a1z) +

z
(a1y a2x) = 0
13. Two vector elds that have positive divergence everywhere are in Figures 20.1 and 20.2.
x
y
Figure 20.1
x
y
Figure 20.2
Problems
17. Since div F(1, 2, 3) is the ux density out of a small region surrounding the point (1, 2, 3), we have
div

F (1, 2, 3)
Flux out of small region around (1, 2, 3)
Volume of region.
So
Flux out of region (div

F (1, 2, 3)) Volume of region
= 5
4
3
(0.01)
3
=
0.00002
3
.
21. Using ux: On S1, x = a and normal is in negative x-direction, so
A = ((3a + 2)
i + 4a
j + (5a + 1)
k ) (A
i ) = (3a + 2)A
Thus
_
S
1
F d
A =
_
S
1
(3a + 2)dA = (3a + 2)(Area of S1) = (3a + 2)w
2
.
On S2, x = a + w and normal is in the positive x-direction, so
A = [(3(a + w) + 2)
i + 4(a + w)
j + (5(a + w) + 1)
k ] (A
i ) = (3a + 3w + 2)A.
Thus
_
S
2
F d
A =
_
S
2
(3a + 3w + 2)dA = (3a + 3w + 2)(Area of S2) = (3a + 3w + 2)w
2
.
424 Chapter Twenty /SOLUTIONS
Next, we have
_
S
3
F d
A =
_
S
3
4xdA and
_
S
4
F d
A =
_
S
4
4xdA. Since these two are integrated over the same
region in the xz-plane, the two integrals cancel. Similarly,
_
S
5
F d
A =
_
S
5
(5x + 1)d
A cancels out
_
S
6
F d
A =
_
S
6
(5x + 1)d
A . Therefore,
Total ux
=
_
S
1
F d
A +
_
S
2
F d
A +
_
S
3
F d
A +
_
S
4
F d
A +
_
S
5
F d
A +
_
S
6
F d
A
= (3a + 2)w
2
+ (3a + 3w + 2)w
2
+
_
S
3
4xdA +
_
S
4
4xdA
+
_
S
5
(5x + 1)dA +
_
S
6
(5x + 1)dA = 3w
3
.
To nd div

F at the point (a, b, c), let the box shrink to the point by letting w 0. Then
div

F = lim
w0
_
Flux through box
Volume of box
_
= lim
w0
_
3w
3
w
3
_
= 3.
Using partial derivatives:
div

F =

x
(3x + 2) +

y
(4x) +

z
(5x + 1) = 3
25. (a) div

B =

x
(y) +

y
(x) +

z
(x + y) = 0, so this could be a magnetic eld.
(b) div

B =

x
(z) +

y
(y) +

z
(x) = 0 + 1 + 0 = 1, so this could not be a magnetic eld.
(c) div

B =

x
(x
2
y
2
x) +

y
(y 2xy) +

z
(0) = 2x1 +1 2x+0 = 0, so this could be a magnetic eld.
29. Let

F = F1
i + F2
j + F3
k . Then
div(g
F ) = div(gF1
i + gF2
j + gF3
k )
=

x
(gF1) +

y
(gF2) +

z
(gF3)
=
g
x
F1 + g
F1
x
+
g
y
F2 + g
F2
y
+
g
z
F3 + g
F3
z
=
g
x
F1 +
g
y
F2 +
g
z
F3 + g
_
F1
x
+
F2
y
+
F3
z
_
= (grad g)

F + g div

F .
33. Using div(g
F ) = (grad g)

F + g div

F , we have
div

G = grad(
b r ) (a r ) +
b r div(a r ) =
b (a r ) +v r 0 =
b (a r ).
37. (a) The velocity vector for the trafc ow would look like:
(b) When 0 x < 2000, the velocity is decreasing linearly from 55 to 15, so its formula is (55 x/50)
i mph. Then,
when 2000 x < 7000, the speed is constant, so v (x) = 15
i mph. Next, when 7000 x < 8000, the velocity

is increasing linearly from 15 to 55, so v (x) = (15 + (x 7000)/25)
i mph. Finally, when x 8000, the speed is

constant, so v (x) = 55
i mph.
20.2 SOLUTIONS 425
(c) divv = dv(x)/dx.
At x = 1000, v(x) = 55 x/50, so divv = 1/50.
At x = 5000, v(x) = 15, so divv = 0.
At x = 7500, v(x) = 15 + (x 7000)/25, so divv = 1/25.
At x = 10, 000, v(x) = 55, so divv = 0.
In each case the units of divv are
miles/hour
feet
.
41. (a) At any point r = x
i + y
j , the direction of the vector eld v is pointing toward the origin, which means it is of
the formv = fr for some negative function f whose value can vary depending on r . The magnitude of v depends
only on the distance r, thus f must be a function depending only on r, which is equivalent to depending only on r
2
since r 0. So v = f(r
2
)r =
_
f(x
2
+ y
2
)
_
(x
i + y
j ).
(b) At (x, y) = (0, 0) the divergence of v is
divv =
(K(x
2
+ y
2
)
1
x)
x
+
(K(x
2
+ y
2
)
1
y)
y
=
Ky
2
Kx
2
(x
2
+ y
2
)
2
+
Kx
2
Ky
2
(x
2
+ y
2
)
2
= 0.
Therefore, v is a point sink at the origin.
(c) The magnitude of v is
v = |K|(x
2
+ y
2
)
1
|x
i + y
j | = |K|(x
2
+ y
2
)
1
(x
2
+ y
2
)
1/2
= |K|(x
2
+ y
2
)
1/2
=
|K|
r
.
(remember, K < 0)
(d) The vector eld looks like the following:
Figure 20.3
(e) We need to show that grad = v .
grad =

x
(
K
2
log(x
2
+ y
2
))
i +

y
(
K
2
log(x
2
+ y
2
))
j
=
Kx
x
2
+ y
2
i +
Ky
x
2
+ y
2
j
= K(x
2
+ y
2
)
1
(x
i + y
j )
= v
Exercises
1. First directly: On the faces x = 0, y = 0, z = 0, the ux is zero. On the face x = 2, a unit normal is
i and d
A = dA
i .
So
_
S
x=2
r d
A =
_
S
x=2
(2
i + y
j + z
k ) (dA
i )
(since on that face, x = 2)
=
_
S
x=2
2dA = 2 (Area of face) = 2 4 = 8.
In exactly the same way, you get
_
S
y=2
r d
A =
_
S
z=2
r d
A = 8,
so
_
S
r
dA = 3 8 = 24.
Now using divergence:
div

F =
x
x
+
y
y
+
z
z
= 3,
so
Flux =
_
2
0
_
2
0
_
2
0
3 dx dy dz = 3 (Volume of Cube) = 3 8 = 24
5. The location of the pyramid has not been completely specied. For instance, where is it centered on the xy plane? How
is base oriented with respect to the axes? Thus, we cannot compute the ux by direct integration with the information we
have. However, we can calculate it using the divergence theorem. First we calculate the divergence of

F .
div

F =
(z)
x
+
0
y
+
x
z
= 0 + 0 + 0 = 0
Thus for any closed surface the ux will be zero, so the ux through our pyramid, regardless of its location or orientation,
is zero.
9. Since div

G = 1, if W is the interior of the box, the Divergence Theorem gives
Flux =
_
W
1 dV = 1 Volume of box = 1 2 3 4 = 24.
13. We have
div((3x + 4y)
i + (4y + 5z)
j + (5z + 3x)
k ) = 3 + 4 + 5 = 12.
Let W be the interior of the cube. Then by the divergence theorem,
_
S
((3x + 4y)
i + (4y + 5z)
j + (5z + 3x)
k ) d
A =
_
W
12 dV = 12 Volume of cube = 12 (2 3 4) = 288.
Problems
17. Since div

F = 3x
2
+ 3y
2
+ 3z
2
, the Divergence Theorem gives
Flux =
_
S
F d
A =
_
W
(3x
2
+ 3y
2
+ 3z
2
) dV.
In spherical coordinates, the region W lies between the spheres = 2 and = 3 and inside the cone = /4. Since
3x
2
+ 3y
2
+ 3z
2
= 3
2
, we have
Flux =
_
S
F d
A =
_
2
0
_
/4
0
_
3
2
3
2

2
sin d dd
= 2
3
5
3
2
(cos )
/4
0
=
633(2
2)
5
= 232.98.
21. By the Divergence Theorem,
_
S
F d
A =
_
W
div
F dV =
_
W
0dV = 0 for a closed surface S, where W is the region
enclosed by S.
25. We consider a sphere of radius R centered at the origin and compute the ux of r through its surface. Since r and d

A
both point radially outward, r d
A = r d
A = RdA on the surface of the sphere, so

_
S
F d
A =
_
S
RdA = R
_
S
dA = R(Surface area of a sphere) = R(4R
2
) = 4R
3
.
Therefore, volume of sphere =
1
3
_
S
F d
A =
4
3
R
3
.
20.2 SOLUTIONS 427
29. (a) Since

F is radial, it is everywhere parallel to the area vector,

A . Also, ||
F || = 1 on the surface of the sphere

x
2
+ y
2
+ z
2
= 1, so
Flux through the sphere =
_
S
F d
A = lim
||
A ||0
A
= lim
||
A ||0
||
F || ||
A|| = lim
||
A ||0
||
A||
= Surface area of sphere = 4 1
2
= 4.
(b) In Cartesian coordinates,
F (x, y, z) =
x
(x
2
+ y
2
+ z
2
)
3/2
i +
y
(x
2
+ y
2
+ z
2
)
3/2
j +
z
(x
2
+ y
2
+ z
2
)
3/2
k .
So,
div

F (x, y, z) =
_
1
(x
2
+ y
2
+ z
2
)
3/2

3x
2
(x
2
+ y
2
+ z
2
)
5/2
_
+
_
1
(x
2
+ y
2
+ z
2
)
3/2

3y
2
(x
2
+ y
2
+ z
2
)
5/2
_
+
_
1
(x
2
+ y
2
+ z
2
)
3/2

3z
2
(x
2
+ y
2
+ z
2
)
5/2
_
=
_
x
2
+ y
2
+ z
2
(x
2
+ y
2
+ z
2
)
5/2

3x
2
(x
2
+ y
2
+ z
2
)
5/2
_
+
_
x
2
+ y
2
+ z
2
(x
2
+ y
2
+ z
2
)
5/2

3y
2
(x
2
+ y
2
+ z
2
)
5/2
_
+
_
x
2
+ y
2
+ z
2
(x
2
+ y
2
+ z
2
)
5/2

3z
2
(x
2
+ y
2
+ z
2
)
5/2
_
=
3(x
2
+ y
2
+ z
2
) 3(x
2
+ y
2
+ z
2
)
(x
2
+ y
2
+ z
2
)
5/2
= 0.
(c) We cannot apply the Divergence Theorem to the whole region within the box, because the vector eld

F is not
dened at the origin. However, we can apply the Divergence Theorem to the region, W, between the sphere and the
box. Since div

F = 0 there, the theorem tells us that
_
Box
(outward)
F d
A +
_
Sphere
(inward)
F d
A =
_
W
div

F dV = 0.
Therefore, the ux through the box and the sphere are equal if both are oriented outward:
_
Box
(outward)
F d
A =
_
Sphere
(inward)
F d
A =
_
Sphere
(outward)
F d
A = 4.
33. (a) At the point (1, 2, 1), we have div

F = 1 2 1
2
= 2.
(b) Since the box is small, we use the approximation
div

F = Flux density
Flux out of box
Volume of box
.
Thus
Flux out of box (div

F ) (Volume of box) = 2(0.2)
3
= 0.016.
(c) To calculate the ux exactly, we use the Divergence Theorem,
Flux out of box =
_
Box
div

F dV =
_
Box
xyz
2
dV.
Since the box has side 0.2, it is given by 0.9 < x < 1.1, 1.9 < y < 2.1, 0.9 < z < 1.1, so
Flux =
_
1.1
0.9
_
2.1
1.9
_
1.1
0.9
xyz
2
dzdydx =
_
1.1
0.9
_
2.1
1.9
xy
z
3
3
1.1
0.9
dydx
=
(1.1)
3
(0.9)
3
3
_
1.1
0.9
xy
2
2
2.1
1.9
dx =
(1.1)
3
(0.9)
3
3

(2.1)
2
(1.9)
2
2

x
2
2
1.1
0.9
=
(1.1)
3
(0.9)
3
3

(2.1)
2
(1.9)
2
2

(1.1)
2
(0.9)
2
2
= 0.016053 . . . .
Notice that you can calculate the ux without knowing the vector eld,

F .
37. (a) The rate at which heat is generated at any point in the earth is div

F at that point. So div

F = 30 watts/km
3
.
(b) Differentiating gives div((x
i +y
j +z
k )) = (1 + 1 + 1) = 3 so = 30/3 = 10 watts/km
3
. Thus,

F = r
has constant divergence. Note that

F = r has ow lines going radially outward, and symmetric about the origin.
(c) The vector grad T gives the direction of greatest increase in temperature. Thus, grad T gives the direction of
greatest decrease in temperature. The equation

F = k grad T says that heat will ow in the direction of greatest
decrease in temperature (i.e. from hot regions to cold), and at a rate proportional to the temperature gradient.
(d) We assume that

F is given by the answer to part (b). Then, using part (c), we have
F = 10(x
i + y
j + z
k ) = 30,000 gradT,
so
grad T =
10
30,000
(x
i + y
j + z
k ).
Integrating we get
T =
10
2(30,000)
(x
2
+ y
2
+ z
2
) + C.
At the surface of the earth, x
2
+ y
2
+ z
2
= 6400
2
, and T = 20
C, so
T =
1
6000
(6400
2
) + C = 20.
Thus,
C = 20 +
6400
2
6000
= 6847.
At the center of the earth, x
2
+ y
2
+ z
2
= 0, so
T = 6847
C.
41. Check that div

E = 0 by taking partial derivatives. For instance,
E1
x
=

x
[q(x x0)[(x x0)
2
+ (y y0)
2
+ (z z0)
2
]
3/2
]
= q[(y y0)
2
+ (z z0)
2
2(x x0)
2
][(x x0)
2
+ (y y0)
2
+ (z z0)
2
]
5/2
and similarly,
E2
y
= q[(x x0)
2
+ (z z0)
2
2(y y0)
2
][(x x0)
2
+ (y y0)
2
+ (z z0)
2
]
5/2
E3
z
= q[(x x0)
2
+ (y y0)
2
2(z z0)
2
][(x x0)
2
+ (y y0)
2
+ (z z0)
2
]
5/2
.
Therefore,
E1
x
+
E2
y
+
E3
z
= 0.
The vector eld

E is dened everywhere but at the point with position vector r 0. If this point lies outside the surface S,
the Divergence Theorem can be applied to the region R enclosed by S, yielding:
_
S
E d
A =
_
R
div

E dV = 0.
20.3 SOLUTIONS 429
If the charge q is located inside S, consider a small sphere Sa centered at q and contained in R. The Divergence Theorem
for the region R
between the two spheres yields:

_
S
E d
A +
_
Sa
E d
A =
_
R
div

E dV = 0.
In this formula, the Divergence Theorem requires S to be given the outward orientation, and Sa the inward orientation. To
compute
_
Sa
E d
A , we use the fact that on the surface of the sphere,

E and
A are parallel and in opposite directions,

so
A =
A
since on the surface of a sphere of radius a,
E = q
r r 0
r r 0
3
=
q
a
2
.
Then,
_
Sa
E d
A =
_

q
a
2
d
A =
q
a
2
Surface area of sphere =
q
a
2
4a
2
= 4q.
_
Sa
E d
A = 4q.
_
S
E d
A
_
Sa
E d
A = 4q.
Exercises
1. Vector. We have
curl(z
i x
j + y
k ) =
i

j

k
z
z x y
i +
k .
5. Using the denition of Cartesian coordinates,
curl

F =
i

j

k
z
(x + y) (y + z) (z + x)
=
_

y
(z + x)

z
(y + z)
_
i +
_

x
(z + x) +

z
(x + y)
_
j
+
_

x
(y + z)

y
(x + y)
_
k
=
k .
9. This vector eld points radically outward and has unit length everywhere (except the origin). Thus, we would expect its
curl to be
0 . Computing the curl directly we get

curl
_
r
r
_
=
i

j

k
z
x
(x
2
+y
2
+z
2
)
1/2
y
(x
2
+y
2
+z
2
)
1/2
z
(x
2
+y
2
+z
2
)
1/2
The
i -component is given by =
_
1
2

2yz
(x
2
+ y
2
+ z
2
)
3/2

_
1
2

2yz
(x
2
+ y
2
+ z
2
)
1/2
__
i
=

0
Similarly, the
j and
k components are also both
0 .
13. This vector eld shows no rotation, and the circulation around any closed curve appears to be zero, so the vector eld has
zero curl.
Problems
17. Yes.

F = (1 + y
2
)
i has constant direction and curl

F = 2y
k =

0 . The ow lines of a vector eld do not have to
bend for it to have nonzero curl.
21. The curl is dened in such a way that if n is a unit vector and C is a small circle in the plane perpendicular to n and with
orientation induced by n , then
(curl

G ) n = Circulation density
_
C
G dr
Area inside C
so
Circulation =
_
C
G dr
_
(curl

G ) n
_
Area inside C.
(a) Let C be the circle in the xy-plane, and let n =

k . Then
Circulation (2
i 3
j + 5
k )
k (0.01)
2
= 0.0005.
(b) By a similar argument to part (a), with n =
i , we nd the circulation around the circle in the yz-plane:

Circulation (2
i 3
j + 5
k )
i (0.01)
2
= 0.0002.
(c) Similarly for circulations around the circle in the xz-plane,
Circulation 0.0003.
25. Let

C = a
i + b
j + c
k . Then
curl(
F +

C ) =
_

y
(F3 + c)

z
(F2 + b)
_
i +
_

z
(F1 + a)

x
(F3 + c)
_
j
+
_

x
(F2 + b)

y
(F1 + a)
_
k
=
_
F3
y

F2
z
_
i +
_
F1
z

F3
x
_
j +
_
F2
x

F1
y
_
k
= curl

F .
29. By Problem 28, curl

F = grad f grad g + f curl grad g = grad f gradg, since curl grad g = 0. Since the cross
product of two vectors is perpendicular to both vectors, curl

F is perpendicular to grad g. But

F is a scalar times grad g,
so curl

F is perpendicular to

F .
33. Investigate the velocity vector eld of the atmosphere near the re. If the curl of this vector eld is non-zero, there is
circulatory motion. Consequently, if the magnitude of the curl of this vector eld is large near the re, a re storm has
probably developed.
37. (a) Since

F is in the xy-plane, curl

F is parallel to

k (because F3 = 0 and F1, F2 have no z-dependence). Imagine
computing the circulation of

F counterclockwise around a small rectangle R at the point P with sides of length h
parallel to

F and sides of length t perpendicular to

F as shown in Figure 20.4. Since

F is perpendicular to C2 and
C4, the line integral over these two sides is zero. Assuming that

F is approximately constant on C1 and C3, its value
on these sides is F(Q)
T and F(P)
T , respectively. Thus, since

F is parallel to C1 and C3, the line integral over
C1 is approximately F(Q)h and the line integral over C3 is approximately F(P)h. Finally
curl

F (P)
Circulation around R
Area of R

F(Q)h F(P)h
ht
=
F(Q) F(P)
t
Directional derivative of F in the direction of
PQ.
20.4 SOLUTIONS 431
Q
P
C
1
h
h
t
t
C
4
C
2
C
3
Q
F
Figure 20.4: Path R used to nd curl

F at P
(b) Since

F = F(x, y)
T = F(x, y)a
i + F(x, y)b
j , with a, b constant, we have

curl

F = (bFx aFy)
k .
Also

T
k = (a
i + b
j )
k = b
i a
j , so
bFx aFy = (gradF) (b
i a
j ) = gradF ((a
i + b
j )
k ) = F
k
,
where F
k
is the directional derivative of F in the direction of the unit vector

T
k , which is perpendicular to
F . The right-hand rule applied to

T
k shows that

T
k is obtained by a clockwise rotation of

T through 90
.
Exercises
1. To calculate
_
C
F dr directly, we compute the integral along each of the sides C1, C2, C3 in Figure 20.5. Now C1 is
parameterized by
x(t) = t, y(t) = 0, z(t) = 0 for 0 t 5, so r
(t) =
i .
Similarly, C2 is parameterized by
x(t) = 5, y(t) = t, z(t) = 0 for 0 t 5, so r

(t) =
j .
Also, C3 is parameterized by
x(t) = 5 t, y(t) = 5 t, z(t) = 0 for 0 t 5, so r

(t) =
j .
Thus
_
C
F dr =
_
C
1
F dr +
_
C
2
F dr +
_
C
3
F dr
=
_
5
0
t(
i +
j )
i dt +
_
5
0
(5 t)(
i +
j )
j dt +
_
5
0
((5 t) (5 t))(
i +
j ) (
j ) dt
=
_
5
0
t dt +
_
5
0
5 t dt +
_
5
0
0 dt =
_
5
0
5 dt = 25.
To calculate
_
C
F dr using Stokes Theorem, we nd

curl

F =
i

j

k
z
x y + z x y + z 0
i +
j + (1 (1))
k =
i +
j + 2
k .
For Stokes Theorem, the triangular region S in Figure 20.5 is oriented upward, so d

A =

k dxdy. Thus
_
C
F dr =
_
S
curl

F d
A =
_
S
(
i +
j + 2
k )
k dxdy
=
_
S
2 dxdy = 2 Area of triangle = 2
1
2
5 5 = 25.
(5, 5, 0)
(5, 0, 0)
(0, 0, 0)
S
C
3
C
2
C
1
x
y
Figure 20.5
5. The boundary of S is C, the circle x
2
+y
2
= 1, z = 0, oriented counterclockwise and parameterized in polar coordinates
by
r () = cos
i + sin
j , 0 2,
so,
r

() = sin
i + cos
j .
Hence
_
C
F dr =
_
2
0
(sin
i + 0
j + cos
k ) (sin
i + cos
j + 0
k )d
=
_
2
0
sin
2
d = .
Now consider the integral
_
S
curl

F d
A . Here curl

F =
k and the area vector d
A , oriented upward, is
given by
d
A = 2x
i + 2y
j +
k dxdy.
If R is the disk x
2
+ y
2
1, then we have
_
S
curl

F d
A =
_
R
(
k ) (2x
i + 2y
j +
k )dxdy.
Converting to polar coordinates gives:
_
S
curl

F d
A =
_
2
0
_
1
0
(
k ) (2r cos
i + 2r sin
j +
k )rdrd
=
_
2
0
_
1
0
(2r cos 2r sin 1)rdrd
=
_
2
0
_
2
3
(cos sin )
1
2
_
d
= .
Thus, we conrm that
_
C
F dr =
_
S
curl

F d
A .
20.4 SOLUTIONS 433
9. The circulation is the line integral
_
C
F dr which can be evaluated directly by parameterizing the circle, C. Or, since

C is the boundary of a at disk S, we can use Stokes Theorem:
_
C
F dr =
_
S
curl

F d
A
where S is the disk x
2
+y
2
1, z = 2 and is oriented upward (using the right hand rule). Then curl

F = y
i x
j +
k
and the unit normal to S is
k . So
_
S
curl

F d
A =
_
S
(y
i x
j +
k )
k dxdy
=
_
S
1 dxdy
= Area of S =
Problems
13. Since
curl

F =
i

j

k
z
y x y x
=

y
(y x)
i

x
(y x)
j +
_

x
(x)

y
(y)
_
k =
i +
j 2
k ,
writing S for the disk in the plane enclosed by the circle, Stokes Theorem gives
_
C
F dr =
_
S
curl

F d
A =
_
S
(
i +
j 2
k ) d
A .
Now d
A = n dA, where n is the unit vector perpendicular to the plane, so

n =
1
3
(
i +
j +
k ).
Thus
_
C
F dr =
_
S
(
i +
j 2
k )
i +
j +
3
dA =
_
S
0
3
dA = 0.
17. (a) The equation of the rim, C, is x
2
+y
2
= 9, z = 2. This is a circle of radius 3 centered on the z-axis, and lying in the
plane z = 2.
(b) Use Stokes Theorem, with C oriented clockwise when viewed from above:
_
S
curl(y
i + x
j + z
k ) d
A =
_
C
(y
i + x
j + z
k ) dr .
Since C is horizontal, the
k component does not contribute to the integral. The remaining vector eld, y
i +x
j , is
tangent to C, of constant magnitude || y
i +x
j || = 3 on C, and points in the opposite direction to the orientation.

Thus
_
S
curl(y
i + x
j + z
k ) d
A =
_
(y
i + x
j ) dr = 3 Length of curve = 3 23 = 18.

21. (a) We have
curl

F =
i

j

k
z
y z x
i (1)
j (1) +
k (1) =
k .
(b) (i) Using Stokes Theorem, with S representing the disk inside the circle, oriented upward, we have
_
C
F dr =
_
S
curl

F d
A =
_
S
(
k )
k dA = Area of disk = 4.
(ii) This is a right triangle in the plane x = 2; it has height 5 and base length 3. Using Stokes Theorem, with S
representing the triangle, oriented toward the origin (in the direction
i ), we have
_
C
F dr =
_
S
curl

F d
A =
_
S
(
k ) (
i dA) =
_
S
dA = Area of triangle =
1
2
3 5 =
15
2
.
25. (a) It appears that div

F < 0, and div

G < 0; div

G is larger in magnitude (more negative) if the scales are the same.
(b) curl

F and curl

G both appear to be zero at the origin (and elsewhere).
(c) Yes, the cylinder with axis along the z-axis will have negative ux through it (ends parallel to xy-plane).
(d) Same as part(c).
(e) No, you cannot draw a closed curve around the origin such that

F has a non-zero circulation around it because curl
is zero. By Stokes theorem, circulation equals the integral of the curl over the surface bounded by the curve.
(f) Same as part(e)
29. (a)

F has only
i and
j components, and they do not depend on z. Thus

F is everywhere parallel to the xy-plane, and
takes the same values for every value of z.
(b) We have
curl

F =
i

j

k
z
F1(x, y) F2(x, y) 0
=
_
F2
x

F1
y
_
k .
(c) Since C is in the xy-plane, oriented counterclockwise when viewed from above, for an area element d

A in S, we
have d
A =

k dxdy. Thus Stokes Theorem says
_
C
F dr =
_
S
curl

F d
A =
_
S
_
F2
x

F1
y
_
k dxdy =
_
S
_
F2
x

F1
y
_
dxdy.
(d) Greens Theorem.
Exercises
1. Since curl

F =
0 and

F is dened everywhere, we know by the curl test that

F is a gradient eld. In fact,

F = gradf,
where f(x, y, z) = xyz + yz
2
, so f is a potential function for

F .
Problems
5. Let v = a
i + b
j + c
k and try
F = v r = (a
i + b
j + c
k ) (x
i + y
j + z
k ) = (bz cy)
i + (cx az)
j + (ay bx)
k .
Then
curl

F =
i

j

k
z
bz cy cx az ay bx
= 2a
i + 2b
j + 2c
k .
Taking a = 1, b =
3
2
, c = 2 gives curl

F = 2
i 3
j +4
k , so the desired vector eld is

F = (
3
2
z 2y)
i +(2x
z)
j + (y +
3
2
x)
k .
9. Since div

G = 2x + 2y + 2z = 0, there is not a vector potential for

G .
SOLUTIONS to Review Problems for Chapter Twenty 435
13. (a) Using the product rule from Problem 28 on page 430, we nd
curl

E = curl
_
r
r
p
_
=
1
r
p
curl r + grad
_
1
r
p
_
r .
Now curl r =
0 and grad
_
1
r
p
_
is parallel to r , so both terms are zero. Thus curl

E =
0 .
(b) The domain of

E is 3-space minus the origin if p > 0, and it is all of 3-space if p 0.
(c) Both domains have the property that any closed curve can be contracted to a point without hitting the origin, so

E
satises the curl test for all p. Since

E has constant magnitude r
1p
on the sphere of radius r centered at the origin,
and is parallel to the outward normal at every point of the sphere, the sphere must be a level surface of the potential
function , that is, is a function of r alone. Further, since
E = r
1p
, a good guess is
(r) =
_
r
1p
dr,
that is,
(r) =
_
r
2p
2p
if p = 2
ln r if p = 2.
You can check that this is indeed a potential function for

E by checking that grad =

E .
Exercises
1. Scalar. div((2 sin(xy) + tan z)
i + (tan y)
j + (e
x
2
+y
2
)
k ) = 2y cos(xy) + 1/ cos
2
y.
5. We have
div

F =

x
(cos x) +

y
(e
y
) +

z
(x + y + z) = sin x + e
y
+ 1
curl

F =
i

j

k
z
cos x e
y
x + y + z
j .
So

F is not solenoidal and not irrotational.
9. C2, C3, C4, C6, since line integrals around C1 and C5 are clearly nonzero. You can see directly that
_
C
2
F dr and
_
C
6
F dr are zero, because C2 and C6 are perpendicular to their elds at every point.
13. (a) Direct method:
curl

F =
i

j

k
z
0 xz xy
= 2x
i (y)
j + z
k = 2x
i + y
j + z
k .
On the surface, d
A has no
i -component, so the
i -component of curl

F does not contribute to the ux. Thus
_
S
curl

F d
A =
_
S
(y
j + z
k ) d
A .
Since y
j + z
k is perpendicular to S and ||y
j + z
k || =
_
y
2
+ z
2
=
5 on S, we have
_
S
curl

F d
A =
5 Area of S =
5 2
5 3 = 30.
(b) Using Stokes theorem, we replace the ux integral by two line integrals around the circular boundaries, C1 and C2,
of S. See Figure 20.6.
_
S
curl

F ds =
_
C
1
F dr +
_
C
2
F dr .
On C1, the left boundary, x = 0, so

F =

0 , and therefore
_
C
1
F dr = 0. On C2, the right boundary, x = 3, so
F = 3z
j 3y
k . This vector eld has ||
F || =
_
(3z)
2
+ (3y)
2
=
_
9(z
2
+ y
2
) = 3
5. and

F is tangent to
the boundary C2 and pointing in the same direction as C2. Thus
_
C
2
F dr = ||
F || Length of C2 = 3
5 2
5 = 30.
x
z
C
1
C
2
Figure 20.6
17. Since div
F = 3x
2
+ 3y
2
, using cylindrical coordinates to calculate the triple integral gives
_
S
F d
A =
_
Interior
of cylinder
(3x
2
+ 3y
2
) dV = 3
_
2
0
_
5
0
_
2
0
r
2
r dr dz d = 3 2 5
r
4
4
2
0
= 120.
21. If C is the rectangular path around the rectangle, traversed counterclockwise when viewed from above, Stokes Theorem
gives
_
S
curl

F d
A =
_
C
F dr .
The
k component of

F does not contribute to the line integral, and the
j component contributes to the line integral only

along the segments of the curve parallel to the y-axis. Thus, if we break the line integral into four parts
_
S
curl

F d
A =
_
(3,0)
(0,0)
F dr +
_
(3,2)
(3,0)
F dr +
_
(0,2)
(3,2)
F dr +
_
(0,0)
(0,2)
F dr ,
we see that the rst and third integrals are zero, and we can replace

F by its
j component in the other two

_
S
curl

F d
A =
_
(3,2)
(3,0)
(x + 7)
j dr +
_
(0,0)
(0,2)
(x + 7)
j dr .
Now x = 3 in the rst integral and x = 0 in the second integral and the variable of integration is y in both, so
_
S
curl

F d
A =
_
2
0
10 dy +
_
0
2
7 dy = 20 14 = 6.
Problems
25. (a) The cube is in Figure 20.7. The vector eld is parallel to the x-axis and zero on the yz-plane. Thus the only contri-
bution to the ux is from S2. On S2, x = c, the normal is outward. Since

F is constant on S2, the ux through face
S2 is
_
S
2
F d
A =

F

A S
2
= c
i c
2
i
= c
3
.
Thus, total ux through box = c
3
.
(b) Using the geometric denition of divergence
div

F = lim
c0
_
Flux through box
Volume of box
_
= lim
c0
_
c
3
c
3
_
= 1
(c) Using partial derivatives,
div

F =

x
(x) +

y
(0) +

z
(0) = 1 + 0 + 0 = 1.
x
y
z
S
2
S
6
S
3
c
S
1
(back)
s
S
5
(bottom)
'
S
4
(back)
Figure 20.7
29. We close the cylinder, S, by adding the circular disk, S1, at the top, z = 3. The surface S +S1 is oriented outward, so S1
is oriented upward. Applying the Divergence Theorem to the closed surface S + S1 enclosing the region W, we have
_
S+S
1
F d
A =
_
W
div

F dV
_
S
F d
A +
_
S
1
F d
A =
_
div

F dV.
Since
div

F = div(z
2
i + x
2
j + 5
k ) = 0,
we have
_
S
F d
A =
_
S
1
F d
A .
Only the
k -component of

F contributes to the ux through S1, and d
A =

k dxdy on S1, so
_
S
F d
A =
_
S
1
(z
2
i + x
2
j + 5
k )
k dxdy =
_
S
1
5 dxdy = 5 Area of S1 = 5(
2)
2
= 10.
33. We close the cylinder, S, by adding the circular disk, S1, at the top, z = 3. The surface S +S1 is oriented outward, so S1
is oriented upward. Applying the Divergence Theorem to the closed surface S + S1 enclosing the region W, we have
_
S+S
1
F d
A =
_
W
div

F dV
_
S
F d
A +
_
S
1
F d
A =
_
div

F dV.
Since
div

F = div(x
3
i + y
3
j +
k ) = 3x
2
+ 3y
2
,
we have
_
S
F d
A =
_
W
(3x
2
+ 3y
2
) dV
_
S
1
F d
A .
To nd the integral over W, we use cylindrical coordinates. For the integral over S1, we use the fact that d
A =

k dxdy,
so only the k-component of

F contributes to the ux.
_
S
F d
A =
_
2
0
_
3
3
_

2
0
3r
2
r dr dz d
_
S
1
(x
3
i + y
3
j +
k )
k dxdy
=
2
0
z
3
3
3
4
r
4
2
0
_
S
1
dxdy
= 2 6
3
4
(
2)
4
Area of S1
= 36 (
2)
2
= 34.
37. Since
div(3x
i + 4y
j + xy
k ) = 3 + 4 + 0 = 7,
we calculate the ux using the Divergence Theorem:
Flux =
_
S
(3x
i + 4y
j + xy
k ) d
A =
_
W
7 dV = 7 Volume of box = 7 3 5 2 = 210.
41. We use Stokes Theorem. Since
curl

F =
i

j

k
z
x + y y + 2z z + 3x
= 2
i 3
k ,
if S is the interior of the square, then
_
C
F dr =
_
S
curl

F d
A =
_
S
(2
i 3
k ) d
A
Since the area vector of S is 49
j , we have
_
C
F dr =
_
S
(2
i 3
k ) d
A = 3
j 49
j = 147.
45. By the Divergence Theorem, since div

F = 0, the ux through the cone equals the ux upward through the disk r 4
in the plane z = 4. The area vector of the disk is

A = 4
2
k . Since the ux is negative,

F = c(
i +
k ) with c > 0.
Thus
Flux =

F

A = c(
i +
k ) 4
2
k = 16c = 7.
so
c =
7
16
and

F =
7
16
(
i +
k ).
49. Since div

F = 1 + 1 + 1 = 3, the ux through the closed cylinder, S1, with interior W, is
_
S
1
F d
A =
_
W
3 dV = 3 Volume of cylinder = 3.
With the base of the cylinder oriented downward and the top of the cylinder oriented upward,
_
S
F d
A =
_
S
1
F d
A
_
Base
F d
A
_
Top
F d
A .
Since

F is parallel to the base, the ux through the base is 0. The ux through the top is contributed entirely by the

k
component. Since z = 1, we have
Flux through top =
_
Top
F d
A =
_
Top
(x
i + y
j +
k ) d
A =
_
Top
k d
A = .
Thus
_
S
F d
A = 3 = 2.
53. (a) Can be computed. If W is the interior of the sphere, by the Divergence Theorem, we have
_
S
F d
A =
_
W
div

F dV = 4 Volume of sphere = 4
4
3
2
3
=
128
3
.
(b) Cannot be computed.
(c) Can be computed. Use the fact that div(curl

F ) = 0. If W is the inside of the sphere, then by the Divergence
Theorem,
_
S
curl

F d
A =
_
W
div(curl

F )dV =
_
W
0 dV = 0
57. (a) We have
curl

F =
i

j

k
z
y
x
2
+ y
2
x
x
2
+ y
2
0
= 0
i + 0
j +
_

x
_
x
x
2
+ y
2
_
+

y
_
y
x
2
+ y
2
__
k .
Since
x
_
x
x
2
+ y
2
_
=
1
x
2
+ y
2

x(2x)
(x
2
+ y
2
)
2
=
x
2
+ y
2
2x
2
(x
2
+ y
2
)
2
=
y
2
x
2
(x
2
+ y
2
)
2
and similarly

y
_
y
x
2
+ y
2
_
=
x
2
y
2
(x
2
+ y
2
)
2
, we have, provided x
2
+ y
2
= 0,
curl

F =
0 .
The domain of curl

F is all points in 3-space except the z-axis.
(b) On C1, the unit circle x
2
+ y
2
= 1 in the xy-plane, the vector eld

F is tangent to the circle and ||
F || = 1. Thus
Circulation =
_
C
1
F dr = ||
F || Perimeter of circle = 2.
Note that Stokes Theorem cannot be used to calculate this circulation since the z-axis pierces any surface which has
this circle as boundary.
(c) Consider the disk (x 3)
2
+y
2
1 in the plane z = 4. This disk has C2 as boundary and curl

F =
0 everywhere
on this disk. Thus, by Stokes Theorem
_
C
2
F dr = 0.
(d) The square S has an interior region which is pierced by the z-axis, so we cannot use Stokes Theorem. We consider
the region, D, between the circle C1 and the square S. See Figure 20.8. Stokes Theorem applies to the region D,
provided C1 is oriented clockwise. Then we have
_
C
1
(clockwise)
F dr +
_
S
F dr =
_
D
curl

F d
A = 0.
Thus,
_
S
F dr =
_
C
1
(clockwise)
F dr =
_
C
1
(counterclockwise)
F dr = 2.
(e) If a simple closed curve goes around the z-axis, then it contains a circle C of the form x
2
+y
2
= a
2
. The circulation
around C is 2 or 2, depending on its orientation. A calculation similar to that in part (d) then shows that the
circulation around the curve is 2 or 2, again depending on its orientation. If the closed curve does not go around
the z-axis, then curl

F =
0 everywhere on its interior and the circulation is zero.

(2, 2, 0) (2, 2, 0)
(2, 2, 0) (2, 2, 0)
C
1
S
D
x
y
Figure 20.8
1. True. By Stokes Theorem, the circulation of

F around C is the ux of curl

F through the at disc S in the xy-plane
enclosed by the circle. An area element for S is d
A =
k dA, where the sign depends on the orientation of the circle.

Since curl

F is perpendicular to the z-axis, curl

F d
A = (curl

F
k )dA = 0, so the ux of curl

F through S is
zero, hence the circulation of

F around C is zero.
5. True. div

F is a scalar whose value depends on the point at which it is calculated.
9. False. The divergence is a scalar function that gives ux density at a point.
13. True. curl

F is a vector whose value depends on the point at which it is calculated.
17. False. Since

F can be written

F (x, y, z) = x
i + y
j + z
k , the divergence of

F is 3.
21. True. By the Divergence theorem,
_
S
F d
A =
_
W
div

F dV , where W is the solid interior of S and the negative
sign is due to the inward orientation of S. Since div

F = 0, we have
_
S
F d
A = 0.
25. True. The boundary of the cube W consists of six squares, but four of them are parallel to the xz or yz-planes and so
contribute zero ux for this particular vector eld. The only two surfaces of the boundary with nonzero ux are S1 and
S2, which are parallel to the xy-plane.
29. True. The circulation density is obtained by dividing the circulation around a circle C (a scalar) by the area enclosed by
C (also a scalar), in the limit as the area tends to zero.
33. False. The left-hand side of the equation does not make sense. The quantity (
F

G ) is a scalar, so we cannot compute
the curl of it.
37. False. For example, take

F = z
i and

G = x
j . Then

F

G = xz
k , and curl(
F

G ) = z
j . However,
(curl
F ) (curl
G ) =
k =
i .
41. True. By Stokes theorem, both ux integrals are equal to the line integral
_
C
F dr , where C is the circle x

2
+y
2
= 1,
oriented counterclockwise when viewed from the positive z-axis.
45. True. By Stokes theorem,
_
S
curl

F d
A =
_
C
F dr , where C is one or more closed curves that form the boundary

of S. Since

F is a gradient eld, its line integral over any closed curve will be zero.
2.1 SOLUTIONS 23
CHAPTER TWO
Exercises
1. For t between 2 and 5, we have
Average velocity =
s
t
=
400 135
5 2
=
265
3
km/hr.
The average velocity on this part of the trip was 265/3 km/hr.
5. The average velocity over a time period is the change in position divided by the change in time. Since the function s(t)
gives the distance of the particle from a point, we read off the graph that s(1) = 2 and s(3) = 6. Thus,
Average velocity =
s(t)
t
=
s(3) s(1)
3 1
=
6 2
2
= 2 meters/sec.
9. (a) Let s = f(t).
(i) We wish to nd the average velocity between t = 1 and t = 1.1. We have
Average velocity =
f(1.1) f(1)
1.1 1
=
7.84 7
0.1
= 8.4 m/sec.
(ii) We have
Average velocity =
f(1.01) f(1)
1.01 1
=
7.0804 7
0.01
= 8.04 m/sec.
(iii) We have
Average velocity =
f(1.001) f(1)
1.001 1
=
7.008004 7
0.001
= 8.004 m/sec.
(b) We see in part (a) that as we choose a smaller and smaller interval around t = 1 the average velocity appears to be
getting closer and closer to 8, so we estimate the instantaneous velocity at t = 1 to be 8 m/sec.
Problems
13. Using h = 0.1, 0.01, 0.001, we see
7
0.1
1
0.1
= 2.148
7
0.01
1
0.01
= 1.965
7
0.001
1
0.001
= 1.948
7
0.0001
1
0.0001
= 1.946.
This suggests that lim
h0
7
h
1
h
1.9.
17. See Figure 2.1.
distance
time
Figure 2.1
24 Chapter Two /SOLUTIONS
21. Since f(t) is concave down between t = 1 and t = 3, the average velocity between the two times should be less than the
instantaneous velocity at t = 1 but greater than the instantaneous velocity at time t = 3, so D < A < C. For analogous
reasons, F < B < E. Finally, note that f is decreasing at t = 5 so E < 0, but increasing at t = 0, so D > 0. Therefore,
the ordering from smallest to greatest of the given quantities is
F < B < E < 0 < D < A < C.
25. lim
h0
(2 + h)
2
4
h
= lim
h0
4 + 4h + h
2
4
h
= lim
h0
(4 + h) = 4
Exercises
1. The derivative, f
(2), is the rate of change of x

3
at x = 2. Notice that each time x changes by 0.001 in the table, the value
of x
3
changes by 0.012. Therefore, we estimate
f
(2) =
Rate of change
of f at x = 2

0.012
0.001
= 12.
The function values in the table look exactly linear because they have been rounded. For example, the exact value of
x
3
when x = 2.001 is 8.012006001, not 8.012. Thus, the table can tell us only that the derivative is approximately 12.
Example 5 on page 89 shows how to compute the derivative of f(x) exactly.
5. In Table 2.1, each x increase of 0.001 leads to an increase in f(x) by about 0.031, so
f
(3)
0.031
0.001
= 31.
Table 2.1
x 2.998 2.999 3.000 3.001 3.002
x
3
+ 4x 38.938 38.969 39.000 39.031 39.062
9. Since f
(x) = 0 where the graph is horizontal, f
(x) = 0 at x = d. The derivative is positive at points b and c, but the

graph is steeper at x = c. Thus f
(x) = 0.5 at x = b and f
(x) = 2 at x = c. Finally, the derivative is negative at points

a and e but the graph is steeper at x = e. Thus, f
(x) = 0.5 at x = a and f
(x) = 2 at x = e. See Table 2.2.

Thus, we have f
(d) = 0, f
(b) = 0.5, f
(c) = 2, f
(a) = 0.5, f
(e) = 2.
Table 2.2
x f
(x)
d 0
b 0.5
c 2
a 0.5
e 2
Problems
13. The coordinates of A are (4, 25). See Figure 2.2. The coordinates of B and C are obtained using the slope of the tangent
line. Since f
(4) = 1.5, the slope is 1.5

2.2 SOLUTIONS 25
A = (4, 25)
B
C
0.1
0.15
0.2
1.5(0.2) = 0.3
Tangent line
Figure 2.2
From A to B, x = 0.2, so y = 1.5(0.2) = 0.3. Thus, at C we have y = 25 + 0.3 = 25.3. The coordinates of
B are (4.2, 25.3).
From A to C, x = 0.1, so y = 1.5(0.1) = 0.15. Thus, at C we have y = 25 0.15 = 24.85. The
coordinates of C are (3.9, 24.85).
17. Figure 2.3 shows the quantities in which we are interested.
2 3
x
Slope = f
(2)
Slope = f
(3)
f(x)
Slope =
f(3)f(2)
32
= f(3) f(2)
T
f(x)
Figure 2.3
The quantities f
(2), f
(3) and f(3) f(2) have the following interpretations:

f
(2) = slope of the tangent line at x = 2

f
(3) = slope of the tangent line at x = 3

f(3) f(2) =
f(3)f(2)
32
= slope of the secant line from f(2) to f(3).
From Figure 2.3, it is clear that 0 < f(3) f(2) < f
(2). By extending the secant line past the point (3, f(3)), we can
see that it lies above the tangent line at x = 3.
Thus
0 < f
(3) < f(3) f(2) < f
(2).
21. (a) Figure 2.4 shows the graph of an even function. We see that since f is symmetric about the y-axis, the tangent line at
x = 10 is just the tangent line at x = 10 ipped about the y-axis, so the slope of one tangent is the negative of that
of the other. Therefore, f
(10) = f
(10) = 6.
(b) From part (a) we can see that if f is even, then for any x, we have f
(x) = f
(x). Thus f
(0) = f
(0), so
f
(0) = 0.
10 10 10
60
60
x
f(x)
Figure 2.4
25. We want f
(2). The exact answer is

f
(2) = lim
h0
f(2 + h) f(2)
h
= lim
h0
(2 + h)
2+h
4
h
,
but we can approximate this. If h = 0.001, then
(2.001)
2.001
4
0.001
6.779
and if h = 0.0001 then
(2.0001)
2.0001
4
0.0001
6.773,
so f
(2) 6.77.
29. The quantity f(0) represents the population on October 17, 2006, so f(0) = 300 million.
The quantity f
(0) represents the rate of change of the population (in millions per year). Since
1 person
11 seconds
=
1/10
6
million people
11/(60 60 24 365) years
= 2.867 million people/year,
so we have f
(0) = 2.867.
33. lim
h0
(2 h)
3
8
h
= lim
h0
8 12h + 6h
2
h
3
8
h
= lim
h0
h(12 + 6h h
2
)
h
= lim
h0
12 + 6h h
2
= 12.
37.
1
4 + h
1
2
=
2
4 + h
2
4 + h
=
(2
4 + h)(2 +
4 + h)
2
4 + h(2 +
4 + h)
=
4 (4 + h)
2
4 + h(2 +
4 + h)
.
Therefore lim
h0
1
h
4 + h
1
2
= lim
h0
1
2
4 + h(2 +
4 + h)
=
1
16
41.
f
(1) = lim
h0
f(1 + h) f(1)
h
= lim
h0
((1 + h)
3
+ 5) (1
3
+ 5)
h
= lim
h0
1 + 3h + 3h
2
+ h
3
+ 5 1 5
h
= lim
h0
3h + 3h
2
+ h
3
h
= lim
h0
(3 + 3h + h
2
) = 3.
45. As we saw in the answer to Problem 39, the slope of the tangent line to f(x) = x
3
at x = 2 is 12. When x = 2,
f(x) = 8 so we know the point (2, 8) is on the tangent line. Thus the equation of the tangent line is y = 12(x +
2) 8 = 12x + 16.
2.3 SOLUTIONS 27
Exercises
1. (a) We use the interval to the right of x = 2 to estimate the derivative. (Alternately, we could use the interval to the left
of 2, or we could use both and average the results.) We have
f
(2)
f(4) f(2)
4 2
=
24 18
4 2
=
6
2
= 3.
We estimate f
(2) 3.
(b) We know that f
(x) is positive when f(x) is increasing and negative when f(x) is decreasing, so it appears that
f
(x) is positive for 0 < x < 4 and is negative for 4 < x < 12.
5. The slope of this curve is approximately 1 at x = 4 and at x = 4, approximately 0 at x = 2.5 and x = 1.5, and
approximately 1 at x = 0. See Figure 2.5.
4 4
4
4
x
Figure 2.5
4 4
4
4
x
Figure 2.6
9. See Figure 2.6.
13. Since 1/x = x
1
, using the power rule gives
k
(x) = (1)x
2
=
1
x
2
.
Using the denition of the derivative, we have
k
(x) = lim
h0
k(x + h) k(x)
h
= lim
h0
1
x+h

1
x
h
= lim
h0
x (x + h)
h(x + h)x
= lim
h0
h
h(x + h)x
= lim
h0
1
(x + h)x
=
1
x
2
.
17. See Figure 2.7.
2 1 1 2
10
6
2
6
10 f(x) = 5x
x
2 1 1 2
5
f
(x)
x
Figure 2.7
21.
2
3
2
2
1
1 f(x)
x
2
3
2
2
1
1
f
(x)
x
Problems
25. We know that f
(x)
f(x + h) f(x)
h
. For this problem, well take the average of the values obtained for h = 1
and h = 1; thats the average of f(x + 1) f(x) and f(x) f(x 1) which equals
f(x + 1) f(x 1)
2
. Thus,
f
(0) f(1) f(0) = 13 18 = 5.

f
(1) [f(2) f(0)]/2 = [10 18]/2 = 4.

f
(2) [f(3) f(1)]/2 = [9 13]/2 = 2.

f
(3) [f(4) f(2)]/2 = [9 10]/2 = 0.5.

f
(4) [f(5) f(3)]/2 = [11 9]/2 = 1.

f
(5) [f(6) f(4)]/2 = [15 9]/2 = 3.

f
(6) [f(7) f(5)]/2 = [21 11]/2 = 5.

f
(7) [f(8) f(6)]/2 = [30 15]/2 = 7.5.

f
(8) f(8) f(7) = 30 21 = 9.

The rate of change of f(x) is positive for 4 x 8, negative for 0 x 3. The rate of change is greatest at about
x = 8.
29. See Figure 2.8.
2 4
f
(x)
x
Figure 2.8
33. See Figure 2.9.
3 3
20
20
f
(x)
x
Figure 2.9
37. (a) Graph II
(b) Graph I
(c) Graph III
41. (a) The function f is increasing where f
is positive, so for x1 < x < x3.

(b) The function f is decreasing where f
is negative, so for 0 < x < x1 or x3 < x < x5.

45. From the given information we know that f is increasing for values of x less than 2, is decreasing between x = 2 and
x = 2, and is constant for x > 2. Figure 2.10 shows a possible graphyours may be different.
2.4 SOLUTIONS 29
4 2 2 4
x
y
Figure 2.10
49. If g(x) is odd, its graph remains the same if you rotate it 180
about the origin. So the tangent line to g at x = x0 is the

tangent line to g at x = x0, rotated 180
.
x
y
y = g(x)
x
y
y = g
(x)
But the slope of a line stays constant if you rotate it 180
. So g
(x0) = g
(x0); g
is even.
Exercises
1. (a) The function f takes quarts of ice cream to cost in dollars, so 200 is the amount of ice cream, in quarts, and $600 is
the corresponding cost, in dollars. It costs $600 to produce 200 quarts of ice cream.
(b) Here, 200 is in quarts, but the 2 is in dollars/quart. After producing 200 quarts of ice cream, the cost to produce one
additional quart is about $2.
5. (a) The statement f(5) = 18 means that when 5 milliliters of catalyst are present, the reaction will take 18 minutes.
Thus, the units for 5 are ml while the units for 18 are minutes.
(b) As in part (a), 5 is measured in ml. Since f
tells how fast T changes per unit a, we have f
measured in minutes/ml.
If the amount of catalyst increases by 1 ml (from 5 to 6 ml), the reaction time decreases by about 3 minutes.
9. (a) This means that investing the $1000 at 5% would yield $1649 after 10 years.
(b) Writing g
(r) as dB/dr, we see that the units of dB/dr are dollars per percent (interest). We can interpret dB as
the extra money earned if interest rate is increased by dr percent. Therefore g
(5) =
dB
dr
|r=5 165 means that
the balance, at 5% interest, would increase by about $165 if the interest rate were increased by 1%. In other words,
g(6) g(5) + 165 = 1649 + 165 = 1814.
Problems
13. (a) Since W = f(c) where W is weight in pounds and c is the number of Calories consumed per day:
f(1800) = 155 means that
consuming 1800 Calories per day
results in a weight of 155 pounds.
f
(2000) = 0 means that

consuming 2000 Calories per day causes
neither weight gain nor loss.
f
1
(162) = 2200 means that
a weight of 162 pounds is caused by
a consumption of 2200 Calories per day.
(b) The units of dW/dc are pounds/(Calories/day).
17. Let p be the rating points earned by the CBS Evening News, let R be the revenue earned in millions of dollars, and let
R = f(p). When p = 4.3,
Rate of change of revenue =
$5.5 million
0.1 point
= 55 million dollars/point.
Thus
f
(4.3) = 55.
21. Units of g
(55) are mpg/mph. The statement g
(55) = 0.54 means that at 55 miles per hour the fuel efciency (in miles
per gallon, or mpg) of the car decreases at a rate of approximately one half mpg as the velocity increases by one mph.
25. (a) The company hopes that increased advertising always brings in more customers instead of turning them away. There-
fore, it hopes f
(a) is always positive.

(b) If f
(100) = 2, it means that if the advertising budget is $100,000, each extra dollar spent on advertising will bring
in $2 worth of sales. If f
(100) = 0.5, each dollar above $100 thousand spent on advertising will bring in $0.50
worth of sales.
(c) If f
(100) = 2, then as we saw in part (b), spending slightly more than $100,000 will increase revenue by an amount
greater than the additional expense, and thus more should be spent on advertising. If f
(100) = 0.5, then the increase

in revenue is less than the additional expense, hence too much is being spent on advertising. The optimum amount
to spend is an amount that makes f
(a) = 1. At this point, the increases in advertising expenditures just pay for
themselves. If f
(a) < 1, too much is being spent; if f
(a) > 1, more should be spent.

29. Solving for dp/d, we get
dp
d
=
p
+ (p/c
2
)
.
(a) For 10 g/cm
3
, we have log 1, so, from Figure 2.40 in the text, we have 2.6 and log p 13.
Thus p 10
13
, so p/c
2
10
13
/(9 10
20
) 10
8
, and
dp
d

10
13
10 + 10
8
2.6 2.6 10
12
.
The derivative can be interpreted as the ratio between a change in pressure and the corresponding change in density.
The fact that it is so large says that a very large change in pressure brings about a very small change in density. This
says that cold iron is not a very compressible material.
(b) For 10
6
, we have log 6, so, from Figure 2.40 in the text, 1.5 and log p 23.
Thus p 10
23
, so p/c
2
10
23
/(9 10
20
) 10
2
, and
dp
d

10
23
10
6
+ 10
2
1.5 1.5 10
17
.
This tells us that the matter in a white dwarf is even less compressible than cold iron.
Exercises
1. (a) Since the graph is below the x-axis at x = 2, the value of f(2) is negative.
(b) Since f(x) is decreasing at x = 2, the value of f
(2) is negative.
(c) Since f(x) is concave up at x = 2, the value of f
(2) is positive.
5. The graph must be everywhere decreasing and concave up on some intervals and concave down on other intervals. One
possibility is shown in Figure 2.11.
2.5 SOLUTIONS 31
x
Figure 2.11
9. f
(x) < 0
f
(x) = 0
13. The velocity is the derivative of the distance, that is, v(t) = s
(t). Therefore, we have

v(t) = lim
h0
s(t + h) s(t)
h
= lim
h0
(5(t + h)
2
+ 3) (5t
2
+ 3)
h
= lim
h0
10th + 5h
2
h
= lim
h0
h(10t + 5h)
h
= lim
h0
(10t + 5h) = 10t
The acceleration is the derivative of velocity, so a(t) = v
(t):
a(t) = lim
h0
10(t + h) 10t
h
= lim
h0
10h
h
= 10.
Problems
4 4
4
4
x
y
Figure 2.12
4 4
x
y
Figure 2.13
25. Suppose p(t) is the average price level at time t. Then, if t0 = April 1991,
Prices are still rising means p
(t0) > 0.
Prices rising less fast than they were means p
(t0) < 0.
Prices rising not as much less fast as everybody had hoped means H < p
(t0), where H is the rate of change in rate of

change of prices that people had hoped for.
29. Since f
is everywhere positive, f is everywhere increasing. Hence the greatest value of f is at x6 and the least value of f is
at x1. Directly fromthe graph, we see that f
is greatest at x3 and least at x2. Since f
gives the slope of the graph of f
, f
is greatest where f
is rising most rapidly, namely at x6, and f
is least where f
is falling most rapidly, namely at x1.

Exercises
1. (a) Function f is not continuous at x = 1.
(b) Function f appears not differentiable at x = 1, 2, 3.
Problems
5. Yes, f is differentiable at x = 0, since its graph does not have a corner at x = 0. See below.
0.4 0.4
1.64
Another way to see this is by computing:
lim
h0
f(h) f(0)
h
= lim
h0
(h +|h|)
2
h
= lim
h0
h
2
+ 2h|h| +|h|
2
h
.
Since |h|
2
= h
2
, we have:
lim
h0
f(h) f(0)
h
= lim
h0
2h
2
+ 2h|h|
h
= lim
h0
2(h +|h|) = 0.
So f is differentiable at 0 and f
(0) = 0.
SOLUTIONS to Review Problems for Chapter Two 33
9. We want to look at
lim
h0
(h
2
+ 0.0001)
1/2
(0.0001)
1/2
h
.
As h 0 from positive or negative numbers, the difference quotient approaches 0. (Try evaluating it for h = 0.001,
0.0001, etc.) So it appears there is a derivative at x = 0 and that this derivative is zero. How can this be if f has a corner
at x = 0?
The answer lies in the fact that what appears to be a corner is in fact smoothwhen you zoom in, the graph of f
looks like a straight line with slope 0! See Figure 2.14.
2 1 0 1 2
1
2
x
f(x)
0.2 0.1 0 0.1 0.2
0.1
0.2
x
f(x)
Figure 2.14: Close-ups of f(x) = (x
2
+ 0.0001)
1/2
showing differentiability at x = 0
13. (a) Since
lim
rr
0
E = kr0
and
lim
rr
+
0
E =
kr
2
0
r0
= kr0
and
E(r0) = kr0,
we see that E is continuous at r0.
(b) The function E is not differentiable at r = r0 because the graph has a corner there. The slope is positive for r < r0
and the slope is negative for r > r0.
r
0
kr
0
r
E
Figure 2.15
Exercises
gives the position of the particle, we nd the values of s(3) = 12 3 3
2
= 27 and s(1) = 12 1 1
2
= 11. Using
these values, we nd
Average velocity =
s(t)
t
=
s(3) s(1)
3 1
=
27 11
2
= 8 mm/sec.
gives the position of the particle, we nd the values on the graph of s(3) = 2 and s(1) = 3. Using these values, we nd
Average velocity =
s(t)
t
=
s(3) s(1)
3 1
=
2 3
2
=
1
2
mm/sec.
9. See Figure 2.16.
1
1
2 3 4
x
f
(x)
Figure 2.16
4 4
x
y
Figure 2.17
17. We need to look at the difference quotient and take the limit as h approaches zero. The difference quotient is
f(3 + h) f(3)
h
=
[(3 + h)
2
+ 1] 10
h
=
9 + 6h + h
2
+ 1 10
h
=
6h + h
2
h
=
h(6 + h)
h
.
Since h = 0, we can divide by h in the last expression to get 6 + h. Now the limit as h goes to 0 of 6 + h is 6, so
f
(3) = lim
h0
h(6 + h)
h
= lim
h0
(6 + h) = 6.
So at x = 3, the slope of the tangent line is 6. Since f(3) = 3
2
+ 1 = 10, the tangent line passes through (3, 10), so its
equation is
y 10 = 6(x 3), or y = 6x 8.
21. lim
h0
1
h
1
(a + h)
2

1
a
2
= lim
h0
a
2
(a
2
+ 2ah + h
2
)
(a + h)
2
a
2
h
= lim
h0
(2a h)
(a + h)
2
a
2
=
2
a
3
Problems
25. First note that the line y = t has slope 1. From the graph, we see that
0 < Slope at C < Slope at B < Slope between A and B < 1 < Slope at A.
Since the instantaneous velocity is represented by the slope, we have
0 < Instantaneous velocity at C < Instantaneous velocity at B < Average velocity between A and B < 1 < Instantaneous velocity at A.
SOLUTIONS to Review Problems for Chapter Two 35
29. (a) A possible example is f(x) = 1/|x 2| as lim
x2
1/|x 2| = .
(b) A possible example is f(x) = 1/(x 2)
2
as lim
x2
1/(x 2)
2
= .
33. (a)
1 2 3 4 5 6
4
5
6
Student Cs answer
=slope of this line
'
Student As answer
=slope of this line
'
Student Bs
answer
= slope
of this line
x
(b) The slope of f appears to be somewhere between student As answer and student Bs, so student Cs answer, halfway
in between, is probably the most accurate.
(c) Student As estimate is f
(x)
f(x+h)f(x)
h
, while student Bs estimate is f
(x)
f(x)f(xh)
h
. Student Cs
estimate is the average of these two, or
f
(x)
1
2
f(x + h) f(x)
h
+
f(x) f(x h)
h
=
f(x + h) f(x h)
2h
.
This estimate is the slope of the chord connecting (x h, f(x h)) to (x + h, f(x + h)). Thus, we estimate that
the tangent to a curve is nearly parallel to a chord connecting points h units to the right and left, as shown below.
37. (a) The yam is cooling off so T is decreasing and f
(t) is negative.
(b) Since f(t) is measured in degrees Fahrenheit and t is measured in minutes, df/dt must be measured in units of
F/min.
41. (a) Slope of tangent line = lim
h0
4+h
4
h
. Using h = 0.001,
4.001
4
0.001
= 0.249984. Hence the slope of the
tangent line is about 0.25.
(b)
y y1 = m(x x1)
y 2 = 0.25(x 4)
y 2 = 0.25x 1
y = 0.25x + 1
(c) f(x) = kx
2
If (4, 2) is on the graph of f, then f(4) = 2, so k 4
2
= 2. Thus k =
1
8
, and f(x) =
1
8
x
2
.
(d) To nd where the graph of f crosses then line y = 0.25x + 1, we solve:
1
8
x
2
= 0.25x + 1
x
2
= 2x + 8
x
2
2x 8 = 0
(x 4)(x + 2) = 0
x = 4 or x = 2
f(2) =
1
8
(4) = 0.5
Therefore, (2, 0.5) is the other point of intersection. (Of course, (4, 2) is a point of intersection; we know that from
the start.)
45. (a) The graph looks straight because the graph shows only a small part of the curve magnied greatly.
(b) The month is March: We see that about the 21
st
of the month there are twelve hours of daylight and hence twelve
hours of night. This phenomenon (the length of the day equaling the length of the night) occurs at the equinox, midway
between winter and summer. Since the length of the days is increasing, and Madrid is in the northern hemisphere, we
are looking at March, not September.
(c) The slope of the curve is found from the graph to be about 0.04 (the rise is about 0.8 hours in 20 days or 0.04
hours/day). This means that the amount of daylight is increasing by about 0.04 hours (about 2
1
2
minutes) per calendar
day, or that each day is 2
1
2
minutes longer than its predecessor.
49. The CAS says the derivative is zero. This can be explained by the fact that f(x) = sin
2
x + cos
2
x = 1, so f
(x) is the
derivative of the constant function 1. The derivative of a constant function is zero.
53. (a) The computer algebra system gives
d
dx
(x
2
+ 1)
2
= 4x(x
2
+ 1)
d
dx
(x
2
+ 1)
3
= 6x(x
2
+ 1)
2
d
dx
(x
2
+ 1)
4
= 8x(x
2
+ 1)
3
(b) The pattern suggests that
d
dx
(x
2
+ 1)
n
= 2nx(x
2
+ 1)
n1
.
Taking the derivative of (x
2
+ 1)
n
with a CAS conrms this.
1. False. For example, the car could slow down or even stop at one minute after 2 pm, and then speed back up to 60 mph at
one minute before 3 pm. In this case the car would travel only a few miles during the hour, much less than 50 miles.
5. True. By denition, Average velocity = Distance traveled/Time.
9. False. If f
(x) is increasing then f(x) is concave up. However, f(x) may be either increasing or decreasing. For example,
the exponential decay function f(x) = e
x
is decreasing but f
(x) is increasing because the graph of f is concave up.

13. False. The function f(x) may be discontinuous at x = 0, for instance f(x) =
0 if x 0
1 if x > 0
. The graph of f may have a
vertical tangent line at x = 0, for instance f(x) = x
1/3
.
17. True. Instantaneous acceleration is a derivative, and all derivatives are limits of difference quotients. More precisely,
instantaneous acceleration a(t) is the derivative of the velocity v(t), so
a(t) = lim
h0
v(t + h) v(t)
h
.
21. True; f(x) = x
3
is increasing over any interval.
25. False. Being continuous does not imply differentiability. For example, f(x) = |x| is continuous but not differentiable at
x = 0.
3.1 SOLUTIONS 37
CHAPTER THREE
Exercises
1. The derivative, f
(x), is dened as
f
(x) = lim
h0
f(x +h) f(x)
h
.
If f(x) = 7, then
f
(x) = lim
h0
7 7
h
= lim
h0
0
h
= 0.
5. y
= x
1
. (power rule)
9. y
= 11x
12
.
13. y
=
3
4
x
7/4
.
17. Since y =
1
r
7/2
= r
7/2
, we have
dy
dx
=
7
2
r
9/2
.
21. Since f(x) =
_
1
x
3
=
1
x
3/2
= x
3/2
, we have f
(x) =
3
2
x
5/2
.
25. y
= 17 + 12x
1/2
.
29. y
= 18x
2
+ 8x 2.
33. Since y = t
3/2
(2 +
t) = 2t
3/2
+t
3/2
t
1/2
= 2t
3/2
+t
2
, we have
dy
dx
= 3t
1/2
+ 2t.
37. f(z) =
z
3
+
1
3
z
1
=
1
3
_
z +z
1
_
, so f
(z) =
1
3
_
1 z
2
_
=
1
3
_
z
2
1
z
2
_
.
41. Since f(x) =
ax +b
x
=
ax
x
+
b
x
= a +bx
1
, we have f
(x) = bx
2
.
45. Since w is a constant times q, we have dw/dq = 3ab
2
.
Problems
49. The x is in the exponent and we have not learned how to handle that yet.
53. y
= 2/3z
3
. (power rule and sum rule)
57. The slopes of the tangent lines to y = x
2
2x + 4 are given by y
= 2x 2. A line through the origin has equation

y = mx. So, at the tangent point, x
2
2x + 4 = mx where m = y
= 2x 2.
x
2
2x + 4 = (2x 2)x
x
2
2x + 4 = 2x
2
2x
x
2
+ 4 = 0
(x + 2)(x 2) = 0
x = 2, 2.
Thus, the points of tangency are (2, 4) and (2, 12). The lines through these points and the origin are y = 2x and
y = 6x, respectively. Graphically, this can be seen in Figure 3.1.
38 Chapter Three /SOLUTIONS
(2, 12)
(2, 4)
y = 6x
y
y = x
2
2x + 4
y = 2x
x
Figure 3.1
61. (a) Since the power of x will go down by one every time you take a derivative (until the exponent is zero after which the
derivative will be zero), we can see immediately that f
(8)
(x) = 0.
(b) f
(7)
(x) = 7 6 5 4 3 2 1 x
0
= 5040.
65. (a) The average velocity between t = 0 and t = 2 is given by
Average velocity =
f(2) f(0)
2 0
=
4.9(2
2
) + 25(2) + 3 3
2 0
=
33.4 3
2
= 15.2 m/sec.
(b) Since f
(t) = 9.8t + 25, we have

Instantaneous velocity = f
(2) = 9.8(2) + 25 = 5.4 m/sec.

(c) Acceleration is given f
(t) = 9.8. The acceleration at t = 2 (and all other times) is the acceleration due to gravity,
which is 9.8 m/sec
2
.
(d) We can use a graph of height against time to estimate the maximum height of the tomato. See Figure 3.2. Alternately,
we can nd the answer analytically. The maximumheight occurs when the velocity is zero and v(t) = 9.8t+25 = 0
when t = 2.6 sec. At this time the tomato is at a height of f(2.6) = 34.9. The maximum height is 34.9 meters.
2.6 5.2
34.9
t (sec)
height (m)
Figure 3.2
(e) We see in Figure 3.2 that the tomato hits ground at about t = 5.2 seconds. Alternately, we can nd the answer
analytically. The tomato hits the ground when
f(t) = 4.9t
2
+ 25t + 3 = 0.
We solve for t using the quadratic formula:
t =
25
_
(25)
2
4(4.9)(3)
2(4.9)
t =
25
683.8
9.8
t = 0.12 and t = 5.2.
We use the positive values, so the tomato hits the ground at t = 5.2 seconds.
3.1 SOLUTIONS 39
69. V =
4
3
r
3
. Differentiating gives
dV
dr
= 4r
2
= surface area of a sphere.
The difference quotient
V (r+h)V (r)
h
is the volume between two spheres divided by the change in radius. Further-
more, when h is very small, the difference between volumes, V (r + h) V (r), is like a coating of paint of depth h
applied to the surface of the sphere. The volume of the paint is about h (Surface Area) for small h: dividing by h gives
back the surface area.
Thinking about the derivative as the rate of change of the function for a small change in the variable gives another
way of seeing the result. If you increase the radius of a sphere a small amount, the volume increases by a very thin layer
whose volume is the surface area at that radius multiplied by that small amount.
73. (a)
d(x
1
)
dx
= lim
h0
(x +h)
1
x
1
h
= lim
h0
1
h
_
1
x +h

1
x
_
= lim
h0
1
h
_
x (x +h)
x(x +h)
_
= lim
h0
1
h
_
h
x(x +h)
_
= lim
h0
1
x(x +h)
=
1
x
2
= 1x
2
.
d(x
3
)
dx
= lim
h0
(x +h)
3
x
3
h
= lim
h0
1
h
_
1
(x +h)
3

1
x
3
_
= lim
h0
1
h
_
x
3
(x +h)
3
x
3
(x +h)
3
_
= lim
h0
1
h
_
x
3
(x
3
+ 3hx
2
+ 3h
2
x +h
3
)
x
3
(x +h)
3
_
= lim
h0
1
h
_
3hx
2
3xh
2
h
3
x
3
(x +h)
3
_
= lim
h0
3x
2
3xh h
2
x
3
(x +h)
3
=
3x
2
x
6
= 3x
4
.
(b) For clarity, let n = k, where k is a positive integer. So x
n
= x
k
.
d(x
k
)
dx
= lim
h0
(x +h)
k
x
k
h
= lim
h0
1
h
_
1
(x +h)
k

1
x
k
_
= lim
h0
1
h
_
x
k
(x +h)
k
x
k
(x +h)
k
_
= lim
h0
1
h
_
x
k
x
k
khx
k1
terms involving h
2
and higher powers of h
..
. . . h
k
x
k
(x +h)
k
_
=
kx
k1
x
k
(x)
k
=
k
x
k+1
= kx
(k+1)
= kx
k1
.
Exercises
1. f
(x) = 2e
x
+ 2x.
5. y
= 10x + (ln 2)2

x
.
9. Since y = 2
x
+
2
x
3
= 2
x
+ 2x
3
, we have
dy
dx
= (ln 2)2
x
6x
4
.
13. f
(t) = (ln(ln 3))(ln 3)

t
.
17. f
(x) = 3x
2
+ 3
x
ln 3
21. Since e and k are constants, e
k
is constant, so we have f
(x) = (ln k)k

x
.
25. y
(x) = a
x
ln a +ax
a1
.
29. We can take the derivative of the sum x
2
+ 2
x
, but not the product.
33. The exponent is x
2
, and we havent learned what to do about that yet.
Problems
37. f(s) = 5
s
e
s
= (5e)
s
, so f
(s) = ln(5e) (5e)

s
= (1 + ln 5)5
s
e
s
.
41. (a) P = 10.186(0.997)
20
9.592 million.
(b) Differentiating, we have
dP
dt
= 10.186(ln 0.997)(0.997)
t
so
dP
dt
t=20
= 10.186(ln 0.997)(0.997)
20
0.0325 million/year.
Thus in 2020, Hungarys population will be decreasing by 32,500 people per year.
45. Since y = 2
x
, y
= (ln 2)2
x
. At (0, 1), the tangent line has slope ln 2 so its equation is y = (ln 2)x + 1. At c, y = 0, so
0 = (ln 2)c + 1, thus c =
1
ln 2
.
49. We are interested in when the derivative
d(a
x
)
dx
is positive and when it is negative. The quantity a
x
is always positive.
However ln a > 0 for a > 1 and ln a < 0 for 0 < a < 1. Thus the function a
x
is increasing for a > 1 and decreasing
for a < 1.
Exercises
1. By the product rule, f
(x) = 2x(x
3
+ 5) + x
2
(3x
2
) = 2x
4
+ 3x
4
+ 10x = 5x
4
+ 10x. Alternatively, f
(x) =
(x
5
+ 5x
2
)
= 5x
4
+ 10x. The two answers should, and do, match.
5. y
=
1
2
x
2
x
+
x(ln 2)2
x
.
9. f
(y) = (ln 4)4

y
(2 y
2
) + 4
y
(2y) = 4
y
((ln 4)(2 y
2
) 2y).
13.
dy
dx
=
1 2
t
(t + 1)(ln 2)2
t
(2
t
)
2
=
2
t
(1 (t + 1) ln 2)
(2
t
)
2
=
1 (t + 1) ln 2
2
t
17.
dz
dt
=
3(5t + 2) (3t + 1)5
(5t + 2)
2
=
15t + 6 15t 5
(5t + 2)
2
=
1
(5t + 2)
2
.
21. w = y
2
6y + 7. w
= 2y 6, y = 0.
25. h
(r) =
d
dr
_
r
2
2r + 1
_
=
(2r)(2r + 1) 2r
2
(2r + 1)
2
=
2r(r + 1)
(2r + 1)
2
.
3.3 SOLUTIONS 41
29.
f
(x) =
(2 + 3x + 4x
2
)(1) (1 +x)(3 + 8x)
(2 + 3x + 4x
2
)
2
=
2 + 3x + 4x
2
3 11x 8x
2
(2 + 3x + 4x
2
)
2
=
4x
2
8x 1
(2 + 3x + 4x
2
)
2
.
Problems
33. Using the quotient rule, we know that j
(x) = (g
(x) f(x) g(x) f
(x))/(f(x))
2
. We use slope to compute the
derivatives. Since f(x) is linear on the interval 0 < x < 2, we compute the slope of the line to see that f
(x) = 2 on this
interval. Similarly, we compute the slope on the interval 2 < x < 4 to see that f
(x) = 2 on the interval 2 < x < 4.

Since f(x) has a corner at x = 2, we know that f
(2) does not exist.

Similarly, g(x) is linear on the interval shown, and we see that the slope of g(x) on this interval is 1 so we have
g
(x) = 1 on this interval.

(a) We have
j
(1) =
g
(1) f(1) g(1) f
(1)
(f(1))
2
=
(1)2 3 2
2
2
=
2 6
4
=
8
4
= 2.
(b) We have j
(2) = (g
(2) f(2) g(2) f
(2))/(f(2)
2
). Since f(x) has a corner at x = 2, we know that f
(2) does
not exist. Therefore, j
(2) does not exist.

(c) We have
j
(3) =
g
(3) f(3) g(3) f
(3)
(f(3))
2
=
(1)2 1(2)
2
2
=
2 + 2
4
= 0.
37. Estimates may vary. From the graphs, we estimate f(2) 0.3, f
(2) 1.1, g(2) 1.6, and g
(2) 0.5. By the

quotient rule, to one decimal place
k
(2) =
f
(2) g(2) f(2) g
(2)
(g(2))
2

1.1(1.6) 0.3(0.5)
(1.6)
2
= 0.7.
41. f(x) = e
x
e
x
f
(x) = e
x
e
x
+e
x
e
x
= 2e
2x
.
45. Since f(0) = 5/1 = 5, the tangent line passes through the point (0, 5), so its vertical intercept is 5. To nd the
slope of the tangent line, we nd the derivative of f(x) using the quotient rule:
f
(x) =
(x + 1) 2 (2x 5) 1
(x + 1)
2
=
7
(x + 1)
2
.
At x = 0, the slope of the tangent line is m = f
(0) = 7. The equation of the tangent line is y = 7x 5.

49. (a) G
(z) = F
(z)H(z) +H
(z)F(z), so
G
(3) = F
(3)H(3) +H
(3)F(3) = 4 1 + 3 5 = 19.
(b) G
(w) =
F
(w)H(w) H
(w)F(w)
[H(w)]
2
, so G
(3) =
4(1) 3(5)
1
2
= 11.
53. (a) If the museum sells the painting and invests the proceeds P(t) at time t, then t years have elapsed since 2000, and
the time span up to 2020 is 20 t. This is how long the proceeds P(t) are earning interest in the bank. Each year the
money is in the bank it earns 5% interest, which means the amount in the bank is multiplied by a factor of 1.05. So,
at the end of (20 t) years, the balance is given by
B(t) = P(t)(1 + 0.05)
20t
= P(t)(1.05)
20t
.
(b)
B(t) = P(t)(1.05)
20
(1.05)
t
= (1.05)
20
P(t)
(1.05)
t
.
(c) By the quotient rule,
B
(t) = (1.05)
20
_
P
(t)(1.05)
t
P(t)(1.05)
t
ln 1.05
(1.05)
2t
_
.
So,
B
(10) = (1.05)
20
_
5000(1.05)
10
150,000(1.05)
10
ln 1.05
(1.05)
20
_
= (1.05)
10
(5000 150,000 ln 1.05)
3776.63.
57. From the answer to Problem 56, we nd that
f
(x) = (x r1)(x r2) (x rn1) 1

+(x r1)(x r2) (x rn2) 1 (x rn)
+(x r1)(x r2) (x rn3) 1 (x rn1)(x rn)
+ + 1 (x r2)(x r3) (x rn)
= f(x)
_
1
x r1
+
1
x r2
+ +
1
x rn
_
.
Exercises
1. f
(x) = 99(x + 1)
98
1 = 99(x + 1)
98
.
5.
d
dx
(
e
x
+ 1) =
d
dx
(e
x
+ 1)
1/2
=
1
2
(e
x
+ 1)
1/2
d
dx
(e
x
+ 1) =
e
x
2
e
x
+ 1
.
9. k
(x) = 4(x
3
+e
x
)
3
(3x
2
+e
x
).
13. f() = (2
1
)
= (
1
2
)
so f
() = (ln
1
2
)2
.
17. p
(t) = 4e
4t+2
.
21. y
= 4e
4t
.
25. z
(x) =
(ln 2)2
x
3
3
_
(2
x
+ 5)
2
.
29. We can write this as f(z) =

ze
z
, in which case it is the same as problem 24. So f
(z) =
1
2
z
e
z
ze
z
.
33.
dy
dx
=
2e
2x
(x
2
+ 1) e
2x
(2x)
(x
2
+ 1)
2
=
2e
2x
(x
2
+ 1 x)
(x
2
+ 1)
2
37. w
= (2t + 3)(1 e
2t
) + (t
2
+ 3t)(2e
2t
).
41.
f
(w) = (e
w
2
)(10w) + (5w
2
+ 3)(e
w
2
)(2w)
= 2we
w
2
(5 + 5w
2
+ 3)
= 2we
w
2
(5w
2
+ 8).
45. f
(y) = e
e
(y
2
)
_
(e
y
2
)(2y)
_
= 2ye
[e
(y
2
)
+y
2
]
.
49. We use the product rule. We have
f
(x) = (ax)(e
bx
(b)) + (a)(e
bx
) = abxe
bx
+ae
bx
.
3.4 SOLUTIONS 43
Problems
53. Using the chain rule, we know that v
(x) = f
(f(x)) f
(x). We use slope to compute the derivatives. Since f(x) is

linear on the interval 0 < x < 2, we compute the slope of the line to see that f
(x) = 2 on this interval. Similarly, we

compute the slope on the interval 2 < x < 4 to see that f
(x) = 2 on the interval 2 < x < 4. Since f(x) has a corner

at x = 2, we know that f
(2) does not exist.

(a) We have v
(1) = f
(f(1)) f
(1) = f
(2) 2. Since f(x) has a corner at x = 2, we know that f
(2) does not exist.

Therefore, v
(1) does not exist.

(b) We have v
(2) = f
(f(2)) f
(2). Since f(x) has a corner at x = 2, we know that f
(2) does not exist. Therefore,

v
(2) does not exist.

(c) We have v
(3) = f
(f(3)) f
(3) = (f
(2))(2). Since f(x) has a corner at x = 2, we know that f
(2) does not

exist. Therefore, v
(3) does not exist.

57. The chain rule gives
d
dx
g(f(x))
x=30
= g
(f(30))f
(30) = g
(20)f
(30) = (1/2)(2) = 1.
61. The graph is concave down when f
(x) < 0.
f
(x) = e
x
2
(2x)
f
(x) =
_
e
x
2
(2x)
_
(2x) +e
x
2
(2)
=
4x
2
e
x
2

2
e
x
2
=
4x
2
2
e
x
2
< 0
The graph is concave down when 4x
2
< 2. This occurs when x
2
<
1
2
, or
1
2
< x <
1
2
.
65. (a) The rate of change of the population is P
(t). If P
(t) is proportional to P(t), we have

P
(t) = kP(t).
(b) If P(t) = Ae
kt
, then P
(t) = kAe
kt
= kP(t).
69. We see that m
(x) is nearly of the form f
(g(x)) g
(x) where
f(g) = e
g
and g(x) = x
6
,
but g
(x) is off by a multiple of 6. Therefore, using the chain rule, let

m(x) =
f(g(x))
6
=
e
(x
6
)
6
.
73. We have h(c) = f(g(c)) = f(b) = 0. From the chain rule,
h
(c) = f
(g(c))g
(c).
Since g is increasing at x = c, we know that g
(c) > 0. We have

f
(g(c)) = f
(b),
and since f is decreasing at x = b, we have f
(g(c)) < 0. Thus,

h
(c) = f
(g(c))
. .
(c)
. .
+
< 0,
so h is decreasing at x = c.
77. We have f(0) = 6 and f(10) = 6e
0.013(10)
= 6.833. The derivative of f(t) is
f
(t) = 6e
0.013t
0.013 = 0.078e
0.013t
,
and so f
(0) = 0.078 and f
(10) = 0.089.
These values tell us that in 1999 (at t = 0), the population of the world was 6 billion people and the population was
growing at a rate of 0.078 billion people per year. In the year 2009 (at t = 10), this model predicts that the population of
the world will be 6.833 billion people and growing at a rate of 0.089 billion people per year.
81. (a) For t < 0, I =
dQ
dt
= 0.
For t > 0, I =
dQ
dt
=
Q0
RC
e
t/RC
.
(b) For t > 0, t 0 (that is, as t 0
+
),
I =
Q0
RC
e
t/RC
Q0
RC
.
Since I = 0 just to the left of t = 0 and I = Q0/RC just to the right of t = 0, it is not possible to dene I at
t = 0.
(c) Q is not differentiable at t = 0 because there is no tangent line at t = 0.
85. Using the chain and product rule:
d
2
dx
2
(f(g(x))) =
d
dx
_
d
dx
(f(g(x)))
_
=
d
dx
_
f
(g(x)) g
(x)
_
= f
(g(x)) g
(x) g
(x) +f
(g(x)) g
(x)
= f
(g(x))
_
g
(x)
_
2
+f
(g(x)) g
(x).
Exercises
1.
Table 3.1
x cos x Difference Quotient sin x
0 1.0 0.0005 0.0
0.1 0.995 0.10033 0.099833
0.2 0.98007 0.19916 0.19867
0.3 0.95534 0.296 0.29552
0.4 0.92106 0.38988 0.38942
0.5 0.87758 0.47986 0.47943
0.6 0.82534 0.56506 0.56464
5. f
(x) = cos(3x) 3 = 3 cos(3x).

9. f
(x) = (2x)(cos x) +x
2
(sin x) = 2xcos x x
2
sin x.
13. z
= e
cos
(sin )e
cos
.
17.
f(x) = (1 cos x)
1
2
f
(x) =
1
2
(1 cos x)
1
2
((sin x))
=
sin x
2
1 cos x
.
3.5 SOLUTIONS 45
21. f
(x) = 2 [sin(3x)] + 2x[cos(3x)] 3 = 2 sin(3x) + 6xcos(3x)

25. y
= 5 sin
4
cos .
29. h
(t) = 1 (cos t) +t(sin t) +

1
cos
2
t
= cos t t sin t +
1
cos
2
t
.
33. y
= 2 cos wsin w sin(w

2
)(2w) = 2(cos wsin w +wsin(w
2
))
37. Using the power and quotient rules gives
f
(x) =
1
2
_
1 sin x
1 cos x
_
1/2
_
cos x(1 cos x) (1 sin x) sin x
(1 cos x)
2
_
=
1
2
_
1 cos x
1 sin x
_
cos x(1 cos x) (1 sin x) sin x
(1 cos x)
2
_
=
1
2
_
1 cos x
1 sin x
_
1 cos x sin x
(1 cos x)
2
_
.
Problems
41. The pattern in the table below allows us to generalize and say that the (4n)
th
derivative of cos x is cos x, i.e.,
d
4
y
dx
4
=
d
8
y
dx
8
= =
d
4n
y
dx
4n
= cos x.
Thus we can say that d
48
y/dx
48
= cos x. From there we differentiate twice more to obtain d
50
y/dx
50
= cos x.
n 1 2 3 4 48 49 50
n
th
derivative sin x cos x sin x cos x cos x sin x cos x
45. (a) v(t) =
dy
dt
=
d
dt
(15 + sin(2t)) = 2 cos(2t).
(b)
1 2 3
14
15
16
y = 15 + sin 2t
t
y
1 2 3
2
2
v = 2 cos 2t
t
v
49. (a) Using triangle OPD in Figure 3.3, we see
OD
a
= cos so OD = a cos
PD
a
= sin so PD = a sin .
O
P
Q
l
a
D d
Figure 3.3
Using triangle PQD, we have
(PD)
2
+d
2
= l
2
so
a
2
sin
2
+d
2
= l
2
, d =
_
l
2
a
2
sin
2
.
Thus,
x = OD +DQ
= a cos +
_
l
2
a
2
sin
2
.
(b) Differentiating, regarding a and l as constants,
dx
d
= a sin +
1
2
(2a
2
sin cos )
_
l
2
a
2
sin
2
= a sin
a
2
sin cos
_
l
2
a
2
sin
2
.
We want to nd dx/dt. Using the chain rule and the fact that d/dt = 2, we have
dx
dt
=
dx
d

d
dt
= 2
dx
d
.
(i) Substituting = /2, we have
dx
dt
= 2
dx
d
=/2
= 2a sin
_
2
_
2
a
2
sin(
2
) cos(
2
)
_
l
2
a
2
sin
2
(
2
)
= 2a cm/sec.
(ii) Substituting = /4, we have
dx
dt
= 2
dx
d
=/4
= 2a sin
_
4
_
2
a
2
sin(
4
) cos(
4
)
_
l
2
a
2
sin
2
(
4
)
= a
2
a
2
_
l
2
a
2
/2
cm/sec.
53. (a) If f(x) = sin x, then
f
(x) = lim
h0
sin(x +h) sin x
h
= lim
h0
(sin xcos h + sin hcos x) sin x
h
= lim
h0
sin x(cos h 1) + sin hcos x
h
= sin x lim
h0
cos h 1
h
+ cos x lim
h0
sin h
h
.
(b)
cos h1
h
0 and
sin h
h
1, as h 0. Thus, f
(x) = sin x 0 + cos x 1 = cos x.

(c) Similarly,
g
(x) = lim
h0
cos(x +h) cos x
h
= lim
h0
(cos xcos h sin xsin h) cos x
h
= lim
h0
cos x(cos h 1) sin xsin h
h
= cos x lim
h0
cos h 1
h
sin x lim
h0
sin h
h
= sin x.
3.6 SOLUTIONS 47
Exercises
1. f
(t) =
2t
t
2
+ 1
.
5. f
(x) =
1
1 e
x
(e
x
)(1) =
e
x
1 e
x
.
9. j
(x) =
ae
ax
(e
ax
+b)
13. f
(w) =
1
cos(w 1)
[sin(w 1)] = tan(w 1).
[This could be done easily using the answer from Problem 10 and the chain rule.]
17. g() = , so g
() = 1.
21. h
(w) = arcsin w +
w
1 w
2
.
25. f
(x) = sin(arctan 3x)

_
1
1 + (3x)
2
_
(3) =
3 sin(arctan 3x)
1 + 9x
2
.
29. Using the chain rule gives f
(x) =
cos x sin x
sin x + cos x
.
33. f
(x) = sin(arcsin(x + 1))(

1
_
1 (x + 1)
2
) =
(x + 1)
_
1 (x + 1)
2
.
Problems
37. Estimates may vary. From the graphs, we estimate f(2) 0.3, f
(2) 1.1, and g
(0.3) 1.7. Thus, by the chain rule,

k
(2) = g
(f(2)) f
(2) g
(0.3) f
(2) 1.7 1.1 1.9.

41. pH = 2 = log x means log x = 2 so x = 10
2
. Rate of change of pH with hydrogen ion concentration is
d
dx
pH =
d
dx
(log x) =
1
x(ln 10)
=
1
(10
2
) ln 10
= 43.4
45. (a)
f
(x) =
1
1 +x
2
+
1
1 +
1
x
2
(
1
x
2
)
=
1
1 +x
2
+
_
1
x
2
+ 1
_
=
1
1 +x
2

1
1 +x
2
= 0
(b) f is a constant function. Checking at a few values of x,
Table 3.2
x arctan x arctan x
1
f(x) = arctan x + arctan x
1
1 0.785392 0.7853982 1.5707963
2 1.1071487 0.4636476 1.5707963
3 1.2490458 0.3217506 1.5707963
49. Since the chain rule gives h
(x) = n
(m(x))m
(x) = 1 we must nd values a and x such that a = m(x) and

n
(a)m
(x) = 1.
Calculating slopes from the graph of n gives
n
(a) =
_
1 if 0 < a < 50
1/2 if 50 < a < 100.
Calculating slopes from the graph of m gives
m
(x) =
_
2 if 0 < x < 50
2 if 50 < x < 100.
The only values of the derivative n
are 1 and 1/2 and the only values of the derivative m
are 2 and 2. In order to

have n
(a)m
(x) = 1 we must therefore have n
(a) = 1/2 and m
(x) = 2. Thus 50 < a < 100 and 50 < x < 100.

Now a = m(x) and from the graph of m we see that 50 < m(x) < 100 for 0 < x < 25 or 75 < x < 100.
The two conditions on x we have found are both satised when 75 < x < 100. Thus h
(x) = 1 for all x in the

interval 75 < x < 100. The question asks for just one of these x values, for example x = 80.
53. Since the point (2, 5) is on the curve, we know f(2) = 5. The point (2.1, 5.3) is on the tangent line, so
Slope tangent =
5.3 5
2.1 2
=
0.3
0.1
= 3.
Thus, f
(2) = 3. Since g is the inverse function of f and f(2) = 5, we know f

1
(5) = 2, so g(5) = 2.
Differentiating, we have
g
(2) =
1
f
(g(5))
=
1
f
(2)
=
1
3
.
57. To nd (f
1
)
(3), we rst look in the table to nd that 3 = f(9), so f

1
(3) = 9. Thus,
(f
1
)
(3) =
1
f
(f
1
(3))
=
1
f
(9)
=
1
5
.
61. We must have
(f
1
)
(5) =
1
f
(f
1
(5))
=
1
f
(10)
=
1
8
.
Exercises
1. We differentiate implicitly both sides of the equation with respect to x.
2x + 2y
dy
dx
= 0 ,
dy
dx
=
2x
2y
=
x
y
.
x
1/2
= 5y
1/2
1
2
x
1/2
=
5
2
y
1/2
dy
dx
dy
dx
=
1
2
x
1/2
5
2
y
1/2
=
1
5
_
y
x
=
1
25
.
We can also obtain this answer by realizing that the original equation represents part of the line x = 25y which has slope 1/25.
3.7 SOLUTIONS 49
9.
2ax 2by
dy
dx
= 0
dy
dx
=
2ax
2by
=
ax
by
13. Using the relation cos
2
y + sin
2
y = 1, the equation becomes:
1 = y + 2 or y = 1. Hence,
dy
dx
= 0.
(x a)
2
+y
2
= a
2
2(x a) + 2y
dy
dx
= 0
2y
dy
dx
= 2a 2x
dy
dx
=
2a 2x
2y
=
a x
y
.
21. Differentiating with respect to x gives
3x
2
+ 2xy
+ 2y + 2yy
= 0
so that
y
=
3x
2
+ 2y
2x + 2y
At the point (1, 1) the slope is
5
4
.
25. First, we must nd the slope of the tangent,
dy
dx
(4,2)
. Implicit differentiation yields:
2y
dy
dx
=
2x(xy 4) x
2
_
x
dy
dx
+y
_
(xy 4)
2
.
Given the complexity of the above equation, we rst want to substitute 4 for x and 2 for y (the coordinates of the point
where we are constructing our tangent line), then solve for
dy
dx
. Substitution yields:
2 2
dy
dx
=
(2 4)(4 2 4) 4
2
_
4
dy
dx
+ 2
_
(4 2 4)
2
=
8(4) 16(4
dy
dx
+ 2)
16
= 4
dy
dx
.
4
dy
dx
= 4
dy
dx
,
Solving for
dy
dx
, we have:
dy
dx
= 0.
The tangent is a horizontal line through (4, 2), hence its equation is y = 2.
Problems
29. (a) Taking derivatives implicitly, we get
2
25
x +
2
9
y
dy
dx
= 0
dy
dx
=
9x
25y
.
(b) The slope is not dened anywhere along the line y = 0. This ellipse intersects that line in two places, (5, 0) and
(5, 0). (These are the ends of the ellipse where the tangent is vertical.)
33. (a) Differentiating both sides of the equation with respect to P gives
d
dP
_
4f
2
P
1 f
2
_
=
dK
dP
= 0.
By the product rule
d
dP
_
4f
2
P
1 f
2
_
=
d
dP
_
4f
2
1 f
2
_
P +
_
4f
2
1 f
2
_
1
=
_
(1 f
2
)(8f) 4f
2
(2f)
(1 f
2
)
2
_
df
dP
P +
_
4f
2
1 f
2
_
=
_
8f
(1 f
2
)
2
_
df
dP
P +
_
4f
2
1 f
2
_
= 0.
So
df
dP
=
4f
2
/(1 f
2
)
8fP/(1 f
2
)
2
=
1
2P
f(1 f
2
).
(b) Since f is a fraction of a gas, 0 f 1. Also, in the equation relating f and P we cant have f = 0, since that
would imply K = 0, and we cant have f = 1, since the left side is undened there. So 0 < f < 1. Thus 1f
2
> 0.
Also, pressure cant be negative, and from the equation relating f and P, we see that P cant be zero either, so P > 0.
Therefore df/dP = (1/2P)f(1f
2
) < 0 always. This means that at larger pressures less of the gas decomposes.
Exercises
1. Using the chain rule,
d
dx
(cosh(2x)) = (sinh(2x)) 2 = 2 sinh(2x).
5. Using the chain rule,
d
dt
_
cosh
2
t
_
= 2 cosh t sinh t.
.
9. Using the chain rule twice,
d
dy
(sinh (sinh(3y))) = cosh (sinh(3y)) cosh(3y) 3
= 3 cosh(3y) cosh (sinh(3y)) .
13. Substituting x for x in the formula for sinh x gives
sinh(x) =
e
x
e
(x)
2
=
e
x
e
x
2
=
e
x
e
x
2
= sinh x.
Problems
17. The graph of sinh x in the text suggests that
As x , sinh x
1
2
e
x
.
As x , sinh x
1
2
e
x
.
3.8 SOLUTIONS 51
Using the facts that
As x , e
x
0,
As x , e
x
0,
we can obtain the same results analytically:
As x , sinh x =
e
x
e
x
2

1
2
e
x
.
As x , sinh x =
e
x
e
x
2

1
2
e
x
.
21. Recall that
sinh A =
1
2
(e
A
e
A
) and cosh A =
1
2
(e
A
+e
A
).
Now substitute, expand and collect terms:
cosh Acosh B + sinh Bsinh A =
1
2
(e
A
+e
A
)
1
2
(e
B
+e
B
) +
1
2
(e
B
e
B
)
1
2
(e
A
e
A
)
=
1
4
_
e
A+B
+e
AB
+e
A+B
+e
(A+B)
+e
B+A
e
BA
e
B+A
+e
AB
_
=
1
2
_
e
A+B
+e
(A+B)
_
= cosh(A+B).
25. Using the denition of cosh x and sinh x, we have cosh x
2
=
e
x
2
+e
x
2
2
and sinh x
2
=
e
x
2
e
x
2
2
. Therefore
lim
x
sinh(x
2
)
cosh(x
2
)
= lim
x
e
x
2
e
x
2
e
x
2
+e
x
2
= lim
x
e
x
2
(1 e
2x
2
)
e
x
2
(1 +e
2x
2
)
= lim
x
1 e
2x
2
1 +e
2x
2
= 1.
29. We want to show that for any A, B with A > 0, B > 0, we can nd K and c such that
y = Ae
x
+Be
x
=
Ke
(xc)
+Ke
(xc)
2
=
K
2
e
x
e
c
+
K
2
e
x
e
c
=
_
Ke
c
2
_
e
x
+
_
Ke
c
2
_
e
x
.
Thus, we want to nd K and c such that
Ke
c
2
= A and
Ke
c
2
= B.
Dividing, we have
Ke
c
Ke
c
=
B
A
e
2c
=
B
A
c =
1
2
ln
_
B
A
_
.
If A > 0, B > 0, then there is a solution for c. Substituting to nd K, we have
Ke
c
2
= A
K = 2Ae
c
= 2Ae
(ln(B/A))/2
= 2Ae
ln
B/A
= 2A
_
B
A
= 2
AB.
Thus, if A > 0, B > 0, there is a solution for K also.
The fact that y = Ae
x
+ Be
x
can be rewritten in this way shows that the graph of y = Ae
x
+ Be
x
is the graph
of cosh x, shifted over by c and stretched (or shrunk) vertically by a factor of K.
Exercises
1. Since f(1) = 1 and we showed that f
(1) = 2, the local linearization is

f(x) 1 + 2(x 1) = 2x 1.
5. With f(x) = 1/(
1 +x), we see that the tangent line approximation to f near x = 0 is

f(x) f(0) +f
(0)(x 0),
which becomes
1
1 +x
1 +f
(0)x.
Since f
(x) = (1/2)(1 +x)

3/2
, f
(0) = 1/2. Thus our formula reduces to

1
1 +x
1 x/2.
This is the local linearization of
1
1 +x
near x = 0.
9. From Figure 3.4, we see that the error has its maximum magnitude at the end points of the interval, x = 1. The
magnitude of the error can be read off the graph as less than 0.2 or estimated as
|Error| |1 sin 1| = 0.159 < 0.2.
The approximation is an overestimate for x > 0 and an underestimate for x < 0.
3.9 SOLUTIONS 53
1 1
1
1
x
sin x
T
c
Error
x
y
Figure 3.4
Problems
13. (a) Since
d
dx
(cos x) = sin x,
the slope of the tangent line is sin(/4) = 1/
2. Since the tangent line passes through the point (/4, cos(/4)) =
(/4, 1/
2), its equation is

y
1
2
=
1
2
_
x

4
_
y =
1
2
x +
1
2
_
4
+ 1
_
.
Thus, the tangent line approximation to cos x is
cos x
1
2
x +
1
2
_
4
+ 1
_
.
(b) From Figure 3.5, we see that the tangent line approximation is an overestimate.
(c) From Figure 3.5, we see that the maximum error for 0 x /2 is either at x = 0 or at x = /2. The error can
either be estimated from the graph, or as follows. At x = 0,
|Error| =
cos 0
1
2
_
4
+ 1
_
= 0.262 < 0.3.

At x = /2,
|Error| =
cos

2
+
1
2

1
2
_
4
+ 1
_
= 0.152 < 0.2.

Thus, for 0 x /2, we have
|Error| < 0.3.
/4 /2
1/
2
T
c
Error
cos x
Tangent Line
2
x +
1
2
_
4
+ 1
_
x
Figure 3.5
17. We have f(x) = x + ln(1 +x), so f
(x) = 1 + 1/(1 +x). Thus f
(0) = 2 so
Local linearization near x = 0 is f(x) f(0) +f
(0)x = 2x.
We get an approximate solution using the local linearization instead of f(x), so the equation becomes
2x = 0.2, with solution x = 0.1.
A computer or calculator gives the actual value as x = 0.102.
21. (a) Suppose g is a constant and
T = f(l) = 2
_
l
g
.
Then
f
(l) =
2
g
1
2
l
1/2
=

gl
.
Thus, local linearity tells us that
f(l + l) f(l) +

gl
l.
Now T = f(l) and T = f(l + l) f(l), so
T

gl
l = 2
_
l
g

1
2
l
l
=
T
2
l
l
.
(b) Knowing that the length of the pendulum increases by 2% tells us that
l
l
= 0.02.
Thus,
T
T
2
(0.02) = 0.01T.
So
T
T
0.01.
Thus, T increases by 1%.
25. (a) The range is f(20) = 16,398 meters.
(b) We have
f
() =

90
25510 cos

90
f
(20) = 682.
Thus, for angles, , near 20
, we have
Range = f() f(20) +f
(20)( 20) = 16398 + 682( 20) meters.

(c) The true range for 21
is f(21) = 17,070 meters. The linear approximation gives

Approximate range = 16398 + 682(21 20) = 17080 meters
which is a little too high.
29. We have f(0) = 1 and f
(0) = 0. Thus
E(x) = cos x 1.
Values for E(x)/(x 0) near x = 0 are in Table 3.3.
Table 3.3
x 0.1 0.01 0.001
E(x)/(x 0) 0.050 0.0050 0.00050
3.9 SOLUTIONS 55
From the table, we can see that
E(x)
(x 0)
0.5(x 0),
so k = 1/2 and
E(x)
1
2
(x 0)
2
=
1
2
x
2
.
In addition, f
(0) = 1, so
E(x)
1
2
x
2
=
f
(0)
2
x
2
.
33. The local linearization of e
x
near x = 0 is 1 + 1x so
e
x
1 +x.
Squaring this yields, for small x,
e
2x
= (e
x
)
2
(1 +x)
2
= 1 + 2x +x
2
.
Local linearization of e
2x
directly yields
e
2x
1 + 2x
for small x. The two approximations are consistent because they agree: the tangent line approximation to 1 + 2x +x
2
is
just 1 + 2x.
The rst approximation is more accurate. One can see this numerically or by noting that the approximation for
e
2x
given by 1 + 2x is really the same as approximating e
y
at y = 2x. Since the other approximation approximates
e
y
at y = x, which is twice as close to 0 and therefore a better general estimate, its more likely to be correct.
37. Note that
[f(g(x))]
= lim
h0
f(g(x +h)) f(g(x))
h
.
Using the local linearizations of f and g, we get that
f(g(x +h)) f(g(x)) f
_
g(x) +g
(x)h
_
f(g(x))
f (g(x)) +f
(g(x))g
(x)h f(g(x))
= f
(g(x))g
(x)h.
Therefore,
[f(g(x))]
= lim
h0
f(g(x +h)) f(g(x))
h
= lim
h0
f
(g(x))g
(x)h
h
= lim
h0
f
(g(x))g
(x) = f
(g(x))g
(x).
A more complete derivation can be given using the error term discussed in the section on Differentiability and Linear
Approximation in Chapter 2. Adapting the notation of that section to this problem, we write
f(z +k) = f(z) +f
(z)k +E
f
(k) and g(x +h) = g(x) +g
(x)h +Eg(h),
where lim
h0
Eg(h)
h
= lim
k0
E
f
(k)
k
= 0.
Now we let z = g(x) and k = g(x +h) g(x). Then we have k = g
(x)h +Eg(h). Thus,

f(g(x +h)) f(g(x))
h
=
f(z +k) f(z)
h
=
f(z) +f
(z)k +E
f
(k) f(z)
h
=
f
(z)k +E
f
(k)
h
=
f
(z)g
(x)h +f
(z)Eg(h)
h
+
E
f
(k)
k

_
k
h
_
= f
(z)g
(x) +
f
(z)Eg(h)
h
+
E
f
(k)
k
_
g
(x)h +Eg(h)
h
_
= f
(z)g
(x) +
f
(z)Eg(h)
h
+
g
(x)E
f
(k)
k
+
Eg(h) E
f
(k)
h k
Now, if h 0 then k 0 as well, and all the terms on the right except the rst go to zero, leaving us with the term
f
(z)g
(x). Substituting g(x) for z, we obtain

[f(g(x))]
= lim
h0
f(g(x +h)) f(g(x))
h
= f
(g(x))g
(x).
Exercises
1. False. The derivative, f
(x), is not equal to zero everywhere, because the function is not continuous at integral values of
x, so f
(x) does not exist there. Thus, the Constant Function Theorem does not apply.
5. False. Let f(x) = x
3
on [1, 1]. Then f(x) is increasing but f
(x) = 0 for x = 0.
9. No. This function does not satisfy the hypotheses of the Mean Value Theorem, as it is not continuous.
However, the function has a point c such that
f
(c) =
f(b) f(a)
b a
.
Thus, this satises the conclusion of the theorem.
Problems
13. A polynomial p(x) satises the conditions of Rolles Theorem for all intervals a x b.
Suppose a1, a2, a3, a4, a5, a6, a7 are the seven distinct zeros of p(x) in increasing order. Thus p(a1) = p(a2) = 0,
so by Rolles Theorem, p
(x) has a zero, c1, between a1 and a2.

Similarly, p
(x) has 6 distinct zeros, c1, c2, c3, c4, c5, c6, where
a1 < c1 < a2
a2 < c2 < a3
a3 < c3 < a4
a4 < c4 < a5
a5 < c5 < a6
a6 < c6 < a7.
The polynomial p
(x) is of degree 6, so p
(x) cannot have more than 6 zeros.

17. The Decreasing Function Theorem is: Suppose that f is continuous on [a, b] and differentiable on (a, b). If f
(x) < 0 on
(a, b), then f is decreasing on [a, b]. If f
(x) 0 on (a, b), then f is nonincreasing on [a, b].

To prove the theorem, we note that if f is decreasing then f is increasing and vice-versa. Similarly, if f is non-
increasing, then f is nondecreasing. Thus if f
(x) < 0, then f
(x) > 0, so f is increasing, which means f is

decreasing. And if f
(x) 0, then f
(x) 0, so f is nondecreasing, which means f is nonincreasing.

21. By the Mean Value Theorem, Theorem 3.7, there is a number c, with 0 < c < 1, such that
f
(c) =
f(1) f(0)
1 0
.
Since f(1) f(0) > 0, we have f
(c) > 0.
Alternatively if f
(c) 0 for all c in (0, 1), then by the Increasing Function Theorem, f(0) f(1).
25. Let h(x) = f(x) g(x). Then h
(x) = f
(x) g
(x) = 0 for all x in (a, b). Hence, by the Constant Function Theorem,
there is a constant C such that h(x) = C on (a, b). Thus f(x) = g(x) +C.
SOLUTIONS to Review Problems for Chapter Three 57
Exercises
1. w
= 100(t
2
+ 1)
99
(2t) = 200t(t
2
+ 1)
99
.
5. Using the quotient rule,
h
(t) =
(1)(4 +t) (4 t)
(4 +t)
2
=
8
(4 +t)
2
.
9. Since h() = (
1/2
2
) =
1/2
2
=
1/2
1
, we have h
() =
1
2
1/2
+
2
.
13. y
= 0
17. s
() =
d
d
sin
2
(3 ) = 6 cos(3 ) sin(3 ).
21. f
() = 1(1 +e
)
2
(e
)(1) =
e
(1 +e
)
2
.
25. Using the chain rule and simplifying,
q
() =
1
2
(4
2
sin
2
(2))
1/2
(8 2 sin(2)(2 cos(2))) =
4 2 sin(2) cos(2)
_
4
2
sin
2
(2)
.
29. Using the chain rule, we get:
m
(n) = cos(e
n
) (e
n
)
33.
d
dx
xe
tan x
= e
tan x
+xe
tan x
1
cos
2
x
.
37. h(x) = ax ln e = ax, so h
(x) = a.
41. Using the product rule gives
H
(t) = 2ate
ct
c(at
2
+b)e
ct
= (cat
2
+ 2at bc)e
ct
.
45. Using the quotient rule gives
w
(r) =
2ar(b +r
3
) 3r
2
(ar
2
)
(b +r
3
)
2
=
2abr ar
4
(b +r
3
)
2
.
49. Since g(w) = 5(a
2
w
2
)
2
, g
(w) = 10(a
2
w
2
)
3
(2w) =
20w
(a
2
w
2
)
3
53. Using the quotient rule gives
g
(t) =
_
k
kt
+ 1
_
(ln(kt) t) (ln(kt) +t)
_
k
kt
1
_
(ln(kt) t)
2
g
(t) =
_
1
t
+ 1
_
(ln(kt) t) (ln(kt) +t)
_
1
t
1
_
(ln(kt) t)
2
g
(t) =
ln(kt)/t 1 + ln(kt) t ln(kt)/t 1 + ln(kt) +t
(ln(kt) t)
2
g
(t) =
2 ln(kt) 2
(ln(kt) t)
2
.
57. g
(x) =
1
2
(5x
4
+ 2).
61. g
(x) =
d
dx
(2x x
1/3
+ 3
x
e) = 2 +
1
3x
4
3
+ 3
x
ln 3.
65. r
() =
d
d
sin[(3 )
2
] = cos[(3 )
2
] 2(3 ) 3 = 6(3 ) cos[(3 )
2
].
69. f
(x) =
d
dx
(2 4x 3x
2
)(6x
e
3) = (4 6x)(6x
e
3) + (2 4x 3x
2
)(6ex
e1
).
73.
h
(x) =
_
1
x
2
+
2
x
3
_
_
2x
3
+ 4
_
+
_
1
x

1
x
2
_
_
6x
2
_
= 2x + 4
4
x
2
+
8
x
3
+ 6x 6
= 4x 2 4x
2
+ 8x
3
77. We wish to nd the slope m = dy/dx. To do this, we can implicitly differentiate the given formula in terms of x:
x
2
+ 3y
2
= 7
2x + 6y
dy
dx
=
d
dx
(7) = 0
dy
dx
=
2x
6y
=
x
3y
.
Thus, at (2, 1), m = (2)/3(1) = 2/3.
Problems
81.
f
(x) = 8 + 2
2x
f
(r) = 8 + 2
2r = 4
r =
12
2
2
= 3
2.
85. Since r(x) = s(t(x)), the chain rule gives r
(x) = s
(t(x)) t
(x). Thus,
r
(0) = s
(t(0)) t
(0) s
(2) (2) (2)(2) = 4.

Note that since t(x) is a linear function whose slope looks like 2 from the graph, t
(x) 2 everywhere. To nd
s
(2), draw a line tangent to the curve at the point (2, s(2)), and estimate the slope.
89. We have r(1) = s(t(1)) s(0) 2. By the chain rule, r
(x) = s
(t(x)) t
(x), so
r
(1) = s
(t(1)) t
(1) s
(0) (2) 2(2) = 4.

Thus the equation of the tangent line is
y 2 = 4(x 1)
y = 4x + 6.
Note that since t(x) is a linear function whose slope looks like 2 from the graph, t
(x) 2 everywhere. To nd
s
(0), draw a line tangent to the curve at the point (0, s(0)), and estimate the slope.
93. Since W is proportional to r
3
, we have W = kr
3
for some constant k. Thus, dW/dr = k(3r
2
) = 3kr
2
. Thus, dW/dr
is proportional to r
2
.
97. (a) f(x) = x
2
4g(x)
f(2) = 4 4(3) = 8
f
(2) = 20
Thus, we have a point (2, 8) and slope m = 20. This gives
8 = 2(20) +b
b = 48, so
y = 20x 48.
SOLUTIONS to Review Problems for Chapter Three 59
(b) f(x) =
x
g(x)
f(2) =
2
3
f
(2) =
11
9
Thus, we have point (2,
2
3
) and slope m =
11
9
. This gives
2
3
= (
11
9
)(2) +b
b =
2
3

22
9
=
16
9
, so
y =
11
9
x
16
9
.
(c) f(x) = x
2
g(x)
f(2) = 4 g(2) = 4(3) = 12
f
(2) = 4
Thus, we have point (2, 12) and slope m = 4. This gives
12 = 2(4) +b
b = 20, so
y = 4x + 20.
(d) f(x) = (g(x))
2
f(2) = (g(2))
2
= (3)
2
= 9
f
(2) = 24
Thus, we have point (2, 9) and slope m = 24. This gives
9 = 2(24) +b
b = 57, so
y = 24x + 57.
(e) f(x) = xsin(g(x))
f(2) = 2 sin(g(2)) = 2 sin 3
f
(2) = sin 3 8 cos 3

We will use a decimal approximation for f(2) and f
(2), so the point (2, 2 sin 3) (2, 0.28) and m 8.06. Thus,
0.28 = 2(8.06) +b
b = 15.84, so
y = 8.06x 15.84.
(f) f(x) = x
2
ln g(x)
f(2) = 4 ln g(2) = 4 ln 3 4.39
f
(2) = 4 ln 3
16
3
0.94.
Thus, we have point (2, 4.39) and slope m = 0.94. This gives
4.39 = 2(0.94) +b
b = 6.27, so
y = 0.94x + 6.27.
101. The curves meet when 1 x
3
/3 = x 1, that is when x
3
+ 3x 6 = 0. So the roots of this equation give us the
x-coordinates of the intersection point. By numerical methods, we see there is one solution near x = 1.3. See Figure 3.6.
Let
y1(x) = 1
x
3
3
and y2(x) = x 1.
So we have
y1
= x
2
and y2
= 1.
However, y2
(x) = +1, so if the curves are to be perpendicular when they cross, then y1
must be 1. Since y1
= x
2
,
y1
= 1 only at x = 1 which is not the point of intersection. The curves are therefore not perpendicular when they
cross.
2 1
1
2
20
15
10
5
5
y = x
3
+ 3x 6
x
y
Figure 3.6
105. Using the denition of cosh x and sinh x, we have cosh x
2
=
e
x
2
+e
x
2
2
and sinh x
2
=
e
x
2
e
x
2
2
. Therefore
lim
x
sinh(x
2
)
cosh(x
2
)
= lim
x
e
x
2
e
x
2
e
x
2
+e
x
2
= lim
x
e
x
2
(1 e
2x
2
)
e
x
2
(1 +e
2x
2
)
= lim
x
1 e
2x
2
1 +e
2x
2
= 1.
109. (a)
dg
dr
= GM
d
dr
_
1
r
2
_
= GM
d
dr
_
r
2
_
= GM(2)r
3
=
2GM
r
3
.
(b)
dg
dr
is the rate of change of acceleration due to the pull of gravity. The further away from the center of the earth, the
weaker the pull of gravity is. So g is decreasing and therefore its derivative,
dg
dr
, is negative.
(c) By part (a),
dg
dr
r=6400
=
2GM
r
3
r=6400
=
2(6.67 10
20
)(6 10
24
)
(6400)
3
3.05 10
6
.
(d) It is reasonable to assume that g is a constant near the surface of the earth.
113. (a) Differentiating, we see
v =
dy
dt
= 2y0 sin(2t)
a =
dv
dt
= 4
2
2
y0 cos(2t).
(b) We have
y = y0 cos(2t)
v = 2y0 sin(2t)
a = 4
2
2
y0 cos(2t).
So
Amplitude of y is |y0|,
Amplitude of v is |2y0| = 2|y0|,
Amplitude of a is |4
2
2
y0| = 4
2
2
|y0|.
The amplitudes are different (provided 2 = 1). The periods of the three functions are all the same, namely 1/.
(c) Looking at the answer to part (a), we see
d
2
y
dt
2
= a = 4
2
2
(y0 cos(2t))
= 4
2
2
y.
So we see that
d
2
y
dt
2
+ 4
2
2
y = 0.
117. (a) The statement f(2) = 4023 tells us that when the price is $2 per gallon, 4023 gallons of gas are sold.
(b) Since f(2) = 4023, we have f
1
(4023) = 2. Thus, 4023 gallons are sold when the price is $2 per gallon.
(c) The statement f
(2) = 1250 tells us that if the price increases from $2 per gallon, the sales decrease at a rate of
1250 gallons per $1 increase in price.
(d) The units of (f
1
)
(4023) are dollars per gallon. We have

(f
1
)
(4023) =
1
f
(f
1
(4023))
=
1
f
(2)
=
1
1250
= 0.0008.
Thus, when 4023 gallons are already sold, sales decrease at the rate of one gallon per price increase of 0.0008 dollars.
In others words, an additional gallon is sold if the price drops by 0.0008 dollars.
121. (a) We multiply through by h = f g and cancel as follows:
f
f
+
g
g
=
h
h
_
f
f
+
g
g
_
fg =
h
h
fg
f
f
fg +
g
g
fg =
h
h
h
f
g +g
f = h
,
which is the product rule.
(b) We start with the product rule, multiply through by 1/(fg) and cancel as follows:
f
g +g
f = h
(f
g +g
f)
1
fg
= h
1
fg
(f
g)
1
fg
+ (g
f)
1
fg
= h
1
fg
f
f
+
g
g
=
h
h
,
which is the additive rule shown in part (a).
125. (a) A CAS gives g
(r) = 0.
(b) Using the product rule,
g
(r) =
d
dr
(2
2r
) 4
r
+ 2
2r
d
dr
(4
r
) = 2 ln 2 2
2r
4
r
+ 2
2r
ln 4 4
r
= ln 4 2
2r
4
r
+ ln 4 2
2r
4
r
= (ln 4 + ln 4)2
2r
4
r
= 0 2
2r
4
r
= 0.
(c) By the laws of exponents, 4
r
= (2
2
)
r
= 2
2r
, so 2
2r
4
r
= 2
2r
2
2r
= 2
0
= 1. Therefore, its derivative is zero.
1. True. Since d(x
n
)/dx = nx
n1
, the derivative of a power function is a power function, so the derivative of a polynomial
is a polynomial.
5. True. Since f
(x) is the limit

f
(x) = lim
h0
f(x +h) f(x)
h
,
the function f must be dened for all x.
9. True; differentiating the equation with respect to x, we get
2y
dy
dx
+y +x
dy
dx
= 0.
Solving for dy/dx, we get that
dy
dx
=
y
2y +x
.
Thus dy/dx exists where 2y + x = 0. Now if 2y + x = 0, then x = 2y. Substituting for x in the original equation,
y
2
+xy 1 = 0, we get
y
2
2y
2
1 = 0.
This simplies to y
2
+ 1 = 0, which has no solutions. Thus dy/dx exists everywhere.
13. False. Since (sinh x)
= cosh x > 0, the function sinh x is increasing everywhere so can never repeat any of its values.
17. False; the fourth derivative of cos t+C, where C is any constant, is indeed cos t. But any function of the form cos t+p(t),
where p(t) is a polynomial of degree less than or equal to 3, also has its fourth derivative equal to cos t. So cos t +t
2
will
work.
21. False; for example, if both f and g are constant functions, then the derivative of f(g(x)) is zero, as is the derivative of
f(x). Another example is f(x) = 5x + 7 and g(x) = x + 2.
25. False. Let f(x) = e
x
and g(x) = x
2
. Let h(x) = f(g(x)) = e
x
2
. Then h
(x) = 2xe
x
2
and h
(x) = (2 +
4x
2
)e
x
2
. Since h
(0) < 0, clearly h is not concave up for all x.

29. False. For example, let f(x) = x + 5, and g(x) = 2x 3. Then f
(x) g
(x) for all x, but f(0) > g(0).

33. Let f be dened by
f(x) =
_
x if 0 x < 2
19 if x = 2
Then f is differentiable on (0, 2) and f
(x) = 1 for all x in (0, 2). Thus there is no c in (0, 2) such that
f
(c) =
f(2) f(0)
2 0
=
19
2
.
The reason that this function does not satisfy the conclusion of the Mean Value Theorem is that it is not continuous
at x = 2.
4.1 SOLUTIONS 63
CHAPTER FOUR
Exercises
1. f
(x) = 6x
2
+ 6x 36. To nd critical points, we set f
(x) = 0. Then
6(x
2
+x 6) = 6(x + 3)(x 2) = 0.
Therefore, the critical points of f are x = 3 and x = 2. To the left of x = 3, f
(x) > 0. Between x = 3 and x = 2,

f
(x) < 0. To the right of x = 2, f
(x) > 0. Thus f(3) is a local maximum, f(2) a local minimum. See Figure 4.1.
3 2
f(x) = 2x
3
+ 3x
2
36x
x
y
Figure 4.1
1 1
f(x) =
x
x
2
+ 1
x
y
Figure 4.2
5.
f
(x) =
x
2
+ 1 x 2x
(x
2
+ 1)
2
=
1 x
2
(x
2
+ 1)
2
=
(1 x)(1 +x)
(x
2
+ 1)
2
.
Critical points are x = 1. To the left of x = 1, f
(x) < 0.
Between x = 1 and x = 1, f
(x) > 0.
To the right of x = 1, f
(x) < 0.
So, f(1) is a local minimum, f(1) a local maximum. See Figure 4.2.
9. There are many possible answers. One possible graph is shown in Figure 4.3.
2 4 6 8
Critical point
Inflection point
x
Figure 4.3
64 Chapter Four /SOLUTIONS
13. From the graph of f(x) in the gure below, we see that the function must have two inection points. We calculate
f
(x) = 4x
3
+ 3x
2
6x, and f
(x) = 12x
2
+ 6x 6. Solving f
(x) = 0 we nd that:
x1 = 1 and x2 =
1
2
.
Since f
(x) > 0 for x < x1, f
(x) < 0 for x1 < x < x2, and f
(x) > 0 for x2 < x, it follows that both points are

inection points.
2
2
2
2
4
Inflection point
E
Inflection point
'
f(x)
x
17. (a) Decreasing for x < 1, increasing for 1 < x < 0, decreasing for 0 < x < 1, and increasing for x > 1.
(b) f(1) and f(1) are local minima, f(0) is a local maximum.
21. The inection points of f are the points where f
changes sign. See Figure 4.4.

x
f
(x) inflection points of f

Figure 4.4
Problems
25. To nd the critical points, we set the derivative equal to zero and solve for t.
F
(t) = Ue
t
+V e
t
(1) = 0
Ue
t
V
e
t
= 0
Ue
t
=
V
e
t
Ue
2t
= V
e
2t
=
V
U
2t = ln(V/U)
t =
ln(V/U)
2
.
The derivative F
(t) is never undened, so the only critical point is t = 0.5 ln(V/U).

29. See Figure 4.5.
4.1 SOLUTIONS 65
x
1
x
2
x
3
x
y
y
= 0 y
= 0 y
= 0
y
> 0 y
< 0 y
> 0 y
< 0
y = f(x)
y
< 0 everywhere
Figure 4.5
33. (a) It appears that this function has a local maximum at about x = 1, a local minimum at about x = 4, and a local
maximum at about x = 8.
(b) The table now gives values of the derivative, so critical points occur where f
(x) = 0. Since f
is continuous, this
occurs between 2 and 3, so there is a critical point somewhere around 2.5. Since f
is positive for values less than

2.5 and negative for values greater than 2.5, it appears that f has a local maximum at about x = 2.5. Similarly, it
appears that f has a local minimum at about x = 6.5 and another local maximum at about x = 9.5.
37. Differentiating using the product rule gives
f
(x) = 3x
2
(1 x)
4
4x
3
(1 x)
3
= x
2
(1 x)
3
(3(1 x) 4x) = x
2
(1 x)
3
(3 7x).
The critical points are the solutions to
f
(x) = x
2
(1 x)
3
(3 7x) = 0
x = 0, 1,
3
7
.
For x < 0, since 1 x > 0 and 3 7x > 0, we have f
(x) > 0.
For 0 < x <
3
7
, since 1 x > 0 and 3 7x > 0, we have f
(x) > 0.
For
3
7
< x < 1, since 1 x > 0 and 3 7x < 0, we have f
(x) < 0.
For 1 < x, since 1 x < 0 and 3 7x < 0, we have f
(x) > 0.
Thus, x = 0 is neither a local maximum nor a local minimum; x = 3/7 is a local maximum; x = 1 is a local
minimum.
41. First, we wish to have f
(6) = 0, since f(6) should be a local minimum:

f
(x) = 2x +a = 0
x =
a
2
= 6
a = 12.
Next, we need to have f(6) = 5, since the point (6, 5) is on the graph of f(x). We can substitute a = 12 into our
equation for f(x) and solve for b:
f(x) = x
2
12x +b
f(6) = 36 72 +b = 5
b = 31.
Thus, f(x) = x
2
12x + 31.
45. From the rst condition, we get that x = 2 is a local minimum for f. From the second condition, it follows that x = 4 is
an inection point. A possible graph is shown in Figure 4.6.
2 1 1 2 3 4 5 6
Local min
Point of inflection
x
Figure 4.6
49. Since the derivative of an even function is odd and the derivative of an odd function is even, f and f
are either both

odd or both even, and f
is the opposite. Graphs I and II represent odd functions; III represents an even function, so III
is f
. Since the maxima and minima of III occur where I crosses the x-axis, I must be the derivative of f
, that is, f
. In
addition, the maxima and minima of II occur where III crosses the x-axis, so II is f.
53. (a) Since f
(x) > 0 and g
(x) > 0 for all x, then f
(x) +g
(x) > 0 for all x, so f(x) +g(x) is concave up for all x.

(b) Nothing can be concluded about the concavity of (f + g)(x). For example, if f(x) = ax
2
and g(x) = bx
2
with
a > 0 and b < 0, then (f + g)
(x) = a + b. So f + g is either always concave up, always concave down, or a

straight line, depending on whether a > |b|, a < |b|, or a = |b|. More generally, it is even possible that (f + g)(x)
may have one or more changes in concavity.
(c) It is possible to have innitely many changes in concavity. Consider f(x) = x
2
+ cos x and g(x) = x
2
. Since
f
(x) = 2 cos x, we see that f(x) is concave up for all x. Clearly g(x) is concave down for all x. However,
f(x) +g(x) = cos x, which changes concavity an innite number of times.
Exercises
1. See Figure 4.7.
1 2 3 4 5
2
4
6
8
Global and local min
Global and local max
Local min
x
y
Figure 4.7
5. Since f(x) = x
3
3x
2
+ 20 is continuous and the interval 1 x 3 is closed, there must be a global maximum
and minimum. The candidates are critical points in the interval and endpoints. Since there are no points where f
(x) is
undened, we solve f
(x) = 0 to nd all the critical points:

f
(x) = 3x
2
6x = 3x(x 2) = 0,
so x = 0 and x = 2 are the critical points; both are in the interval. We then compare the values of f at the critical points
and the endpoints:
f(1) = 16, f(0) = 20, f(2) = 16, f(3) = 20.
Thus the global maximum is 20 at x = 0 and x = 3, and the global minimum is 16 at x = 1 and x = 2.
4.2 SOLUTIONS 67
9. Since f(x) = e
x
sin x is continuous and the interval 0 x 2 is closed, there must be a global maximum and
minimum. The possible candidates are critical points in the interval and endpoints. Since there are no points where f
is
undened, we solve f
(x) = 0 to nd all critical points:

f
(x) = e
x
sin x +e
x
cos x = e
x
(sin x + cos x) = 0.
Since e
x
= 0, the critical points are when sin x = cos x; the only solutions in the given interval are x = /4 and
x = 5/4. We then compare the values of f at the critical points and the endpoints:
f(0) = 0, f(/4) = e
/4
_
2
2
_
= 0.322, f(5/4) = e
5/4
_
2
2
_
= 0.0139, f(2) = 0.
Thus the global maximum is 0.322 at x = /4, and the global minimum is 0.0139 at x = 5/4.
13. (a) We have f
(x) = 10x
9
10 = 10(x
9
1). This is zero when x = 1, so x = 1 is a critical point of f. For values of
x less than 1, x
9
is less than 1, and thus f
(x) is negative when x < 1. Similarly, f
(x) is positive for x > 1. Thus

f(1) = 9 is a local minimum.
We also consider the endpoints f(0) = 0 and f(2) = 1004. Since f
(0) < 0 and f
(2) > 0, we see x = 0 and

x = 2 are local maxima.
(b) Comparing values of f shows that the global minimum is at x = 1, and the global maximum is at x = 2.
Problems
17. Differentiating gives
f
(x) = 1
1
x
2
,
so the critical points satisfy
1
1
x
2
= 0
x
2
= 1
x = 1 (We want x > 0).
Since f
is negative for 0 < x < 1 and f
is positive for x > 1, there is a local minimum at x = 1.

Since f(x) as x 0
+
and as x , the local minimum at x = 1 is a global minimum; there is no global
maximum. See Figure 4.8. The the global minimum is f(1) = 2.
1
2 f(x) = x + 1/x
x
Figure 4.8
21. Differentiating using the product rule gives
f
(t) = 2 sin t cos t cos t (sin

2
t + 2) sin t = 0
sin t(2 cos
2
t sin
2
t 2) = 0
sin t(2(1 sin
2
t) sin
2
t 2) = 0
sin t(3 sin
2
t) = 3 sin
3
t = 0.
Thus, the critical points are where sin t = 0, so
t = 0, , 2, 3, . . . .
Since f
(t) = 3 sin
3
t is negative for < t < 0, positive for 0 < t < , negative for < t < 2, and so on, we nd
that t = 0, 2, . . . give local minima, while t = , 3, . . . give local maxima. Evaluating gives
f(0) = f(2) = (0 + 2)1 = 2
f() = f(3) = (0 + 2)(1) = 2.
Thus, the global maximum of f(t) is 2, occurring at t = 0, 2, . . ., and the global minimum of f(t) is 2, occurring at
t = , 3, . . . . See Figure 4.9.
32 3 2
2
2
t
Figure 4.9
25. We want to maximize the height, y, of the grapefruit above the ground, as shown in the gure below. Using the derivative
we can nd exactly when the grapefruit is at the highest point. We can think of this in two ways. By common sense, at the
peak of the grapefruits ight, the velocity, dy/dt, must be zero. Alternately, we are looking for a global maximum of y,
so we look for critical points where dy/dt = 0. We have
dy
dt
= 32t + 50 = 0 and so t =
50
32
1.56 sec.
Thus, we have the time at which the height is a maximum; the maximum value of y is then
y 16(1.56)
2
+ 50(1.56) + 5 = 44.1 feet.
1 2 3
20
40
60
t
y
29. We set f
(r) = 0 to nd the critical points:

2A
r
3

3B
r
4
= 0
2Ar 3B
r
4
= 0
2Ar 3B = 0
r =
3B
2A
.
The only critical point is at r = 3B/(2A). If r > 3B/(2A), we have f
> 0 and if r < 3B/(2A), we have f
< 0.
Thus, the force between the atoms is minimized at r = 3B/(2A).
33. For x > 0, the line in Figure 4.10 has
Slope =
y
x
=
x
2
e
3x
x
= xe
3x
.
If the slope has a maximum, it occurs where
d
dx
(Slope) = 1 e
3x
3xe
3x
= 0
e
3x
(1 3x) = 0
x =
1
3
.
4.3 SOLUTIONS 69
For this x-value,
Slope =
1
3
e
3(1/3)
=
1
3
e
1
=
1
3e
.
Figure 4.10 shows that the slope tends toward 0 as x ; the formula for the slope shows that the slope tends toward 0
as x 0. Thus the only critical point, x = 1/3, must give a local and global maximum.
E '
x
T
c
y
(x, x
2
e
3x
)
x
y
Figure 4.10
200 400 600
1
2
3
Parasite
Drag
'
Induced
Drag
E
Total
Drag
c
speed
(mph)
drag
(thousands
of pounds)
Figure 4.11
37. (a) Figure 4.11 contains the graph of total drag, plotted on the same coordinate system with induced and parasite drag. It
was drawn by adding the vertical coordinates of Induced and Parasite drag.
(b) Airspeeds of approximately 160 mph and 320 mph each result in a total drag of 1000 pounds. Since two distinct
airspeeds are associated with a single total drag value, the total drag function does not have an inverse. The parasite
and induced drag functions do have inverses, because they are strictly increasing and strictly decreasing functions,
respectively.
(c) To conserve fuel, y the at the airspeed which minimizes total drag. This is the airspeed corresponding to the lowest
point on the total drag curve in part (a): that is, approximately 220 mph.
41. (a) We know that h
(x) < 0 for 2 x < 1, h
(1) = 0, and h
(x) > 0 for x > 1. Thus, h
(x) decreases to its

minimum value at x = 1, which we know to be zero, and then increases; it is never negative.
(b) Since h
(x) is non-negative for 2 x 1, we know that h(x) is never decreasing on [2, 1]. So a global
maximum must occur at the right hand endpoint of the interval.
(c) The graph below shows a function that is increasing on the interval 2 x 1 with a horizontal tangent and an
inection point at (1, 2).
2 1 1
(1, 2)
x
y
h(x)
Exercises
(b) We see in Figure 4.12 that in each case the graph of f is a parabola with one critical point, its vertex, on the positive
x-axis. The critical point moves to the right along the x-axis as a increases.
(c) To nd the critical points, we set the derivative equal to zero and solve for x.
f
(x) = 2(x a) = 0
x = a.
The only critical point is at x = a. As we saw in the graph, and as a increases, the critical point moves to the right.
c
large a
small a
x
Figure 4.12
c
large a
c
small a
x
Figure 4.13
(b) We see in Figure 4.13 that in each case f appears to have two critical points. One critical point is a local minimum at
the origin and the other is a local maximum in quadrant I. As the parameter a increases, the critical point in quadrant I
appears to move down and to the left, closer to the origin.
(c) To nd the critical points, we set the derivative equal to zero and solve for x. Using the product rule, we have:
f
(x) = x
2
e
ax
(a) + 2x e
ax
= 0
xe
ax
(ax + 2) = 0
x = 0 and x =
2
a
.
There are two critical points, at x = 0 and x = 2/a. As we saw in the graph, as a increases the nonzero critical point
moves to the left.
9. (a) Figure 4.14 shows the effect of varying a with b = 1.
(b) Figure 4.15 shows the effect of varying b with a = 1.
c
large a
T
small a
x
Figure 4.14
c
large b
E
small b
x
Figure 4.15
(c) In each case f appears to have two critical points, a local maximum in quadrant I and a local minimum on the
positive x-axis. From Figure 4.14 it appears that increasing a moves the local maximum up and does not move the
local minimum. From Figure 4.15 it appears that increasing b moves the local maximum up and to the right and
moves the local minimum to the right along the x-axis.
(d) To nd the critical points, we set the derivative equal to zero and solve for x. Using the product rule, we have:
f
(x) = ax 2(x b) +a (x b)
2
= 0
a(x b)(2x + (x b)) = 0
a(x b)(3x b) = 0
x = b or x =
b
3
.
There are two critical points, at x = b and at x = b/3. Increasing a does not move either critical point horizontally.
Increasing b moves both critical points to the right. This conrms what we saw in the graphs.
4.3 SOLUTIONS 71
Problems
13. (a) Writing y = L(1 +Ae
kt
)
1
, we nd the rst derivative by the chain rule
dy
dt
= L(1 +Ae
kt
)
2
(Ake
kt
) =
LAke
kt
(1 +Ae
kt
)
2
.
Using the quotient rule to calculate the second derivative gives
d
2
y
dt
2
=
_
LAk
2
e
kt
(1 +Ae
kt
)
2
2LAke
kt
(1 +Ae
kt
)(Ake
kt
)
_
(1 +Ae
kt
)
4
=
LAk
2
e
kt
(1 +Ae
kt
)
(1 +Ae
kt
)
3
.
(b) Since L, A > 0 and e
kt
> 0 for all t, the factor LAk
2
e
kt
and the denominator are never zero. Thus, possible
inection points occur where
1 +Ae
kt
= 0.
Solving for t gives
t =
ln A
k
.
(c) The second derivative is positive to the left of t = ln(A)/k and negative to the right, so the function changes from
concave up to concave down at t = ln(A)/k.
17. Cubic polynomials are all of the form f(x) = Ax
3
+ Bx
2
+ Cx + D. There is an inection point at the origin (0, 0)
if f
(0) = 0 and f(0) = 0. Since f(0) = D, we must have D = 0. Since f
(x) = 6Ax + 2B, giving f
(0) = 2B,
we must have B = 0. The family of cubic polynomials with inection point at the origin is the two parameter family
f(x) = Ax
3
+Cx.
10 10
100
100
a = 1
a = 20
x
Figure 4.16
(b) The function f(x) = x
2
+ a sin x is concave up for all x if f
(x) > 0 for all x. We have f
(x) = 2 a sin x.
Because sin x varies between 1 and 1, we have 2 a sin x > 0 for all x if 2 < a < 2 but not otherwise. Thus
f(x) is concave up for all x if 2 < a < 2.
25. Since the horizontal asymptote is y = 5, we know a = 5. The value of b can be any number. Thus y = 5(1 e
bx
) for
any b > 0.
29. Since y(0) = a/(1 +b) = 2, we have a = 2 + 2b. To nd a point of inection, we calculate
dy
dt
=
abe
t
(1 +be
t
)
2
,
and using the quotient rule,
d
2
y
dx
2
=
abe
t
(1 +be
t
)
2
abe
t
2(1 +be
t
)(be
t
)
(1 +be
t
)
4
=
abe
t
(1 +be
t
)
(1 +be
t
)
3
.
The second derivative is equal to 0 when be
t
= 1, or b = e
t
. When t = 1, we have b = e. The second derivative
changes sign at this point, so we have an inection point. Thus
y =
2 + 2e
1 +e
1t
.
33. Differentiating y = ae
x
+bx gives
dy
dx
= ae
x
+b.
Since the global minimum occurs for x = 1, we have a/e +b = 0, so b = a/e.
The value of the function at x = 1 is 2, so we have 2 = a/e +b, which gives
2 =
a
e
+
a
e
=
2a
e
,
so a = e and b = 1. Thus
y = e
1x
+x.
We compute d
2
y/dx
2
= e
1x
, which is always positive, so this conrms that x = 1 is a local minimum. Because
the value of e
1x
+x as x grows without bound, this local minimum is a global minimum.
37. Since lim
t
N = a, we have a = 200,000. Note that while N(t) will never actually reach 200,000, it will become
arbitrarily close to 200,000. Since N represents the number of people, it makes sense to round up long before t .
When t = 1, we have N = 0.1(200,000) = 20,000 people, so plugging into our formula gives
N(1) = 20,000 = 200,000
_
1 e
k(1)
_
.
Solving for k gives
0.1 = 1 e
k
e
k
= 0.9
k = ln 0.9 0.105.
41. (a) The graph of r has a vertical asymptote if the denominator is zero. Since (x b)
2
is nonnegative, the denominator
can only be zero if a 0. Then
a + (x b)
2
= 0
(x b)
2
= a
x b =
a
x = b
a.
In order for there to be a vertical asymptote, a must be less than or equal to zero. There are no restrictions on b.
(b) Differentiating gives
r
(x) =
1
(a + (x b)
2
)
2
2(x b),
so r
= 0 when x = b. If a 0, then r
is undened at the same points at which r is undened. Thus the only critical
point is x = b. Since we want r(x) to have a maximum at x = 3, we choose b = 3. Also, since r(3) = 5, we have
r(3) =
1
a + (3 3)
2
=
1
a
= 5 so a =
1
5
.
45. Let f(x) = Ae
Bx
2
. Since
f(x) = Ae
Bx
2
= Ae
(x0)
2
(1/B)
,
this is just the family of curves y = e
(xa)
2
b
multiplied by a constant A. This family of curves is discussed in the text;
here, a = 0, b =
1
B
. When x = 0, y = Ae
0
= A, so A determines the y-intercept. A also serves to atten or stretch the
graph of e
Bx
2
vertically. Since f
(x) = 2ABxe
Bx
2
, f(x) has a critical point at x = 0. For B > 0, the graphs are
bell-shaped curves centered at x = 0, and f(0) = A is a global maximum.
4.3 SOLUTIONS 73
To nd the inection points of f, we solve f
(x) = 0. Since f
(x) = 2ABxe
Bx
2
,
f
(x) = 2ABe
Bx
2
+ 4AB
2
x
2
e
Bx
2
.
Since e
Bx
2
is always positive, f
(x) = 0 when
2AB + 4AB
2
x
2
= 0
x
2
=
2AB
4AB
2
x =
_
1
2B
.
These are points of inection, since the second derivative changes sign here. Thus for large values of B, the inection
points are close to x = 0, and for smaller values of B the inection points are further from x = 0. Therefore B affects
the width of the graph.
In the graphs in Figure 4.17, A is held constant, and variations in B are shown.
x
'
Large B
'
Small B
Figure 4.17: f(x) = Ae
Bx
2
for varying B
49.
2
1
0
1
x
y
T
b = 4
T
b = 3
T
b = 2
c
b = 1
The larger the value of b, the narrower the humps and more humps per given region there are in the graph.
53. (a) To nd lim
r0
+ V (r), rst rewrite V (r) with a common denominator:
lim
r0
+
V (r) = lim
r0
+
A
r
12

B
r
6
= lim
r0
+
ABr
6
r
12
A
0
+
+.
As the distance between the two atoms becomes small, the potential energy diverges to +.
(b) The critical point of V (r) will occur where V
(r) = 0:
V
(r) =
12A
r
13
+
6B
r
7
= 0
12A+ 6Br
6
r
13
= 0
12A+ 6Br
6
= 0
r
6
=
2A
B
r =
_
2A
B
_
1/6
To determine whether this is a local maximum or minimum, i we can use the rst derivative test. Since r is positive,
the sign of V
(r) is determined by the sign of 12A+6Br

6
. Notice that this is an increasing function of r for r > 0,
so V
(r) changes sign from to + at r = (2A/B)

1/6
. The rst derivative test yields
r
_
2A
B
_
1/6
(r) neg. zero pos.

Thus V (r) goes from decreasing to increasing at the critical point r = (2A/B)
1/6
, so this is a local minimum.
(c) Since F(r) = V
(r), the force is zero exactly where V
(r) = 0, i.e. at the critical points of V . The only critical

point was the one found in part (b), so the only such point is r = (2A/B)
1/6
.
(d) Since the numerator in r = (2A/B)
1/6
is proportional to A
1/6
, the equilibrium size of the molecule increases when
the parameter A is increased. Conversely, since B is in the denominator, when B is increased the equilibrium size of
the molecules decrease.
57. (a) The vertical intercept is W = Ae
e
bc0
= Ae
e
b
. There is no horizontal intercept since the exponential function
is always positive. There is a horizontal asymptote. As t , we see that e
bct
= e
b
/e
ct
0, since t is positive.
Therefore W Ae
0
= A, so there is a horizontal asymptote at W = A.
(b) The derivative is
dW
dt
= Ae
e
bct
(e
bct
)(c) = Ace
e
bct
e
bct
.
Thus, dW/dt is always positive, so W is always increasing and has no critical points. The second derivative is
d
2
W
dt
2
=
d
dt
(Ace
e
bct
)e
bct
+Ace
e
bct
d
dt
(e
bct
)
= Ac
2
e
e
bct
e
bct
e
bct
+Ace
e
bct
(c)e
bct
= Ac
2
e
e
bct
e
bct
(e
bct
1).
Now e
bct
decreases frome
b
> 1 when t = 0 toward 0 as t . The second derivative changes sign from positive
to negative when e
bct
= 1, i.e., when b ct = 0, or t = b/c. Thus the curve has an inection point at t = b/c,
where W = Ae
e
b(b/c)c
= Ae
1
.
A = 50, b = 2, c = 1
A = 50, b = 2, c = 5
A = 20, b = 2, c = 1
t
W
Figure 4.18
(d) The nal size of the organism is given by the horizontal asymptote W = A. The curve is steepest at its inection
point, which occurs at t = b/c, W = Ae
1
. Since e = 2.71828 . . . 3, the size the organism when it is growing
fastest is about A/3, one third its nal size. So yes, the Gompertz growth function is useful in modeling such growth.
Exercises
1. We look for critical points of M:
dM
dx
=
1
2
wL wx.
Now dM/dx = 0 when x = L/2. At this point d
2
M/dx
2
= w so this point is a local maximum. The graph of M(x)
is a parabola opening downward, so the local maximum is also the global maximum.
4.4 SOLUTIONS 75
5. We only consider > 0. For such , the value of v as and as 0
+
. Thus, v does not have a maximum
velocity. It will have a minimum velocity. To nd it, we set dv/d = 0:
dv
d
= k
1
2
_
c
+
c
_
1/2
_
1
c

c
2
_
= 0.
Solving, and remembering that > 0, we obtain
1
c

c
2
= 0
1
c
=
c
2
= c
2
,
so
= c.
Thus, we have one critical point. Since
dv
d
< 0 for < c
and
dv
d
> 0 for > c,
the rst derivative test tells us that we have a local minimum of v at x = c. Since = c is the only critical point, it gives
the global minimum. Thus the minimum value of v is
v = k
_
c
c
+
c
c
=
2k.
9. A graph of F against is shown in Figure 4.19.
Taking the derivative:
dF
d
=
mg(cos sin )
(sin +cos )
2
.
At a critical point, dF/d = 0, so
cos sin = 0
tan =
1
= arctan
_
1
_
.
If = 0.15, then = arctan(1/0.15) = 1.422 81.5
. To calculate the maximum and minimum values of F, we

evaluate at this critical point and the endpoints:
At = 0, F =
0.15mg
sin 0 + 0.15 cos 0
= 1.0mg newtons.
At = 1.422, F =
0.15mg
sin(1.422) + 0.15 cos(1.422)
= 0.148mg newtons.
At = /2, F =
0.15mg
sin(
2
) + 0.15 cos(
2
)
= 0.15mg newtons.
Thus, the maximum value of F is 1.0mg newtons when = 0 (her arm is vertical) and the minimum value of F is
0.148mg newtons is when = 1.422 (her arm is close to horizontal). See Figure 4.20.
2
1.0mg
F =
0.15mg
sin +0.15 cos
F (newtons)
Figure 4.19
1.422
0.148mg
1.0mg
F =
0.15mg
sin +0.15 cos
F
Figure 4.20
13. Examination of the graph suggests that 0 x
3
e
x
2. The lower bound of 0 is the best possible lower bound since
f(0) = (0)
3
e
0
= 0.
To nd the best possible upper bound, we nd the critical points. Differentiating, using the product rule, yields
f
(x) = 3x
2
e
x
x
3
e
x
Setting f
(x) = 0 and factoring gives

3x
2
e
x
x
3
e
x
= 0
x
2
e
x
(3 x) = 0
So the critical points are x = 0 and x = 3. Note that f
(x) < 0 for x > 3 and f
(x) > 0 for x < 3, so f(x) has a local

maximum at x = 3. Examination of the graph tells us that this is the global maximum. So 0 x
3
e
x
f(3).
f(3) = 3
3
e
3
1.34425
So 0 x
3
e
x
3
3
e
3
1.34425 are the best possible bounds for the function.
1
3
3
e
3
1.34425
2
f(x) = x
3
e
x
Figure 4.21
Problems
17. (a) The rectangle on the left has area xy, and the semicircle on the right with radius y/2 has area (1/2)(y/2)
2
= y
2
/8.
We have
Area of entire region = xy +

8
y
2
.
(b) The perimeter of the gure is made of two horizontal line segments of length x each, one vertical segment of length
y, and a semicircle of radius y/2 of length y/2. We have
Perimeter of entire region = 2x +y +

2
y = 2x + (1 +

2
)y.
(c) We want to maximize the area xy +y
2
/8, with the perimeter condition 2x + (1 +/2)y = 100. Substituting
x = 50
_
1
2
+

4
_
y = 50
2 +
4
y
4.4 SOLUTIONS 77
into the area formula, we must maximize
A(y) =
_
50
_
1
2
+

4
_
y
_
y +

8
y
2
= 50y
_
1
2
+

8
_
y
2
on the interval
0 y 200/(2 +) = 38.8985
where y 0 because y is a length and y 200/(2 +) because x 0 is a length.
The critical point of A occurs where
A
(y) = 50
_
1 +

4
_
y = 0
at
y =
200
4 +
= 28.005.
A maximum for A must occur at the critical point y = 200/(4 + ) or at one of the endpoints y = 0 or y =
200/(2 +). Since
A(0) = 0 A
_
200
4 +
_
= 700.124 A
_
200
2 +
_
= 594.2
the maximum is at
y =
200
4 +
.
Hence
x = 50
2 +
4
y = 50
2 +
4
200
4 +
=
100
4 +
.
The dimensions giving maximum area with perimeter 100 are
x =
100
4 +
= 14.0 y =
200
4 +
= 28.0.
The length y is twice as great as x.
21. We want to minimize the surface area S of the box, shown in Figure 4.22. The box has 5 faces: the bottom which has area
x
2
and the four sides, each of which has area xh. Thus we want to minimize
S = x
2
+ 4xh.
The volume of the box is 8 = x
2
h, so h = 8/x
2
. Substituting this expression in for h in the formula for S gives
S = x
2
+ 4x
8
x
2
= x
2
+
32
x
.
Differentiating gives
dS
dx
= 2x
32
x
2
.
To minimize S we look for critical points, so we solve 0 = 2x 32/x
2
. Multiplying by x
2
gives
0 = 2x
3
32,
so x = 16
1/3
cm. Then we can nd
h =
8
x
2
=
8
16
2/3
=
16
2 16
2/3
=
16
1/3
2
cm.
We can check that this critical point is a minimum of S by checking the sign of
d
2
S
dx
2
= 2 +
64
x
3
which is positive when x > 0. So S is concave up at the critical point and therefore x = 16
1/3
gives a minimum value
of S.
x
x
h
Figure 4.22
25. The geometry shows that as x increases from 0 to 10 the length y decreases. The minimum value of y is 5 when x = 10,
and the maximum value of y is
10
2
+ 5
2
=
125 when x = 0.
We can also use calculus. By the Pythagorean theorem we have
y =
_
5
2
+ (10 x)
2
.
The extreme values of
y = f(x) =
_
5
2
+ (10 x)
2
for 0 x 10 occur at the endpoints x = 0 or x = 10 or at a critical point of f. The only solution of the equation
f
(x) =
x 10
_
5
2
+ (10 x)
2
= 0
is x = 10, which is the only critical point of f. We have
f(0) =
125 = 11.18 f(10) = 5.

Therefore the minimum value of y is 5 and the maximum value is
125 = 5
5.
29. The rectangle in Figure 4.23 has area, A, given by
A = 2xy =
2x
1 +x
2
for x 0.
At a critical point,
dA
dx
=
2
1 +x
2
+ 2x
_
2x
(1 +x
2
)
2
_
= 0
2(1 +x
2
2x
2
)
(1 +x
2
)
2
= 0
1 x
2
= 0
x = 1.
Since A = 0 for x = 0 and A 0 as x , the critical point x = 1 is a local and global maximum for the area. Then
y = 1/2, so the vertices are
(1, 0) , (1, 0) ,
_
1,
1
2
_
,
_
1,
1
2
_
.
T
c
y
E '
x
y =
1
1 +x
2
x
y
Figure 4.23
4.4 SOLUTIONS 79
33. From the triangle shown in Figure 4.24, we see that
_
w
2
_
2
+
_
h
2
_
2
= 30
2
w
2
+h
2
= 4(30)
2
= 3600.
h/2
30
w/2
Figure 4.24
The strength, S, of the beam is given by
S = kwh
2
,
for some constant k. To make S a function of only one variable, substitute for h
2
, giving
S = kw(3600 w
2
) = k(3600w w
3
).
Differentiating and setting dS/dw = 0,
dS
dw
= k(3600 3w
2
) = 0.
Solving for w gives
w =
1200 = 34.64 cm,

so
h
2
= 3600 w
2
= 3600 1200 = 2400
h =
2400 = 48.99 cm.

Thus, w = 34.64 cm and h = 48.99 cm give a critical point. To check that this is a local maximum, we compute
d
2
S
dw
2
= 6w < 0 for w > 0.
Since d
2
S/dw
2
< 0, we see that w = 34.64 cm is a local maximum. It is the only critical point, so it is a global
maximum.
37. Any point on the curve can be written (x, x
2
). The distance between such a point and (3, 0) is given by
s(x) =
_
(3 x)
2
+ (0 x
2
)
2
=
_
(3 x)
2
+x
4
.
Plotting this function in Figure 4.25, we see that there is a minimum near x = 1.
To nd the value of x that minimizes the distance we can instead minimize the function Q = s
2
(the derivative is
simpler). Then we have
Q(x) = (3 x)
2
+x
4
.
Differentiating Q(x) gives
dQ
dx
= 6 + 2x + 4x
3
.
Plotting the function 4x
3
+2x6 shows that there is one real solution at x = 1, which can be veried by substitution; the
required coordinates are therefore (1, 1). Because Q
(x) = 2 + 12x
2
is always positive, x = 1 is indeed the minimum.
See Figure 4.26.
4 2 0 2 4
2.5
5
7.5
10
12.5
15
17.5
x
y
Figure 4.25
4 2 2 4
100
50
50
100
x
y
Figure 4.26
41. (a) Suppose n passengers sign up for the cruise. If n 100, then the cruises revenue is R = 2000n, so the maximum
revenue is
R = 2000 100 = 200,000.
If n > 100, then the price is
p = 2000 10(n 100)
and hence the revenue is
R = n(2000 10(n 100)) = 3000n 10n
2
.
To nd the maximum revenue, we set dR/dn = 0, giving 20n = 3000 or n = 150. Then the revenue is
R = (2000 10 50) 150 = 225,000.
Since this is more than the maximum revenue when n 100, the boat maximizes its revenue with 150 passengers,
each paying $1500.
(b) We approach this problem in a similar way to part (a), except now we are dealing with the prot function . If
n 100, we have
= 2000n 80,000 400n,
so is maximized with 100 passengers yielding a prot of
= 1600 100 80,000 = $80,000.
If n > 100, we have
= n(2000 10(n 100)) (80,000 + 400n).
We again set d/dn = 0, giving 2600 = 20n, so n = 130. The prot is then $89,000. So the boat maximizes prot
by boarding 130 passengers, each paying $1700. This gives the boat $89,000 in prot.
45. (a) The line in the left-hand gure has slope equal to the rate worms arrive. To understand why, see line (1) in the
right-hand gure. (This is the same line.) For any point Q on the loading curve, the line PQ has slope
QT
PT
=
QT
PO +OT
=
load
traveling time + searching time
.
(b) The slope of the line PQ is maximized when the line is tangent to the loading curve, which happens with line (2).
The load is then approximately 7 worms.
(c) If the traveling time is increased, the point P moves to the left, to point P
, say. If line (3) is tangent to the curve, it

will be tangent to the curve further to the right than line (2), so the optimal load is larger. This makes sense: if the
bird has to y further, youd expect it to bring back more worms each time.
P O
4
8
Number of worms
time
load
(number of worms)
P
P
Traveling time
O T
Searching time
8
Q
Number of worms
(1)
(2)
(3)
time
load
(number of worms)
4.4 SOLUTIONS 81
49. (a) Since the speed of light is a constant, the time of travel is minimized when the distance of travel is minimized. From
Figure 4.27,
Distance
OP =
_
x
2
+ 1
2
=
_
x
2
+ 1
Distance
PQ =
_
(2 x)
2
+ 1
2
=
_
(2 x)
2
+ 1
Thus,
Total distance traveled = s =
_
x
2
+ 1 +
_
(2 x)
2
+ 1.
The total distance is a minimum if
ds
dx
=
1
2
(x
2
+ 1)
1/2
2x +
1
2
((2 x)
2
+ 1)
1/2
2(2 x)(1) = 0,
giving
x
x
2
+ 1
2 x
_
(2 x)
2
+ 1
= 0
x
x
2
+ 1
=
2 x
_
(2 x)
2
+ 1
Squaring both sides gives
x
2
x
2
+ 1
=
(2 x)
2
(2 x)
2
+ 1
.
Cross multiplying gives
x
2
((2 x)
2
+ 1) = (2 x)
2
(x
2
+ 1).
Multiplying out
x
2
(4 4x +x
2
+ 1) = (4 4x +x
2
)(x
2
+ 1)
4x
2
4x
3
+x
4
+x
2
= 4x
2
4x
3
+x
4
+ 4 4x +x
2
.
Collecting terms and canceling gives
0 = 4 4x
x = 1.
We can see that this value of x gives a minimum by comparing the value of s at this point and at the endpoints,
x = 0, x = 2.
At x = 1,
s =
_
1
2
+ 1 +
_
(2 1)
2
+ 1 = 2.83.
At x = 0,
s =
_
0
2
+ 1 +
_
(2 0)
2
+ 1 = 3.24.
At x = 2,
s =
_
2
2
+ 1 +
_
(2 2)
2
+ 1 = 3.24.
Thus the shortest travel time occurs when x = 1; that is, when P is at the point (1, 1).
x
(2 x) 2
A
O
1
P = (x, 1)
x
y
Figure 4.27
(b) Since x = 1 is halfway between x = 0 and x = 2, the angles 1 and 2 are equal.
Exercises
1. The xed costs are $5000, the marginal cost per item is $2.40, and the price per item is $4.
5. (a) Total cost, in millions of dollars, C(q) = 3 + 0.4q.
(b) Revenue, in millions of dollars, R(q) = 0.5q.
(c) Prot, in millions of dollars, (q) = R(q) C(q) = 0.5q (3 + 0.4q) = 0.1q 3.
9. (a) At q = 5000, MR > MC, so the marginal revenue to produce the next item is greater than the marginal cost. This
means that the company will make money by producing additional units, and production should be increased.
(b) Prot is maximized where MR = MC, and where the prot function is going from increasing (MR > MC) to
decreasing (MR < MC). This occurs at q = 8000.
Problems
13. (a) (q) is maximized when R(q) > C(q) and they are as far apart as possible. See Figure 4.28.
(b)
(q0) = R
(q0) C
(q0) = 0 implies that C
(q0) = R
(q0) = p.
Graphically, the slopes of the two curves at q0 are equal. This is plausible because if C
(q0) were greater than p

or less than p, the maximum of (q) would be to the left or right of q0, respectively. In economic terms, if the cost
were rising more quickly than revenues, the prot would be maximized at a lower quantity (and if the cost were rising
more slowly, at a higher quantity).
q
0
C(q)
R(q)
s
maximum (q)
q
$
Figure 4.28
q
0
p
q
C
(q)
$
Figure 4.29
17. For each month,
Prot = Revenue Cost
= pq wL = pcK
wL
The variable on the right is L, so at the maximum
d
dL
= pcK
L
1
w = 0
Now 1 is negative, since 0 < < 1, so 1 is positive and we can write
pcK
L
1
= w
giving
L =
_
pcK
w
_
1
1
Since 1 is negative, when L is just above 0, the quantity L
1
is huge and positive, so d/dL > 0. When L is large,
L
1
is small, so d/dL < 0. Thus the value of L we have found gives a global maximum, since it is the only critical
point.
4.5 SOLUTIONS 83
21. (a) a(q) = C(q)/q, so C(q) = 0.01q
3
0.6q
2
+ 13q.
(b) Taking the derivative of C(q) gives an expression for the marginal cost:
C
(q) = MC(q) = 0.03q

2
1.2q + 13.
To nd the smallest MC we take its derivative and nd the value of q that makes it zero. So: MC
(q) = 0.06q1.2 =
0 when q = 1.2/0.06 = 20. This value of q must give a minimum because the graph of MC(q) is a parabola
opening upward. Therefore the minimum marginal cost is MC(20) = 1. So the marginal cost is at a minimum when
the additional cost per item is $1.
(c) a
(q) = 0.02q 0.6

Setting a
(q) = 0 and solving for q gives q = 30 as the quantity at which the average is minimized, since the graph
of a is a parabola which opens upward. The minimum average cost is a(30) = 4 dollars per item.
(d) The marginal cost at q = 30 is MC(30) = 0.03(30)
2
1.2(30) + 13 = 4. This is the same as the average cost at
this quantity. Note that since a(q) = C(q)/q, we have a
(q) = (qC
(q) C(q))/q
2
. At a critical point, q0, of a(q),
we have
0 = a
(q0) =
q0C
(q0) C(q0)
q
2
0
,
so C
(q0) = C(q0)/q0 = a(q0). Therefore C
(30) = a(30) = 4 dollars per item.

Another way to see why the marginal cost at q = 30 must equal the minimum average cost a(30) = 4 is to view
C
(30) as the approximate cost of producing the 30

th
or 31
st
good. If C
(30) < a(30), then producing the 31

st
good would lower the average cost, i.e. a(31) < a(30). If C
(30) > a(30), then producing the 30

th
good would
raise the average cost, i.e. a(30) > a(29). Since a(30) is the global minimum, we must have C
(30) = a(30).
2 4
2
4
C = 4
C = 3
C = 2
Q = x
1/2
y
1/2
x
y
Figure 4.30
(b) Comparing the lines C = 2, C = 3, C = 4, we see that the cost increases as we move away from the origin. The
line C = 2 does not cut the curve Q = 1; the lines C = 3 and C = 4 cut twice.
The minimum cost occurs where a cost line is tangent to the production curve.
(c) Using implicit differentiation, the slope of x
1/2
y
1/2
= 1 is given by
1
2
x
1/2
y
1/2
+
1
2
x
1/2
y
1/2
y
= 0
y
=
x
1/2
y
1/2
x
1/2
y
1/2
=
y
x
.
The cost lines all have slope 2. Thus, if the curve is tangent to a line, we have
y
x
= 2
y = 2x.
Substituting into Q = x
1/2
y
1/2
= 1 gives
x
1/2
(2x)
1/2
= 1
2x = 1
x =
1
2
y = 2
1
2
=
2.
Thus the minimum cost is
C = 2
1
2
+
2 = 2
2.
Exercises
1. The rate of growth, in billions of people per year, was
dP
dt
= 6.7(0.011)e
0.011t
.
On January 1, 2007, we have t = 0, so
dP
dt
= 6.7(0.011)e
0
= 0.0737 billion/year = 73.7 million people/year.
5. The rate of change of the power dissipated is given by
dP
dR
=
81
R
2
.
9. The rate of change of velocity is given by
dv
dt
=
mg
k
_
k
m
e
kt/m
_
= ge
kt/m
.
When t = 0,
dv
dt
t=0
= g.
When t = 1,
dv
dt
t=1
= ge
k/m
.
These answers give the acceleration at t = 0 and t = 1. The acceleration at t = 0 is g, the acceleration due to gravity,
and at t = 1, the acceleration is ge
k/m
, a smaller value.
13. We know dR/dt = 0.2 when R = 5 and V = 9 and we want to know dI/dt. Differentiating I = V/R with V constant
gives
dI
dt
= V
_
1
R
2
dR
dt
_
,
so substituting gives
dI
dt
= 9
_
1
5
2
0.2
_
= 0.072 amps per second.
17. We have
dA
dt
=
9
16
[4 4 cos(4)]
d
dt
.
So
dA
dt
=/4
=
9
16
_
4 4 cos
_
4

4
__
0.2 =
9
16
(4 + 4)0.2 = 0.9 cm
2
/min.
Problems
21. Since the radius is 3 feet, the volume of the gas when the depth is h is given by
V = 3
2
h = 9h.
We want to nd dV/dt when h = 4 and dh/dt = 0.2. Differentiating gives
dV
dt
= 9
dh
dt
= 9(0.2) = 5.655 ft
3
/sec.
Notice that the value H = 4 is not used. This is because V is proportional to h so dV/dt does not depend on h.
4.6 SOLUTIONS 85
25. The rate of change of temperature with distance, dH/dy, at altitude 4000 ft approximated by
dH
dy

H
y
=
38 52
6 4
= 7
F/thousand ft.
A speed of 3000 ft/min tells us dy/dt = 3000, so
Rate of change of temperature with time =
dH
dy

dy
dt
7
F
thousand ft

3 thousand ft
min
= 21
F/min.
Other estimates can be obtained by estimating the derivative as
dH
dy

H
y
=
52 60
4 2
= 4
F/thousand ft
or by averaging the two estimates
dH
dy

7 4
2
= 5.5
F/thousand ft.
If the rate of change of temperature with distance is 4
/thousand ft, then

Rate of change of temperature with time =
dH
dy

dy
dt
4
F
thousand ft

3 thousand ft
min
= 12
F/min.
Thus, estimates for the rate at which temperature was decreasing range from 12
F/min to 21
F/min.
29. (a) The surface of the water is circular with radius r cm. Applying Pythagoras Theorem to the triangle in Figure 4.31
shows that
(10 h)
2
+r
2
= 10
2
so
r =
_
10
2
(10 h)
2
=
_
20h h
2
cm.
(b) We know dh/dt = 0.1 cm/hr and we want to know dr/dt when h = 5 cm. Differentiating
r =
_
20h h
2
gives
dr
dt
=
1
2
(20h h
2
)
1/2
_
20
dh
dt
2h
dh
dt
_
=
10 h
20h h
2
dh
dt
.
Substituting dh/dt = 0.1 and h = 5 gives
dr
dt
h=5
=
5
20 5 5
2
(0.1) =
1
2
75
= 0.0577 cm/hr.
Thus, the radius is decreasing at 0.0577 cm per hour.
T
c
h
T
c
(10 h)
10
s
r
Figure 4.31
33. Let the volume of clay be V. The clay is in the shape of a cylinder, so V = r
2
L. We know dL/dt = 0.1 cm/sec and we
want to know dr/dt when r = 1 cm and L = 5 cm. Differentiating with respect to time t gives
dV
dt
= 2rL
dr
dt
+r
2
dL
dt
.
However, the amount of clay is unchanged, so dV/dt = 0 and
2rL
dr
dt
= r
2
dL
dt
,
therefore
dr
dt
=
r
2L
dL
dt
.
When the radius is 1 cm and the length is 5 cm, and the length is increasing at 0.1 cm per second, the rate at which
the radius is changing is
dr
dt
=
1
2 5
0.1 = 0.01 cm/sec.
Thus, the radius is decreasing at 0.01 cm/sec.
37. From Figure 4.32, Pythagoras Theorem shows that the ground distance, d, between the train and the point, B, vertically
below the plane is given by
d
2
= x
2
+y
2
.
Figure 4.33 shows that
z
2
= d
2
+ 4
2
so
z
2
= x
2
+y
2
+ 4
2
.
We know that when x = 1, dx/dt = 80, y = 5, dy/dt = 500, and we want to know dz/dt. First, we nd z:
z
2
= 1
2
+ 5
2
+ 4
2
= 42, so z =
42.
Differentiating z
2
= x
2
+y
2
+ 4
2
gives
2z
dz
dt
= 2x
dx
dt
+ 2y
dy
dt
.
Canceling 2s and substituting gives
42
dz
dt
= 1(80) + 5(500)
dz
dt
=
2580
42
= 398.103 mph.
A
B
d
T
c
y
E '
x
Train
Plane: 4 miles above
this point
Figure 4.32: View from air
A d
B
4 miles
Plane
z
Train
Figure 4.33: Vertical view
4.7 SOLUTIONS 87
41. The volume of a cube is V = x
3
. So
dV
dt
= 3x
2
dx
dt
,
and
1
V
dV
dt
=
3
x
dx
dt
.
The surface area of a cube is A = 6x
2
. So
dA
dt
= 12x
dx
dt
,
and
1
A
dA
dt
=
2
x
dx
dt
.
Thus the percentage rate of change of the volume of the cube,
1
V
dV
dt
, is larger.
45. (a) Since the elevator is descending at 30 ft/sec, its height from the ground is given by h(t) = 300 30t, for 0
t 10.
(b) From the triangle in the gure,
tan =
h(t) 100
150
=
300 30t 100
150
=
200 30t
150
.
Therefore
= arctan
_
200 30t
150
_
and
d
dt
=
1
1 +
_
20030t
150
_
2

_
30
150
_
=
1
5
_
150
2
150
2
+ (200 30t)
2
_
.
Notice that
d
dt
is always negative, which is reasonable since decreases as the elevator descends.
(c) If we want to know when changes (decreases) the fastest, we want to nd out when d/dt has the largest magnitude.
This will occur when the denominator, 150
2
+ (200 30t)
2
, in the expression for d/dt is the smallest, or when
200 30t = 0. This occurs when t =
200
30
seconds, and so h(
200
30
) = 100 feet, i.e., when the elevator is at the level
of the observer.
Exercises
1. Since f
(a) > 0 and g
(a) < 0, lHopitals rule tells us that

lim
xa
f(x)
g(x)
=
f
(a)
g
(a)
< 0.
5. The denominator approaches zero as x goes to zero and the numerator goes to zero even faster, so you should expect that
the limit to be 0. You can check this by substituting several values of x close to zero. Alternatively, using lHopitals rule,
we have
lim
x0
x
2
sin x
= lim
x0
2x
cos x
= 0.
9. The larger power dominates. Using lHopitals rule
lim
x
x
5
0.1x
7
= lim
x
5x
4
0.7x
6
= lim
x
20x
3
4.2x
5
= lim
x
60x
2
21x
4
= lim
x
120x
84x
3
= lim
x
120
252x
2
= 0
so 0.1x
7
dominates.
Problems
13. We want to nd lim
x
f(x), which we do by three applications of lHopitals rule:
lim
x
2x
3
+ 5x
2
3x
3
1
= lim
x
6x
2
+ 10x
9x
2
= lim
x
12x + 10
18x
= lim
x
12
18
=
2
3
.
So the line y = 2/3 is the horizontal asymptote.
17. We have limx1 x = 1 and limx1(x 1) = 0, so lHopitals rule does not apply.
21. This is an
0
form. With y = limx(1 +x)
1/x
, we take logarithms to get
ln y = lim
x
1
x
ln(1 +x).
This limit is a 0 form,
lim
x
1
x
ln(1 +x),
which can be rewritten as the /form
lim
x
ln(1 +x)
x
,
to which lHopitals rule applies.
25. Let f(x) = ln x and g(x) = 1/x so f
(x) = 1/x and g
(x) = 1/x
2
and
lim
x0
+
ln x
1/x
= lim
x0
+
1/x
1/x
2
= lim
x0
+
x
1
= 0.
29. To get this expression in a form in which lHopitals rule applies, we combine the fractions:
1
x

1
sin x
=
sin x x
xsin x
.
Letting f(x) = sin x x and g(x) = xsin x, we have f(0) = 0 and g(0) = 0 so lHopitals rule can be used.
Differentiating gives f
(x) = cos x 1 and g
(x) = xcos x + sin x, so f
(0) = 0 and g
(0) = 0, so f
(0)/g
(0) is
undened. Therefore, to apply lHopitals rule we differentiate again to obtain f
(x) = sin x and g
(x) = 2 cos x
xsin x, for which f
(0) = 0 and g
(0) = 2 = 0. Then
lim
x0
_
1
x

1
sin x
_
= lim
x0
_
sin x x
xsin x
_
= lim
x0
_
cos x 1
xcos x + sin x
_
= lim
x0
_
sin x
2 cos x xsin x
_
=
0
2
= 0.
33. Since limt0 sin
2
At = 0 and limt0 cos At 1 = 1 1 = 0, this is a 0/0 form. Applying lHopitals rule we get
lim
t0
sin
2
At
cos At 1
= lim
t0
2Asin At cos At
Asin At
= lim
t0
2 cos At = 2.
37. Let f(x) = cos x and g(x) = x. Observe that since f(0) = 1, lHopitals rule does not apply. But since g(0) = 0,
lim
x0
cos x
x
does not exist.
41. Let k = n/2, so k as n . Thus,
lim
n
_
1 +
2
n
_
n
= lim
k
_
1 +
1
k
_
2k
= lim
k
_
_
1 +
1
k
_
k
_
2
= e
2
.
4.8 SOLUTIONS 89
45. This limit is of the form 0
0
so we apply lHopitals rule to
ln f(t) =
ln
_
(3
t
+ 5
t
)/2
_
t
.
We have
lim
t
ln f(t) = lim
t
_
(ln 3)3
t
+ (ln 5)5
t
_
/
_
3
t
+ 5
t
_
1
= lim
t
(ln 3)3
t
+ (ln 5)5
t
3
t
+ 5
t
= lim
t
ln 3 + (ln 5)(5/3)
t
1 + (5/3)
t
=
ln 3 + 0
1 + 0
= ln 3.
Thus
lim
t
f(t) = lim
t
e
ln f(t)
= e
lim
t
ln f(t)
= e
ln 3
= 3.
49. To evaluate, we use lHopitals Rule:
lim
x0
1 cosh 3x
x
= lim
x0
3 sinh 3x
1
= 0.
53. Since the limit is of the form 0/0, we can apply lHopitals rule. We have
lim
x/2
1 sin x + cos x
sin x + cos x 1
= lim
x/2
cos x sin x
cos x sin x
=
1
1
= 1.
Exercises
1. Between times t = 0 and t = 1, x goes at a constant rate from 0 to 1 and y goes at a constant rate from 1 to 0. So the
particle moves in a straight line from (0, 1) to (1, 0). Similarly, between times t = 1 and t = 2, it goes in a straight line
to (0, 1), then to (1, 0), then back to (0, 1). So it traces out the diamond shown in Figure 4.34.
1 1
1
1
x
y
t = 0, t = 4
t = 1
t = 2
t = 3
Figure 4.34
5. For 0 t

2
, we have x = sin t increasing and y = cos t decreasing, so the motion is clockwise for 0 t

2
.
Similarly, we see that the motion is clockwise for the time intervals

2
t , t
3
2
, and
3
2
t 2.
9. Let f(t) = ln t. Then f
(t) =
1
t
. The particle is moving counterclockwise when f
(t) > 0, that is, when t > 0. Any

other time, when t 0, the position is not dened.
13. We have
dx
dt
= 2 sin 2t,
dy
dt
= cos t.
The speed is
v =
_
4 sin
2
(2t) + cos
2
t.
Thus, v = 0 when sin(2t) = cos t = 0, and so the particle stops when t = /2, 3/2, . . . or t = (2n+1)
2
, for any
integer n.
17. One possible answer is x = 2, y = t.
21. The ellipse x
2
/25 +y
2
/49 = 1 can be parameterized by x = 5 cos t, y = 7 sin t, 0 t 2.
25. We have
dy
dx
=
dy/dt
dx/dt
=
4 cos(4t)
3 cos(3t)
.
Thus when t = , the slope of the tangent line is 4/3. Since x = 0 and y = 0 when t = , the equation of the tangent
line is y = (4/3)x.
29. We see from the parametric equations that the particle moves along a line. It sufces to plot two points: at t = 0, the
particle is at point (1, 4), and at t = 1, the particle is at point (4, 3). Since x increases as t increases, the motion is
left to right on the line as shown in Figure 4.35.
Alternately, we can solve the rst equation for t, giving t = (x 1)/3, and substitute this into the second equation
to get
y =
x 1
3
4 =
1
3
x
13
3
.
The line is y =
1
3
x
13
3
.
2 2 4 6
4
2
t = 0
x
y
Figure 4.35
3 3
3
3
t = 0
x
y
Figure 4.36
33. The graph is a circle centered at the origin with radius 3. The equation is
x
2
+y
2
= (3 cos t)
2
+ (3 sin t)
2
= 9.
The particle is at the point (3, 0) when t = 0, and motion is counterclockwise. See Figure 4.36.
Problems
37. (a) Eliminating t between
x = 2 +t, y = 4 + 3t
gives
y 4 = 3(x 2),
y = 3x 2.
4.8 SOLUTIONS 91
Eliminating t between
x = 1 2t, y = 1 6t
gives
y 1 = 3(x 1),
y = 3x 2.
Since both parametric equations give rise to the same equation in x and y, they both parameterize the same line.
(b) Slope = 3, y-intercept = 2.
41. It is a straight line through the point (3, 5) with slope 1. A linear parameterization of the same line is x = 3 + t,
y = 5 t.
45. (a) (i) A horizontal tangent occurs when dy/dt = 0 and dx/dt = 0. Thus,
dy
dt
= 6e
2t
2e
2t
= 0
6e
2t
= 2e
2t
e
4t
=
1
3
4t = ln
1
3
t =
1
4
ln
1
3
= 0.25 ln 3 = 0.275.
We need to check that dx/dt = 0 when t = 0.25 ln 3. Since dx/dt = 2e
2t
+2e
2t
is always positive, dx/dt
is never zero.
(ii) A vertical tangent occurs when dx/dt = 0 and dy/dt = 0. Since dx/dt = 2e
2t
+ 2e
2t
is always positive,
there is no vertical tangent.
(b) The chain rule gives
dy
dx
=
dy/dt
dx/dt
=
6e
2t
2e
2t
2e
2t
+ 2e
2t
=
3e
2t
e
2t
e
2t
+e
2t
.
(c) As t , we have e
2t
0. Thus,
lim
t
dy
dx
= lim
t
3e
2t
e
2t
e
2t
+e
2t
= lim
t
3e
2t
e
2t
= 3.
As t , the fraction gets closer and closer to 3.
49. (a) To determine if the particles collide, we check whether they are ever at the same point at the same time. We rst set
the two x-coordinates equal to each other:
4t 4 = 3t
t = 4.
When t = 4, both x-coordinates are 12. Now we check whether the y-coordinates are also equal at t = 4:
yA(4) = 2 4 5 = 3
yB(4) = 4
2
2 4 1 = 7.
Thus, the particles do not collide since they are not at the same point at the same time.
(b) For the particles to collide, we need both x- and y-coordinates to be equal. Since the x-coordinates are equal at t = 4,
we nd the k value making yA(4) = yB(4).
Substituting t = 4 into yA(t) = 2t k and yB(t) = t
2
2t 1, we have
8 k = 16 8 1
k = 1.
(c) To nd the speed of the particles, we differentiate.
For particle A,
x(t) = 4t 4, so x
(t) = 4, and x
(4) = 4
y(t) = 2t 1, so y
(t) = 2, and y
(4) = 2
Speed
A
=
_
(x
(t))
2
+ (y
(t))
2
=
_
4
2
+ 2
2
=
20.
For particle B,
x(t) = 3t, so x
(t) = 3, and x
(4) = 3
y(t) = t
2
2t 1, so y
(t) = 2t 2, and y
(4) = 6
Speed
B
=
_
(x
(t))
2
+ (y
(t))
2
=
_
3
2
+ 6
2
=
45.
Thus, when t = 4, particle B is moving faster.
53. (a) The particle touches the x-axis when y = 0. Since y = cos(2t) = 0 for the rst time when 2t = /2, we have
t = /4. To nd the speed of the particle at that time, we use the formula
Speed =
_
_
dx
dt
_
2
+
_
dy
dt
_
2
=
_
(cos t)
2
+ (2 sin(2t))
2
.
When t = /4,
Speed =
_
(cos(/4))
2
+ (2 sin(/2))
2
=
_
(
2/2)
2
+ (2 1)
2
=
_
9/2.
(b) The particle is at rest when its speed is zero. Since
_
(cos t)
2
+ (2 sin(2t))
2
0, the speed is zero when
cos t = 0 and 2 sin(2t) = 0.
Now cos t = 0 when t = /2 or t = 3/2. Since 2 sin(2t) = 4 sin t cos t, we see that this expression also
equals zero when t = /2 or t = 3/2.
(c) We need to nd d
2
y/dx
2
. First, we must determine dy/dx. We know
dy
dx
=
dy/dt
dx/dt
=
2 sin 2t
cos t
=
4 sin t cos t
cos t
= 4 sin t.
Since dy/dx = 4 sin t, we can now use the formula:
d
2
y
dx
2
=
dw/dt
dx/dt
where w =
dy
dx
d
2
y
dx
2
=
4 cos t
cos t
= 4.
Since d
2
y/dx
2
is always negative, our graph is concave down everywhere.
Using the identity y = cos(2t) = 1 2 sin
2
t, we can eliminate the parameter and write the original equation
as y = 1 2x
2
, which is a parabola that is concave down everywhere.
57. For 0 t 2, we get Figure 4.37.
x
y
1
Figure 4.37
SOLUTIONS to Review Problems for Chapter Four 93
Exercises
1. See Figure 4.38.
1 2 3 4 5 6
10
20
30
40
50
f(x)
x
Local max
Local min
Local and global max
Local and global min
Local max
Figure 4.38
5. (a) First we nd f
and f
:
f
(x) = e
x
sin x +e
x
cos x
f
(x) = e
x
sin x e
x
cos x
e
x
cos x e
x
sin x
= 2e
x
cos x
(b) The critical points are x = /4, 5/4, since f
(x) = 0 here.
(c) The inection points are x = /2, 3/2, since f
changes sign at these points.

(d) At the endpoints, f(0) = 0, f(2) = 0. So we have f(/4) = (e
/4
)(
2/2) as the global maximum; f(5/4) =

e
5/4
(
2/2) as the global minimum.

(e) See Figure 4.39.
2
3
2 2
|
|
conc. down
|
incr. decreasing
| concave up
|
|
increasing
conc. down |
|
x
Figure 4.39
9. As x , e
x
, so xe
x
. Thus limxxe
x
= .
As x ,
x
e
x
0, since e
x
grows much more quickly than x. Thus limxxe
x
= 0.
Using the product rule,
f
(x) = e
x
xe
x
= (1 x)e
x
,
which is zero when x = 1, negative when x > 1, and positive when x < 1. Thus f(1) = 1/e
1
= 1/e is a local
maximum.
Again, using the product rule,
f
(x) = e
x
e
x
+xe
x
= xe
x
2e
x
= (x 2)e
x
,
which is zero when x = 2, positive when x > 2, and negative when x < 2, giving an inection point at (2,
2
e
2
). With the
above, we have the following diagram:
concave up
y
> 0
x = 2
concave down
y
< 0
x = 1
y
> 0
increasing
y
< 0
decreasing
The graph of f is shown in Figure 4.40.
1 2
x
f(x) = xe
x
Figure 4.40
and f(x) has one global maximum at 1/e and no local or global minima.
13. Since g(t) is always decreasing for t 0, we expect it to a global maximum at t = 0 but no global minimum. At t = 0,
we have g(0) = 1, and as t , we have g(t) 0.
Alternatively, rewriting as g(t) = (t
3
+ 1)
1
and differentiating using the chain rule gives
g
(t) = (t
3
+ 1)
2
3t
2
.
Since 3t
2
= 0 when t = 0, there is a critical point at t = 0, and g decreases for all t > 0. See Figure 4.41.
1
g(t) =
1
t
3
+ 1
x
Figure 4.41
17. lim
x+
f(x) = +, and lim
x0
+
f(x) = +.
Hence, x = 0 is a vertical asymptote.
f
(x) = 1
2
x
=
x 2
x
, so x = 2 is the only critical point.
f
(x) =
2
x
2
, which can never be zero. So there are no inection points.
x 2
f
0 +
f
+ + +
f
2
f(x) = x ln x
x
Thus, f(2) is a local and global minimum.
21. We see from the parametric equations that the particle moves along a line. It sufces to plot two points: at t = 0, the
particle is at point (4, 1), and at t = 1, the particle is at point (2, 5). Since x decreases as t increases, the motion is right
to left and the curve is shown in Figure 4.42.
2 2 4 6
2
2
4
6
t = 0
x
y
Figure 4.42
Alternately, we can solve the rst equation for t, giving t = (x4)/2, and substitute this into the second equation
to get
y = 4
_
(x 4)
2
_
+ 1 = 2x + 9.
The line is y = 2x + 9.
25. We have
dM
dt
= (3x
2
+ 0.4x
3
)
dx
dt
.
If x = 5, then
dM
dt
= [3(5
2
) + 0.4(5
3
)](0.02) = 2.5 gm/hr.
Problems
29. The local maxima and minima of f correspond to places where f
is zero and changes sign or, possibly, to the endpoints

of intervals in the domain of f. The points at which f changes concavity correspond to local maxima and minima of f
.
The change of sign of f
, from positive to negative corresponds to a maximum of f and change of sign of f
from negative
to positive corresponds to a minimum of f.
33. Since the x
3
term has coefcient of 1, the cubic polynomial is of the form y = x
3
+ax
2
+bx +c. We now nd a, b, and
c. Differentiating gives
dy
dx
= 3x
2
+ 2ax +b.
The derivative is 0 at local maxima and minima, so
dy
dx
x=1
= 3(1)
2
+ 2a(1) +b = 3 + 2a +b = 0
dy
dx
x=3
= 3(3)
2
+ 2a(3) +b = 27 + 6a +b = 0
Subtracting the rst equation from the second and solving for a and b gives
24 + 4a = 0 so a = 6
b = 3 2(6) = 9.
Since the y-intercept is 5, the cubic is
y = x
3
6x
2
+ 9x + 5.
Since the coefcient of x
3
is positive, x = 1 is the maximum and x = 3 is the minimum. See Figure 4.43. To conrm
that x = 1 gives a maximum and x = 3 gives a minimum, we calculate
d
2
y
dx
2
= 6x + 2a = 6x 12.
At x = 1,
d
2
y
dx
2
= 6 < 0, so we have a maximum.
At x = 3,
d
2
y
dx
2
= 6 > 0, so we have a minimum.
1 3
5
x
y
Figure 4.43: Graph of y = x
3
6x
2
+ 9x + 5
37. First notice that since this function approaches 0 as x approaches either plus or minus innity, any local extrema that we
nd are also global extrema.
Differentiating y = axe
bx
2
gives
dy
dx
= ae
bx
2
2abx
2
e
bx
2
= ae
bx
2
(1 2bx
2
).
Since we have a critical points at x = 1 and x = 1, we know 1 2b = 0, so b = 1/2.
The global maximum is 2 at x = 1, so we have 2 = ae
1/2
which gives a = 2e
1/2
. Notice that this value of a also
gives the global minimum at x = 1.
Thus,
y = 2xe
(
1x
2
2
)
.
41. We want to maximize the volume V = x
2
h of the box, shown in Figure 4.44. The box has 5 faces: the bottom, which has
area x
2
and the four sides, each of which has area xh. Thus 8 = x
2
+ 4xh, so
h =
8 x
2
4x
.
Substituting this expression in for h in the formula for V gives
V = x
2
8 x
2
4x
=
1
4
(8x x
3
).
dV
dx
=
1
4
(8 3x
2
).
To maximize V we look for critical points, so we solve 0 = (8 3x
2
)/4, getting x =
_
8/3. We discard the negative
solution, since x is a positive length. Then we can nd
h =
8 x
2
4x
=
8
8
3
4
_
8
3
=
16
3
4
_
8
3
=
4
3
2
_
2
3
=
_
2
3
.
Thus x =
_
8/3 cm and h =
_
2/3 cm.
We can check that this critical point is a maximum of V by checking the sign of
d
2
V
dx
2
=
3
2
x,
which is negative when x > 0. So V is concave down at the critical point and therefore x =
_
8/3 gives a maximum
value of V .
x
x
h
Figure 4.44
45. Let f(x) = xsin x. Then f
(x) = xcos x + sin x.

f
(x) = 0 when x = 0, x 2, and x 5. The latter two estimates we can get from the graph of f
(x).
Zooming in (or using some other approximation method), we can nd the zeros of f
(x) with more precision. They

are (approximately) 0, 2.029, and 4.913. We check the endpoints and critical points for the global maximumand minimum.
f(0) = 0, f(2) = 0,
f(2.029) 1.8197, f(4.914) 4.814.
Thus for 0 x 2, 4.81 f(x) 1.82.
49. (a) We set the derivative equal to zero and solve for t to nd critical points. Using the product rule, we have:
f
(t) = (at
2
)(e
bt
(b)) + (2at)e
bt
= 0
ate
bt
(bt + 2) = 0
t = 0 or t =
2
b
.
There are two critical points: t = 0 and t = 2/b.
(b) Since we want a critical point at t = 5, we substitute and solve for b:
5 = 2/b
b =
2
5
= 0.4.
To nd the value of a, we use the fact that f(5) = 12, so we have:
a(5
2
)e
0.4(5)
= 12
a 25e
2
= 12
a =
12e
2
25
= 3.547.
(c) To show that f(t) has a local minimum at t = 0 and a local maximum at t = 5, we can use the rst derivative test or
the second derivative test. Using the rst derivative test, we evaluate f
at values on either side of t = 0 and t = 5.

Since f
(t) = 3.547te
0.4t
(0.4t + 2), we have
f
(1) = 3.547e
0.4
(2.4) = 12.700 < 0
and
f
(1) = 3.547e
0.4
(1.6) = 3.804 > 0,
and
f
(6) = 3.547(6)e
2.4
(0.4) = 0.772 < 0.
The function f is decreasing to the left of t = 0, increasing between t = 0 and t = 5, and decreasing to the right of
t = 5. Therefore, f(t) has a local minimum at t = 0 and a local maximum at t = 5. See Figure 4.45.
5
12
f(t)
t
Figure 4.45
53. Let V be the volume of the ice, so that V = 3(r
2
1
2
). Now,
dV
dt
= 6r
dr
dt
.
Thus for r = 1.5, we have
dV
dt
= 6(1.5)(0.03) = 0.848 cm
3
/hr.
57. (a) The business must reorder often enough to keep pace with sales. If reordering is done every t months, then,
Quantity sold in t months = Quantity reordered in each batch
rt = q
t =
q
r
months.
(b) The amount spent on each order is a+bq, which is spent every q/r months. To nd the monthly expenditures, divide
by q/r. Thus, on average,
Amount spent on ordering per month =
a +bq
q/r
=
ra
q
+rb dollars.
(c) The monthly cost of storage is kq/2 dollars, so
C = Ordering costs + Storage costs
C =
ra
q
+rb +
kq
2
dollars.
(d) The optimal batch size minimizes C, so
dC
dq
=
ra
q
2
+
k
2
= 0
ra
q
2
=
k
2
q
2
=
2ra
k
so
q =
_
2ra
k
items per order.
61. Since the volume is xed at 200 ml (i.e. 200 cm
3
), we can solve the volume expression for h in terms of r to get (with h
and r in centimeters)
h =
200 3
7r
2
.
Using this expression in the surface area formula we arrive at
S = 3r
_
r
2
+
_
600
7r
2
_
2
By plotting S(r) we see that there is a minimum value near r = 2.7 cm.
65. (a) The length of the piece of wire made into a circle is x cm, so the length of the piece made into a square is (Lx) cm.
See Figure 4.46.
x L x
r
Wire
Circle:
Perimeter x
Square:
Perimeter L x
s
Figure 4.46
The circumference of the circle is x, so its radius, r cm, is given by
r =
x
2
cm.
The perimeter of the square is (L x), so the side length, s cm, is given by
s =
L x
4
cm.
Thus, the sum of areas is given by
A = r
2
+s
2
=
_
x
2
_
2
+
_
L x
4
_
2
=
x
2
4
+
(L x)
2
16
, for 0 x L.
Setting dA/dx = 0 to nd the critical points gives
dA
dx
=
x
2

(L x)
8
= 0
8x = 2L 2x
(8 + 2)x = 2L
x =
2L
8 + 2
=
L
4 +
0.44L.
To nd the maxima and minima, we substitute the critical point and the endpoints, x = 0 and x = L, into the area
function.
For x = 0, we have A =
L
2
16
.
For x =
L
4 +
, we have L x = L
L
4 +
=
4L
4 +
. Then
A =

2
L
2
4(4 +)
2
+
1
16
_
4L
4 +
_
2
=
L
2
4(4 +)
2
+
L
2
(4 +)
2
=
L
2
+ 4L
2
4(4 +)
2
=
L
2
4(4 +)
=
L
2
16 + 4
.
For x = L, we have A =
L
2
4
.
Thus, x =
L
4 +
gives the minimum value of A =
L
2
16 + 4
.
Since 4 < 16, we see that x = L gives the maximum value of A =
L
2
4
.
This corresponds to the situation in which we do not cut the wire at all and use the single piece to make a circle.
(b) At the maximum, x = L, so
Length of wire in square
Length of wire in circle
=
0
L
= 0.
Area of square
Area of circle
=
0
L
2
/4
= 0.
At the minimum, x =
L
4 +
, so L x = L
L
4 +
=
4L
4 +
.
=
4L/(4 +)
L/(4 +)
=
4
.
Area of square
Area of circle
=
L
2
/(4 +)
2
L
2
/(4(4 +)
2
)
=
4
.
(c) For a general value of x,
=
L x
x
.
Area of square
Area of circle
=
(L x)
2
/16
x
2
/(4)
=

4

(L x)
2
x
2
.
If the ratios are equal, we have
L x
x
=

4

(L x)
2
x
2
.
So either L x = 0, giving x = L, or we can cancel (L x) and multiply through by 4x
2
, giving
4x = (L x)
x =
L
4 +
.
Thus, the two values of x found in part (a) are the only values of x in 0 x L making the ratios in part (b) equal.
(The ratios are not dened if x = 0.)
69. Evaluating the limits in the numerator and the denominator we get 0/e
0
= 0/1 = 0, so this is not an indeterminate form.
lHopitals rule does not apply.
73. If f(x) = 1 cosh(5x) and g(x) = x
2
, then f(0) = g(0) = 0, so we use lHopitals Rule:
lim
x0
1 cosh 5x
x
2
= lim
x0
5 sinh 5x
2x
= lim
x0
25 cosh 5x
2
=
25
2
.
77. The radius r is related to the volume by the formula V =
4
3
r
3
. By implicit differentiation, we have
dV
dt
=
4
3
3r
2
dr
dt
= 4r
2
dr
dt
.
The surface area of a sphere is 4r
2
, so we have
dV
dt
= s
dr
dt
,
but since
dV
dt
=
1
3
s was given, we have
dr
dt
=
1
3
.
81. We want to nd dP/dV . Solving PV = k for P gives
P = k/V
so,
dP
dV
=
k
V
2
.
85. (a) A CAS gives
d
dx
arcsinh x =
1
1 +x
2
(b) Differentiating both sides of sinh(arcsinh x) = x, we get
cosh(arcsinh x)
d
dx
(arcsinh x) = 1
d
dx
(arcsinh x) =
1
cosh(arcsinh x)
.
Since cosh
2
x sinh
2
x = 1, cosh x =
_
1 + sinh
2
x. Furthermore, since cosh x > 0 for all x, we take
the positive square root, so cosh x =
_
1 + sinh
2
x. Therefore, cosh(arcsinh x) =
_
1 + (sinh(arcsinh x))
2
=
1 +x
2
. Thus
d
dx
arcsinh x =
1
1 +x
2
.
89. (a) E B A
C
D
E '
x
E '

3 x
u
1m
Figure 4.47
We want to maximize the sum of the lengths EC and CD in Figure 4.47. Let x be the distance AE. Then x can
be between 0 and 1, the length of the left rope. By the Pythagorean theorem,
EC =
_
1 x
2
.
The length of the rope from B to C can also be found by the Pythagorean theorem:
BC =
_
EC
2
+EB
2
=
_
1 x
2
+ (
3 x)
2
=
_
4 2
3x.
Since the entire rope from B to D has length 3 m, the length from C to D is
CD = 3
_
4 2
3x.
The distance we want to maximize is
f(x) = EC +CD =
_
1 x
2
+ 3
_
4 2
3x, for 0 x 1.
f
(x) =
2x
2
1 x
2
3
2
_
4 2
3x
.
Setting f
(x) = 0 gives the cubic equation

2
3x
3
7x
2
+ 3 = 0.
Using a computer algebra system to solve the equation gives three roots: x = 1/
3, x =
3/2, x =
3. We
discard the negative root. Since x cannot be larger than 1 meter (the length of the left rope), the only critical point of
interest is x =
3/2, that is, halfway between A and B.

To nd the global maximum, we calculate the distance of the weight from the ceiling at the critical point and at
the endpoints:
f(0) =
1 + 3
4 = 2
f
_
3
2
_
=
_
1
3
4
+ 3
_
4 2
3
2
= 2.5
f(1) =
0 + 3
_
4 2
3 = 4
3 = 2.27.
Thus, the weight is at the maximum distance from the ceiling when x =
3/2; that is, the weight comes to rest at a

point halfway between points A and B.
(b) No, the equilibrium position depends on the length of the rope. For example, suppose that the left-hand rope was 1
cm long. Then there is no way for the pulley at its end to move to a point halfway between the anchor points.
1. True. Since the domain of f is all real numbers, all local minima occur at critical points.
5. False. For example, if f(x) = x
3
, then f
(0) = 0, but f(x) does not have either a local maximum or a local minimum at
x = 0.
9. Let f(x) = ax
2
, with a = 0. Then f
(x) = 2ax, so f has a critical point only at x = 0.

13. False. For example, if f(x) = x
3
, then f
(0) = 0, so x = 0 is a critical point, but x = 0 is neither a local maximum nor

a local minimum.
17. True. If the maximum is not at an endpoint, then it must be at critical point of f. But x = 0 is the only critical point of
f(x) = x
2
and it gives a minimum, not a maximum.
21. False. The circumference A and radius r are related by A = r
2
, so dA/dt = 2rdr/dt. Thus dA/dt depends on r and
since r is not constant, neither is dA/dt.
25. f(x) = x
2
+ 1 is positive for all x and concave up.
29. This is impossible. Since f
exists, so must f
, which means that f is differentiable and hence continuous. If f(x) were

positive for some values of x and negative for other values, then by the Intermediate Value Theorem, f(x) would have
to be zero somewhere, but this is impossible since f(x)f
(x) < 0 for all x. Thus either f(x) > 0 for all values of x,
in which case f
(x) < 0 for all values of x, that is f is concave down. But this is impossible by Problem 26. Or else
f(x) < 0 for all x, in which case f
(x) > 0 for all x, that is f is concave up. But this is impossible by Problem 28.
5.1 SOLUTIONS 103
CHAPTER FIVE
Exercises
1. (a) (i) Since the velocity is increasing, for an upper estimate we use a right sum. Using n = 4, we have t = 3, so
Upper estimate = (37)(3) + (38)(3) + (40)(3) + (45)(3) = 480.
(ii) Using n = 2, we have t = 6, so
Upper estimate = (38)(6) + (45)(6) = 498.
(b) The answer using n = 4 is more accurate as it uses the values of v(t) when t = 3 and t = 9.
(c) Since the velocity is increasing, for a lower estimate we use a left sum. Using n = 4, we have t = 3, so
Lower estimate = (34)(3) + (37)(3) + (38)(3) + (40)(3) = 447.
5. (a) The velocity is always positive, so the particle is moving in the same direction throughout. However, the particle is
speeding up until shortly before t = 0, and slowing down thereafter.
(b) The distance traveled is represented by the area under the curve. Using whole grid squares, we can overestimate the
area as 3 + 3 + 3 + 3 + 2 + 1 = 15, and we can underestimate the area as 1 + 2 + 2 + 1 + 0 + 0 = 6.
9. Using t = 0.2, our upper estimate is
1
1 + 0
(0.2) +
1
1 + 0.2
(0.2) +
1
1 + 0.4
(0.2) +
1
1 + 0.6
(0.2) +
1
1 + 0.8
(0.2) 0.75.
The lower estimate is
1
1 + 0.2
(0.2) +
1
1 + 0.4
(0.2) +
1
1 + 0.6
(0.2) +
1
1 + 0.8
(0.2)
1
1 + 1
(0.2) 0.65.
Since v is a decreasing function, the bug has crawled more than 0.65 meters, but less than 0.75 meters. We average the
two to get a better estimate:
0.65 + 0.75
2
= 0.70 meters.
13. From t = 0 to t = 5 the velocity is positive so the change in position is to the right. The area under the velocity graph
gives the distance traveled. The region is a triangle, and so has area (1/2)bh = (1/2)5 10 = 25. Thus the change in
position is 25 cm to the right.
Problems
17. Since f is increasing, the right-hand sum is the upper estimate and the left-hand sum is the lower estimate. We have
f(a) = 13, f(b) = 23 and t = (b a)/n = 2/100. Thus,
|Difference in estimates| = |f(b) f(a)|t
= |23 13|
1
50
=
1
5
.
104 Chapter Five /SOLUTIONS
21. The change in position is calculated from the area between the velocity graph and the t-axis, with the region below the
axis corresponding to negatives velocities and counting negatively.
Figure 5.1 shows the graph of f(t). From t = 0 to t = 3 the velocity is positive. The region under the graph of f(t)
is a triangle with height 6 cm/sec and base 3 seconds. Thus, from t = 0 to t = 3, the particle moves
Distance moved to right =
1
2
3 6 = 9 centimeters.
From t = 3 to t = 4, the velocity is negative. The region between the graph of f(t) and the t-axis is a triangle with height
2 cm/sec and base 1 second, so in this interval the particle moves
Distance moved to left =
1
2
1 2 = 1 centimeter.
Thus, the total change in position is 9 1 = 8 centimeters to the right.
3
4
2
6
Motion to right: 9 cm
Motion to left: 1 cm
f(t)
t
y
Figure 5.1
25. The graph of her velocity against time is a straight line from 0 mph to 60 mph; see Figure 5.2. Since the distance traveled
is the area under the curve, we have
Shaded area =
1
2
t 60 = 10 miles
Solving for t gives
t =
1
3
hr = 20 minutes .
60
velocity (mph)
time (hr)
Area = 10
'
t
Figure 5.2
Exercises
1. (a) Left-hand sum. Right-hand sum would be smaller.
(b) We have a = 0, b = 2, n = 6, x =
2
6
=
1
3
.
5.2 SOLUTIONS 105
5. With x = 5, we have
Left-hand sum = 5(0 + 100 + 200 + 100 + 200 + 250 + 275) = 5625,
Right-hand sum = 5(100 + 200 + 100 + 200 + 250 + 275 + 300) = 7125.
The average of these two sums is our best guess for the value of the integral;
_
20
15
f(x) dx
5625 + 7125
2
= 6375.
9. With x = 3, we have
Left-hand sum = 3(32 + 22 + 15 + 11) = 240,
Right-hand sum = 3(22 + 15 + 11 + 9) = 171.
The average of these two sums is our best guess for the value of the integral;
_
12
0
f(x) dx
240 + 171
2
= 205.5.
13. We use a calculator or computer to see that
_
1
1
e
x
2
dx = 1.4936.
17. A graph of y = ln x shows that this function is non-negative on the interval x = 1 to x = 4. Thus,
Area =
_
4
1
ln xdx = 2.545.
The integral was evaluated on a calculator.
21. The graph of y = x
4
8 has intercepts x =
4
8. See Figure 5.3. Since the region is below the x-axis, the integral is
negative, so
Area =
_ 4
8
(x
4
8) dx = 21.527.
8
4
8
y = x
4
8
x
y
Figure 5.3
Problems
25. Left-hand sum gives: 1
2
(1/4) + (1.25)
2
(1/4) + (1.5)
2
(1/4) + (1.75)
2
(1/4) = 1.96875.
Right-hand sum gives: (1.25)
2
(1/4) + (1.5)
2
(1/4) + (1.75)
2
(1/4) + (2)
2
(1/4) = 2.71875.
We estimate the value of the integral by taking the average of these two sums, which is 2.34375. Since x
2
is mono-
tonic on 1 x 2, the true value of the integral lies between 1.96875 and 2.71875. Thus the most our estimate could
be off is 0.375. We expect it to be much closer. (And it isthe true value of the integral is 7/3 2.333.)
29. Looking at the graph of e
x
sin x for 0 x 2 in Figure 5.4, we see that the area, A1, below the curve for 0 x
is much greater than the area, A2, above the curve for x 2. Thus, the integral is
_
2
0
e
x
sin xdx = A1 A2 > 0.
2
x
A
1
A
2
c
e
x
sin x
Figure 5.4
33. (a)
_
0
3
f(x) dx = 2.
(b)
_
4
3
f(x) dx =
_
0
3
f(x) dx +
_
3
0
f(x) dx +
_
4
3
f(x) dx = 2 + 2
A
2
=
A
2
.
37. (a) If the interval 1 t 2 is divided into n equal subintervals of length t = 1/n, the subintervals are given by
1 t 1 +
1
n
, 1 +
1
n
t 1 +
2
n
, . . . , 1 +
n 1
n
t 2.
The left-hand sum is given by
Left sum =
n1
r=0
f
_
1 +
r
n
_
1
n
=
n1
r=0
1
1 + r/n

1
n
=
n1
r=0
1
n + r
and the right-hand sum is given by
Right sum =
n
r=1
f
_
1 +
r
n
_
1
n
=
n
r=1
1
n + r
.
Since f(t) = 1/t is decreasing in the interval 1 t 2, we know that the right-hand sum is less than
_
2
1
1/t dt and
the left-hand sum is larger than this integral. Thus we have
n
r=1
1
n + r
<
_
t
1
1
t
dt <
n1
r=0
1
n + r
.
(b) Subtracting the sums gives
n1
r=0
1
n + r

n
r=1
1
n + r
=
1
n

1
2n
=
1
2n
.
(c) Here we need to nd n such that
1
2n
5 10
6
, so n
1
10
10
6
= 10
5
.
5.3 SOLUTIONS 107
Exercises
1. The units of measurement are dollars.
5. The integral
_
6
0
a(t) dt represents the change in velocity between times t = 0 and t = 6 seconds; it is measured in km/hr.
9. Average value =
1
2 0
_
2
0
(1 + t) dt =
1
2
(4) = 2.
13. Since the average value is given by
Average value =
1
b a
_
b
a
f(x) dx,
the units for dx inside the integral are canceled by the units for 1/(b a) outside the integral, leaving only the units
for f(x). This is as it should be, since the average value of f should be measured in the same units as f(x).
Problems
17. (a) Using rectangles under the curve, we get
Acres defaced (1)(0.2 + 0.4 + 1 + 2) = 3.6 acres.
(b) Using rectangles above the curve, we get
Acres defaced (1)(0.4 + 1 + 2 + 3.5) = 6.9 acres.
(c) The number of acres defaced is between 3.6 and 6.9, so we estimate the average, 5.25 acres.
21. (a) An overestimate is 7 tons. An underestimate is 5 tons.
(b) An overestimate is 7 +8 +10 +13 +16 +20 = 74 tons. An underestimate is 5 +7 +8 +10 +13 +16 = 59 tons.
(c) If measurements are made every t months, then the error is |f(6) f(0)| t. So for this to be less than 1 ton, we
need (20 5) t < 1, or t < 1/15. So measurements every 2 days or so will guarantee an error in over- and
underestimates of less than 1 ton.
25. The area under the curve represents the number of cubic feet of storage times the number of days the storage was used.
This area is given by
Area under graph = Area of rectangle + Area of triangle
= 30 10,000 +
1
2
30(30,000 10,000)
= 600,000.
Since the warehouse charges $5 for every 10 cubic feet of storage used for a day, the company will have to pay
(5)(60,000) = $300,000.
29. We know that the the integral of F, and therefore the work, can be obtained by computing the areas in Figure 5.5.
4 8 10
14 16
2
1
1
2
A
1
A
2
A
3
A
4
x (meters)
F (newtons)
Figure 5.5
W =
_
16
0
F(x) dx = Area above x-axis Area below x-axis
= A
1
+ A
2
+ A
3
A
4
=
1
2
4 2 + 4 2 +
1
2
2 2
1
2
2 2
= 12 newton meters.
33. (a) Average value of f =
1
5
_
5
0
f(x) dx.
(b) Average value of |f| =
1
5
_
5
0
|f(x)| dx =
1
5
(
_
2
0
f(x) dx
_
5
2
f(x) dx).
37. Suppose F(t) represents the total quantity of water in the water tower at time t, where t is in days since April 1. Then the
graph shown in the problem is a graph of F
(t). By the Fundamental Theorem,

F(30) F(0) =
_
30
0
F
(t)dt.
We can calculate the change in the quantity of water by calculating the area under the curve. If each box represents about
300 liters, there is about one box, or 300 liters, from t = 0 to t = 12, and 6 boxes, or about +1800 liters, from t = 12
to t = 30. Thus
_
30
0
F
(t)dt = 1800 300 = 1500,

so the nal amount of water is given by
F(30) = F(0) +
_
30
0
F
(t)dt = 12,000 + 1500 = 13,500 liters.

41. (a) Since t = 0 to t = 31 covers January:
Average number of
daylight hours in January
=
1
31
_
31
0
[12 + 2.4 sin(0.0172(t 80))] dt.
Using left and right sums with n = 100 gives
Average
306
31
9.9 hours.
(b) Assuming it is not a leap year, the last day of May is t = 151(= 31 + 28 + 31 + 30 + 31) and the last day of June
is t = 181(= 151 + 30). Again nding the integral numerically:
Average number of
daylight hours in June
=
1
30
_
181
151
[12 + 2.4 sin(0.0172(t 80))] dt
431
30
14.4 hours.
(c)
Average for whole year =
1
365
_
365
0
[12 + 2.4 sin(0.0172(t 80))] dt
4381
365
12.0 hours.
(d) The average over the whole year should be 12 hours, as computed in (c). Since Madrid is in the northern hemisphere,
the average for a winter month, such as January, should be less than 12 hours (it is 9.9 hours) and the average for a
summer month, such as June, should be more than 12 hours (it is 14.4 hours).
5.4 SOLUTIONS 109
Exercises
1. (a) A graph of f
(x) = sin(x
2
) is shown in Figure 5.6. Since the derivative f
(x) is positive between x = 0 and x = 1,

the change in f(x) is positive, so f(1) is larger than f(0). Between x = 2 and x = 2.5, we see that f
(x) is negative,
so the change in f(x) is negative; thus, f(2) is greater than f(2.5).
1 2 3
1
1
x
Figure 5.6: Graph of f
(x) = sin(x
2
)
(b) The change in f(x) between x = 0 and x = 1 is given by the Fundamental Theorem of Calculus:
f(1) f(0) =
_
1
0
sin(x
2
)dx = 0.310.
Since f(0) = 2, we have
f(1) = 2 + 0.310 = 2.310.
Similarly, since
f(2) f(0) =
_
2
0
sin(x
2
)dx = 0.805,
we have
f(2) = 2 + 0.805 = 2.805.
Since
f(3) f(0) =
_
3
0
sin(x
2
)dx = 0.774,
we have
f(3) = 2 + 0.774 = 2.774.
The results are shown in the table.
x 0 1 2 3
f(x) 2 2.310 2.805 2.774
5. The graph of y = 5 ln(2x) is above the line y = 3 for 3 x 5. See Figure 5.7. Therefore
Area =
_
5
3
(5 ln(2x) 3) dx = 14.688.
3 5
y = 5 ln(2x)
y = 3
x
y
Figure 5.7
9. The graph of y = cos t is above the graph of y = sin t for 0 t /4 and y = cos t is below y = sin t for
/4 < t < . See Figure 5.8. Therefore, we nd the area in two pieces:
Area =
_
/4
0
(cos t sin t) dt +
_

/4
(sin t cos t) dt = 2.828.
/4

y = cos x
y = sin x
x
y
Figure 5.8
13. We have f(t) = F
(t) = 6t + 4, so by the Fundamental Theorem of Calculus,

_
5
2
(6t + 4) dt = F(5) F(2) = 95 20 = 75.
17. We have f(t) = F
(t) = 1/(cos
2
t), so by the Fundamental Theorem of Calculus,
_

0
1/(cos
2
t) dt = F() F(0) = 0 0 = 0.
Problems
21. Note that
_
b
a
g(x) dx =
_
b
a
g(t) dt. Thus, we have
_
b
a
(f(x) + g(x)) dx =
_
b
a
f(x) dx +
_
b
a
g(x) dx = 8 + 2 = 10.
25. We write
_
b
a
_
c1g(x) + (c2f(x))
2
_
dx =
_
b
a
_
c1g(x) + c
2
2
(f(x))
2
_
dx
=
_
b
a
c1g(x) dx +
_
b
a
c
2
2
(f(x))
2
dx
= c1
_
b
a
g(x) dx + c
2
2
_
b
a
(f(x))
2
dx
= c1(2) + c
2
2
(12) = 2c1 + 12c
2
2
.
29. Since f is even,
_
2
0
f(x) dx = (1/2)6 = 3 and
_
5
0
f(x) dx = (1/2)14 = 7. Therefore
_
5
2
f(x) dx =
_
5
0
f(x) dx
_
2
0
f(x) dx = 7 3 = 4.
5.4 SOLUTIONS 111
33. (a) 0, since the integrand is an odd function and the limits are symmetric around 0.
(b) 0, since the integrand is an odd function and the limits are symmetric around 0.
37. See Figure 5.9. Since
_
1
0
f(x) dx = A1 and
_
2
1
f(x) dx = A2 and
_
3
2
f(x) dx = A3, we know that
0 <
_
1
0
f(x) dx <
_
2
1
f(x) dx <
_
3
2
f(x) dx.
In addition,
_
2
0
f(x) dx = A1 A2, which is negative, but smaller in magnitude than
_
2
1
f(x) dx. Thus
_
2
1
f(x) dx <
_
2
0
f(x) dx < 0.
The area A3 lies inside a rectangle of height 20 and base 1, so A3 < 20. The area A2 lies inside a rectangle below the
x-axis of height 10 and width 1, so 10 < A2. Thus:
(viii) < (ii) < (iii) < (vi) < (i) < (v) < ( iv) < (vii).
1
2 3
10
10
f(x)
c
A
1
c
A
2
A
3
T
c
10
E '
1
E '
1
T
c
20
x
Figure 5.9
41. (a) Yes.
(b) No, because the sum of the left sums has 20 subdivisions. The result is the left sum approximation with 20 subdivi-
sions to
_
3
1
f(x) dx.
a
b
Slope=
f(b)f(a)
ba
f(x)
T
c
f(b) f(a)
x
Figure 5.10
49. By the given property,
_
a
a
f(x) dx =
_
a
a
f(x) dx, so 2
_
a
a
f(x) dx = 0. Thus
_
a
a
f(x) dx = 0.
Exercises
1. (a) Suppose f(t) is the ow rate in m
3
/hr at time t. We are only given two values of the ow rate, so in making our
estimates of the ow, we use one subinterval, with t = 3/1 = 3:
Left estimate = 3[f(6 am)] = 3 100 = 300 m
3
(an underestimate)
Right estimate = 3[f(9 am)] = 3 280 = 840 m
3
(an overestimate).
The best estimate is the average of these two estimates,
Best estimate =
Left + Right
2
=
300 + 840
2
= 570 m
3
.
(b) Since the ow rate is increasing throughout, the error, i.e., the difference between over- and under-estimates, is given
by
Error t [f(9 am) f(6 am)] = t[280 100] = 180t.
We wish to choose t so that the the error 180t 6, or t 6/180 = 1/30. So the ow rate gauge should be
read every 1/30 of an hour, or every 2 minutes.
5. We know that
_
5
3
f(x)dx = Area above the axis Area below the axis.
The area above the axis is about 3 boxes. Since each box has area (1)(5) = 5, the area above the axis is about (3)(5) = 15.
The area below the axis is about 11 boxes, giving an area of about (11)(5) = 55. We have
_
5
3
f(x)dx 15 55 = 40.
9. Since x intercepts are x = 0, , 2, . . .,
Area =
_

0
sin xdx = 2.
1
x
y
y = sin x
13. The graph of y = e
x
+e
2(x1)
has intercepts where e
x
= e
2(x1)
, or where x = 2(x 1), so x = 2. See Figure 5.11.
Since the region is below the x-axis, the integral is negative, so
Area =
_
2
0
e
x
+ e
2(x1)
dx = 2.762.
2
y = e
x
+ e
2(x1)
x
y
Figure 5.11
SOLUTIONS to Review Problems for Chapter Five 113
17. Distance traveled =
_
1.1
0
sin(t
2
) dt 0.40 miles.
Problems
21. (a) By the chain rule,
d
dx
_
1
2
sin
2
t
_
=
1
2
2 sin t cos t = sin t cos t.
(i) Using a calculator,
_
0.4
0.2
sin t cos t dt = 0.056
(ii) The Fundamental Theorem of Calculus tells us that the integral is
_
0.4
0.2
sin t cos t dt = F(0.4) F(0.2) =
1
2
_
sin
2
(0.4) sin
2
(0.2)
_
= 0.05609.
25. On the interval 2 x 5,
Average value
of f
=
1
5 2
_
5
2
f(x) dx = 4,
so
_
5
2
f(x) dx = 12.
Thus
_
5
2
(3f(x) + 2) dx = 3
_
5
2
f(x) dx + 2
_
5
2
1 dx = 3(12) + 2(5 2) = 42.
29. (a) Train A starts earlier than Train B, and stops later. At every moment Train A is going faster than Train B. Both trains
increase their speed at a constant rate through the rst half of their trip and slow down during the second half. Both
trains reach their maximum speed at the same time. The area under the velocity graph for Train A is larger than the
area under the velocity graph for Train B, meaning that Train A travels fartheras would be expected, given that its
speed is always higher than Bs.
(b) (i) The maximum velocity is read off the vertical axis. The graph for Train A appears to go about twice as high as
the graph for Train B; see Figure 5.12. So
Maximum velocity of Train A
Maximum velocity of Train B
=
vA
vB
2.
E '
t
B
E '
t
A
v
B
v
A
Train A
Train B
t (hr)
v (km/hr)
Figure 5.12
(ii) The time of travel is the horizontal distance between the start and stop times (the two t-intercepts). The horizontal
distance for Train A appears to be about twice the corresponding distance for Train B; see Figure 5.12. So
Time traveled by Train A
Time traveled by Train B
=
tA
tB
2.
(iii) The distance traveled by each train is given by the area under its graph. Since the area of triangle is
1
2
Base
Height, and since the base and height for Train A is approximately twice that for Train B, we have
Distance traveled by Train A
Distance traveled by Train B
=
1
2
vA tA
1
2
vB tB
2 2 = 4.
33. (a) At t = 20 minutes, she stops moving toward the lake (with v > 0) and starts to move away from the lake (with
v < 0). So at t = 20 minutes the cyclist turns around.
(b) The cyclist is going the fastest when v has the greatest magnitude, either positive or negative. Looking at the graph,
we can see that this occurs at t = 40 minutes, when v = 25 and the cyclist is pedaling at 25 km/hr away from the
lake.
(c) From t = 0 to t = 20 minutes, the cyclist comes closer to the lake, since v > 0; thereafter, v < 0 so the cyclist
moves away from the lake. So at t = 20 minutes, the cyclist comes the closest to the lake. To nd out how close she
is, note that between t = 0 and t = 20 minutes the distance she has come closer is equal to the area under the graph
of v. Each box represents 5/6 of a kilometer, and there are about 2.5 boxes under the graph, giving a distance of about
2 km. Since she was originally 5 km away, she then is about 5 2 = 3 km from the lake.
(d) At t = 20 minutes she turns around , since v changes sign then. Since the area below the t-axis is greater than the
area above, the farthest she is from the lake is at t = 60 minutes. Between t = 20 and t = 60 minutes, the area under
the graph is about 10.8 km. (Since 13 boxes 5/6 = 10.8.) So at t = 60 she will be about 3 +10.8 = 13.8 km from
the lake.
37. All the integrals have positive values, since f 0. The integral in (ii) is about one-half the integral in (i), due to the
apparent symmetry of f. The integral in (iv) will be much larger than the integral in (i), since the two peaks of f
2
rise to
10,000. The integral in (iii) will be smaller than half of the integral in (i), since the peaks in f
1/2
will only rise to 10. So
_
2
0
(f(x))
1/2
dx <
_
1
0
f(x) dx <
_
2
0
f(x) dx <
_
2
0
(f(x))
2
dx.
41. (a) V, since the slope is constant.
(b) IV, since the net area under this curve is the most negative.
(c) III, since the area under the curve is largest.
(d) II, since the steepest ascent at t = 0 occurs on this curve.
(e) III, since average velocity is (total distance)/5, and III moves the largest total distance.
(f) I, since average acceleration is
1
5
_
5
0
v
(t) dt =
1
5
(v(5) v(0)), and in I, the velocity increases the most from start
(t = 0) to nish (t = 5).
45. (a) About 300 meter
3
/sec.
(b) About 250 meter
3
/sec.
(c) Looking at the graph, we can see that the 1996 ood reached its maximum just between March and April, for a high
of about 1250 meter
3
/sec. Similarly, the 1957 ood reached its maximum in mid-June, for a maximum ow rate of
3500 meter
3
/sec.
(d) The 1996 ood lasted about 1/3 of a month, or about 10 days. The 1957 ood lasted about 4 months.
(e) The area under the controlled ood graph is about 2/3 box. Each box represents 500 meter
3
/sec for one month. Since
1 month = 30
days
month
24
hours
day
60
minutes
hour
60
seconds
minute
= 2.592 10
6
2.6 10
6
seconds,
each box represents
Flow (500 meter
3
/sec) (2.6 10
6
sec) = 13 10
8
meter
3
of water.
So, for the articial ood,
Additional ow
2
3
13 10
8
= 8.7 10
8
meter
3
10
9
meter
3
.
(f) The 1957 ood released a volume of water represented by about 12 boxes above the 250 meter/sec baseline. Thus,
for the natural ood,
Additional ow 12 13 10
8
= 1.95 10
10
2 10
10
meter
3
.
So, the natural ood was nearly 20 times larger than the controlled ood and lasted much longer.
SOLUTIONS to Review Problems for Chapter Five 115
49. In (a), f
(1) is the slope of a tangent line at x = 1, which is negative. As for (c), the rate of change in f(x) is given by
f
(x), and the average value of this over 0 x a is

1
a 0
_
a
0
f
(x) dx =
f(a) f(0)
a 0
.
This is the slope of the line through the points (0, 1) and (a, 0), which is less negative that the tangent line at x = 1.
Therefore, (a) < (c) < 0. The quantity (b) is
__
a
0
f(x) dx
_
/a and (d) is
_
a
0
f(x) dx, which is the net area under the
graph of f (counting the area as negative for f below the x-axis). Since a > 1 and
_
a
0
f(x) dx > 0, we have 0 <(b)<(d).
Therefore
(a) < (c) < (b) < (d).
53. (a) Since the length of the interval of integration is 2 1 = 1, the width of each subdivision is t = 1/n. Thus the
endpoints of the subdivision are
t0 = 1, t1 = 1 + t = 1 +
1
n
, t2 = 1 + 2t = 1 +
2
n
, . . . ,
ti = 1 + it = 1 +
i
n
, . . . , tn1 = 1 + (n 1)t = 1 +
n 1
n
.
Thus, since the integrand is f(t) = t,
Left-hand sum =
n1
i=0
f(ti)t =
n1
i=0
tit =
n1
i=0
_
1 +
i
n
_
1
n
=
n1
i=0
n + i
n
2
.
(b) The CAS nds the formula for the Riemann sum
n1
i=0
n + i
n
2
=
(1+n) n
2
+ n
2
n
2
=
3
2

1
2n
.
(c) Taking the limit as n
lim
n
_
3
2

1
2n
_
= lim
n
3
2
lim
n
1
2n
=
3
2
+ 0 =
3
2
.
(d) The shape under the graph of y = t between t = 1 and t = 2 is a trapezoid of width 1, height 1 on the left and 2 on
the right. So its area is 1 (1 + 2)/2 = 3/2. This is the same answer we got by computing the denite integral.
57. (a) Different systems may give different answers. A typical answer is
_
c
a
x
1 + bx
2
dx =
ln
_
c
2
b + 1
a
2
b + 1
_
2b
.
Some CASs may not have the absolute values in the answer; since b > 0, the answer is correct without the absolute
values.
(b) Using the properties of logarithms, we can rewrite the answer to part (a) as
_
c
a
x
1 + bx
2
dx =
ln
c
2
b + 1
ln
a
2
b + 1
2b
=
ln
c
2
b + 1
2b

ln
a
2
b + 1
2b
.
If F(x) is an antiderivative of x/(1 + bx
2
), then the Fundamental Theorem of Calculus says that
_
c
a
x
1 + bx
2
dx = F(c) F(a).
Thus
F(c) F(a) =
ln
c
2
b + 1
2b

ln
a
2
b + 1
2b
.
This suggests that
F(x) =
ln
1 + bx
2
2b
.
(Since b > 0, we know
1 + bx
2
= 1 + bx
2
.) Taking the derivative conrms this:
d
dx
_
ln(1 + b x
2
)
2 b
_
=
x
1 + b x
2
.
1. False. The units of the integral are the product of the units for f(x) times the units for x.
5. False. The integral is the change in position from t = a to t = b. If the velocity changes sign in the interval, the total
distance traveled and the change in position will not be the same.
9. True, since
_
2
0
2f(x)dx = 2
_
2
0
f(x)dx.
13. False. Let f(x) = 7 and g(x) = 9 for all x.
Then
_
2
1
f(x) dx +
_
3
2
g(x) dx = 7 + 9 = 16, but
_
3
1
(f(x) + g(x)) dx =
_
3
1
16 dx = 32.
17. False. Let f(x) = x and g(x) = 5. Then
_
6
2
f(x) dx = 16 and
_
6
2
g(x) dx = 20, so
_
6
2
f(x) dx
_
6
2
g(x) dx, but
f(x) > g(x) for 5 < x < 6.
21. True. We have by the properties of integrals in Theorem 5.3,
_
9
1
f(x)dx =
_
4
1
f(x)dx +
_
9
4
f(x)dx.
Since (1/(4 1))
_
4
1
f(x)dx = A and (1/(9 4))
_
9
4
f(x)dx = B, we have
_
9
1
f(x)dx = 3A + 5B.
Dividing this equation through by 8, we get that the average value of f on the interval [1, 9] is (3/8)A + (5/8)B.
25. False. A counterexample is given by the functions f and g in Figure 5.13. The function f is decreasing, g is increasing,
and we have
_
2
1
f(x) dx =
_
2
1
g(x) dx,
because both integrals equal 1/2, the area of of the same sized triangle.
1 2
1
f(x)
x
1 2
1
g(x)
x
Figure 5.13
6.1 SOLUTIONS 117
CHAPTER SIX
Exercises
1. See Figure 6.1.
1
1
x
F(0) = 0
F(0) = 1
Figure 6.1
5. Since dP/dt is negative for t < 3 and positive for t > 3, we know that P is decreasing for t < 3 and increasing for
t > 3. Between each two integer values, the magnitude of the change is equal to the area between the graph dP/dt and
the t-axis. For example, between t = 0 and t = 1, we see that the change in P is 1. Since P = 2 at t = 0, we must
have P = 1 at t = 1. The other values are found similarly, and are shown in Table 6.1.
Table 6.1
t 1 2 3 4 5
P 1 0 1/2 0 1
Problems
9. See Figure 6.2. Note that since f(x1) = 0 and f
(x1) < 0, F(x1) is a local maximum; since f(x3) = 0 and f
(x3) > 0,
F(x3) is a local minimum. Also, since f
(x2) = 0 and f changes from decreasing to increasing about x = x2, F has an

inection point at x = x2.
x
1
x
2
x
3
F(x)
x
Figure 6.2
118 Chapter Six /SOLUTIONS
13. Between t = 0 and t = 1, the particle moves at 10 km/hr for 1 hour. Since it starts at x = 5, the particle is at x = 15
when t = 1. See Figure 6.3. The graph of distance is a straight line between t = 0 and t = 1 because the velocity is
constant then.
Between t = 1 and t = 2, the particle moves 10 km to the left, ending at x = 5. Between t = 2 and t = 3, it moves
10 km to the right again. See Figure 6.3.
1 2 3 4 5 6
5
10
15
t (hr)
x (km)
Figure 6.3
As an aside, note that the original velocity graph is not entirely realistic as it suggests the particle reverses direction
instantaneously at the end of each hour. In practice this means the reversal of direction occurs over a time interval that is
short in comparison to an hour.
17. Looking at the graph of g
in Figure 6.4, we see that the critical points of g occur when x = 15 and x = 40, since
g
(x) = 0 at these values. Inection points of g occur when x = 10 and x = 20, because g
(x) has a local maximum or

minimum at these values. Knowing these four key points, we sketch the graph of g(x) in Figure 6.5.
We start at x = 0, where g(0) = 50. Since g
is negative on the interval [0, 10], the value of g(x) is decreasing there.
At x = 10 we have
g(10) = g(0) +
_
10
0
g
(x) dx
= 50 (area of shaded trapezoid T1)
= 50
_
10 + 20
2
10
_
= 100.
Similarly,
g(15) = g(10) +
_
15
10
g
(x) dx
= 100 (area of triangle T2)
= 100
1
2
(5)(20) = 150.
Continuing,
g(20) = g(15) +
_
20
15
g
(x) dx = 150 +
1
2
(5)(10) = 125,
and
g(40) = g(20) +
_
40
20
g
(x) dx = 125 +
1
2
(20)(10) = 25.
We now nd concavity of g(x) in the intervals [0, 10], [10, 15], [15, 20], [20, 40] by checking whether g
(x) increases
or decreases in these same intervals. If g
(x) increases, then g(x) is concave up; if g
(x) decreases, then g(x) is concave

down. Thus we nally have the graph of g(x) in Figure 6.5.
6.2 SOLUTIONS 119
15 40
10
(10, 20)
(20, 10)
g
(x)
T
1
c
T
2
x
Figure 6.4
x
g(x) (0, 50)
(10, 100)
(15, 150)
(20, 125)
(40, 25)
Figure 6.5
21. (a) Critical points of F(x) are the zeros of f: x = 1 and x = 3.
(b) F(x) has a local minimum at x = 1 and a local maximum at x = 3.
(c) See Figure 6.6.
1 2 3 4
F(x)
x
Figure 6.6
Notice that the graph could also be above or below the x-axis at x = 3.
25. Both F(x) and G(x) have roots at x = 0 and x = 4. Both have a critical point (which is a local maximum) at x = 2.
However, since the area under g(x) between x = 0 and x = 2 is larger than the area under f(x) between x = 0 and
x = 2, the y-coordinate of G(x) at 2 will be larger than the y-coordinate of F(x) at 2. See below.
1 3 2 4
F(x)
G(x)
x
Exercises
1. 5x
5. sin t
9.
1
2z
2
13.
t
4
4

t
3
6

t
2
2
17. F(t) =
_
6t dt = 3t
2
+ C
21. F(z) =
_
(z + e
z
) dz =
z
2
2
+ e
z
+ C
25. P(t) =
_
(2 + sin t) dt = 2t cos t + C
29. f(x) = 3, so F(x) = 3x+C. F(0) = 0 implies that 3 0+C = 0, so C = 0. Thus F(x) = 3x is the only possibility.
33. f(x) = x
2
, so F(x) =
x
3
3
+C. F(0) = 0 implies that
0
3
3
+C = 0, so C = 0. Thus F(x) =
x
3
3
is the only possibility.
37.
_
5xdx =
5
2
x
2
+ C.
41.
_
(t
2
+
1
t
2
) dt =
t
3
3

1
t
+ C
45. 2t
2
+ 7t + C
49. cos t + C
53.
_
3
0
(x
2
+ 4x + 3) dx =
_
x
3
3
+ 2x
2
+ 3x
_
3
0
= (9 + 18 + 9) 0 = 36
57.
_
5
2
(x
3
x
2
) dx =
_
x
4
4

x
3
3
_
5
2
=
609
4
39 29.728.
61.
_
/4
0
(sin t + cos t) dt = (cos t + sin t)
/4
0
=
_
2
2
+
2
2
_
(1 + 0) = 1.
Problems
65. The graph crosses the x-axis where
7 8x + x
2
= 0
(x 7)(x 1) = 0;
so x = 1 and x = 7. See Figure 6.7. The parabola opens upward and the region is below the x-axis, so
Area =
_
7
1
(7 8x + x
2
) dx
=
_
7x 4x
2
+
x
3
3
_
7
1
= 36.
1 7
y = 7 8x + x
2
x
Figure 6.7
69. The graph is shown in Figure 6.8. Since cos sin for 0 /4, we have
Area =
_
/4
0
(cos sin ) d
= (sin + cos )
/4
0
=
1
2
+
1
2
1 =
2 1.
6.2 SOLUTIONS 121
4
c
y = cos
y = sin
Figure 6.8
73. The area under f(x) = 8x between x = 1 and x = b is given by
_
b
1
(8x)dx. Using the Fundamental Theorem to evaluate
the integral:
Area = 4x
2
b
1
= 4b
2
4.
Since the area is 192, we have
4b
2
4 = 192
4b
2
= 196
b
2
= 49
b = 7.
Since b is larger than 1, we have b = 7.
77. (a) The average value of f(t) = sin t over 0 t 2 is given by the formula
Average =
1
2 0
_
2
0
sin t dt
=
1
2
(cos t)
2
0
=
1
2
(cos 2 (cos 0)) = 0.
We can check this answer by looking at the graph of sin t in Figure 6.9. The area below the curve and above the
t-axis over the interval 0 t , A1, is the same as the area above the curve but below the t-axis over the interval
t 2, A2. When we take the integral of sin t over the entire interval 0 t 2, we get A1 A2 = 0.
2
A
1
A
2
t
Figure 6.9
(b) Since
_

0
sin t dt = cos t
0
= cos (cos 0) = (1) (1) = 2,
the average value of sin t on 0 t is given by
Average value =
1
_

0
sin t dt =
2
.
81. The rate at which water is entering the tank is
dV
dt
= 120 6t.
Thus, the total quantity of water in the tank at time t = 4 is
V =
_
4
0
(120 6t) dt.
Since an antiderivative to 120 6t is
120t 3t
2
,
we have
V =
_
4
0
(120 6t) dt = (120t 3t
2
)
4
0
= (120 4 6 4
2
) (120 0 3 0
2
)
= 384 ft
3
.
The radius is 5 feet, so if the height is h ft, the volume is V = 5
2
h = 25h. Thus, at time t = 4, we have V = 384, so
384 = 25h
h =
384
25
= 4.889 ft.
3 7
16
t
CCl
4
dumped
Figure 6.10
(b) 7 years, because t
2
14t + 49 = (t 7)
2
indicates that the rate of ow was zero after 7 years.
(c)
Area under the curve = 3(16) +
_
7
3
(t
2
14t + 49) dt
= 48 +
_
1
3
t
3
7t
2
+ 49t
_
7
3
= 48 +
343
3
343 + 343 9 + 63 147
=
208
3
= 69.333 cubic yards.
Exercises
1. y =
_
(x
3
+ 5) dx =
x
4
4
+ 5x + C
5. Since y = x +sin x , we differentiate to see that dy/dx = 1 +cos x, so y satises the differential equation. To show
that it also satises the initial condition, we check that y() = 0:
y = x + sin x
y() = + sin = 0.
6.3 SOLUTIONS 123
9. Integrating gives
_
dq
dz
dz =
_
(2 + sin z) dz = 2z cos z + C.
If q = 5 when z = 0, then 2(0) cos(0) + C = 5 so C = 6. Thus q = 2z cos z + 6.
Problems
13. (a) We nd F for each piece, 0 x 1 and 1 x 2.
For 0 x 1, we have f(x) = x + 1, so F is of the form
_
(x + 1) dx =
x
2
2
+ x + C.
Since we want F(1) = 1, we need C = 1/2. See Figure 6.11.
For 1 x 2, we have f(x) = x 1, so F is of the form
_
(x 1) dx =
x
2
2
x + C.
Again, since we want F(1) = 1, we have C = 3/2. See Figure 6.11.
(b) Evaluating
F(2) F(0) =
_
2
2
2
2 +
3
2
_
0
2
2
+ 0 +
1
2
_
= 1.
The region under the graph of f consists of two triangles, whose area is
Area =
1
2
+
1
2
= 1.
(c) The Fundamental Theorem of Calculus says
_
2
0
f(x) dx = F(2) F(0).
Since the value of the integral is just the area under the curve, we have shown this in part (b).
1 2
1
x
F(x)
Figure 6.11
17. (a) To nd the height of the balloon, we integrate its velocity with respect to time:
h(t) =
_
v(t) dt
=
_
(32t + 40) dt
= 32
t
2
2
+ 40t + C.
Since at t = 0, we have h = 30, we can solve for C to get C = 30, giving us a height of
h(t) = 16t
2
+ 40t + 30.
(b) To nd the average velocity between t = 1.5 and t = 3, we nd the total displacement and divide by time.
Average velocity =
h(3) h(1.5)
3 1.5
=
6 54
1.5
= 32 ft/sec.
The balloons average velocity is 32 ft/sec downward.
(c) First, we must nd the time when h(t) = 6. Solving the equation 16t
2
+ 40t + 30 = 6, we get
6 = 16t
2
+ 40t + 30
0 = 16t
2
+ 40t + 24
0 = 2t
2
5t 3
0 = (2t + 1)(t 3).
Thus, t = 1/2 or t = 3. Since t = 1/2 makes no physical sense, we use t = 3 to calculate the balloons
velocity. At t = 3, we have a velocity of v(3) = 32(3) + 40 = 56 ft/sec. So the balloons velocity is 56 ft/sec
downward at the time of impact.
21. Since the acceleration is constant, a graph of the velocity versus time looks like this:
30
t (sec)
v (mph)
200 mph
A
The distance traveled in 30 seconds, which is how long the runway must be, is equal to the area represented by A.
We have A =
1
2
(base)(height). First we convert the required velocity into miles per second.
200 mph =
200 miles
hour
_
1 hour
60 minutes
__
1 minute
60 seconds
_
=
200
3600
miles
second
=
1
18
miles/second.
Therefore A =
1
2
(30 sec)(200 mph) =
1
2
(30 sec)
_
1
18
miles/sec
_
=
5
6
miles.
25. (a) a(t) = 1.6, so v(t) = 1.6t + v0 = 1.6t, since the initial velocity is 0.
(b) s(t) = 0.8t
2
+ s0, where s0 is the rocks initial height.
Exercises
1. By the Fundamental Theorem, f(x) = F
(x). Since f is positive and increasing, F is increasing and concave up. Since
F(0) =
_
0
0
f(t)dt = 0, the graph of F must start from the origin. See Figure 6.12.
x
F(x)
Figure 6.12
6.4 SOLUTIONS 125
5. (1 + x)
200
.
9. arctan(x
2
).
13.
Table 6.2
x 0 0.5 1 1.5 2
I(x) 0 0.50 1.09 2.03 3.65
17. If f
(x) =
sin x
x
, then f(x) is of the form
f(x) = C +
_
x
a
sin t
t
dt.
Since f(1) = 5, we take a = 1 and C = 5, giving
f(x) = 5 +
_
x
1
sin t
t
dt.
Problems
21. Using the solution to Problem 20, we see that F(x) is increasing for x <

and F(x) is decreasing for x >
. Thus
F(x) has its maximum value when x =
.
By the Fundamental Theorem of Calculus,
F(
) F(0) =
_

0
f(t) dt =
_

0
sin(t
2
) dt.
Since F(0) = 0, calculating Riemann sums gives
F(
) =
_

0
sin(t
2
) dt = 0.895.
25. Since G
(x) = cos(x
2
) and G(0) = 3, we have
G(x) = G(0) +
_
x
0
cos(t
2
) dt = 3 +
_
x
0
cos(t
2
) dt.
Substituting x = 1 and evaluating the integral numerically gives
G(1) = 3 +
_
1
0
cos(t
2
) dt = 3.905.
29. If we let f(x) =
_
x
2
sin(t
2
) dt and g(x) = x
3
, using the chain rule gives
d
dx
_
x
3
2
sin(t
2
) dt = f
(g(x)) g
(x) = sin((x
3
)
2
) 3x
2
= 3x
2
sin(x
6
).
33. Since
_
3
cos x
e
t
2
dt =
_
cos x
3
e
t
2
dt, if we let f(x) =
_
x
3
e
t
2
dt and g(x) = cos x, using the chain rule gives
d
dx
_
3
cos x
e
t
2
dt =
d
dx
_
cos x
3
e
t
2
dt = f
(g(x)) g
(x) = e
(cos x)
2
(sin x) = sin xe
cos
2
x
.
37. (a) The denition of P gives
P(0) =
_
0
0
arctan(t
2
) dt = 0
and
P(x) =
_
x
0
arctan(t
2
) dt.
Changing the variable of integration by letting t = z gives
_
x
0
arctan(t
2
) dt =
_
x
0
arctan((z)
2
)(dz) =
_
x
0
arctan(z
2
) dz.
Thus P is an odd function.
(b) Using the Second Fundamental Theorem gives P
(x) = arctan(x
2
), which is greater than 0 for x = 0. Thus P is
increasing everywhere.
(c) Since
P
(x) =
2x
1 + x
4
,
we have P concave up if x > 0 and concave down if x < 0.
(d) See Figure 6.13.
5 5
7
7
P(x)
x
Figure 6.13
41. The derivative of s(t t0) is s
(t t0), so s
(t t0) = v(t). Substituting w = t t0, so that t = w + t0, we get

s
(w) = v(w + t0). Renaming w to t, we get s
(t) = v(t + t0).

45. If we let f(x) = erf(x) and g(x) =

x, then we are looking for
d
dx
[f(g(x))]. By the chain rule, this is the same as
g
(x)f
(g(x)). Since
f
(x) =
d
dx
_
2
_
x
0
e
t
2
dt
_
=
2
e
x
2
and g
(x) =
1
2
x
, we have
f
(g(x)) =
2
e
x
,
and so
d
dx
[erf(
x)] =
1
2
x
2
e
x
=
1
x
e
x
.
Exercises
1. (a) The object is thrown from an initial height of y = 1.5 meters.
(b) The velocity is obtained by differentiating, which gives v = 9.8t + 7 m/sec. The initial velocity is v = 7 m/sec
upward.
(c) The acceleration due to gravity is obtained by differentiating again, giving g = 9.8 m/sec
2
, or 9.8 m/sec
2
down-
ward.
SOLUTIONS to Review Problems for Chapter Six 127
Problems
5. In Problem 4 we used the equation 0 = 16t
2
+ 400 to learn that the object hits the ground after 5 seconds. In a more
general form this is the equation y =
g
2
t
2
+v0t +y0, and we know that v0 = 0, y0 = 400 ft. So the moment the object
hits the ground is given by 0 =
g
2
t
2
+ 400. In Problem 4 we used g = 32 ft/sec
2
, but in this case we want to nd a g
that results in the object hitting the ground after only 5/2 seconds. We put in 5/2 for t and solve for g:
0 =
g
2
(
5
2
)
2
+ 400, so g =
2(400)
(5/2)
2
= 128 ft/sec
2
.
9. (a) t =
s
1
2
vmax
, where t is the time it takes for an object to travel the distance s, starting from rest with uniform
acceleration a. vmax is the highest velocity the object reaches. Since its initial velocity is 0, the mean of its highest
velocity and initial velocity is
1
2
vmax.
(b) By Problem 8, s =
1
2
gt
2
, where g is the acceleration due to gravity, so it takes
_
200/32 = 5/2 seconds for the body
to hit the ground. Since v = gt, vmax = 32(
5
2
) = 80 ft/sec. Galileos statement predicts (100 ft)/(40 ft/sec) = 5/2
seconds, and so Galileos result is veried.
(c) If the acceleration is a constant a, then s =
1
2
at
2
, and vmax = at. Thus
s
1
2
vmax
=
1
2
at
2
1
2
at
= t.
Exercises
1. See Figure 6.14
1
x
F(0) = 0
F(0) = 1
Figure 6.14
5.
_
(2 + cos t) dt = 2t + sin t + C
9.
_
8
x
dx = 16x
1/2
+ C
13. tan x + C
17.
1
10
(x + 1)
10
+ C
21. 3 sin x + 7 cos x + C
25. F(x) =
_
1
x
2
dx =
1
x
+ C
29. F(x) =
_
5e
x
dx = 5e
x
+ C
33. We have F(x) =
x
4
4
+2x
3
4x+C. Since F(0) = 4, we have 4 = 0+C, so C = 4. So F(x) =
x
4
4
+2x
3
4x+4.
37. F(x) =
_
cos xdx = sin x + C. If F(0) = 4, then F(0) = 0 + C = 4 and thus C = 4. So F(x) = sin x + 4.
Problems
41.
_
3
0
x
2
dx =
x
3
3
3
0
= 9 0 = 9.
45. (a) See Figure 6.15. Since f(x) > 0 for 0 < x < 2 and f(x) < 0 for 2 < x < 5, we have
Area =
_
2
0
f(x) dx
_
5
2
f(x) dx
=
_
2
0
(x
3
7x
2
+ 10x) dx
_
5
2
(x
3
7x
2
+ 10x) dx
=
_
x
4
4

7x
3
3
+ 5x
2
_
2
0
_
x
4
4

7x
3
3
+ 5x
2
_
5
2
=
__
4
56
3
+ 20
_
(0 0 + 0)
_
__
625
4

875
3
+ 125
_
_
4
56
3
+ 20
__
=
253
12
.
2
5
x
Figure 6.15: Graph of f(x) = x
3
7x
2
+ 10x
(b) Calculating
_
5
0
f(x) dx gives
_
5
0
f(x) dx =
_
5
0
(x
3
7x
2
+ 10x) dx
=
_
x
4
4

7x
3
3
+ 5x
2
_
5
0
=
_
625
4

875
3
+ 125
_
(0 0 + 0)
=
125
12
.
This integral measures the difference between the area above the x-axis and the area below the x-axis. Since the
denite integral is negative, the graph of f(x) lies more below the x-axis than above it. Since the function crosses
the axis at x = 2,
_
5
0
f(x) dx =
_
2
0
f(x) dx +
_
5
2
f(x) dx =
16
3

63
4
=
125
12
,
whereas
Area =
_
2
0
f(x) dx
_
5
2
f(x) dx =
16
3
+
64
4
=
253
12
.
SOLUTIONS to Review Problems for Chapter Six 129
49. The graph of y = c(1 x
2
) has x-intercepts of x = 1. See Figure 6.16. Since it is symmetric about the y-axis, we have
Area =
_
1
1
c(1 x
2
) dx = 2c
_
1
0
(1 x
2
) dx
= 2c
_
x
x
3
3
_
1
0
=
4c
3
.
We want the area to be 1, so
4c
3
= 1, giving c =
3
4
.
1 1
c
y = c(1 x
2
)
x
y
Figure 6.16
53. Since A
(r) = C(r) and C(r) = 2r, we have

A
(r) = 2r.
Thus, we have, for some arbitrary constant K:
A(r) =
_
2r dr = 2
_
r dr = 2
r
2
2
+ K = r
2
+ K.
Since a circle of radius r = 0 has area = 0, we substitute to nd K:
0 = 0
2
+ K
K = 0.
Thus
A(r) = r
2
.
x
1
x
2
x
3
x
4
x
f(x)
c
Point of
inflection
Figure 6.17
61. We have
d
dx
_
x
2
arccos(t
7
) dt = arccos x
7
.
65. If we let f(x) =
_
x
5
cos(t
3
) dt and g(x) = e
x
, using the chain rule gives
d
dx
_
e
x
5
cos(t
3
) dt = f
(g(x)) g
(x) = cos((e
x
)
3
) e
x
= e
x
cos(e
3x
).
69. F(x) represents the net area between (sin t)/t and the t-axis from t =

2
to t = x, with area counted as negative for
(sin t)/t below the t-axis. As long as the integrand is positive F(x) is increasing. Therefore, the global maximum of
F(x) occurs at x = and is given by the area
A1 =
_

/2
sin t
t
dt.
At x = /2, F(x) = 0. Figure 6.18 shows that the area A1 is larger than the area A2. Thus F(x) > 0 for

2
< x
3
2
.
Therefore the global minimum is F(
2
) = 0.
3
2
1
1
t
y
A
1
A
2
T
T
y =
sin t
t
Figure 6.18
73. Let v be the velocity and s be the position of the particle at time t. We know that a = dv/dt, so acceleration is the slope
of the velocity graph. Similarly, velocity is the slope of the position graph. Graphs of v and s are shown in Figures 6.19
and 6.20, respectively.
1 2 3 4 5 6 7
t
v
Figure 6.19: Velocity against time
1 2 3 4 5 6 7
t
s
Figure 6.20: Position against time
77. (a) Using g = 32 ft/sec
2
, we have
t (sec) 0 1 2 3 4 5
v(t) (ft/sec) 80 48 16 16 48 80
(b) The object reaches its highest point when v = 0, which appears to be at t = 2.5 seconds. By symmetry, the object
should hit the ground again at t = 5 seconds.
(c) Left sum = 80(1) + 48(1) + 16(
1
2
) = 136 ft , which is an overestimate.
Right sum = 48(1) + 16(1) + (16)(
1
2
) = 56 ft , which is an underestimate.
Note that we used a smaller third rectangle of width 1/2 to end our sum at t = 2.5.
(d) We have v(t) = 80 32t, so antidifferentiation yields s(t) = 80t 16t
2
+ s0.
But s0 = 0, so s(t) = 80t 16t
2
.
At t = 2.5, s(t) = 100 ft., so 100 ft. is the highest point.
81. (a) We have x =
(b a)
n
and xi = a + i(x) = a + i
_
b a
n
_
, so, since f(xi) = xi
3
,
Riemann sum =
n
i=1
f(xi)x =
n
i=1
_
a + i
_
b a
n
_
_
3
_
b a
n
_
.
(b) A CAS gives
n
i=1
_
a +
i(b a)
n
_
3
(b a)
n
=
(a b)(a
3
(n 1)
2
+ (a
2
b + ab
2
)(n
2
1) + b
3
(n + 1)
3
)
4n
2
.
Taking the limit as n gives
lim
n
n
i=1
_
a + i
_
b a
n
_
_
3
_
b a
n
_
=
(a + b)(a b)(a
2
+ b
2
)
4
.
(c) The answer to part (b) simplies to
b
4
4

a
4
4
. Since
d
dx
_
x
4
4
_
= x
3
, the Fundamental Theorem of Calculus says
that
_
b
a
x
3
dx =
x
4
4
b
a
=
b
4
4

a
4
4
.
85. (a) A CAS gives
_
1
(x 1)(x 3)
dx =
1
2
(ln |x 3| ln |x 1|)
_
1
(x 1)(x 4)
dx =
1
3
(ln |x 4| ln |x 1|)
_
1
(x 1)(x + 3)
dx =
1
4
(ln |x + 3| ln |x 1|).
Although the absolute values are needed in the answer, some CASs may not include them.
(b) The three integrals in part (a) obey the rule
_
1
(x a)(x b)
dx =
1
b a
(ln |x b| ln |x a|).
(c) Checking the formula by calculating the derivative
d
dx
_
1
b a
(ln |x b| ln |x a|)
_
=
1
b a
_
1
x b

1
x a
_
=
1
b a
_
(x a) (x b)
(x a)(x b)
_
=
1
b a
_
b a
(x a)(x b)
_
=
1
(x a)(x b)
.
1. True. A function can have only one derivative.
5. False. Differentiating using the product and chain rules gives
d
dx
_
1
2x
e
x
2
_
=
1
2x
2
e
x
2
+ e
x
2
.
9. True. If y = F(x) is a solution to the differential equation dy/dx = f(x), then F
(x) = f(x), so F(x) is an antideriva-

tive of f(x).
13. False. The solution of the initial value problem dy/dx = 1 with y(0) = 5 is a solution of the differential equation that
is not positive at x = 0.
17. True. All solutions of the differential equation dy/dt = 3t
2
are in the family y(t) = t
3
+C of antiderivatives of 3t
2
. The
initial condition y(1) = tells us that y(1) = = 1
3
+ C, so C = 1. Thus y(t) = t
3
+ 1 is the only solution
of the initial value problem.
21. True. Suppose t is measured in seconds from when the ball was thrown. The acceleration a = dv/dt is 32 ft/sec
2
, so
the velocity of the ball is v = 32t + C feet/second at time t. At t = 0 the velocity is 10, so v = 32t 10. Since
v = ds/dt, an antiderivative gives the height s = 16t
2
10t + K feet of the ball at time t. Since the ball starts at the
top of the building, s = 100 when t = 0. Substituting gives s = 16t
2
10t + 100. The ball hits the ground when
s = 0, so we solve 0 = 16t
2
10t + 100. The positive solution t = 2.2 tells us that the ball hits the ground after
2.2 seconds.
25. True. Since F and G are both antiderivatives of f, they must differ by a constant. In fact, we can see that the constant C
is equal to
_
2
0
f(t)dt since
F(x) =
_
x
0
f(t)dt =
_
x
2
f(t)dt +
_
2
0
f(t)dt = G(x) + C.
7.1 SOLUTIONS 133
CHAPTER SEVEN
Exercises
1. (a) We substitute w = 1 + x
2
, dw = 2xdx.
_
x=1
x=0
x
1 + x
2
dx =
1
2
_
w=2
w=1
1
w
dw =
1
2
ln |w|
2
1
=
1
2
ln 2.
(b) We substitute w = cos x, dw = sin xdx.
_
x=
4
x=0
sin x
cos x
dx =
_
w=
2/2
w=1
1
w
dw
= ln |w|
2/2
1
= ln
2
2
=
1
2
ln 2.
5. We use the substitution w = x, dw = dx.
_
e
x
dx =
_
e
w
dw = e
w
+ C = e
x
+ C.
Check:
d
dx
(e
x
+ C) = (e
x
) = e
x
.
9. We use the substitution w = 3 t, dw = dt.
_
sin(3 t)dt =
_
sin(w)dw = (cos(w)) + C = cos(3 t) + C.
Check:
d
dt
(cos(3 t) + C) = sin(3 t)(1) = sin(3 t).
13. We use the substitution w = t
3
3, dw = 3t
2
dt.
_
t
2
(t
3
3)
10
dt =
1
3
_
(t
3
3)
10
(3t
2
dt) =
_
w
10
_
1
3
dw
_
=
1
3
w
11
11
+ C =
1
33
(t
3
3)
11
+ C.
Check:
d
dt
[
1
33
(t
3
3)
11
+ C] =
1
3
(t
3
3)
10
(3t
2
) = t
2
(t
3
3)
10
.
17. In this case, it seems easier not to substitute.
_
y
2
(1 + y)
2
dy =
_
y
2
(y
2
+ 2y + 1) dy =
_
(y
4
+ 2y
3
+ y
2
) dy
=
y
5
5
+
y
4
2
+
y
3
3
+ C.
Check:
d
dy
_
y
5
5
+
y
4
2
+
y
3
3
+ C
_
= y
4
+ 2y
3
+ y
2
= y
2
(y + 1)
2
.
134 Chapter Seven /SOLUTIONS
21. In this case, it seems easier not to substitute.
_
(x
2
+ 3)
2
dx =
_
(x
4
+ 6x
2
+ 9) dx =
x
5
5
+ 2x
3
+ 9x + C.
Check:
d
dx
_
x
5
5
+ 2x
3
+ 9x + C
_
= x
4
+ 6x
2
+ 9 = (x
2
+ 3)
2
.
25. We use the substitution w = sin , dw = cos d.
_
sin
6
cos d =
_
w
6
dw =
w
7
7
+ C =
sin
7
7
+ C.
Check:
d
d
_
sin
7
7
+ C
_
= sin
6
cos .
29. We use the substitution w = ln z, dw =
1
z
dz.
_
(ln z)
2
z
dz =
_
w
2
dw =
w
3
3
+ C =
(ln z)
3
3
+ C.
Check:
d
dz
_
(ln z)
3
3
+ C
_
= 3
1
3
(ln z)
2
1
z
=
(ln z)
2
z
.
33. We use the substitution w =
y, dw =
1
2
y
dy.
_
e
y
dy = 2
_
e
w
dw = 2e
w
+ C = 2e
y
+ C.
Check:
d
dy
(2e
y
+ C) = 2e
1
2
y
=
e
y
.
37. We use the substitution w = 1 + 3t
2
, dw = 6t dt.
_
t
1 + 3t
2
dt =
_
1
w
(
1
6
dw) =
1
6
ln |w| + C =
1
6
ln(1 + 3t
2
) + C.
(We can drop the absolute value signs since 1 + 3t
2
> 0 for all t).
Check:
d
dt
_
1
6
ln(1 + 3t
2
) + C
_
=
1
6
1
1 + 3t
2
(6t) =
t
1 + 3t
2
.
41. Since d(sinh x)/dx = cosh x, we have
_
cosh xdx = sinh x + C.
45. We use the substitution w = x
2
and dw = 2xdx so
_
xcosh x
2
dx =
1
2
_
cosh w dw =
1
2
sinh w + C =
1
2
sinh x
2
+ C.
Check this answer by taking the derivative:
d
dx
_
1
2
sinh x
2
+ C
_
= xcosh x
2
.
49. Make the substitution w = x
2
, dw = 2xdx. We have
_
2xcos(x
2
) dx =
_
cos wdw = sin w + C = sin x
2
+ C.
53. Make the substitution w = x
2
+ 1, dw = 2xdx. We have
_
x
x
2
+ 1
dx =
1
2
_
dw
w
=
1
2
ln |w| + C =
1
2
ln(x
2
+ 1) + C.
(Notice that since x
2
+ 1 0, |x
2
+ 1| = x
2
+ 1.)
57.
_
/2
0
e
cos
sin d = e
cos
/2
0
= e
cos(/2)
e
cos(0)
= 1
1
e
7.1 SOLUTIONS 135
61. We substitute w =
x. Then dw =
1
2
x
1/2
dx.
_
x=4
x=1
cos
x
dx =
_
w=2
w=1
cos w(2 dw)
= 2(sin w)
2
1
= 2(sin 2 sin 1).
65.
_
3
1
1
x
dx = ln x
3
1
= ln 3.
69. Let w =
y + 1, so y = w
2
1 and dy = 2wdw. Thus
_
y
_
y + 1 dy =
_
(w
2
1)w2wdw = 2
_
w
4
w
2
dw
=
2
5
w
5
2
3
w
3
+ C =
2
5
(y + 1)
5/2
2
3
(y + 1)
3/2
+ C.
73. Let w =
x 2, so x = w
2
+ 2 and dx = 2wdw. Thus
_
x
2
x 2 dx =
_
(w
2
+ 2)
2
w2wdw = 2
_
w
6
+ 4w
4
+ 4w
2
dw
=
2
7
w
7
+
8
5
w
5
+
8
3
w
3
+ C
=
2
7
(x 2)
7/2
+
8
5
(x 2)
5/2
+
8
3
(x 2)
3/2
+ C.
Problems
77. If we let y = 3x in the rst integral, we get dy = 3dx. Also, the limits x = 0 and x = /3 become y = 0 and
y = 3(/3) = . Thus
_
/3
0
3 sin
2
(3x) dx =
_
/3
0
sin
2
(3x) 3dx =
_

0
sin
2
(y) dy.
81. As x goes from
a to
b the values of w = x
2
increase from a to b. Since x =
w we have dx = dw/(2
w). Hence
_

b
a
dx =
_
b
a
1
2
w
dw =
_
b
a
g(w)dw
with
g(w) =
1
2
w
.
85. For the rst integral, let w = sin x, dw = cos xdx. Then
_
e
sin x
cos xdx =
_
e
w
dw.
For the second integral, let w = arcsin x, dw =
1
1x
2
dx. Then
_
e
arcsin x
1 x
2
dx =
_
e
w
dw.
89. (a) This integral can be evaluated using integration by substitution. We use w = x
2
, dw = 2xdx.
_
xsin x
2
dx =
1
2
_
sin(w)dw =
1
2
cos(w) + C =
1
2
cos(x
2
) + C.
(b) This integral cannot be evaluated using a simple integration by substitution.
(c) This integral cannot be evaluated using a simple integration by substitution.
(d) This integral can be evaluated using integration by substitution. We use w = 1 + x
2
, dw = 2xdx.
_
x
(1 + x
2
)
2
dx =
1
2
_
1
w
2
dw =
1
2
(
1
w
) + C =
1
2(1 + x
2
)
+ C.
(e) This integral cannot be evaluated using a simple integration by substitution.
(f) This integral can be evaluated using integration by substitution. We use w = 2 + cos x, dw = sin xdx.
_
sin x
2 + cos x
dx =
_
1
w
dw = ln |w| + C = ln |2 + cos x| + C.
93. Since (e
+1
)
3
= e
3+3
= e
3
e
3
, we have
Area =
_
2
0
(e
+1
)
3
d =
_
2
0
e
3
e
3
d
= e
3
_
2
0
e
3
d = e
3
1
3
e
3
2
0
=
e
3
3
(e
6
1).
97. If f(x) =
1
x + 1
, the average value of f on the interval 0 x 2 is dened to be
1
2 0
_
2
0
f(x) dx =
1
2
_
2
0
dx
x + 1
.
Well integrate by substitution. We let w = x + 1 and dw = dx, and we have
_
x=2
x=0
dx
x + 1
=
_
w=3
w=1
dw
w
= ln w
3
1
= ln 3 ln 1 = ln 3.
Thus, the average value of f(x) on 0 x 2 is
1
2
ln 3 0.5493. See Figure 7.1.
2
0.54931
f(x) =
1
1+x
x
Figure 7.1
101. (a) The Fundamental Theorem gives
_

cos
2
sin d =
cos
3
=
(1)
3
3

(1)
3
3
= 0.
This agrees with the fact that the function f() = cos
2
sin is odd and the interval of integration is centered at
x = 0, thus we must get 0 for the denite integral.
(b) The area is given by
Area =
_

0
cos
2
sin d =
cos
3
0
=
(1)
3
3

(1)
3
3
=
2
3
.
7.1 SOLUTIONS 137
105. We substitute w = 1 x into Im,n. Then dw = dx, and x = 1 w.
When x = 0, w = 1, and when x = 1, w = 0, so
Im,n =
_
1
0
x
m
(1 x)
n
dx =
_
0
1
(1 w)
m
w
n
(dw)
=
_
0
1
w
n
(1 w)
m
dw =
_
1
0
w
n
(1 w)
m
dw = In,m.
109. (a) At time t = 0, the rate of oil leakage = r(0) = 50 thousand liters/minute.
At t = 60, rate = r(60) = 15.06 thousand liters/minute.
(b) To nd the amount of oil leaked during the rst hour, we integrate the rate from t = 0 to t = 60:
Oil leaked =
_
60
0
50e
0.02t
dt =
_
50
0.02
e
0.02t
_
60
0
= 2500e
1.2
+ 2500e
0
= 1747 thousand liters.
113. Since v is given as the velocity of a falling body, the height h is decreasing, so v =
dh
dt
, and it follows that h(t) =
_
v(t) dt and h(0) = h0. Let w = e
t
gk
+ e
t
gk
. Then
dw =
_
gk
_
e
t
gk
e
t
gk
_
dt,
so
dw
gk
= (e
t
gk
e
t
gk
) dt. Therefore,
_
v(t)dt =
_
_
g
k
_
e
t
gk
e
t
gk
e
t
gk
+ e
t
gk
_
dt
=
_
g
k
_
1
e
t
gk
+ e
t
gk
_
e
t
gk
e
t
gk
_
dt
=
_
g
k
_
_
1
w
_
dw
gk
=
_
g
gk
2
ln |w| + C
=
1
k
ln
_
e
t
gk
+ e
t
gk
_
+ C.
Since
h(0) =
1
k
ln(e
0
+ e
0
) + C =
ln 2
k
+ C = h0,
we have C = h0 +
ln 2
k
. Thus,
h(t) =
1
k
ln
_
e
t
gk
+ e
t
gk
_
+
ln 2
k
+ h0 =
1
k
ln
_
e
t
gk
+ e
t
gk
2
_
+ h0.
Exercises
1. (a) Since we can change x
2
into a multiple of x
3
by integrating, let v
= x
2
and u = e
x
. Using v = x
3
/3 and u
= e
x
we get
_
x
2
e
x
dx =
_
uv
dx = uv
_
u
vdx
=
x
3
e
x
3

1
3
_
x
3
e
x
dx.
(b) Since we can change x
2
into a multiple of x by differentiating, let u = x
2
and v
= e
x
. Using u
= 2x and v = e
x
we have
_
x
2
e
x
dx =
_
uv
dx = uv
_
u
vdx
= x
2
e
x
2
_
xe
x
dx.
5. Let u = t and v
= e
5t
, so u
= 1 and v =
1
5
e
5t
.
Then
_
te
5t
dt =
1
5
te
5t
_
1
5
e
5t
dt =
1
5
te
5t
1
25
e
5t
+ C.
9. Let u = ln y, v
= y. Then, v =
1
2
y
2
and u
=
1
y
. Integrating by parts, we get:
_
y ln y dy =
1
2
y
2
ln y
_
1
2
y
2
1
y
dy
=
1
2
y
2
ln y
1
2
_
y dy
=
1
2
y
2
ln y
1
4
y
2
+ C.
13. Let u = sin and v
= sin , so u
= cos and v = cos . Then

_
sin
2
d = sin cos +
_
cos
2
d
= sin cos +
_
(1 sin
2
) d
= sin cos +
_
1 d
_
sin
2
d.
By adding
_
sin
2
d to both sides of the above equation, we nd that 2
_
sin
2
d = sin cos + + C, so
_
sin
2
d =
1
2
sin cos +

2
+ C
.
17. Let u = t + 2 and v
2 + 3t, so u
= 1 and v =
2
9
(2 + 3t)
3/2
. Then
_
(t + 2)
2 + 3t dt =
2
9
(t + 2)(2 + 3t)
3/2
2
9
_
(2 + 3t)
3/2
dt
=
2
9
(t + 2)(2 + 3t)
3/2
4
135
(2 + 3t)
5/2
+ C.
21. Let u = y and v
=
1
5y
, so u
= 1 and v = 2(5 y)
1/2
.
_
y
5 y
dy = 2y(5 y)
1/2
+ 2
_
(5 y)
1/2
dy = 2y(5 y)
1/2
4
3
(5 y)
3/2
+ C.
7.2 SOLUTIONS 139
25. Let u = arctan 7z and v
= 1, so u
=
7
1+49z
2
and v = z. Now
_
7z dz
1+49z
2
can be evaluated by the substitution
w = 1 + 49z
2
, dw = 98z dz, so
_
7z dz
1 + 49z
2
= 7
_
1
98
dw
w
=
1
14
_
dw
w
=
1
14
ln |w| + C =
1
14
ln(1 + 49z
2
) + C
So
_
arctan 7z dz = z arctan 7z
1
14
ln(1 + 49z
2
) + C.
29. Let u = x, u
= 1 and v
= sinh x, v = cosh x. Integrating by parts, we get

_
xsinh x dx = xcosh x
_
cosh x dx
= xcosh x sinh x + C.
33. We use integration by parts. Let u = z and v
= e
z
, so u
= 1 and v = e
z
. Then
_
10
0
ze
z
dz = ze
z
10
0
+
_
10
0
e
z
dz
= 10e
10
+ (e
z
)
10
0
= 11e
10
+ 1
0.9995.
37. We use integration by parts. Let u = arcsin z and v
= 1, so u
=
1
1 z
2
and v = z. Then
_
1
0
arcsin z dz = z arcsin z
1
0
_
1
0
z
1 z
2
dz =

2

_
1
0
z
1 z
2
dz.
To nd
_
1
0
z
1 z
2
dz, we substitute w = 1 z
2
, so dw = 2z dz.
Then
_
z=1
z=0
z
1 z
2
dz =
1
2
_
w=0
w=1
w
1
2
dw =
1
2
_
w=1
w=0
w
1
2
dw = w
1
2
1
0
= 1.
Thus our nal answer is

2
1 0.571.
Problems
41. Using integration by parts with u
= e
t
, v = t, so u = e
t
and v
= 1, we have
Area =
_
2
0
te
t
dt = te
t
2
0
_
2
0
1 e
t
dt
= (te
t
e
t
)
2
0
= 2e
2
e
2
+ 1 = 1 3e
2
.
45. Since the graph of f(t) = ln(t
2
1) is above the graph of g(t) = ln(t 1) for t > 1, we have
Area =
_
3
2
(ln(t
2
1) ln(t 1)) dt =
_
3
2
ln
_
t
2
1
t 1
_
dt =
_
3
2
ln(t + 1) dt.
We can cancel the factor of (t 1) in the last step above because the integral is over 2 t 3, where (t 1) is not zero.
We use
_
ln xdx = xln x 1 with the substitution x = t + 1. The limits t = 2, t = 3 become x = 3, x = 4,
respectively. Thus
Area =
_
3
2
ln(t + 1) dt =
_
4
3
ln xdx = (xln x x)
4
3
= 4 ln 4 4 (3 ln 3 3) = 4 ln 4 3 ln 3 1.
49. First, let u = e
x
and v
= sin x, so u
= e
x
and v = cos x.
Thus
_
e
x
sin xdx = e
x
cos x +
_
e
x
cos xdx. To calculate
_
e
x
cos xdx, we again need to use integration by parts.
Let u = e
x
and v
= cos x, so u
= e
x
and v = sin x.
Thus
_
e
x
cos xdx = e
x
sin x
_
e
x
sin xdx.
This gives
_
e
x
sin xdx = e
x
sin x e
x
cos x
_
e
x
sin xdx.
By adding
_
e
x
sin xdx to both sides, we obtain
2
_
e
x
sin xdx = e
x
(sin x cos x) + C.
Thus
_
e
x
sin xdx =
1
2
e
x
(sin x cos x) + C.
This problem could also be done in other ways; for example, we could have started with u = sin x and v
= e
x
as well.
53. We integrate by parts. Since we know what the answer is supposed to be, its easier to choose u and v
. Let u = x
n
and
v
= e
x
, so u
= nx
n1
and v = e
x
. Then
_
x
n
e
x
dx = x
n
e
x
n
_
x
n1
e
x
dx.
57. (a) One way to avoid integrating by parts is to take the derivative of the right hand side instead. Since
_
e
ax
sin bxdx is
the antiderivative of e
ax
sin bx,
e
ax
sin bx =
d
dx
[e
ax
(Asin bx + B cos bx) + C]
= ae
ax
(Asin bx + B cos bx) + e
ax
(Ab cos bx Bb sin bx)
= e
ax
[(aAbB) sin bx + (aB + bA) cos bx].
Thus aAbB = 1 and aB + bA = 0. Solving for A and B in terms of a and b, we get
A =
a
a
2
+ b
2
, B =
b
a
2
+ b
2
.
Thus
_
e
ax
sin bx = e
ax
_
a
a
2
+ b
2
sin bx
b
a
2
+ b
2
cos bx
_
+ C.
7.2 SOLUTIONS 141
(b) If we go through the same process, we nd
ae
ax
[(aAbB) sin bx + (aB + bA) cos bx] = e
ax
cos bx.
Thus aAbB = 0, and aB + bA = 1. In this case, solving for A and B yields
A =
b
a
2
+ b
2
, B =
a
a
2
+ b
2
.
Thus
_
e
ax
cos bx = e
ax
(
b
a
2
+b
2
sin bx +
a
a
2
+b
2
cos bx) + C.
61. We have
Bioavailability =
_
3
0
15te
0.2t
dt.
We rst use integration by parts to evaluate the indenite integral of this function. Let u = 15t and v
= e
0.2t
dt, so
u
= 15dt and v = 5e
0.2t
. Then,
_
15te
0.2t
dt = (15t)(5e
0.2t
)
_
(5e
0.2t
)(15dt)
= 75te
0.2t
+ 75
_
e
0.2t
dt = 75te
0.2
375e
0.2t
+ C.
Thus,
_
3
0
15te
0.2t
dt = (75te
0.2t
375e
0.2t
)
3
0
= 329.29 + 375 = 45.71.
The bioavailability of the drug over this time interval is 45.71 (ng/ml)-hours.
65. (a) We want to compute C1, with C1 > 0, such that
_
1
0
(1(x))
2
dx =
_
1
0
(C1 sin(x))
2
dx = C
2
1
_
1
0
sin
2
(x) dx = 1.
We use integration by parts with u = v
= sin(x).
So u
= cos(x) and v =
1
cos(x). Thus
_
1
0
sin
2
(x) dx =
1
sin(x) cos(x)
1
0
+
_
1
0
cos
2
(x) dx
=
1
sin(x) cos(x)
1
0
+
_
1
0
(1 sin
2
(x)) dx.
Moving
_
1
0
sin
2
(x) dx from the right side to the left side of the equation and solving, we get
2
_
1
0
sin
2
(x) dx =
1
sin(x) cos(x)
1
0
+
_
1
0
1 dx = 0 + 1 = 1,
so
_
1
0
sin
2
(x) dx =
1
2
.
Thus, we have
_
1
0
(1(x))
2
dx = C
2
1
_
1
0
sin
2
(x) dx =
C
2
1
2
.
So, to normalize 1, we take C1 > 0 such that
C
2
1
2
= 1 so C1 =
2.
(b) To normalize n, we want to compute Cn, with Cn > 0, such that
_
1
0
(n(x))
2
dx = C
2
n
_
1
0
sin
2
(nx) dx = 1.
The solution to part (a) shows us that
_
sin
2
(t) dt =
1
2
sin(t) cos(t) +
1
2
_
1 dt.
In the integral for n, we make the substitution t = nx, so dx =
1
n
dt. Since t = 0 when x = 0 and t = n when
x = 1, we have
_
1
0
sin
2
(nx) dx =
1
n
_
n
0
sin
2
(t) dt
=
1
n
_
1
2
sin(t) cos(t)
n
0
+
1
2
_
n
0
1 dt
_
=
1
n
_
0 +
n
2
_
=
1
2
.
Thus, we have
_
1
0
(n(x))
2
dx = C
2
n
_
1
0
sin
2
(nx) dx =
C
2
n
2
.
So to normalize n, we take Cn such that
C
2
n
2
= 1 so Cn =
2.
Exercises
1.
1
10
e
(3)
(3 cos + sin ) + C.
(Let a = 3, b = 1 in II-9.)
5. Note that you cant use substitution here: letting w = x
3
+ 5 does not work, since there is no dw = 3x
2
dx in the
integrand. What will work is simply multiplying out the square: (x
3
+ 5)
2
= x
6
+ 10x
3
+ 25. Then use I-1:
_
(x
3
+ 5)
2
dx =
_
x
6
dx + 10
_
x
3
dx + 25
_
1 dx =
1
7
x
7
+ 10
1
4
x
4
+ 25x + C.
9.
1
3
arctan
y
3
+ C.
(Let a =
3 in V-24).
13.
5
16
sin 3 sin 5 +
3
16
cos 3 cos 5 + C.
(Let a = 3, b = 5 in II-12.)
17.
_
1
3
x
4
4
9
x
3
+
4
9
x
2
8
27
x +
8
81
_
e
3x
+ C.
(Let a = 3, p(x) = x
4
in III-14.)
21. Substitute w = x
2
, dw = 2xdx. Then
_
x
3
sin x
2
dx =
1
2
_
wsin wdw. By III-15, we have
_
wsin wdw =
1
2
wcos w +
1
2
sin w + C =
1
2
x
2
cos x
2
+
1
2
sin x
2
+ C.
7.3 SOLUTIONS 143
25. Use IV-21 twice to get the exponent down to 1:
_
1
cos
5
x
dx =
1
4
sin x
cos
4
x
+
3
4
_
1
cos
3
x
dx
_
1
cos
3
x
dx =
1
2
sin x
cos
2
x
+
1
2
_
1
cos x
dx.
Now use IV-22 to get
_
1
cos x
dx =
1
2
ln
(sin x) + 1
(sin x) 1
+ C.
Putting this all together gives
_
1
cos
5
x
dx =
1
4
sin x
cos
4
x
+
3
8
sin x
cos
2
x
+
3
16
ln
(sin x) + 1
(sin x) 1
+ C.
29.
_
1
x
2
+ 4x + 3
dx =
_
1
(x + 1)(x + 3)
dx =
1
2
(ln |x + 1| ln |x + 3|) + C.
(Let a = 1 and b = 3 in V-26).
33. arctan(z + 2) + C.
(Substitute w = z + 2 and use V-24, letting a = 1.)
37.
_
sin
3
3 cos
2
3 d =
_
(sin 3)(cos
2
3)(1 cos
2
3) d
=
_
sin 3(cos
2
3 cos
4
3) d.
Using an extension of the tip given in rule IV-23, we let w = cos 3, dw = 3 sin 3 d.
_
sin 3(cos
2
3 cos
4
3) d =
1
3
_
(w
2
w
4
) dw
=
1
3
(
w
3
3

w
5
5
) + C
=
1
9
(cos
3
3) +
1
15
(cos
5
3) + C.
41. Let a =
3 in VI-30 and VI-28:

_
1
0
_
3 x
2
dx = (
1
2
x
_
3 x
2
+
3
2
arcsin
x
3
)
1
0
1.630.
45.
_
1
0
1
x
2
+ 2x + 1
dx =
_
1
0
1
(x + 1)
2
dx.
We substitute w = x + 1, so dw = dx. Note that when x = 1, we have w = 2, and when x = 0, we have w = 1.
_
x=1
x=0
1
(x + 1)
2
dx =
_
w=2
w=1
1
w
2
dw =
1
w
w=2
w=1
=
1
2
+ 1 =
1
2
.
49. Let w = x
2
, dw = 2xdx. When x = 0, w = 0, and when x =
1
2
, w =
1
2
. Then
_ 1
2
0
xdx
1 x
4
=
_ 1
2
0
1
2
dw
1 w
2
=
1
2
arcsin w
1
2
0
=
1
2
(arcsin
1
2
arcsin 0) =

12
.
Problems
53. (a)
1
1 0
_
1
0
V0 cos(120t)dt =
V0
120
sin(120t)
1
0
=
V0
120
[sin(120) sin(0)]
=
V0
120
[0 0] = 0.
(b) Lets nd the average of V
2
rst.
V
2
= Average of V
2
=
1
1 0
_
1
0
V
2
dt
=
1
1 0
_
1
0
(V0 cos(120t))
2
dt
= V
2
0
_
1
0
cos
2
(120t)dt
Now, let 120t = x, and dt =
dx
120
. So
V
2
=
V
2
0
120
_
120
0
cos
2
xdx.
=
V
2
0
120
_
1
2
cos xsin x +
1
2
x
_
120
0
II-18
=
V
2
0
120
60 =
V
2
0
2
.
So, the average of V
2
is
V
2
0
2
and V =
_
average of V
2
=
V0
2
.
(c) V0 =
2 V = 110
2 156 volts.
Exercises
1. Since 25 x
2
= (5 x)(5 + x), we take
20
25 x
2
=
A
5 x
+
B
5 + x
.
So,
20 = A(5 + x) + B(5 x)
20 = (AB)x + 5A + 5B,
giving
AB = 0
5A + 5B = 20.
Thus A = B = 2 and
20
25 x
2
=
2
5 x
+
2
5 + x
.
7.4 SOLUTIONS 145
5. Since s
4
1 = (s
2
1)(s
2
+ 1) = (s 1)(s + 1)(s
2
+ 1), we have
2
s
4
1
=
A
s 1
+
B
s + 1
+
Cs + D
s
2
+ 1
.
Thus,
2 = A(s + 1)(s
2
+ 1) + B(s 1)(s
2
+ 1) + (Cs + D)(s 1)(s + 1)
2 = (A + B + C)s
3
+ (AB + D)s
2
+ (A + B C)s + (AB D),
giving
A + B + C = 0
AB + D = 0
A + B C = 0
AB D = 2.
From the rst and third equations we nd A + B = 0 and C = 0. From the second and fourth we nd A B = 1 and
D = 1. Thus A = 1/2 and B = 1/2 and
2
s
4
1
=
1
2(s 1)

1
2(s + 1)

1
s
2
+ 1
.
9. Using the result of Problem 2, we have
_
x + 1
6x + x
2
dx =
_
1/6
x
dx +
_
5/6
6 + x
dx =
1
6
(ln |x| + 5 ln |6 + x|) + C.
13. Using the result of Problem 6, we have
_
2y
y
3
y
2
+ y 1
dy =
_
1
y 1
dy +
_
1 y
y
2
+ 1
dy = ln |y 1| + arctan y
1
2
ln
y
2
+ 1
+ C.
17. We let
10x + 2
x
3
5x
2
+ x 5
=
10x + 2
(x 5)(x
2
+ 1)
=
A
x 5
+
Bx + C
x
2
+ 1
giving
10x + 2 = A(x
2
+ 1) + (Bx + C)(x 5)
10x + 2 = (A + B)x
2
+ (C 5B)x + A5C
so
A + B = 0
C 5B = 10
A5C = 2.
Thus, A = 2, B = 2, C = 0, so
_
10x + 2
x
3
5x
2
+ x 5
dx =
_
2
x 5
dx
_
2x
x
2
+ 1
dx = 2 ln |x 5| ln
x
2
+ 1
+ K.
21. Completing the square gives x
2
+ 4x + 5 = 1 + (x + 2)
2
. Since x + 2 = tan t and dx = (1/ cos
2
t)dt, we have
_
1
x
2
+ 4x + 5
dx =
_
1
1 + tan
2
t

1
cos
2
t
dt =
_
dt = t + C = arctan(x + 2) + C.
Problems
25. (a) We have
3x + 6
x
2
+ 3x
=
2(x + 3)
x(x + 3)
+
x
x(x + 3)
=
2
x
+
1
x + 3
.
Thus
_
3x + 6
x
2
+ 3x
dx =
_
_
2
x
+
1
x + 3
_
dx = 2 ln |x| + ln |x + 3| + C.
(b) Let a = 0, b = 3, c = 3, and d = 6 in V-27.
_
3x + 6
x
2
+ 3x
dx =
_
3x + 6
x(x + 3)
dx
=
1
3
(6 ln |x| + 3 ln |x + 3|) + C = 2 ln |x| + ln |x + 3| + C.
29. Since y
2
+ 3y + 3 = (y + 3/2)
2
+ (3 9/4) = (y + 3/2)
2
+ 3/4, we have
_
dy
y
2
+ 3y + 3
=
_
dy
(y + 3/2)
2
+ 3/4
.
Substitute y + 3/2 = tan , so y = (tan ) 3/2.
33. Since t
2
+ 4t + 7 = (t + 2)
2
+ 3, we have
_
(t + 2) sin(t
2
+ 4t + 7) dt =
_
(t + 2) sin((t + 2)
2
+ 3) dt.
Substitute w = (t + 2)
2
+ 3, so dw = 2(t + 2) dt.
This integral can also be computed without completing the square, by substituting w = t
2
+ 4t + 7, so dw =
(2t + 4) dt.
37. We write
1
(x + 7)(x 2)
=
A
x + 7
+
B
x 2
,
giving
1 = A(x 2) + B(x + 7)
1 = (A + B)x + (2A + 7B)
so
A + B = 0
2A + 7B = 1.
Thus, A = 1/9, B = 1/9, so
_
1
(x + 7)(x 2)
dx =
_
1/9
x + 7
dx +
_
1/9
x 2
dx =
1
9
ln |x + 7| +
1
9
ln |x 2| + C.
41. We use partial fractions and write
1
3P 3P
2
=
A
3P
+
B
1 P
,
multiply through by 3P(1 P), and then solve for A and B, getting A = 1 and B = 1/3. So
_
dP
3P 3P
2
=
_ _
1
3P
+
1
3(1 P)
_
dP =
1
3
_
dP
P
+
1
3
_
dP
1 P
=
1
3
ln |P|
1
3
ln |1 P| + C =
1
3
ln
P
1 P
+ C.
7.4 SOLUTIONS 147
45. Since x
2
+ x
4
= x
2
(1 + x
2
) cannot be factored further, we write
x 2
x
2
+ x
4
=
A
x
+
B
x
2
+
Cx + D
1 + x
2
.
Multiplying by x
2
(1 + x
2
) gives
x 2 = Ax(1 + x
2
) + B(1 + x
2
) + (Cx + D)x
2
x 2 = (A + C)x
3
+ (B + D)x
2
+ Ax + B,
so
A + C = 0
B + D = 0
A = 1
B = 2.
Thus, A = 1, B = 2, C = 1, D = 2, and we have
_
x 2
x
2
+ x
4
dx =
_
_
1
x

2
x
2
+
x + 2
1 + x
2
_
dx =
_
dx
x
2
_
dx
x
2

_
xdx
1 + x
2
+ 2
_
dx
1 + x
2
= ln |x| +
2
x

1
2
ln
1 + x
2
+ 2 arctan x + K.
We use K as the constant of integration, since we already used C in the problem.
49. Since (4 z
2
)
3/2
= (
4 z
2
)
3
, we substitute z = 2 sin , so dz = 2 cos d. We get
_
dz
(4 z
2
)
3/2
=
_
2 cos d
(4 4 sin
2
)
3/2
=
_
2 cos d
8 cos
3
=
1
4
_
d
cos
2
=
1
4
tan + C
Since sin = z/2, we have cos =
_
1 (z/2)
2
= (
4 z
2
)/2, so
_
dz
(4 z
2
)
3/2
=
1
4
tan + C =
1
4
sin
cos
+ C =
1
4
z/2
(
4 z
2
)/2
+ C =
z
4
4 z
2
+ C
53. Notice that because
3x
(x1)(x4)
is negative for 2 x 3,
Area =
_
3
2
3x
(x 1)(x 4)
dx.
Using partial fractions gives
3x
(x 1)(x 4)
=
A
x 1
+
B
x 4
=
(A + B)x B 4A
(x 1)(x 4)
.
Multiplying through by (x 1)(x 4) gives
3x = (A + B)x B 4A
so A = 1 and B = 4. Thus
_
3
2
3x
(x 1)(x 4)
dx =
_
3
2
_
1
x 1
+
4
x 4
_
dx = (ln |x 1| 4 ln |x 4|)
3
2
= 5 ln 2.
57. We have
Area =
_
3
0
1
x
2
+ 9
dx.
Let x = 3 tan so dx = (3/ cos
2
)d and
_
x
2
+ 9 =
_
9 sin
2
cos
2
+ 9 =
3
cos
.
When x = 0, = 0 and when x = 3, = /4. Thus
_
3
0
1
x
2
+ 9
dx =
_
/4
0
1
9 tan
2
+ 9
3
cos
2
d =
_
/4
0
1
3/ cos

3
cos
2
d =
_
/4
0
1
cos
d
=
1
2
ln
sin + 1
sin 1
/4
0
=
1
2
ln
1/
2 + 1
1/
2 1
=
1
2
ln
_
1 +
2 1
_
.
This answer can be simplied to ln(1 +
2) by multiplying the numerator and denominator of the fraction by (
2 + 1)
and using the properties of logarithms. The integral
_
(1/ cos )d is done using the Table of Integrals.
61. Using partial fractions, we write
3x
2
+ 1
x
3
+ x
=
3x
2
+ 1
x(x
2
+ 1)
=
A
x
+
Bx + C
x
2
+ 1
3x
2
+ 1 = A(x
2
+ 1) + (Bx + C)x = (A + B)x
2
+ Cx + A.
So, A + B = 3, C = 0 and A = 1, giving B = 2. Thus
_
3x
2
+ 1
x
3
+ x
dx =
_
_
1
x
+
2x
x
2
+ 1
_
dx = ln |x| + ln
x
2
+ 1
+ K.
Using the substitution w = x
3
+ x, we get dw = (3x
2
+ 1)dx, so we have
_
3x
2
+ 1
x
3
+ x
dx =
_
dw
w
= ln |w| + K = ln
x
3
+ x
+ K.
The properties of logarithms show that the two results are the same:
ln |x| + ln
x
2
+ 1
+ K = ln
x(x
2
+ 1)
+ K = ln
x
3
+ x
+ K.
65. (a) If a > 0, then
x
2
a = (x
a)(x +
a).
This means that we can use partial fractions:
1
x
2
a
=
A
x
a
+
B
x +
a
,
giving
1 = A(x +
a) + B(x
a),
so A + B = 0 and (AB)
a = 1. Thus, A = B = 1/(2
a).
So
_
1
x
2
a
dx =
_
1
2
a
_
1
x
a

1
x +
a
_
dx =
1
2
a
(ln
ln
x +
) + C.
(b) If a = 0, we have
_
1
x
2
dx =
1
x
+ C.
(c) If a < 0, then a > 0 so x
2
a = x
2
+ (a) cannot be factored. Thus
_
1
x
2
a
dx =
_
1
x
2
+ (a)
dx =
1
a
arctan
_
x
a
_
+ C.
7.5 SOLUTIONS 149
Exercises
1. (a) The approximation LEFT(2) uses two rectangles, with the height of each rectangle determined by the left-hand
endpoint. See Figure 7.2. We see that this approximation is an underestimate.
a
b
x
Figure 7.2
a
b
x
Figure 7.3
(b) The approximation RIGHT(2) uses two rectangles, with the height of each rectangle determined by the right-hand
endpoint. See Figure 7.3. We see that this approximation is an overestimate.
(c) The approximation TRAP(2) uses two trapezoids, with the height of each trapezoid given by the secant line connect-
ing the two endpoints. See Figure 7.4. We see that this approximation is an overestimate.
a
b
x
Figure 7.4
(d) The approximation MID(2) uses two rectangles, with the height of each rectangle determined by the height at the
midpoint. Alternately, we can view MID(2) as a trapezoid rule where the height is given by the tangent line at
the midpoint. Both interpretations are shown in Figure 7.5. We see from the tangent line interpretation that this
approximation is an underestimate
a
b
x
a
b
x
Figure 7.5
5. (a) The approximation LEFT(2) uses two rectangles, with the height of each rectangle determined by the left-hand
endpoint. See Figure 7.6. We see that this approximation is an underestimate (that is, it is more negative).
a b
x
Figure 7.6
a b
x
Figure 7.7
(b) The approximation RIGHT(2) uses two rectangles, with the height of each rectangle determined by the right-hand
endpoint. See Figure 7.7. We see that this approximation is an overestimate (that is, it is less negative).
(c) The approximation TRAP(2) uses two trapezoids, with the height of each trapezoid given by the secant line connect-
ing the two endpoints. See Figure 7.8. We see that this approximation is an overestimate (that is, it is less negative).
a b
x
Figure 7.8
(d) The approximation MID(2) uses two rectangles, with the height of each rectangle determined by the height at the
midpoint. Alternately, we can view MID(2) as a trapezoid rule where the height is given by the tangent line at
the midpoint. Both interpretations are shown in Figure 7.9. We see from the tangent line interpretation that this
approximation is an underestimate (that is, it is more negative).
a b
x
a b
x
Figure 7.9
9. (a)
MID(2) = 2 f(1) + 2 f(3)
= 2 2 + 2 10
= 24
TRAP(2) =
LEFT(2) + RIGHT(2)
2
=
12 + 44
2
(see Problem 8)
= 28
7.5 SOLUTIONS 151
(b)
2 4
f(x) = x
2
+ 1
Area shaded
=MID(2)
x
2 4
f(x) = x
2
+ 1
Area shaded
=TRAP(2)
x
MID(2) is an underestimate, since f(x) = x
2
+ 1 is concave up and a tangent line will be below the curve.
TRAP(2) is an overestimate, since a secant line lies above the curve.
Problems
13. Since the function is decreasing, LEFT is an overestimate and RIGHT is an underestimate. Since the graph is concave
down, secant lines lie below the graph so TRAP is an underestimate and tangent lines lie above the graph so MID is an
overestimate. We can see that MID and TRAP are closer to the exact value than LEFT and RIGHT. In order smallest to
largest, we have:
RIGHT(n) < TRAP(n) < Exact value < MID(n) < LEFT(n).
17. f(x) is decreasing and concave up, so LEFT and TRAP give overestimates and RIGHT and MID give underestimates.
21. (a)
_
2
0
sin d = cos
2
0
= 0.
(b) See Figure 7.10. MID(1) is 0 since the midpoint of 0 and 2 is , and sin = 0. Thus MID(1) = 2(sin ) = 0.
The midpoints we use for MID(2) are /2 and 3/2, and sin(/2) = sin(3/2). Thus MID(2) = sin(/2) +
sin(3/2) = 0.
2
Figure 7.10
(c) MID(3) = 0.
In general, MID(n) = 0 for all n, even though your calculator (because of round-off error) might not return
it as such. The reason is that sin(x) = sin(2 x). If we use MID(n), we will always take sums where we are
adding pairs of the form sin(x) and sin(2 x), so the sum will cancel to 0. (If n is odd, we will get a sin in the
sum which does not pair up with anything but sin is already 0.)
25.
TRAP(n) =
LEFT(n) + RIGHT(n)
2
=
LEFT(n) + LEFT(n) + f(b)x f(a)x
2
= LEFT(n) +
1
2
(f(b) f(a))x
Exercises
1. We saw in Problem 7 in Section 7.5 that, for this denite integral, we have LEFT(2) = 27, RIGHT(2) = 135, TRAP(2) =
81, and MID(2) = 67.5. Thus,
SIMP(2) =
2MID(2) + TRAP(2)
3
=
2(67.5) + 81
3
= 72.
Notice that
_
6
0
x
2
dx =
x
3
3
6
0
=
6
3
3

0
3
3
= 72,
and so SIMP(2) gives the exact value of the integral in this case.
Problems
5. (a)
_
4
0
e
x
dx = e
x
4
0
= e
4
e
0
53.598 . . ..
(b) Computing the sums directly, since x = 2, we have
LEFT(2)= 2 e
0
+ 2 e
2
2(1) + 2(7.389) = 16.778; error = 36.820.
RIGHT(2)= 2 e
2
+ 2 e
4
2(7.389) + 2(54.598) = 123.974; error = 70.376.
TRAP(2)=
16.778 + 123.974
2
= 70.376; error = 16.778.
MID(2)= 2 e
1
+ 2 e
3
2(2.718) + 2(20.086) = 45.608; error = 7.990.
SIMP(2)=
2(45.608) + 70.376
3
= 53.864; error = 0.266.
(c) Similarly, since x = 1, we have LEFT(4)= 31.193; error = 22.405
RIGHT(4)= 84.791; error = 31.193
TRAP(4)= 57.992; error = 4.394
MID(4)= 51.428; error = 2.170
SIMP(4)= 53.616; error = 0.018
(d) For LEFT and RIGHT, we expect the error to go down by 1/2, and this is very roughly what we see. For MID and
TRAP, we expect the error to go down by 1/4, and this is approximately what we see. For SIMP, we expect the error
to go down by 1/2
4
= 1/16, and this is approximately what we see.
9. (a) If f(x) = 1, then
_
b
a
f(x) dx = (b a).
Also,
h
3
_
f(a)
2
+ 2f(m) +
f(b)
2
_
=
b a
3
_
1
2
+ 2 +
1
2
_
= (b a).
So the equation holds for f(x) = 1.
If f(x) = x, then
_
b
a
f(x) dx =
x
2
2
b
a
=
b
2
a
2
2
.
Also,
h
3
_
f(a)
2
+ 2f(m) +
f(b)
2
_
=
b a
3
_
a
2
+ 2
a + b
2
+
b
2
_
=
b a
3
_
a
2
+ a + b +
b
2
_
=
b a
3
_
3
2
b +
3
2
a
_
7.7 SOLUTIONS 153
=
(b a)(b + a)
2
=
b
2
a
2
2
.
So the equation holds for f(x) = x.
If f(x) = x
2
, then
_
b
a
f(x) dx =
x
3
3
b
a
=
b
3
a
3
3
. Also,
h
3
_
f(a)
2
+ 2f(m) +
f(b)
2
_
=
b a
3
_
a
2
2
+ 2
_
a + b
2
_
2
+
b
2
2
_
=
b a
3
_
a
2
2
+
a
2
+ 2ab + b
2
2
+
b
2
2
_
=
b a
3
_
2a
2
+ 2ab + 2b
2
2
_
=
b a
3
_
a
2
+ ab + b
2
_
=
b
3
a
3
3
.
So the equation holds for f(x) = x
2
.
(b) For any quadratic function, f(x) = Ax
2
+ Bx + C, the Facts about Sums and Constant Multiples of Integrands
give us:
_
b
a
f(x) dx =
_
b
a
(Ax
2
+ Bx + C) dx = A
_
b
a
x
2
dx + B
_
b
a
xdx + C
_
b
a
1 dx.
Now we use the results of part (a) to get:
_
b
a
f(x) dx = A
h
3
_
a
2
2
+ 2m
2
+
b
2
2
_
+ B
h
3
_
a
2
+ 2m +
b
2
_
+ C
h
3
_
1
2
+ 2 1 +
1
2
_
=
h
3
_
Aa
2
+ Ba + C
2
+ 2(Am
2
+ Bm + C) +
Ab
2
+ Bb + C
2
_
=
h
3
_
f(a)
2
+ 2f(m) +
f(b)
2
_
Exercises
1. (a) See Figure 7.11. The area extends out innitely far along the positive x-axis.
1
x
y
Figure 7.11
1
x
y
Figure 7.12
(b) See Figure 7.12. The area extends up innitely far along the positive y-axis.
5. We have
_

1
1
5x + 2
dx = lim
b
_
b
1
1
5x + 2
dx = lim
b
_
1
5
ln (5x + 2)
_
b
1
= lim
b
_
1
5
ln (5b + 2)
1
5
ln (7)
_
.
As b , we know that ln (5b + 2) , and so this integral diverges.
9. We have
_

0
xe
x
2
dx = lim
b
_
b
0
xe
x
2
dx = lim
b
_
1
2
e
x
2
_
b
0
= lim
b
_
1
2
e
b
2
1
2
_
= 0 +
1
2
=
1
2
.
This integral converges to 1/2.
13.
_
0
e
x
1 + e
x
dx = lim
b
_
0
b
e
x
1 + e
x
dx
= lim
b
ln |1 + e
x
|
0
b
= lim
b
[ln |1 + e
0
| ln |1 + e
b
|]
= ln(1 + 1) ln(1 + 0) = ln 2.
17. This integral is improper because 1/v is undened at v = 0. Then
_
1
0
1
v
dv = lim
b0
+
_
1
b
1
v
dv = lim
b0
+
_
ln v
1
b
_
= ln b.
As b 0
+
, this goes to innity and the integral diverges.
21. We use V-26 with a = 4 and b = 4:
_
4
0
1
u
2
16
du = lim
b4
_
b
0
1
u
2
16
du
= lim
b4
_
b
0
1
(u 4)(u + 4)
du
= lim
b4
(ln |u 4| ln |u + 4|)
8
b
0
= lim
b4
1
8
(ln |b 4| + ln 4 ln |b + 4| ln 4) .
As b 4
, ln |b 4| , so the limit does not exist and the integral diverges.

25. This is a proper integral; use V-26 in the integral table with a = 4 and b = 4.
_
20
16
1
y
2
16
dy =
_
20
16
1
(y 4)(y + 4)
dy
=
ln |y 4| ln |y + 4|
8
20
16
=
ln 16 ln 24 (ln 12 ln 20)
8
=
ln 320 ln 288
8
=
1
8
ln(10/9) = 0.01317.
7.7 SOLUTIONS 155
29.
_
2
0
1
4 x
2
dx = lim
b2
_
b
0
1
4 x
2
dx
= lim
b2
arcsin
x
2
b
0
= lim
b2
arcsin
b
2
= arcsin 1 =

2
.
33. The integrand is undened at y = 3 and y = 3. To consider the limits one at a time, divide the integral at y = 0;
_
3
0
y dy
_
9 y
2
= lim
b3
_
b
0
y
_
9 y
2
dy = lim
b3
_
(9 y
2
)
1/2
_
b
0
= lim
b3
_
3 (9 b
2
)
1/2
_
= 3.
A similar argument shows that
_
0
3
y dy
_
9 y
2
= lim
b3
+
_
0
b
y
_
9 y
2
dy = lim
b3
+
_
(9 y
2
)
1/2
_
0
b
= lim
b3
+
_
3 + (9 b
2
)
1/2
_
= 3.
Thus the original integral converges to a value of 0:
_
3
3
y dy
_
9 y
2
=
_
0
3
y dy
_
9 y
2
+
_
3
0
y dy
_
9 y
2
= 3 + 3 = 0.
Problems
37. (a) There is no simple antiderivative for this integrand, so we use numerical methods. We nd
P(1) =
1
_
1
0
e
t
2
dt = 0.421.
(b) To calculate this improper integral, use numerical methods. If you cannot input innity into your calculator, increase
the upper limit until the value of the integral settles down. We nd
P() =
1
_

0
e
t
2
dt = 0.500.
41. The factor ln x grows slowly enough (as x 0
+
) not to change the convergence or divergence of the integral, although
it will change what it converges or diverges to.
The integral is always improper, because ln x is not dened for x = 0. Integrating by parts (or, alternatively, the
integral table) yields
_
e
0
x
p
ln xdx = lim
a0
+
_
e
a
x
p
ln xdx
= lim
a0
+
_
1
p + 1
x
p+1
ln x
1
(p + 1)
2
x
p+1
_
e
a
= lim
a0
+
__
1
p + 1
e
p+1
1
(p + 1)
2
e
p+1
_
_
1
p + 1
a
p+1
ln a
1
(p + 1)
2
a
p+1
__
.
If p < 1, then (p + 1) is negative, so as a 0
+
, a
p+1
and ln a , and therefore the limit does not
exist.
If p > 1, then (p +1) is positive and its easy to see that a
p+1
0 as a 0. Looking at graphs of x
p+1
ln x (for
different values of p) shows that a
p+1
ln a 0 as a 0. This is not so easy to see analytically. Its true because if we
let t =
1
a
then
lim
a0
+
a
p+1
ln a = lim
t
_
1
t
_
p+1
ln
_
1
t
_
= lim
t
ln t
t
p+1
.
This last limit is zero because ln t grows very slowly, much more slowly than t
p+1
. So if p > 1, the integral converges
and equals e
p+1
[1/(p + 1) 1/(p + 1)
2
] = pe
p+1
/(p + 1)
2
.
What happens if p = 1? Then we get
_
e
0
ln x
x
dx = lim
a0
+
_
e
a
ln x
x
dx
= lim
a0
+
(ln x)
2
2
e
a
= lim
a0
+
_
1 (ln a)
2
2
_
.
Since ln a as a 0
+
, this limit does not exist.
To summarize,
_
e
0
x
p
ln x converges for p > 1 to the value pe
p+1
/(p + 1)
2
.
45. We calculate
m1 =
1
2
_

xe
x
2
/2
dx.
Since the integrand is odd, for any b, the integral
_
b
b
xe
x
2
/2
dx = 0.
Thus,
m1 =
1
2
_

xe
x
2
/2
dx =
1
2
lim
b
_
b
b
xe
x
2
/2
dx = 0.
49. (a)
(1) =
_

0
e
t
dt
= lim
b
_
b
0
e
t
dt
= lim
b
e
t
b
0
= lim
b
[1 e
b
] = 1.
Using Problem 11,
(2) =
_

0
te
t
dt = 1.
(b) We integrate by parts. Let u = t
n
, v
= e
t
. Then u
= nt
n1
and v = e
t
, so
_
t
n
e
t
dt = t
n
e
t
+ n
_
t
n1
e
t
dt.
So
(n + 1) =
_

0
t
n
e
t
dt
7.8 SOLUTIONS 157
= lim
b
_
b
0
t
n
e
t
dt
= lim
b
_
t
n
e
t
b
0
+ n
_
b
0
t
n1
e
t
dt
= lim
b
b
n
e
b
+ lim
b
n
_
b
0
t
n1
e
t
dt
= 0 + n
_

0
t
n1
e
t
dt
= n(n).
(c) We already have (1) = 1 and (2) = 1. Using (n + 1) = n(n) we can get
(3) = 2(2) = 2
(4) = 3(3) = 3 2
(5) = 4(4) = 4 3 2.
So it appears that (n) is just the rst n 1 numbers multiplied together, so
(n) = (n 1)!.
Exercises
1. For large x, the integrand behaves like 1/x
2
because
x
2
x
4
+ 1

x
2
x
4
=
1
x
2
.
Since
_

1
dx
x
2
converges, we expect our integral to converge. More precisely, since x
4
+ 1 > x
4
, we have
x
2
x
4
+ 1
<
x
2
x
4
=
1
x
2
.
Since
_

1
dx
x
2
is convergent, the comparison test tells us that
_

1
x
2
x
4
+ 1
dx converges also.
5. The integrand is continuous for all x 1, so whether the integral converges or diverges depends only on the behavior of
the function as x . As x , polynomials behave like the highest powered term. Thus, as x , the integrand
x
x
2
+ 2x + 4
behaves like
x
x
2
or
1
x
. Since
_

1
1
x
dx diverges, we predict that the given integral will diverge.
9. The integrand is continuous for all x 1, so whether the integral converges or diverges depends only on the behavior of
the function as x . As x , polynomials behave like the highest powered term. Thus, as x , the integrand
x
2
+ 4
x
4
+ 3x
2
+ 11
behaves like
x
2
x
4
or
1
x
2
. Since
_

1
1
x
2
dx converges, we predict that the given integral will converge.
13. The integrand is unbounded as t 5. We substitute w = t 5, so dw = dt. When t = 5, w = 0 and when t = 8,
w = 3.
_
8
5
6
t 5
dt =
_
3
0
6
w
dw.
Since
_
3
0
6
w
dw = lim
a0
+
6
_
3
a
1
w
dw = 6 lim
a0
+
2w
1/2
3
a
= 12 lim
a0
+
(
a) = 12
3,
our integral converges.
17. Since
1
u + u
2
<
1
u
2
for u 1, and since
_

1
du
u
2
converges,
_

1
du
u + u
2
converges.
21. Since
1
1 + e
y

1
e
y
= e
y
and
_

0
e
y
dy converges, the integral
_

0
dy
1 + e
y
converges.
25. Since
3 + sin
for 4, and since

_

4
2
d diverges, then
_

4
3 + sin
d diverges.
Problems
29. The convergence or divergence of an improper integral depends on the long-term behavior of the integrand, not on its
short-term behavior. Figure 7.13 suggests that g(x) f(x) for all values of x beyond x = k. Since
_
k
f(x) dx
converges, we expect
_
k
g(x) dx converges also.
However we are interested in
_
a
g(x) dx. Breaking the integral into two parts enables us to use the fact that
_
k
g(x) dx is nite:
_

a
g(x) dx =
_
k
a
g(x) dx +
_

k
g(x) dx.
The rst integral is also nite because the interval from a to k is nite. Therefore, we expect
_
a
g(x) dx converges.
a
k
f(x)
g(x)
x
Figure 7.13
33. (a) The tangent line to e
t
has slope (e
t
)
= e
t
. Thus at t = 0, the slope is e
0
= 1. The line passes through (0, e
0
) =
(0, 1). Thus the equation of the tangent line is y = 1 + t. Since e
t
is everywhere concave up, its graph is always
above the graph of any of its tangent lines; in particular, e
t
is always above the line y = 1 + t. This is tantamount to
saying
1 + t e
t
,
with equality holding only at the point of tangency, t = 0.
(b) If t =
1
x
, then the above inequality becomes
1 +
1
x
e
1/x
, or e
1/x
1
1
x
.
Since t =
1
x
, t is never zero. Therefore, the inequality is strict, and we write
e
1/x
1 >
1
x
.
(c) Since e
1/x
1 >
1
x
,
1
x
5
(e
1/x
1)
<
1
x
5
_
1
x
_ =
1
x
4
.
Since
_

1
dx
x
4
converges,
_

1
dx
x
5
(e
1/x
1)
converges.
SOLUTIONS to Review Problems for Chapter Seven 159
Exercises
1.
1
3
(t + 1)
3
5. Using the power rule gives
3
2
w
2
+ 7w + C.
9. Let 5z = w, then 5dz = dw, which means dz =
1
5
dw, so
_
e
5z
dz =
_
e
w
1
5
dw =
1
5
_
e
w
dw =
1
5
e
w
+ C =
1
5
e
5z
+ C,
where C is a constant.
13. The power rule gives
2
5
x
5/2
+
3
5
x
5/3
+ C
17. Dividing by x
2
gives
_ _
x
3
+ x + 1
x
2
_
dx =
_
_
x +
1
x
+
1
x
2
_
dx =
1
2
x
2
+ ln |x|
1
x
+ C.
21. Integration by parts twice gives
_
x
2
e
2x
dx =
x
2
e
2x
2

_
2xe
2x
dx =
x
2
2
e
2x
x
2
e
2x
+
1
4
e
2x
+ C
= (
1
2
x
2
1
2
x +
1
4
)e
2x
+ C.
Or use the integral table, III-14 with p(x) = x
2
and a = 1.
25. We integrate by parts, using u = (ln x)
2
and v
= 1. Then u
= 2
ln x
x
and v = x, so
_
(ln x)
2
dx = x(ln x)
2
2
_
ln xdx.
But, integrating by parts or using the integral table,
_
ln xdx = xln x x + C. Therefore,
_
(ln x)
2
dx = x(ln x)
2
2xln x + 2x + C.
Check:
d
dx
_
x(ln x)
2
2xln x + 2x + C
= (ln x)
2
+ x
2 ln x
x
2 ln x 2x
1
x
+ 2 = (ln x)
2
.
29. Substitute w = 4 x
2
, dw = 2xdx:
_
x
_
4 x
2
dx =
1
2
_

wdw =
1
3
w
3/2
+ C =
1
3
(4 x
2
)
3/2
+ C.
Check
d
dx
_
1
3
(4 x
2
)
3/2
+ C
_
=
1
3
_
3
2
(4 x
2
)
1/2
(2x)
_
= x
_
4 x
2
.
33. Denote
_
cos
2
d by A. Let u = cos , v
= cos . Then, v = sin and u
= sin . Integrating by parts, we get:

A = cos sin
_
(sin ) sin d.
Employing the identity sin
2
= 1 cos
2
, the equation above becomes:
A = cos sin +
_
d
_
cos
2
d
= cos sin + A + C.
Solving this equation for A, and using the identity sin 2 = 2 cos sin we get:
A =
_
cos
2
d =
1
4
sin 2 +
1
2
+ C.
[Note: An alternate solution would have been to use the identity cos
2
=
1
2
cos 2 +
1
2
.]
37. Multiplying out, dividing, and then integrating yields
_
(t + 2)
2
t
3
dt =
_
t
2
+ 4t + 4
t
3
dt =
_
1
t
dt +
_
4
t
2
dt +
_
4
t
3
dt = ln |t|
4
t

2
t
2
+ C,
41. Let cos = w, then sin d = dw, so
_
tan d =
_
sin
cos
d =
_
1
w
dw
= ln |w| + C = ln | cos | + C,
45. Let w = 2z, so dw = 2dz. Then, since
d
dw
arctan w =
1
1 + w
2
, we have
_
dz
1 + 4z
2
=
_
1
2
dw
1 + w
2
=
1
2
arctan w + C =
1
2
arctan 2z + C.
49. If u = t 10, t = u + 10 and dt = 1 du, so substituting we get
_
(u + 10)u
10
du =
_
(u
11
+ 10u
10
) du =
1
12
u
12
+
10
11
u
11
+ C
=
1
12
(t 10)
12
+
10
11
(t 10)
11
+ C.
53. Let x
2
= w, then 2xdx = dw, x = 1 w = 1, x = 3 w = 9. Thus,
_
3
1
x(x
2
+ 1)
70
dx =
_
9
1
(w + 1)
70
1
2
dw
=
1
2

1
71
(w + 1)
71
9
1
=
1
142
(10
71
2
71
).
57. Let w = ln x, then dw = (1/x)dx which gives
_
1
x
tan(ln x) dx =
_
tan wdw =
_
sin w
cos w
dw = ln(| cos w|) + C = ln(| cos(ln x)|) + C.
61. Dividing and then integrating term by term, we get
_
e
2y
+ 1
e
2y
dy =
_ _
e
2y
e
2y
+
1
e
2y
_
dy =
_
(1 + e
2y
) dy =
_
dy +
_
1
2
_
_
e
2y
(2) dy
= y
1
2
e
2y
+ C.
65. Let w =
x
2
+ 1, then dw =
xdx
x
2
+ 1
so that
_
x
x
2
+ 1
cos
_
x
2
+ 1 dx =
_
cos wdw = sin w + C = sin
_
x
2
+ 1 + C.
69.
_
e
2x+3
dx =
1
2
_
e
2x+3
2dx. If u =
2x + 3, du =
2dx, so
1
2
_
e
u
du =
1
2
e
u
+ C =
1
2
e
2x+3
+ C.
73. The integral table yields
_
5x + 6
x
2
+ 4
dx =
5
2
ln |x
2
+ 4| +
6
2
arctan
x
2
+ C
=
5
2
ln |x
2
+ 4| + 3 arctan
x
2
+ C.
Check:
d
dx
_
5
2
ln |x
2
+ 4| +
6
2
arctan
x
2
+ C
_
=
5
2
_
1
x
2
+ 4
(2x) + 3
1
1 + (x/2)
2
1
2
_
=
5x
x
2
+ 4
+
6
x
2
+ 4
=
5x + 6
x
2
+ 4
.
77. Integration by parts will be used twice. First let u = e
ct
and dv = sin(kt)dt, then du = ce
ct
dt and v =
(1/k) cos kt. Then
_
e
ct
sin kt dt =
1
k
e
ct
cos kt
c
k
_
e
ct
cos kt dt
=
1
k
e
ct
cos kt
c
k
_
1
k
e
ct
sin kt +
c
k
_
e
ct
sin kt dt
_
=
1
k
e
ct
cos kt
c
k
2
e
ct
sin kt
c
2
k
2
_
e
ct
sin kt dt
Solving for
_
e
ct
sin kt dt gives
k
2
+ c
2
k
2
_
e
ct
sin kt dt =
e
ct
k
2
(k cos kt + c sin kt) ,
so
_
e
ct
sin kt dt =
e
ct
k
2
+ c
2
(k cos kt + c sin kt) + C.
81. By completing the square, we get
x
2
3x + 2 = (x
2
3x + (
3
2
)
2
) + 2
9
4
= (x
3
2
)
2
1
4
.
Then
_
1
x
2
3x + 2
dx =
_
1
_
(x
3
2
)
2
1
4
dx.
Let w = (x (3/2)), then dw = dx and a
2
= 1/4. Then we have
_
1
x
2
3x + 2
dx =
_
1
w
2
a
2
dw
and from VI-29 of the integral table we have
_
1
w
2
a
2
dw = ln
w +
_
w
2
a
2
+ C
= ln
_
x
3
2
_
+
_
_
x
3
2
_
2
1
4
+ C
= ln
_
x
3
2
_
+
_
x
2
3x + 2
+ C.
85. Let w = ax
2
+ 2bx + c, then dw = (2ax + 2b)dx so that
_
ax + b
ax
2
+ 2bx + c
dx =
1
2
_
dw
w
=
1
2
ln |w| + C =
1
2
ln |ax
2
+ 2bx + c| + C.
89. Multiplying out and integrating term by term gives
_
(x
2
+ 5)
3
dx =
_
(x
6
+ 15x
4
+ 75x
2
+ 125)dx =
1
7
x
7
+ 15
x
5
5
+ 75
x
3
3
+ 125x + C
=
1
7
x
7
+ 3x
5
+ 25x
3
+ 125x + C.
93. If u = 1 + cos
2
w, du = 2(cos w)
1
(sin w) dw, so
_
sin wcos w
1 + cos
2
w
dw =
1
2
_
2 sin wcos w
1 + cos
2
w
dw =
1
2
_
1
u
du =
1
2
ln |u| + C
=
1
2
ln |1 + cos
2
w| + C.
97. If u =
x + 1, u
2
= x + 1 with x = u
2
1 and dx = 2udu. Substituting, we get
_
x
x + 1
dx =
_
(u
2
1)2udu
u
=
_
(u
2
1)2 du = 2
_
(u
2
1) du
=
2u
3
3
2u + C =
2(
x + 1)
3
3
2
x + 1 + C.
101. Letting u = z 5, z = u + 5, dz = du, and substituting, we have
_
z
(z 5)
3
dz =
_
u + 5
u
3
du =
_
(u
2
+ 5u
3
)du =
u
1
1
+ 5
_
u
2
2
_
+ C
=
1
(z 5)
+
5
2(z 5)
2
+ C.
105. Let w = y
2
2y + 1, so dw = 2(y 1) dy. Then
_
_
y
2
2y + 1(y 1) dy =
_
w
1/2
1
2
dw =
1
2

2
3
w
3/2
+ C =
1
3
(y
2
2y + 1)
3/2
+ C.
Alternatively, notice the integrand can be written
_
(y 1)
2
(y 1) = (y 1)
2
. This leads to a different-looking but
equivalent answer.
109. If u = 2 sin x, then du = 2 cos xdx, so
_
cos(2 sin x) cos xdx =
1
2
_
cos(2 sin x)2 cos xdx =
1
2
_
cos udu
=
1
2
sin u + C =
1
2
sin(2 sin x) + C.
113. We use the substitution w = x
2
+ 2x and dw = (2x + 2) dx so
_
(x + 1) sinh(x
2
+ 2x) dx =
1
2
_
sinh w dw =
1
2
cosh w + C =
1
2
cosh(x
2
+ 2x) + C.
Check this answer by taking the derivative:
d
dx
_
1
2
cosh(x
2
+ 2x) + C
_
=
1
2
(2x+2) sinh(x
2
+2x) = (x+1) sinh(x
2
+
2x).
117. Let w = 1 + 5x
2
. We have dw = 10xdx, so
dw
10
= xdx. When x = 0, w = 1. When x = 1, w = 6.
xdx
1 + 5x
2
=
_
6
1
1
10
dw
w
=
1
10
_
6
1
dw
w
=
1
10
ln |w|
6
1
=
1
10
(ln 6 ln 1) =
ln 6
10
121. Integrating by parts, we take u = e
2x
, u
= 2e
2x
, v
= sin 2x, and v =

1
2
cos 2x, so
_
e
2x
sin 2xdx =
e
2x
2
cos 2x +
_
e
2x
cos 2xdx.
Integrating by parts again, with u = e
2x
, u
= 2e
2x
, v
= cos 2x, and v =

1
2
sin 2x, we get
_
e
2x
cos 2xdx =
e
2x
2
sin 2x
_
e
2x
sin 2xdx.
Substituting into the previous equation, we obtain
_
e
2x
sin 2xdx =
e
2x
2
cos 2x +
e
2x
2
sin 2x
_
e
2x
sin 2xdx.
Solving for
_
e
2x
sin 2xdx gives
_
e
2x
sin 2xdx =
1
4
e
2x
(sin 2x cos 2x) + C.
This result can also be obtained using II-8 in the integral table. Thus
_

e
2x
sin 2x = [
1
4
e
2x
(sin 2x cos 2x)]
=
1
4
(e
2
e
2
) 133.8724.
We get 133.37 using Simpsons rule with 10 intervals. With 100 intervals, we get 133.8724. Thus our answer matches
the approximation of Simpsons rule.
125.
_
1
0
dx
x
2
+ 1
= tan
1
x
1
0
= tan
1
1 tan
1
0 =

4
0 =

4
.
129. (a) We split
1
x
2
x
=
1
x(x 1)
into partial fractions:
1
x
2
x
=
A
x
+
B
x 1
.
Multiplying by x(x 1) gives
1 = A(x 1) + Bx = (A + B)x A,
so A = 1 and A + B = 0, giving A = 1, B = 1. Therefore,
_
1
x
2
x
dx =
_
_
1
x 1

1
x
_
dx = ln |x 1| ln |x| + C.
(b)
_
1
x
2
x
dx =
_
1
(x 1)(x)
dx. Using a = 1 and b = 0 in V-26, we get ln |x 1| ln |x| + C.
133. Splitting the integrand into partial fractions with denominators x and (x + 5), we have
1
x(x + 5)
=
A
x
+
B
x + 5
.
Multiplying by x(x + 5) gives the identity
1 = A(x + 5) + Bx
so
1 = (A + B)x + 5A.
Since this equation holds for all x, the constant terms on both sides must be equal. Similarly, the coefcient of x on both
sides must be equal. So
5A = 1
A + B = 0.
Solving these equations gives A = 1/5, B = 1/5 and the integral becomes
_
1
x(x + 5)
dx =
1
5
_
1
x
dx
1
5
_
1
x + 5
dx =
1
5
(ln |x| ln |x + 5|) + C.
137. Splitting the integrand into partial fractions with denominators (1 + x), (1 + x)
2
and x, we have
1 + x
2
x(1 + x)
2
=
A
1 + x
+
B
(1 + x)
2
+
C
x
.
Multiplying by x(1 + x)
2
gives the identity
1 + x
2
= Ax(1 + x) + Bx + C(1 + x)
2
so
1 + x
2
= (A + C)x
2
+ (A + B + 2C)x + C.
Since this equation holds for all x, the constant terms on both sides must be equal. Similarly, the coefcient of x and x
2
on both sides must be equal. So
C = 1
A + B + 2C = 0
A + C = 1.
Solving these equations gives A = 0, B = 2 and C = 1. The integral becomes
_
1 + x
2
(1 + x)
2
x
dx = 2
_
1
(1 + x)
2
dx +
_
1
x
dx =
2
1 + x
+ ln |x| + K.
141. Using the substitution w = sin x, we get dw = cos xdx, so we have
_
cos x
sin
3
x + sin x
dx =
_
dw
w
3
+ w
.
But
1
w
3
+ w
=
1
w(w
2
+ 1)
=
1
w

w
w
2
+ 1
,
so
_
cos x
sin
3
x + sin x
dx =
_
_
1
w

w
w
2
+ 1
_
dw
= ln |w|
1
2
ln
w
2
+ 1
+ C
= ln |sin x|
1
2
ln
sin
2
x + 1
+ C.
145. To nd
_
we
w
dw, integrate by parts, with u = w and v
= e
w
. Then u
= 1 and v = e
w
.
Then
_
we
w
dw = we
w
+
_
e
w
dw = we
w
e
w
+ C.
Thus
_

0
we
w
dw = lim
b
_
b
0
we
w
dw = lim
b
(we
w
e
w
)
b
0
= 1.
149. We nd the exact value:
_

10
1
z
2
4
dz =
_

10
1
(z + 2)(z 2)
dz
= lim
b
_
b
10
1
(z + 2)(z 2)
dz
= lim
b
1
4
(ln |z 2| ln |z + 2|)
b
10
=
1
4
lim
b
[(ln |b 2| ln |b + 2|) (ln 8 ln 12)]
=
1
4
lim
b
__
ln
b 2
b + 2
_
+ ln
3
2
_
=
1
4
(ln 1 + ln 3/2) =
ln 3/2
4
.
153. The integrand
x
x + 1
1 as x , so theres no way
_

1
x
x + 1
dx can converge.
Problems
157. Since (e
x
)
2
= e
2x
, we have
Area =
_
1
0
(e
x
)
2
dx =
_
1
0
e
2x
dx =
1
2
e
2x
1
0
=
1
2
(e
2
1).
161. In the interval of 0 x 2, the equation y =
4 x
2
represents one quadrant of the radius 2 circle centered at the
origin. Interpreting the integral as an area gives a value of
1
4
2
2
= .
165. First solution: After the substitution w = x + 1, the rst integral becomes
_
w 1
w
dw = w
_
w
1
dw.
With this same substitution, the second integral becomes
_
w
1
dw.
Second solution: We note that the sum of the integrands is 1, so the sum of the integrals is x. Thus
_
x
x + 1
dx = x
_
1
x + 1
dx.
169. (a) (i) Multiplying out gives
_
(x
2
+ 10x + 25) dx =
x
3
3
+ 5x
2
+ 25x + C.
(ii) Substituting w = x + 5, so dw = dx, gives
_
(x + 5)
2
dx =
_
w
2
dw =
w
3
3
+ C =
(x + 5)
3
3
+ C.
(b) The results of the two calculations are not the same since
(x + 5)
3
3
+ C =
x
3
3
+
15x
2
3
+
75x
3
+
125
3
+ C.
However they differ only by a constant, 125/3, as guaranteed by the Fundamental Theorem of Calculus.
173. (a) i. 0 ii.
2
iii.
1
2
(b) Average value of f(t) < Average value of k(t) < Average value of g(t)
We can look at the three functions in the range
2
x
3
2
, since they all have periods of 2 (| cos t|
and (cos t)
2
also have a period of , but that does not hurt our calculation). It is clear from the graphs of the three
functions below that the average value for cos t is 0 (since the area above the x-axis is equal to the area below it),
while the average values for the other two are positive (since they are everywhere positive, except where they are 0).
2

3
2
f(t)
t
2
3
2
f(t)
t
2
3
2
f(t)
t
It is also fairly clear from the graphs that the average value of g(t) is greater than the average value of k(t); it is
also possible to see this algebraically, since
(cos t)
2
= | cos t|
2
| cos t|
because | cos t| 1 (and both of these s are <s at all the points where the functions are not 0 or 1).
177. (a) f(x) = 1 + e
x
is concave up for 0 x 0.5, so trapezoids will overestimate
_
0.5
0
f(x)dx, and the midpoint
rule will underestimate.
(b) f(x) = e
x
2
is concave down for 0 x 0.5, so trapezoids will underestimate
_
0.5
0
f(x)dx and midpoint will
overestimate the integral.
(c) Both the trapezoid rule and the midpoint rule will give the exact value of the integral. Note that upper and lower sums
will not, unless the line is horizontal.
181. If I(t) is average per capita income t years after 2005, then I
(t) = r(t).
(a) Since t = 10 in 2015, by the Fundamental Theorem,
I(10) I(0) =
_
10
0
r(t) dt =
_
10
0
1556.37e
0.045t
dt =
1556.37
0.045
e
0.045t
10
0
= 19,655.65 dollars.
so I(10) = 34,586 + 19,655.65 = 54,241.65.
(b) We have
I(t) I(0) =
_
t
0
r(t) dt =
_
t
0
1556.37e
0.045t
dt =
1556.37
0.045
e
0.045t
t
0
= 34,586
_
e
0.045t
1
_
.
Thus, since I(0) = 34,586,
I(t) = 34,586 + 34,586(e
0.045t
1) = 34,586e
0.045t
dollars.
185. We want to calculate
_
1
0
Cn sin(nx) Cm sin(mx) dx.
We use II-11 from the table of integrals with a = n, b = m. Since n = m, we see that
_
1
0
n(x) m(x) dx = CnCm
_
1
0
sin(nx) sin(mx) dx
=
CnCm
m
2
2
n
2
2
(n cos(nx) sin(mx) m sin(nx) cos(mx))
1
0
=
CnCm
(m
2
n
2
)
2
_
n cos(n) sin(m) m sin(n) cos(m)
n cos(0) sin(0) + m sin(0) cos(0)
_
= 0
since sin(0) = sin(n) = sin(m) = 0.
189. (a) A possible answer is
_
sin xcos xcos(2x) dx =
cos(4x)
16
.
Different systems may give the answer in a different form.
(b)
d
dx
_
cos(4x)
16
_
=
sin(4x)
4
.
(c) Using the double angle formula sin 2A = 2 sin Acos A twice, we get
sin(4x)
4
=
2 sin(2x) cos(2x)
4
=
2 2 sin xcos xcos(2x)
4
= sin xcos xcos(2x).
1. True. Let w = f(x), so dw = f
(x) dx, then

_
f
(x) cos(f(x)) dx =
_
cos wdw = sin w + C = sin(f(x)) + C.
5. False. Completing the square gives
_
dx
x
2
+ 4x + 5
=
_
dx
(x + 2)
2
+ 1
= arctan(x + 2) + C.
9. True. y
2
1 is concave up, and the midpoint rule always underestimates for a function that is concave up.
13. False. Let f(x) = 1/(x + 1). Then
_

0
1
x + 1
dx = lim
b
ln |x + 1|
b
0
= lim
b
ln(b + 1),
but lim
b
ln(b + 1) does not exist.
17. True. By properties of integrals and limits,
lim
b
_
b
0
af(x) dx = a lim
b
_
b
0
f(x) dx.
Thus, the limit on the left of the equation is nite exactly when the limit on the right side of the equation is nite. Thus
_
0
af(x) dx converges if
_
0
f(x) dx converges.
21. False. The subdivision size x = (1/10)(6 2) = 4/10.
25. True. We have
LEFT(n) RIGHT(n) = (f(x0) + f(x1) + + f(xn1))x (f(x1) + f(x2) + + f(xn))x.
On the right side of the equation, all terms cancel except the rst and last, so:
LEFT(n) RIGHT(n) = (f(x0) f(xn))x = (f(2) f(6))x.
This is also discussed in Section 5.1.
29. False. This is true if f is an increasing function or if f is a decreasing function, but it is not true in general. For example,
suppose that f(2) = f(6). Then LEFT(n) = RIGHT(n) for all n, which means that if
_
6
2
f(x)dx lies between LEFT(n)
and RIGHT(n), then it must equal LEFT(n), which is not always the case.
For example, if f(x) = (x 4)
2
and n = 1, then f(2) = f(6) = 4, so
LEFT(1) = RIGHT(1) = 4 (6 2) = 16.
However
_
6
2
(x 4)
2
dx =
(x 4)
3
3
6
2
=
2
3
3

_
2
3
3
_
=
16
3
.
In this example, since LEFT(n) = RIGHT(n), we have TRAP(n) = LEFT(n). However trapezoids overestimate the
area, since the graph of f is concave up. This is also discussed in Section 7.5.
8.1 SOLUTIONS 169
CHAPTER EIGHT
Exercises
1. Each strip is a rectangle of length 3 and width x, so
Area of strip = 3x, so
Area of region =
_
5
0
3 dx = 3x
5
0
= 15.
Check: This area can also be computed using Length Width = 5 3 = 15.
5. The strip has width y, so the variable of integration is y. The length of the strip is x. Since x
2
+y
2
= 10 and the region
is in the rst quadrant, solving for x gives x =
_
10 y
2
. Thus
Area of strip xy =
_
10 y
2
dy.
The region stretches from y = 0 to y =
10, so
Area of region =
_

10
0
_
10 y
2
dy.
Evaluating using VI-30 from the Table of Integrals, we have
Area =
1
2
_
y
_
10 y
2
+ 10 arcsin
_
y
10
__
10
0
= 5(arcsin 1 arcsin 0) =
5
2
.
Check: This area can also be computed using the formula
1
4
r
2
=
1
4
(
10)
2
=
5
2
.
9. Each slice is a circular disk with radius r = 2 cm.
Volume of disk = r
2
x = 4x cm
3
.
Summing over all disks, we have
Total volume
4x cm
3
.
Taking a limit as x 0, we get
Total volume = lim
x0
4x =
_
9
0
4 dx cm
3
.
Evaluating gives
Total volume = 4x
9
0
= 36 cm
3
.
Check: The volume of the cylinder can also be calculated using the formula V = r
2
h = 2
2
9 = 36 cm
3
.
13. Each slice is a circular disk. See Figure 8.1. The radius of the sphere is 5 mm, and the radius r at height y is given by the
Pythagorean Theorem
y
2
+r
2
= 5
2
.
Solving gives r =
_
5
2
y
2
mm. Thus,
Volume of disk r
2
y = (5
2
y
2
)y mm
3
.
170 Chapter Eight /SOLUTIONS
Summing over all disks, we have
Total volume
(5
2
y
2
)y mm
3
.
Taking the limit as y 0, we get
Total volume = lim
y0
(5
2
y
2
)y =
_
5
0
(5
2
y
2
) dy mm
3
.
Evaluating gives
Total volume =
_
25y
y
3
3
_
5
0
=
250
3
mm
3
.
Check: The volume of a hemisphere can be calculated using the formula V =
2
3
r
3
=
2
3
5
3
=
250
3
mm
3
.
y
5
r
Figure 8.1
Problems
17. Quarter circle of radius r =
15. See Figure 8.2.
15 h
2
E '
h
E '
h
h
y
Figure 8.2
3
T
c
y
6
3 y/2
y
Figure 8.3
21. Cone with height 6 and radius 3. See Figure 8.3.
25.
x
y
Radius =
r(hy)
h
This cone is what you get when you rotate the line x = r(h
y)/h about the yaxis. So slicing perpendicular to the yaxis
yields
V =
_
y=h
y=0
x
2
dy =
_
h
0
_
(h y)r
h
_
2
dy
=
r
2
h
2
_
h
0
(h
2
2hy +y
2
) dy
=
r
2
h
2
_
h
2
y hy
2
+
y
3
3
_
h
0
=
r
2
h
3
.
8.2 SOLUTIONS 171
Exercises
1. The volume is given by
V =
_
1
0
y
2
dx =
_
1
0
x
4
dx =
x
5
5
1
0
=

5
.
5. The volume is given by
V =
_
1
1
y
2
dx =
_
1
1
(e
x
)
2
dx =
_
1
1
e
2x
dx =

2
e
2x
1
1
=

2
(e
2
e
2
).
9. Since the graph of y = x
2
is below the graph of y = x for 0 x 1, the volume is given by
V =
_
1
0
x
2
dx
_
1
0
(x
2
)
2
dx =
_
1
0
(x
2
x
4
) dx =
_
x
3
3

x
5
5
_
1
0
=
2
15
.
13. Since f
(x) = 1/(x + 1), we evaluate the integral numerically to get

Arc length =
_
2
0
_
1 +
_
1
x + 1
_
2
dx = 2.302.
17. The length is
_
2
1
_
(x
(t))
2
+ (y
(t))
2
+ (z
(t))
2
dt =
_
2
1
_
5
2
+ 4
2
+ (1)
2
dt =
42.
This is the length of a straight line from the point (8, 5, 2) to (13, 9, 1).
Problems
21. The two functions intersect at (0, 0) and (8, 2). We slice the volume with planes perpendicular to the line x = 9. This
divides the solid into thin washers with volume
Volume of slice = r
2
out
y r
2
in
y.
The outer radius is the horizontal distance from the line x = 9 to the curve x = y
3
, so rout = 9 y
3
. Similarly, the inner
radius is the horizontal distance from the line x = 9 to the curve x = 4y, so rin = 9 4y. Integrating from y = 0 to
y = 2 we have
V =
_
2
0
[(9 y
3
)
2
(9 4y)
2
] dy.
25. One arch of the sine curve lies between x = /2 and x = /2. Since d(cos x)/dx = sin x, evaluating the integral
numerically gives
Arc length =
_
/2
/2
_
1 + sin
2
xdx = 3.820.
29.
x
y
(y = 1)
Radius = 1 + x
3
We slice the region perpendicular to the xaxis. The Rie-
mann sum we get is
(x
3
+ 1)
2
x. So the volume V is
the integral
V =
_
1
1
(x
3
+ 1)
2
dx
=
_
1
1
(x
6
+ 2x
3
+ 1) dx
=
_
x
7
7
+
x
4
2
+x
_
1
1
= (16/7) 7.18.
33. Slice the object into rings vertically, as is Figure 8.4. A typical ring has thickness x and outer radius y = 1 and inner
radius y = x
2
.
Volume of slice 1
2
x y
2
x = (1 x
4
) x.
Volume of solid = lim
x0
(1 x
4
) x =
_
1
0
(1 x
4
) dx =
_
x
x
5
5
_
1
0
=
4
5
.
1
1
1
E'
x
Inner radius
T
Outer radius
T
y = x
2
x
y
Figure 8.4: Cross-section of solid
1
1
Tc
y
'
Base of triangle (standing on paper)
E '
x
y = x
2
x
y
Figure 8.5: Base of solid
37. An equilateral triangle of side s has height
3s/2 and
Area =
1
2
s
3s
2
=
3
4
s
2
.
Slicing perpendicularly to the y-axis gives equilateral triangles whose thickness is y and whose side is x =

y.
See Figure 8.5. Thus
Volume of triangular slice
3
4
(
y)
2
y =
3
4
y y.
Volume of solid =
_
1
0
3
4
y dy =
3
4
y
2
2
1
0
=
3
8
.
41. We now slice perpendicular to the x-axis. As stated in the problem, the cross-sections obtained thereby will be squares,
with base length e
x
. The volume of one square slice is (e
x
)
2
dx. (See Figure 8.6.) Adding the volumes of the slices yields
Volume =
_
x=1
x=0
y
2
dx =
_
1
0
e
2x
dx =
e
2x
2
1
0
=
e
2
1
2
= 3.195.
x
y
z
'
'
e
x
'
'
e
x
Figure 8.6
8.2 SOLUTIONS 173
45. If y = e
x
2
/2
, then x =
2 ln y. (Note that since 0 < y 1, ln y 0.) A typical slice has thickness y and radius
x. See Figure 8.7. So
Volume of slice = x
2
y = 2 ln y y.
Thus,
Total volume = 2
_
1
0
ln y dy.
Since ln y is not dened at y = 0, this is an improper integral:
Total Volume = 2
_
1
0
ln y dy = 2 lim
a0
_
1
a
ln y dy
= 2 lim
a0
(y ln y y)
1
a
= 2 lim
a0
(1 a ln a +a).
By looking at the graph of xln x on a calculator, we see that lim
a0
a ln a = 0. Thus,
Total volume = 2(1) = 2.
E
x
y
Figure 8.7
49. We can nd the volume of the tree by slicing it into a series of thin horizontal cylinders of height dh and circumference
C. The volume of each cylindrical disk will then be
V = r
2
dh =
_
C
2
_
2
dh =
C
2
dh
4
.
Summing all such cylinders, we have the total volume of the tree as
Total volume =
1
4
_
120
0
C
2
dh.
We can estimate this volume using a trapezoidal approximation to the integral with h = 20:
LEFT estimate =
1
4
[20(31
2
+ 28
2
+ 21
2
+ 17
2
+ 12
2
+ 8
2
)] =
1
4
(53660).
RIGHT estimate =
1
4
[20(28
2
+ 21
2
+ 17
2
+ 12
2
+ 8
2
+ 2
2
)] =
1
4
(34520).
TRAP =
1
4
(44090) 3509 cubic inches.
53. As can be seen in Figure 8.8, the region has three straight sides and one curved one. The lengths of the straight sides are
1, 1, and e. The curved side is given by the equation y = f(x) = e
x
. We can nd its length by the formula
_
1
0
_
1 +f
(x)
2
dx =
_
1
0
_
1 + (e
x
)
2
dx =
_
1
0
_
1 +e
2x
dx.
Evaluating the integral numerically gives 2.0035. The total length, therefore, is about 1 + 1 +e + 2.0035 6.722.
1
1
f(x) = e
x
x
y
Figure 8.8
57. (a) For n = 1, we have
|x| +|y| = 1.
In the rst quadrant, the equation is the line
x +y = 1.
By symmetry, the graph in the other quadrants gives the square in Figure 8.9.
For n = 2, the equation is of a circle of radius 1, centered at the origin:
x
2
+y
2
= 1.
For n = 4, the equation is
x
4
+y
4
= 1.
The graph is similar to a circle, but bulging out more. See Figure 8.9.
(b) For n = 1, the arc length is the perimeter of the square. Each side is the hypotenuse of a right triangle of sides
1, 1,
2. Thus
Arc length = 4
2 = 5.657.
For n = 2, the arc length is the perimeter of the circle of radius 1. Thus
Arc length = 2 1 = 2 = 6.283.
For n = 4, we nd the arc length using the formula
L =
_
b
a
_
1 + (f
(x))
2
dx.
We nd the arc length of the top half of the curve, given by y = (1 x
4
)
1/4
, and double it. Since
dy
dx
=
1
4
(1 x
4
)
3/4
(4x
3
) =
x
3
(1 x
4
)
3/4
,
Arc length = 2
_
1
1
1 +
_
x
3
(1 x
4
)
3/4
_
2
dx = 2
_
1
1
_
1 +
x
6
(1 x
4
)
3/2
dx.
The integral is improper because the integral is not dened at x = 1. Using numerical methods, we nd
Arc length = 2
_
1
1
_
1 +
x
6
(1 x
4
)
3/2
dx = 7.018.
1 1
1
1
'
|x| +|y| = 1
'
x
2
+ y
2
= 1
'
x
4
+ y
4
= 1
x
y
Figure 8.9
8.3 SOLUTIONS 175
Exercises
1. With r = 1 and = 2/3, we nd x = r cos = 1 cos(2/3) = 1/2 and y = r sin = 1 sin(2/3) =
3/2.
The rectangular coordinates are (1/2,
3/2).
5. With x = 1 and y = 1, nd r from r =
_
x
2
+y
2
=
1
2
+ 1
2
=
2. Find from tan = y/x = 1/1 = 1. Thus,

= tan
1
(1) = /4. Since (1, 1) is in the rst quadrant this is a correct . The polar coordinates are (
2, /4).
9. (a) Table 8.1 contains values of r = 1 sin , both exact and rounded to one decimal.
Table 8.1
0 /3 /2 2/3 4/3 3/2 5/3 2 7/3 5/2 8/3
r 1 1
3/2 0 1
3/2 1 1 +
3/2 2 1 +
3/2 1 1
3/2 0 1
3/2
r 1 0.134 0 0.134 1 1.866 2 1.866 1 0.134 0 0.134
2 1 1 2
2
1
1
2
x
y
Figure 8.10
2 1 1 2
2
1
1
2
x
y
Figure 8.11
(c) The circle has equation r = 1/2. The cardioid is r = 1 sin . Solving these two simultaneously gives
1/2 = 1 sin ,
or
sin = 1/2.
Thus, = /6 or 5/6. This gives the points (x, y) = ((1/2) cos /6, (1/2) sin /6) = (
3/4, 1/4) and (x, y) =

((1/2) cos 5/6, (1/2) sin 5/6) = (
3/4, 1/4) as the location of intersection.

(d) The curve r = 1 sin 2, pictured in Figure 8.11, has two regions instead of the one region that r = 1 sin has.
This is because 1 sin 2 will be 0 twice for every 2 cycle in , as opposed to once for every 2 cycle in for
1 sin .
13. See Figures 8.12 and 8.13. The rst curve will be similar to the second curve, except the cardioid (heart) will be rotated
clockwise by 90
(/2 radians). This makes sense because of the identity sin = cos( /2).
Figure 8.12: r = 1 cos Figure 8.13: r = 1 sin
17. The region is given by
8 r
18 and /4 /2.
21. We have cos = x/
_
x
2
+y
2
for each point. By denition cos
1
(x/
_
x
2
+y
2
) is an angle between 0 and , in
quadrant I or II, which matches points A and B. The formula is valid for A and B.
Problems
25. The region between the spirals is shaded in Figure 8.14.
Area =
1
2
_
2
0
((2)
2
2
) d =
1
2
_
2
0
3
2
d =
1
2
2
0
= 4
3
.
2 4
x
y
Figure 8.14: Region between the
inner spiral, r = , and the outer
spiral, r = 2
1
2
1
2
r = 1 sin
r = 1/2
/6
x
y
Figure 8.15
29. The two curves intersect where
1 sin =
1
2
sin =
1
2
=

6
,
5
6
.
See Figure 8.15. We nd the area of the right half and multiply that answer by 2 to get the entire area. The integrals can be
computed numerically with a calculator or, as we show, using integration by parts or formula IV-17 in the integral tables.
Area of right half =
1
2
_
/6
/2
_
(1 sin )
2
_
1
2
_
2
_
d
=
1
2
_
/6
/2
_
1 2 sin + sin
2

1
4
_
d
=
1
2
_
/6
/2
_
3
4
2 sin + sin
2
_
d
=
1
2
_
3
4
+ 2 cos
1
2
sin cos +
1
2
/6
/2
=
1
2
_
5
6
+
7
3
8
_
.
Thus,
Total area =
5
6
+
7
3
8
.
8.3 SOLUTIONS 177
(b) The curves intersect when r
2
= 2
4 cos 2 = 2
cos 2 =
1
2
.
In the rst quadrant:
2 =

3
so =

6
.
Using symmetry, the area in the rst quadrant can be multiplied by 4 to nd the area of the total bounded region.
Area = 4
_
1
2
_
_
/6
0
(4 cos 2 2) d
= 2
_
4 sin 2
2
2
_
/6
0
= 4 sin

3

2
3
= 4
3
2

2
3
= 2
3
2
3
= 1.370.
2 2
1.5
1.5
r =
2
r
2
= 4 cos 2
x
y
Figure 8.16
37. (a) Expressing x and y parametrically in terms of , we have
x = r cos =
cos
and y = r sin =
sin
.
The slope of the tangent line is given by
dy
dx
=
dy/d
dx/d
=
_
cos sin
2
___
sin cos
2
_
=
sin cos
cos + sin
.
At = /2, we have
dy
dx
=/2
=
1 (/2)0
0 + (/2)1
=
2
.
At = /2, we have x = 0, y = 2/, so the equation of the tangent line is
y =
2
x +
2
.
(b) As 0,
x =
cos
and y =
sin
1.
Thus, y = 1 is a horizontal asymptote. See Figure 8.17.
y = 1
r = 1/
x
y
Figure 8.17
41. Parameterized by , the curve r = f() is given by x = f() cos and y = f() sin . Then
Arc length =
_

_
_
dx
d
_
2
+
_
dy
d
_
2
d
=
_

_
(f
() cos f() sin )

2
+ (f
() sin +f() cos )

2
d
=
_

_
(f
())
2
cos
2
2f
()f() cos sin + (f())

2
sin
2
+(f
())
2
sin
2
+ 2f
()f() sin cos + (f())

2
cos
2
d
=
_

_
(f
())
2
(cos
2
+ sin
2
) + (f())
2
(sin
2
+ cos
2
) d
=
_

_
(f
())
2
+ (f())
2
d.
Exercises
1. Since density is e
x
gm/cm,
Mass =
_
10
0
e
x
dx = e
x
10
0
= 1 e
10
gm.
5. (a) Figure 8.18 shows a graph of the density function.
20
300
900
x
Figure 8.18
(b) Suppose we choose an x, 0 x 20. We approximate the density of the number of the cars between x and x +x
miles as (x) cars per mile. Therefore, the number of cars between x and x + x is approximately (x)x. If we
slice the 20 mile strip into N slices, we get that the total number of cars is
C
N
i=1
(xi)x =
N
i=1
_
600 + 300 sin(4
xi + 0.15)
x,
where x = 20/N. (This is a right-hand approximation; the corresponding left-hand approximation is
N1
i=0
(xi)x.)
8.4 SOLUTIONS 179
(c) As N , the Riemann sum above approaches the integral
C =
_
20
0
(600 + 300 sin 4
x + 0.15) dx.
If we calculate the integral numerically, we nd C 11513. We can also nd the integral exactly as follows:
C =
_
20
0
(600 + 300 sin 4
x + 0.15) dx
=
_
20
0
600 dx +
_
20
0
300 sin 4
x + 0.15 dx
= 12000 + 300
_
20
0
sin 4
x + 0.15 dx.
Let w =
x + 0.15, so x = w
2
0.15 and dx = 2wdw. Then
_
x=20
x=0
sin 4
x + 0.15 dx = 2
_
w=
20.15
w=
0.15
wsin 4wdw, (using integral table III-15)
= 2
_
1
4
wcos 4w +
1
16
sin 4w
_
20.15
0.15
1.624.
Using this, we have C 12000 + 300(1.624) 11513, which matches our numerical approximation.
9. We slice the block horizontally. Aslice has area 103 = 30 and thickness z. On such a slice, the density is approximately
constant. Thus
Mass of slice Density Volume (2 z) 30z,
so we have
Mass of block
(2 z)30z.
In the limit as z 0, the sum becomes an integral and the approximation becomes exact. Thus
Mass of block =
_
1
0
(2 z)30 dz = 30
_
2z
z
2
2
_
1
0
= 30
_
2
1
2
_
= 45.
Problems
13. (a) We form a Riemann sum by slicing the region into concentric rings of radius r and width r. Then the volume
deposited on one ring will be the height H(r) multiplied by the area of the ring. A ring of width r will have an area
given by
Area = (r + r)
2
(r
2
)
= (r
2
+ 2rr + (r)
2
r
2
)
= (2rr + (r)
2
).
Since r is approaching zero, we can approximate
Area of ring (2rr + 0) = 2rr.
From this, we have
V H(r) 2rr.
Thus, summing the contributions from all rings we have
V
H(r) 2rr.
Taking the limit as r 0, we get
V =
_
5
0
2r
_
0.115e
2r
_
dr.
(b) We use integration by parts:
V = 0.23
_
5
0
_
re
2r
_
dr
= 0.23
_
re
2r
2

e
2r
4
_
5
0
0.181(millimeters) (kilometers)
2
= 0.181 10
3
10
6
meters
3
= 181 cubic meters.
17. We slice time into small intervals. Since t is given in seconds, we convert the minute to 60 seconds. We consider water
loss over the time interval 0 t 60. We also need to convert inches into feet since the velocity is given in ft/sec. Since
1 inch = 1/12 foot, the square hole has area 1/144 square feet. For water owing through a hole with constant velocity v,
the amount of water which has passed through in some time, t, can be pictured as the rectangular solid in Figure 8.19,
which has volume
Area Height = Area Velocity Time.
ccc
Area =
1
144
ft
2
c
Height = g(t)t
E
Figure 8.19: Volume of water passing
through hole
Over a small time interval of length t, starting at time t, water ows with a nearly constant velocity v = g(t)
through a hole 1/144 square feet in area. In t seconds, we know that
Water lost
_
1
144
ft
2
_
(g(t) ft/sec)(t sec) =
_
1
144
_
g(t) t ft
3
.
Adding the water from all subintervals gives
Total water lost
1
144
g(t) t ft
3
.
As t 0, the sum tends to the denite integral:
Total water lost =
_
60
0
1
144
g(t) dt ft
3
.
21. The center of mass is
x =
_
0
x(2 + sin x) dx
_
0
(2 + sin x) dx
.
The numerator is
_
0
(2x +xsin x) dx = (x
2
xcos x + sin x)
0
=
2
+.
The denominator is
_
0
(2 + sin x) dx = (2x cos x)
0
= 2 + 2. So the center of mass is at
x =

2
+
2 + 2
=
( + 1)
2( + 1)
=

2
.
8.4 SOLUTIONS 181
25. (a) Since the density is constant, the mass is the product of the area of the plate and its density.
Area of the plate =
_
1
0
x
2
dx =
1
3
x
3
1
0
=
1
3
cm
2
.
Thus the mass of the plate is 2 1/3 = 2/3 gm.
(b) See Figure 8.20. Since the region is fatter closer to x = 1, x is greater than 1/2.
E '
x
1
x
2
x
y
Figure 8.20
(c) To nd the center of mass along the x-axis, we slice the region into vertical strips of width x. See Figure 8.20. Then
Area of strip = Ax(x)x x
2
x
Then, since the density is 2 gm/cm
2
, we have
x =
_
1
0
2x
3
dx
2/3
=
3
2

2x
4
4
1
0
= 3
_
1
4
_
=
3
4
cm.
This is greater than 1/2, as predicted in part (b).
29. Since the density is constant, the total mass of the solid is the product of the volume of the solid and its density: (1
e
2
)/2. By symmetry, y = 0. To nd x, we slice the solid into disks of width x, perpendicular to the x-axis. See
Figure 8.21. A disk at x has radius y = e
x
, so
Volume of disk = Ax(x)x = y
2
x = e
2x
x.
Since the density is , we have
x =
_
1
0
x e
2x
dx
Total mass
=
_
1
0
xe
2x
dx
(1 e
2
)/2
=
2
1 e
2
_
1
0
xe
2x
dx.
The integral
_
xe
2x
dx can be done by parts: let u = x and v
= e
2x
. Then u
= 1 and v = e
2x
/(2). So
_
xe
2x
dx =
xe
2x
2

_
e
2x
2
dx =
xe
2x
2

e
2x
4
.
and then
_
1
0
xe
2x
dx =
_
xe
2x
2

e
2x
4
_
1
0
=
_
e
2
2

e
2
4
_
_
0
1
4
_
=
1 3e
2
4
.
The nal result is:
x =
2
1 e
2

1 3e
2
4
=
1 3e
2
2 2e
2
0.343.
Notice that x is less that 1/2, as we would expect from the fact that the solid is wider near the origin.
x
y
y = e
x
E '
x
Radius = y
Figure 8.21
Exercises
1. The work done is given by
W =
_
2
1
3xdx =
3
2
x
2
2
1
=
9
2
joules.
5. The force exerted on the satellite by the earth (and vice versa!) is GMm/r
2
, where r is the distance from the center of
the earth to the center of the satellite, m is the mass of the satellite, M is the mass of the earth, and G is the gravitational
constant. So the total work done is
_
8.410
6
6.410
6
F dr =
_
8.410
6
6.410
6
GMm
r
2
dr =
_
GMm
r
_
8.410
6
6.410
6
1.489 10
10
joules.
Problems
9. The bucket moves upward at 40/10 = 4 meters/minute. If time is in minutes, at time t the bucket is at a height of x = 4t
meters above the ground. See Figure 8.22.
The water drips out at a rate of 5/10 = 0.5 kg/minute. Initially there is 20 kg of water in the bucket, so at time t
minutes, the mass of water remaining is
m = 20 0.5t kg.
Consider the time interval between t and t + t. During this time the bucket moves a distance x = 4t meters. So,
during this interval,
Work done mgx = (20 0.5t)g4t joules.
Total work done = lim
t0
(20 0.5t)g4t = 4g
_
10
0
(20 0.5t) dt
= 4g(20t 0.25t
2
)
10
0
= 700g = 700(9.8) = 6860 joules.
8.5 SOLUTIONS 183
x
T
c
40 m
Platform
Ground
Figure 8.22
13. Let x be the distance measured from the bottom
the tank. See Figure 8.23. To pump a layer of wa-
ter of thickness x at x feet from the bottom, the
work needed is
(62.4)6
2
(20 x)x.
Therefore, the total work is
W =
_
10
0
36 (62.4)(20 x)dx
= 36 (62.4)(20x
1
2
x
2
)
10
0
= 36 (62.4)(200 50)
1,058,591.1 ft-lb.
c
T
c
T
E
c
T
c
T
T
x
20 x
Volume of Slice = (6
2
)x
x
10
20
Figure 8.23
17. We slice the water horizontally as in Figure 8.24. We use similar triangles to nd the radius r of the slice at height h in
terms of h:
r
h
=
4
12
so r =
1
3
h.
At height h,
Volume of slice r
2
h =
_
1
3
h
_
2
h ft
3
.
The density of water is lb/ft
3
, so
Weight of slice
_
1
3
h
_
3
h lb.
The water at height h must be lifted a distance of 12 h ft, so
Work to move slice = Volume Distance lifted

_
_
1
3
h
_
2
h
_
(12 h) ft-lb.
The work done, W, to lift all the water is the sum of the work done on the pieces:
W
(
1
3
h)
2
h(12 h) ft-lb.
As h 0, we obtain a denite integral. Since h varies from h = 0 to h = 9, and = 62.4, we have:
W =
_
9
0
(
1
3
h)
2
(12 h)dh =
62.4
9
_
9
0
(12h
2
h
3
)dh =
62.4
9
_
4h
3
h
4
4
_
9
0
= 27,788 ft-lb.
The work to pump all the water out is 27,788 ft-lbs.
T
c
12 ft
T
c
h
T
c
h
T
c
12 h
E '
r
E '
4 ft
Figure 8.24
21. (a) Divide the wall into N horizontal strips, each
of which is of height h. See Figure 8.25. The
area of each strip is 1000h, and the pressure at
depth hi is 62.4hi, so we approximate the force
on the strip as 1000(62.4hi)h.
h
50
1000
Figure 8.25
Therefore,
Force on the Dam
N1
i=0
1000(62.4hi)h.
(b) As N , the Riemann sum becomes the integral, so the force on the dam is
_
50
0
(1000)(62.4h) dh = 62400
h
2
2
50
0
= 78,000,000 pounds.
25. (a) Since the density of water is = 1000 kg/m
3
, at the base of the dam, water pressure gh = 1000 9.8 180 =
1.76 10
6
nt/m
2
.
(b) To set up a denite integral giving the force, we divide the dam into horizontal strips. We use horizontal strips
because the pressure along each strip is approximately constant, since each part is at approximately the same depth.
See Figure 8.26.
Area of strip = 2000h m
2
.
Pressure at depth of h meters = gh = 9800h nt/m
2
. Thus,
Force on strip Pressure Area = 9800h 2000h = 1.96 10
7
hh nt.
Summing over all strips and letting h 0 gives:
Total force = lim
h0
1.96 10
7
hh = 1.96 10
7
_
180
0
hdh newtons.
Evaluating gives
Total force = 1.96 10
7
h
2
2
180
0
= 3.2 10
11
newtons.
T
c
180 m
E '
2000 m
T
c
h
T
c
h
Figure 8.26
8.5 SOLUTIONS 185
29. We need to divide the disk up into circular rings of charge and integrate their contributions to the potential (at P) from 0
to a. These rings, however, are not uniformly distant from the point P. A ring of radius z is
R
2
+z
2
away from point
P (see Figure 8.27).
The ring has area 2z z, and charge 2z z. The potential of the ring is then
2z z
R
2
+z
2
and the total potential
at point P is
_
a
0
2z dz
R
2
+z
2
=
_
a
0
2z dz
R
2
+z
2
.
We make the substitution u = z
2
. Then du = 2z dz. We obtain
_
a
0
2z dz
R
2
+z
2
=
_
a
2
0
du
R
2
+u
= (2
_
R
2
+u)
a
2
0
= (2
_
R
2
+z
2
)
a
0
= 2(
_
R
2
+a
2
R).
(The substitution u = R
2
+z
2
or
R
2
+z
2
works also.)
z
P
R
R
2
+z
2
Figure 8.27
33. This time, lets split the second rod into small slices of length dr. See Figure 8.28. Each slice is of mass
M
2
l
2
dr, since the
density of the second rod is
M
2
l
2
. Since the slice is small, we can treat it as a particle at distance r away from the end of
the rst rod, as in Problem 32. By that problem, the force of attraction between the rst rod and particle is
GM1
M
2
l
2
dr
(r)(r +l1)
.
So the total force of attraction between the rods is
_
a+l
2
a
GM1
M
2
l
2
dr
(r)(r +l1)
=
GM1M2
l2
_
a+l
2
a
dr
(r)(r +l1)
=
GM1M2
l2
_
a+l
2
a
1
l1
_
1
r

1
r +l1
_
dr.
=
GM1M2
l1l2
(ln |r| ln |r +l1|)
a+l
2
a
=
GM1M2
l1l2
[ln |a +l2| ln |a +l1 +l2| ln |a| + ln |a +l1|]
=
GM1M2
l1l2
ln
_
(a +l1)(a +l2)
a(a +l1 +l2)
_
.
This result is symmetric: if you switch l1 and l2 or M1 and M2, you get the same answer. That means its not important
which rod is rst, and which is second.
l2
l1
a
dr
r
Figure 8.28
Exercises
1. At any time t, in a time interval t, an amount of 1000t is deposited into the account. This amount earns interest for
(10 t) years giving a future value of 1000e
(0.08)(10t)
. Summing all such deposits, we have
Future value =
_
10
0
1000e
0.08(10t)
dt = $15,319.30.
Problems
5.
1 2
t (years from present)
$/year
The graph reaches a peak each summer, and a trough each winter. The graph shows sunscreen sales increasing from
cycle to cycle. This gradual increase may be due in part to ination and to population growth.
9. You should choose the payment which gives you the highest present value. The immediate lump-sum payment of $2800
obviously has a present value of exactly $2800, since you are getting it now. We can calculate the present value of the
installment plan as:
PV = 1000e
0.06(0)
+ 1000e
0.06(1)
+ 1000e
0.06(2)
$2828.68.
Since the installment payments offer a (slightly) higher present value, you should accept this option.
13. (a) Suppose the oil extracted over the time period [0, M] is S. (See Figure 8.29.) Since q(t) is the rate of oil extraction,
we have:
S =
_
M
0
q(t)dt =
_
M
0
(a bt)dt =
_
M
0
(10 0.1t) dt.
To calculate the time at which the oil is exhausted, set S = 100 and try different values of M. We nd M = 10.6
gives
_
10.6
0
(10 0.1t) dt = 100,
so the oil is exhausted in 10.6 years.
0 M
t
q(t)
Extraction Curve
Area below
the extraction curve
is the total oil extracted
E
Figure 8.29
(b) Suppose p is the oil price, C is the extraction cost per barrel, and r is the interest rate. We have the present value of
the prot as
Present value of prot =
_
M
0
(p C)q(t)e
rt
dt
8.7 SOLUTIONS 187
=
_
10.6
0
(20 10)(10 0.1t)e
0.1t
dt
= 624.9 million dollars.
17.
_
q
0
(p
S(q)) dq =
_
q
0
p
dq
_
q
0
S(q) dq
= p
_
q
0
S(q) dq.
Using Problem 16, this integral is the extra amount consumers pay (i.e., suppliers earn over and above the minimum they
would be willing to accept for supplying the good). It results from charging the equilibrium price.
Exercises
1. The two humps of probability in density (a) correspond to two intervals on which its cumulative distribution function is
increasing. Thus (a) and (II) correspond.
A density function increases where its cumulative distribution funciton is concave up, and it decreases where its
cumulative distribution function is concave down. Density (b) matches the distribution with both concave up and concave
down sections, which is (I). Density (c) matches (III) which has a concave down section but no interval over which it is
concave up.
5. Since the function takes on the value of 4, it cannot be a cdf (whose maximum value is 1). In addition, the function
decreases for x > c, which means that it is not a cdf. Thus, this function is a pdf. The area under a pdf is 1, so 4c = 1
giving c =
1
4
. The pdf is p(x) = 4 for 0 x
1
4
, so the cdf is given in Figure 8.30 by
P(x) =
_
_
0 for x < 0
4x for 0 x
1
4
1 for x >
1
4
1
4
1
x
P(x)
Figure 8.30
9. This function increases and levels off to c. The area under the curve is not nite, so it is not 1. Thus, the function must be
a cdf, not a pdf, and 3c = 1, so c = 1/3.
The pdf, p(x) is the derivative, or slope, of the function shown, so, using c = 1/3,
p(x) =
_
_
0 for x < 0
(1/3 0)/(2 0) = 1/6 for 0 x 2
(1 1/3)/(4 2) = 1/3 for 2 < x 4
0 for x > 4.
See Figure 8.31.
2 4
1
3
1
6
x
p(x)
Figure 8.31
Problems
13. (a) F(7) = 0.6 tells us that 60% of the trees in the forest have height 7 meters or less.
(b) F(7) > F(6). There are more trees of height less than 7 meters than trees of height less than 6 meters because every
tree of height 6 meters also has height 7 meters.
17. (a) The area under the graph of the height density function p(x) is concentrated in two humps centered at 0.5 mand 1.1 m.
The plants can therefore be separated into two groups, those with heights in the range 0.3 m to 0.7 m, corresponding
to the rst hump, and those with heights in the range 0.9 m to 1.3 m, corresponding to the second hump. This grouping
of the grasses according to height is probably close to the species grouping. Since the second hump contains more
area than the rst, there are more plants of the tall grass species in the meadow.
(b) As do all cumulative distribution functions, the cumulative distribution function P(x) of grass heights rises from 0
to 1 as x increases. Most of this rise is achieved in two spurts, the rst as x goes from 0.3 m to 0.7 m, and the second
as x goes from 0.9 m to 1.3 m. The plants can therefore be separated into two groups, those with heights in the range
0.3 m to 0.7 m, corresponding to the rst spurt, and those with heights in the range 0.9 m to 1.3 m, corresponding to
the second spurt. This grouping of the grasses according to height is the same as the grouping we made in part (a),
and is probably close to the species grouping.
(c) The fraction of grasses with height less than 0.7 m equals P(0.7) = 0.25 = 25%. The remaining 75% are the tall
grasses.
21. (a) We must have
_

0
f(t)dt = 1, for even though it is possible that any given person survives the disease, everyone
eventually dies. Therefore,
_

0
cte
kt
dt = 1.
Integrating by parts gives
_
b
0
cte
kt
dt =
c
k
te
kt
b
0
+
_
b
0
c
k
e
kt
dt
=
_
c
k
te
kt
c
k
2
e
kt
_
b
0
=
c
k
2

c
k
be
kb
c
k
2
e
kb
.
As b , we see
_

0
cte
kt
dt =
c
k
2
= 1 so c = k
2
.
(b) We are told that
_
5
0
f(t)dt = 0.4, so using the fact that c = k
2
and the antiderivatives from part (a), we have
_
5
0
k
2
te
kt
dt =
_
k
2
k
te
kt
k
2
k
2
e
kt
_
5
0
= 1 5ke
5k
e
5k
= 0.4
8.8 SOLUTIONS 189
so
5ke
5k
+e
5k
= 0.6.
Since this equation cannot be solved exactly, we use a calculator or computer to nd k = 0.275. Since c = k
2
, we
have c = (0.275)
2
= 0.076.
(c) The cumulative death distribution function, C(t), represents the fraction of the population that have died up to time
t. Thus,
C(t) =
_
t
0
k
2
xe
kx
dx =
_
kxe
kx
e
kx
_
t
0
= 1 kte
kt
e
kt
.
Exercises
1.
0.24
0
0.08
0.12
p(x)
A1
A2
A4
A3
2 6 8
x (tons of sh)
Splitting the gure into four pieces, we see that
Area under the curve = A1 +A2 +A3 +A4
=
1
2
(0.16)4 + 4(0.08) +
1
2
(0.12)2 + 2(0.12)
= 1.
We expect the area to be 1, since
_

p(x) dx = 1 for any probability density function, and p(x) is 0 except when
2 x 8.
Problems
5. (a) We can nd the proportion of students by integrating the density p(x) between x = 1.5 and x = 2:
P(2) P(1.5) =
_
2
1.5
x
3
4
dx
=
x
4
16
2
1.5
=
(2)
4
16

(1.5)
4
16
= 0.684,
so that the proportion is 0.684 : 1 or 68.4%.
(b) We nd the mean by integrating x times the density over the relevant range:
Mean =
_
2
0
x
_
x
3
4
_
dx
=
_
2
0
x
4
4
dx
=
x
5
20
2
0
=
2
5
20
= 1.6 hours.
(c) The median will be the time T such that exactly half of the students are nished by time T, or in other words
1
2
=
_
T
0
x
3
4
dx
1
2
=
x
4
16
T
0
1
2
=
T
4
16
T =
4
8 = 1.682 hours.
9. (a) Since = 100 and = 15:
p(x) =
1
15
2
e
1
2
(
x100
15
)
2
.
(b) The fraction of the population with IQ scores between 115 and 120 is (integrating numerically)
_
120
115
p(x) dx =
_
120
115
1
15
2
e
(x100)
2
450
dx
=
1
15
2
_
120
115
e
(x100)
2
450
dx
0.067 = 6.7% of the population.
13. It is not (a) since a probability density must be a non-negative function; not (c) since the total integral of a probability
density must be 1; (b) and (d) are probability density functions, but (d) is not a good model. According to (d), the
probability that the next customer comes after 4 minutes is 0. In real life there should be a positive probability of not
having a customer in the next 4 minutes. So (b) is the best answer.
Exercises
1. Vertical slices are circular. Horizontal slices would be similar to ellipses in cross-section, or at least ovals (a word derived
from ovum, the Latin word for egg).
SOLUTIONS to Review Problems for Chapter Eight 191
Figure 8.32
5. We slice the region vertically. Each rotated slice is approximately a cylinder with radius y = x
2
+ 1 and thickness x.
See Figure 8.33. The volume of a typical slice is (x
2
+ 1)
2
x. The volume, V , of the object is the sum of the volumes
of the slices:
V
(x
2
+ 1)
2
x.
As x 0 we obtain an integral.
V =
_
4
0
(x
2
+ 1)
2
dx =
_
4
0
(x
4
+ 2x
2
+ 1)dx =
_
x
5
5
+
2x
3
3
+x
_
4
0
=
3772
15
= 790.006.
4
E '
x
T
c
y
y = x
2
+ 1
x
y
Figure 8.33
9. We divide the region into vertical strips of thickness x. As a slice is rotated about the x-axis, it creates a disk of radius
rout from which has been removed a smaller circular disk of inside radius rin. We see in Figure 8.34 that rout = 2x and
rin = x. Thus,
Volume of a slice (rout)
2
x (rin)
2
x = (2x)
2
x (x)
2
x.
To nd the total volume, V , we integrate this quantity between x = 0 and x = 3:
V =
_
3
0
((2x)
2
(x)
2
)dx =
_
3
0
(4x
2
x
2
) dx =
_
3
0
3x
2
dx = x
3
3
0
= 27 = 84.823.
3
y = x
y = 2x
T
c
rout = 2x
T
c
r
in
= x
x
y
Figure 8.34
13. The region is bounded by x = 4y, the x-axis and x = 8. The two lines x = 4y and x = 8 intersect at (8, 2). We slice the
volume with planes that are perpendicular to the line x = 10. This divides the solid into thin washers with
Volume r
2
out
dy r
2
in
dy.
The inner radius is the distance from the line x = 10 to the line x = 8 and the outer radius is the distance from the line
x = 10 to the line x = 4y. Integrating from y = 0 to y = 2 we have
V =
_
2
0
_
(10 4y)
2
(2)
2
dy.
17. We slice the cone horizontally into cylindrical disks with radius r and thickness h. See Figure 8.35. The volume of each
disk is r
2
h. We use the similar triangles in Figure 8.36 to write r as a function of h:
r
h
=
3
12
so r =
1
4
h.
The volume of the disk at height h is (
1
4
h)
2
h. To nd the total volume, we integrate this quantity from h = 0 to
h = 12.
V =
_
12
0
_
1
4
h
_
2
dh =

16
h
3
3
12
0
= 36 = 113.097 m
3
.
T
c
12 m
E '
6 m
Tc
h
T
c
h
Figure 8.35
T
c
12 m
E '
r
T
c
h
E '
3
Figure 8.36
21. Well nd the arc length of the top half of the ellipse, and multiply that by 2. In the top half of the ellipse, the equation
(x
2
/a
2
) + (y
2
/b
2
) = 1 implies
y = +b
_
1
x
2
a
2
.
Differentiating (x
2
/a
2
) + (y
2
/b
2
) = 1 implicitly with respect to x gives us
2x
a
2
+
2y
b
2
dy
dx
= 0,
so
dy
dx
=
2x
a
2
2y
b
2
=
b
2
x
a
2
y
.
Substituting this into the arc length formula, we get
Arc Length =
_
a
a
1 +
_
b
2
x
a
2
y
_
2
dx
=
_
a
a
_
1 +
_
b
4
x
2
a
4
(b
2
)(1
x
2
a
2
)
_
dx
=
_
a
a
1 +
_
b
2
x
2
a
2
(a
2
x
2
)
_
dx.
Hence the arc length of the entire ellipse is
2
_
a
a
1 +
_
b
2
x
2
a
2
(a
2
x
2
)
_
dx.
25. The arc length is given by
L =
_
2
1
_
1 +e
2x
dx 4.785.
Note that
1 +e
2x
does not have an obvious elementary antiderivative, so we use an approximation method to nd an
approximate value for L.
Problems
29. (a) The points of intersection are x = 0 to x = 2, so we have
Area =
_
2
0
(2x x
2
)dx = x
2
x
3
3
2
0
=
4
3
= 1.333.
(b) The outside radius is 2x and the inside radius is x
2
, so we have
Volume =
_
2
0
((2x)
2
(x
2
)
2
)dx =
_
2
0
(4x
2
x
4
)dx =

15
(20x
3
3x
5
)
2
0
=
64
15
= 13.404.
(c) The length of the perimeter is equal to the length of the top plus the length of the bottom. Using the arclength formula,
and the fact that the derivative of 2x is 2 and the derivative of x
2
is 2x, we have
L =
_
2
0
_
1 + 2
2
dx +
_
2
0
_
1 + (2x)
2
dx = 4.4721 + 4.6468 = 9.119.
33. (a) Since y = e
bx
is non-negative, we integrate to nd the area:
Area =
_
1
0
(e
bx
)dx =
1
b
e
bx
1
0
=
1
b
(1 e
b
).
(b) Each slice of the object is approximately a cylinder with radius e
bx
and thickness x. We have
Volume =
_
1
0
(e
bx
)
2
dx =
_
1
0
e
2bx
dx =

2b
e
2bx
1
0
=

2b
(1 e
2b
).
37. Slice the object into disks vertically, as in Figure 8.37. A typical disk has thickness x and radius y =
1 x
2
. Thus
Volume of disk y
2
x = (1 x
2
) x.
Volume of solid = lim
x0
(1 x
2
) x =
_
1
0
(1 x
2
) dx =
_
x
x
3
3
_
1
0
=
2
3
.
Note: As we expect, this is the volume of a half sphere.
x
1
E '
x
y
1
Figure 8.37
41. Slicing perpendicularly to the y-axis gives semicircles whose thickness is y and whose diameter is x =
_
1 y
2
. See
Figure 8.38. Thus
Volume of semicircular slice

2
_
_
1 y
2
2
_
2
y =

8
(1 y
2
) y.
Volume of solid =
_
1
0
8
(1 y
2
) dy =

8
_
y
y
3
3
_
1
0
=

8

2
3
=

24
.
1
1
T
c
y
E '
x
Diameter of semicircle (standing on paper)

x
y
Figure 8.38: Base of Solid
45. (a) The line y = ax must pass through (l, b). Hence b = al, so a = b/l.
(b) Cut the cone into N slices, slicing perpendicular to the xaxis. Each piece is almost a cylinder. The radius of the ith
cylinder is r(xi) =
bxi
l
, so the volume
V
N
i=1
_
bxi
l
_
2
x.
Therefore, as N , we get
V =
_
l
0
b
2
l
2
x
2
dx
=
b
2
l
2
_
x
3
3
_
l
0
=
_
b
2
l
2
__
l
3
3
_
=
1
3
b
2
l.
49. Multiplying r = 2a cos by r, converting to Cartesian coordinates, and completing the square gives
r
2
= 2ar cos
x
2
+y
2
= 2ax
x
2
2ax +a
2
+y
2
= a
2
(x a)
2
+y
2
= a
2
.
This is the standard form of the equation of a circle with radius a and center (x, y) = (a, 0).
To check the limits on note that the circle is in the right half plane, where /2 /2. Rays from the origin
at all these angles meet the circle because the circle is tangent to the y-axis at the origin.
53. Writing C in parametric form gives
x = 2a cos
2
and y = 2a cos sin ,
so
Arc length =
_
/2
/2
_
(4a cos sin )
2
+ (2a sin
2
+ 2a cos
2
)
2
d
= 2a
_
/2
/2
_
4 cos
2
sin
2
+ sin
4
2 sin
2
cos
2
+ cos
4
d
= 2a
_
/2
/2
_
sin
4
+ 2 sin
2
cos
2
+ cos
4
d
= 2a
_
/2
/2
_
(sin
2
+ cos
2
)
2
d
= 2a
_
/2
/2
d = 2a.
57. Let x be the distance from the bucket to the surface of the water. It follows that 0 x 40. At x feet, the bucket weighs
(30
1
4
x), where the
1
4
x term is due to the leak. When the bucket is x feet from the surface of the water, the work done
by raising it x feet is (30
1
4
x) x. So the total work required to raise the bucket to the top is
W =
_
40
0
(30
1
4
x)dx
=
_
30x
1
8
x
2
_
40
0
= 30(40)
1
8
40
2
= 1000 ft-lb.
61. Let h be height above the bottom of the dam. Then
Water force =
_
25
0
(62.4)(25 h)(60) dh
= (62.4)(60)
_
25h
h
2
2
_
25
0
= (62.4)(60)(625 312.5)
= (62.4)(60)(312.5)
= 1,170,000 lbs.
65. We divide up time between 1971 and 1992 into intervals of length t, and calculate how much of the strontium-90
produced during that time interval is still around.
Strontium-90 decays exponentially, so if a quantity S0 was produced t years ago, and S is the quantity around today,
S = S0e
kt
. Since the half-life is 28 years,
1
2
= e
k(28)
, giving k = ln(1/2)/28 0.025.
We measure t in years from 1971, so that 1992 is t = 21.
E E '
1971 (t = 0) 1992 (t = 21)
(21 t)
t
t
Since strontium-90 is produced at a rate of 3 kg/year, during the interval t, a quantity 3t kg was produced. Since
this was (21 t) years ago, the quantity remaining now is (3t)e
0.025(21t)
. Summing over all such intervals gives
Strontium remaining
in 1992
_
21
0
3e
0.025(21t)
dt =
3e
0.025(21t)
0.025
21
0
= 49 kg.
[Note: This is like a future value problem from economics, but with a negative interest rate.]
69. Graph B is more spread out to the right, and so it represents a gas in which more of the molecules are moving at faster
velocities. Thus the average velocity in gas B is larger.
73. (a) Slicing horizontally, as shown in Figure 8.39, we see that the volume of one disk-shaped slab is
V x
2
y =
y
a
y.
Thus, the volume of the water is given by
V =
_
h
0
a
ydy =

a
y
2
2
h
0
=
h
2
2a
.
x
y
E '
x
T
c
y
T
c
h Tc
y
y = ax
2
Figure 8.39
(b) The surface of the water is a circle of radius x. Since at the surface, y = h, we have h = ax
2
. Thus, at the surface,
x =
_
(h/a). Therefore the area of the surface of water is given by
A = x
2
=
h
a
.
(c) If the rate at which water is evaporating is proportional to the surface area, we have
dV
dt
= kA.
(The negative sign is included because the volume is decreasing.) By the chain rule,
dV
dt
=
dV
dh

dh
dt
. We know
dV
dh
=
h
a
and A =
h
a
so
h
a
dh
dt
= k
h
a
giving
dh
dt
= k.
(d) Integrating gives
h = kt +h0.
Solving for t when h = 0 gives
t =
h0
k
.
77. (a) We need to check that the point with the given coordinates is on the curve, i.e., that
x = a sin
2
t, y =
a sin
3
t
cos t
satises the equation
y =
_
x
3
a x
.
This can be done by substituting into the computer algebra system and asking it to simplify the difference between
the two sides, or by hand calculation:
Right-hand side =
_
(a sin
2
t)
3
a a sin
2
t
=
_
a
3
sin
6
t
a(1 sin
2
t)
=
_
a
3
sin
6
t
a cos
2
t
=
_
a
2
sin
6
t
cos
2
t
=
a sin
3
t
cos t
= y = Left-hand side.
We chose the positive square root because both sin t and cos t are nonnegative for 0 t /2. Thus the point always
lies on the curve. In addition, when t = 0, x = 0 and y = 0, so the point starts at x = 0. As t approaches /2, the
value of x = a sin
2
t approaches a and the value of y = a sin
3
t/ cos t increases without bound (or approaches ),
so the point on the curve approaches the vertical asymptote at x = a.
(b) We calculate the volume using horizontal slices. See the graph of y =
_
x
3
/(a x) in Figure 8.40.
a
T
c
x a
y
x
y
Figure 8.40
The slice at y is a disk of thickness y and radius x a, hence it has volume (x a)
2
y. So the volume is
given by the improper integral
Volume =
_

0
(x a)
2
dy.
(c) We substitute
x = a sin
2
t, y =
a sin
3
t
cos t
and
dy =
d
dt
_
a sin
3
t
cos t
_
dt = a
_
3 sin
2
t +
sin
4
t
cos
2
t
_
dt.
Since t = 0 where y = 0 and t = /2 at the asymptote where y , we get
Volume =
_
/2
0
(a sin
2
t a)
2
a
_
3 sin
2
t +
sin
4
t
cos
2
t
_
dt
= a
3
_
/2
0
(3 sin
2
t cos
4
t + sin
4
t cos
2
t) dt =

2
a
3
8
.
You can use a CAS to calculate this integral; it can also be done using trigonometric identities.
1. True. Since y =
9 x
2
represent the top and bottom halves of the sphere, slicing disks perpendicular to the x-axis
gives
Volume of slice y
2
x = (9 x
2
)x
Volume =
_
3
3
(9 x
2
) dx.
5. False. Volume is always positive, like area.
9. True. One way to look at it is that the center of mass should not change if you change the units by which you measure the
masses. If you double the masses, that is no different than using as a new unit of mass half the old unit. Alternatively, let
the masses be m1, m2, and m3 located at x1, x2, and x3. Then the center of mass is given by:
x =
x1m1 +x2m2 +x3m3
m1 +m2 +m3
.
Doubling the masses does not change the center of mass, since it doubles both the numerator and the denominator.
13. False. Work is the product of force and distance moved, so the work done in either case is 200 ft-lb.
17. False. The pressure is positive and when integrated gives a positive force.
21. False. It is true that p(x) 0 for all x, but we also need
_
p(x)dx = 1. Since p(x) = 0 for x 0, we need only

check the integral from 0 to . We have
_

0
xe
x
2
dx = lim
b
_
1
2
e
x
2
_
b
0
=
1
2
.
25. False. Since f is concave down, this means that f
(x) is decreasing, so f
(x) f
(0) = 3/4 on the interval [0, 4].

However, it could be that f
(x) becomes negative so that (f
(x))
2
becomes large, making the integral for the arc length
large also. For example, f(x) = (3/4)x x
2
is concave down and f
(0) = 3/4, but f(0) = 0 and f(4) = 13, so the

graph of f on the interval [0, 4] has arc length at least 13.
29. False. Note that p is the density function for the population, not the cumulative density function. Thus p(10) = p(20)
means that x values near 10 are as likely as x values near 20.
9.1 SOLUTIONS 199
CHAPTER NINE
Exercises
1. The terms look like powers of 2 so we guess sn = 2
n
. This makes the rst term 2
1
= 2 rather than 4. We try instead
sn = 2
n+1
. If we now check, we get the terms 4, 8, 16, 32, 64, . . ., which is right.
5. The numerator is n. The denominator is then 2n + 1, so sn = n/(2n + 1).
9. The rst term is 2 1/(2 1 + 1) = 2/3. The second term is 2 2/(2 2 + 1) = 4/5. The rst ve terms are
2/3, 4/5, 6/7, 8/9, 10/11.
Problems
13. Since 2
n
increases without bound as n increases, the sequence diverges.
17. We have:
lim
n
_
n
10
+
10
n
_
= lim
n
n
10
+ lim
n
10n.
Since n/10 gets arbitrarily large and 10/n approaches 0 as n , the sequence diverges.
21. Since the exponential function 2
n
dominates the power function n
3
as n , the series diverges.
25. (a) matches (IV), since the sequence increases toward 1.
(b) matches (III), since the odd terms increase toward 1 and the even terms decrease toward 1.
(c) matches (II), since the sequence decreases toward 0.
(d) matches (I), since the sequence decreases toward 1.
29. We have s2 = s1 + 2 = 3 and s3 = s2 + 3 = 6. Continuing, we get
1, 3, 6, 10, 15, 21.
33. The rst 6 terms of the sequence for the sampling is
cos 0.5, cos 1.0, cos 1.5, cos 2.0, cos 2.5, cos 3.0
= 0.878, 0.540, 0.071, 0.416, 0.801, 0.990.
37. The rst smoothing gives
1.5, 2, 3, 4, 5, 6, 7 . . .
The second smoothing gives
1.75, 2.17, 3, 4, 5, 6 . . .
Terms which are already the same as their average with their neighbors are not changed.
41. The differences between consecutive terms are 4, 9, 16, 25, so, for example, s2 = s1 +4 and s3 = s2 +9. Thus, a possible
recursive denition is sn = sn1 + n
2
for n > 1 and s1 = 1.
45. For n > 1, if sn = n(n + 1)/2, then sn1 = (n 1)(n 1 + 1)/2 = n(n 1)/2. Since
sn =
1
2
(n
2
+ n) =
n
2
2
+
n
2
and sn1 =
1
2
(n
2
n) =
n
2
2

n
2
,
we have
sn sn1 =
n
2
+
n
2
= n,
so
sn = sn1 + n.
In addition, s1 = 1(2)/2 = 1.
200 Chapter Nine /SOLUTIONS
49. The sequence appears to be decreasing toward 0, but at a slower and slower rate. Even after 100 terms, it is hard to guess
what the series will eventually do. It can be shown that it converges to 0.
53. In year 1, the payment is
p1 = 10,000 + 0.05(100,000) = 15,000.
The balance in year 2 is 100,000 10,000 = 90,000, so
p2 = 10,000 + 0.05(90,000) = 14,500.
The balance in year 3 is 80,000, so
p3 = 10,000 + 0.05(80,000) = 14,000.
Thus,
pn = 10,000 + 0.05(100,000 (n 1) 10,000)
= 15,500 500n.
57. (a) The rst 12 terms are
1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144.
(b) The sequence of ratios is
1, 2,
3
2
,
5
3
,
8
5
,
13
8
,
21
13
,
34
21
,
55
34
,
89
55
. . . .
To three decimal places, the rst ten ratios are
1, 2, 1.500, 1.667, 1.600, 1.625, 1.615, 1.619, 1.618, 1.618.
It appears that the sequence of ratios is converging to r = 1.618. We nd (1.618)
2
= 2.618 = 1.618 + 1 so
r seems to satisfy r
2
= r + 1. Alternatively, by the quadratic formula, the positive root of x
2
x 1 = 0 is
(1 +
5)/2 = 1.618.
(c) If we multiply both sides of the equation r
2
= r + 1 by Ar
n2
, we obtain
Ar
n
= Ar
n1
+ Ar
n2
.
Thus, if sn = Ar
n
, then sn1 = Ar
n1
and sn2 = Ar
n2
, so the sequence satises sn = sn1 + sn2.
Exercises
1. Yes, a = 5, ratio = 2.
5. No. Ratio between successive terms is not constant:
2x
2
x
= 2x, while
3x
3
2x
2
=
3
2
x.
9. Yes, a = 1, ratio = y
2
.
13. The series has 9 terms. The rst term is a = 0.00002 and the constant ratio is x = 0.1, so
Sum =
0.00002(1 x
9
)
(1 x)
=
0.00002(1 (0.1)
9
)
0.9
= 0.0000222.
17. Using the formula for the sum of a nite geometric series,
20
n=4
_
1
3
_
n
=
_
1
3
_
4
+
_
1
3
_
5
+ +
_
1
3
_
20
=
_
1
3
_
4
_
1 +
1
3
+
_
1
3
_
2
+
_
1
3
_
16
_
=
(1/3)
4
(1 (1/3)
17
)
1 (1/3)
=
3
17
1
2 3
20
.
9.2 SOLUTIONS 201
21. This is a geometric series with rst term 2 and ratio 2z,
2 4z + 8z
2
16z
3
+ =
2
1 (2z)
=
2
1 + 2z
.
This series converges for | 2z| < 1, that is for 1/2 < z < 1/2.
Problems
25. Since the amount of ampicillin excreted during the time interval between tablets is 250 mg, we have
Amount of ampicillin excreted = Original quantity Final quantity
250 = Q(0.04)Q.
Solving for Q gives, as before,
Q =
250
1 0.04
260.42.
29. If a payment M in the future has present value P, then
M = P(1 + r)
t
,
where t is the number of periods in the future and r is the interest rate. Thus
P =
M
(1 + r)
t
.
The monthly interest rate here is 0.09/12 = 0.0075, so the present value of rst payment, made at the end of the rst
month, is M/(1.0075). The present value of the second payment is M/(1.0075)
2
. Continuing in this way, the sum of the
present value of all of the payments is
M
(1.0075)
+
M
(1.0075)
2
+ +
M
(1.0075)
240
.
This is a nite geometric series with 240 terms, with sum
M
1.0075
_
1 1.0075
240
1 1.0075
1
_
= 111.145M.
Setting this equal to the loan amount of 200,000 gives a monthly payment of M = $200,000/111.145 = $1799.45.
33. If the half-life is T hours, then the exponential decay formula Q = Q0e
kt
gives k = ln 2/T. If we start with Q0 = 1
tablet, then the amount of drug present in the body after 5T hours is
Q = e
5kT
= e
5 ln 2
= 0.03125,
so 3.125% of a tablet remains. Thus, immediately after taking the rst tablet, there is one tablet in the body. Five half-lives
later, this has reduced to 1 0.03125 = 0.03125 tablets, and immediately after the second tablet there are 1 + 0.03125
tablets in the body. Continuing this forever leads to
Number of tablets in body = 1 + 0.03125 + (0.03125)
2
+ + (0.03125)
n
+ .
This is an innite geometric series, with common ratio x = 0.03125, and sum 1/(1 x). Thus
Number of tablets in body =
1
1 0.03125
= 1.0323.
37. (a) The acceleration of gravity is 32 ft/sec
2
so acceleration = 32 and velocity v = 32t + C. Since the ball is dropped,
its initial velocity is 0 so v = 32t. Thus the position is s = 16t
2
+ C. Calling the initial position s = 0, we have
s = 6t. The distance traveled is h so h = 16t. Solving for t we get t =
1
4
h.
(b) The rst drop from 10 feet takes
1
4
10 seconds. The rst full bounce (to 10 (

3
4
) feet) takes
1
4
_
10 (
3
4
) seconds
to rise, therefore the same time to come down. Thus, the full bounce, up and down, takes 2(
1
4
)
_
10 (
3
4
) seconds.
The next full bounce takes 2(
1
4
)10 (
3
4
)
2
= 2(
1
4
)
10
_
_
3
4
_
2
seconds. The n
th
bounce takes 2(
1
4
)
10
_
_
3
4
_
n
seconds. Therefore the
Total amount of time
=
1
4
10 +
2
4
10
_
3
4
+
2
4
10
_
_
3
4
_
2
+
2
4
10
_
_
3
4
_
3
. .
Geometric series with a =
2
4
10
_
3
4
=
1
2
10
_
3
4
and x =
_
3
4
+
=
1
4
10 +
1
2
10
_
3
4
_
1
1
_
3/4
_
seconds.
Exercises
1. The series is 1 + 2 + 3 + 4 + 5 + . The sequence of partial sums is
S1 = 1, S2 = 1 + 2, S3 = 1 + 2 + 3, S4 = 1 + 2 + 3 + 4, S5 = 1 + 2 + 3 + 4 + 5, . . .
which is
1, 3, 6, 10, 15 . . . .
5. We use the integral test with f(x) = x/(x
2
+ 1) to determine whether this series converges or diverges. We determine
whether the corresponding improper integral
_

1
x
x
2
+ 1
dx converges or diverges:
_

1
x
x
2
+ 1
dx = lim
b
_
b
1
x
x
2
+ 1
dx = lim
b
1
2
ln(x
2
+ 1)
b
1
= lim
b
_
1
2
ln(b
2
+ 1)
1
2
ln 2
_
= .
Since the integral
_

1
x
x
2
+ 1
dx diverges, we conclude from the integral test that the series
n=1
n
n
2
+ 1
diverges.
9. The improper integral
_

0
1
x
2
+ 1
dx converges to

2
, since
_
b
0
1
x
2
+ 1
dx = arctan x|
b
0
= arctan b arctan 0 = arctan b,
and lim
b
arctan b =

2
. The terms of the series
n=1
1
n
2
+ 1
form a right hand sum for the improper integral; each term
represents the area of a rectangle of width 1 tting completely under the graph of the function
1
x
2
+ 1
. (See Figure 9.1.)
Thus the sequence of partial sums is bounded above by

2
. Since the partial sums are increasing (every new term added is
positive), the series is guaranteed to converge to some number less than or equal to /2 by Theorem 9.1.
9.3 SOLUTIONS 203
1 2 3 4 5 6 7 8
1/10
1/5
1/2
1
x
y
y = 1/(x
2
+ 1)
Figure 9.1
Problems
13. Using the integral test, we compare the series with
_

0
3
x + 2
dx = lim
b
_
b
0
3
x + 2
dx = 3 ln |x + 2|
b
0
.
Since ln(b + 2) is unbounded as b , the integral diverges and therefore so does the series.
17. Since the terms in the series are positive and decreasing, we can use the integral test. We calculate the corresponding
improper integral using the substitution w = 1 + x
2
:
_

0
2x
(1 + x
2
)
2
dx = lim
b
_
b
0
2x
(1 + x
2
)
2
dx = lim
b
1
(1 + x
2
)
b
0
= lim
b
_
1
1 + b
2
+ 1
_
= 1.
Since the limit exists, the integral converges, so the series
n=0
2n
(1 + n
2
)
2
converges.
21. Using the integral test, we compare the series with
_

0
3
x
2
+ 4
dx = lim
b
_
b
0
3
x
2
+ 4
dx =
3
2
lim
b
arctan
_
x
2
_
b
0
=
3
2
lim
b
arctan
_
b
2
_
=
3
4
,
by integral table V-24. Since the integral converges so does the series.
25. Both
n=1
_
1
2
_
n
and
n=1
_
2
3
_
n
are convergent geometric series. Therefore, by Property 1 of Theorem 9.2, the series
n=1
_
1
2
_
n
+
_
2
3
_
n
converges.
29. Since the terms in the series are positive and decreasing, we can use the integral test. We calculate the corresponding
improper integral using the substitution w = 1 + ln x:
_

1
1
x(1 + ln x)
dx = lim
b
_
b
1
1
x(1 + ln x)
dx = lim
b
ln(1 + ln x)
b
1
= lim
b
ln(1 + ln b).
Since the limit does not exist, the integral diverges, so the series
n=1
1
n(1 + ln n)
diverges.
33. Using ln(2
n
) = nln 2, we see that
1
ln(2
n
)
=
1
(ln 2)n
.
The series on the right is the harmonic series multiplied by 1/ ln 2. Since the harmonic series diverges,

n=1
1/ ln(2
n
)
diverges.
37. (a) A common denominator is k(k + 1) so
1
k

1
k + 1
=
k + 1
k(k + 1)

k
k(k + 1)
=
k + 1 k
k(k + 1)
=
1
k(k + 1)
.
(b) Using the result of part (a), the partial sum can be written as
S3 =
1
1 2
+
1
2 3
+
1
3 4
=
1
1

1
2
+
1
2

1
3
+
1
3

1
4
= 1
1
4
.
All of the intermediate terms cancel out, leaving only the rst and last terms. Thus S10 = 1
1
11
and Sn = 1
1
n + 1
.
(c) The limit of Sn as n is lim
n
_
1
1
n + 1
_
= 1 0 = 1. Thus the series
k=1
1
k(k + 1)
converges to 1.
41. We have an = Sn Sn1. If

an converges, then S = limnSn exists. Hence limnSn1 exists and is equal
to S also. Thus
lim
n
an = lim
n
(Sn Sn1) = lim
n
Sn lim
n
Sn1 = S S = 0.
45. (a) Let N an integer with N c. Consider the series
i=N+1
ai. The partial sums of this series are increasing because all
the terms in the series are positive. We show the partial sums are bounded using the right-hand sum in Figure 9.2. We
see that for each positive integer k
f(N + 1) + f(N + 2) + + f(N + k)
_
N+k
N
f(x) dx.
Since f(n) = an for all n, and c N, we have
aN+1 + aN+2 + + a
N+k

_
N+k
c
f(x) dx.
Since f(x) is a positive function,
_
N+k
c
f(x) dx
_
b
c
f(x) dx for all b N + k. Since f is positive and
_
c
f(x) dx is convergent,
_
N+k
c
f(x) dx <
_
c
f(x) dx, so we have
aN+1 + aN+2 + + a
N+k

_

c
f(x) dx for all k.
Thus, the partial sums of the series
i=N+1
ai are bounded and increasing, so this series converges by Theorem 9.1.
Now use Theorem 9.2, property 2, to conclude that
i=1
ai converges.
c
N N + 1
f(x)
Area = f(N + 3)
Area = f(N + 1)
Area = f(N + 2)
x
Figure 9.2
c
N N + 1
Area = f(N)
Area = f(N + 1)
Area = f(N + 2)
f(x)
x
Figure 9.3
9.3 SOLUTIONS 205
(b) We now suppose
_

c
f(x) dx diverges. In Figure 9.3 we see that for each positive integer k
_
N+k+1
N
f(x) dx f(N) + f(N + 1) + + f(N + k).
Since f(n) = an for all n, we have
_
N+k+1
N
f(x) dx aN + aN+1 + + a
N+k
.
Since f(x) is dened for all x c, if
_
c
f(x) dx is divergent, then
_
N
f(x) dx is divergent. So as k , the the
integral
_
N+k+1
N
f(x) dx diverges, so the partial sums of the series
i=N
ai diverge. Thus, the series
i=1
ai diverges.
More precisely, suppose the series converged. Then the partial sums would be bounded. (The partial sums would
be less than the sum of the series, since all the terms in the series are positive.) But that would imply that the integral
converged, by Theorem 9.1 on Convergence of Monotone Bounded Sequences. This contradicts the assumption that
_
N
f(x) dx is divergent.
49. (a) A calculator or computer gives
20
1
1
n
2
=
1
1
2
+
1
2
2
+ +
1
20
2
= 1.596.
(b) Since
1
1
n
2
=

2
6
, the answer to part (a) gives
2
6
1.596

6 1.596 = 3.09
(c) A calculator or computer gives
100
1
1
n
2
=
1
1
2
+
1
2
2
+ +
1
100
2
= 1.635,
so
2
6
1.635

6 1.635 = 3.13.
(d) The error in approximating
2
/6 by
20
1
1/n
2
is the tail of the series
21
1/n
2
. From Figure 9.4, we see that
21
1
n
2
<
_

20
dx
x
2
=
1
x
20
=
1
20
= 0.05.
A similar argument leads to a bound for the error in approximating
2
/6 by
100
1
1/n
2
as
101
1
n
2
<
_

100
dx
x
2
=
1
x
100
=
1
100
= 0.01.
20 21 22
. . .
1/x
2
x
Figure 9.4
Exercises
1. Let an = 1/(n 3), for n 4. Since n 3 < n, we have 1/(n 3) > 1/n, so
an >
1
n
.
The harmonic series
n=4
1
n
diverges, so the comparison test tells us that the series
n=4
1
n 3
also diverges.
5. As n gets large, polynomials behave like the leading term, so for large n,
n + 4
n
3
+ 5n 3
behaves like
n
n
3
=
1
n
2
.
Since the series
n=1
1/n
2
converges, we predict that the given series will converge.
9. Let an = 1/(n
4
+ e
n
). Since n
4
+ e
n
> n
4
, we have
1
n
4
+ e
n
<
1
n
4
,
so
0 < an <
1
n
4
.
Since the p-series
n=1
1
n
4
converges, the comparison test tells us that the series
n=1
1
n
4
+ e
n
also converges.
13. Let an = (2
n
+ 1)/(n2
n
1). Since n2
n
1 < n2
n
+ n = n(2
n
+ 1), we have
2
n
+ 1
n2
n
1
>
2
n
+ 1
n(2
n
+ 1)
=
1
n
.
Therefore, we can compare the series
n=1
2
n
+ 1
n2
n
1
with the divergent harmonic series
n=1
1
n
. The comparison test tells
us that
n=1
2
n
+ 1
n2
n
1
also diverges.
17. Since an = 1/(r
n
n!), replacing n by n + 1 gives an+1 = 1/(r
n+1
(n + 1)!). Thus
|an+1|
|an|
=
1
r
n+1
(n + 1)!
1
r
n
n!
=
r
n
n!
r
n+1
(n + 1)!
=
1
r(n + 1)
,
so
L = lim
n
|an+1|
|an|
=
1
r
lim
n
1
n + 1
= 0.
Since L = 0, the ratio test tells us that
n=1
1
r
n
n!
converges for all r > 0.
21. Since cos(n) = (1)
n
, this is an alternating series.
25. Let an = 1/(2n + 1). Then replacing n by n + 1 gives an+1 = 1/(2n + 3). Since 2n + 3 > 2n + 1, we have
0 < an+1 =
1
2n + 3
<
1
2n + 1
= an.
We also have limnan = 0. Therefore, the alternating series test tells us that the series
n=1
(1)
n1
2n + 1
converges.
9.4 SOLUTIONS 207
29. The series
(1)
n
2n
converges by the alternating series test. However
1
2n
diverges because it is a multiple of the
harmonic series. Thus
(1)
n
2n
is conditionally convergent.
33. We rst check absolute convergence by deciding whether

arcsin(1/n) converges. Since sin for small angles ,
writing x = sin we see that
arcsin x x.
Since arcsin x behaves like x for small x, we expect arcsin(1/n) to behave like 1/n for large n. To conrm this we
calculate
lim
n
arcsin(1/n)
1/n
= lim
x0
arcsin x
x
= lim
0
sin
= 1.
Thus
arcsin(1/n) diverges by limit comparison with the harmonic series

1/n.
We now check conditional convergence. The original series is alternating, so we check whether an+1 < an. Consider
an = f(n), where f(x) = arcsin(1/x). Since
d
dx
arcsin
1
x
=
1
_
1 (1/x)
2
_
1
x
2
_
is negative for x > 1, we know that an is decreasing for n > 1. Therefore, for n > 1
an+1 = arcsin(1/(n + 1)) < arcsin(1/n) = an.
Since arcsin(1/n) 0 as n , we see that
(1)
n1
arcsin(1/n) is conditionally convergent.
37. The n
th
term is an = 1 cos(1/n) and we are taking bn = 1/n
2
. We have
lim
n
an
bn
= lim
n
1 cos(1/n)
1/n
2
.
This limit is of the indeterminate form 0/0 so we evaluate it using lHopitals rule. We have
lim
n
1 cos(1/n)
1/n
2
= lim
n
sin(1/n)(1/n
2
)
2/n
3
= lim
n
1
2
sin(1/n)
1/n
= lim
x0
1
2
sin x
x
=
1
2
.
The limit comparison test applies with c = 1/2. The p-series

1/n
2
converges because p = 2 > 1. Therefore
(1 cos(1/n)) also converges.

41. The n
th
term an = 2
n
/(3
n
1) behaves like 2
n
/3
n
for large n, so we take bn = 2
n
/3
n
. We have
lim
n
an
bn
= lim
n
2
n
/(3
n
1)
2
n
/3
n
= lim
n
3
n
3
n
1
= lim
n
1
1 3
n
= 1.
The limit comparison test applies with c = 1. The geometric series

2
n
/3
n
=
(2/3)
n
converges. Therefore
2
n
/(3
n
1) also converges.
Problems
45. The comparison test requires that an = sin n be positive for all n. It is not.
49. The alternating series test requires an = sin n be positive, which it is not. This is not an alternating series.
53. The partial sums look like: S1 = 1, S2 = 0, S3 = 0.5, S4 = 0.3333, S5 = 0.375, S10 = 0.3679, S20 = 0.3679, and
higher partial sums agree with these rst 4 decimal places. The series appears to be converging to about 0.3679.
Since an = 1/n! is positive and decreasing and limn1/n! = 0, the alternating series test conrms the conver-
gence of this series.
57. We use the ratio test and calculate
lim
n
|an+1|
|an|
= lim
n
n!/(n + 1)
2
(n 1)!/n
2
= lim
n
_
n!
(n 1)!

n
2
(n + 1)
2
_
= lim
n
_
n
n
2
(n + 1)
2
_
.
Since the limit does not exist (it is ), the series diverges.
61. We compare the series with the convergent series
1/n
2
. From the graph of tan x, we see that tan x < 2 for 0 x 1,
so tan(1/n) < 2 for all n. Thus
1
n
2
tan
_
1
n
_
<
1
n
2
2,
so the series converges, since 2
1/n
2
converges. Alternatively, we try the integral test. Since the terms in the series
are positive and decreasing, we can use the integral test. We calculate the corresponding integral using the substitution
w = 1/x:
_

1
1
x
2
tan
_
1
x
_
dx = lim
b
_
b
1
1
x
2
tan
_
1
x
_
dx = lim
b
ln
_
cos
1
x
_
b
1
= lim
b
_
ln
_
cos
_
1
b
__
ln(cos 1)
_
= ln(cos 1).
Since the limit exists, the integral converges, so the series
n=1
1
n
2
tan (1/n) converges.
65. Since the exponential, 2
n
, grows faster than the power, n
2
, the terms are growing in size. Thus, lim
n
an = 0. We conclude
that this series diverges.
69. As n , we see that
n + 2
n
2
1

n
n
2
=
1
n
.
Since
(1/n) diverges, we expect our series to have the same behavior.

More precisely, for all n 2, we have
0
1
n
=
n
n
2

n + 2
n
2
1
,
so
n=2
n + 2
n
2
1
diverges by comparison with the divergent series
1
n
.
73. Factoring gives
n=1
_
2
n
_
a
= 2
a
n=1
1
n
a
. This is a constant times a p-series that converges if a > 1 and diverges if
a 1.
77. To use the alternating series test, consider an = f(n), where f(x) = arctan(a/x). We need to show that f(x) is
decreasing. Since
f
(x) =
1
1 + (a/x)
2
_
a
x
2
_
,
we have f
(x) < 0 for a > 0, so f(x) is decreasing for all x. Thus an+1 < an for all n, and as lim
n
arctan(a/n) = 0
for all a, by the alternating series test,
n=1
(1)
n
arctan(a/n)
converges.
81. Since 0 cn 2
n
for all n, and since

2
n
is a convergent geometric series,

cn converges by the Comparison
Test. Similarly, since 2
n
an, and since

2
n
is a divergent geometric series,

an diverges by the Comparison Test.
We do not have enough information to determine whether or not

bn and
dn converge.
85. Each term in
bn is greater than or equal to a1 times a term in the harmonic series:

b1 = a1 1
b2 =
a1 + a2
2
> a1
1
2
b3 =
a1 + a2 + a3
3
> a1
1
3
.
.
.
bn =
a1 + a2 + + an
n
> a1
1
n
Adding these inequalities gives
bn > a1
1
n
.
9.5 SOLUTIONS 209
Since the harmonic series

1/n diverges, a1 times the harmonic series also diverges. Then, by the comparison test, the
series
bn diverges.
89. The limit
lim
n
n
an = lim
n
5n + 1
3n
2
= 0 < 1,
so the series converges.
Exercises
1. Yes.
5. The general term can be written as
1 3 5 (2n 1)
2
n
n!
x
n
for n 1. Other answers are possible.
9. The general term can be written as
(x a)
n
2
n1
n!
for n 1. Other answers are possible.
13. Since Cn = (n + 1)/(2
n
+ n), replacing n by n + 1 gives Cn+1 = (n + 2)/(2
n+1
+ n + 1). Using the ratio test, we
have
|an+1|
|an|
= |x|
|Cn+1|
|Cn|
= |x|
(n + 2)/(2
n+1
+ n + 1)
(n + 1)/(2
n
+ n)
= |x|
n + 2
2
n+1
+ n + 1

2
n
+ n
n + 1
= |x|
n + 2
n + 1

2
n
+ n
2
n+1
+ n + 1
.
Since
lim
n
n + 2
n + 1
= 1
and
lim
n
_
2
n
+ n
2
n+1
+ n + 1
_
=
1
2
lim
n
_
2
n
+ n
2
n
+ (n + 1)/2
_
=
1
2
,
because 2
n
dominates n as n , we have
lim
n
|an+1|
|an|
=
1
2
|x|.
Thus the radius of convergence is R = 2.
17. The coefcient of the n
th
term is Cn = (1)
n+1
/n
2
. Now consider the ratio
an+1
an
n
2
x
n+1
(n + 1)
2
x
n
|x| as n .
Thus, the radius of convergence is R = 1.
21. We write the series as
x
x
3
3
+
x
5
5

x
7
7
+ + (1)
n1
x
2n1
2n 1
+ ,
so
an = (1)
n1
x
2n1
2n 1
.
Replacing n by n + 1, we have
an+1 = (1)
n+11
x
2(n+1)1
2(n + 1) 1
= (1)
n
x
2n+1
2n + 1
.
Thus
|an+1|
|an|
=
(1)
n
x
2n+1
2n + 1
2n 1
(1)
n1
x
2n1
=
2n 1
2n + 1
x
2
,
so
L = lim
n
|an+1|
|an|
= lim
n
2n 1
2n + 1
x
2
= x
2
.
By the ratio test, this series converges if L < 1, that is, if x
2
< 1, so R = 1.
Problems
25. We use the ratio test:
an+1
an
(x 3)
n+1
n + 1

n
(x 3)
n
=
n
n + 1
|x 3|.
Since n/(n + 1) 1 as n , we have
lim
n
an+1
an
= |x 3|.
The series converges for |x 3| < 1. The radius of convergence is 1 and the series converges for 2 < x < 4.
We check the endpoints. For x = 2, we have
n=2
(x 3)
n
n
=
n=2
(2 3)
n
n
=
n=2
(1)
n
n
.
This is the alternating harmonic series and converges. For x = 4, we have
n=2
(x 3)
n
n
=
n=2
(4 3)
n
n
=
n=2
1
n
.
This is the harmonic series and diverges. The series converges at x = 2 and diverges at x = 4. Therefore, the interval of
convergence is 2 x < 4.
29. We use the ratio test to nd the radius of convergence
an+1
an
(n + 1)!x
n+1
n!x
n
= |x|(n + 1).
Since limn|x|(n + 1) = for all x = 0, the radius of convergence is R = 0. There are no endpoints to check. The
series converges only for x = 0, so the interval of convergence is the single point x = 0.
33. To compare 8/(4 +y) with a/(1 x), divide the numerator and denominator by 4. This gives 8/(4 +y) = 2/(1 +y/4),
which is the sum of a geometric series with a = 2, x = y/4. The power series is
2 + 2(y/4) + 2(y/4)
2
+ 2(y/4)
3
+ =
n=0
2(y/4)
n
.
This series converges for |y/4| < 1, that is, for 4 < y < 4.
37. The series is centered at x = 7. Since the series converges at x = 0, which is a distance of 7 from x = 7, the radius
of convergence, R, is at least 7. Since the series diverges at x = 17, which is a distance of 10 from x = 7, the radius
of convergence is no more than 10. That is, 7 R 10.
41. (a) We have
f(x) = 1 + x +
x
2
2
+ ,
so
f(0) = 1 + 0 + 0 + = 1.
(b) To nd the domain of f, we nd the interval of convergence.
lim
n
|an+1|
|an|
= lim
n
|x
n+1
/(n + 1)!|
|x
n
/n!|
= lim
n
_
|x|
n+1
n!
|x|
n
(n + 1)!
_
= |x| lim
n
1
n + 1
= 0.
Thus the series converges for all x, so the domain of f is all real numbers.
(c) Differentiating term-by-term gives
f
(x) =
d
dx
_

n=0
x
n
n!
_
=
d
dx
_
1 + x +
x
2
2!
+
x
3
3!
+
x
4
4!
+
_
= 0 + 1 + 2
x
2!
+ 3
x
2
3!
+ 4
x
3
4!
+
= 1 + x +
x
2
2!
+
x
3
3!
+ .
SOLUTIONS to Review Problems for Chapter Nine 211
Thus, the series for f and f
are the same, so

f(x) = f
(x).
(d) We guess f(x) = e
x
.
Exercises
1. If b = 1, then the sum is 6. If b = 1, we use the formula for the sum of a nite geometric series. This is a six-term
geometric series (n = 6) with initial term a = b
5
and constant ratio x = b :
Sum =
a(1 x
n
)
1 x
=
b
5
(1 b
6
)
1 b
.
5. We have:
lim
n
_
n + 1
n
_
= 1.
The terms of the sequence do not approach 0, and oscillate between values that are getting closer to +1 and 1. Thus the
sequence diverges.
9. We use the integral test to determine whether this series converges or diverges. To do so we determine whether the
corresponding improper integral
_

1
3x
2
+ 2x
x
3
+ x
2
+ 1
dx converges or diverges. The integral can be calculated using the
substitution w = x
3
+ x
2
+ 1, dw = (3x
2
+ 2x) dx.
_

1
3x
2
+ 2x
x
3
+ x
2
+ 1
dx = lim
b
_
b
1
3x
2
+ 2x
x
3
+ x
2
+ 1
dx
= lim
b
ln |x
3
+ x
2
+ 1|
b
1
= lim
b
_
ln |b
3
+ b
2
+ 1| ln 3
_
= .
Since the integral
_

1
3x
2
+ 2x
x
3
+ x
2
+ 1
dx diverges, we conclude from the integral test that the series
n=1
3n
2
+ 2n
n
3
+ n
2
+ 1
diverges.
13. Since an = 1/(2
n
n!), replacing n by n + 1 gives an+1 = 1/(2
n+1
(n + 1)!). Thus
|an+1|
|an|
=
1
2
n+1
(n + 1)!
1
2
n
n!
=
2
n
n!
2
n+1
(n + 1)!
=
1
2(n + 1)
,
so
L = lim
n
|an+1|
|an|
= lim
n
1
2n + 2
= 0.
Since L < 1, the ratio test tells us that
n=1
1
2
n
n!
converges.
17. Let an = 1/
n
2
+ 1. Then replacing n by n + 1 we have an+1 = 1/
_
(n + 1)
2
+ 1. Since
_
(n + 1)
2
+ 1 >
n
2
+ 1, we have
1
_
(n + 1)
2
+ 1
<
1
n
2
+ 1
,
so
0 < an+1 < an.
In addition, limnan = 0 so
n=0
(1)
n
n
2
+ 1
converges by the alternating series test.
21. Since
lim
n
an = lim
n
1
arctan n
=
2
= 0
we know that
(1)
n1
arctan n
diverges by Property 3 of Theorem 9.2.
25. The n
th
term an = (n
3
2n
2
+ n + 1)/(n
5
2) behaves like n
3
/n
5
= 1/n
2
for large n, so we take bn = 1/n
2
. We
have
lim
n
an
bn
= lim
n
(n
3
2n
2
+ n + 1)/(n
5
2)
1/n
2
= lim
n
n
5
2n
4
+ n
3
+ n
2
n
5
2
= 1.
The limit comparison test applies with c = 1. The p-series

1/n
2
converges because p = 2 > 1. Therefore the series
_
n
3
2n
2
+ n + 1
_
/
_
n
5
2
_
also converges.
29. This is a p-series with p > 1, so it converges.
33. We use the integral test to determine whether this series converges or diverges. To do so we determine whether the
corresponding improper integral
_

1
x
2
x
3
+ 1
dx converges or diverges:
_

1
x
2
x
3
+ 1
dx = lim
b
_
b
1
x
2
x
3
+ 1
dx = lim
b
1
3
ln |x
3
+ 1|
b
1
= lim
b
_
1
3
ln(b
3
+ 1)
1
3
ln 2
_
.
Since the limit does not exist, the integral
_

1
x
2
x
3
+ 1
dx diverges an so we conclude from the integral test that the
series
n=1
n
2
n
3
+ 1
diverges. The limit comparison test with bn = 1/n can also be used.
37. Let an = 2
n
(n + 1)
(n + 2)
=
_
n + 1
n + 2
__
1
2
n
_
. Since
(n + 1)
(n + 2)
< 1 and
1
2
n
=
_
1
2
_
n
, we have
0 < an <
_
1
2
_
n
,
so that we can compare the series
n=1
2
n
(n + 1)
(n + 2)
with the convergent geometric series
n=1
_
1
2
_
n
. The comparison test
tells us that
n=1
2
n
(n + 1)
(n + 2)
also converges.
41. Writing an = 1/(2 + sin n), we have limnan does not exist, so the series diverges by Property 3 of Theorem 9.2.
45. Since ln(1 + 1/k) = ln((k + 1)/k) = ln(k + 1) ln k, the n
th
partial sum of this series is
Sn =
n
k=1
ln
_
1 +
1
k
_
=
n
k=1
ln(k + 1)
n
k=1
ln k
= (ln 2 + ln 3 + + ln(n + 1)) (ln 1 + ln 2 + + ln n)
= ln(n + 1) ln 1
= ln(n + 1).
Thus, the partial sums, Sn, grow without bound as n , so the series diverges by the denition.
49. Let Cn =
(2n)!
(n!)
2
. Then replacing n by n + 1, we have Cn+1 =
(2n + 2)!
((n + 1)!)
2
. Thus, with an = (2n)!x
n
/(n!)
2
, we have
|an+1|
|an|
= |x|
|Cn+1|
|Cn|
= |x|
(2n + 2)!/((n + 1)!)
2
(2n)!/(n!)
2
= |x|
(2n + 2)!
(2n)!

(n!)
2
((n + 1)!)
2
.
SOLUTIONS to Review Problems for Chapter Nine 213
Since (2n + 2)! = (2n + 2)(2n + 1)(2n)! and (n + 1)! = (n + 1)n! we have
|Cn+1|
|Cn|
=
(2n + 2)(2n + 1)
(n + 1)(n + 1)
,
so
lim
n
|an+1|
|an|
= |x| lim
n
|Cn+1|
|Cn|
= |x| lim
n
(2n + 2)(2n + 1)
(n + 1)(n + 1)
= |x| lim
n
4n + 2
n + 1
= 4|x|,
so the radius of convergence of this series is R = 1/4.
53. Here the coefcient of the n
th
term is Cn = n/(2n + 1). Now we have
an+1
an
((n + 1)/(2n + 3))x

n+1
(n/(2n + 1))x
n
=
(n + 1)(2n + 1)
n(2n + 3)
|x| |x| as n .
Thus, by the ratio test, the radius of convergence is R = 1.
57. We use the ratio test to nd the radius of convergence;
an+1
an
x
2(n+1)+1
(n + 1)!

n!
x
2n+1
x
2
n + 1
Since limn|x
2
|/(n + 1) = 0 for all x, the radius of convergence is R = . There are no endpoints to check. The
interval of convergence is all real numbers < x < .
Problems
61. To get $20,000 the day he retires he needs to invest a present value P such that P(1 +5/100)
20
= $20,000. Solving for
P gives the present value P = $20,000 1.05
20
. To fund the second payment he needs to invest $ 20,000 1.05
21
, and
so on. To fund the payment 10 years after his retirement he needs to invest $ 20,000 1.05
30
. There are 11 payments in
all, so
Total investment = 20,000 1.05
20
+ 20,000 1.05
21
+ . . . + 20,000 1.05
30
= 20,000 1.05
20
_
1 + 1.05
1
+ . . . + 1.05
10
_
= 20,000 1.05
20
_
1 1.05
11
1 1.05
1
_
= 20,000 1.05
20
8.7217
= $65,742.60.
65.
Present value of rst coupon =
50
1.04
Present value of second coupon =
50
(1.04)
2
, etc.
Total present value =
50
1.04
+
50
(1.04)
2
+ +
50
(1.04)
10
. .
coupons
+
1000
(1.04)
10
. .
principal
=
50
1.04
_
1 +
1
1.04
+ +
1
(1.04)
9
_
+
1000
(1.04)
10
=
50
1.04
_
1
_
1
1.04
_
10
1
1
1.04
_
+
1000
(1.04)
10
= 405.545 + 675.564
= $1081.11
69. A person should expect to pay the present value of the bond on the day it is bought.
Present value of rst payment =
10
1.04
Present value of second payment =
10
(1.04)
2
, etc.
Therefore,
10
1.04
+
10
(1.04)
2
+
10
(1.04)
3
+ .
This is a geometric series with a =
10
1.04
and x =
1
1.04
, so
10
1.04
1
1
1.04
= 250.
73. No. If the series
n=1
(1)
n1
an converges then, using Theorem 9.2, part 3, we have lim
n
(1)
n1
an = 0, which cannot
happen if lim
n
an = 0.
77. Since
an converge, we know that limnan = 0. Thus limn(1/an) does not exist, and it follows that
(1/an)
diverges by Property 3 of Theorem 9.2.
81. The series
n=1
_
1
n
+
1
n
_
=
n=1
2
n
diverges by Theorem 9.2 and the fact that
n=1
1
n
diverges.
The series
n=1
_
1
n

1
n
_
=
n=1
0 = 0
converges. But
n=1
1
n
diverges by Theorem 9.2 and the fact that
n=1
1
n
diverges.
Thus, if an = 1/n and bn = 1/n, so that
an and
bn both diverge, we see that
(an + bn) may diverge.

If, on the other hand, an = 1/n and bn = 1/n, so that

an and

bn both diverge, we see that

(an + bn)
may converge.
Therefore, if

an and

bn both diverge, we cannot tell whether

(an + bn) converges or diverges. Thus the
statement is true.
1. False. The rst 1000 terms could be the same for two different sequences and yet one sequence converges and the other
diverges. For example, sn = 0 for all n is a convergent sequence, but
tn =
_
0 if n 1000
n if n > 1000
is a divergent sequence.
5. False. The terms sn tend to the limit of the sequence which may not be zero. For example, sn = 1 + 1/n is a convergent
sequence and sn tends to 1 as n increases.
9. False. The sequence 1, 1, 1, 1, . . . given by sn = (1)
n
alternates in sign but does not converge.
13. False. This power series has an interval of convergence about x = 0. Knowing the power series converges for x = 1 does
not tell us whether the series converges for x = 2. Since the series converges at x = 1, we know the radius of convergence
is at least 1. However, we do not know whether the interval of convergence extends as far as x = 2, so we cannot say
whether the series converges at x = 2.
For example,
x
n
2
n
converges for x = 1 (it is a geometric series with ratio of 1/2), but does not converge for
x = 2 (the terms do not go to 0).
Since this statement is not true for all Cn, the statement is false.
17. True. Consider the series
(bn) and
(an). The series
(bn) converges, since
bn converges, and
0 an bn.
By the comparison test,
(an) converges, so
an converges.
21. False, since if we write out the terms of the series, using the fact that cos 0 = 1, cos = 1, cos(2) = 1, cos(3) = 1,
and so on, we have
(1)
0
cos 0 + (1)
1
cos + (1)
2
cos 2 + (1)
3
cos 3 +
= (1)(1) + (1)(1) + (1)(1) + (1)(1) +
= 1 + 1 + 1 + 1 + .
This is not an alternating series.
25. False. For example, if an = (1)
n1
/n, then
an converges by the alternating series test. But (1)

n
an = (1)
n
(1)
n1
/n =
(1)
2n1
/n = 1/n. Thus,
(1)
n
an is the negative of the harmonic series and does not converge.
29. True. Since the series is alternating, Theorem 9.9 gives the error bound. Summing the rst 100 terms gives S100, and if
the true sum is S,
|S S100| < a101 =
1
101
< 0.01.
33. False. Consider the series
n=1
1/n. This series does not converge, but 1/n 0 as n .
37. False. The alternating harmonic series
(1)
n
n
is conditionally convergent because it converges by the Alternating
Series test, but the harmonic series
(1)
n
n
1
n
is divergent. The alternating harmonic series is not absolutely
convergent.
41. True. Since the power series converges at x = 10, the radius of convergence is at least 10. Thus, x = 9 must be within
the interval of convergence.
45. True. The interval of convergence is centered on x = a, so a = (11 + 1)/2 = 5.
APPENDIX A SOLUTIONS 441
APPENDIX
Solutions for Section A
1. The graph is
-3 -2 -1 1 2 3
-40
-20
20
40
(a) The range appears to be y 30.
(b) The function has two zeros.
5. The largest root is at about 2.5.
9. Using a graphing calculator, we see that when x is around 0.45, the graphs intersect.
13. (a) Only one real zero, at about x = 1.15.
(b) Three real zeros: at x = 1, and at about x = 1.41 and x = 1.41.
17. (a) Since f is continuous, there must be one zero between = 1.4 and = 1.6, and another between = 1.6 and
= 1.8. These are the only clear cases. We might also want to investigate the interval 0.6 0.8 since f()
takes on values close to zero on at least part of this interval. Now, = 0.7 is in this interval, and f(0.7) = 0.01 < 0,
so f changes sign twice between = 0.6 and = 0.8 and hence has two zeros on this interval (assuming f is not
really wiggly here, which its not). There are a total of 4 zeros.
(b) As an example, we nd the zero of f between = 0.6 and = 0.7. f(0.65) is positive; f(0.66) is negative. So
this zero is contained in [0.65, 0.66]. The other zeros are contained in the intervals [0.72, 0.73], [1.43, 1.44], and
[1.7, 1.71].
(c) Youve found all the zeros. A picture will conrm this; see Figure A.1.
0.2 0.4 0.6 0.8 1 1.2 1.4
1.6
1.8
f() = sin 3 cos 4 + 0.8
Figure A.1
21.
1
2 4
4
5
x
f(x) = 4x x
2
Bounded and 5 f(x) 4.
Solutions for Section B
1. 2e
i/2
5. 0e
i
, for any .
9. 3 4i
13.
1
4

9i
8
17. 5
3
(cos
3
2
+ i sin
3
2
) = 125i
21. One value of
3
i is
3
e
i
2 = (e
i
2
)
1
3
= e
i
6
= cos

6
+ i sin

6
=
3
2
+
i
2
25. One value of (4 + 4i)
2/3
is [
32e
(i3/4)
]
(2/3)
= (
32)
2/3
e
(i/2)
= 2
5/3
cos

2
+ i2
5/3
sin

2
= 2i
3
4
29. We have
i
1
=
1
i
=
1
i

i
i
= i,
i
2
=
1
i
2
= 1,
i
3
=
1
i
3
=
1
i

i
i
= i,
i
4
=
1
i
4
= 1.
The pattern is
i
n
=
i n = 1, 5, 9,
1 n = 2, 6, 10,
i n = 3, 7, 11,
1 n = 4, 8, 12, .
Since 36 is a multiple of 4, we know i
36
= 1.
Since 41 = 4 10 + 1, we know i
41
= i.
33. To conrm that z =
a + bi
c + di
, we calculate the product
z(c + di) =
ac + bd
c
2
+ d
2
=
bc ad
c
2
+ d
2
i
(c + di)
=
ac
2
+ bcd bcd + ad
2
+ (bc
2
acd + acd + bd
2
)i
c
2
+ d
2
=
a(c
2
+ d
2
) + b(c
2
+ d
2
)i
c
2
+ d
2
= a + bi.
37. True, since
a is real for all a 0.

APPENDIX C SOLUTIONS 443
41. True. We can write any nonzero complex number z as re
i
, where r and are real numbers with r > 0. Since r > 0, we
can write r = e
c
for some real number c. Therefore, z = re
i
= e
c
e
i
= e
c+i
= e
w
where w = c + i is a complex
number.
45. Using Eulers formula, we have:
e
i(2)
= cos 2 + i sin 2
On the other hand,
e
i(2)
= (e
i
)
2
= (cos + i sin )
2
= (cos
2
sin
2
) + i(2 cos sin )
Equating real parts, we nd
cos 2 = cos
2
sin
2
.
49. Replacing by (x + y) in the formula for sin :
sin(x + y) =
1
2i
e
i(x+y)
e
i(x+y)
=
1
2i
e
ix
e
iy
e
ix
e
iy
=
1
2i
((cos x + i sin x) (cos y + i sin y) (cos (x) + i sin (x)) (cos (y) + i sin (y)))
=
1
2i
((cos x + i sin x) (cos y + i sin y) (cos x i sin x) (cos y i sin y))
= sin xcos y + cos xsin y.
Solutions for Section C
1. (a) f
(x) = 3x
2
+ 6x + 3 = 3(x + 1)
2
. Thus f
(x) > 0 everywhere except at x = 1, so it is increasing everywhere

except perhaps at x = 1. The function is in fact increasing at x = 1 since f(x) > f(1) for x > 1, and
f(x) < f(1) for x < 1.
(b) The original equation can have at most one root, since it can only pass through the x-axis once if it never decreases.
It must have one root, since f(0) = 6 and f(1) = 1.
(c) The root is in the interval [0, 1], since f(0) < 0 < f(1).
(d) Let x0 = 1.
x0 = 1
x1 = 1
f(1)
f
(1)
= 1
1
12
=
11
12
0.917
x2 =
11
12

f
11
12
11
12
0.913
x3 = 0.913
f(0.913)
f
(0.913)
0.913.
Since the digits repeat, they should be accurate. Thus x 0.913.
5. Let f(x) = sin x 1 + x; we want to nd all zeros of f, because f(x) = 0 implies sin x = 1 x.
Graphing sin x and 1 x in Figure C.2, we see that f(x) has one solution at x
1
2
.
/2
1
1
y = sin x
y = 1 x
Figure C.2
Letting x0 = 0.5, and using Newtons method, we have f
(x) = cos x + 1, so that

x1 = 0.5
sin(0.5) 1 + 0.5
cos(0.5) + 1
0.511,
x2 = 0.511
sin(0.511) 1 + 0.511
cos(0.511) + 1
0.511.
Thus sin x = 1 x has one solution at x 0.511.
9. Let f(x) = ln x
1
x
, so f
(x) =
1
x
+
1
x
2
.
Now use Newtons method with an initial guess of x0 = 2.
x1 = 2
ln 2
1
2
1
2
+
1
4
1.7425,
x2 1.763,
x3 1.763.
Thus x 1.763 is a solution. Since f
(x) > 0 for positive x, f is increasing: it must be the only solution.

Solutions for Section D
Exercises
1. The magnitude is 3
i =
3
2
+ 0
2
= 3.
The angle of 3
i is 0 because the vector lies along the positive x-axis.

5. 2v + w = (2 2)
i + (4 + 3)
j = 7
j .
9. Two vectors have opposite direction if one is a negative scalar multiple of the other. Since
5
j =
5
6
(6
j )
the vectors 5
j and 6
j have opposite direction. Similarly, 6
j and
j have opposite direction.

13. Scalar multiplication by 2 doubles the magnitude of a vector without changing its direction. Thus, the vector is 2(4
i
3
j ) = 8
i 6
j .
17. In components, the vector from (7, 7) to (9, 11) is (9 7)
i + (11 9)
j = 2
i + 2
j .
In components, the vector from (8, 10) to (10, 12) is (10 8)
i + (12 10)
j = 2
i + 2
j .
The two vectors are equal.
21. The velocity is v (t) = e
t
i + (1/(1 + t))
j . When t = 0, the velocity vector is v =
i +
j .
The speed is v =
1
2
+ 1
2
=
2.
The acceleration is a (t) = e
t
i 1/((1 + t)
2
)
j . When t = 0, the acceleration vector is a =
j .

Chapter 1 Calc Sinlge Multi

Uploaded by

Copyright:

Available Formats

Chapter 1 Calc Sinlge Multi

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 1 Calc Sinlge Multi

Uploaded by

Copyright:

Available Formats

APPENDIX A SOLUTIONS 269

a is real for all a 0.

(x) > 0 everywhere except at x = 1, so it is increasing everywhere

(x) = cos x + 1, so that

(x) > 0 for positive x, f is increasing: it must be the only solution.

i is 0 because the vector lies along the positive x-axis.

j have opposite direction. Similarly, 6

j have opposite direction.

j . When t = 0, the velocity vector is v =

j . When t = 0, the acceleration vector is a =

(0) = coefcient of x which equals 1, so f

(0); and since

(x) = (3/2)(1 +x)

(x) = (1/2)(3/2)(1 +x)

(x) = (1/2)(3/4)(1 +x)

(r) > 0 for r > r0 and V

(r) < 0 for r < r0. Thus,

1 +x. We use a Taylor polynomial with x = 1 to approximate

f(x) cos xdx =

f(x) cos nx = 0.)

f(x) sin xdx =

f(x) sin 2xdx =

f(x) sin 3xdx =

sin x and F3(x) =

denotes the biggest integer smaller than or equal to

cos kxsin kxdx +b

f(x) cos kxdx =

f(x) sin kxdx =

(x) = 6. The Taylor polynomial about x = 1 is

(0) = 0. To be specic let g

(0) > 0. Then

f(x) sin x(mx) dx = 0.

= 2x into the differential equation, we get

= 1 and x +y = x + (x 1) = 1, so this checks as a solution.

= 0.5(1 +y)(2 y), we will nd that y

> 0 if 1 < y < 2. Notice that if the

= 0, which only happens if y = 2 or y = 1. This also can be seen on

= 2x, with y(0) = 1

F is the outside temperature.

F. This is probably incorrect because the

F between 1 pm and 10 pm, the outside

F at 7 am, but not by muchnot enough to freeze.

y, where k is a positive constant. Separating variables,

2gR, we say that v0 is the escape velocity.

= 2 sin t + 3 cos t and y

= 2 cos t 3 sin t. Thus, y

+x = 0 are x = Acos t +Bsin t;

+ 4x = 0 are x = Acos 2t +Bsin 2t;

+ 16x = 0 are x = Acos 4t +Bsin 4t.

(0) = 0 = 4Asin 0 + 4Bcos 0, so B = 0. Thus, the solution is

= 3 sin 3t, we see that y

+ 4x = 0 represents an undamped oscillator, and so goes with (IV).

4x = 0 has characteristic equation r

+ 1.01x = 0 has characteristic equation r

term represents an amplifying force. This is reected in the solution by e

+ 1.01x has characteristic equation r

+ cs = 0 is overdamped if the discriminant b

= sin x, whereas (II) matches

= cos x. The slope in (V) tends to zero as x , so this must match y

= 4; so y = 4(0.2) = 0.8. Thus, y1 = 1 + 0.8 = 1.8.