Calculus II

Exercises

Brian E. Blank

December 10, 2015

first president university press

ii
Preface

This document contains exercises to accompany the second semester calculus course taught at First President
University.
The intention of these exercises is not only to help students test their understanding of both the theory
and practice of the course material, but also to serve as a guide of the material that will be covered on the
exams. If the course syllabus indicates that a section of the committee-chosen course textbook is covered,
and if this document has no exercises related to a particular topic in that section, then that particular topic
will not be the source of problems on the exams. For instance, Applications to Physics and Engineering is
listed as being covered in the course syllabus. But there are no problems here that pertain to the subsection
concerning hydrostatic pressure. That means that hydrostatic pressure will not appear on the examinations.
The preceding paragraph is not a promise that exam questions will be just like the exercises found here.
There is always the hope that a course will nurture creativity and intellectual independence. Even at First
President University. Former Yale professor William Deresiewicz, in his critique1 of American university ed-
ucation, wrote

I think it [Yale] probably deserves its reputation as the best among elite universities (as distinct
from liberal arts colleges) at nuturing creativity and intellectual independence. Notoriously pre-
professional places like Penn, Duke, or [sic] First President University, or [sic] notoriously anti-
intellectual ones like Princeton or [sic] Dartmouth, are clearly far worse.

Given that the creativity and intellectual independence nurtured at Yale is of the sort that allows its Jane
Austen scholars to confuse the conjunctions “or” and “and”, it is hard to be certain that First President
University is intended to be included among the institutions that are far worse than Yale. Whatever the
intent, we need not be bound by any perception. It is likely that some exam problems will have some novelty.
Certain topics are like one-trick ponies. Do one problem and you can do them all. Other topics require a
variety of techniques. The number of exercises per section reflects that. For instance, the section on Improper
Integrals contains many exercises. They all have something in common, the evaluation of an improper integral,
but there are also significant differences between the exercises because very different integration techniques
are needed for the evaluations.
A few words about notation are in order. These notes use strict syntax for the evaluation of functions.
There are three parts to the notation for the value of a function—the name of the function, a pair of paren-
theses, and the argument at which the function is evaluated. (In Calculus I and II, sometimes called single
variable calculus, we study functions of one variable, so the word argument is singular. In Calculus III,
sometimes called multivariable calculus, the word argument is replaced by sequence of arguments.)
The name of a function is usually a letter such as f , or g, or F , or G. Popular Greek letters for function
names are ϕ, or ψ, or Φ, or Ψ.
1 Excellent Sheep: The Miseducation of the American Elite and the Way to a Meaningful Life, Free Press, 2014.

iii
iv

The argument is the object at which the function is evaluated. The argument might be a number, a
variable, or an algebraic expression. Generally, the argument is placed inside
( a pair
) of parentheses that serve
as delimiters. Thus, for the function f , we might have f (2.75), f (x), or f x2 + 3 . In this document, and on
your exams, the argument will always be enclosed inside parentheses. No exceptions!. In the mathematical
literature, trigonometric and certain other functions such as logarithms are treated differently: commonly the
functional parentheses are omitted. Not so here. Thus, sin x will always be written as sin(x).
When strict functional notation is used, certain other parentheses can be omitted. Consider, for example,
the expression sin(x)2 . There is only one valid way to interpret this when strict functional notation is in force:
(x)2 cannot be the argument of the sine function because (x)2 is not surrounded by parentheses. Therefore,
the square must be applied to sin(x). Because there is no other way to interpret sin(x)2 , we need not write it
( )2
as sin(x) . Similarly, f (x)2 means the square of f (x). This square may be regarded as the evaluation at x
of a function that we denote by f 2 . In other words, f 2 (x) is an alternative notation for f (x)2 . To illustrate
with the sine function, we have
( )2 ( )2
sin2 (x) = sin(x)2 = sin(x) = sin x ,

but the last of these expressions will not be used in this document.
The expression ex signifies the constant e raised to the power x. The quantity ex is a function of x. This
function is known as the exponential function. Sometimes the name natural exponential function is used. As
with the trigonometric functons, a three letter abbreviation, exp has been introduced for its name. Thus,

exp(x) = ex .

This notation has the advantage of emphasizing the functional relationship of the expression. Its main benefit,
however is typographical: it brings the argument x down to the base level of the line, where symbols are not
rendered with a smaller
( ) font. This is particularly desirable when the argument is itself a power. Compare,
2
for example, exp x2 with ex .
One particularly important function is the primary object of study in Calculus I. It is the derivative
function. Unlike the functions discussed above, the derivative function is not evaluated at real numbers: it
is evaluated at functions. Likewise, a value of the derivative function is not a number: it is a function. The
derivative function is often denoted by D. If f (x) = x2 , for example, then the value of D(f ) is the function
g where g(x) = 2 x. That is, D(f ) = g and D(f )(x) = g(x) = 2 x. Using the prime notation for a derivative,
we have D(f )(x) = f ′ (x), but the former notation better emphasizes the functional nature of the derivative.
Solutions to most of the exercises, and answers to the rest, are included in this document, but you will
have to page forward to find them. Solutions to an exercise set are not located after that exercise set. All
solutions are at the back of the book. That is, the first solution follows the very last exercise.
0
Areas and Distances (Corresponds to
Stewart 5.1)

1. The graph of a function f appears in the figure below.

In accordance with the specific instructions given, use rectangles to approximate the area of the region
that is under the graph of f and above the interval [0, 6] of the x-axis. (For your interest, the exact area
is 34.83571429, to eight decimal places.)

i) Use 6 equal width intervals. The height of each rectangle is the value of f at the left endpoint of the
rectangle’s base.

ii) Use 6 equal width intervals. The height of each rectangle is the value of f at the right endpoint of
the rectangle’s base.

iii) Use 3 equal width intervals. The height of each rectangle is the value of f at the midpoint of the
rectangle’s base.

2. The graph of a function f appears in the figure below.

1
2

In accordance with the specific instructions given, use rectangles to approximate the area of the region
that is under the graph of f and above the interval [1, 3] of the x-axis. (For your interest, the exact area
is 11.05516755, to eight decimal places.)

i) Use 8 equal width intervals. The height of each rectangle is the value of f at the left endpoint of the
rectangle’s base.

ii) Use 8 equal width intervals. The height of each rectangle is the value of f at the right endpoint of
the rectangle’s base.

iii) Use 4 equal width intervals. The height of each rectangle is the value of f at the midpoint of the
rectangle’s base.
3. The speed v(t) (in ft/s) of a car measured at 10s intervals is:

t 0 10 20 30 40 50 60
v(t) 54 57 61 66 65 63 62

In accordance with the specific instructions given, approximate the distance the car travelled during the
minute in which the speed measurements were taken.

i) Use 3 equal width intervals and the values of v(t) at the left endpoints.

ii) Use 3 equal width intervals and the values of v(t) at the right endpoints.

iii) Use 3 equal width intervals and the values of v(t) at midpoints.
3

iv) Use 6 equal width intervals and the values of v(t) at the left endpoints.

v) Use 6 equal width intervals and the values of v(t) at the right endpoints.

vi) Use 6 equal width intervals and the averages of the values of v(t) at left and right endpoints.
4
Riemann Sums, Riemann Integrals,
Definite Integrals, Midpoint Rule
(Corresponds to Stewart 5.2)

√ ∑3
1. For f (x) = x on the interval [3, 4], calculate the Riemann sums j f (sj ) ∆x where the sample points
s1 , s2 , s3 are as follows:

i) left endpoints
ii) right points
iii) midpoints
iv) s1 = 47/15 and the distance between consecutive sample points is ∆x.

∫ 4
1
2. Estimate the Riemann integral dx by the Riemann sum that arises from 4 subintervals and the
1 x
1
choice of sample points s1 , s2 , s3 , s4 in which each sample point sj is a distance ∆x from the left
3
endpoint xj−1 of the subinterval [xj−1 , xj ] that contains it.

∫ 2
3. Estimate the definite integral 3 x2 dx by using the Midpoint Rule with n = 5. (The exact value is
1
7.)

4. The plot of f (x) = (x − 1) (x − 2) (x − 4) = x3 − 7 x2 + 14 x − 8, 1 ≤ x ≤ 4 is shown in the figure

below.

5
6

Notice that the verical axis shown is the line x = 1, not the y-axis.

i) The area of the region under the graph of f and above the interval [1, 2] of the x-axis is 5/12. The
area of the region above the graph of f and under the interval [2, 4] of the x-axis is 8/3. Use the given
∫4
information to evaluate 1 f (x) dx.

∫2
ii) Estimate the definite integral 1
f (x) dx using the Midpoint Rule with n = 3.

∫4
iii) Estimate the definite integral 2
f (x) dx using the Midpoint Rule with n = 6.

∫4
iv) Estimate the definite integral 1
f (x) dx using the Midpoint Rule with n = 9.

∫ 3
5. Let f (x) = 8 x − 43 x + 70 x − 21. In this exercise, the definite integral
3 2
f (x) dx is to be estimated
1/2
∑
5
by two Riemann sums f (sj ) ∆x. The sample points used will be specified momentarily. As an aid
j=1
to visualizing the locations of the sample points, the following graph of f over the interval of integration
may be helpful. For each node xj , a vertical line segment x = xj has been drawn from the x-axis to the
graph of f .
7

i) For each j, the sample point sj in the j th subinterval satisfies f (x) ≤ f (sj ) for all x in the j th
subinterval. The resulting Riemann sum R∗ is called an upper Riemann sum. It is an overestimate
of the definite integral it approximates.

ii) For each j, the sample point sj in the j th subinterval satisfies f (sj ) ≤ f (x) for all x in the j th
subinterval. The resulting Riemann sum R∗ is called a lower Riemann sum. It is an underestimate
of the definite integral it approximates.

6. The interval [1, 5] is partitioned into N subintervals of equal width. Let s1 denote the midpoint of the
first subinterval, s2 the midpoint of the second subinterval, and so on, with sN denoting the midpoint
of the last subinterval. Consider the limit
∑
N
4 s2j − 1
lim .
N →∞ N
j=1

(i) Identify the sum inside the limit as a Riemann sum, and identify the limit of these Riemann sums
as a Riemann integral.
(ii) After the first part of the Fundamental Theorem of Calculus has been covered (in the next section),
evaluate the given limit.
7. Identify
N (
∑ )2
2j − 1 6
1+
j=1
N N
as a Riemann sum for a certain function f over the interval [1, 3] partitioned into N subintervals of
equal width. After the first part of the Fundamental Theorem of Calculus has been covered (in the next
section), evaluate the limit of these Riemann sums,
∑N ( )2
2j − 1 6
lim 1+ .
N →∞ N N
j=1
8
The Fundamental Theorem of
Calculus, Parts 1 and 2 (Corresponds
to Stewart 5.3)
∫ 2 ( 2 )
1. Calculate 6x + 2x − 4 dx.
1
∫ 9 √
2. Calculate x dx.
4
∫ π
3. Calculate sin(θ) dθ.
π/3
∫ 4
4+x
4. Calculate √ dx.
1 x
∫ 2 √
5. Calculate (2 − w) 2w dw.
0
∫ 1
6. Calculate (exp(t) + te ) dt. (Note: exp(t) = et . The notation is typographically handy.)
0
∫ 4
√
4
7. Calculate ds.
1 s
∫ π/3
8. Calculate sec2 (α) dα.
π/6

∫ π/3
9. Calculate 3 csc(t) cot(t) dt.
π/4
∫ 1 ( )−1
10. Calculate 1 + x2 dx.
0
∫ 6
2
11. Calculate dz.
3 z
∫ √
3/2
1
12. Calculate √ dx.
1/2 1 − x2

9
10

∫ 1
3x2 − 1 B
13. Calculate dx. (Rewrite the integrand as A + for certain constants A and B).
0 x2 + 1 x2
+1
∫ π/3
14. Calculate tan2 (α) dα. (Use a trigonometric identity and then Exercise 8.)
π/6
∫ 1
15. Calculate 32+t dt.
0
∫ x √
16. Calculate F ′ (x) and F ′ (2) for F (x) = 1 + t3 dt.
0
∫ s
′ ′ 19 + u4
17. Calculate G (s) and G (3) for G(s) = 2
du.
0 1+u
∫ 2π √
18. Calculate D(F )(x) and D(F )(π/6) for F (x) = 17 + 16 sin(s) ds.
x
Note: D(F )(x) is an alternative notation for F ′ (x).
∫ √
x2 √
19. Calculate D(F )(x) and D(F )(3) for F (x) = 1 + t dt.
1
∫ 2π (π )
20. Calculate G′ (x) and G′ (2) for G(x) = dt. sin
t 3x
∫ x3 ( )
′ ′ 16
21. Calculate F (x) and F (2) for F (x) = exp dz.
x2 z
The So-Called Net Change Theorem
(Corresponds to Stewart 5.4)

Note: This section of the text is primarily a repetition of material from previous sections. The exercises
provided for this section are limited to the Net Change Theorem. In fact, even the “Net Change Theorem”
is merely a reformulation of the part of the Fundamental Theorem of Calculus that concerns the evaluation
of definite integrals.

1. For t ≥ 0 measured in minutes, the rate at which water flows from a tank is 120 exp(−t) liters per
minute. By how many liters has the volume of water in the tank decreased between t = 2 and t = 3.

2. Suppose that the tank in the preceding exercise initially contained 120 liters of water. After how many
minutes is the volume in the tank equal to 60 liters?

3. At time t = 0, Carl, the very hungry caterpillar, encountered a tasty leaf. For the first two minutes
thereafter, Carl’s rate of consumption was 60 mg/min. At time t = 2 min, Carl, no longer so very
hungry, consumed leafy matter at the rate of 240/t2 mg/min. How many mg of leaf did Carl eat in the
first 4 min of his feast?

4. The rate of change of energy with respect to time is called power. The usual metric system unit of
energy is the Joule, abbreviated by J. The usual metric system unit of power is the watt, abbreviated by
w, which is a change of 1 J per second (i.e., 1 w = 1 J/s). A megawatt, abbreviated MW, is one million
watts (i.e., 1 MW = 106 w = 106 J/s. Notice that

J
1 MW × 1 hr = 106 × 3600 s = 3.6 × 109 J.
s

This quantity is called a megawatt-hour, and it is clear from the right side that it is a unit of energy.

The figure below shows the graph of electric power P (t) (measured in megawatts) as a function of time
t (measured in hours) for one day in San Francisco.

11
12

(Confession: For your calculating pleasure, the author rounded each of P (0), P (3), P (6), P (9), P (12),
P (15), P (18), P (21), P (24) to the nearest multiple of 50. Hands up—who thought the figure is meant
to depict a chapeau? An elephant eaten by a snake?) Approximate the electric energy E used in San
Francisco that day. Use the megawatt-hour as the unit of measurement.
The Substitution Rule (Corresponds
to Stewart 5.5)
∫ 7 √
3
1. Calculate 1 + 9 x dx.
0
∫ 4 √
2. Calculate 5 − x dx.
1
∫ 4 √
3. Calculate x 5 − x dx.
1
∫ 4 √
4. Calculate x2 5 − x dx.
1
∫ 2 √
5. Calculate x 4 − x2 dx.
0
∫ 2 √
6. Calculate x3 4 − x2 dx.
0
∫ 1
2x − 1
7. Calculate dx.
0 x2 + 1
∫ 1
x2 + 4 x + 2
8. Calculate dx.
0 x2 + 1
∫ 1 √
8 − 7 x dx.
3
9. Calculate x
0
∫ 0 √
3
10. Calculate x2 1 + x dx.
−1
∫ 2
ln(x)
11. Calculate dx.
1 x
∫
1
12. Calculate dx.
x ln(x)
∫ e ( )
1 1
13. Calculate ln dx.
1 x x

13
14

∫ 4 √
exp ( x)
14. Calculate √ dx.
1 x
∫ π/4
15. Calculate tan(x) sec2 (x) dx.
0
∫ π/3
16. Calculate tan2 (x) sec2 (x) dx.
0
Areas Between Curves (Corresponds
to Stewart 6.1)

1. Let R be the region that is bounded above by the horizontal line segment y = 1, 1 ≤ x ≤ e, bounded
below by y = ln(x), 1 ≤ x ≤ e, and bounded on the left by the vertical line segment of x = 1, 0 ≤ y ≤ 1.
Express the area of R in two ways: as an integral with respect to x, and as an integral with respect to
y. Evaluate the latter. (The method for evaluating the former will be learned in the next chapter.)
2. Calculate the area of the region that is between arcs of y = sec(x) and y = cos(x) for 0 ≤ x ≤ π/3.
3. Calculate the area of the region that is bounded above by a segment of the line y = x + 1 and below by
an arc of the parabola y = x2 + 1.
4. Calculate the area of the region that is between arcs of the curves y = x3 − x + 1 and y = 2 x2 − 1 for
−1 ≤ x ≤ 2.
5. Calculate the area of the region that is between arcs of the curves y = cos(x) and y = sin(x) for
π/4 ≤ x ≤ 5π/4.
6. Let R be the region with boundary that is composed of a segment of the line y = 3x, a segment of the
line y = x + 2, and an arc of the parabola y = x2 . Express the area of R as a sum of integrals of the
∫ b
form h(x) dx and evaluate the area.
a

7. Let R be the region described in the preceding exercise. Express the area of R as a sum of integrals of
∫ d
the form h(y) dy and evaluate the area.
c

8. Calculate the area of the region that is bounded above by an arc of the graph of y = ln(x) and below
by a segment of the line y = (x − 1)/(e − 1).
9. Let R be the region that is bounded above by a segment of the horizontal line y = 1, below by a segment
of the horizontal line y = −1, on the left by an arc of the graph of the parabola x = y 2 − 2, and on the
∫ d
right by an arc of the graph of y = ln(x). Express the area of R as an integral of the form h(y) dy
c
and evaluate the area.
10. Let R be the region described in the preceding exercise. Express the area of R as a sum of integrals of
∫ b
the form h(x) dx.
a

11. Let R be the region that is inside the circle x2 + y 2 = 25 and above the line segment joining the point
∫ b
(−4, −3) to the point (3, 4). Express the area of R as a sum of integrals of the form h(x) dx.
a

15
16

12. Depths of a swimming pool are measured in meters every 2 m.

(i) Use midpoints to estmate the volume of the pool. Upon how many subintervals is the Riemann sum
based? What length do the subintervals have? What is the volume estimate? (ii) Use the averages of
the left and right endpoints to estimate the volume of the pool. Upon how many subintervals estimate
based? What length do the subintervals have? What is the volume estimate?
13. From west to east, a city is 5.1 km wide. North-south measurements in km are given at 0.85 km intervals
in the figure.

(i) Use midpoints to estmate the city’s area. Upon how many subintervals is the Riemann sum based?
What length do the subintervals have? What is the area estimate? (ii) Use left endpoints to estimate the
city’s area. Upon how many subintervals is the Riemann sum based? What length do the subintervals
have? What is the area estimate?
Volumes by Disks and Washers
(Corresponds to Stewart 6.2)

1. Let R be the region bounded above by y = x2 , 1 ≤ x ≤ 2, below by the x-axis, and laterally by
segments of the lines x = 1 and x = 2. What is the volume of the solid that results when R is rotated
about the x-axis?
√
2. Let R be the region bounded above by y = cos(x), −π/2 ≤ x ≤ π/2 and below by the x-axis. What
is the volume of the solid that results when R is rotated about the x-axis?

3. Let R be the region bounded above by y = sec(x), 0 ≤ x ≤ π/4, below by the x-axis, and laterally
by segments of the lines x = 0 and x = π/4. What is the volume of the solid that results when R is
rotated about the x-axis?
√
4. Let R be the region bounded above by y = x, 4 ≤ x ≤ 9, below by the x-axis, and laterally by
segments of the lines x = 4 and x = 9. What is the volume of the solid that results when R is rotated
about the x-axis?
√
5. Let R be the region bounded on the right by y = x, above by y = 4, on the left by the y-axis, and
below by y = 3. What is the volume of the solid that results when R is rotated about the y-axis?

6. Let R be the region bounded on the right by y = x4 , above by y = 81, on the left by the y-axis, and
below by y = 16. What is the volume of the solid that results when R is rotated about the y-axis?

7. Let R be the region bounded on the right by y = 1/x, above by y = 5, on the left by the y-axis, and
below by y = 1/3. What is the volume of the solid that results when R is rotated about the y-axis?

8. Let R be the region in the first quadrant bounded by y = x3 and y = 2x2 . What is the volume of the
solid that results when R is rotated about the x-axis?
√
9. Let R be the region bounded above by y = exp(x), below by y = x + 1, and on the right by a segment
of the line x = 1. What is the volume of the solid that results when R is rotated about the x-axis?

10. Let R be the region bounded on the left by y = 4 − x2 , 0 ≤ x ≤ 2, on the right by y = (x − 2)2 , 2 ≤
x ≤ 4, and on the top by a segment of the line y = 4. What is the volume of the solid that results when
R is rotated about the y-axis?
√
11. Let R be the region bounded by y = 3 x and y = x. What is the volume of the solid that results when
R is rotated about the y-axis?

12. Let R be the region bounded on the left by y = ln(x), below by the x-axis, and on the right by a
segment of the line y = 2 − x/e. What is the volume of the solid that results when R is rotated about
the y-axis?

17
18

13. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line x = −2. Express V by means of an integral.

14. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line x = 2. Express V by means of an integral.

15. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line y = 1/2. Express V by means of an integral.

16. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line y = 6. Express V by means of an integral.

17. Let R be the region bounded above by y = 4 − x2 and below by y = x + 2. Let V be the volume of the
solid that results when R is rotated about the line y = 5. Express V by means of an integral.

18. Let R be the region bounded on the left by y = 4 − x2 , on the right by y = x + 2, and above by y = 3.
Let V be the volume of the solid that results when R is rotated about the line x = 5. Express V by
means of an integral.
Volumes by Cylindrical Shells
(Corresponds to Stewart 6.3)
√
1. Let R be the region bounded by y = x − x2 , 0 ≤ x ≤ 1 and the x-axis. What is the volume of the
solid that results when R is rotated about the y-axis?
( )
2. Let R be the region bounded above by y = 4/ 2 + x2 , by x = 2, on the right, by the y-axis on the
left, and by the x-axis below. What is the volume of the solid that results when R is rotated about the
y-axis?
3. Let R be the region bounded above by y = 4, bounded on the right by y = x2 , and bounded by the
y-axis on the left. What is the volume of the solid that results when R is rotated about the x-axis?
√
4. Let R be the region bounded by y = x, 1 ≤ x ≤ 4, by y = 2, 4 ≤ x ≤ 5, by y = x − 3, 4 ≤ x ≤ 5,
and by y = 1, 1 ≤ x ≤ 4. What is the volume of the solid that results when R is rotated about the
x-axis?

5. Let R be the region bounded by y = sin(x), 0 ≤ x ≤ π and the x-axis. Let V be the volume of the
solid that results when R is rotated about the line x = 4. Express V by means of an integral.

6. Let R be the region bounded by y = sin(x), 0 ≤ x ≤ π/4, by y = cos(x), 0 ≤ x ≤ π/4, and by the
y-axis. Let V be the volume of the solid that results when R is rotated about the line x = −π. Express
V by means of an integral.

7. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line x = −2. Express V by means of an integral.

8. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line x = 2. Express V by means of an integral.

9. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line y = 1/2. Express V by means of an integral.

10. Let R be the region bounded by y = x2 , 1 ≤ x ≤ 2 and y = 3 x − 2, 1 ≤ x ≤ 2. Let V be the volume

of the solid that results when R is rotated about the line y = 6. Express V by means of an integral.

19
20
Work (Corresponds to Stewart 6.4)

I like work; it fascinates me. I can sit and look at it for hours.
Jerome K. Jerome

1. If a spring with spring constant 8 pounds per inch is stretched 7 inches beyond its equilibrium position,
then how much work is done?

2. If a spring with spring constant 20 newtons per centimeter is stretched 12 cm beyond its equilibrium
position, then how much work is done?

3. If the amount of work in stretching a spring 0.02 meter beyond its equilibrium position is 8 J, then what
force is necessary to maintain the spring at that position?

4. If 5 J work is done in extending a spring 0.2 m beyond its equilibrium position, then how much additional
work is required to extend it a further 0.2 m?

5. A spring is stretched 10 cm beyond its equilibrium position. If the force required to maintain it in its
stretched position is 240 N, then how much work was done?

6. A force of 280 pounds compresses a spring 4 inches from its natural length. How much work is done
compressing it a further 4 inches?

7. If 40 joules of work are done in stretching a spring 8 cm beyond its natural length, then how much work
is done stretching a further 4 cm?

8. The work done in extending a spring at rest two inches beyond its equilibrium position is equal to 2/3
ft-lb. How much force is required to maintain the spring in that position?

9. A tank completely filled with water has square base of side 15 feet. It is 10 feet deep and is being
pumped dry. The pump floats on the surface of the water and pumps the water to the top of the pool,
at which point the water runs off. How much work is done in emptying the pool?

10. A swimming pool has rectangular base with side lengths of 6 m and 10 m. The pool is 3 m deep but
not completely filled: the depth of the water it contains is 2 m. A pump floats on the surface of the
water and pumps the water to the top of the pool, at which point the water runs off. How much work
is done emptying the pool?

11. A swimming pool has rectangular base with side lengths of 16 and 30 feet. The pool is 10 feet deep but
not completely filled: the depth of the water it contains is 8 feet. A pump floats on the surface of the
water and pumps the water to the top of the pool, at which point the water runs off. How much work
has the pump performed when half the water has been pumped?

21
22

12. A semi-cylindrical tank filled with water is 16 feet long. It has rectangular horizontal cross-sections and
vertical cross-sections that are semicircular of radius 6 feet. See the figure below. A pump which floats
on the surface of the water pumps the water to the top, at which point the water runs off. How much
work is done in emptying the tank?

13. A filled reservoir is in the shape of a frustum of a cone. The radius of the circular base of the frustum
is 50 ft. The radius of the circular top of the frustum is 25 ft. The depth of the frustum at the center is
50 ft. A pump that floats on the surface of the water pumps water to the top of the reservoir, at which
point the water runs off. How much work is done emptying the reservoir? (A frustum of a cone is a
cone with the top lopped off by a chop parallel to the base. King Henry VIII once had several cones,
but he called out, “Off with their heads.” Then he had frustums.)

14. Francis stands at the top of a tall building and pulls a chain up the side of the building. The chain is
50 feet long and weighs 3 pounds per linear foot. How much work does Francis do in pulling the chain
to the top?
15. Francis stands at the top of a tall building and pulls a chain up the side of the building. The chain is 50
feet long and weighs 3 pounds per linear foot. How much work does Francis do in pulling up the first
30 feet of the chain to the top of the building?

16. A heavy uniform cable hangs over the side of a building that is 40 m high. From the roof, it is used to
lift a 300 kg load from ground level to a height 24 meters above ground level. If the cable weighs 220
newtons per linear meter, then how much work is done?
17. A heavy uniform cable is used to lift a 300 pound load from ground level to the top of a 100 foot high
building. If the cable weighs 20 pounds per linear foot then how much work is done?
18. With regard to the preceding exercise, how much work is done in lifting the load from a height 30 feet
above ground level to a height 80 feet above ground level?
Average Value (Corresponds to
Stewart 6.5)

In each of Exercises 1–7, a function f and an interval I = [a, b] are given. Calculate favg , the average value
of f over the interval I. Find a number c in the interval (a, b) for which f (c) is equal to favg .

1. f (x) = sin(x) I = [0, π/3]

2. f (x) = √
1/x I = [2, 4]
3. f (x) = x/4 I = [0, 4]
4. f (x) = x2 + 2 x + 2; I = [0, 3]
5. f (x) = 60/x2 I = [1, 3]
6. f (x) = (x√− 1)1/2 I = [1, 5]
7. f (x) = x 169 − x2 I = [5, 12]

23
24
Integration by Parts (Corresponds to
Stewart 7.1)
∫
1. Calculate x exp(x) dx
∫
2. Calculate x exp(−x) dx
∫
3. Calculate (2x + 5) ex/3 dx
∫
4. Calculate x sin (x) dx
∫
5. Calculate (4x + 2) sin (2x) dx
∫
6. Calculate 9x cos (3x) dx
∫
7. Calculate x ln (4x) dx
∫
8. Calculate 9 x2 ln (x) dx
∫
ln (x)
9. Calculate dx
x2
∫
10. Calculate 16 x3 ln (2x) dx
∫2
11. Calculate 0
xe−x/2 dx
∫π
12. Calculate 0
x cos (x) dx
∫ π/2
13. Calculate 0
4x sin (x/3) dx
∫e
14. Calculate 1
ln (x) dx
∫1
15. Calculate 1/3
x ln (3x) dx
∫ 4
ln (x)
16. Calculate √ dx
1 x
∫1
17. Calculate 0
x3x dx
∫
18. Integrate by parts successively to evaluate x2 e−x dx
∫
19. Integrate by parts successively to evaluate x2 cos (x) dx
∫
20. Integrate by parts successively to evaluate 16 x3 ln2 (x) dx

25
26

∫1
21. Calculate 0
2 arctan (x) dx
∫1
22. Calculate 0
arcsin (x) dx

Suppose that a is a nonzero constant. In the next two exercises, use the reduction formula
∫ ∫
1 p ax p
x e dx = x e −
p ax
xp−1 eax dx
a a
to evaluate the given integral.
∫1
23. 0 x2 ex/2 dx
∫1
24. 0 4x2 e−2x dx
∫ √
25. Simplify the integrand before integrating by parts: 2 ln ( x) dx
∫ ( )
26. Simplify the integrand before integrating by parts: 4 x ln x5 dx
∫
27. Simplify the integrand before integrating by parts: 16x3 ln (1/x) dx
∫ √
28. Calculate 9 x ln (x) dx
∫
29. Calculate x sec2 (x) dx
∫1
30. Evaluate −2 x (x + 3)−1/2 dx. For the first step, integrate by parts with u = x.
∫2
31. Evaluate x (x − 1)1/2 dx. For the first step, integrate by parts with u = x.
1
∫ √
32. Make an appropriate substitution before integrating by parts to evaluate sin ( x) dx
∫ ( )
33. Make an appropriate substitution before integrating by parts to evaluate 2x3 exp x2 dx
∫ ( )
34. Make an appropriate substitution before integrating by parts to evaluate 3x5 sin x3 dx
Trigonometric Integrals (Corresponds
to Stewart 7.2)

∫ π/2
1. Evaluate sin2 (x/2) dx.
0
∫ π/4
2. Evaluate 48 cos2 (2x) dx
π/6

∫ 2π/3
3. Evaluate 3 tan2 (x/2) dx
0
∫
( )
4. Evaluate cos2 (x) − sin2 (x) dx.

∫ π/4
5. Evaluate 48 cos3 (2x + π) dx.
π/6
∫
6. Evaluate 7 sin7 (x/10) dx
∫ (x) (x)
7. Evaluate 5 sin2 cos3 dx.
3 3
∫ ( )√ ( )
3 2x 2x
8. Evaluate 7 sin cos dx.
3 3
∫
cos3 (x)
9. Evaluate dx.
sin (x)
∫
5 cos3 (2x)
10. Evaluate √ dx
sin(2x)
∫ π
11. Evaluate sin3 (x) cos4 (x) dx.
0
∫ π/2 √
12. Evaluate sin (x) cos3 (x) dx.
0
∫ 3π/2
13. Evaluate 5 sin2 (x/3) cos7 (x/3) dx.
0

27
28

14. Use the reduction formula

∫ ∫
1 n−1
sin (u) du = − cos(u) sin
n n−1
(u) + sinn−2 (u) du
n n
to show that ∫
1 3 3
sin4 (u) du = − sin3 (u) cos(u) − cos(u) sin(u) + u + C.
4 8 8
15. Use the reduction formula
∫ ∫
n 1 n−1 n−1
cos (u) du = cos (u) sin(u) + cosn−2 (u) du
n n
to show that ∫
1 3 3
cos4 (u) du = cos3 (u) sin(u) + cos(u) sin(u) + u + C.
4 8 8
∫ 1
16. Use the preceding exercise to evaluate cos4 (πx) dx
0
∫ π/4
17. Use either Exercise 14 or Exercise 15 to evaluate cos2 (t) sin2 (t) dt.
0
∫ π
18. Use Exercise 15 two times (or three, if you prefer) to evaluate cos6 (x/2) dx.
0
∫ π/3
19. Use Exercise 14 two times (or three, if you prefer) to evaluate 64 cos2 (x) sin4 (x) dx.
π/6
∫
20. Use the identity tan2 (x) = sec2 (x) − 1 to evaluate tan3 (x) dx.

21. Use the identity tan2 (x) = sec2 (x) − 1 and the integral formula
∫
1 1
sec3 (u) du = sec(u) tan(u) + ln ( | sec(u) + tan(u) | ) + C
2 2
∫
to evaluate tan2 (x) sec(x) dx.
∫
22. Use the identity tan (x) = sec (x) − 1 to evaluate
2 2
tan3 (x) sec(x) dx.
Trigonometric Substitution
(Corresponds to Stewart 7.3)

∫ √
3/2
24 x2
1. Evaluate √ dx.
0 1 − x2
∫ 3 √
2. Evaluate 2 25 − x2 dx.
0
∫
( )3/2
3. Evaluate 4 − x2 dx.
∫
2
4. Evaluate √ dx.
1 + 4x2
∫
2x3
5. Evaluate dx.
9 + x2
∫
16 x2
6. Evaluate √ dx. The integral formula
3 + 4 x2
∫
1 1
sec3 (θ) dθ = sec(θ) tan(θ) + ln ( | sec(θ) + tan(θ) | ) + C
2 2
might be useful.
∫
36
7. Evaluate 2 dx.
(9 + 4 x2 )
∫
1
8. Evaluate √ dx.
x − 25
2
∫ √
9. Evaluate 9x2 − 4 dx.
∫
1
10. Calculate √ dx.
4 x − x2
∫ −2
1
11. Calculate dx.
−4 x2 + 8 x + 20
∫
x2
12. Convert 3/2
dx to a trigonometric integral by making a substitution of the form
(25 x2 + 20 x + 3)
x = a t(θ) + b where a and b are nonzero constants, possibly negative, and t(θ) is a trigonometric
function. Do not evaluate the trigonometric integral that results.

29
30
Using Partial Fractions to Integrate
Rational Functions (Corresponds to
Stewart 7.4)
∫
x − 11
1. Evaluate dx.
x2 − x − 2
∫
2
2. Evaluate dx.
x3 − x
∫
4 x2 + 3 x + 2
3. Evaluate dx.
x3 + x2
∫
6 x2 + 17 x − 8
4. Evaluate dx.
(x + 4)(x2 + 4)
∫
x3 + 5 x2 − x + 30
5. Evaluate dx.
(x2 + 4) (x2 + 9)
∫
2 x3 + 5 x − 1
6. Evaluate 2 dx.
(x2 + 1)
∫
5 x4 + 6 x3 + 30 x2 + 45
7. Evaluate 2 dx.
x2 (x2 + 3)
∫
6
8. Evaluate dx.
x2 + 6 x + 18

31
32
Numerical Integration (Corresponds
to Stewart’s Section 7.7)

∫5
1. Approximate 1
(6/x) dx by the Trapezoidal Rule, the Midpoint Rule, and Simpson’s Rule with N =
2, 4, 6.

2. In 2011, the distribution by quintile of pre-tax household income was: lowest quintile 5.13, second quintile
9.45, middle quintile 13.85, fourth quintile 20.33, highest quintile 51.24. Let L(x) be the percentage of
pre-tax household income earned by the lowest x% of households. Use the Trapezoidal Rule to alculate
the coefficient of inequality (or Gini coefficient),

∫ 100 (
1 )
γ= x − L(x) dx.
5000 0

3. Refer to the definition of the Gini coefficient γ that was given in the preceding problem. Approximate
γ using the Trapezoidal Rule and Simpson’s Rule with the following data:

x 10 20 30 40 50 60 70 80 90
L(x) 3 7 12 18 24 32 41 54 73

4. After the injection of 5 mg of dye into a heart, readings of the concentration c(t) of ejected dye (in
mg/L) were taken at 1.5 second intervals. The equation for cardiac output r is

∫ T
r c(t) dt = 5
0

where T is a time that is large enough for essentially all of the injected dye to be ejected, but not so long
that recirculation occurs. Use Simpson’s Rule to estimate cardiac output r, stated in the customary
units of L/min, based on the following tabulated readings:

t 0 1.5 3 4.5 6 7.5 9

c(t) 0 2.4 6.3 9.7 7.1 2.3 0

5. Depths of a swimming pool are measured in meters every 2 m.

33
34

Use Simpson’s Rule to estimate the volume of the pool.

6. Speed measurements (in meters per second) of a runner taken at half-second intervals during the first 5
seconds of a sprint are tabulated below.
Time 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Speed 0 5.26 6.67 7.41 8.33 8.33 9.52 9.52 10.64 10.64 10.87
About how many meters did the athlete run during that 5 second interval? Use Simpson’s Rule.

7. Use the Trapezoidal Rule to approximate the area between the curves shown in the figure below. Then
find the Simpson Rule aaproximation. (The exact area is, to seven correct decimal places, 22.1710099.)
Improper Integrals (Corresponds to
Stewart 7.8)
∫ 4
1. Evaluate (4 − x)−0.9 dx.
2
∫ 2
2. Evaluate (x − 2)−1/5 dx.
1
∫ 3
1
3. Evaluate √ dx.
−1 x+1
∫ 1
x
4. Evaluate 1/4
dx.
0 (1 − x2 )
∫ 1
1
5. Evaluate √ dx.
0 1 − x2
∫ 3
6. Evaluate x−1/2 (1 + x) dx.
0
∫ e
1
7. Evaluate dx.
1 x · ln 1/3
(x)
∫1
8. Evaluate 0
ln (x) dx.
∫ √
2
1
9. Evaluate √ dx.
1 x x2 − 1
∫4
10. Evaluate −2
(x + 1)−2/3 dx.
∫ e
1
11. Evaluate 2/7
dx.
1/e x ln (x)
∫ 2
1
12. Evaluate √ dx.
−2 4 − x2
∫1
13. Evaluate 0
ln2 (x) dx.
∫4
14. Evaluate x−1/2 ln(x) dx.
0
∫1 ( )
15. Evaluate −1 ln 1 − x2 dx.

35
36

∫∞
16. Evaluate 4
x−3/2 dx.
∫ ∞
1
17. Evaluate dx.
0 1 + x2
∫ ∞
x
18. Evaluate 2 dx.
0 (1 + x2 )
∫∞
19. Evaluate 2
e−x/2 dx.
∫ ∞
1
20. Evaluate dx.
2 x (x − 1)
∫ ∞
ex
21. Evaluate dx.
0 e2x+1
∫ ∞
ex
22. Evaluate 3 dx.
0 (ex + 1)
∫ ∞( )2
x+1
23. Evaluate dx.
0 x2 + 1
∫∞
x e−3x dx.
2
24. Evaluate 1
∫ ∞
1
25. Evaluate dx.
e x ln2 (x)
∫∞ x
26. Evaluate 0
(2/3) dx.
∫ −2 −3
27. Evaluate −∞
x dx.
∫ −2
1
28. Evaluate 4/3
dx.
−∞ (1 + x)
∫ 1
x
29. Evaluate 2 dx.
−∞ (1 + x2 )
∫0
30. Evaluate −∞
x exp (x/2) dx.
∫ ∞
1
31. Evaluate dx.
−∞ 4 + x2
∫ ∞
1
32. Evaluate dx.
−∞ x2 + 2x + 10
∫ ∞
x+2
33. Evaluate 2 dx.
−∞ (x2 + 1)
Arc Length (Corresponds to Stewart
8.1)

1
1. Calculate the arc length of the graph of y = x2 − ln(x), 1 ≤ x ≤ 2. Hint: Express the terms inside
8
the square root as one perfect square.
2. Watch one, do one, create one. Routine algebraic manipulation in the preceding exercise takes advan-
tage of what may seem from casual observation to be a fortunate accident. In fact, the terms of the
preceding exercise were meticulously determined in a math lab. Watch your spelling.
( )( )
d 3 d p
i) Let p be a constant. For what value of p is the product x x a constant?
dx dx
ii) For the remainder of this exercise, let p have the specific value determined in part (i). This value
is necessarily negative. Let a and b be nonzero constants and set f (x) = a x3 + b xp . Notice that,
( )2
d d
because p is negative, f (x) has the form u − v and that f (x) has the form u2 − 2uv + v 2 .
dx dx
( )2
d
Calculate f (x) , making sure to expand the square, an operation that yields three summands.
dx
What relationship must a and b satisfy in order that the cross term −2uv in the expanded square be
exactly -1/2?
iii) For the remainder of this exercise, let a = 1 and let b have the value determined by the relationship
of part (ii). Then the cross term −2uv of part (ii) is equal to -1/2. It follows that

( )2
d
1+ f (x) = 1 + (u − v)2
dx
( )
= 1 + u2 − 2uv + v 2
( )
1
= 1 + u2 − + v 2
2
( )
1
= u2 + + v 2
2
= (u + v)2 .
Use these observations to calculate the arc length of the graph of y = f (x), 1 ≤ x ≤ 2.
3. Approximate the arc length of y = x2 , 1/2 ≤ x ≤ 1 using Simpson’s Rule with n = 4.
4. The plot of the equation x2 + 2 x = y 3 + y + 6 is shown in the viewing window [0, 2] × [−1.5, 1].

37
38

Express the length of this curve as an unevaluated integral.

a
5. Let a be a positive constant. The graph of y = (exp(x/a) + exp(−x/a)) is called a catenary. Calculate
2
its length between the two points with ±a for abscissas.
Surface Area (Corresponds to Stewart
8.2)

The integral formula ∫ √

1 √ 1 ( √ )
1 + u2 du = u 1 + u2 + ln u + 1 + u2 + C
2 2
can be useful for evaluating a few surface areas.

1. The graph of y = f (x) = x3 , 0 ≤ x ≤ 1 is rotated about the x-axis. What is the area of the resulting
surface?

2. The graph of y = f (x) = x1/3 , 1 ≤ x ≤ 8 is rotated about the y-axis. What is the area of the resulting
surface?
3. The graph of y = exp(−x), 0 ≤ x ≤ ln(2) is rotated about the x-axis. What is the area of the resulting
surface?
√
4. The graph of y = 25 − x2 , −3 ≤ x ≤ 4 is rotated about the x-axis. What is the area of the resulting
surface?
5. The graph of y = x2 , 0 ≤ x ≤ 2 is rotated about the y-axis. By integrating with respect to y, find the
area of the resulting surface.
6. Redo the previous exercise, this time integrating with respect to x.
1
7. The graph of y = f (x) = x2 − ln(x), 1 ≤ x ≤ 2 is rotated about the x-axis. What is the area of the
8
resulting surface? (The first step is to show that 1 + f ′ (x)2 is the square of a rational function R(x).
On expanding f (x) R(x), you will obtain four summands, of which two are trivial to integrate and two
can be handled by techniques learned earlier.)

39
40
Center of Mass (Corresponds to
Stewart 8.3)

1. Find the center of mass of the region in the first quadrant that is bounded above by y = ln(x) and on
the right by x = e. (The region has uniform mass density.)
2. Find the center of mass of the region that is bounded above by y = 2 x and below by y = x2 . (The
region has uniform mass density.)
√
3. Find the center of mass of the region that is bounded above by y = 3 x, 1 ≤ x ≤ 4 and below by
y = x + 2, 1 ≤ x ≤ 4. Assume that the mass density of the region is constant.
4. The region shown in the figure is a rectangle surmounted by half a disk. The rectangle has height h,
width 2, and vertical central axis x = 1. The region has uniform mass density and its center of mass
lies on the horizontal diameter of the disk.

What is the value of h? (With some thought and simple geometry, you can do this exercise by calculating
only one very easy integral.)

41
42
Separable Differential Equations
(Corresponds to Stewart 9.3)

1. Solve the IVP

dy
(1 + x) = y, y(3) = 8.
dx
What is y(5)?

2. If
dy 2x + 1
= 2 , y(2) = 1,
dx 3y + 2
then for what positive value of x is y(x) equal to 0?

3. Solve the initial value problem y ′ (x) = (1 + 2 x)/y 2 , y(1) = 2.

√
4. Solve the initial value problem y ′ (x) = 6 x y, y(1) = 9.

dy 1
5. Solve the initial value problem = y + , y (0) = 2
dx y
6. A mixing tank initially contains 100 kg of salt dissolved in 1000 L water. At time t = 0 an inlet valve
and an outlet valve are opened. Salt-free water flows in at the rate of 20 L per minute and is instantly
mixed. The salt solution flows out at the rate of 20 L per minute. At what time is the mass of salt in
the tank equal to 10 kg?

7. A mixing tank initially contains 5 kg of salt dissolved in 200 L water. At time t = 0 an inlet valve and
an outlet valve are opened. A salt water solution (brine) of constant concentration 2 kg salt per 10 L
water flows in at the rate of 10 L per minute and is instantly mixed. The salt solution flows out of the
mixing tank at the rate of 10 L per minute. What is the concentration of salt in the mixing tank one
hour after the opening of the valves?

8. A mixing tank contains a brine solution containing 100 kg of salt dissolved in 2000 liters of water.
Another brine solution, this one with a concentration of 1 kg salt per 5 liters, flows in at the rate of
10 liters per minute. The well-mixed solution flows out of the tank at the same rate. How long in
minutes does it take for the salt concentration in the mixing tank to double? If the inflow and outflow
are maintained indefinitely, about how much salt will be dissolved in the mixing tank’s solution?

9. The mass m(t) of a radioactive isotope decays according to the Law of Radioactive Decay: D(m)(t) =
−λ m(t) where λ is a positive constant, the decay constant, that depends on the particular substance.
If m0 is the mass of the radioactive substance at time t = 0, then what is the mass m(t) at an arbitrary
time t?

43
44

10. (Continuation of Exercise 9.) Let τ be the positive constant that is related to the decay constant by the
equatuion λ τ = ln(2). Show that m(t + τ ) = m(t)/2 for every value of t. The constant τ is called the
half-life of the isotope. If you measure the mass of the substance and then return τ units of time later
and measure again, your second measurement will be half the first.

11. (Continuation of Exercise 10.) Use the constant τ to express the Law of Radioactive Decay as exponential
decay with base 2.

12. Eight grams of a certain radioactive isotope decay exponentially to six grams in 100 years. After how
many more years will only four grams remain? (Use either Exercise 10 or Exercise 11.)

13. (Radiocarbon Dating) Two isotopes of carbon, 12 C, which is stable, and 14 C, which decays exponentially
with a 5700 year half-life, are found in a known fixed ratio in living matter. After death, carbon is no
longer metabolized and the amount m(t) of 14 C decreases due to radioactive decay. In the analysis of a
sample performed T years after death, the mass of 12 C, unchanged since death, can be used to determine
the mass m0 of 14 C that the sample had at the moment of death. The time T since death can then
be calculated from the law of exponential decay and the measurement of m(T ). In 1991, radiocarbon
dating established that fossils of dwarfed woolly mammoths, then recently found on Wrangel Island,
which is located in the Arctic Ocean off the coast of Siberia, were 3700 years old. What percentage of
m0 was the measured amount of 14 C in these fossils? (Prior to the discovery of these fossils, mammoths
were thought to have become extinct 10,000 years ago.)

14. In 1994, a parka-clad mummified body of a girl was found in a subterranean meat cellar near Barrow,
Alaska. Radiocarbon analysis showed that the girl died around ce 1200. What percentage of m0 was
the amount of 14 C in the mummy?
15. In recent years, prehistoric cave art has been found in a number of new locations. At Chauvet, France,
the amount of 14 C in some charcoal samples was determined to be 0.0879 · m0 . About how old are the
cave drawings?

16. Let y (t) and v (t) = dy/dt be the height and velocity of a projectile that has been shot straight up from
the surface of Earth with initial velocity v0 . Newton’s Law of Gravitation states that

dv gR2
=− 2,
dt (R + y)

where R is the radius of the Earth and g is the acceleration due to gravity at Earth’s surface. Using the
Chain Rule to express dv/dt in terms of dv/dy, we have

dv dv dy dv
= · = · v.
dt dy dt dy

By equating this expression for dv/dt with the one given by Newton’s Law of Gravitation, we obtain

dv g R2
v· =− 2.
dy (R + y)

Observe that we are now regarding v as a function of y rather than t. The initial value condition is that
v = v0 for y = 0. Find √ the unique solution v(y) to this initial value problem. (You can leave it as a
formula for v 2 .) If v0 < 2gR, then what is the maximum height attained by the projectile?
Models of Population Growth
(Corresponds to Stewart 9.4)

1. Suppose that a population P (t) grows according to the differential equation D(P )(t) = k P (t) for
some positive constant k. Let T be defined by the equation k T = ln(2). Show that, the equation
P (t + T ) = 2 P (t) holds for every value of t. The significance of this equation is that if we take a census
and then take a second census T units of time later, the second count is double the first. The constant
T is called the doubling time.

2. (Continuation of Exercise 1) Show that if P (t) is a population satisfying the differential equation of
Exercise 1, and if T is the doubling time for the population, then P (t) = P (0) 2t/T .

3. In a yeast nutrient broth, the population P (t) of a colony of the bacterium Escherichia Coli (E. Coli)
grows according to the differential equation D(P )(t) = k P (t) for some positive constant k. At 10:00 AM
there were 6000 bacterial colonists and at 10:20 AM there were 12 000. How many will there be at
12:30 PM?

4. Day-care centers are germ centers. Little kids coughing through uncovered mouths, little kids sneezing
on other little kids, little kids drooling on playmates, little kids touching everything with their unwashed
hands. In any given day-care center, at any given time, an infection is spreading among the children
doing time there. Let P (t) be the fraction (between 0 and 1) of infected children. It is reasonable
that the rate D(P )(t) at which the infection is spreading is proportional to P (t): if one infected child
can infect k children, then two infected children can infect 2k children. So it may seem reasonable to
model the outbreak by the exponential growth differential equation D(P )(t) = k P (t). Not so fast!
As P (t) grows and becomes closer to 1, there are fewer uninfected children remaining who are capable
of becoming infected. The decreasing pool of potential infectees will moderate ( the
) rate at which the
infection spreads. We therefore emend our model to D(P )(t) = k P (t) 1 − P (t) : the rate at which
the infection spreads is jointly proportional to the fraction P (t) of children capable of passing on the
infection and the fraction 1 − P (t) of children capable of becoming infected. Suppose that P0 infected
children arrive at the day care center Monday morning. What is P (t) in terms of t, k, and P0 ?
5. (Continuation) Suppose that 100 children attend a day-care center. Suppose that at 9 AM Monday
morning, only one child is infected. Suppose that at 9 AM Tuesday morning three children are infected.
By what time will the infection have spread to half the day-care population?

6. Suppose that a and b are positive constants. If a population P (t) satisfies the differential equation
D(P )(t) = a P − b P 2 , then what is the carrying capacity of the population?

45
46
Linear Differential Equations
(Corresponds to Stewart 9.5)

1. Solve the initial value problem

dy
+ 2y = 1, y(0) = 2
dx
2. Solve the initial value problem
dy y
=1+ , y(0) = 3
dx 2
3. Solve the initial value problem
dy 2xy
= 15x2 − , y(1) = 6
dx 1 + x2

4. Solve the initial value problem

dy y
=1− , y(4) = 5
dx x
5. Solve the initial value problem
dy 2 √
+ y − 7 x = 0, y(1) = 9
dx x
6. Solve the initial value problem
dy
+ xy = x, y(0) = 4
dx
7. Solve the initial value problem
dy
− 2 x y = 4 x, y(0) = 1
dx
8. Solve the initial value problem
dy
= 2 x (2 + y), y(0) = 1
dx
9. Solve the initial value problem
dy
− 3y = e3x , y(2) = 0
dx
10. Solve the initial value problem

dy
= csc (x) + y cot (x) , y(π/2) = 4
dx

47
48

11. Solve the initial value problem

dy y
+ = 1, y(e) = 2 e
dx x ln (x)

12. A large capacity mixing tank is partially filled with 120 gallons of water in which 80 lbs of a solute is
dissolved. Two valves are simultaneously opened. Water containing the same solute as in the mixing
tank flows into the tank at the rate of 24 gallons per minute. The concentration of the solute in the
water flowing in is 1 lb per gallon. The second open valve allows the thoroughly mixed solution to flow
out of the tank at the rate of 16 gallons per minute. Find y(t), the weight of the solute in the mixing
tank t minutes after the valves have been opened (and during the time period during which the tank is
not filled to capacity). How many pounds of the solute are in the mixing tank one-quarter of an hour
after the opening of the valves?
Sequences (Corresponds to Stewart
11.1)

3n
1. What are the first five terms of the sequence an = , 0 ≤ n < ∞?
2n + 1
∞
2. Let f (x) = x3 − 4. Consider the recursive sequence {xn }n=1 defined by

f (xn )
x1 = 1, xn+1 = xn − (1 ≤ n < ∞)
f ′ (xn )
Calculate the first four terms of this sequence exactly as rational numbers. Calculate the fifth term
to five decimal places. (This sequence is the Newton-Raphson approximation of the root of f (x) = 0.
Compare x5 with 41/3 .)
3 + 5 n2
3. Find the limit of the sequence an = .
5 + 3 n2
√
n5 + 9 n6
4. Find the limit of the sequence an = .
1 + 4 n6
5. Find the limit of the sequence an = 3n π −n .
√ √
6. Find the limit of the sequence an = exp(1/ n) + exp(−1/ n).
(2 n − 1)! n2
7. Find the limit of the sequence an = .
(2 n + 1)!
sin(n)
8. Find the limit of the sequence an = . Two intuitive results may help. First, if |an | → 0, then
n
an → 0. Also, if an ≤ bn ≤ cn for all n and if limn→∞ an = limn→∞ cn = L, then limn→∞ an = L. This
is sometimes called The Pinching Theorem, sometimes The Squeeze Theorem.
9. Find the limit of the sequence an = n sin(1/n).
10. Find the limit of the sequence an = e2 n/(n+3) .
n2
11. Find the limit of the sequence an = √ .
n5 + 1
ln(n)
12. Find the limit of the sequence an = √ .
n
ln(n)
13. Find the limit of the sequence an = .
ln(3 n)

49
50

ln(n)
14. Find the limit of the sequence an = .
ln (n3 )
15. Find the limit of the sequence an = ln(n + 10) − ln(n + 1).
16. Find the limit of the sequence an = n3 · (1.1)−n .
2n + n5
17. Calculate lim .
n→∞ 2n + n2

3 × 7n + 6 × 11n
18. Calculate lim .
n→∞ 5n + 2 × 11n
( )n
1
19. Find the limit of the sequence an = 1− .
n
(√ )1/n
20. Find the limit of the sequence an = 1 + n2 .

21. It can be shown that the recursive sequence

√ √
a1 = 2, an+1 = 2 an (1 ≤ n < ∞)

has a limit L. What is L?

22. It can be shown that the recursive sequence
1
a1 = 1, an+1 = 1 + (1 ≤ n < ∞)
an + 1
converges. To what number L does the sequence converge?
Series (Corresponds to Stewart 11.2)

∞
∑ ( )
n
1. Consider the infinite series ln .
n=2
n+2
(i) Calculate the first five terms s1 , s2 , s3 , s4 , s5 of the sequence of partial sums. (Use a property of the
logarithm to simplify these terms.)
∑M ( )
n
(ii) For what integer M is ln equal to the N th partial sum sN of the given series?
n=2
n + 2
(iii) Find a formula for sN that holds for N > 1. (No formal proof needed. Your formula should fit the
∑∞ ( )
n
pattern established in part (i).) (iv) Does ln converge or diverge?
n=2
n+2
∞
∑ { }∞
2N + 1
2. For what series an is the sequence of partial sums equal to ? (It might help to first
n=1
N + 2 N =1
figure out the terms a1 , a2 , a3 , . . . if you do not immediately see how to find a formula for the general
term an .) Show that the series you have found is convergent and determine its value.
3. Evaluate 5 − 2.5 + 1.25 − 0.625 + 0.3125 − 0.15625 + · · · .
∑∞
(−2)n−1
4. Evaluate .
n=2
3n

∑∞
3n+1
5. Evaluate .
n=1
4n−1

∑∞
22n+1
6. Evaluate .
n=1
5n
∞
∑ 2
7. Use partial fractions to find the sequence of partial sums of the series 2−1
. Determine the value
n=2
n
of the series.
∑∞
8. The “Test for Divergence” states that if limn→∞ an ̸= 0, then n=1 an diverges. (We will encounter
other tests that can lead to the conclusion “divergent,” but the other tests have other names.) To which
of the
∑∞following series can the Test for Divergence be applied.
a) n=1 √1n
∑∞
b) ∑ n=2 ln(n)
n
∞ 1
c) n=1 n1/n
∑∞ 1
d) n=1 arctan(n)
∑∞ ( 1+2/n )n
e) n=1 7

51
52

9. In each sentence, use either the word “may” or the word “must” to fill in the blank so that the completed
sentence is correct.
(i) A series with summands tending to zero converge.
(ii) A series that converges have summands that tend to zero.
(iii) If a series diverges, then the Test for Divergence succeed in proving the divergence.
(iv) If the partial sums of an infinite series are bounded, then the series converge.
(v) If a series diverges, then its terms diverge.
(vi) If the partial sums of an infinite series diverge, then the series diverge.
The Integral Test (Corresponds to
Stewart 11.3)

1. This exercise begins with a bit of (easy) theory. Let K be any fixed integer, which will assume the role
∞
∑
of an initial value for an index of summation of a series SK = an . Commonly used values of K are
n=K
0 and 1, but K is not limited to these two choices. Let M be an integer greater than K. It too will be
∞
∑
used as the initial value of an index of summation of a series: to be specific, the series SM = an
n=M
obtained by deleting the terms aK , aK+1 , aK+2 , . . . , aM −1 from the series

SK = aK + aK+1 + aK+2 + · · · + aM −1 + aM + aM +1 + · · · .

Then we may write SK as the sum of the finite series

∑
M −2
F = an = aK + aK+1 + aK+2 + · · · + aM −1
n=K

and the infinite series SM : Thus, SK = F + SM . In this context, SM is said to be a tail of the series
SK : a tail of a series is the infinite series that results by dropping a finite number, any finite number
no matter how large, of terms from the beginning of an infinite series. A finite series of real numbers
always converges—you just add up all the numbers and the process, being finite, comes to a happy end.
It follows that, if a tail SM of a series SK converges, then so too does the full series SK = F + SM . Here
is the point of this observation: there are series for which a convergence test might be conclusive by
applying it to one of its tails even if the test cannot be applied to the full series. This can happen if the
hypotheses required for the application of a testhold for the tail but not for the entire series. Consider,
for example, the series
∑∞
n2
S1 = 3
.
n=1
n + 32
The Integral test cannot be applied to S1 . To understand why, salculate the first three terms in the
sum. What hypothesis of the Integral Test is violated? Nevertheless, the Integral Test can be applied
to tails
∑∞
n2
SM = 3
n=1
n + 32
of S1 beginning with what value of M ?
∞
∑ n2
2. Does converge, or does it diverge? Use the Integral Test.
n=1
n3 + 32

53
54

∞
∑ n2
3. Does 2 converge, or does it diverge? Use the Integral Test.
n=1 (n3 + 32)
∞
∑
4. Suppose that r is a positive constant that is less than 1. Use the Integral Test to show that n rn
n=1
converges.
∑∞
5. The value of S = n=1 1/n2 is known to be exactly π 2 /6, or 1.644934068 to nine decimal∫ places. The
∞
partial sum s10 is 1.549767731 to nine decimal places. Calculate the improper integrals 10 (1/x2 ) dx
∫∞
and 11 (1/x2 ) dx. Use these two calculations together with s10 to find the lower estimate ℓ and the
upper estimate m of S that arises from the Integral Test. Estimate S by using the midpoint of the
interval [ℓ, m] containing S.
∑∞ ∑5
6. The exact value of S = n=1 (1/n3 ) is not known. It is not that much work to calculate n=1 (1/n3 ) =
256103/216000 = 1.185662037 to nine decimal places. Use this sum and the estimates of the Integral
Test to find an interval containing the exact value of S. If the midpoint of this interval is used to
estimate S, how big can the error be?
Series with Positive Terms
(Corresponds to Stewart 11.4 and
11.3)

In each of Exercises 1–12, use the Comparison Test to show that the given series converges. State the series
that you use for comparison and the reason for its convergence.
∞
∑ n
1. 3+1
n=1
n

∑∞
5n − 4
2.
n=1
n9/4
∞
∑ 1
3. ( √ )2
n=1 n+ 2

∑∞
2 + sin (n)
4.
n=1
n4

∑∞
2n − 1
5.
n=1
nen
∞
∑ n+2
6. 5/2 + 3
n=1
2n

∑∞
n2 + 2n + 10
7.
n=1
2n4

∑∞
2n
8.
n=1
n3n
∞
∑ n2
9. (1/3)
n=1

∞
∑ 1
10. √
n=1
2n n−1

55
56

∞
∑ n3
11.
n=1
n4.01 +1

∞
∑ 2
12.
n=1
2n + 1/2n

In each of Exercises 13–20, use the Comparison Test to show that the given series diverges. State the
series that you use for comparison and the reason for its divergence.
∞
∑ 1
13. √
n=1
n + n

∞
∑ 1
14.
n=2
ln (n)

∑∞
1 + ln (n)
15.
n=1
3n − 2

∑∞
2n + 5
16.
n=1
n2 + 1

∞
∑ 1
17. √
n=1
10 + n2

∑∞
3n + 1
18.
n=1
3n + 2n

∑∞
3n + n
19. √ n
n=1
n3 + 1

∑∞
2n + 1
20.
n=1
2n + n2
∑
In Exercises 21–36 use the Limit Comparison Test to determine whether the given series an converges
or diverges.
∞
∑ 3n
21.
n=1
2n2 − 1

∞ √
∑ n+1
22.
n=1
n2

∞
∑ 2
(n − 1)
23. 4
n=1 (n + 1)
∞
∑ n2 + 2
24. 2
n=1 (n2 + 1)
57

∑∞ √
n+1
25.
n=1
2n2 − n

∞
∑ 3
26.
n=1
2n +3

∑∞
2n + 11
27.
n=1
3n − 1

∑∞
arctan(n)
28.
n=1
n2

∞ √
∑ 1 + n2
29. 1/4
n=1 (1 + n8 )
∞
∑ n1/3
30. √
n=1 1 + n5/2
∞
∑ ( )
31. sin 1/n2
n=1
sin(x)
(Hint: Use your knowledge of lim .)
x→0 x
∑∞
n + ln (n)
32.
n=1
n3

∞
∑ 1
33.
n=1
n + ln (n)

∑∞
2 n + n3
34.
n=1
3 n + n2

∑∞
7n + n10
35.
n=1
10n

∞ (
∑ )n
2n + 5
36.
n=1
3n + 7

In Exercises 37–48 use the Integral Test to determine whether the given series converges or diverges. (You
may notice that, in several instances, a comparison test would provide a more efficient method of determining
the behavior of the series.)
∞
∑ n
37. 2+1
n=1
n

∞
∑ 1
38. 2+4
n=1
n
58

∞
∑ 1
39.
n=1
n(n + 2)
∞
∑ ( )
40. n2 exp −n3
n=1

∑∞
2n2
41. 3
n=2
n +4
∞
∑ 1
42.
n=2
n ln (n)
∞
∑ 1
43. 2
n=2
n ln (n)
∞
∑ exp(n)
44.
n=1
(1 + exp(n))2

∑∞
n
45. n
n=1
10

∑∞
ln (n)
46.
n=1
n
∞
∑
47. n exp (−n/2)
n=2

∞
∑ 1
48. √
n=2
n n2 − 1
Alternating Series (Corresponds to
Stewart Section 11.5)

∑∞ n+1
In each of Exercises 1–6 a convergent alternating series S = n=1 (−1) an is given. It is easy to see
that, because the terms an decrease to 0, the Alternating Series Test can be∑ applied to each of these series,
N n+1
establishing convergence. Find a value of N such that the partial sum SN = n=1 (−1) an approximates
the given infinite series to within 0.01. That is, find N so that |S − SN | ≤ 0.01 holds. The point of these
exercises is to obtain the estimate of S efficiently. If a particular value of N works, then any larger value of N
also works, but the calculation of the larger partial sum involves more computation. That extra computation
is not needed for the desired approximation. Therefore, find the smallest value of N that the Alternating
Series Test guarantees will do the job.
∞
∑ 1
1. (−1)n+1 √
n=1
n

∞
∑ 1
2. (−1)n+1
n=1
n3

∞
∑ 4
3. (−1)n+1
n=1
2n − 1

∞
∑ 1
4. (−1)n+1
n=1
n!

∞
∑ 1
5. (−1)n+1
n=1
n2 + 15n

∞
∑ 1
6. (−1)n+1
n=1
n ln (n + 1)
∑∞ n
In each of Exercises 7–10, an alternating series n=1 (−1) an is given. For each series, limn→∞ an = 0.
n=∞
However, {an }n=1 is not a decreasing sequence. Nevertheless, in each case, there is an integer M such
n=∞
∑∞ {an }n=M
that
n
is a decreasing sequence. Therefore, the Alternating Series Test may be applied to the tail
n=M (−1) a n , establishing that the given series converges. Find the smallest value of M ,

∞
∑ (−1)n+1
7.
n=1
n2 − 20n + 101

59
60

∞
∑ (−1)n+1
8.
n=1
n4 − 32n2 + 400
∞
∑ n5
9. (−1)n+1
n=1
2n
∞ ( )
∑ n3
n+1 ln
10. (−1) √
n=1
n

In each of Exercises 11 and 12 calculate the given alternating series to three decimal places. (That means
that the truncation error should not exceed 5 × 10−4 , or 0.0005.)

∑∞
(−1)n+1
11.
n=1
n!
∞
∑ ( )
n+1 2 − 1/n
12. (−1)
n=1
10n
The Ratio Test and the Root Test
(Corresponds to Stewart Section 11.6)

In each of Exercises 1–8 use the Ratio Test to determine the convergence or divergence of the given series.

∑∞
n
1. n
n=1
e

∑∞
n2
2.
n=1
2n

∑∞
2n
3.
n=1
n3

∑∞
10n
4.
n=1
n!

∑∞
n100
5.
n=1
n!

∞
∑ n!
6.
n=1
n5 · 7n

∑∞
3n + n
7.
n=1
2 n + n3

∑∞
2n
8. √
n=1 n!

In each of Exercises 9–18, verify that the Ratio Test yields no information about the convergence of the
given series. Use other methods to determine whether the series converges absolutely, converges conditionally,
or diverges.
∞
∑ n ln (n)
9. (−1)
n=2
n

61
62

∞
∑ n 1
10. (−1)
n=2
ln (n2 )
∞
∑ n ln (n)
11. (−1)
n=1
n2
∞
∑ n n+1
12. (−1) ·
n=1
n3 + 1
∞
∑ ( )
1
13. (−1) · 1 +
n

n=1
n
∞
∑ √
n n
14. (−1)
n=1
n+4
∞
∑ n en + ln (n)
15. (−1)
n=1
e n · n2
∞
∑ n n!
16. (−1) ·
n=1
n! + 2n
∞
∑ n
17. (−1) arctan(n)
n=1

∞
∑ ( )
1n
18. (−1) sin
n=1
n

In each of Exercises 19–25, use the Root Test to determine the convergence or divergence of the given
series.
∑∞ −n/2
19. n=1 n

∑∞
23n
20.
n=1
32n

∑∞
n
21.
n=1
2n

∑∞
10n
22.
n=1
n10
∞
∑ n7
23.
n=2
lnn (n)

∑∞ ( )n
ln (n + 4)
24.
n=1
ln (n2 + 4)
63

∞ (
∑ )n
37
25.
n=1
n

In each of Exercises 26–40 determine whether the series converges absolutely, converges conditionally, or
diverges. You may use any test.
∞
∑ (2n)!
26. (−1)n
n=1
(3n)!
∞
∑ (n!)2
27. (−1)n
n=1
(2n)!

n + 3n
28. sum∞
n=1 (−1)
n
n3 + 2 n
∞
∑ ( )n
29. (−1)n n1/n + 1/2
n=1
∞
∑ 2n
30.
n=1
1 + lnn (n)

∑∞
exp (n)
31.
n=1
lnn (n)
∞
∑ n 1
32. (−1) √
n=1
n + 10
∞
∑ n n!
33. (−1)
n=1
3n
∞
∑ n n
34. (−1)
n=1
n2 − 11
∞
∑ n 1
35. (−1)
n=2
n ln3 (n)
∞
∑ n n2 + 2n + 2
36. (−1)
n=1
3n3 + 7
∞
∑ n e1/n
37. (−1)
n=1
ne
∞
∑ n
38. (−1) ln (1 + 1/n)
n=1
∞
∑ n 1
39. (−1) n
n=1
(1 + 1/n)
64

∞
∑ n ln (n)
40. (−1) √
n=1
n
Intervals of Convergence (Corresponds
to Stewart 11.8)

∞
∑ n
In Exercises 1–16, determine the interval of convergence of the given power series an (x − c) . You can
n=0
complete the calculations by carrying out the following four step algorithm (which reduces to two steps if the
interval of convergence is either a single point or the entire real axis).

Step 1 Identify the sequence of coefficients {an } and the base point c. This may require rewriting
the power series so that it is in exactly the generic form written above. This means that the coefficient of the
variable in parentheses is 1, the operation between the variable and the base point is subtraction, and the
power to which the difference of the variable and the base pointbase point is raised is the index
( of summation.
)n
For example, (3x − 2)n should be rewritten as 3n (x − 2/3)n , (x + 2)n should be rewritten as x − (−2) , and
(x − 5)2n should be rewritten as un with u = (x − 5)2 replacing x as the variable of the power series.

an
Step 2 Calculate the radius of convergence R using either of the following two formulas, R = lim
n→∞ an+1
1
or R = lim , whichever seems easiest to use. Notice that the quantities in the limits are the reciprocals
n→∞ |a |1/n
n
of the fractions used in the Ratio and Root Tests. At this stage, if R = 0 or R = ∞, the calculation is finished.
If R = 0, then the interval is [c, c], consisting of only one point: the base point. If R = ∞, then the interval
is (−∞, ∞), which is to say the entire real line.

Step 3 Check to see if the left endpoint of the interval of convergence belongs to the interval. Do so
by substituting x = c − R into the power series and use any suitable test for convergence. (However, nei-
ther the Ratio Test nor he Root Test will yield a conclusion for an endpoint of an interval of convergence, so
there is no point trying them.) For an endpoint, the type of convergence, conditional or absolute, is immaterial.

Step 4 Check to see if the right endpoint of the interval of convergence belongs to the interval. Do so
by substituting x = c + R into the power series and use any suitable test for convergence. As in Step 3, do
not bother with either the Ratio Test or the Root Test here. Again, the type of convergence, conditional or
absolute, is immaterial.
∞ (
∑ )n
−x
1.
n=0
3
∑∞
2. n=1 n xn

65
66

∑∞
10 n
3. x
n=0
n+1

∞
∑ n+1 (10x)n
4. (−1) √
n=1
n

∑∞
(4x)n
5.
n=1
n2

∑∞
n
6. n
(2x + 1)n
n=0
4

∞ (
∑ )n
−x + 1
7.
n=1
n

∞
∑ 3
8. (−1)n √ xn
n=0
n2 + 1
∞
∑ ( )n
n2 1
9. · x −
n=0
n3 + 1 2

∞
∑ 3n
10. 3+1
(x + 2)n
n=0
n

∑∞ ( )n
1 + 2n 1
11. · x+
n=0
1 + 3n 2

∑∞
(x + 6)n
12. √
n=1
n

∑∞
(π − x)n
13.
n=1
n

∑∞
(x + 4)n
14.
n=0
n5 + 1
∑∞ n
15. n=0 (3x + 2)
∞
∑ n+1
16. (−1)n (2x + 5)n
n=0
3n + 1

In Exercises 17 and 18, the powers that appear in the given power series are not the indices of summations
of the series. To use the algorithm as it is given in the instructions to Exercises 1–16, you must make a
∑∞ ∞
∑ ∑∞ ∑∞
2n 2n+1
change of variable. Change an (x − c) to an un and an (x − c) to (x − c) an un where
n=0 n=0 n=0 n=0
u = (x − c)2 . Treat u as the variable of the power series and use the algorithm with u replacing x. Find an
interval of convergence for u and unwind what it means for x.
67

∑∞
4n 2n
17. (x + 3)
n=0
n + 1

∑∞
4 n + n2 2n+1
18. n
(x − 1)
n=0
(n + 1)9
68
Series Related to 1/(1 ± x)
(Corresponds to Stewart 11.8)

In each of Exercises 1–6 express the given function as a power series in x with base point (center of convergence)
0. Instead of a brute force computation, use the formula
∞
∑ 1
un = (−1 < u < 1)
n=0
1−u

and report terms through at least order 4.

1
1.
2+x
1
2.
1 − 2x
x
3.
4 + 3x
1
4.
4 + x2
1 + x2
5.
1 − x2
x
6.
2 + 3 x2

Exercises 7–12 all use the power series for 1/(2 + x) found in Exercise 1. The techniques for working with
series include multiplication, division, differentiation, and integration.

7. Express 1/(2 + x)2 as a power series with base point 0. For the method of calculation, multiply the
series for 1/(2 + x), which you found in Exercise 1, by itself. Carry out the multiplication term-by-term
and collect all terms through x3 .

8. Express 1/(2 + x)2 as a power series with base point 0, but use a method of(calculation
) that differs from
1 d 1
that of the preceding problem. This time, observe that =− . Use the series for
(2 + x)2 dx 2 + x
1/(2 + x), which you found in Exercise 1, and differentiate term-by-term all terms through x4 .

69
70

9. Factor 2 − x − x2 . One of the factors is 2 + x. Then use the( method )of Exercise 7 together with the
calculation of Exercise 1 to calculate the power series of 1/ 2 − x − x2 with base point 0. Report all
terms through x3 .
10. Express ln(2 + x) as a power series with base point 0. Report terms through x5 .
∞
∑
11. The power series n xn converges for −1 < x < 1. Let f (x) be the sum of this series for x in the
n=1
interval of convergence. It can be shown that f (x) = x/(x − 1)2 but this exercise does not rely on this
f (x)
evaluation of f (x). Calculate the power series of q(x) = . Do the calculation in two different ways.
2+x
First, multiply the given series for f (x) term-by-term with the series for 1/(2 + x). Report all terms
through x4 . For the second calculation, write q(x) = a0 + a1 x + a2 x2 + a3 x3 + a4 x4 + · · · . Then
∞
∑ ( )
n xn = f (x) = (2 + x) a0 + a1 x + a2 x2 + a3 x3 + a4 x4 + · · · .
n=1

Multiply the two factors on the right term-by-term. Collect all terms through x4 . Equate the constant
terms of the series on each side to find a0 . It will be 0. Then equate the coefficients of x on each side
to find a1 . Continue with x2 , x3 , and x4 .
∞
∑
12. The power series 2 + n xn converges for −1 < x < 1. Let f (x) be the sum of this series for x in
n=1 ( )
the interval of convergence. It can be shown that f (x) = 2x2 − 3x + 2 /(x − 1)2 but this exercise
2+x
does not rely on this evaluation of f (x). Calculate the power series of q(x) = . To do so, write
f (x)
q(x) = a0 + a1 x + a2 x + a3 x + · · · . Then
2 3

( ∞
)
∑ ( )
2+x= 2+ nx n
a0 + a1 x + a2 x2 + a3 x3 + a4 x4 + · · · .
n=1

Multiply the two factors on the right term-by-term. Collect all terms through x4 . Equate like powers
of x on each side of the equation through x3 . Successively solve for a0 , a1 , a2 and a3 .

In Exercises 13 and 14, use the power series

1 1 1
ln(1 + u) = u − u2 + u3 − u4 + · · · (−1 < u < 1)
2 3 4
to express the given function as a power series in x with base point 0. Report terms through at least order 4.

13. ln(2 − x)
14. x ln(3 + 4x2 )
Taylor Polynomials (Corresponds to
Stewart 11.9)
√
1. Calculate the order 3 Taylor polynomial of f (x) = x with base point 4.

2. Calculate the order 4 Taylor polynomial of f (x) = 1/x with base point 1/2.
3. Calculate the order 4 Taylor polynomial of f (x) = sin(x) with base point π.

4. Calculate the order 4 Taylor polynomial of f (x) = exp(1 + x/2) with base point −2.
5. Calculate the order 3 Taylor polynomial of f (x) = (4 + x)−3/2 with base point −3.
√
3+x
6. Calculate the order 2 Taylor polynomial of f (x) = with base point 1.
x
7. Calculate the orders 1, 2, and 3 Taylor polynomials of f (x) = cos(x) with base point π/3.

71
72
Taylor and Maclaurin Series

In each of Exercises 1–4 use the Maclaurin series

1 3 1
sin(u) = u − u + u5 − · · · (∞ < u < ∞)
3! 5!
for sin(u),
1 2 1
cos(u) = 1 − u + u4 − · · · (∞ < u < ∞)
2! 4!
for cos(u), and/or

1 2 1 1
exp(u) = 1 + u + u + u3 + u4 + · · · (∞ < u < ∞)
2! 3! 4!
for exp(u) to obtain Maclaurin series for the given function. Report terms through at least order 4.
( )
sin x2
1.
x
1 − cos (2x)
2.
x
x
3.
exp (x2 )
( )
4. exp x2 + 2x

In Exercises 5 and 6, use Newton’s Binomial Series,

α(α − 1) 2 α(α − 1)(α − 2) 3 α(α − 1)(α − 2)(α − 3) 4

(1 + u)α = 1 + αu + u + u + u + ···
2! 3! 4!
to find the first few terms of the Maclaurin series of the given functions.
1
5. √
1 + 2x2
6. (8 + x3 )1/3

In Exercises 7 and 8, use a Taylor polynomial to calculate the given integral with an error less than 10−5 .

73
74

∫ 0.2 ( )
7. sin x2 dx
0
∫ 0.2
1
8. dx
0 1 + x3
Polar Coordinates (Corresponds to
Stewart 10.3)

1. The polar coordinates of three points are given. State the Cartesian coordinates of each point. Then
find two other pairs of polar coordinates of each point: one with r > 0 and one with r < 0.
P = (3, π/4), Q = (−4, 3π/2), R = (10, −π/3)
2. The Cartesian coordinates of three points are given. For each point, find two pairs of polar coordinates
(r, θ): one pair with r > 0 and 0 ≤ θ < 2π and one pair with r < 0 and 0 ≤ θ < 2π:
√ √
P = (−3, 3), Q = (2, 2 3), R = (− 3, 1)
3. Find the distance between the points with polar coordinates (2, −π/6) and (−6, π/3).

4. Sketch the region in the plane that consists of points with polar coordinates that satisfy the inequalities
1 ≤ r ≤ 2 and π/4 ≤ θ ≤ 2π/3.

5. Identify the curves with the given polar equations:

(i) r = 5 sec(θ)
(ii) r2 sin(2θ) = 6.

6. Sketch the curve with polar equation r = −3 cos(θ).

7. Sketch the curve with polar equation r = (1.2)θ , 0 ≤ θ ≤ 7 π/2.

8. Sketch the curve with polar equation r = 3 + 2 sin(θ).

9. Sketch the curve with polar equation r = 2 cos(3θ).

10. Sketch the curve with polar equation r = cos(θ/2).

11. Sketch the curve with polar equation r = 1 + 2 sin(θ).

12. Sketch the curve with polar equation r = cos(2θ + π/4).

75
76
Areas in Polar Coordinates
(Corresponds to Stewart 10.4)

1. Calculate the area of the region inside the cardioid r = 1 + sin(θ).

2. Calculate the area of the region between the x-axis and the polar curve r = θ, 0 ≤ θ ≤ π.

√
3. Find the area inside the polar curve r = sin(2 θ).

77
78

4. Find the area inside the polar curve r = sin(2 θ).

5. The figure below shows a cross-section of a lump of green kryptonite, a once-rare mineral first discovered
on Earth by Jerry Siegel in the 1940s, but which has become increasingly common thanks to meteor
showers
√ and the
( proliferation
) of superhero films. The boundary curve is given in polar coordinates by
r = sin(θ)/ 5 − 4 cos(θ) . What is the area of the enclosed region?
79

6. Find the area of the region inside one leaf of the three-leaf rose r = cos(3 θ).

√
√ the area of the region inside the cardioid r = 1 − sin(θ) and above the line y = −x/ 3, 0 ≤ x ≤
7. Find
3 3/4.
80

8. Find the area of the region inside the circle r = 2 sin(θ) and outside the circle r = 1.

9. Find the area of the region inside the cardioid r = 1 − sin(θ) and outside the circle r = 1/2.
81

10. Find the area of the region inside the circle r = 3 cos(θ) and outside the cardioid r = 1 + cos(θ).

√
11. The figure below shows the circle with polar equation r = 2 sin(θ) and the curve with polar equation
r2 = sin(2 θ), 0 ≤ θ ≤ π/2. Find the area of the region inside both curves.
82

12. The figure below√ shows √ the circle with polar equation r = 2 and the secant line segment between the
two points (1, 3) and ( 3, 1) on the circle. √
(a)
[ √Show] that the line segment has the rectangular equation y = −x + 1 + 3 for x in the interval
1, 3 .
√
1+ 3
(b) Show that the line segment has the polar equation r = for θ in the interval [π/6, π/3].
cos(θ) +√sin(θ)
1+ 3 ( π)
c) Show that the line segment has the polar equation r = √ csc θ + for θ in the interval
2 4
[π/6, π/3].
d) Using polar coordinates, calculate the area of the sliver of the illustrated circle that lies above the
illustrated secant.
e) (Optional. Only for those who wish to prolong the enjoyment of this exercise set) It is a fact known
to few humans that ( )
5π √
cot = 2 − 3.
12
Most of the handful of humans who have derived this formula have obtained it from trigonometric
identities. Join the select few who have espied this cotangent evaluation by redoing the calculation of
part (d) using Cartesian coordinates and comparing the outcomes.
Solutions: Areas and Distances
(Corresponds to Stewart 5.1)

1. i) Let ∆x denote the width of each rectangle. Then ∆x = 1. From left to right, the heights of the
rectangles are f (0) = 2, f (1) = 4, f (2) = 5, f (3) = 7, f (4) = 8, f (5) = 7. The approximate area is the
sum of the areas of the rectangles: (2 + 4 + 5 + 7 + 8 + 7) ∆x, or 33.

ii) Let ∆x denote the width of each rectangle. Then ∆x = 1. From left to right, the heights of the
rectangles are f (1) = 4, f (2) = 5, f (3) = 7, f (4) = 8, f (5) = 7, f (6) = 4. The approximate area is the
sum of the areas of the rectangles: (4 + 5 + 7 + 8 + 7 + 4) ∆x, or 35.

iii) Let ∆x denote the width of each rectangle. Then ∆x = 2. From left to right, the heights of the
rectangles are f (1) = 4, f (3) = 7, f (5) = 7. The approximate area is the sum of the areas of the
rectangles: (4 + 7 + 7) ∆x, or 36.
2. i) Let ∆x denote the width of each rectangle. Then ∆x = 1/4. From left to right, the heights of the rect-
angles are f (1) = 5, f (5/4) = 4, f (3/2) = 7, f (7/4) = 8, f (2) = 6, f (9/4) = 6, f (5/2) = 5, f (11/4) = 3.
The approximate area is the sum of the areas of the rectangles: (5 + 4 + 7 + 8 + 6 + 6 + 5 + 3) ∆x, or 11.

ii) Let ∆x denote the width of each rectangle. Then ∆x = 1/4. From left to right, the heights of the rect-
angles are f (5/4) = 4, f (3/2) = 7, f (7/4) = 8, f (2) = 6, f (9/4) = 6, f (5/2) = 5, f (11/4) = 3, f (3) = 2.
The approximate area is the sum of the areas of the rectangles: (4 + 7 + 8 + 6 + 6 + 5 + 3 + 2) ∆x, or
10.25.

iii) Let ∆x denote the width of each rectangle. Then ∆x = 1/2. From left to right, the heights of the
rectangles are f (5/4) = 4, f (7/4) = 8, f (9/4) = 6, f (11/4) = 3. The approximate area is the sum of the
areas of the rectangles: (4 + 8 + 6 + 3) ∆x, or 10.5.
3. i) (54 + 61 + 65) fts × 20 s, or 3600 ft.

ii) (61 + 65 + 62) fts × 20 s, or 3760 ft.

iii) (57 + 66 + 63) fts × 20 s, or 3720 ft.

iv) (54 + 57 + 61 + 66 + 65 + 63) fts × 10 s, or 3660 ft.

83
84

v) (57 + 61 + 66 + 65 + 63 + 62) fts × 10 s, or 3740 ft.

vi) One can start from scratch and average the left and right endpoints of all six subintervals. However,
we can use the calculations already performed in parts (iv) and (v). The answer is the average of the
estimates found in those two parts: (3660 + 3740)/2 ft, or 3700 ft.
Solutions: Riemann Sums, Riemann
Integrals, Definite Integrals, Midpoint
Rule (Corresponds to Stewart 5.2)

1. i) Here, ∆x = (4−3)/3 = 1/3. The nodes are x0 = 3, x1 = x0 +∆x = 10/3, x2 = x1 +∆x = 11/3, x3 = 4.
The sample points are s1 = x0 = 3, s2 = x1 = 10/3, s3 = x2 = 11/3. We calculate f (s1 ) =
1.73205, f (s2 ) = 1.82574, f (s3 ) = 1.91485. The requested Riemann sum is ( f (s1 ) + f (s2 ) + f (s3 ) ) ∆x =
1.8242.

ii) The sample points are now s1 = x1 = 10/3, s2 = x2 = 11/3, s3 = x3 = 4. We calculate f (s1 ) =
1.82574, f (s2 ) = 1.91485, f (s3 ) = 2. The requested Riemann sum is ( f (s1 ) + f (s2 ) + f (s3 ) ) ∆x =
1.9135.

iii) The sample points are now s1 = x0 + (∆x)/2 = 19/6, s2 = x1 + (∆x)/2 = 7/2, s3 = x2 + (∆x)/2 =
23/6. We calculate f (s1 ) = 1.77952, f (s2 ) = 1.87083, f (s3 ) = 1.95789. The requested Riemann sum is
( f (s1 ) + f (s2 ) + f (s3 ) ) ∆x = 1.8694.

iv) The sample points are now s1 = 47/15, s2 = s1 + ∆x = 52/15, s3 = s2 + ∆x = 19/5. We calculate
f (s1 ) = 1.77012, f (s2 ) = 1.86189, f (s3 ) = 1.94936. The requested Riemann sum is ( f (s1 ) + f (s2 ) + f (s3 ) ) ∆x =
1.8605. The rectangles corresponding to this Riemann sum are shown in the figure below.

85
86

2. Here, ∆x = (4 − 1)/4 = 3/4. The nodes are x0 = 1, x1 = x0 + ∆x = 7/4, x2 = x1 + ∆x = 5/2, x3 = x2 +

∆x = 13/4x4 = x3 +∆x = 4. The sample points are s1 = x0 +(∆x)/3 = 5/4, s2 = x1 +(∆x)/3 = 2, s3 =
x2 + (∆x)/3 = 11/4, s4 = x3 + (∆x)/3 = 7/2. We calculate f (s1 ) = 0.80000, f (s2 ) = 0.50000, f (s3 ) =
0.36364, f (s4 ) = 0.28571. The requested Riemann sum is ( f (s1 ) + f (s2 ) + f (s3 ) + f (s4 ) ) ∆x = 1.4620.

3. The width of each subinterval is ∆x = (2−1)/5 = 0.2. The nodes of the partition are x0 = 1, x1 = x0 + ∆x = 1.
1 1
The sample points arising from the Midpoint Rule are s1 = x0 + ∆x = 1 + 0.1 = 1.1, s2 = x1 + ∆x = 1.2 +
( 2 ) 2
The Midpoint Rule approximation is 3 (1.1)2 + (1.3)2 + (1.5)2 + (1.7)2 + (1.9)2 × 0.2 or 6.99. With
n = 100, the approximation is 6.999975, which is quite accurate for an extremely modest amount of
computer calculation.

4. i) The area of the region under the graph of f and above the interval [1, 2] of the x-axis is 5/12. The
area of the region above the graph of f and under the interval [2, 4] of the x-axis is 8/3. Use the given
∫4
information to evaluate 1 f (x) dx.

∫2
ii) Estimate the definite integral 1
f (x) dx using the Midpoint Rule with n = 3.

∫4
iii) Estimate the definite integral 2
f (x) dx using the Midpoint Rule with n = 6.

∫4
iv) Estimate the definite integral 1
f (x) dx using the Midpoint Rule with n = 9.
∫ 4
5 8 9
i) f (x) dx = − =− .
1 12 3 4

ii) Here, ∆x = (2 − 1)/3 = 1/3. The nodes are x0 = 1, x1 = 4/3, x2 = 5/3, x3 = 2. The sam-
ple points are s1 = x0 + (∆x)/2 = 7/6, s2 = x1 + (∆x)/2 = 9/6, s3 = x2 + (∆x)/2 = 11/6. We
calculate f (s1 ) = 0.39351, f (s2 ) = 0.62500, f (s3 ) = 0.30092. The Midpoint Rule approximation is
87

( f (s1 ) + f (s2 ) + f (s3 ) ) ∆x = 0.4398.

iii) Here, ∆x = (4 − 2)/6 = 1/3. The nodes are x0 = 2, x1 = 7/3, x2 = 8/3, x3 = 3, x4 = 10/3, x5 =
11/3, x6 = 4. The sample points are s1 = x0 + (∆x)/2 = 13/6, s2 = x1 + (∆x)/2 = 15/6, s3 =
x2 + (∆x)/2 = 17/6, s4 = x3 + (∆x)/2 = 19/6, s5 = x4 + (∆x)/2 = 21/6, s6 = x5 + (∆x)/2 = 23/6.
We calculate f (s1 ) = −0.35648, f (s2 ) = −1.12500, f (s3 ) = −1.78240, f (s4 ) = −2.10648, f (s5 ) =
−1.87500, f (s6 ) = −0.86574. The Midpoint Rule approximation is ( f (s1 ) + f (s2 ) + f (s3 ) + f (s4 ) + f (s5 ) + f (s6 ) ) ∆x =
−2.7037.

iv) We can calculate the requested Midpoint Rule approximation from scratch, but that would duplicate
all the work already done. Notice that, for the requested Midpoint Rule, ∆x = (4−1)/9 = 1/3. Because
∆x = 1/3 in both parts (ii) and (iii), we can combine the two calculations: the nodes found in part
(ii) partition the subinterval [1, 2] into 3 sub-subintervals and the nodes found in part (iii) partition the
subinterval [2, 4] into 6 sub-subintervals—altogether there are 9 equal-width subintervals of [1, 4]. We
find that the Midpoint Rule approximation is 0.4398 + (−2.7037), or −2.2639.

( )
5. The local extrema can be located by the first derivative test. We calculate f ′ (x) = 2 12 x2 − 43 x + 35 .
If you are skilled at factoring, you can write this as f ′ (x) = 2 (4x − 5) (3x − 7), which shows that 5/4 and
7/3 are the roots of f ′ . Without factoring, you can determine these roots using the Quadratic Formula.
With one possible exception, we can now determine the sample points for both parts (i) and (ii). The
possible exception is the sample point s2 in the subinterval [1, 3/2] for the lower Riemann sum. We
calculate f (1) = 14 and f (1/2) = 57/4 = 14 + 1/4 > f (1). The sample points for the upper Riemann
sum are s1 = 1, s2 = 5/4, s3 = 3/2, s4 = 2, s5 = 3. Because ∆x = (3 − 1/2)/5 = 1/2, the upper Riemann
sum R∗ is:

( ( ) ( ) ) ( )
5 3 239 57 1 1155
R∗ = f (1) + f +f + f (2) + f (3) ∆x = 14 + + + 11 + 18 = = 36.09375.
4 2 16 4 2 32

The sample points for the lower Riemann sum are s1 = 1/2, s2 = 1, s3 = 2, s4 = 7/3, s5 = 5/2. The
lower Riemann sum R∗ is:

( ( ) ( ) ( )) ( )
1 7 5 17 266 41 1 2665
R∗ = f + f (1) + f (2) + f +f ∆x = + 14 + 11 + + = ≈ 24.67593.
2 3 2 4 27 4 2 108

∫ 3
365
The exact evaluation of the integral is f (x) dx =
, just in case you ever wondered about the
1/2 12
significance of the ratio of the number of days in a year to the number of months. This ratio, about
30.41666667, is between the upper and lower Riemann sums, of which we could be certain even if we
could not determine its exact value.

The upper and lower Riemann sums are illustrated in the two figures that follow
88

6. For every j from 1 to N , the point sj is the midpoint of the j th subinterval. All we require for this
problem is that sj is a point in the j th subinterval. Had sj been an endpoint of the subinterval or
a randomly selected point inside the subinterval, there would have been no change whatsoever in the
solution that follows.
89

∑
N
4 s2j − 1 ∑
N
We must identify as f (sj ) ∆x for some function f and width ∆x. The latter quantity
j=1
N j=1
follows easily from the given information a = 1 and b = 5:

b−a 5−1 4
∆x = = = .
N N N

We express the given sum with the factor ∆x appearing explicitly:

∑
N ∑N ( ) ∑N ( )
4 s2j − 1 1 4 1
= sj −
2
= sj −
2
∆x.
j=1
N j=1
4 N j=1
4

If we define f (x) = x2 − 1/4, then we have

∑
N
4 s2j − 1 ∑
N
= f (sj ) ∆x.
j=1
N j=1

Thus,

∑
N
4 s2j − 1 ∑
N
lim = lim f (sj ) ∆x
N →∞ N N →∞
j=1 j=1
∫ 5
= f (x) dx
1
∫ 5 ( )
1
= x −
2
dx
1 4
( 3 )5
x x
= −
3 4 1
121
= .
3

7. Students possessing a very good facility for working with algebraic expressions might not need the first
step of this written solution, but even they may benefit from a reminder that studying a particular
case can provide an insight into the general situation. We will write the sum explicitly for N = 4.
Why this particular choice of N ? There is nothing special about N = 4, but it makes for a Goldilocks
choice: 4 subintervals are not so few as to inhibit detection of a pattern nor so large that calculation
is tedious. In other words, the number N of subintervals is just right when N = 4. For this value
of N , we have ∆x = (b − a)/N = (3 − 1)/4 = 1/2. The nodes x0 , x1 , x2 , x3 , x4 of the partition are
x0 = a = 1, x1 = x0 + ∆x = 1 + 1/2 = 3/2, x2 = x1 + ∆x = 3/2 + 1/2 = 2, x3 = x2 + ∆x = 2 + 1/2 =
5/2, x4 = x3 + ∆x = 5/2 + 1/2 = 3. The subintervals [x0 , x1 ], [x1 , x2 ], [x2 , x3 ], [x3 , x4 ] of the partition are
therefore [1, 3/2], [3/2, 2], [2, 5/2], [5/2, 3]. We will be looking for 4 sample points s1 , s2 , s3 , s3 , s4 with
s1 in the first subinterval of the partition, s2 in the second subinterval of the partition, s3 in the third,
and s4 in the fourth. To identify these sample points, we will rewrite the sum explicitly, without sigma
90

notation, making sure that the subinterval width ∆x, in this case 1/2, appears as a common factor:
N (
∑ )2 ( )2 ( )2 ( )2 ( )2
2j − 1 6 2−1 6 4−1 6 6−1 6 8−1 6
1+ = 1+ + 1+ + 1+ + 1+
j=1
N N 4 4 4 4 4 4 4 4
( )2 ( )2 ( )2 ( )2
5 6 7 6 9 6 11 6
= + + +
4 4 4 4 4 4 4 4
( )2 ( )2 ( )2 ( )2
5 1 7 1 9 1 11 1
= 3 +3 +3 +3 .
4 2 4 2 4 2 4 2

Let us now observe that 5/4, 7/4, 9/4, and 11/4 are points that are in, respectively, the subintervals
[1, 3/2], [3/2, 2], [2, 5/2], [5/2, 3]. Indeed, the four listed points are the midpoints of the listed subintervals.
It appears that we have our sample points s1 = 5/4, s2 = 7/4, s3 = 9/4, s4 = 11/4, and

∑N ( )2
2j − 1 6 2 2 2 2
1+ = 3 (s1 ) ∆x + 3 (s2 ) ∆x + 3 (s3 ) ∆x + 3 (s4 ) ∆x.
j=1
N N

We need a function f such that f (s1 ) = 3 s21 , f (s2 ) = 3 s22 , f (s3 ) = 3 s23 , f (s4 ) = 3 s24 . Clearly f (x) =
3 x2 does the job. In summary, for this function f , for N = 4, and for the midpoints of the subintervals
as the choice of sample points, we have

∑N ( )2 ∑N
2j − 1 6
1+ = f (sj ) ∆x.
j=1
N N j=1

Now we know what we’re looking for. Let us proceed to the general case. Letting a = 1 and b = 3, the
width of the entire interval is b − a, or 2. Therefore, the subinterval width ∆x is given by ∆x = 2/N .
We can rewrite the given sum in terms of ∆x as follows:

∑N ( )2 ∑
N ( ( ) )2
2j − 1 6 1 2 2
1+ = 3 1+ j−
j=1
N N j=1
2 N N

∑
N ( ( ) )2
1
= 3 1+ j− ∆x ∆x
j=1
2

∑
N ( )2
1
= 3 1 + ∆x + (j − 1)∆x ∆x.
j=1
2

For j = 1, we see that

1 1
s1 = 1 + ∆x + (1 − 1)∆x = a + ∆x,
2 2
which is in the first subinterval. For j = 2, we see that
1 1 1 1
s2 = 1 + ∆x + (2 − 1)∆x = 1 + ∆x + ∆x = (a + ∆x) + ∆x = x1 + ∆x,
2 2 2 2
which is in the second subinterval. For j = 3, we see that
1 1 1 1
s3 = 1 + ∆x + (3 − 1)∆x = a + ∆x + 2 ∆x = (a + 2 ∆x) + ∆x = x2 + ∆x,
2 2 2 2
91

which is in the third subinterval. And so on. For f (x) = 3 x2 and for midpoints as the choice of sample
points, we have
∑N ( )2 ∑N
2j − 1 6
1+ = f (sj ) ∆x.
j=1
N N j=1

Thus,
N (
∑ )2 ∑N ∫ 3 ∫ 3
2j − 1 6
lim 1+ = lim f (sj ) ∆x = f (x) dx = 3 x2 dx = 26.
N →∞ N N N →∞ 1 1
j=1 j=1
92
Solutions: The Fundamental Theorem
of Calculus (Corresponds to Stewart
5.3)

1. 13

2. 38/3

3. 3/2

4. 38/3

5. 32/15

6. e2 /(e + 1)

7. 4
√
8. 2 3/3
√ √
9. 3 2 − 2 3

10. π/4

11. 2 ln(2)

12. π/6

13. We have ( 2 ) ( )
3x2 − 1 3x + 3 − 4 3 x2 + 1 − 4 4
= = =3− 2 .
x2 + 1 x2 + 1 x2 + 1 x +1
Therefore,
∫ ∫ ( )
1
3x2 − 1 1
4 x=1
dx = 3− dx = (3 x − 4 arctan(x))|x=0 = 3 − π.
0 x2 + 1 0 x2 + 1

14. The identity 1 + tan2 (α) = sec2 (α) tells us that tan2 (α) = sec2 (α) − 1. Therefore,
∫ ∫ ∫ ∫ √
π/3 π/3 ( ) π/3 π/3
2 3 π
2
tan (α) dα = sec (α) − 1 dα =
2
sec (α) dα −
2
1 dα = − .
π/6 π/6 π/6 π/6 3 6

93
94

15. We begin by expanding the exponential:

∫ 1 ∫ 1 ∫ t=1
2+t 2 t
1
t 3t 18
3 dt = 3 3 dt = 9 3 dt = 9 = .
0 0 0 ln(3) t=0 ln(3)
√ √
16. F ′ (x) = 1 + t3 t=x = 1 + x3 and F ′ (2) = 3

19 + u4 19 + s4
17. G′ (s) = 2 = and G′ (3) = 100/10 = 10
1 + u u=s 1 + s2
∫ x√
18. First observe that F (x) = − 17 + 16 sin(s) ds. Therefore
2π
√ √

D(F )(x) = −
17 + 16 sin(s) = − 17 + 16 sin(x)
s=x
√ √
and D(F )(π/6) = − 17 + 16 sin(π/6) = − 17 + 8 = −5.
19. For u = x2 , we have
∫ x2 √ √
d
D(F )(x) = 1 + t dt
dx 1
∫ u√ √
d
= 1 + t dt
dx
( 1∫ u √ )( )
d √ du
= 1 + t dt
du 1 dx
(√ )
√
= 1 + t (2x)
(√ )
t=u
√
= 1 + u (2x)
(√ √ )
= 1 + x2 (2x)
√
= 2 x 1 + | x |.
√
and D(F )(3) = 2 · 3 · 4 = 12.
∫ 2π (π )
20. Calculate G′ (x) and G′ (2) for G(x) = sin dt. For u = 3x, we have
3x t
∫ 2π (π)
′ d
G (x) = sin dt
dx 3x t
∫ 2π (π)
d
= sin dt
dx t
u∫
u
(π)
d
= − sin dt
dx t
( 2π ∫ u ( π ) ) ( du )
d
= − sin dt
du 2π t dx
( ( π ) )

= − sin (3)
( πt) t=u
= −3 sin
( uπ )
= −3 sin
3x
95

and G′ (2) = −3 sin (π/6) = −3/2.

21. Let u = x2 and w = x3 . Choose any constant, such as 7. Ultimately we are interested in the value
x = 2. It so happens that when x = 2, we have u = 4, w = 8, and 7 is between 4 and 8. However, had
we chosen a different constant not between 4 and 8, a constant such as −13 or 100000, the calculation
would be exactly the same and the answer would be the same. We have
∫ x3 ( ) ∫ 7 ( ) ∫ x3 ( ) ∫ x3 ( ) ∫ x2 ( )
16 16 16 16 16
F (x) = exp dz = exp dz+ exp dz = exp dz− exp dz.
x2 z x2 z 7 z 7 z 7 z

Therefore
∫ x3 ( ) ∫ x2 ( )
′ d 16 d 16
F (x) = exp dz − exp dz
dx 7 z dx 7 z
∫ w ( ) ∫ u ( )
d 16 d 16
= exp dz − exp dz
dx z dx z
( 7∫ w ( ) )( 7 ) ( ∫ u ( ) )( )
d 16 dw d 16 du
= exp dz − exp dz
dw 7 z dx du 7 z dx
( ( )) ( ( ))
16 ( 2) 16
= exp 3x − exp (2x)
w u
( ( )) ( ( ))
16 ( 2) 16
= exp 3
3x − exp (2x)
x x2

and ( ( )) ( ( ))
′ 16 16
F (2) = exp (12) − exp (4) = 12 exp (2) − 4 exp (4) .
8 4
96
Solutions: The So-Called Net Change
Theorem (Corresponds to Stewart 5.4)

1. By the Net Change Theorem, the decrease in volume is

∫ 3 3

120 exp(−t) dt = − 120 exp(−t) = −120 exp(−3) + 120 exp(−2) = 120 e−3 (e − 1) .
2 2

2. The question asked, in other words, is, For what value of T is the decrease of volume during the time
interval [0, T ] equal to 120 - 60, or 60, liters? That is, we must solve the equation
∫ T
120 exp(−t) dt = 60
0
T

for T . The left side is − 120 exp(−t) , or −120 exp(−T ) + 120. Thus, −120 exp(−T ) + 120 = 60,
0
or 120 exp(−T ) = 60, or (exp(−T
) ) = 1/2. Applying the natural logarithm to each side, we obtain
−T = ln(1/2), or T = − ln 2−1 . Using the identity ln (up ) = p ln(u) with u = 2 and p = −1, we arrive
at the simplified value, T = ln(2).

3. Let f (t) be the rate of consumption at time t. Then

{
60 if 0 ≤ t ≤ 2
f (t) = 240
t2 if 2 ≤ t

The mass of leaf consumption in mg was

∫ ∫ ∫ ( 4 )
4 2 4
240 240
f (t) dt = 60 dt + dt = 120 + − = 120 + 60 = 180.
0 0 2 t2 t 2

∫ 24
4. By the Net Change Theorem, E = 0 P (t) dt. Notice that if P (t) is measured in megawatts and t is
measured in hors, then P (t) dt and hence E is measured in megawatt-hours. No directions for estimating
the integral were given, so several answers are acceptable. If the Midpoint Rule is to be used, then we
can either use ∆ t = 3 and estimate the values P (1.5), P (4.5), P (7.5), . . . , P (22.5), or we can use ∆ t = 6
and read off the values P (3), P (9), P (15), P (21). We will opt for the latter:

E ≈ (400 + 800 + 850 + 650) × 6 = 16200.

If we use left endpoints for the sample points, then we set ∆ t = 3 and obtain the estimate

E ≈ (450 + 400 + 500 + 800 + 850 + 850 + 750 + 650) × 3 = 15750.

97
98

Because P (0) = P (24), right endpoints for the sample points lead to the same estimate, 15750, as do
left endpoints.
As it happens, the actual value of E for the function graphed in the figure is 15789.433, to three decimal
places. That function will not be given here for space reasons: it is a “cubic spline” made up of 8
different degree 3 polynomials used on different subintervals.
Solutions: The Substitution Rule
(Corresponds to Stewart 5.5)

1. Letting u = 1 + 9 x, du = 9 dx, we have

∫ 7 ∫ u=64
√
3
64
1√3
1 4/3 85
1 + 9 x dx = u du = u = .
0 1 9 12 u=1 4

2. Letting u = 5 − x, du = −dx, we have

∫ 4 ∫ 4 u=4
√ √ 2 3/2 14
5 − x dx = u du = u = .
1 1 3 u=1 3

3. Letting u = 5 − x, du = −dx, we have

∫ 4 ∫
√ 4 √
x 5 − x dx = (5 − u) u du
1 1
∫ 4( )
= 5 u1/2 − u3/2 du
1
( )u=4
10 3/2 2 5/2
= u − u
3 5 u=1
( )u=4
2 3/2
= u (25 − 3 u)
15 u=1
164
= .
15
4. Letting u = 5 − x, du = −dx, we have
∫ 4 ∫
√ 4
2 √
x2 5 − x dx = (5 − u) u du
1 1
∫ 4( )
√
= 25 u − 10 u3/2 + u5/2 du
1
u=4
50 3/2 2
= u − 4 u5/2 + u7/2
3 7 u=1

1 3/2 ( )u=4
2
= u 350 − 84 u + 6 u
21 u=1
608
= .
21

99
100

5. Letting u = 4 − x2 , du = −2 x dx, we have

∫ 2 √ ∫
1 0√ 8
x 4 − x2 dx = − u du = .
0 2 4 3

6. Letting u = 4 − x2 , du = −2 x dx, we have x2 = 4 − u and

∫ 2 √ ∫ 2 √ ∫ ∫
1 0 √ 1 0 ( 1/2 ) 64
x3
4 − x dx =
2 x 2
4 − x x dx = −
2 (4 − u) u du = − 4 u − u3/2 du = .
0 0 2 4 2 4 15

7. Letting u = x2 + 1, du = 2 x dx, we have

∫ 1 ∫ 1( )
2x − 1 2x 1
2
dx = − dx
0 x +1 0 x2 + 1 x2 + 1
∫ 1 ∫ 1
2x 1
= 2+1
dx − 2+1
dx
0 x 0 x
∫ 2 ∫ 1
1 1
= du − 2+1
dx
1 u 0 x
u=2 x=1

= ln(u) − arctan(x)
u=1 x=0
π
= ln(2) − .
4
8. We begin with the observation
( 2 )
x2 + 4 x + 2 x + 1 + 4x + 1 4x 1
= =1+ 2 + .
x2 + 1 x2 + 1 x + 1 x2 + 1
Therefore, letting u = x2 + 1, du = 2 x dx, we have
∫ 1 2 ∫ 1( )
x + 4x + 2 4x 1
dx = 1+ 2 + dx
0 x2 + 1 0 x + 1 x2 + 1
∫ 1 ∫ 1 ∫ 1
2x 1
= 1 dx + 2 2
dx + 2
dx
0 0 x +1 0 x +1
∫ 2 ∫ 1
1 1
= 1+2 du + 2
dx
1 u 0 x +1
u=2 x=1

= 1 + 2 ln(u) − arctan(x)
u=1 x=0
π
= 1 + 2 ln(2) + .
4
9. Letting u = 8 − 7 x, du = −7 dx, we have
∫ 1 ∫ 1 ∫ 8( )
√ 1 1 249
x 8 − 7 x dx = − (8 − u) u du = 8 u1/3 − u4/3 du =
3 1/3
.
0 49 8 49 1 343

10. 27/140
11. Letting u = ln(x), du = (1/x) dx, we have
∫ 2 ∫ ln(2)
ln(x) 1 2
dx = u du = (ln (2)) .
1 x 0 2
101

12. Letting u = ln(x), du = (1/x) dx, we have

∫ ∫
1 1
dx = du = ln(| u |) + C = ln(| ln(x) |) + C.
x ln(x) u

13. Letting u = ln(x), du = (1/x) dx, we have

∫ ( ) ∫ e ∫ e ∫ 1
e
1 1 1 ( −1 ) 1 1
ln dx = ln x dx = − ln (x) dx = − u du = − .
1 x x 1 x 1 x 0 2
√ ( √ )
14. Letting u = x, du = 1/(2 x) dx, we have
∫ 4 √ ∫ 2
exp( x)
√ dx = 2 exp(u) du = 2 e (e − 1).
1 x 1

15. Letting u = tan(x), du = sec2 (x) dx, we have

∫ π/4 ∫ 1
2 1
tan(x) sec (x) dx = u du = .
0 0 2

16. Letting u = tan(x), du = sec2 (x) dx, we have

∫ ∫ √
π/3 3 √
tan2 (x) sec2 (x) dx = u2 du = 3.
0 0
102
Solutions: Areas Between Curves
(Corresponds to Stewart 6.1)

1. The region is shown in the figure.

The area is
∫ ∫
e ( ) 1 ( )
1 − ln(x) dx = exp(y) − 1 dy = e − 2.
1 0

2. The region is shown in the figure.

103
104

The area is
∫ π/3 π/3 ( √
√ ) 3

sec (x) − cos (x) dx = ln ( | sec (x) + tan (x) | ) − sin (x) = ln 2 + 3 − .
0 0 2

3. The region is shown in the figure.

The area is ∫ 2( ) 2
( ) 1 1 9
(x + 1) − x2 − 1 dx = x2 + 2 x − x3 = .
−1 2 3 −1 2

4. The two given curves intersect at the point (−1, 1) for the left value x = −1 of the region and at the
point (2, 7) for the right value x = 2 of the region. Because a cubic may intersect a quadratic in as
105

many as three points, we must investigate whether there is an additional point of intersection between
the two already observed. We can answer this question by determining if there is a linear term x − c
such that

( )
x3 − x + 1 − 2 x2 − 1 = x3 − 2x2 − x + 2 = (x + 1)(x − 2)(x − c).

The constant term of the middle expression is 2, and the constant term of the expression on the right
is (1)(−2)(−c), or 2c. We see that, if there is such a factoring, then c must be 1. We can then verify
by expansion that the factoring does indeed hold for c = 1. It follows that (1, 1) is another point of
intersection of the two curves. It is where we will break up the integration. The region is shown in the
figure.

The area is

∫ 1( ) ∫ 2( )
( 3 ) ( 2 ) ( 2 ) ( 3 ) 37
x − x + 1 − 2 x − 1 dx + 2 x − 1 − x − x + 1 dx = .
−1 1 12

5. The region is shown in the figure.

106

The area is ∫ 5 π/4 ( ) √

sin(x) − cos(x) dx = 2 2.
π/4

6. The region is shown in the figure.

The area is ∫ ∫
1 ( ) 2 ( ) 7
3 x − x2 dx + (x + 2) − x2 dx = .
0 1 3

7. The area is ∫ ∫
3 (√ y) 4
√ 7
y− dy + ( y − (y − 2)) dy = .
0 3 3 3
107

8. The region is shown in the figure.

∫ ( )
e
x−1
The area is ln(x) − dx. If you know the aniderivative of ln(x), and it is not assumed that
1 e−1
you do at this point, then you can calculate the area using this integral formula. The area is (3 − e)/2.
Otherwise, you can calculate the area this way:
∫ 1( ) ( )y=1
( ) 1 y
1
(e − 1) y + 1 − exp(y) dy = (e − 1) y + y − e
2
= (3 − e).
0 2 y=0 2

9. The region is shown in the figure.

108

The area is ∫ ( )
1 ( 2 ) 10 1
e − y − 2 dy =
y
+e− .
−1 3 e

10. The area is

∫ −1 (√ ( √ )
) ∫ 1/e ( )
∫ e ( )
x + 2 − − x + 2 dx + 1 − (−1) dx + 1 − ln(x) dx.
−2 −1 1/e

11. The region is shown in the figure.

The area is ∫ −4 (√ ( √ )) ∫ 3 (√ )
25 − x − − 25 − x
2 2 dx + 25 − x − (x + 1) dx.
2
−5 −4

12. (i) 4, 4, 254.4; (ii) 8, 2, 253.62

13. (i) 3, 1.7, 61.455; (ii) 6, 0.85, 48.586
Solutions: Volumes by Disks and
Washers (Corresponds to Stewart 6.2)

1. The requested volume V is

∫ 2 ( )2 31
V = π x2 dx = π.
1 5

2. The requested volume V is

∫ π/2 (√ )2
V = π cos(x) dx = 2 π.
−π/2

109
110

3. The requested volume V is

∫ π/4
2
V = π (sec(x)) dx = π.
0

4. The requested volume V is

∫ 9 (√ )2 65
V = π x dx = π.
4 2

5. The requested volume V is

∫ 4 ( )2 781
V = π y 2 dy = π.
3 5

6. The requested volume V is

∫ 81 ( )2 1330
V = π y 1/4 dy = π.
16 3
111

7. The requested volume V is

( )2
∫ 5
1 14
V = π dy = π.
1/3 y 5

8. The requested volume V is

∫ 2 ( )2 ( )2 256
V = π 2 x2 − π x3 dx = π.
0 35
112

9. The requested volume V is

∫ 1
2 (√ )2 1 ( )
V = π (ex ) − π x + 1 dx = π e2 − 4 .
0 2
113

√
10. Notice that y = 4 − x2 is, for 0 ≤ x ≤ 2, equivalent to x = 4 − y, 0 ≤ y ≤ 4 and y = (x − 2)2 is
√
equivalent to x = 2 + y. The requested volume V is
∫ (√ )2
4
√ 2 112
V = π (2 + y) − π 4 − y dy = π.
0 3

√
11. Notice that y = 3 x is equivalent to x = y 2 /3. The requested volume V is
∫ 3( ( 2 )2 )
y 18
V = πy −π
2
dy = π.
0 3 5

12. Notice that y = ln(x) is equivalent to x = exp(y) and y = 2 − x/e is equivalent to x = e (2 − y). The
requested volume V is
∫ 1( )
( )2 1 ( )
V = π e (2 − y) − π exp(2y) dy = π 11 e2 + 3 .
0 6
114

∫ ( ( )2 )
4
√ 2 1
13. π ( y + 2) − π (y + 2) + 2 dy
1 3
∫ ( ( )2 )
4
1 √ 2
14. π 2 − (y + 2) − π (2 − y) dy
1 3
∫ ( ( )2 ( )2 )
2
1 1
15. π (3 x − 2) − −π x −
2
dx
1 2 2
∫ ( ( )
2 )2 2
16. π 6 − x2 − π (6 − (3 x − 2)) dx
1
∫ ( (
1 )2 ( ( ))2 )
17. π 5 − (x + 2) − π 5 − 4 − x2 dx
−2
∫ ( ( )2 )
3 √ 2
18. π 5 + 4 − y − π (5 − (y − 2)) dy
0
Solutions: Volumes by Cylindrical
Shells (Corresponds to Stewart 6.3)

1. The requested volume V is

∫ 1 (√ ) 3
V = 2πx x − x2 dx = π.
0 10

2. The requested volume V is

∫ 2
4
V = 2πx dx = 4 π ln(3).
0 2 + x2

115
116

3. The requested volume V is

∫ 4
√ 128
V = 2πy y dy = π.
0 5

√
4. First note that y = x − 3 is equivalent to x = y + 3 and y = x is equivalent to x = y 2 . The requested
volume V is ∫ 2
( ) 37
V = 2 π y (y + 3) − y 2 dy = π.
1 6
∫ π
5. 2 π (4 − x) sin(x) dx
0
∫ π/4
6. 2 π (x + π) (cos(x) − sin(x)) dx
0
∫ 2 ( )
7. 2 π (x + 2) (3 x − 2) − x2 dx
1
117

∫ 2 ( )
8. 2 π (2 − x) (3 x − 2) − x dx
2
1
∫ ( )( )
4
1 √ 1
9. y−
2π y − (y + 2) dy
1 2 3
∫ 4 ( )
√ 1
10. 2 π (6 − y) y − (y + 2) dy
1 3
118
Solutions: Work (Corresponds to
Stewart 6.4)
∫ 7 in
lb
1. The work done is 8 x dx, or 196 ft-in, or 49/3 ft-lb.
0 in
∫ 12 cm
N
2. The work done is 20 x dx, or 1440 N-cm, or 14.4 J.
0 cm
3. The given information allows us to solve for the spring constant:
∫ 0.02 m x=0.02 m
1 2
8J = kx dx = kx = 0.0002k m2 ,
0m 2 x=0 m

which implies that k = 40000 N/m. The force necessary to maintain the spring at x = 0.02 m is
k × 0.02 m, or 40000 × 0.02 N, or 800 N.
4. The given information allows us to solve for the spring constant:
∫ 0.2 m x=0.2 m
1 2
5J = kx dx = kx = 0.02k m2
0m 2 x=0 m

or
5
J · m−2 = 250 N · m · m−2 = 250 N · m−1 .
k=
0.02
The required work W is calculated as follows:
∫ 0.4 m x=0.4 m
1 1 ( )
W = kx dx = kx2 = k 0.42 m2 − 0.22 m2 = 0.06k m2 .
0.2 m 2 x=0.2 m 2

The answer can now be obtained by substituting in the value for the spring constant:

W = (0.06) (250 N · m−1 ) m2 = 15 N · m = 15 J.

5. The equation ( )
1
240N = F = k x = k (10) cm = k m
10
gives us k = 2400 N/m. The requested work W is
∫ ∫ 1/10 m
10 cm
N 1/10 m
N 2 N
W = k x dx = 2400 x dx = 1200 x = 12 m2 = 12 J.
0 m 0 m x=0 m

119
120

6. The equation
( )
4
280lb = F = k x = k (4) in = k ft
12
gives us k = 840 lb/ft. The requested work W is
∫ ∫ 2/3 ft
8 in
lb 8/12 ft
lb 2 22 − 12 lb 2
W = k x dx = 840 x dx = 420 x = 420 × ft = 140 ft lb.
4 in ft 4/12 ft ft 1/3 ft 32 ft

7. The given information can be expressed by the equation

∫ 8/100 m
8 cm
1
2 64 −4
40 J = k x dx = k x = 10 k m2 ,
0 cm 2 0m 2

or k = 12500 N/m. The work W done stretching the spring from 8 cm to 12 cm is given by
∫ 12/100 m
12 cm
1 1 N 122 − 82 2
W = k x dx = k x2 = 12500 m = 50 J.
8 cm 2 8/100 m 2 m 104

8. The given information can be expressed by the equation

∫ 2/12 ft
2 2 in
1 1
ft lb = k x dx = k x2 = k ft2 ,
3 0 in 2 0 ft 72

or k = 48 lb/ft. The force F required to maintain the spring at x = 2 in is

lb lb 1
F = k × 2 in = 48 × 2 in = 48 × ft = 8 lb.
ft ft 6

9. The work W done is

∫ 10 ft
lb
W = 62.42796 3 152 ft2 y dy
ft 0 ft
y=10 ft
62.42796 × 152 lb 2
= y
2 ft y=0 ft

= 62.42796 × 152 × 50 ft lb
= 702314.55 ft lb.

10. The work W done is

∫ 3m
m 3 kg
W = 9.80665 10 6 × 10m2 y dy
s2 m3 1m
y=3 m
9806.65 × 6 × 10 kg 2
= y
2 s2 y=1 m
( ) kg m2
= 9806.65 × ×30 × 32 − 12
s2
= 2353596 J.
121

11. The work W done is

∫ 2+8/2 ft
lb
W = 62.42796 16 × 30ft2 y dy
ft3 2 ft
y=6 ft
62.42796 × 480 lb 2
= y
2 ft y=2 ft
( )
= 62.42796 × 240 62 − 22 ft lb
= 479446.7 ft lb.

12. The work W done is

∫ 6 ft √
lb
W = 62.42796 16 ft × 2 × 36 − y 2 y dy
ft3 0 ft
y=6 ft
( )
2 3/2 lb
= − 665.89824 36 − y
ft2 y=0 ft
= 665.89824 × 63 ft lb
= 143834 ft lb.

13.

The work W done is

∫ 50 ft (y )2 ( )y=50
lb 1 4 25 3 625 2
W = 62.42796 3 π× + 25 ft y dy = 62.42796 π y + y + y ) = 434126920 ft lb.
ft 0 ft 2 16 3 2 y=0

14. The work W done is given by

∫ 50 ft
50 ft
lb 3 lb 2 3 lb 2 2
W = 3 y dy = y = 50 ft = 3750 ft lb.
0 ft ft 2 ft 0 ft 2 ft
122

15. The 20 ft closest the ground weighs 60 lb and is lifted 30 ft. Therefore, the work done on the 20 ft
closest the ground is (60)(30) ft lb. The total work W done is given by
∫ 30 ft
30 ft
lb 3 lb 2 3 lb 2 2
W = (60)(30) ft lb + 3 y dy = 1800 ft lb + y = 1800 ft lb + 30 ft = 3150 ft lb.
0 ft ft 2 ft 0 ft 2 ft

16. There are three components to the total work: the work done lifting the load, the work done lifting
the 40 - 24, or 16, meters of cable closest the ground, and the work done lifting the 24 meters of cable
closest to the roof. The total work W done is given by
∫ 24 m
N
W = 300 × 9.80665 × 24J + 220 × 16 × 24J + 220 y dy = 218447.88 J.
0m m

(Make sure you understand why one summand has the factor 9.80665 but the other two do not.)
17. The total work W done is given by
∫ 100 ft
lb
W = (300)(100) ft lb + 20 y dy = 130 000 ft lb.
0 ft ft

18. When the lift begins, there are 100 − 30, or 70, feet of cable hanging over the side of the building. The
load and the lowest 20 feet of cable are lifted 50 feet. There are three components to the total work: the
work done lifting the load, the work done lifting the 20 feet of cable closest the ground, and the work
done lifting the 50 feet of cable closest to the roof. The total work W done is given by
∫ 50 ft
lb lb
W = 300 × 50ft lb + 20 ft × 20 × 50 ft + 20 y dy = 60 000 ft-lb.
ft 0m ft
Solutions: Average Value
(Corresponds to Stewart 6.5)

1. favg = 3/(2 π), c = arcsin(3/(2 π)) ≈ 0.497767.

1 2
2. favg = ln(2), c= ≈ 2.88539.
2 ln(2)
3. favg = 2/3, c = 4/9.
√
4. favg = 8, c = 7 − 1 ≈ 1.64575.
√
5. favg = 20, c = 3.

6. favg = 4/3, c = 25/9.

7. Making the substitution u = 169 − x2 , du = −2 x dx, we have
∫ 12 √ ∫ 25
1 1 1 229
favg = x 169 − x2 dx = − · · u1/2 du = .
12 − 5 5 7 2 144 3
√ ( )
We seek a value c between 5 and 12 such that c 169 − c2 = 229/3, or c2 169 − c2 = (229/3)2 , or
( 2 )2
9 c − (9)(169) c2 + (229)2 = 0. This is a quadratic equation with respect to the variable c2 . The
169 7 √
quadratic formula gives c2 = ± 965. Both of these values give rise to a c in the interval (5, 12).
2 6
Therefore, we have √
169 7 √
c= ± 965
2 6
with both values of c in the interval (5, 12) and satisfying f (c) = favg . Twins! The values of c are,
approximately, 6.946808 for the minus sign and 10.98826 for the plus sign.)

123
124
Solutions: Integration by Parts
(Corresponds to Stewart 7.1)

1. Let u = x, dv = exp(x) dx. Then du = dx, v = exp(x) and

∫ ∫ ∫
x exp(x) dx = u v − v du = x exp(x) − exp(x) dx = x exp(x) − exp(x) + C = (x − 1) ex + C.

2. Let u = x, dv = exp(−x) dx. Then du = dx, v = − exp(−x) and

∫ ∫ ∫
x exp(−x) dx = u v− v du = −x exp(−x)+ exp(−x) dx = −x exp(−x)−exp(−x)+C = − (x + 1) e−x +C.

3. Let u = 2x + 5, dv = exp(x/3) dx. Then du = 2 dx, v = 3 exp(x/3) and

∫ ∫
(2x + 5) ex/3 dx = u v − v du
∫
= 3 (2x + 5) exp(x/3) − (3)(2) exp(x/3) dx

= 3 (2x + 5) exp(x/3) − 18 exp(x/3) + C

= 3 (2x − 1) exp(x/3) + C.

4. Let u = x, dv = sin(x) dx. Then du = dx, v = − cos(x) and

∫ ∫ ∫
x sin (x) dx = u v − v du = −x cos(x) + cos(x) dx = −x cos(x) + sin(x) + C.

5. Let u = 4 x + 2, dv = sin(2x) dx. Then du = 4 dx, v = −(1/2) cos(2x), and

∫ ∫ ∫
1 1
(4 x+2) sin(2 x) dx = u v− v du = − (4 x+2) cos(2x)+4 cos(2x) dx = −(2 x+1) cos(2x)+sin(2x)+C.
2 2

6. Let u = x, dv = 9 cos(3x) dx. Then du = dx, v = 3 sin(3x), and

∫ ∫ ∫
9x cos (3x) dx = u v − v du = 3x sin(3x) − 3 sin(3x) dx = 3x sin(3x) + cos(3x) + C.

125
126

7. Let u = ln(4x) = ln(4) + ln(x), dv = x dx. Then du = (1/x) dx, v = (1/2) x2 and
∫ ∫
x ln (4x) dx = u v − v du
2

∫ 2
1 2 1 x
= x ln(4x) − dx
2 2 x
∫
1 2 1
= x ln(4x) − x dx
2 2
1 2 1
= x ln(4x) − x2 + C
2 4
1 2( )
= x 2 ln(4x) − 1 + C.
4
Actually, this calculation could have been done in much greater generality. Suppose that a and p are
constants with a ̸= 0 and p ̸= −1. Setting u = ln(ax), dv = xp dx, we have du = (1/x) dx, v =
(1/(p + 1)) xp+1 , and
∫ ∫
xp ln (a x) dx = u v − v du
∫ p+1
1 1 x
= xp+1 ln (a x) − dx
p+1 p+1 x
∫
1 1
= xp+1 ln (a x) − xp dx
p+1 p+1
1 1
= xp+1 ln (a x) − xp+1 + C
p+1 (p + 1)2
( )
1 1
= x p+1
ln (a x) − + C.
p+1 p+1

8. Let u = ln(x), dv = 9x2 dx. Then du = (1/x) dx, v = 3 x3 and

∫
9 x2 ln (x) dx = x3 (3 ln (x) − 1) + C,

which agrees with the result of the general calculation with which the solution to Exercise 7 concludes.
9. Let u = ln(x), dv = x−2 dx. Then du = (1/x) dx, v = −x−1 , and
∫
ln (x) (1 + ln(x))
dx = − + C,
x2 x
which agrees with the result of the general calculation with which the solution to Exercise 7 concludes.
10. ∫Setting a = 2 and p = 3, follow the general calculation with which the solution to Exercise 7 concludes:
16 x3 ln (2x) dx = x4 (4 ln (2x) − 1) + C.
11. With u = x and dv = e−x/2 , we have du = dx, v = −2 e−x/2 , and
∫ 2 x=2 ∫ 2 ( )x=2
−x/2 −x/2 8
xe dx = −2x e +2 e−x/2 dx = −2x e−x/2 − 4 e−x/2 =4− .
0 x=0 0 x=0 e

12. With u = x and dv = cos (x), we have du = dx, v = sin (x), and
∫ π x=π ∫ π ( )x=π

x cos (x) dx = x sin(x) − sin(x) dx = x sin(x) + cos(x) = −1 − 1 = −2.
0 x=0 0 x=0
127

13. With u = x and dv = 4 sin (x/3), we have du = dx, v = −12 cos (x/3), and
∫ π/2 (x) ( x ) x=π/2 ∫ π/2 (x)

4 x sin dx = −12 x cos + 12 cos dx
0 3 3 x=0 0 3
( (x) ( x ))x=π/2

= −12 x cos + 36 sin
( 3
) ( ) 3 x=0
π π π
= −12 cos + 36 sin
2√ 6 6
= 18 − 3 3 π.

14. Integrate by parts with u = ln (x) and dv = dx. Then

∫ e e ∫ e ( )e
1
ln (x) dx = x ln(x) − x dx = x ln(x) − x = (e − e) − (0 − 1) = 1.
1 1 1 x 1

15. Integrating by parts with u = ln(3x) and dv = x dx, we have

∫ 1 ∫
1
1 2 1 1 21
x ln (3x) dx =
x ln(3x) − x dx
1/3 2 1/3 2 1/3 x
( )1
1 2 1
= x ln(3x) − x2
2 4 1/3
1 1 1 1
= ln(3) − + ·
2 4 4 9
1 2
= ln(3) − .
2 9
ln (3) /2 − 2/9
√ √
16. Let u = ln(x) and dv = (1/ x) dx. Then, du = (1/x) dx, v = 2 x dx, and
∫ ∫ ∫
ln (x) √ 1√ √ √
√ dx = 2 x ln(x) − 2 x dx = 2 x ln(x) − 2 x−1/2 dx = 2 x (ln(x) − 2) .
x x

It follows that
∫ 4 √ √
ln (x)
√ dx = 2 4 (ln(4) − 2) − 2 1 (ln(1) − 2) = 8 ln(2) − 4.
1 x

17. Integrate by parts with dv = 3x dx. Then v = (3x ) / ln(3) and

∫ x=1 ∫ 1
1
1 x
1
x
x 3 dx = x3 − 3x dx
0 ln(3) x=0 ln(3) 0
( )x=1
1 1 x

= x3 − 2
x
3
ln(3) ln (3)
( ) ( x=0
)
3 3 1
= − − 0− 2
ln(3) ln2 (3) ln (3)
1
= (3 ln(3) − 2) .
ln2 (3)
128

18. We integrate by parts twice. In each application we take dv = e−x dx:

∫ ∫
2 −x 2 −x
x e dx = −x e +2 x e−x dx
( ∫ )
= −x2 e−x + 2 −x e−x + e−x dx
( )
= −x2 e−x + 2 −x e−x − e−x + C
( )
= − x2 + 2x + 2 e−x + C.

19. In the first integration by parts we set u = x2 . In the second, we set u = x:

∫ ∫
x cos (x) dx = x sin(x) − 2
2 2
x sin(x) dx
( ∫ )
= x sin(x) − 2 −x cos(x) +
2
cos(x) dx

= x2 sin(x) − 2 (−x cos(x) + sin(x)) + C

( )
= x2 − 2 sin(x) + 2 x cos(x) + C.

20. In the first integration by parts we set u = ln2 (x). In the second, we set u = ln (x):
∫ ∫
16 x3 ln2 (x) dx = 4 x4 ln2 (x) − 8 x3 ln(x) dx
(∫ )
1 4 1
= 4 x ln (x) − 8
4 2
x ln(x) − 3
x dx
4 4
( )
1 4 1 4
= 4 x ln (x) − 8
4 2
x ln(x) − x +C
4 16
( )
1
= x4 4 ln2 (x) − 2 ln(x) + + C.
2

( )
21. Let u = 2 arctan (x) , dv = dx. Then du = 2/ 1 + x2 dx, v = x, and
∫ 1 x=1 ∫ 1
2x
2 arctan (x) dx = 2 x arctan (x) − 2
dx
0 x=0 0 1+x
( ( ))x=1
= 2x arctan (x) − ln 1 + x2
x=0
= 2 arctan(1) − ln(2) − (0 − 0)
π
= − ln(2).
2

( √ )
22. Integrate by parts with u = arcsin (x) , dv = dx. Then v = x, du = 1/ 1 − x2 dx, and
∫ x=1 ∫ (
1 √
1
x ( )1/2 )x=1 π
arcsin (x) 1− x2 dx = x arcsin(x) − √ dx = x arcsin(x) + 1 − x2 = −1.
0 x=0 0 1−x2 x=0 2

23. We use the reduction formula twice, setting a = 1/2 in both applications of the formula. We use p = 2
129

in the first application and p = 1 in the second:

∫ ∫
1 2 x/2 2
x2 ex/2 dx = x e − x1 ex/2 dx
1/2 1/2
∫
= 2x e 2 x/2
− 4 x1 ex/2 dx
( ∫ )
1 1 x/2 1
= 2 x2 ex/2 − 4 x e − x0 ex/2 dx
1/2 1/2
( )
= 2 x2 ex/2 − 4 2 x ex/2 − 4 ex/2 + C
( )
= 2 x2 − 4x + 8 ex/2 + C.

Therefore, ∫ 1 √
x2 ex/2 dx = 2 (1 − 4 + 8) e1/2 − 2 (8) = 10 e − 16.
0

24. We use the reduction formula twice, setting a = −2 in both applications of the formula. We use p = 2
in the first application and p = 1 in the second:
∫ 1 ∫
1 2
4 x2 e−2x dx = 4 · x2 e−2x − 4 · x1 e−2x dx
(−2) (−2)
0
∫
2 −2x
= −2 x e + 4 x1 e−2x dx
( ∫ )
2 −2x 1 −2x 1 −2x
= −2 x e +4 xe − e dx
(−2) (−2)
∫
= −2 x2 e−2x − 2 x e−2x + 2 e−2x dx

= −2 x2 e−2x − 2 x e−2x − e−2x + C

( )
= − 2 x2 + 2 x + 1 e−2x + C.

Therefore, ∫ 1
4x2 e−2x dx = − (2 + 2 + 1) e−2 + (0 + 0 + 1) e0 = 1 − 5 e−2 .
0

25.
∫ ∫ ( )
(√ )
2 ln x dx = 2 ln x1/2 dx
∫
1
= 2ln(x) dx
2
= x (ln(x) − 1) + C.

26. We first use a basic property of the logarithm:

∫ ∫
( 5)
4x ln x dx = 20 x ln (x) dx.

Now set a = 1 and p = 1, and follow the general calculation with which the solution to Exercise 7
concludes: ∫ ( )
( ) 1 1
4x ln x5 dx = 20 x2 ln (x) − + C = 5x2 (2 ln(x) − 1) + C.
2 4
130

27. We first use a basic property of the logarithm:

∫ ( ) ∫
1
16x3 ln dx = −16 x3 ln (x) dx.
x

Now set a = 1 and p = 3, and follow the general calculation with which the solution to Exercise 7
concludes:
∫ ( ) ( )
1 1
16x3 ln dx = −4 x4 ln (x) − + C = x4 (1 − 4 ln(x)) + C.
x 4

28. Set a = 1 and p = 1/2, and follow the general calculation with which the solution to Exercise 7 concludes:
∫ ( )
√ 9 3/2 1
9 x ln (x) dx = x ln (x) − + C = 2 x3/2 (3 ln (x) − 2) + C.
3/2 3/2

29. Integrating by parts with u = x, dv = sec2 (x) dx, we have du = dx, v = tan(x), and
∫ ∫
x sec2 (x) dx = x tan(x) − tan(x) dx

= x tan(x) − ln (|sec(x)|) + C.

30. Integrating by parts with u = x, dv = (x + 3)−1/2 dx, we have du = dx, v = 2 (x + 3)1/2 , and
∫ ∫
−1/2
x (x + 3) dx = 2 x (x + 3) 1/2
−2 (x + 3)1/2 dx
4
= 2 x (x + 3)1/2 − (x + 3)3/2 + C
3
2
= (x + 3)1/2 (x − 6) + C.
3

Therefore,
∫ 1
2 2 2 4
x (x + 3)−1/2 dx = (2) (−5) − (1) (−8) = (−10 + 8) = − .
−2 3 3 3 3

31. Integrating by parts with u = x, dv = (x − 1)1/2 dx, we have du = dx, v = (2/3) (x − 1)3/2 , and
∫ ∫
2 2
x (x − 1)1/2 dx = x (x − 1)3/2 − (x − 1)3/2 dx
3 3
( )
2 2 2
= x (x − 1)3/2 − (x − 1)5/2
+C
3 3 5
2
= (x − 1)3/2 (3x + 2) + C.
15

Therefore,
∫ 2
2 16
x (x − 1)1/2 dx = (6 + 2) − 0 = .
1 15 15
131
√
32. Let y = (x) = x1/2 . Then dy = (1/2) x−1/2 dx. It follows that dx = 2 y dy. Then, integration by
parts is applied with u = y and dv = sin(y) dy. Thus,
∫ ∫
(√ )
sin x dx = 2 y sin(y) dy
∫
= −2y cos(y) + 2 , cos(y) dy

= −2y cos(y) + 2 sin(y) + C

√ (√ ) (√ )
= −2 x cos x + 2 sin x + C.

33. Begin with the substitution y = x2 , dy = 2x dx. From these two equations we see that 2 x3 dx =
x2 2x dx = y dy. After making this subsitution, use integration by parts with u = y, dv = exp(y) dy.
Then du = dy v = exp(y). Thus,
∫ ∫
3
( 2)
2x exp x dx = y exp(y) dy
∫
= y exp(y) − , exp(y) dy

= y exp(y) − exp(y) + C
( ) ( )
= x2 − 1 exp x2 + C

34. Begin with the substitution y = x3 , dy = 3x2 dx. From these two equations we see that 3 x5 dx =
x3 3 x2 dx = y dy. After making this subsitution, use integration by parts with u = y, dv = sin(y) dy.
Then du = dy v = − cos(y). Thus,
∫ ∫
5
( 3)
3x sin x dx = y sin(y) dy
∫
= −y cos(y) + , cos(y) dy

= −y cos(y) + sin(y) + C
( ) ( )
= −x3 cos x3 + sin x3 + C
132
Solutions: Trigonometric Integrals
(Corresponds to Stewart 7.2)

∫ π/4
1. First make the change of variable u = x/2, du = (1/2) dx. The result is 2 sin(u)2 du. The an-
0
tiderivative of the integrand is u − cos(u) sin(u) + C. The answer is (π − 2)/4.

∫ π/2
2. First make the change of variable u = 2 x, du = 2 dx. The result is 24 cos(u)2 du. The antiderivative
π/3
√
of the integrand is 12 u + 12 cos(u) sin(u) + C. The answer is 2π − 3 3.

∫ ∫
π/3 π/3 ( )
3. First make the change of variable u = x/2, du = (1/2) dx. The result is 2
6 tan(u) du, or 6 sec(u)2 − 1 du.
0 0 √
The antiderivative of the integrand is 6 tan (u) − 6u + C and the definite integral evaluates to 6 3 − 2 π.

4. The integrand may be expanded and each term integrated. Easier is to use the identity cos2 (x) −
sin2 (x) = cos(2x). The answer is C + sin(2x)/2, or cos (x) sin (x) + C.

∫ ∫
3π/2 3π/2 ( )
5. First make the change of variable t = 2x+π, dt = 2 dx. The result is 24 cos(t)3 dt, or 24 1 − sin(t)2 cos(t) dt.
√ 4π/3 4π/3
∫ − 3/2 ( )
The change of variable u = sin(t), du = cos(t) dt results in 24 u2 − 1 du, which evaluates to
√ −1
9 3 − 16.

6. Making the changes of variable u = x/10, du = (1/10) dx first and w = cos(u), dw = − sin(u) du later,

133
134

we have

∫ (x) ∫
7 sin7 dx = 70 sin7 (u) du
10
∫
( )3
= 70 sin2 (u) sin(u) du
∫
( )3
= 70 1 − cos2 (u) sin(u) du
∫
( )3
= −70 1 − w2 dw
∫
( )
= −70 w − 3 w2 + 3 w4 − w6 dw
( )
3 5 1 7
= C − 70 w − w + w − w
3
5 7
= C − 70 cos(u) + 70 cos(u)3 − 42 cos(u)5 + 10 cos(u)7
(x) ( x )3 ( x )5 ( x )7
= C − 70 cos + 70 cos − 42 cos + 10 cos .
10 10 10 10

7. Making the changes of variable u = x/3, du = (1/3) dx first and w = sin(u), dw = cos(u) du later, we
have

∫ (x) (x) ∫
5 sin2 cos3 dx = 15 sin2 (u) cos3 (u) du
3 3
∫
= 15 sin2 (u) cos2 (u) cos (u) du
∫
( )
= 15 sin2 (u) 1 − sin2 (u) cos (u) du
∫
( )
= 15 w2 1 − w2 dw
∫
( )
= 15 w2 − w4 dw
( 3 )
w w5
= 15 − +C
3 5
= 5 sin3 (u) − 3 sin5 (u) + C
(x) (x)
= 5 sin3 − 3 sin5 + C.
3 3

8. Making the changes of variable u = 2x/3, du = (2/3) dx first and w = cos(u), dw = − sin(u) du later, we
135

have
∫ ( )√ ( ) ∫
3 2x 2x 21 √
7 sin cos dx = sin3 (u) cos(u) du
3 3 2
∫ √
21
= sin2 (u) cos(u) sin(u) du
2
∫
21 ( )√
= 1 − cos2 (u) cos(u) sin(u) du
2
∫
21 ( )√
= − 1 − w2 w dw
2
∫ ( )
21
= − w1/2 − w5/2 dw
2
( )
21 2 3/2 2 7/2
= C− w − w
2 3 7
( ) ( )
2x 2x
= C − 7 cos 3/2
+ 3 cos 7/2
.
3 3

9. We have
∫ ∫ ∫ ∫ ( )
cos3 (x) cos2 (x) 1 − sin2 (x) 1
dx = cos(x) dx = cos(x) dx = − sin(x) cos(x) dx.
sin(x) sin(x) sin(x) sin(x)

Make the substitution u = sin(x), du = cos(x) dx to obtain

∫
cos3 (x) 1
dx = C − sin2 (x) + ln (|sin (x)|) .
sin(x) 2

10. Write
( )
cos3 (2x) = cos2 (2x) cos (2x) = 1 − sin2 (2x) cos (2x) .
∫( ) √
The change of variable u = sin (2x) , du = 2 cos (2x) dx converts the integral to (5/2) 1 − u2 / u du,
√ √
or 5 u − u5/2 + C. After resubstituting u = sin (2x), we obtain 5 sin (2x) − sin5/2 (2x) + C.

11. Answer: 4/35

12. Write
( )
cos3 (x) = cos2 (x) cos (x) = 1 − sin2 (x) cos (x) .
∫1√ ( )
The change of variable u = sin (x) , du = cos (x) dx converts the integral to u 1 − u2 du, or
∫ 1 ( 1/2 ) 0

0
u − u5/2 du, or 8/21.

13. 12
Answer: 16/21

16. Answer: 3/8

17. To begin, we change cos2 (x) to 1 − sin2 (x), expand the integrand, and integrate term-by-term. This
results in two integrals: one with sin2 (x) as the integrand and one with sin4 (x) as the integrand. The
136

integral of sin2 (x) is handled as in Exercise 1. We will omit the details of that calculation. For the
integral of sin4 (x), we use the formula in Exercise 14. Here are some of the details:

∫ ∫
π/4 π/4 ( )
2 2
cos (t) sin (t) dt = 1 − sin2 (t) sin2 (t) dt
0 0
∫ π/4 ( )
= sin2 (t) − sin4 (t) dt
0
∫ π/4 ∫ π/4
= sin2 (t) dt − sin4 (t) dt
0 0
( ) ∫ π/4
π 1
= − − sin4 (t) dt
8 4 0
( ) ( π/4 )
π 1 1 3 3
= − − − sin3 (t) cos(t) − cos(t) sin(t) + t
8 4 4 8 8 0
( ) ( ( )4 ( )2 )
π 1 1 1 3 1 3 π
= − − − √ − √ +
8 4 4 2 8 2 8 4
π
= .
32

18. The first step is to make the substitution u = x/2, du = (1/2) dx. Then we use the reduction formula
of Exercise 15 with n = 6. Finally, we use the formula for the integral of cos(u)4 found in Exercise 15.
Here are the details:

∫ π ∫ π/2
cos6 (x/2) dx = 2 cos6 (u) du
0
(0 u=π/2 ∫ )
1 5 π/2
= 2 5
cos (u) sin(u) + 4
cos (u) du
6 u=0 6 0
∫ π/2
5
= cos4 (u) du
3
(0 π/2 )
5 1 3 3 3
= cos (u) sin(u) + cos(u) sin(u) + u
3 4 8 8 0
5 3 π
=
3 8 2
5π
= .
16

19. To begin, we change cos2 (x) to 1 − sin2 (x), expand the integrand, and integrate term-by-term. This
results in two integrals: one with sin4 (x) as the integrand and one with sin6 (x) as the integrand. We
work with the higher power, using the reduction formula in Exercise 14 with n = 6 (and u = x). After
combining terms, we use the formula for the integral of sin4 (x) developed in Exercise 14. (The alternative
would be to continue to use the reduction formula.) In the evaluations at the limits of integration, we
137
√
use cos(π/6) = sin(π/3) = 3/2 and sin(π/6) = cos(π/3) = 1/2. Here are the details:
∫ π/3 ∫ π/3
( )
64 cos2 (x) sin4 (x) dx = 64 1 − sin2 (x) sin4 (x) dx
π/6 π/6
(∫ π/3 ∫ π/3 )
= 64 sin (x) dx −
4 6
sin (x) dx
π/6 π/6
(∫ ( x=π/3 ∫ ))
π/3
1 5 π/3 4
= 64 sin (x) dx − − cos(x) sin (x)
4 5 + sin (x) dx
π/6 6 x=π/6 6 π/6
(∫ √ ∫ )
π/3
3 5 π/3 4
= 64 4
sin (x) dx + − sin (x) dx
π/6 48 6 π/6
( ∫ √ )
1 π/3 4 3
= 64 sin (x) dx +
6 π/6 48
∫ π/3 √
32 4 3
= sin4 (x) dx +
3 π/6 3
(( )x=π/3 ) √
32 1 3 3 4 3
= − sin (x) cos(x) − cos(x) sin(x) + x
3 +
3 4 8 8 x=π/6 3
( √ ) √
32 π 3 4 3
= − +
3 16 32 3
2π √
= + 3.
3
20. Letting u = tan(x), du = sec2 (x), we have
∫ ∫
tan3 (x) dx = tan(x) tan2 (x) dx
∫
( )
= tan(x) sec2 (x) − 1 dx
∫ ∫
= tan(x) sec2 (x) dx − tan(x) dx
∫ ∫
= u du − tan(x) dx
1 2
= u − ln( | sec(x)| ) + C
2
1
= tan2 (x) − ln( | sec(x)| ) + C.
2
21. We have,
∫ ∫
( 2 )
2
tan (x) sec(x) dx = sec (x) − 1 sec(x) dx
∫ ∫
= sec3 (x) dx − sec(x) dx
( )
1 1
= sec(x) tan(x) + ln ( | sec(x) + tan(x) | ) − ln ( | sec(x) + tan(x) | ) + C
2 2
1 1
= sec(x) tan(x) − ln ( | sec(x) + tan(x) | ) + C.
2 2
138

22. Making the change of variable u = sec(x), du = sec(x) tan(x) dx, we have,
∫ ∫
tan3 (x) sec(x) dx = tan2 (x) sec(x) tan(x) dx
∫
( 2 )
= sec (x) − 1 sec(x) tan(x) dx
∫
( 2 )
= u − 1 du
1 3
= u −u+C
3
1
= sec3 (x) − sec(x) + C.
3
Solutions: Trigonometric Substitution
(Corresponds to Stewart 7.3)

1. Make the change of variable x = sin(θ), dx = cos(θ) dθ. Then

∫ ∫
24 x2 sin2 (θ)
√ dx = 24 √ cos(θ) dθ
1 − x2 1 − sin2 (θ)
∫
sin2 (θ)
= 24 cos(θ) dθ
cos(θ)
∫
= 24 sin2 (θ) dθ
∫
1
= 24 (1 − cos(2 θ)) dθ
2
( )
1
= 12 θ − sin(2 θ)
2
( )
= 12 θ − sin(θ) cos(θ)
( √ )
= 12 arcsin (x) − x 1 − x2 .

Therefore,

 
∫ √ (√ ) √ v u ( √ )2 √
3/2 2
24 x  3 u
3t 3  π 3 1 √
√ dx = 12 arcsin − 1−  = 12 · − 12 · · = 4 π − 3 3.
0 1−x 2 2 2 2 3 2 2

139
140

2. Make the change of variable x = 5 sin(θ), dx = 5 cos(θ) dθ. Then

∫ √ ∫ √
2 25 − x dx = 2 × 25
2 1 − sin2 (θ) cos(θ) dθ
∫
= 2 × 25 cos2 (θ) dθ
∫
1
= 2 × 25 (1 + cos(2 θ)) dθ
2
( )
1
= 25 θ + sin(2 θ)
2
( )
= 25 θ + sin(θ) cos(θ)
( )
( x ) x √25 − x2
= 25 arcsin +
5 5 5
(x) √
= 25 arcsin + x 25 − x2 .
5
Therefore, ∫ 3 √ ( )
3
2 25 − x2 dx = 25 arcsin + 12.
0 5
3. Make the change of variable x = 2 sin(θ), dx = 2 cos(θ) dθ. Then
∫ ∫
( )3/2 ( )3/2
4 − x2 dx = 2 × 43/2 1 − sin2 (θ) cos(θ) dθ
∫
= 16 cos3 (θ) cos(θ) dθ
( )
1 3 3 3
= 16 cos (θ) sin(θ) + cos(θ) sin(θ) + θ + C
4 8 8
(√ )3 √
4 − x2 x 4 − x2 x (x)
= 4 +6 + 6 arcsin +C
2 2 2 2 2
1 ( )3/2 3 ( )1/2 (x)
= x 4 − x2 + x 4 − x2 + 6 arcsin + C.
4 2 2
( )
4. Make the change of variable x = tan(θ)/2, dx = sec2 (θ)/2 dθ. Then
∫ ∫
2 1
√ dx = √ sec2 (θ) dθ
1 + 4x2 1 + tan2 (θ)
∫
1
= sec2 (θ) dθ
sec(θ)
∫
= sec(θ) dθ

= ln ( | sec(θ) + tan(θ)| ) + C
( √ )
= ln | 1 + 4x2 + 2x| + C
(√ )
= ln 1 + 4x2 + 2x + C.

In
√ the last line, the absolute value function has been omitted because it is superfluous: the expression
1 + 4x2 + 2x cannot assume a negative value for real x.
141

5. Make the change of variable x = 3 tan(θ), dx = 3 sec2 (θ) dθ. Then, using the formula
∫
1
tan3 (θ) dθ = tan2 (x) − ln( | sec(x)| ) + C
2

from the last set of exercises, we have

∫ ∫
2x3 2 × 33 tan3 (θ)
dx = 3 sec2 (θ) dθ
9 + x2 9 + 9 tan2 (θ)
∫
2 × 3 tan3 (θ)
= 3 sec2 (θ) dθ
1 + tan2 (θ)
∫
tan3 (θ)
= 18 sec2 (θ) dθ
sec2 (θ)
∫
= 18 tan3 (θ) dθ

= 9 tan2 (θ) − 18 ln( | sec(θ)| ) + C

(√ )
9 + x 2
= x2 − 18 ln +C
3
(√ )
= x2 − 18 ln 9 + x2 + C
( )
= x2 − 9 ln 9 + x2 + C.

√
6. Let a = 3. Make the change of variable x = (a/2) tan(θ), dx = (a/2) sec2 (θ) dθ. We have
∫ ( a )3 ∫
16 x2 tan2 (θ)
√ dx = 16 √ sec2 (θ) dθ
3 + 4 x2 2 3 + 3 tan2 (θ)
( a )3 1 ∫ tan2 (θ)
= 16 √ sec2 (θ) dθ
2 a 1 + tan2 (θ)
( a )3 1 ∫ tan2 (θ)
= 16 sec2 (θ) dθ
2 a sec(θ)
∫
2
= 2a tan2 (θ) sec(θ) dθ
∫
( 2 )
= 6 sec (θ) − 1 sec(θ) dθ
∫ ∫
= 6 sec (θ) dθ − 6
3
sec(θ) dθ

= 3 sec(θ) tan(θ) + 3 ln ( | sec(θ) + tan(θ) | ) − 6 ln ( | sec(θ) + tan(θ) | ) + C

= 3 sec(θ) tan(θ) − 3 ln ( | sec(θ) + tan(θ) | ) + C
√ (√ )
a2 + 4x2 2x a2 + 4x2 2x
= 3 − 3 ln + +C
a a a a
√ (√ )
= 2 x a2 + 4x2 − 3 ln a2 + 4x2 + 2x + C
√ (√ )
= 2 x 3 + 4x2 − 3 ln 3 + 4x2 + 2x + C.
142

7. Make the change of variable x = (3/2) tan(θ), dx = (3/2) sec2 (θ) dθ. Then
∫ ∫
36 1 3 2
2 dx = 36 ( )2 sec (θ) dθ
2
(9 + 4 x ) 2
9 + 9 tan (θ)) 2 2
∫
36 × 3 1 2
= ( )2 sec (θ) dθ
9 ×2
2
1 + tan2 (θ)
∫
2 1
= sec2 (θ) dθ
3 sec4 (θ)
∫
2 1
= dθ
3 sec2 (θ)
∫
2
= cos2 (θ) dθ
3
∫
2 1
= (1 + cos(2θ)) dθ
3 2
( )
1
= θ + sin(θ) cos(θ) + C
3
( ( ) )
1 2x 2x 3
= arctan +√ √ +C
3 3 9 + 4 x2 9 + 4 x2
( )
1 2x 2x
= arctan + + C.
3 3 9 + 4 x2

The method of this problem can be used to derive the following general formula:
∫ ( )
2 1 x 1 bx
2 dx = + arctan + C.
(a2 + b2 x2 ) a2 (a2 + b2 x2 ) a3 b a

8. Make the change of variable x = 5 sec(θ), dx = 5 sec(θ) tan(θ) dθ and, to simplify the final answer,
observe that the addition of a specific constant such as − ln(5) to an arbitrary constant results in an
arbitrary constant:
∫ ∫
1 1
√ dx = √ 5 sec(θ) tan(θ) dθ
x2 − 25 25 sec2 (θ) − 25
∫
5 1
= √ √ sec(θ) tan(θ) dθ
25 sec (θ) − 1
2
∫
1
= √ sec(θ) tan(θ) dθ
tan2 (θ)
∫
= sec(θ) dθ

= ln( | sec(θ) + tan(θ) | ) + C

( )
x √x2 − 25

= ln + +C
5 5
( √ )
x + x2 − 25

= ln +C
5
( √ )

= ln x + x2 − 25 + C − ln(5)
( √ )

= ln x + x2 − 25 + C.
143

9. Make the change of variable x = (2/3) sec(θ), dx = (2/3) sec(θ) tan(θ) dθ and use the integral formula
of Exercise 21 of the preceding section:
∫ √ ∫ √
2
9x − 4 dx =
2 4 sec2 (θ) − 4 sec(θ) tan(θ) dθ
3
∫ √
4
= sec2 (θ) − 1 sec(θ) tan(θ) dθ
3
∫ √
4
= tan2 (θ) sec(θ) tan(θ) dθ
3
∫
4
= sec(θ) tan2 (θ) dθ
3
( )
4 1
= · sec(θ) tan(θ) − ln ( | sec(θ) + tan(θ) | ) + C
3 2
( √ ( ))
2 3x 9x2 − 4 3x √9x2 − 4

= − ln + +C
3 2 2 2 2
( )
1 √ 2 3x + √9x2 − 4
2
= x 9x − 4 − ln +C
2 3 2
√ ( √ ) 2
1 2
= x 9x2 − 4 − ln 3x + 9x2 − 4 + ln(2) + C
2 3 3
1 √ 2 2 ( √ )

= x 9x − 4 − ln 3x + 9x2 − 4 + C.
2 3

10. We complete the square before obtaining the expression 1 − stuff2 that can be reduced by the trigono-
metric substitution stuff = sin(θ). The expression we will identify as “stuff”, namely (x − 2)/2, finalizes
two lines above the last line of the following chain of equalities:
( )
4 x − x2 = 0 + −x2 + 4 x
( )
= 0 − x2 − 4 x
( )2 ( ( )2 )
4 4
= 0+ − x − 4x +
2
2 2
( 2 )
= 4 − x − 4x + 4
2
= 4 − (x − 2)
( )
1 2
= 4 1 − (x − 2)
4
( )
1 2
= 4 1 − (x − 2)
4
( ( )2 )
x−2
= 4 1−
2
( )
= 4 1 − sin2 (θ)
= 4 cos2 (θ),

where (x − 2)/2 = sin(θ), or x = 2 sin(θ) + 2, dx = 2 cos(θ) dθ. Thus,

∫ ∫ ∫ ( )
1 1 x−2
√ dx = √ 2 cos(θ) dθ = 1 dθ = θ + C = arcsin + C.
4 x − x2 4 cos2 (θ) 2
144

11. We complete the square before obtaining the expression 1 + stuff2 that can be reduced by the trigono-
metric substitution stuff = tan(θ). The expression we will identify as “stuff”, namely (x + 4)/2, finalizes
two lines above the last line of the following chain of equalities:

( )
x2 + 8 x + 20 = x2 + 8 x + 20
( ( )2 ) ( )2
8 8
= 2
x + 8x + + 20 −
2 2
( 2 )
= x + 8 x + 16 + 4
2
= (x + 4) + 4
( )
1 2
= 4 (x + 4) + 1
4
(( )2 )
x+4
= 4 +1
2
( )
= 4 tan2 (θ) + 1
= 4 sec2 (θ),

where (x + 4)/2 = tan(θ), or x = 2 (tan(θ) − 4, 2

) dx = 2 sec (θ) dθ. We also note that the equation
(
( −4 we have
(x +)4)/2 = tan(θ) leads to θ = arctan (x + 4)/2 . Corresponding to x = ) θ = arctan (−4 +
4)/2 = arctan(0) = 0 and corresponding to x = −24 we have θ = arctan (−2+4)/2 = arctan(1) = π/4.
Thus,

∫ ∫ ∫ π/4
−2
1 π/4
1 1 π/4
1 π
dx = 2 sec2 (θ) dθ = 1 dθ = θ = .
−4 x2 + 8 x + 20 0 4 sec2 (θ) 2 0 2 0 8

12. We complete the square before obtaining the expression stuff2 − 1 that can be reduced by the trigono-
metric substitution stuff = sec(θ). The expression we will identify as “stuff”, namely 5 x + 2, finalizes
two lines above the last line of the following chain of equalities:

( )
2 2 4
25 x + 20 x + 3 = 25 x + x +3
5
( ( )2 ) ( )2
4 1 4 1 4
= 25 2
x + x+ · + 3 − 25 · +3
5 2 5 2 5
( ( )2 )
4 2
= 25 2
x + x+ +3−4
5 5
( )2
2
= 25 x+ −1
5
2
= (5 x + 2) − 1
= sec2 (θ) − 1
= tan2 (θ),
145

1 2 1
where 5 x + 2 = sec(θ), or x = sec(θ) − , dx = sec(θ) tan(θ) dθ. Thus,
5 5 5
∫ ∫ ( )2
x2 1 2 1 1
3/2
dx = sec(θ) − 3 sec(θ) tan(θ) dθ
2
(25 x + 20 x + 3) 5 5 tan (θ) 5
∫
1 2 sec(θ)
= (sec(θ) − 2) dθ
125 tan2 (θ)
∫
1 ( 2 ) sec(θ)
= sec (θ) − 4 sec(θ) + 4 dθ
125 tan2 (θ)
∫
1 sec3 (θ) − 4 sec2 (θ) + 4 sec(θ)
= dθ.
125 tan2 (θ)

No more was asked, but, for the curious, here is the evaluation of the integral:
∫ ( ( √ ) )
x2 1 2 + 20 x + 3 − √
25 x + 6
3/2
dx = ln 5 x + 2 + 25 x .
(25 x2 + 20 x + 3) 125 25 x2 + 20 x + 3

This evaluation can be accomplished using methods of integration that are covered in Calculus II, but
the task is somewhat tricky.
146
Solutions: Using Partial Fractions to
Integrate Rational Functions
(Corresponds to Stewart 7.4)

1. The partial fraction expansion is

x − 11 4 3
= − ,
x2 − x − 2 x+1 x−2
which leads to the evaluation
∫
x − 11
dx = 4 ln( | x + 1 | ) − 3 ln( | x − 2 | ).
x2−x−2
2. The partial fraction expansion is
2 1 2 1
= − + ,
x3 −x x+1 x x−1
which leads to the evaluation
∫
2
dx = ln( | x + 1 | ) − 2 ln( | x | ) + ln( | x − 1 | ).
x3 − x
3. The partial fraction expansion is
4 x2 + 3 x + 2 1 2 3
3 2
= + 2+ ,
x +x x x x+1
which leads to the evaluation
∫
4 x2 + 3 x + 2 2
3 2
dx = ln( | x | ) − + 3 ln( | x + 1 | ).
x +x x
4. The partial fraction expansion is
6 x2 + 17 x − 8 1 5x − 3
= + 2 ,
(x + 4)(x2 + 4) x+4 x +4
which leads to the evaluation
∫ ∫ ∫
6 x2 + 17 x − 8 1 5x − 3
dx = dx + dx
(x + 4)(x2 + 4) x+4 x2 + 4
∫ ∫
5 2x 1
= ln( | x + 4 | ) + dx − 3 dx
2 x2 + 4 x2 + 4
5 ( ) 3 (x)
= ln( | x + 4 | ) + ln x2 + 4 − arctan + C.
2 2 2

147
148

5. The partial fraction expansion is

x3 + 5 x2 − x + 30 2−x 2x + 3
= 2 + 2 ,
(x2 + 4) (x2 + 9) x +4 x +9

which leads to the evaluation

∫ 3 (x) (x)
x + 5 x2 − x + 30 1 ( 2 ) ( 2 )
dx = − ln x + 4 + arctan + ln x + 9 + arctan + C.
(x2 + 4) (x2 + 9) 2 2 3

6. The partial fraction expansion is

2 x3 + 5 x − 1 2x 3x − 1
2 = + ,
(x2 + 1) x2 + 1 (x2 + 1)2

which leads to the evaluation

∫
2 x3 + 5 x − 1 ( ) 1 x+3 1
2 dx = ln x2 + 1 − 2
− arctan(x) + C.
(x2 + 1) 2 x +1 2

7. The partial fraction expansion is

5 x4 + 6 x3 + 30 x2 + 45 5 6x
2 = + 2,
x2 (x2 + 3) x2 (x2 + 3)

which leads to the evaluation

∫
5 x4 + 6 x3 + 30 x2 + 45 5 3
2 dx = − − 2 + C.
x2 (x2 + 3) x x +3

8. Completing the square, we have

( ( )2 ) ( )2
2
( 2
) 2 6 6 ( ) 2
x +6 x+18 = x + 6 x +18 = x + 6 x + +18− = x2 + 6 x + 9 +18−9 = (x + 3) +9,
2 2

which becomes u2 + a2 by letting a = 3 and making the change of variable u = x + 3, du = dx. Thus,
∫ ∫ (u) (x )
6 1 6
dx = 6 du = arctan = 2 arctan + 1 .
x2 + 6 x + 18 a2 + u 2 a a 3
Solutions: Numerical Integration
(Corresponds to Stewart’s Section 7.7)

1. Let f (x) = 6/x. For N = 2, we have ∆x = (5 − 1)/2 = 2, and the Trapezoidal, Midpoint, and Simpson’s
approximations, T2 , M2 , S2 , are

∆x
T2 = (f (1) + 2 f (3) + f (5))
2
56
= ,
5
M2 = ∆x (f (2) + f (4))
= 9,
∆x
S2 = (f (1) + 4 f (3) + f (5))
3
152
= .
15

For N = 4, we have ∆x = (5 − 1)/4 = 1, and the Trapezoidal, Midpoint, and Simpson’s approximations,
T4 , M4 , S4 , are

∆x
T4 = (f (1) + 2 f (2) + 2 f (3) + 2 f (4) + f (5))
2
101
= ,
10
M4 = ∆x (f (3/2) + f (5/2) + f (7/2) + f (9/2))
993
= ,
105
∆x
S4 = (f (1) + 4 f (2) + 2 f (3) + 4 f (4) + f (5))
3
146
= .
15

For N = 6, we have ∆x = (5 − 1)/6 = 2/3, and the Trapezoidal, Midpoint, and Simpson’s approxima-

149
150

tions, T6 , M6 , S6 , are
∆x
T6 = (f (1) + 2 f (1 + ∆) + 2 f (1 + 2 ∆x) + 2 f (1 + 3 ∆x) + 2 f (1 + 4 ∆x) + 2 f (1 + 5 ∆x) + f (5))
2( ( ) ( ) ( ) ( ) )
1 5 7 11 13
= f (1) + 2 f + 2f + 2 f (3) + 2 f + 2f + f (5)
3 3 3 3 3
148072
= ,
15015
M6 = ∆x (f (1 + ∆/2) + f (1 + 3 ∆/2) + f (1 + 5 ∆/2) + f (1 + 7 ∆/2) + f (1 + 9 ∆/2) + f (1 + 11 ∆/2))
669
= ,
70
∆x
S6 = (f (1) + 4 f (1 + ∆x) + 2 f (1 + 2 ∆x) + 4 f (1 + 3 ∆x) + 2 f (1 + 4 ∆x) + 4 f (1 + 5 ∆x) + f (5))
3( ( ) ( ) ( ) ( ) )
2 5 7 11 13
= f (1) + 4 f + 2f + 4 f (3) + 2 f + 4f + f (5)
9 3 3 3 3
435976
= .
45045

For the numerical evaluations of these 9 rational numbers, see the following table. For reference, the
exact value of the given integral is 6 ln(5) = 9.6566... .

N 2 4 6
Trapezoidal Rule 56/5, or 11.2 10.1 9.862
Midpoint Rule 9 9.448 9.557
Simpson’s Rule 152/15, or 10.133 9.733 9.679

2. The values of L needed are L(0) = 0, L(20) = 5.13, L(40) = L(20)+9.45 = 14.58, L(60) = L(40)+13.85 =
28.43, L(80) = L(60) + 20.33 = 48.76, L(100) = 100. Because ∆x = (100 − 0)/5 = 20, we have
∫ 100 (
1 )
γ = x − L(x) dx
5000 0
∫ 100 ∫ 100
1 1
= x dx − L(x) dx
5000 0 5000 0
100 ∫
1
= 1− L(x) dx
5000 0
1 20 ( )
≈ 1− · 0 + 2 × 5.13 + 2 × 14.58 + 2 × 28.43 + 2 × 48.76 + 100
5000 2
= 0.4124.

The World Bank gives 0.411 for the 2010 Gini coefficient. Source: https://en.wikipedia.org/wiki/
List_of_countries_by_income_equality
3. Because ∆x = 10, we have, using the Trapezoidal Rule,
1 10
γ ≈ 1− · (0 +2 × 3 + 2 × 7 + 2 × 12 + 2 × 18 + 2 × 24 + 2 × 32 + 2 × 41 + 2 × 54 + 2 × 73 + 100) = 0.372
5000 2
and, using Simpson’s Rule,
1 10
γ ≈ 1− · (0+4×3+2×7+4×12+2×18+4×24+2×32+4×41+2×54+4×73+100) = 0.3773.
5000 3
151

4. First,
∫ 9
1.5
c(t) dt ≈ (0 + 4 × 2.4 + 2 × 6.3 + 4 × 9.7 + 2 × 7.1 + 4 × 2.3 + 0)
0 3
mg s
= 42.2
L
42.2 mg min
=
60 L
mg min
= 0.7033 .
L

Therefore,
5 mg L L
r= = 7.109 .
0.7033 mg min min
5. The cross-sectional area A is approximately given by
2
A≈ (0.77 + 4 × 1.5 + 2 × 2.0 + 4 × 2.4 + 2 × 2.8 + 4 × 3.1 + 2 × 3.4 + 4 × 3.6 + 3.9) = 42.3133 m2 .
3
The volume V is therefore approximately given by V ≈ 6 m × 42.3133 m2 = 253.88 m3 .

6. The distance s (in meters) that the athlete ran was

∫ 5
s = s′ (t) dt
(
0
1/2
≈ 0 + 4 × 5.26 + 2 × 6.67 + 4 × 7.41 + 2 × 8.33
3
)
+4 × 8.33 + 2 × 9.52 + 4 × 9.52 + 2 × 10.64 + 4 × 10.64 + 10.87
= 40.97.

7. The nodes are x0 = 0, x1 = 1, x2 = 2, x3 = 3, x4 = 4, x5 = 5, , x6 = 6. Thus, N = 6 and ∆ x =

(6 − 0)/N = 1. The heights of the vertical line segments between the curves and over the nodes are
y0 = 0, y1 = 5, y2 = 3, y3 = 4, y4 = 6, y5 = 3, y6 = 0. The trapezoidal Rule approximation is
∆x
T6 = (1 × y0 + 2 × y1 + 2 × y2 + 2 × y3 + 2 × y4 + 2 × y5 + 1 × y6 )
2
1
= (1 × 0 + 2 × 5 + 2 × 3 + 2 × 4 + 2 × 6 + 2 × 3 + 1 × 0)
2
= 21.

and the Simpson Rule approximation is

∆x
S6 = (1 × y0 + 4 × y1 + 2 × y2 + 4 × y3 + 2 × y4 + 4 × y5 + 1 × y6 )
3
1
= (1 × 0 + 4 × 5 + 2 × 3 + 4 × 4 + 2 × 6 + 4 × 3 + 1 × 0)
3
= 22.
152
Solutions: Improper Integrals
(Corresponds to Stewart 7.8)

1. The following calculation both shows that the given improper integral is convergent and evaluates it:
∫ ∫ ( )4−ϵ ( )
4
−0.9
4−ϵ
−0.9 (4 − x)0.1 (2)0.1 (ϵ)0.1
(4−x) dx = lim+ (4−x) dx = lim+ − = lim+ − + = 10·20.1 .
2 ϵ→0 2 ϵ→0 0.1 2 ϵ→0 0.1 0.1

2. The following calculation both shows that the given improper integral is convergent and evaluates it:
∫ ∫ ( )2−ϵ ( )
2
−1/5
2−ϵ
−1/5 (x − 2)4/5 5 × ϵ4/5 5 × 14/5 5
(x−2) dx = lim+ (x−2) dx = lim+ = lim+ − =− .
1 ϵ→0 1 ϵ→0 4/5 1 ϵ→0 4 4 4

3. The following calculation both shows that the given improper integral is convergent and evaluates it:
∫ ∫ 3
3
1 3
1 √ √
√ dx = lim+ √ dx = lim+ 2 x + 1 = lim+ 4 − ϵ = 4.
−1 x+1 ϵ→0 −1+ϵ x+1 ϵ→0 −1+ϵ ϵ→0

4. The following calculation both shows that the given improper integral is convergent and evaluates it:
∫ ∫ ( ) ( )
1
x 1−ϵ
x 2( )3/4 1−ϵ 2( ) 2 2
dx = lim+ dx = lim+ − 1 − x2 = lim − 2ϵ − ϵ 2 3/4
+ = .
1/4 1/4
0 (1 − x2 ) ϵ→0 0 (1 − x2 ) ϵ→0 3 0 ϵ→0 + 3 3 3

5. The following calculation both shows that the given improper integral is convergent and evaluates it:
∫ ∫
1
1 1−ϵ
1 1−ϵ π
√ dx = lim+ √ dx = lim+ arcsin(x) 0 = lim+ arcsin(1−ϵ) = arcsin(1) = .
0 1 − x2 ϵ→0 0 1−x 2 ϵ→0 ϵ→0 2

6. Expand the integrand:

∫ 3 ∫ 3 ∫ 3
−1/2
x (1 + x) dx = x1/2
dx + lim+ x−1/2 dx
0 0 ϵ→0 ϵ
√ 3

= 2 3 + lim+ 2 x1/2
ϵ→0 ϵ
√ √ √
= 2 3+2 3−2 ϵ
√
= 4 3.

153
154

7. Make the change of variable u = ln(x), du = (1/x) dx and observe that the limits of integration for u
are from u = ln(1) = 0 to u = ln(e) = 1:

∫ e ∫ 1
1
dx = u−1/3 du
1 x · ln1/3 (x) 0
∫ 1
= lim+ u−1/3 du
ϵ→0 ϵ
1
3 2/3
= lim u
ϵ→0+ 2 ϵ
3 3 2/3
= − ϵ
2 2
3
= .
2

∫
8. We use the integral formula ln (x) dx = x ln(x) − x + C, which was obtained in the section on inte-
gration by parts. In the third line of the calculation that follows, the indeterminate form 0 × ∞ arises.
A change of variable is made to convert the indeterminate form to the form ∞/∞, which we evaluate
by applying L’Hôpital’s Rule:

∫ 1 ∫ 1
ln (x) dx = lim+ ln (x) dx
0 ϵ→0 ϵ
1

= lim+ (x ln(x) − x)
ϵ→0 ϵ
= (1 × ln(1) − 1) − lim (ϵ × ln(ϵ) − ϵ)
ϵ→0+
= −1 − lim+ ϵ × ln(ϵ)
ϵ→0
ln(ϵ)
= −1 − lim+
ϵ→01/ϵ
ln(1/ϵ)
= −1 + lim+
ϵ→0 1/ϵ
ln(t)
= −1 + lim
t→+∞ t
d
ln(t)
= −1 + lim dt d
dt t
t→+∞

1/t
= −1 + lim
t→+∞ 1
= −1.
155

9. The solution uses a basic integration formula involving the arcsecant as the antiderivative:
∫ √ ∫ √
2 2
1 1
√ dx = lim+ √ dx
1 x x2 − 1 ϵ→0 1+ϵ x x − 1
2
√2

= lim+ arcsec (x)
ϵ→0 1+ϵ
( (√ ) )
= lim+ arcsec 2 − arcsec (1 + ϵ)
ϵ→0
π
= −0
4
π
= .
4

10. The singularity is at x = −1 in the interior of the interval of integration, which means that the given
integral must be decomposed into two improper integrals. Both are of types already treated in detail,
so no further detail will be given here:
∫ 4 ∫ −1 ∫ 4 ( )
(x + 1)−2/3 dx = (x + 1)−2/3 dx + (x + 1)−2/3 dx = 3 + 3 · 51/3 = 3 51/3 + 1 .
−2 −2 −1

11. The integrand is very similar to the one in Exercise 7, and it is handled in exactly the same way. The
current exercise does have, however, an additional complication: the simgularity is at x = 1, which is
inside the interval of integration. So the given integral must be decomposed into two improper integrals,
as was done in the preceding exercise. We have
∫ e ∫ 1 ∫ e
1 1 1 7 7 14
2/7
dx = 2/7
dx + 2/7
dx = + = .
1/e x ln (x) 1/e x ln (x) 1 x ln (x) 5 5 5

12. Issues at both limits of integration must be treated:

∫ 2 ∫ 0 ∫ 2
1 1 1
√ dx = √ dx + √ dx
−2 4−x 2
−2 4−x 2
0 4 − x2
∫ 0 ∫ 2−ϵ
1 1
= lim+ √ dx + lim+ √ dx
ϵ→0 −2+ϵ 4 − x2 ϵ→0 0 4 − x2
( x )0 ( x )2−ϵ

= lim+ arcsin + lim+ arcsin
ϵ→0 2 −2+ϵ ϵ→0 2 0
( ϵ ) ( ϵ)
= − lim+ arcsin −1 + + lim+ arcsin 1 −
ϵ→0 2 ϵ→0 2
= − arcsin(−1) + arcsin(1)
( π) π
= − − +
2 2
= π.

13. Integrating by parts twice yields the integral formula

∫
( )
ln2 (x) dx = x ln2 (x) − 2 ln(x) + 2 + C.

In addition to this integral formula, we will need two limit formulas. One of them,

lim ϵ ln(ϵ) = 0
ϵ→0+
156

was derived in Exercise 7. The other is obtained in a similar way:

ln2 (ϵ)
lim+ ϵ ln2 (ϵ) = lim+
ϵ→0 ϵ→0 1/ϵ
2
(− ln(1/ϵ))
= lim+
ϵ→0 1/ϵ
2
(ln(t))
= lim
t→+∞ t
d 2
(ln(t))
= lim dt d
dt t
t→+∞

ln(t)
= 2 lim
t→+∞ t
d
ln(t)
= 2 lim dt d
dt t
t→+∞

1
= 2 lim
t→+∞ t
= 0.

Using these three equations, we have

∫ 1 ∫ 1
ln2 (x) dx = lim ln2 (x) dx
0 ϵ→0+ ϵ
( ( 2 )) 1
= x ln (x) − 2 ln(x) + 2
lim
ϵ→0+ ϵ
( 2 )
= 2 − lim+ ϵ ln (ϵ) − 2 ln(ϵ) + 2
ϵ→0
= 2.

14. Integrating by parts, we have

∫ ∫
−1/2 1
x ln(x) dx = 2 x 1/2
ln(x) − 2 x−1/2 dx = 2 x1/2 ln(x) − 4 x1/2 + C = 2 x1/2 (ln(x) − 2).
x

Additionally, we will need one limit formula, which we obtain from a limit formula already derived
( )
lim+ ϵ1/2 ln(ϵ) = 2 lim+ ϵ1/2 ln ϵ1/2 = 2 lim+ u ln (u) = 0.
ϵ→0 ϵ→0 u→0

Therefore
∫ 4 ∫ 4
−1/2
x ln(x) dx = lim+ x−1/2 ln(x) dx
0 ϵ→0 ϵ
4

= lim+ 2 x1/2 (ln(x) − 2)
ϵ→0 ϵ

= 24 1/2
(ln(4) − 2) − lim+ 2 ϵ1/2 (ln(ϵ) − 2)
ϵ→0

= 4 (2 ln(2) − 2) − 2 lim+ ϵ1/2 ln(ϵ)

ϵ→0
= 8 (ln(2) − 1).
157

15. Issues at both limits of integration must be treated. The method is to expand the logarithm. Doing so
leads to two similar integrations for which we use the formula
∫ ( )
1
ln(1 + a x) dx = x + ln(1 + a x) − x + C
a

with a = −1 and a = 1. The cited formula may be obtained by integrating by parts with u = ln(1 + a x)
and dv = dx. For a = −1 it yields
∫
ln(1 − x) dx = (x − 1) ln(1 − x) − x + C

an for a = 1 it yields ∫
ln(1 + x) dx = (x + 1) ln(1 + x) − x + C.

The following evaluation makes use of these integral formulas, as well as the limit formula limϵ→0+ ϵ ln(ϵ) =
0, which was derived in Exercise 8:

∫ ∫
1 ( ) 1 ( )
ln 1 − x 2
dx = ln (1 − x)(1 + x) dx
−1 −1
∫ 1 ∫ 1
= ln(1 − x) dx + ln(1 + x) dx
−1 −1
∫ 1−ϵ ∫ 1
= lim+ ln(1 − x) dx + lim+ ln(1 + x) dx
ϵ→0 −1 ϵ→0 −1+ϵ
( ) 1−ϵ ( ) 1
= lim+ (x − 1) ln(1 − x) − x + lim+ (x + 1) ln(1 + x) − x
ϵ→0 −1 ϵ→0 −1+ϵ
= lim+ (−ϵ ln(ϵ) − 1 + ϵ) − (−2 ln(2) + 1) + 2 ln(2) − 1 − lim+ (ϵ ln(ϵ) + 1 − ϵ)
ϵ→0 ϵ→0
= 4 (ln(2) − 1) .

16. We have
∫ ∞ ∫ N N
−3/2 1
x dx = lim x−3/2 dx = lim − 2 x−1/2 = 1 − 2 lim = 1.
4 N →∞ 4 N →∞ 4 N →∞ N 1/2

17. We have
∫ ∞ ∫ N N
1 1 π
dx = lim dx = lim arctan(x) = lim arctan(N ) = .
0 1 + x2 N →∞ 0 1 + x2 N →∞ 0 N →∞ 2

18. We have
∫ ∞ ∫ ( )
x N
x 1 ( )−1 N 1 1 1
2 dx = lim 2 dx = − lim 1 + x2 = 1 − lim 2
= .
0 (1 + x2 ) N →∞ 0 (1 + x2 ) 2 N →∞ 0 2 N →∞ 1 + N 2

19. We have
∫ ∞ ∫ N N ( ) 2

e−x/2 dx = lim e−x/2 dx = −2 lim e−x/2 = 2 e−1 − lim e−N/2 = .
2 N →∞ 2 N →∞ 2 N →∞ e
158

20. A partial fraction decomposition, the details of which are omitted, is key. We have
∫ ∞ ∫ N
1 1
dx = lim dx
2 x (x − 1) 2 N →∞
x (x − 1)
∫ N( )
1 1
= lim − dx
N →∞ 2 x−1 x
N

= lim (ln( |x − 1| ) − ln( |x| ))
N →∞ 2
( )N
|x − 1|
= lim ln
N →∞ |x|
( ) 2 ( )
N −1 1
= lim ln − ln
N →∞ N 2
( )
1
= lim ln 1 − + ln (2)
N →∞ N
= ln (1) + ln (2)
= ln (2) .

21. Begin with the substitution u = ex , du = ex dx. The limits of integration for u are u = e0 = 1 and
u = e∞ = ∞. It follows that
∫ ∞ ∫ ∞
ex 1
2x + 1
dx = 2+1
du
0 e 1 u
∫ N
1
= lim 2
du
N →∞ 1 u +1
N

= lim arctan(u)
N →∞ 1
= lim arctan(N ) − arctan(1)
N →∞
π π
= −
2 4
π
= .
4

22. Begin with the substitution u = ex , du = ex dx. The limits of integration for u are u = e0 = 1 and
u = e∞ = ∞. It follows that
∫ ∞ ∫ ∞
ex 1
3 dx = du
0 (ex + 1) 1 (u + 1)3
∫ N
1
= lim du
N →∞ 1 (u + 1)3
N
1
= − lim (u + 1)−2
2( N →∞ 1
)
1 1 1
= − lim
2 4 N →∞ (N + 1)2
1 1
= · −0
2 4
1
= .
8
159

23. A partial fraction decomposition, the details of which are omitted, is key. We have
∫ ∞( )2 ∫ N( )2
x+1 x+1
dx = lim dx
0 x2 + 1 N →∞ 0 x2 + 1
∫ N ( )
1 2x
= lim + dx
N →∞ 0 x2 + 1 (x2 + 1)2
( )N
1
= lim arctan(x) − 2
N →∞ x + 1 0
( ) ( )
1 1
= lim arctan(N ) − 2 − arctan(0) − 2
N →∞ N +1 0 +1
π
= + 1.
2

24. Begin with the substitution u = 3x2 , du = 6x dx. The limits of integration for u are u = 3 and u = ∞.
It follows that
∫ ∞ ∫
1 ∞ −u
x e−3x dx =
2
e du
1 6 3
∫ N
1
= lim e−u du
6 N →∞ 3
N
1
= − lim e−u
6 N →∞ 3
1 ( −3 )
= e − lim e−N
6 N →∞
1
= .
6e3

25. Begin with the substitution u = ln(x), du = (1/x) dx. The limits of integration for u are u = 1 and
u = ∞. It follows that
∫ ∞ ∫ ∞
1 1
2 dx = 2
du
e x ln (x) 1 u
∫ N
1
= lim du
N →∞ 1 u2
N
1
= − lim
N →∞ u
1
1
= 1 − lim
N →∞ N
= 1.

26. We have
∫ ∞ ( )x ∫ N ( )x ( )x N ( ( ))−1
2 2 1 2 1 3
dx = lim dx = lim = − = ln .
0 3 N →∞ 0 3 ln (2/3) N →∞ 3 ln (2/3) 2
0

27. We have ∫ ∫
−2 −2 −2
1 1
x−3 dx = lim x−3 dx = − lim x−2 =− .
−∞ N →∞ −N 2 N →∞ −N 8
160

28. We have
∫ −2 ∫ −2 −2
1 1 −1/3
4/3
dx = lim 4/3
dx = −3 lim (1 + x) = 3.
−∞ (1 + x) N →∞ −N (1 + x) N →∞ −N

29. We have ∫ ∫ 1
1
x 1
x 1 1 1
2 dx = lim 2 dx = − lim 2 =− .
−∞ (1 + x2 ) N →∞ −N (1 + x2 ) 2 N →∞ 1 + x −N 4

30. We integrate by parts to obtain

∫
x exp (x/2) dx = 2 (x − 2) exp(x/2) + C.

We use this antiderivative formula to evaluate the given improper integral. To obtain the last line of
the calculation, we apply L’Hôpital’s Rule to show that limN →∞ (−N − 2) exp(−N/2) = 0:
∫ 0 ∫ 0 0

x exp (x/2) dx = lim x exp (x/2) dx = lim 2 (x − 2) exp(x/2) = −4.
−∞ N →∞ −N N →∞ −N

31. We have ∫ ∞
1
∫ N
1 1 ( x )N
dx = lim dx = lim arctan = π.
0 4 + x2 N →∞ 0 4 + x2 N →∞ 2 2 0 4
A similar calculation provides the same value for the integration over the interval (−∞, 0]. Thus,
∫ ∞ ∫ 0 ∫ ∞
1 1 1 π π π
dx = dx + dx = + = .
−∞ 4 + x2 −∞ 4 + x2 0 4+x 2 4 4 2
∫ ∞
1
Evaluate dx.
−∞ x2 + 2x + 10
We will need the basic integral formula
∫ (u)
1 1
du = arctan + C.
u2 + a2 a a
With u = x + 1 and a = 3, it follows that
∫ ∫ ( )
1 1 1 x+1
dx = dx = arctan + C.
x2 + 2x + 10 (x + 1)2 + 32 3 3

Thus,
∫ ∞ ∫ N
1 1
2
dx = lim dx
0 x + 2x + 10 N →∞ 0 x2
+ 2x + 10
( )N
1 x + 1
= lim arctan
N →∞ 3 3
( )0 ( )
1 N +1 1
= lim arctan − arctan
N →∞ 3 3 3
( )
π 1
= − arctan .
6 3
161

Similarly,
∫ 0 ∫ 0
1 1
2
dx = lim dx
−∞ x + 2x + 10 N →∞ −N + 2x + 10 x2
( )0
1 x + 1
= lim arctan
N →∞ 3 3 −N
( ) ( )
1 1 −N + 1
= arctan − lim arctan
3 N →∞ 3 3
( ) ( )
1 −π
= arctan − .
3 6

It follows that
∫ ∞ ( ) ( ) ( )
1 π 1 1 −π π
dx = − arctan + arctan − = .
−∞ x2 + 2x + 10 6 3 3 6 3

32. The integrand is a rational fraction that is already in partial fraction form. We expand the numerator
and split the interval of integration—the result is a sum of four improper integrals:
∫ ∞ ∫ 0 ∫ 0 ∫ ∞ ∫ ∞
x+2 x 2 x 2
2 dx = 2 dx + 2 dx + 2 dx + 2 dx.
−∞ (x2 + 1) −∞ (x2 + 1) −∞ (x2 + 1) 0 (x2 + 1) 0 (x2 + 1)

We will refer to the four improper integrals on the right side of this equation as, from left to right,
I1 , I2 , I3 , and I4 . We have
∫ 0
x
I1 = lim dx
N →∞ (x 2 + 1)2
−N
0
1 1
= − lim 2
2 N →∞ x + 1
−N
( )
1 1
= − lim 1 − 2
2 N →∞ N +1
1
= − .
2
From this calculation we deduce that I3 is also convergent and equal to −I1 , or 1/2. Thus, I1 + I2 = 0.
Next, using the general formula that concludes the solution of Exercise 7 from the set on Trigonometric
Substitution, we have
∫ 0
2
I2 = lim 2 dx
N →∞ −N (x2 + 1)
( )0
x
= lim 2
+ arctan(x)
N →∞ 1+x −N
( )
−N
= 0 − lim + arctan(−N )
N →∞ 1 + N 2
( ( π ))
= 0− 0+ −
2
π
= .
2
162

From this calculation we deduce that I4 is also convergent and equal to I2 , or π/2. Thus,
∫ ∞
x+2 1 π 1 π
2 2 dx = I1 + I2 + I3 + I4 = − 2 + 2 + 2 + 2 = π.
−∞ (x + 1)
Solutions: Arc Length (Corresponds
to Stewart 8.1)

1. Let f (x) = x2 /8 − ln(x), we have

( )2 ( )2
d x 1
1+ f (x) = 1+ −
dx 4 x
( ( )2 )
( x )2 x 1 1
= 1+ −2 +
4 4 x x
( )
( x )2 1 ( 1 )2
= 1+ − +
4 2 x
( ( )2 )
( x )2 1 1
= + +
4 2 x
( )2
x 1
= + .
4 x

The requested arc length L is therefore given by

√ ( )2 √
∫ 2 ∫ 2 ( )2 ∫ 2( )
d x 1 x 1 3
L= 1+ f (x) dx = + dx = + dx = + ln(2).
1 dx 1 4 x 1 4 x 8

2. i) We have
( )( )
d 3 d p ( )( )
x x = 3x2 p xp−1 = 3 p xp+1 ,
dx dx

which is a constant only if p + 1 = 0, or p = −1.

ii) Let f (x) = a x3 + b x−1 . Then

( )2 (
d ( )2 ( )( ) ( )2 )
f (x) = 3 a x2 + 2 3 a x2 −b x−2 + −b x−2 .
dx

( )( ) 1
The cross term 2 3 a x2 −b x−2 simplifies to −6 a b, and for this to equal -1/2 we must have b = .
12 a

163
164

iii) Let a = 1 and b = 1/12. Then f (x) = x3 + 1/(12 x). Then

( )2 ( )2
d 1
1+ f (x) = 1 + 3x − 2
dx 12 x2
( )2
( )2 1 1
= 1 + 3x2 − +
2 12 x2
( )2
( 2 )2 1 1
= 3x + +
2 12 x2
( )2
1
= 3x2 + .
12 x2
The requested arc length is
√ ( )2 √
∫ 2 ∫ 2 ( )2 ∫ 2( )
d 1 2 1 169
1+ f (x) dx = 3x2 + 2
dx = 3x + 2
dx = .
1 dx 1 12 x 1 12 x 24
∫1 √
3. The arc length is 1/2 1 + 4x2 dx. With n = 4, we have ∆x = 1/8 and nodes 1/2, 5/8, 3/4, 7/8, 1.
√
Letting g(x) = 1 + 4x2 , the requested Simpson’s Rule approximation is
( ( ) ( ) ( ) ( ) )
∆x 1 5 3 7
g + 4g + 2g + 4g + g (1) , or 0.9050506.
3 2 8 4 8
The exact length is 0.9050460693, to 10 decimal places.
4. We have two options: we can either write the equation of the curve as y = f (x) or as x = g(y). The
former involves solving a cubic, whereas the latter involves solving a quadratic. Being sensible, we
choose the latter. We can either use the Quadratic Formula, or we can √complete the square by adding
1 to both side of the equation. Either way, we find x = g(y) = −1 + y 3 + y + 7. (The negative sign
in the ± of the solution for y results in a branch for which y ≤ −1, and we want a branch that includes
the values −1.5 ≤ y ≤ 1 ). The requested length L is given by
√
∫ 1 2
(3 y 2 + 1)
L= 1+ dy.
−1.5 4 (y 3 + y + 7)
An accurate numerical integration, not required, yields the value 3.099775857.
Life is filled with choices. Three types of choices, in fact: good choices and bad choices. Had we solved
the equation x2 + 2 x = y 3 + y + 6 for y instead of x, we would have obtained, with a little bit of work,
the details of which are omitted, the following formula for y = f (x):

√
We will omit the expression for 1 + f ′ (x)2 , which is rather intimidating: the expression for f (x) is
already sufficiently convincing that it presents a more difficult route to the arc length. Nevertheless. as
far as the numerical value of the arc length, it is five of one, five-sixths of half a dozen of the other, as
they say. A screen capture from a Maple session shows the same answer when the length is expressed
as an integral with respect to x:
165

5. We have
( )2 ( )2
d a 1
1+ (exp(x/a) + exp(−x/a)) = 1+ (exp(x/a) − exp(−x/a))
dx 2 2
( )
1 1 1
= 1+ exp(2x/a) − + exp(−x/a)
4 2 4
1 1 1
= exp(2x/a) + + exp(−x/a)
4 2 4
1
= (exp(2x/a) + 2 + exp(−2x/a))
4
1 2
= (exp(x/a) + exp(−x/a)) .
4
The requested arc length L is given by
√ ( )2
∫ a
d a
L = 1+ (exp(x/a) + exp(−x/a)) dx
−a dx 2
∫ a √
1 2
= (exp(x/a) + exp(−x/a)) dx
−a 4
∫ a
1
= (exp(x/a) + exp(−x/a)) dx
−a 2
( )
1
= a e−
e
166
Solutions: Area of a Surface of
Revolution (Corresponds to Stewart
8.2)

1. Making the change of variable u = 1 + 9 x4 , du = 36 x3 dx, we find that the surface area is given by

∫ ∫ 1 ∫ 10
1 √ √ π √ π ( √ )
2 π f (x) 1 + f ′ (x)2 dx = 2 π x3 1 + 9 x4 dx = u du = 10 10 − 1 .
0 0 18 1 27

2. We write x = g(y) = y 3 . The calculation is identical to that of the preceding exercise except for
the limits of integration (and using y for the variable of integration, although the symbol used for the
variable of integration has no effect on the numerical value of a definite integral):

∫ ∫ 2 ∫ 145
2 √
′
√ π √ π ( √ √ )
2
2 π f (y) 1 + g (y) dy = 3 4
2 π y 1 + 9 y dy = u du = 145 145 − 10 10 .
1 1 18 10 27

√ ( √ )
3. Let f (x) = exp(−x) and F (u) = π u 1 + u2 + π ln u + 1 + u2 + C. Making the change of variable
u = exp(−x), du = − exp(−x) dx, we find that the surface area is given by

∫ ∫ ∫ √ ( )
ln(2) √ ln(2) √ 1
1
2 π f (x) 1 + f ′ (x)2 dx = 2 π exp(−x) 1 + exp(−x)2 dx = 2 π 2
1 + u du = F (1)−F = 3.9438245
0 0 1/2 2

√
4. For f (x) = 25 − x2 we have

√ √
√ x2 25 5
1 + f ′ (x)2 = 1+ = =√ .
25 − x2 25 − x 2
25 − x2

The surface area S is given by

∫ 4 √ ∫ 4 √ ∫ 4
5
S= 2 π f (x) 1 + f ′ (x)2 dx = 2 π 25 − x2 √ dx = 10 π 1 dx = 70 π.
−3 −3 25 − x2 −3

167
168

√
5. Solving y = f (x) = x2 for x, we have x = g(y) = y. The surface area S is given by
∫ 4√
S = 2 π g(y) 1 + g ′ (y)2 dy
0
∫ 4 √
√ 1
= 2π y 1+ dy
0 4y
∫ 4√
1
= 2π y + dy
0 4
∫ 4 √
1
= 2π 4 y + 1 dy
0 2
∫ 17
1
= π u1/2 du (u = 4 y + 1, du = 4 dy)
4
π ( 1
√ )
= 17 17 − 1 .
6

6. Let f (x) = x2 . The surface area S is given by

∫ 2 √
S = 2πx 1 + f ′ (x)2 dx
0
∫ 2 √
= 2π x 1 + 4 x2 dx
0
∫
2 π 17 1/2
= u du (u = 1 + 4 x2 , du = 8 x dx)
8 1
π ( √ )
= 17 17 − 1 .
6

7. We calculate
( )2 ( ( )2 ) ( )2 ( )2
′ 1 1 1 1 1 1
1 + f (x) = 1 + 2 x −
2
= 1 + (2 x) −
2
= (2 x)2 + = 2x + .
8x 2 8x 2 8x 8x

The arc length is

∫ 2 √ ∫ 2( )( )
1 1
2 π f (x) 1 + f ′ (x)2 dx = 2 π x2 − ln(x) 2x + dx
1 1 8 8x
∫ 2( )
1 1 1 1
= 2π 2 x3 + x − x ln(x) − ln(x) dx
8 4 64 x
( 1 )
63 1
= π − ln2 (2) − ln(2)
4 64
≈ 47.278914.
Solutions: Center of Mass
(Corresponds to Stewart 8.3)

(( 2 ) )
1. The area of the region is 1. The center of mass is e + 1 /4, (e − 2) /2

2. The area A of the region is

∫ 2 ( ) 4
A= 2 x − x2 dx = .
0 3

Therefore,

∫
1 2 ( )
x̄ = x 2 x − x2 dx
A 0
( )2
1 2 3 1 4
= x − x
A 3 4 0
1 4
=
A 3
= 1

169
170

and
∫ (
1 2 ( )2 )
ȳ = (2 x)2 − x2 dx
2A 0
( )2
1 4 3 1 5
= x − x
2A 3 5 0
1 64
=
2A 15
8
= .
5

3. The area A of the region is calculated as follows:

∫ 4 ( )4
( √ ) 1 2 1
A= 3 x − (x + 2) dx = 2 x − x − 2 x = .
3/2
1 2 1 2
The x-center of mass x̄ is calculated as follows:
∫ 4
1 ( √ )
x̄ = · x 3 x − (x + 2) dx
A 1
∫ 4( )
1
= · 3 x3/2 − x2 − 2 x dx
A 1
( )4
1 6 5/2 1 3 2

= · x − x −x
A 5 3
( 1
)
1 192 64 6 1
= − − 16 − + + 1
A 5 3 5 3
1 6
= ·
1/2 5
12
= .
5
171

The y-center of mass ȳ is calculated as follows:

∫ 4
1 ( √ 2 )
ȳ = · (3 x) − (x + 2)2 dx
2A 1
∫ 4
1 ( )
= · 5 x − x2 − 4 dx
2A 1
( )4
1 5 2 1 3
= · x − x − 4 x
2A 2 3
( 1
)
1 64 5 1
= 40 − − 16 − + + 4
2A 3 2 3
1 9
= ·
2 (1/2) 2
9
= .
2

4. The area A of the region is 2h + π/2, the sum of the areas of a 2 × h rectangle and a half-disk
√ of radius
1. The equation of the semicircle is (x − 1)2 + (y − h)2 = 1, h ≤ y, or y = f (x) = h + 1 − (x − 1)2 .
We calculate
∫ 2 ∫ 2( √ ( ))
f (x)2 dx = h2 + 2 h 1 − (x − 1)2 + 1 − (x − 1)2 dx
0 0
∫ 2
π ( )
2
= 2h + 2h + 1 − (x − 1)2 dx
2 0
2 4
= 2h + hπ + .
3
When this quantity is divided by 2 A, the y-center of mass, namely h, results. Therefore

2 h2 + h π + 4/3 4
=h or 2 h2 + h π + = 4 h2 + h π.
4h + π 3
172
√
2 6
As a result, 6 h = 4 , or h = .
3
Solutions: Separable Differential
Equations (Corresponds to Stewart
9.3)
∫ ∫
1 1
1. We have dy = dx, or ln(|y|) = ln(|1+x|)+C. It follows that |y| = A |1+x| for A = exp(C).
y 1+x
Substituting x = 3 and y = 8 results in A = 2. Thus, |y| = 2 |1 + x|. For values of x such as the
cited values 3 and 5, the absolute values on the right side of the solution are not needed. That is,
|y| = 2 (1 + x) for x > 1/2. The solution y = −2 (1 + x) is not consistent with the given initial value.
Therefore y(x) = 2 (1 + x) and y(5) = 12.
2. After separating variables and integrating, we obtain y(x)3 + 2 y(x) = x2 + x + C. Substituting x = 2
and y(2) = 1 results in C = −3. Substituting y(x) = 0 in the equation
(√ y(x)
)
3
+ 2 y(x) = x2 + x − 3
results in a quadratic equation in x. the positive root of which is 13 − 1 /2.
√
3. Answer: y (x) = 3 3 x + 3 x2 + 2
( )2
4. Answer: y (x) = 2 x3/2 + 1
√
5. Answer: y (x) = 5 exp(2 x) − 1
6. Let y(t) be the mass of salt in the tank (measured in kg) at time t (measured in min). The differential
equation is
dy y
= 0 − 20 · (rate = rate in - rate out)
dt 1000
The solution of the initial value problem is y(t) = 100 exp(−t/50). let T be the time when y(T ) = 10.
Solve the equation 10 = 100 exp(−T /50) to obtain T = 50 ln(10).
7. Let y(t) be the mass of salt in the mixing tank at time t. The rate equation for the change of the mass
of salt in the mixing tank, namely overall rate = rate in - rate out, is, in units of kg per min,
dy 2 y 1
= · 10 − · 10 = (40 − y) .
dt 10 200 20
∫ ∫
1
This differential equation leads to the equation 20 dy = 1 dt, or −20 ln( | 40 − y | ) = t + C.
40 − y
Into this last equation, substitute t = 0 and y = 5, which is the initial condition y(0) = 5, to obtain
C = −20 ln(35). We therefore have −20 ln( | 40 − y | ) = t − 20 ln(35), or ln( | 40 ( − y |)) = ln(35) −
t/20, or, on exponentiating, | 40 − y | = exp (ln(35) − t/20), or | 40 − y | = exp ln(35) exp(−t/20),
or | 40 − y | = 35 exp(−t/20). Are the absolute values needed? Initially y is equal to 5, which is
less than 40, but y is increasing. If y were to grow larger than 40, the concentration in the mixing

173
174

tank would exceed 40 kg per 200 L, or 1 kg per 5 L. Such a concentration would exceed both the initial
concentration in the mixing tank and the concentration of the inflow. Clearly this could happen only if
a brine sorcerer cast a spell on our mixing tank. Absent such magic, we may safely drop the absolute
values. Thus, 40 − y = 35 exp(−t/20), or y = 40 − 35 exp(−t/20). Set t = 60 min (1 hour) to obtain
y(60) = 40 − 35 exp(−3), or y(60) = 38.2575. The concentration is 38.2575 kg per 200 L, or 0.1913 kg/L.

8. The rate equation, overall rate = rate in - rate out, is

dy 1 y 1
= · 10 − · 10 = (400 − y) .
dt 5 2000 200
where y(t) is the mass in kg of the salt in the mixing tank’s solution at time t minutes. To solve this
separable ode, we set ∫ ∫
1 1
dy = dt.
400 − y 200
After integrating we obtain − ln ( | 400 − y | ) = C + t/200, or ln ( | 400 − y | ) = C − t/200 (where we
replaced the constant −C with plain old C), or | 400 − y | = C exp(−t/200) (where we replaced the
constant exp(C) with plain old C). Because y(0) = 100, we solve | 400 − 100 | = C exp(−0/200) for C,
the last manifestation of C, and find C = 300. Thus, | 400 − y | = 300 exp(−t/200). As long as y(t)
remains less than 400, we may dispense with the absolute values. Common sense tells us that this will
be the case. Were y(t) greater than 400 , the concentration of salt would be greater than 400 kg per
2000 liters, or 0.2 kg per liter. How could this possibly happen when the initial concentation is 100 kg
per 2000 liters, or 0.05 kg per liter, and the inflow has concentration 1 kg per 5 liters, or 0.2 kg per liter?
Thus 400 − y = 300 exp(−t/200), or y = 400 − 300 exp(−t/200). Because the volume of water in the
tank is always 2000 liters, the concentration doubles in strength when the salt content doubles. We solve
y(t) = 200 to find t = 200 ln(3/2), or about 81 minutes. Long term, the amount y∞ of salt is about
( ( ))
t
y∞ = lim y(t) = lim 400 − 300 exp − = 400.
t→∞ t→∞ 200

Notice that the long term concentration, 400 kg per 2000 liters is the same concentration of the brine
solution flowing into the mixing tank.

9. The Law of Radioactive Decay, D(m)(t) = −λ m(t), gives us

∫ ∫
1
dm = −λ 1 dt
m

or ln( | m(t) | ) = −λ t + C, or m(t) = exp (−λ t + C), or m(t) = exp(C) exp (−λ t), or m(t) =
C exp (−λ t). Two remarks: 1) the absolute values were dropped because m(t), being a mass, is nonneg-
ative, and 2) the constant exp(C), being an arbitrary (positive) constant rather than a specific constant
was replaced with C, which is a simpler choice for representing an arbitrary constant. The initial value
condition m(0) = m0 allows us to solve for C: C = m0 . Thus, m(t) = m0 exp (−λ t). C

10. We have
( ) 1
m(t + τ ) = m0 exp −λ (t + τ ) = m0 exp (−λ t)) exp (−λ τ ) = m(t) exp(− ln(2)) = m(t).
2

11. We have ( ( ))
t m0
m(t) = m0 exp(−λ t) = m0 exp(− ln(2)) = m0 exp ln 2(−t/τ ) = t/τ .
τ 2
175

12. First note that the answer is not “We lose two grams every hundred years so after 100 more years the
isotope will have decayed from 6 grams to 4 grams.” Instead, we use the formula m(t) = m0 exp(−λt) =
8 exp(−λt) for the amount of radioactive material at time t. Since m(100) = 6, we have
( )100
6 = m(100) = 8 e−100 λ = 8 e−λ .

From this equation we conclude that

( )1/100 ( )1/100
6 3
e−λ = = .
8 4

Thus, the formula for the amount of isotope present at time t is

( )t/100
−λt
( )t 3
m(t) = 8 e = 8 e−λ = 8 .
4

There will be four grams of material present when

( )t/100
3
4 = m(t) = 8 ·
4
or ( )t/100
1 3
= .
2 4

We solve for t by taking the natural logarithm of both sides:

(( ) )
t/100
3 t
ln(1/2) = ln = · ln(3/4),
4 100

or

ln(1/2)
t = 100 · ≈ 240.942.
ln(3/4)
Thus, at t = 240.942, which is to say after 140.942 more years, there will be 4 grams of the isotope
remaining.

13. We will use the form of the Law of Radioactive Decay obtained in Exercise 11: it is more convenient
when a half-life is given as the parameter describing decay. We have

m(3700) 1
= 3700/5730 ≈ 0.63917.
m0 2

14. We have
m(1994 − 1200) 1
= 794/5730 ≈ 0.90842.
m0 2

15. We have
m0 5730 ln(0.0879)
0.0879 · m0 = m(T ) = or T =− ≈ 20 100 years.
2 /5730
T ln(2)
176

16. The differential equation

dv g R2
v· =− 2.
dy (R + y)
is separable. Following the general procedure, we obtain
∫ ∫ ( )
gR2
v dv = − 2 dy
(R + y)
or
1 2 g R2
v = + C.
2 (R + y)
When y = 0 we have v = v0 . Therefore

1 2 g R2
v0 = + C,
2 (R + 0)

or C = 1
2 v02 − g R. It follows that
( )
1 2 g R2 1 2
v = + v −gR ,
2 (R + y) 2 0
or
R ( )
v 2 = (2 g R) + v02 − 2 g R .
(R + y)
√ ( )
Assuming v0 < 2gR, we set v = 0 and solve for y. We find that the maximum height is v02 R/ 2gR − v02 .

√
Note: If v0 ≥ 2gR, then the second summand of the formula for v 2 is nonnegative, and
R
v2 ≥ 2 g R .
(R + y)

It is not possible for y to remain bounded. For if there was a constant K with y ≤ K, then we would
have R + y ≤ R + K, and therefore 1/(R + y) ≥ 1/(R + K) and R/(R + y) ≥ R/(R + K). It would
follow that
R
v2 ≥ 2 g R ,
(R + K)
or √
R
v ≥ 2gR .
(R + K)
Let ρ denote the positive constant on the right side of this inequality, so that v ≥ ρ. A consequence of
this ineqality is that the value of y at time t exceeds ρ t. By letting t tend to infinity in the inequality
y(t) ≥ ρ t, we see that y → ∞ as t → ∞. This contradicts the hypothesis
√ that all values of y satisfy
y ≤ K. However you dice it, y must become unbounded if v0 ≥ √2gR. In other words, the projectile
escapes
√ Earth’s gravity and does not fall back to Earth if v0 ≥ 2gR. For this reason we say that
2gR is Earth’s escape velocity. Now imagine
√ an object BH so massive that the acceleration gBH at the
surface of the object is so great that 2gBH R exceeds the speed of light. What happens to a photon
trying to escape the object’s gravity?
Solutions: Models of Population
Growth (Corresponds to Stewart 9.4)

1. The solution to the differential equation D(P )(t) = k P (t) is P (t) = P0 exp(k t) where P0 = P (0). It
follows that
P (t + T ) = P0 ek (t+T ) = P0 ek t ek T = P (t) ek T = P (t) eln(2) = 2 P (t).
2. We have ( ) ( ( ))
t
P (t) = P0 ek t = P0 exp ln(2) = P0 exp ln 2t/T = P0 2t/T .
T
3. Let us measure time t in hours with t = 0 corresponding to 10:00 AM, time t = 1/3 corresponding to
10:20 AM, and time t = 2.5 corresponding to 12:30 PM. Because P (1/3) = 12000 = 2 × 6000 = 2 × P (0),
we see that T = 1/3 − 0 = 1/3 is the doubling time for the population. From Exercise 2 we have
P (t) = 6000 · 23 t , and P (2.5) = 6000 · 27.5 ≈ 1 086 116. That’s a lot of colonists. We’re going to need
more broth!
4. In integrated form, after separating
( variables
) to their respective sides of the equality sign, the differential
equation D(P )(t) = k P (t) 1 − P (t) becomes
∫ ∫
1
dP = k dt,
P (1 − P )
or, by finding the partial fraction form of the left side integrand,
∫ ( )
1 1
+ dP = k t + C.
P 1−P
This equation leads to ln(P ) − ln( | 1 − P | ) = k t + C. By using the identity ln(u) − ln(v) = ln(u/v) and
discarding the absolute values, which are superfluous because 1−P > 0, we have ln(P/(1−P )) = k t+C.
Applying the natural exponential function results in P/(1 − P ) = exp(C) exp(k t), which may be written
more consicely as P/(1−P ) = C exp(k t). (As usual, there is no need to write an arbitrary constant such
as exp(C) in such a specific form.) A small algebra trick helps us solve for P : equate the reciprocals of
each side of the preceding equation to obtain (1 − P )/P = C exp(−k t) (writing the arbitrary constant
1/C as C). Then 1/P − 1 = C exp(−k t). This is a good time to solve for C. Setting P = P0 and t − 0,
we find C = 1/P0 − 1. Thus, 1/P − 1 = (1/P0 − 1) exp(−k t), or 1/P = 1 + (1/P0 − 1) exp(−k t), or or
1
P (t) = ( ) .
1+ 1
P0 − 1 exp(−k t)

Stewart gives the logistic differential equation as

( )
dP P
= kP 1−
dt M

177
178

with solution
M M − P0
P (t) = where A= .
1 + A exp(−k t) P0
The equation we have used to model the spread of an infection is the logistic differential equation with
carrying capacity M equal to 1. If M is set equal to 1 in Stewart’s solution of the logistic equation,
then it matches the solution we have derived for the spread of an infection.
5. In the formula derived in the preceding problem, there are two unspecified constants: the initial pop-
ulation P0 and the proportionality constant k. Let t = 0 correspond to 9 AM Monday morning. Then
P0 = P (0) = 1/100 and
1
P (t) = .
1 + (100 − 1) exp(−k t)
Let us use days as the unit of time. Then t = 1 corresponds to 9 AM Tuesday morning and P (1) = 3/100,
or
3 1
= .
100 1 + 99 exp(−k)
We obtain k = − ln(97/297). Thus
1
P (t) = ,
1 + 99 exp(t ln(97/297))
or
1
P (t) = ( 97 )t .
1 + 99 297
The equation P (t) = 1/2 has solution t = 4.106... . According to our model, about 4.1 days after 9 AM
Monday, that is to say, by about 11:30 Friday morning, the fiftieth child will be infected.
6. We may write the differential equation as
( ) ( ) ( )
dP b P P
= aP 1 − P = aP 1 − = aP 1 −
dt a a/b M

where M = a/b is the carrying capacity.

Solutions: Linear Differential
Equations (Corresponds to Stewart
9.5)

1. The given differential equation is already in the standard

(∫ form for) a linear differential equation, with
p(x) = 2 and q(x) = 1. The integrating factor is exp p(x) dx , or exp (2x). After mutiplying each
side of the differential equation by the integrating factor, we have

dy
exp (2x) + 2 exp (2x) y = exp (2x) ,
dx
or
d
(exp (2x) y) = exp (2x) ,
dx
or ∫
exp (2x) y = exp (2x) dx,

or
1
exp (2x) y = exp (2x) + C.
2
Now is an auspicious time for using the initial condition y(0) = 2 to determine that C = 3/2. It follows
that
1 3
y = + exp (−2 x) .
2 2
2. In standard form, the given linear differential equation is
( )
dy 1
+ − y=1
dx 2
(∫ )
with p(x) = 1/2 and q(x) = 1. The integrating factor is exp p(x) dx , or exp (−x/2). After mutiplying
each side of the differential equation by the integrating factor, we have
( )
dy 1
exp (−x/2) + − exp (−x/2) y = exp (−x/2) ,
dx 2
or
d
(exp (−x/2) y) = exp (−x/2) ,
dx
or ∫
exp (−x/2) y = exp (−x/2) dx,

179
180

or
exp (−x/2) y = −2 exp (−x/2) + C.

Now is a timely occasion for using the initial condition y(0) = 3 to determine that C = 5. It follows
that
( )
y = exp (x/2) 5 − 2 exp (−x/2) ,

or
y = 5 exp (x/2) − 2.

3. We rewrite the differential equation in the form

dy 2x
+ y = 15x2
dx 1 + x2
( ) (∫ ( ) )
to see
( that
( p(x) ))= 2x/ 1 + x2 and q(x) = 15 x2 . The integrating factor is exp 2x/ 1 + x2 dx , or
exp ln 1 + x2 , or 1 + x2 . Multiplying the rewritten differential equation by the integrating factor
leaves us with
( ) dy ( )
1 + x2 + 2 x y = 15 x2 + x4 ,
dx
or
d (( ) ) ( )
1 + x2 y = 15 x2 + x4 .
dx
Integrating each side gives us
( )
1 + x2 y = 5 x3 + 3 x5 + C.
( )
This is a splendid moment for calculating the constant of integration: 1 + 12 6 = 5 13 + 3 15 + C, or
C = 4. Substituting this value for C and solving for y, we obtain

5 x3 + 3 x5 + 4
y= .
1 + x2

dy 1
4. Writing the differential equation in the standard linear form + y = 1 directs us to the integrating
(∫ ) dx x
factor exp (1/x) dx , or exp (ln(x)), or x. Multiplying the rewritten differential equation by the
integrating factor leaves us with

dy d
x + y = x, or x (x y) = x.
dx dx

Integrating each side results in x y = x2 /2 + C. Now is a dandy time to determine the constant of
integration. Set x = 4 and y = 5 in x y = x2 /2 + C to obtain 20 = 8 + C, or C = 12. It follows that
x y = x2 /2 + 12, or y = x/2 + 12/x.
dy
( ) √
5. Answer: y = 2x3/2 + 7/x 2
(∫ In Standard
) form, the given linear
( (equation
)) is dx + x2 y = 7 x, so the
integrating factor is exp (2/x) dx , or exp (2 ln(x)), or exp ln x2 , or x2 . Multiplying the differential
dy d ( 2 )
equation by the integrating factor results in x2 dx + 2 x y = 7 x5/2 , or x y = 7 x5/2 . Integrating
dx
results in x2 y = 2 x7/2 + C. Using the initial value to solve for C at this juncture would be hunky-dory.
Setting x = 1 and y = 9, we have 12 × 9 = 2 × 17/2 + C, or C = 7. Therefore, x2 y = 2 x7/2 + 7, or
y = 2 x3/2 + 7/x2 .
181

6. The(∫given
) differential
( equation
) is in standard linear form as presented. The integrating factor is
exp x , or exp x2 /2 . We multiply the differential equation by this factor, obtaining
( ) dy ( ) ( ) d ( ( ) ) ( )
exp x2 /2 + x exp x2 /2 y = x exp x2 /2 , or exp x2 /2 y = x exp x2 /2 . Integration gives
( dx) ( ) dx
us exp x2 /2 y = exp x2 /2 + C. Solving for C now is an admirable next step. We substitute x = 0
and y = 4 to obtain 4 = 1 + C, or C (= 3. We ) now replace C with this value in our implicit solution and
solve explicitly for y: y = 1 + 3 exp −x2 /2 .
dy dy
7. Matching the differential equation − 2 x y = 4 x with the standard form + p(x) y = q(x) of the lin-
dx dx (∫ )
ear differential equation, we set p(x) = −2 x to obtain the integrating factor exp (−2 x) , which evalu-
( ) ( ) dy ( ) ( )
ates to exp −x2 . Multiplying the equation by this factor gives us exp −x2 + (−2 x) exp −x2 y = 4 x exp −x2 ,
dx
d ( ( 2) ) ( 2) ( 2) ( )
or exp −x y = 4 x exp −x . Integrating yields exp −x y = C − 2 exp −x2 . Now is an
dx
admirable time for determining( C.)Set x = 0 and y( = 1)in the equation( just) obtained to get 1 = C − 2,
or C = 3. It follows that exp −x2 y = 3 − 2 exp −x2 , or y = 3 exp x2 − 2.

8. The given differential equation ( is)equivalent to the differential equation of the preceding exercise. The
answer is therefore y = 3 exp x2 − 2. However, in the form presented in the current exercise, it takes
some unwinding to see that the equation is linear. On∫the other hand, ∫ in the present form it is obvious
1
that it is separable. If we use that technique we obtain dy = 2 x dx, or ln( | 2 + y | ) = x2 + C.
2+y
) ln(3). It follows that ln( | 2 +(y 2| )) = x + ln(3), or
2
The initial condition
( 2 y(0)
) = 1 results in C( =
| 2 + y | = exp x + ln(3) , or | 2 + y | = exp x exp (ln(3)), or | 2 + y | = 3 exp x . Notice that y
2

cannot be −2 because our equation tells us that | 2 + y | must be positive. Therefore, an arc of a solution
curve must either lie under the horizontal line y = −2 or above the horizontal line y = −2. An arc that
lies under y = −2 is disconnected from an arc that lies above. Therefore, the initial condition y(0) = 1
can only determine an arc of a solution curve that lies above y = −2. For(such ) an arc, 2 + y(is positive
)
and we my discard the troublesome absolute values. Thus, 2 + y = 3 exp x2 , or y = 3 exp x2 − 2.

9. Writing the linear differential equation (∫ in the standard

) form D(y)(x) + p(x) y(x) = q(x), we have
p(x) = −3. The integrating factor is exp p(x) dx , or exp (−3 x). After multiplying each side by the
dy ( −3x )
integrating factor, we obtain e−3x + (−3)e−3x y = 1, or dxd
e y = 1, or e−3x y = x + C. Now is
dx
an estimable time for determining the constant of integration. Set x = 2 and y = 0, the values of the
initial condition, in the equation just obtained. The result is 0 = 2+C, or C = −2. Thus, e−3x y = x−2,
or y = (x − 2) e3x .

10. Writing the linear differential equation in the standard form D(y)(x) + p(x) y(x) = q(x), we have
dy ( ) (∫ )
+ − cot(x) y = csc(x). Thus, p(x) = − cot(x), and the integrating factor is exp p(x) dx , or
dx ( ∫ )
exp − cot(x) dx , or exp (ln(csc(x))), or csc(x). After multiplying each side by the integrating factor,
dy ( )
we obtain csc(x) + − csc(x) cot(x) y = csc2 (x), or dx
d
(csc(x) y) = csc2 (x), or csc(x) y = − cot(x) +
dx
C. Now is a meritorious time for determining the constant of integration. Set x = π/2 and y = 4, the
values of the initial condition, in the equation just obtained. The result is 4 = 0 + C, or C = 4. Thus,
csc(x) y = 4 − cot(x), or, on dividing by csc(x), which is equivalent to multiplying by the reciprocal
sin(x), we obtain y = 4 sin(x) − cos(x).

11. The integrating factor is

(∫ )
1
exp dx , or exp (ln(ln(x))) , or ln(x).
x ln(x)
182

Multiply each side of the differential equation by this integrating factor to obtain
dy 1
ln(x) + y = ln(x),
dx x
or
d
(ln(x) y) = ln(x),
dx
or ∫
ln(x) y = ln(x) dx = x ln(x) − x + C.

This is a propitious time for determining the value of C. Setting

∫ y = 2 e and x = e results in the
equation 2 e = e − e + C, or C = 2 e. It follows that ln(x) y = ln(x) dx = x ln(x) − x + 2 e, or
2e − x
y =x+ .
ln(x)

12. Let y(t), the weight of the solute in the mixing tank t minutes after the valves have been opened. We
are given that y(0) = 80. The differential equation for y(t) is obtained by translating into mathematics
the following observation: “The overall rate of change of the weight of the solute in the mixing tank
equals the rate of the weight of the solute going in minus the rate of the weight of the solute going out.”
The rate of the volume of the fluid entering is 24 gal/min. Because the concentration of this solution
is 1 lb/gal, we see that the rate of the weight of the solute going in is (24 gal/min) × (1 lb/gal), or 24
lb/min. The rate of the volume of the fluid exiting is 16 gal/min. Because the concentration of this
solution is y/V (t) lb/gal, where V (t) is the volume of the solution in the tank at( time t, we
) see that the
rate of the weight of the solute going out is (16gal/min) × (y/V (t) lb/gal), or 16/V (t) y lb/min. We
are given that V (0) = 120 gal. The information we have been given also tells us that the net increase
of the fluid in the tank is (24 − 16) gal per minute. Therefore, V (t) = 120 + 8 t. Thus, the differential
equation for y(t) is
dy 16
= 24 − y.
dt 120 + 8 t
This linear differential equation has standard form
dy 16
+ y = 24.
dt 120 + 8 t
(∫ )
The integrating factor is exp 16/(120 + 8 t) dt , or exp (2 ln(120 + 8 t)), or (120 + 8 t)2 . Thus,

dy
(120 + 8 t)2 + 16 (120 + 8 t) y = 24 (120 + 8 t)2 ,
dt
or
d ( )
(120 + 8 t)2 y = 24 (120 + 8 t)2 .
dt
Integrating each side, we obtain

(120 + 8 t)2 y = (120 + 8 t)3 + C.

We can make the numbers smaller by factoring 8 out of each parenthesis and dividing both sides of the
equation by 82 :
(15 + t)2 y = 8 (15 + t)3 + C.
Now is the opportune time to solve for C using the initial condition y(0) = 80. We find C = −9000.
Thus,
(15 + t)2 y = 8 (15 + t)3 − 9000,
183

or
9000
y = 8 (15 + t) − .
(15 + t)2
Consequently, y(15) = 8 (15 + 15) − 9000/(15 + 15)2 , or y(15) = 230.
184
Solutions: Sequences (Corresponds to
Stewart 11.1)

1. a0 = 30 /(2 × 0 + 1) = 1, a1 = 31 /(2 × 1 + 1) = 1, a2 = 32 /(2 × 2 + 1) = 9/5, a3 = 33 /(2 × 3 + 1) =

27/7, a4 = 34 /(2 × 4 + 1) = 9.
2. Because f ′ (xn ) = 3 x2 , we have
( )
x3n − 4 2 x3n + 2
xn+1 = xn − = .
3 x2n 3 x2n
( 3 ) ( ) ( ) ( )
It follows that x1(= 1, x2 = 2, x3 = 2
) ( 2 + 2 / 3 × 2)
2
= 20/12 = 5/3, x 4 = 2 (5/3)3
+ 2 / 3 × (5/3))2
=
358/225, x5 = 2 (358/225) + 2 / 3 × (358/225)) = 34331981/21627675 ≈ 1.58741. To five decimal
3 2

places, 41/3 = 1.58740.

3. Divide every term in sight by the highest power of n, namely n2 :
3 + 5 n2 3/n2 + 5 0+5 5
lim 2
= lim = = .
n→∞ 5 + 3 n n→∞ 5/n2 + 3 0+3 3

4. Divide every term in the fraction by the highest power of n, namely n6 :

√ √ √
n5 + 9 n6 1/n + 9 0+9 3
lim = lim = = .
n→∞ 1 + 4 n6 n→∞ 1/n6 + 4 0+4 2

5. The given sequence decays exponentially:

1 1 1
lim 3n π −n = lim n
= lim n
= lim .
n→∞ n→∞ (π/3) n→∞ exp (ln ((π/3) )) n→∞ exp (n ln (π/3))
Because π > 3, we have π/3 > 1 and k = ln (π/3) > 0. It follows that limn→∞ exp (n ln (π/3)) =
limn→∞ exp (k n) = ∞. Consequently, the limit of the given sequence is 0.
6. We have
( √ √ ) ( √ ) ( √ )
lim exp(1/ n) + exp(−1/ n) = exp lim 1/ n + exp lim (−1 n) = exp(0) + exp(0) = 2.
n→∞ n→∞ n→∞

7. We deal with a ratio of factorials by expanding the factorial of the larger number until the factorial of
the smaller number appears. In this exercise, we have (2 n + 1)! = (2 n + 1) · 2n · (2 n − 1)!. Thus,

(2 n − 1)! n2 (2 n − 1)! n2 1 n 1 1 1
lim = lim = lim = lim = .
n→∞ (2 n + 1)! n→∞ (2 n + 1) · 2n · (2 n − 1)! 2 n→∞ (2 n + 1) 2 n→∞ (2 + 1/n) 4

185
186

8. We have
| sin(n) | 1
0≤ ≤ .
n n
By the Pinching Theorem, we conclude that | sin(n) |/n → 0. It follows that sin(n)/n → 0.
9. Letting x be a continuous variable tending to infinity and making the change of variable u = 1/x, we
have
sin(1/x) sin(u)
lim n sin(1/n) = lim x sin(1/x) = lim = lim+ = 1.
n→∞ x→∞ x→∞ 1/x u→0 u

10. We again divide each term in the fraction by the highest power of n that occurs, namely n1 :
( ) ( ) ( )
2n 2 2
lim exp = lim exp = exp = e2 .
n→∞ n+3 n→∞ 1 + 3/n 1+0

11. The standard technique of dividing by a highest power works here. In this instance, we will divide by
by the highest power of n in the numerator, namely n2 . We obtain

n2 n2 /n2 1 1
lim √ = lim √ = lim √ =√ = 0.
n→∞ 5
n +1 n→∞ 5 4
n /n + 1/n4 n→∞ n + 1/n4 ∞+0

We could equally well have divided every term by what is effectively the highest power in the denomi-
nator, namely n5/2 , although that power doesn’t really appear:

n2 n2 /n5/2 n−1/2 0
lim √ = lim √ = lim √ =√ = 0.
n→∞ n5 + 1 n→∞ n5 /n5 + 1/n5 n→∞ 1 + 1/n5 1+0

Finally, a rough, intuitive calculation would lead us to the same limit. The 1 added to n5 becomes
negligible as n tends to infinity. For n = 1000, the difference between n5 + 1 and n5 is the differnece
between 1,000,000,000,000,001√and 1,000,000,000,000,000. Some difference! It’s just a rounding error.
It is obvious that limn→∞ n2 / n5 = limn→∞ n2 /n5/2 = limn→∞ 1/n1/2 = 0.
ln(n)
12. We are to evaluate lim for p = 1/2. We will show that
n→∞ np
ln(n)
lim =0
n→∞ np
for every positive constant p. Applying L’Hôpital’s Rule, we have
d
ln(n) dn ln(n) 1/n 1 1
lim = lim d p
= lim = lim = 0.
n→∞ np p np−1 p n→∞ np
dn n
n→∞ n→∞

13. This limit, for which a formal calculation leads to the indeterminate form ∞/∞, can be calculated using
L’Hôpital’s Rule:
d d
ln(n) dn ln(n) dn ln(n) 1/n
lim = lim d
= lim d
( ) = lim = 1.
n→∞ ln(3 n) n→∞ 1/n
dn ln(3 n) dn ln(3) + ln(n)
n→∞ n→∞

An algebraic approach is more elementary and equally effective:

ln(n) ln(n) ln(n)/ ln(n) 1 1

lim = lim = lim = lim = = 1.
n→∞ ln(3 n) n→∞ ln(3) + ln(n) n→∞ ln(3)/ ln(n) + ln(n)/ ln(n) n→∞ ln(3)/ ln(n) + 1 0+1
187

14. L’Hôpital’s Rule is an effective way to deal with the indeterminate form ∞/∞:
d
ln(n) dn ln(n) 1/n 1 1
lim = lim d
= lim = lim = .
n→∞ ln (n3 ) 3 2
3n /n3 3 3
dn ln (n )
n→∞ n→∞ n→∞

An even easier method is to simplify the denominator:

ln(n) ln(n) 1 1
lim = lim = lim = .
n→∞ ln (n3 ) n→∞ 3 ln (n) n→∞ 3 3

15. In this case we can resolve the indeterminate form ∞ − ∞ by appealing to an algebraic property of
logarithms to convert the indeterminate form to the more convenient indeterminate form ∞/∞:
( ) ( ) ( ) ( )
( ) n + 10 n/n + 10/n n/n + 10/n 1+0
lim ln(n+10)−ln(n+1) = lim ln = lim ln = ln lim = ln = 0.
n→∞ n→∞ n+1 n→∞ n/n + 1/n n→∞ n/n + 1/n 1+0

16. The only significance of the number 1.1 is that it is greater than 1. We will write the solution with b
denoting 1.1. The calculation shows that the same result holds for any b > 1. We write n3 · (1.1)−n
as n3 /bn . The formal limit is ∞/∞, which is an indeterminate form. However, it is an indeperminate
form to which we may apply L’Hôpital’s Rule without any preliminary algebraic manipulation (such as
is necessary for the indeterminate forms 0 · ∞, 00 , ∞0 , 1∞ , and ∞ − ∞). In preparation for the use of
d
L’Hôpital’s Rule, we recall that bx = ln(b) · bx . Through repeated use of L’Hôpital’s Rule, it follows
dx
that
d 3
n3 dn n
lim = lim
n→∞ bn n→∞ d bn
dn
3 n2
= lim n
ln(b) n→∞ b
d 2
3 n
= lim dnd
ln(b) n→∞ dn bn
3·2 n
= lim
ln(b)2 n→∞ bn
3·2 d
dn n
= lim d n
ln(b)2 n→∞ dn b
3! 1
= lim
ln(b)3 n→∞ bn
= 0.

It is worth noting that if the exponent 3 is replaced with an arbitrary positive integer k and if we apply
L’Hôpital’s Rule k times, then we obtain

nk k! 1
lim = lim n = 0.
n→∞ bn ln(b) k n→∞ b
The phrase “exponential growth” has become a cliché in English. Nobody speaks of “geometric growth”.
But the geometric growth bn is exponential growth in disguise: bn = exp(n ln(b)). What this exercise
shows is that
polynomial growth in n
lim = 0.
n→∞ exponential growth in n
188

n5 n2
17. The solution to Exercise 16 shows that lim n = 0 and lim n = 0. On dividing by 2n every term in
n→∞ 2 n→∞ 2
the given fraction whose limit we are to determine, we have

2n + n5 2n /2n + n5 /2n 1 + n5 /2n 1+0

lim = lim = lim = = 1.
n→∞ 2n + n2 n→∞ 2n /2n + n2 /2n n→∞ 1 + n2 /2n 1+0

18. The fastest growing term in the given fraction is 11n . Divide every term in the fraction by it:

3 × 7n + 6 × 11n 3 × 7n /11n + 6 × 11n /11n 3 × (7/11)n + 6 0+6

lim = lim = lim = = 3.
n→∞ 5 + 2 × 11
n n n→∞ 5 /11 + 2 × 11 /11
n n n n n→∞ n
(5/11) + 2 0+2

19. Answer: 1/e (Not 1!) This exercise involves the indeterminate form 1∞ . Resist all temptation, no matter
how great it might be, to think that 1∞ is 1. It might be, but often it is not. Generally, indeterminate
forms involving an exponent are handled by a preliminary application of the natural logarithm. Let
( )
1 ln(1 − 1/n)
yn = ln (an ) = n ln 1 − = .
n 1/n

We now have the indeterminate form 0/0, which is L’Hôpital Rule ready. To simplify the differentiation,
let us make the change of variable x = 1/n. Then

ln(1 − x)
lim yn = lim
n→∞ x→0+ x
dx ln(1 − x)
d
= lim d
dx x
x→0+

1
= − lim+
x→0 1−x
= −1.

It follows that

lim an = lim exp (ln (an ))

n→∞ n→∞
= lim exp (yn )
n→∞
( )
= exp lim yn
n→∞
= exp (−1) .

The same calculation shows that limn→∞ (1 + a/n)n = exp(a) for any constant a.

20. The limit of the sequence, calculated with no refinement, is the indeterminate form ∞0 . To evaluate an
indeterminate form involving a power, namely one of the three forms 00 , 1∞ , and ∞0 , first apply the
natural logarithm. The result of this application will be, in all three cases, ±0 · ∞, an indeterminate
form with no power. To handle the indeterminate form ±0 · ∞, write the product as a quotient in which
the denominator is the reciprocal of one of the two factors. The resulting quotient will be one of the
indeterminate forms 0/0 or ∞/∞. Both of these indeterminate forms can be treated with L’Hôpital’s
Rule. Thus, if yn = ln (an ) , then
(( ) ( )
)
2 1/2n 1 ( 2
) ln 1 + n2
yn = ln 1 + n = ln 1 + n =
2n 2n
189

and
d
( )
dn ln 1 + n2
lim yn = lim d
dn 2n
n→∞ n→∞

2n
= lim
n→∞ 2 (1 + n2 )
n/n
= lim
n→∞ (1/n + n2 /n)
1
= lim
n→∞ (1/n + n)
1
=
0+∞
= 0.

It follows that
( )
lim an = lim exp (ln (an )) = lim exp (yn ) = exp lim yn = exp(0) = 1.
n→∞ n→∞ n→∞ n→∞

√ √ √
21. We have L = limn→∞ an+1 = limn→∞ 2 an = 2 limn→∞ an = 2 L, so L2 = 2 L, or L = 2.
22. We have L = limn→∞ an+1 = 1 + 1/ limn→∞ (an + 1) = 1 + 1/ (L + 1), so L (L + 1) = (L + 1) + 1, or
L2 = 2. The terms of the sequence are all greater than or equal to 1, so L must √
be greater than or equal
to 1. We therefore discard the negative root of L2 = 2 and conclude that L = 2.
190
Solutions: Series (Corresponds to
Stewart 11.2)

1. (i) We have
( )
2
s1 = ln ,
4

( ) ( ) ( )
2 3 2×3
s2 = ln + ln = ln ,
4 5 4×5

( ) ( ) ( )
4 2×3×4 2×3
s3 = s2 + ln = ln = ln ,
6 4×5×6 5×6

( ) ( ) ( )
5 2×3×5 2×3
s4 = s3 + ln = ln = ln ,
7 5×6×7 6×7

( ) ( ) ( )
6 2×3×6 2×3
s5 = s4 + ln = ln = ln .
8 6×7×8 7×8

(ii) The integers M and N are unequal because the given series begins with n = 2 instead of n = 1.
The integer N in sN is the number of summands in the partial sum sN . Thus, when N = 1 there is
only one summand. The partial sum begins with n = 2 and ends with n = 2. In other words, M = 2
for N = 1, which is to say that M = N + 1. This equation holds generally:

∑
M ( ) ∑
N +1 ( )
n n
sN = ln = ln .
n=2
n+2 n=2
n+2

Remark: The point of part (ii) is to emphasize that the index N of a partial sum sN refers to the
number of terms an in the sum.∑ It does not refer to the largest index of the summands an that
∞
constitute the sum. For a series n=2 an such as the one in this ∑ exercise, the N th partial sum sN is
∞
given by SN = a2 + a3 + · · · + aN + aN +1 . For a series of the form n=0 an , the N th partial sum sN is
given by SN = a0 + a1 + · · · + aN −2 + aN −1 .

(iii) From the pattern in part (i), we have

( )
2×3
sN = ln .
(N + 2)(N + 3)

191
192

(iv) We calculate
( )
2×3
lim sN = lim ln
N →∞ N →∞ (N + 2)(N + 3)
( ( ))
= lim ln(6) − ln (N + 2)(N + 3)
N →∞
( )
= ln(6) − lim ln (N + 2)(N + 3)
N →∞
= ln(6) − ∞
= −∞.

The series diverges.

2. Starting at the beginning,

2 N + 1 3
a1 = s1 = sN = = = 1,
N =1 N + 2 N =1 3

2 N + 1 2 N + 1 5 3 1
a2 = (a1 + a2 ) − a1 = s2 − s1 = − = − = ,
N + 2 N =2 N + 2 N =1 4 3 4

2 N + 1 2 N + 1 7 5 3
a3 = (a1 + a2 + a3 ) − (a1 + a2 ) = s3 − s2 = − = − = .
N + 2 N =3 N + 2 N =2 5 4 20
In general

an = (a1 + a2 + · · · + an−1 + an ) − (a1 + a2 + · · · + an−1 )

= sn − sn−1

2 N + 1 2 N + 1
= −
N +2 N =n N +2 N =n−1
2n + 1 2n − 1
= −
n+2 n+1
(n + 1)(2 n + 1) − (n + 2)(2 n − 1)
=
(n + 1)(n + 2)
3
= .
(n + 1)(n + 2)
Thus, the series we are dealing with is
∞
∑ ∞
∑ 3 1 3 1 3
an = =1+ + + + ··· + + ···
n=1 n=1
(n + 1)(n + 2) 4 20 10 (n + 1)(n + 2)

It converges to 2:
∞
∑ 2N + 1 2 N/N + 1/N 2 + 1/N
an = lim sN = lim = lim = lim = 2.
N →∞ N →∞ N + 2 N →∞ N/N + 2/N N →∞ 1 + 2/N
n=1

3. We have
∑∞ ( )n
1 5
5 − 2.5 + 1.25 − 0.625 + 0.3125 − 0.15625 + · · · = 5 − = = 10/3.
n=0
2 1 − (−1/2)
193

4. We have
∑∞ ∑∞ ∞ ( )n ( )
(−2)n−1 −1 (−2)n 1 ∑ 2 1 (−2/3)2 2
= (−2) = − − = − =− .
n=2
3n
n=2
3n 2 n=2
3 2 1 − (−2/3) 15

5. We have
∑∞ ∞ ∞ ( )n ( )
3n+1 3 ∑ 3n ∑ 3 (3/4)1
= = 12 = 12 = 36.
n=1
4n−1 4−1 n=1 4n n=1
4 1 − 3/4

6. We have
∑∞ ∑∞ ∑∞ ( )n ( )
22n+1 4n 4 (4/5)1
= 2 = 2 = 2 = 8.
n=1
5n n=1
5n n=1
5 1 − 4/5

7. Partial fractions lead to the identity

2 1 1
= − .
n2 −1 n−1 n+1
Thus,

∑
N +1
2
sN =
n=2
n2 −1
∑
N +1 ( )
1 1
= −
n=2
n−1 n+1
( ) ( ) ( ) ( ) ( ) ( ) ( )
1 1 1 1 1 1 1 1 1 1 1 1 1
= 1− + − + − + − + − + ··· − + −
3 2 4 3 5 4 6 5 6 N −1 N +1 N N +2
1 1 1
= 1+ − − .
2 N +1 N +2

Thus,
∞
∑ ( )
2 1 1 1 3
= lim sN = lim 1 + − − = .
n=2
n − 1 N →∞
2 N →∞ 2 N +1 N +2 2

8. The first four series can be shown to be divergent. However, that conclusion√can be attained by the Test
for Divergence only for the series in parts (c) and (d). Clearly limn→∞ 1/ n = 0, and, by L’Hôpital’s
Rule, limn→∞ ln(n)/n = 0. Therefore, the Test for Divergence cannot be applied to the series in parts
(a) and (b). Because limn→∞ n1/n = 1, the Test for Divergence does apply to the series in (c), and shows
that it is divergent. Because limn→∞ arctan(n) = π/2, the Test for Divergence shows that the series
in part (d) is divergent.
( Finally,
)n for any constant r, we have limn→∞ (1 + r/n)n = exp(r). Therefore,
1 + 2/n
the numerator of tends to exp(2) while the denominator tends to infinity. The ratio tends
7
to 0, and the Test for Divergence cannot be applied. The series is, in fact, convergent. The Test for
Divergence can never be successful when applied to a convergent series.
9. (i) “may” Convergence can occur, as for example with a geometric series with ratio r between -1 and 1.
But divergence can occur, as for example with the harmonic series.
(ii) “must” The Test for Divergence is logically equivalent to the ∑
statement filled in by “must.”
(iii) “may” The Test for Divergence establishes the divergence of (−1)n , but it does not establish the
divergence of the harmonic series.
∑ ∑
(iv) “may” The partial sums of (1/2)n and of (−1)n are bounded. The first of these series converges
194

but the second does not.

∑ ∑
(v) “may” The series (−1)n and 1/n both diverge. The terms of the first of these two series diverge,
but the terms of the second series converge.
(vi) “must” This is by definition: an infinite series is convergent if and only if its partial sums are
convergent.
Solutions: The Integral Test
(Corresponds to Stewart 11.3)

∞
∑ n2
1. The first three terms of are 0.30303, 0.1, and 0.1525. These terms are increasing, not de-
n3 + 32
n=1 ( )
creasing, so the Integral Test cannot be applied. However, let us be persistent. Let f (x) = x2 / x3 + 32 .
Notice that f (n) is the nth term of S1 . After some simplification we find that

( )
x x3 − 64
f ′ (x) = − 2 .
(x3 + 32)

We see that f ′ (x) ≤ 0 for x ≥ 4. This tells us that f (x) decreases for x ≥ 4, and therefore the terms of
S1 are decreasing beginning with M = 4. The Integral Test can be applied to every tail SM of S1 with
M ≥ 4.

2. As noted in the preceding exercise, the Integral Test leads to a conclusion for this series by applying it
∑∞
n2
to a tail. The Integral Test shows that 3
diverges:
n=1
n + 32

∫ ∞ ∫ N
x2 x2
3
dx = lim dx
4 x + 32 N →∞ 4 + 32 x3

1 ( 3 )N
= lim ln x + 32
N →∞ 3
( 4
)
1 ( 3 ) 1
= lim ln N + 32 − ln (96)
3 N →∞ 3
= ∞.

( )2 ( ) ( )3
3. Let f (x) = x2 / nx3 + 32 . Then f ′ (x) = −4 x x3 − 16 / x3 + 32 ≤ 0 for x ≥ 3. The Integral Test

195
196

∞
∑ n2
may be applied to the tail that begins with the third term. It shows that 2 converges:
n=1 (n3 + 32)
∫ ∞ ∫ N
x2 x2
2 dx = lim 2 dx
3 (x3 + 32) N →∞ (x3 + 32)
3
N
1
= − lim
N →∞ 3 (x + 32)
3
( 3
)
1 1
= − lim −
N →∞ 3 (N 3 + 32) 3 × 59
1
=
177
< ∞.

4. For r in the interval (0,1), the factor rn is decreasing. However, the other factor, n, of an = n rn is
increasing. It is not so easy to see which of the two opposing factors wins the battle. Following the
method used in preceding exercises, set f (x) = x rx and differentiate: f ′ (x) = (1 + x ln(r)) rx , which
is negative for 1 + x ln(r) < 0. We deduce that the terms of the series decrease for 1 + n ln(r) < 0,
or n ln(r) < −1, or n > −1/ ln(r). Why has the direction of the inequality changed? It is because
ln(r) < 0 thanks to r being less than 1. Let M denote the smallest integer greater than −1/ ln(r). If
r is slightly less than 1, this ineger might be fairly large. For example, if r = 0.99, then M = 100.
∞
The first 101 terms of the sequence {n (0.99)n }n=1 are 0.99, 1.96, 2.91, 3.84, 4.75, . . . , 36.59949949,
36.60323412,36.60323413, 36.59957. As our calculation predicted, the decrease only begins once the
summation index reaches M = 100. Integrating by parts, we find
∫
1
x rx dx = (x ln(r) − 1) rx .
ln(r)2

L’Hôpital’s Rule shows that limx→∞ x rx = 0 for 0 < r < 1. The improper integral is convergent and,
therefore, so is the series.
∫∞ ∫∞
5. We calculate 10 (1/x2 ) dx = 0.1 and 11 (1/x2 ) dx = 0.909090909 to nine decimal places. Therefore
ℓ = 1.549767731 + 0.909090909 = 1.640676822 and m = 1.5497677318 + 0.1 = 1.6497677318. The
midpoint approximation to S is (1.64067682 + 1.6497677318)/2, or 1.645222276.
∫∞ ∫∞
6. We calculate 6 (1/x3 ) dx = 1/72 and 6 (1/x3 ) dx = 1/50. The value of S, whatever it may be, lies
in the interval [1.185662037 + 1/72, 1.185662037 + 1/50], or [1.199550926, 1.205662037]. The midpoint
is 1.202606482. If this value is used to estimate the exact value of S, then the maximum error is the
distance of the midpoint to an endpoint of the interval, or half the width of the interval, or 0.0030555555.
The (unknown) exact value of S is 1.202056903 to nine decimal places.
Solutions: Series with Positive Terms
(Corresponds to Stewart 11.4 and
11.3)

∑∞
1
1. Converges. The convergent series 2
may be used for comparison.
n=1
n

∑∞
5
2. Converges. The convergent p-series 5/4
may be used for comparison.
n=1
n

∑∞
1
3. Converges. The convergent p-series 2
may be used for comparison.
n=1
n

∑∞
3
4. Converges. The convergent p-series 4
may be used for comparison.
n=1
n

∑∞ ( )n
1
5. Converges. The convergent geometric series 2 may be used for comparison.
n=1
e

∑∞
1
6. Converges. The convergent p-series 3/2
may be used for comparison.
n=1
n

∑∞
1
7. Converges. The convergent p-series may be used for comparison.
n=1
n2
∞ ( )n
∑ 2
8. Converges. The convergent geometric series may be used for comparison.
n=1
3
∞ ( )n
∑ 1
9. Converges. The convergent geometric series may be used for comparison.
n=1
3

∑∞
1
10. Converges. The convergent p-series 3/2
may be used for comparison.
n=1
n
∞
∑ 1
11. Converges. The convergent p-series 1.01
may be used for comparison.
n=1
n

197
198

∞ ( )n
∑ 1
12. Converges. The convergent geometric series may be used for comparison.
n=1
2

∑∞
1
13. Diverges. The divergent p-series may be used for comparison.
n=1
2n

∑∞
1
14. Diverges. The divergent p-series may be used for comparison.
n=2
n

∑∞
1
15. Diverges. The divergent p-series may be used for comparison.
n=2
3n

∑∞
2
16. Diverges. The divergent p-series may be used for comparison.
n=2
n

∑∞
1
17. Diverges. The divergent p-series may be used for comparison.
n=2
n

∑∞
1
18. Diverges. The series , which diverges by the Divergence Test, may be used for comparison:
n=1
2

∑∞ ∑∞ ∑∞
3n + 1 3n 1
n + 2n
> n + 3n
= .
n=1
3 n=1
3 n=1
2

This line of investigation also shows divergence by means of the Divergence Test.

∞
∑ 1
19. Diverges. The divergent p-series √ may be used for comparison. In the following comparison,
n=1
2 n
observe that, at each step with an inequality, a smaller right side is obtained by either using a numerator
that is smaller than the one on the left, or a denominator that is larger than the one on the left:

∑∞ ∞
∑
3n + n 3n
√ n > √ n
n=1
n3 + 1 n=1
n3 + 1
∞
∑ 3n
> √ n √ n
n=1
n3 + n3
∞
∑ 1
= √ .
n=1
2 n

∑∞
1
20. Diverges. The divergent series may be used for comparison. In the following comparison, observe
n=1
2
that, at each step with an inequality, a smaller right side is obtained by either using a numerator that
199

is smaller than the one on the left, or a denominator that is larger than the one on the left:

∑∞ ∞
∑
2n + 1 2n
>
n=1
2n + n2 n=1
2n + n2
∞
∑ 2n
>
n=1
2n + 2n
∑∞
1
= .
n=1
2

This line of investigation also shows divergence by means of the Divergence Test.
∞
∑ ∑∞
1 an 3
21. Diverges. The divergent harmonic series bn = can be used for comparison: lim = .
n=1 n=1
n n→∞ bn 2
∞
∑ ∑∞
1 an
22. Converges. The convergent p-series bn = 3/2
can be used for comparison: lim = 1.
n n→∞ bn
n=1 n=1

∞
∑ ∑∞
1 an
23. Converges. The convergent p-series bn = 2
can be used for comparison: lim = 1.
n=1 n=1
n n→∞ bn

∞
∑ ∑∞
1 an
24. Converges. The convergent p-series bn = 2
can be used for comparison: lim = 1.
n=1 n=1
n n→∞ bn

∞
∑ ∑∞
1 an 1
25. Converges. The convergent p-series bn = 3/2
can be used for comparison: lim = .
n n→∞ bn 2
n=1 n=1

∞
∑ ∞ ( )n
∑ 1 an
26. Converges. The convergent geometric series bn = can be used for comparison: lim = 3.
n=1 n=1
2 n→∞ bn
∞
∑ ∞ ( )n
∑ 2 an
27. Converges. The convergent geometric series bn = can be used for comparison: lim = 1.
n=1 n=1
3 n→∞ bn
∞
∑ ∑∞
1 an π
28. Converges. The convergent p-series bn = 2
can be used for comparison: lim = lim arctan(n) = .
n=1 n=1
n n→∞ bn n→∞ 2
∞
∑ ∑∞
1 an
29. Diverges. The (divergent) harmonic series bn = can be used for comparison: lim = 1.
n=1 n=1
n n→∞ bn

30. Diverges. Observe that the denominator of an is comparable to n5/4 and so an is comparable to
∑∞ ∑∞
n1/3 1 1 an
bn = 5/4 = 11/12 . The divergent p-series bn = 11/12
can be used for comparison: lim = 1.
n n n n→∞ bn
n=1 n=1

∞
∑ ∑∞
1 an
31. Converges. The convergent p-series bn = 2
can be used for comparison: lim = 1.
n=1 n=1
n n→∞ bn
∞
∑ ∑∞ ( )
1 an ln(n)
32. Converges. The convergent p-series bn = can be used for comparison: lim = lim 1 + = 1 + 0.
n=1 n=1
n2 n→∞ bn n→∞ n
200

∞
∑ ∑∞
1
33. Diverges. The (divergent) harmonic series bn = can be used for comparison:
n=1 n=1
n
an 1 1
lim = lim = .
n→∞ bn n→∞ 1 + ln(n)/n 1+0

∞
∑ ∞ ( )n
∑ 2
34. Converges. The convergent geometric series bn = can be used for comparison:
3
( ) n=1 n=1
an 1 + n3 /2n 1+0
lim = lim = .
n→∞ bn n→∞ 1 + n2 /3n 1+0

∞
∑ ∑∞ ( )n
7
35. Converges. The convergent geometric series bn = can be used for comparison:
10
( ) n=1 n=1
an n10
lim = lim 1 + n = 1.
n→∞ bn n→∞ 7

∞
∑ ∑∞ ( )n
2
36. Converges. The convergent geometric series bn = can be used for comparison:
3
( ) n=1 n=1 ( )n
an 1 an 1 + 5/(2 n)
lim = exp . Remark: The limit of an /bn might initially seem to be 1 because = ,
n→∞ bn 6 bn 1 + 7/(3 n)
which has a base tending to 1. After closer inspection, we see that the limit is the indeterminate form
1∞ and must be calculated more carefully (using L’Hôpital’s Rule, for example).

37. Diverges. We calculate

∫ ∞ ∫ N
x x
2
dx = lim dx
1 x +1 N →∞ 1 +1 x2
1 ( ) N
= lim ln x2 + 1 1
2 N →∞
1 ( ) ln(2)
= lim ln N 2 + 1 −
2 N →∞ 2
= ∞.

38. Converges. Using limN →∞ arctan(N ) = π/2, we calculate

∫ ∞ ∫ N ( ) ( )
1 1 1 1 1π 1π π
2
dx = lim 2
dx = lim arctan(N ) − arctan(1) = − = .
1 x +4 N →∞ 1 x +4 N →∞ 2 2 22 24 8
201

39. Converges.
∫ ∞ ∫ N
1 1
dx = lim dx
1 x(x + 2) N →∞ 1 x(x + 2)
∫ N ( )
1 1 1
= lim − dx
2 N →∞ 1 x x+2
1 N
= lim (ln (|x|) − ln (|x + 2|))|1
2 N →∞
1 ln(3)
= lim (ln(N ) − ln(N + 2)) +
2 N →∞ ( ) 2
1 N ln(3)
= lim ln +
2 N →∞ N +2 2
( )
1 1 ln(3)
= lim ln +
2 N →∞ 1 + 2/N 2
ln(3)
= 0+ .
2
40. Converges.
∫ ∞ ∫
( ) N ( )
x exp −x3 dx =
2
lim x2 exp −x3 dx
1 N →∞ 1
1 ( )N
= − lim exp −x3 1
3 N →∞
1 ( )
= lim exp −N 3 + exp (−1)
2 N →∞
= exp (−1) .
( 3)
Remark:
( The
) hypotheses
( ) of the Integral Test are satisfied. If f (x) = x2
exp −x , then f ′ (x) =
x 2 − 3x3 exp −x3 . For x ≥ 1, the first and last factors of f ′ (x) are positive and the middle factor
is negative. Therefore f ′ (x) < 0 for x ≥ 1, and f (x) is decreasing, as is required for the Integral Test to
be applicable.
41. Diverges.
∫ ∞ ∫ N
2x2 2x2
3
dx = lim dx
2 x +4 N →∞ 2 +4 x3
2 ( ) N
= lim ln x3 + 4 2
3 N →∞
2 ( ( ) )
= lim ln N 3 + 4 − ln(12)
3 N →∞
= ∞.
( )
Remark: The hypotheses of the Integral Test are satisfied. If f (x) = 2x2 / x3 + 4 , then f ′ (x) =
( )( )−2
−2x x3 − 8 x3 + 4 . For x ≥ 2, the factors of f ′ (x) involving x are all nonnegative. Therefore
′
f (x) ≤ 0 for x ≥ 2, and f (x) is decreasing, as is required for the Integral Test to be applicable.
42. Diverges.
∫ ∞ ∫ N
1 1
dx = lim dx
2 x ln (x) N →∞ 2 x ln (x)
( ) N
= lim ln ln(x) 1
N →∞
= ∞.
202

Remark: For this exercise, the Integral Test is the best test we have available.
43. Converges.
∫ ∞ ∫ N
1 1
dx = lim dx
2 x ln2 (x) N →∞2 x ln2 (x)
( )N
1
= − lim
N →∞ ln(x) 1
= 1.

Remark: For this exercise, the Integral Test is the best test we have available.

44. Converges. By making the substitution u = exp(x), we obtain the integral formula
∫
exp(x) 1
2
dx = − + C.
(1 + exp(x)) 1 + exp(x)
∫ ∞
exp(x)
From this we deduce that the improper integral dx is convergent.
1 (1 + exp(x))2

Remark: The hypotheses of the Integral Test are satisfied. If f (x) = exp(x)/(1 + exp(x))2 , then
f ′ (x) = − exp(x) (−1+exp(x))/(1+exp(x))3 . For x ≥ 1, all factors of f ′ (x) involving x are nonnegative.
Therefore f ′ (x) ≤ 0 for x ≥ 1, and f (x) is decreasing, as is required for the Integral Test to be applicable.

45. Converges. Integrating by parts, we find

∫
(1 + x ln(10)) 10−x
x 10−x dx = − + C.
ln2 (10)
∫ ∞
x
Using this formula we see that the improper integral x
dx is convergent. Remark: The hypotheses
1 10
of the Integral Test are satisfied. If f (x) = x/10x , then f ′ (x) = (1 − x ln(10)) /10x . For x ≥ 1, the
numerator of f ′ (x) is nnegative. Therefore f ′ (x) ≤ 0 for x ≥ 1, and f (x) is decreasing, as is required
for the Integral Test to be applicable.

46. Divergent. The integral formula ∫

ln (x) 1
dx = ln2 (x) + C,
x 2
∫ ∞
ln (x)
which we obtain by making the substitution u = ln(x), shows that the improper integral dx
1 x
is divergent.
∫ ∞
8
47. Converges. We integrate by parts to find .
x exp (−x/2) dx =
2 e
Remark: The hypotheses of the Integral Test are satisfied. If f (x) = x exp(−x/2), then f ′ (x) =
(1 − x/2) exp(−x/2). For x ≥ 2, the first factor of f ′ (x) is nonpositive. Therefore f ′ (x) ≤ 0 for x ≥ 2,
and f (x) is decreasing, as is required for the Integral Test to be applicable.
∫ ∫ ∞
1 1 π
48. Converges. Use the integral formula √ dx = arcsec(x) + C to obtain √ dx = .
x x −12
2 x x −12 6
Solutions: Alternating Series
∑∞ n+1
In each of Exercises 1–6 a convergent alternating series S = n=1 (−1) an is given. It is easy to see
that, because the terms an decrease to 0, the Alternating Series Test can be∑ applied to each of these series,
N n+1
establishing convergence. Find a value of N such that the partial sum SN = n=1 (−1) an approximates
the given infinite series to within 0.01. That is, find N so that |S − SN | ≤ 0.01 holds. The point of these
exercises

1. N = 9999
In this exercise and the others in this group, the solution is obtained by finding the smallest index ν for
which anu ≤ 0.01. For N = ν − 1 we then have

|S − SN | ≤ aN +1 = aν ≤ 0.01.
√
We therefore must find the least ν such that 1/ ν ≤ 0.01 = 1/100, or
sqrtν ≥ 100, or ν = 1002 = 10000. Thus, we can be certain that N = 10000 − 1 = 9999 does the
job. We can be sure that every larger index results in the required accuracy, but, if that accuracy
suffices, why would we add up more terms than are necessary for that accuracy? In the other direction,
what can we say about partial sums SK with K < 9999? Nothing at all with certainty! There may
well be indices smaller than 9999 that result in the desired accuracy, but theory does not allow us to
state for certain that there are such smaller indices. In summary, N = 9999 provides the required
accuracy and is the smallest index that theory ensures us does the job. It turns out that there are
indices much smaller than 9999 that do the job: S = 0.6048986434 and S2500 = 0.5948996442, so
|S − S2500 | = 0.0099989992 < 0.01. Thus, we see that 2500 does the job, and 2500 is much smaller
than 9999. The problem is, we can find 2500 only by performing a tremendous amount of numerical
computation (including the calculation of S to more accuracy than the exercise required): we have no
theoretical basis for knowing that S2500 is sufficiently accurate.
2. N = 4
3. N = 200
4. N = 4
5. N = 4
6. N = 29

7. M = 10
8. M = 4
9. M = 7
10. M = 7

203
204

11. 1 − 1/2 + 1/6 − 1/24 + 1/120 − 1/720 ≈ 0.63194

12. 1/10 − 3/200 + 1/600 ≈ 0.08667
Solutions: The Ratio Test and the
Root Test
∑∞
n
1. n
converges:
n=1
e
( )
an+1 n + 1 en 1 1 1
lim = lim = lim 1 + = < 1.
n→∞ an n→∞ n en+1 e n→∞ n e

∑∞
n2
2. converges:
n=1
2n
( )2 ( )2
an+1 n+1 2n 1 1 1
lim = lim = lim 1 + = < 1.
n→∞ an n→∞ n 2n+1 2 n→∞ n 2

∑∞
2n
3. diverges:
n=1
n3
( )3 ( )3
an+1 n 2n+1 1
lim = lim = 2 lim = 2 > 1.
n→∞ an n→∞ n+1 2n n→∞ 1 + 1/n

∑∞
10n
4. converges:
n=1
n!

an+1 n! 10n+1 n! 1
lim = lim = 10 lim = 10 lim = 0 < 1.
n→∞ an n→∞ (n + 1)! 10n n→∞ (n + 1) · n! n→∞ (n + 1)

∑∞
n100
5. converges:
n=1
n!
( )100 ( )100 ( )100
an+1 n! n+1 n! 1 1 1
lim = lim = lim 1+ = lim 1+ = 0×1100 = 0 < 1.
n→∞ an n→∞ (n + 1)! n n→∞ (n + 1) · n! n n→∞ (n + 1) n
∞
∑ n!
6. diverges:
n=1
n5 · 7n
( )5 ( )5
an+1 (n + 1)! n 7n 1 (n + 1) · n! 1 1 15
lim = lim = lim = lim (n + 1) = ∞ > 1.
n→∞ an n→∞ n! n+1 7n+1 7 n→∞ n! 1 + 1/n 7 (1 + 0)5 n→∞

205
206

∑∞
3n + n
7. diverges:
n=1
2n + n3

an+1 3n+1 + n + 1 2n + n3
lim = lim n
× n+1
n→∞ an n→∞ 3 +n 2 + (n + 1)3
3 + (n + 1)/3n 1 + n3 /2n
= lim n
×
n→∞ 1 + n/3 2 + (n + 1)3 /2n
3+0 1+0
= ×
1+0 2+0
3
=
2
> 1.

∑∞
2n
8. √ converges:
n=1 n!
√ √
an+1 n! 2n+1 n! 1
lim = lim √ = 2 lim = 2 lim √ = 0 < 1.
n→∞ an n→∞ (n + 1)! 2n n→∞ (n + 1) · n! n→∞ n+1

9. Let p be any exponent and let c be any constant. Then, using L’Hôpital’s Rule, we have
( )p ( )p ( )p ( )p
c lnp (n + 1) ln(n + 1) 1/(n + 1) n 1
lim = lim = lim = lim = lim = 1.
n→∞ c lnp (n) n→∞ ln(n) n→∞ 1/n n→∞ n+1 n→∞ 1 + 1/n

In a similar way, if q is any exponent, we have

( )q ( )q
c (n + 1)q n+1 1
lim = lim = lim 1+ = 1.
n→∞ c nq n→∞ n n→∞ n

lnp (n)
Thus, if an = , then
nq
an+1 lnp (n + 1) nq
lim = lim p × = 1 × 1 = 1.
n→∞ an n→∞ ln (n) (n + 1)q

Using p = q = 1 for this example, we see that the Ratio Test is inconclusive. Using the Alternating
∑∞
ln(n)
Series Test, we see that the given series is convergent. Its series of absolute values, , is seen to
n=2
n
be divergent by using the Comparison Test with the harmonic series for comparison. Hence, the given
series is conditionally convergent.
∞
1 ∑ n 1
10. First, simplify the given series, writing it as (−1) . Use the calculation of Exercise 9 with
2 n=2 ln(n)
p = −1 and q = 0 to see that the Ratio Test is inconclusive. Using the Alternating Series Test, we see
∞
1 ∑ 1
that the given series is convergent. Its series of absolute values, , is seen to be divergent
2 n=2 ln(n)
by using the Comparison Test with the harmonic series for comparison. Hence, the given series is
conditionally convergent.
207

11. Use the calculation of Exercise 9 with p = 1 and q = 2 to see that the Ratio Test is inconclusive.
L’Hôpital’s Rule shows that, for every q > 0,

lnp (n)
lim = 0.
n→∞ nq
By taking p = 1 and q = 1/2, we obtain

ln(n)
lim √ = 0.
n→∞ n

It follows that there exists and M with

ln(n)
√ <1
n
√
for all n with n ≥ M . That is, n > ln(n) for n ≥ M . Therefore
√
1 n ln(n)
bn = 3/2 = 2 >
n n n2
∞
∑ ∑∞
1 ln (n)
for all n with n ≥ M . Using the convergent p-series for comparison, we deduce that
n 3/2
n=1
n2
n=M
is convergent. Therefore, the given series is absolutely convergent.
∑∞ ∑∞
n+1 1
12. The series 3+1
is seen to be convergent by using the Limit Comparison Test with 2
for the
n=1
n n=1
n
comparison series. Therefore, the given series is absolutely convergent.

13. The given series is divergent by the Divergence Test.

14. Using the Alternating Series Test, we see that the given series is convergent. Its series of absolute
∑∞ √
n
values, , is seen to be divergent by using the Limit Comparison Test with the comparison
n=1
n+4
∑∞
1
series √ . Hence, the given series is conditionally convergent.
n=1
n

∑∞
en + ln(n)
15. The series of absolute values, , is seen to be conergent by using the Comparison Test with
n=1
e n · n2
∑∞
2
the comparison series 2
:
n=1
n

en + ln(n) en + en 2 en 2
< n 2 = n 2 = 2.
e ·n
n 2 e ·n e ·n n
Hence, the given series is absolutely convergent.

16. Because
n! 1 1
lim = lim = = 1,
n→∞ n! + 2n n→∞ 1 + 2n /n! 1+0
the given series is divergent by the Divergence Test.

17. Because limn→∞ arctan(n) = π/2, the given series is divergent by the Divergence Test.
208

∞
∑ ( )
1
18. The given series is convergent by the Alternating Series Test. The series of absolute values, sin ,
n=1
n
∑∞
1
is divergent by the Limit Comparison Test with for the comparison series. Hence, the given series
n=1
n
is conditionally convergent.
19. Because ( )1/n
1
lim n−n/2 = lim
= 0 < 1,
n→∞ n1/2 n→∞
∑∞
the given series n=1 n−n/2 converges by the Root Test.
20. Because ( )1/n
23n 23 8
lim = lim = < 1,
n→∞ 32n n→∞ 3 2 9
∞
∑ 3n
2
the given series converges by the Root Test.
n=1
32n

21. Because
( n )1/n n1/n 1 1
lim = lim = lim n1/n = < 1,
n→∞ 2n n→∞ 2 2 n→∞ 2
∑∞
n
the given series converges by the Root Test.
n=1
2n

22. Because ( )1/n

10n 10 10
lim = lim =( )10 = 10 > 1,
n→∞ n10 n→∞ n10/n limn→∞ n1/n
∑∞
10n
the given series diverges by the Root Test.
n=1
n10

23. Because ( )7
( )1/n
n7 limn→∞ n1/n 17
lim = = = 0 < 1,
n→∞
n
ln (n) limn→∞ ln(n) ∞
∞
∑ n7
the given series n converges by the Root Test.
n=2
ln (n)

24. Using L’Hôpital’s Rule, we have

(( )n )1/n
ln(n + 4) ln(n + 4)
lim = lim
n→∞ ln (n2 + 4) n→∞ ln (n2 + 4)
1/(n + 4)
= lim
n→∞ 2n/ (n2 + 4)

n2 + 4
= lim
n→∞ 2n(n + 4)

1 + 4/n2
= lim
n→∞ 2 + 8/n
1
=
2
< 1.
209

∑∞ ( )n
ln (n + 4)
Therefore, the given series converges.
n=1
ln (n2 + 4)

25. Because (( )n )1/n

37 37 37
lim = lim = = 0 < 1,
n→∞ n n→∞ n ∞
∞ (
∑ )n
37
the given series converges by the Root Test.
n=1
n
an+1
26. Absolutely convergent by the Rato Test: if an = (2n)!/(3n)!, then lim = 0 < 1.
n→∞ an
an+1 1
27. Absolutely convergent by the Rato Test: if an = (n!)2 /(2n)!, then lim = < 1.
n→∞ an 4
28. Divergent by the Divergence Test:
n + 3n n/2n + (3/2)n 0+∞
lim = lim = = ∞ ̸= 0.
n→∞ n3 + 2n n→∞ n3 /2n + 1 0+1
29. Divergent by the Root Test:
(( )n )1/n ( ) 3
lim n1/n + 1/2 = lim n1/n + 1/2 = > 1.
n→∞ n→∞ 2
( ) 1/n
2n 2
30. Absolutely convergent by the Root Test: lim = = 0 < 1.
n
n→∞ 1 + ln (n) ∞
( )1/n
exp (n) e
31. Absolutely convergent by the Root Test: lim = lim = 0 < 1.
n→∞ lnn (n) n→∞ ln(n)

32. Conditionally convergent. The given series is convergent by the Alternating Series Test. The series
of absolute values diverges by the Limit Comparison Test using the divergent p-series with p = 1/2 for
comparison.
n!
33. Divergent by the Divergence Test: lim = ∞ ̸= 0.
n→∞ 3n
34. Conditionally convergent
35. Absolutely convergent. Use the Integral Test
∫ ∞
1 1
3 dx = 2 .
2 x ln (x) 2 ln (2)
36. Conditionally convergent
37. Absolutely convergent. Use the limit Comparison Test to compare the series of absolute values with
the convergent p-series with p = e > 1.
38. Conditionally convergent The given series has the behavior of the alternating harmonic series because
ln (1 + 1/n)
lim = 1.
n→∞ 1/n
n
39. Divergent by the Divergence Test: lim (1 + 1/n) = e.
n→∞

40. Conditionally convergent

210
Solutions: Power Series, Intervals of
Convergence (Corresponds to Stewart
11.8)

1. (−3, 3)

2. (−1, 1)

3. [−1, 1)

4. (−1/10, 1/10]

5. [−1/4, 1/4]

6. (−5/2, 3/2)

7. (−∞, ∞)

8. (−1, 1]

9. [−1/2, 3/2)

10. [−7/3, −5/3]

11. (−2, 1)

12. [−7, −5)

13. (π − 1, π + 1]

14. [−5, −3]

15. (−1, −1/3)

16. (−3, −2)

( )2
17. For u = x − (−3) , we have

∑∞ ∑∞
4n 2n 4n
(x + 3) = un .
n=0
n+1 n=0
n+1

211
212

For this power series in u, we have c = 0 and an = 4n /(n + 1). It follows that
an
R = lim
n→∞ an+1
4n n + 2
= lim
n→∞ n + 1 4n+1
4n n + 2
= lim n+1
n→∞ 4 n+1
1 1 + 2/n
= lim
n→∞ 4 1 + 1/n
1
= .
4
For the left endpoint, set u = c − R = 0 − 1/4 = −1/4 in the power series:

∑∞
4n ∑∞
4n −(1)n ∑∞
1
un = = −(1)n ,
n + 1 n + 1 4 n n + 1
n=0 u=−1/4 n=0 n=0

which is convergent by the Alternating Series Test. For the right endpoint, set u = c+R = 0+1/4 = −1/4
in the power series:
∑∞
4n ∑∞
4n 1 ∑∞
n 1
u = = ,
n=0
n+1 n+1 4
n=0
n
n=0
n+1
u=1/4

which is a tail of a divergent p-series. It follows that [−1/4, 1/4) is the interval of convergence for the
( )2
powers series in u. Thus, the given power series in x converges for −1/4 ≤ x − (−3) < 1/4, which
( )2
is the same interval as 0 ≤ x − (−3) < 1/4, or |x − (−3)| < 1/2, or −1/2 < x − (−3) < 1/2, or
−3 − 1/2 < x < −3 + 1/2, or −7/2 < x < −5/2. In interval notation, the answer is (−7/2, −5/2).

18. [−1/2, 5/2)

Solutions: Series Related to 1/(1 ± x)
(Corresponds to Stewart 11.8)

1. With u = x/2 we have

1 1 1
= ×
2+x 2 1+u
1 ( )
= 1 − u + u2 − u3 + u4 − u5 + · · ·
2
1 1 1 1 3 1 4
= − x + x2 − x + x ± ···
2 4 8 16 32

2. With u = 2x we have
1 1
=
1 − 2x 1−u
= 1 + u + u2 + u3 + u4 + u5 + · · ·
= 1 + 2x + 4x2 + 8x3 + 16x4 + 32x5 + · · ·

3. With u = −3x/4 we have

x x 1
=
4 + 3x 4 1−u
x( )
= 1 + u + u2 + u3 + u4 + u5 + · · ·
4( )
x 3 9 2 27 3 81 4
= 1− x+ x − x + x − ···
4 4 16 64 256
1 3 9 27 4 81 5
= x − x2 + x3 − x + x − ···
4 16 64 256 1024

4. With u = −x2 /4 we have

1 1 1
=
4 + x2 4 1−u
1( )
= 1 + u + u2 + u3 + u4 + u5 + · · ·
4( )
1 1 1
= 1 − x2 + x4 − · · ·
4 4 16
1 1 2 1 4
= − x + x − ···
4 16 64

213
214

5. With u = x2 we have
1 1
= = 1 + u + u2 + u3 + u4 + · · · = 1 + x2 + x4 + · · ·
1 − x2 1−u
Therefore,
1 + x2 ( )( )
= 1 + x2 1 + x2 + x4 + · · · = 1 + 2x2 + 2x4 + · · ·
1 − x2

6. With u = 3 x2 /2 we have

x x 1 x ( ) 1 3 9 27 7
2
= · = 1 − u + u2 − u3 + u4 + · · · = x − x3 + x5 − x ± ···
2 + 3x 2 1+u 2 2 4 8 16

7. We have
1 1 1
= ·
(2 + x)2 2+x 2+x
( )( )
1 1 1 1 3 1 4 1 1 1 1 3 1 4
= − x + x2 − x + x ± ··· − x + x2 − x + x ± ···
2 4 8 16 32 2 4 8 16 32
( ) ( )
1 1 1 1 1 1 1 1 1 1 1 1
= · + − · − · x+ · + · + · x2
2 2 2 4 4 2 2 8 4 4 8 2
( )
1 1 1 1 1 1 1 1
+ − · − · − · − · x3 + · · ·
2 16 4 8 8 4 16 2
1 1 3 2 1 3
= − x+ x − x ± ··· .
4 4 16 8

8. We have
( )
1 d 1
= −
(2 + x)2 dx 2 + x
( )
d 1 1 1 1 3 1 4
= − − x + x2 − x + x ± ···
dx 2 4 8 16 32
( )
1 1 3 2 1 3
= − 0− + x− x + x ± ···
4 4 16 8
1 1 3 2 1 3
= − x+ x − x ± ··· .
4 4 16 8

9. We have
1 1 1
= ·
2 − x − x2 1−x 2+x
( )
( ) 1 1 1 1 3 1 4
= 1 + x + x2 + x3 + x4 + · · · − x + x2 − x + x ± ···
2 4 8 16 32
( ) ( ) ( )
1 1 1 1 1 1 1 1 1 1
= + − + x+ − + x2 + − + − + x3 + · · ·
2 4 2 8 4 2 16 8 4 2
1 1 3 5 3
= + x + x2 + x + ··· .
2 4 8 16
215

10.
∫
1
ln(2 + x) = dx
2+x
∫ ( )
1 1 1 2 1 3 1 4
= − x+ x − x + x ± · · · dx
2 4 8 16 32
1 1 1 3 1 4 1 5
= C + x − x2 + x − x + x ± ···
2 8 24 64 160
Setting x = 0, we see that C = ln(2).
11. The method of multiplication follows problems already written up. We turn to the second method:
( )
x + 2 x2 + 3 x3 + 4 x4 + · · · = (2 + x) a0 + a1 x + a2 x2 + a3 x3 + a4 x4 + · · ·
= 2 a0 + (2 a1 + a0 ) x + (2 a2 + a1 ) x2 + (2 a3 + a2 ) x3 + (2 a4 + a3 ) x4 + · · · .

Equating like coefficients of powers of x on each side, we obtain the equations 0 = 2 a0 , 1 = 2 a1 +a0 , 2 =
2 a2 + a1 , 3 = 2 a3 + a2 , 4 = 2 a4 + a3 . One by one, we solve a0 = 0, a1 = 1/2, a2 = 3/4, a3 = 9/8, a4 =
23/16. That is q(x) = x/2 + 3x2 /4 + 9x3 /8 + 23x4 /16 + · · · .
12. We have
( )( )
2+x = 2 + x + 2 x2 + 3 x3 + · · · a0 + a1 x + a2 x2 + a3 x3 + · · ·
= 2a0 + (2a1 + a0 ) x + (2a2 + a1 + 2a0 ) x2 + (2a3 + a2 + 2a1 + 3a0 ) x3 + · · · .

Writing the left side of this equation as 2 + x + 0x2 + 0x3 and equating like powers of x on the two sides
of the equation, we obtain 2 = 2a0 , 1 = 2a1 + a0 , 0 = 2a2 + a1 + 2a0 , 0 = 2a3 + a2 + 2a1 + 3a0 . We solve
successively, a0 = 1, a1 = 0, a2 = −1, a3 = 0. Thus, (2 + x)/f (x) = 1 − x2 + · · · . If you wish to carry
the calculation further, you will get (2 + x)/f (x) = 1 − x2 − x3 − x4 /2 + x5 /4 + · · · .

13. With u = −x/2 we have

( )
ln(2 − x) = ln 2(1 + u)
= ln(2) + ln(1 + u)
1 1 1
= ln(2) + u − u2 + u3 − u4 + · · ·
2 3 4
1 1 2 1 3 1
= ln(2) + x − x − x − x4 − · · ·
2 8 24 64
14. With u = 4x2 /3 we have
( )
x ln(3 + 4x2 ) = x ln 3(1 + u)
( )
= x ln(3) + ln(1 + u)
( )
1 1 1
= x ln(3) + u − u2 + u3 − u4 + · · ·
2 3 4
( )
4 16
= x ln(3) + x2 − x4 + · · ·
3 18
4 3 8 5
= ln(3) · x + x − x + · · ·
3 9
216
Solutions: Taylor Polynomials
(Corresponds to Stewart 11.9)

1.
1 1 1
T3 (x) = 2 + (x − 4) − (x − 4)2 + (x − 4)3
4 64 512

2.
T4 (x) = 2 − 4 (x − 1/2) + 8 (x − 1/2)2 − 16 (x − 1/2)3 + 32 (x − 1/2)4

3.
1
T4 (x) = −(x − π) + (x − π)3
6

4.
1 1 1 1
T4 (x) = 1 + (x + 2) + (x + 2)2 + (x + 2)3 + (x + 2)4
2 8 48 384

5.
3 15 35
T3 (x) = 1 − (x + 3) + (x + 3)2 − (x + 3)3
2 8 16

6.
7 111
T2 (x) = 2 − (x − 1) + (x − 1)2
4 64

7. If a Taylor polynomial of a certain order has been calculated, then one need not begin at the beginning
in order to calculate a higher order Taylor polynomial. If TN is the Taylor polynomial of order N of a
function f with base point c, then the order N + 1 Taylor polynomial of f with base point c is given by

f (N +1) (c)
TN +1 (x) = TN (x) + (x − c)N +1 .
(N + 1)!

The order 1 Taylor polynomial T1 of cos(x) with base point π/3 is written out in the following figure.
The graph of T1 is the tangent line to the graph of y = cos(x) at (π/3, cos(π/3)).

217
218

The order 2 Taylor polynomial T2 of cos(x) with base point π/3 is written out in the following figure.
You will note that T2 is obtained from T1 by adding one more term. You will also note that the addition
of this term results in a quadratic function that approximates the graph of cos(x) better than the linear
graph of T1 .

The order 3 Taylor polynomial T3 of cos(x) with base point π/3 is written out in the following figure.
You will note that T3 is obtained from T2 by adding one more term. You will also note that the addition
of this term results in a cubic polynomial that approximates the graph of cos(x) better than the graph
of the quadratic polynomial T2 .
219

The order 6 Taylor polynomial T6 of cos(x) with base point π/3 is graphed in the following figure. This
exercise did not request the calculation of T6 . It is graphed here to show that Taylor polynomials can
be very accurate approximations of their generating function if the order is sufficiently large.
220
Solutions: Taylor and Maclaurin
Series (Corresponds to Stewart 11.9)

1. ( ) ( )
sin x2 1 ( 2 ) 1 ( 2 )3 1
= x − x + · · · = x − x5 + · · ·
x x 3! 6

2.
1 − cos (2x) 1( ( )) 2 4
= 1 − 1 − (2x)2 /2 + (2x)4 /24 − (2x)6 /720 + · · · = 2x − x3 + x5 − · · ·
x x 3 45

3.
x ( )
= x exp −x2
exp (x2 )
( )
( 2 ) 1 ( 2 )2
= x 1 + −x −x + ···
2!
1
= x − x3 + x5 + · · ·
2

4.
( ) ( )
exp x2 + 2x = exp x2 · exp (2x)
( )( )
( 2 ) 1 ( 2 )2 1 ( 2 )3 1 1 1
= 1+ x + x + x + ··· 1 + (2x) + (2x) + (2x) + (2x) + · · ·
2 3 4
2! 3! 2! 3! 4!
( )( )
1 4 2
= 1 + x2 + x4 + · · · 1 + 2x + 2x2 + x3 + x4 + · · ·
2 3 3
10 3 19 4
= 1 + 2x + 3x + x + x + · · ·
2
3 6

5. Use u = 2x2 and α = −1/2:

3
1 − x2 + x4 − · · ·
2

6. Write (8 + x3 )1/3 = 2(1 + u)α with u = x3 /8 and α = 1/3:

1 3
2+ x + ···
12

221
222

7.
∫ ∫ ( )
0.2 ( ) 0.2
1 6 1
sin x2 dx = x2 − x + x1 0 − · · · dx
0 0 3! 5!
( )0.2
1 3 1 7 1 1
= x − x + x 1 − · · ·
3 42 1320 0
(0.2)3 (0.2)7 (0.2)1 1
= − + − ···
3 42 1320
(0.2)7
We calculate = 0.0000003... < 10−5 . Therefore, the error resulting from dropping the second and
42 ∫ 0.2
( ) (0.2)3
all subsequence terms is less than 10−5 . In other words, sin x2 dx ≈ = 0.0027 with an
0 3
error less than 10−5 . (In fact, the given integral equals 0.002666361920 to 10 correct decimal places.)

8. We calculate,
∫ ∫
0.2
1 0.2 ( )
dx = 1 − x3 + x6 − x9 + · · · dx
0 1 + x3 0
( )1/5
x4 x7 x10
= x− + − + · · ·
4 7 10 0
(0.2)4 (0.2)7 (0.2)10
= (0.2) − + − + ···
4 7 10
= 0.2 − 0.0004 + 0.1828571429 × 10−5 − · · ·

Note that the third summand in the last line is less than 10−5 . Thus
∫ 0.2
1
dx ≈ 0.2 − 0.0004 = .1996
0 1 + x3

with an error less than 10−5 . In fact, the given integral has value 0.1996018184 to 10 correct decimal
places.
Solutions: Polar Coordinates
(Corresponds to Stewart 10.3)
√ √ √
1. P = (3 2/2, 3 2/2), Q = (0, 4), R = (−5, 5 3)
Alternative polar coordinates for P : (3, 9 π/4) and (−3, 5 π/4)
Alternative polar coordinates for Q: (4, π/2) and (−4, −π/2)
Alternative polar coordinates for R: (10, 5π/3) and (−10, 2π/3)
√ √
2. P = (3 2, 3π/4) and (−3 2, 7π/4)
Q = (4, π/3) and (−4, 4 π/3)
R = (2, 5 π/6) and (−2, 11 π/6)
√ √
√ Cartesian coordinates of the points are ( 3, −1) and (−3, −3 3). The distance between them is
3. The
2 10.
4.

5. (i) x = 5
(ii) y = 3/x
6.

223
224

7.

8.

9.

10.
225

11.

12.
226
Solutions: Areas in Polar Coordinates
(Corresponds to Stewart 10.4)

In the calculations that follow, we will use certain integration facts without elaboration: The integrals of
sin(θ) and cos(θ) are 0 over the θ-interval [0, 2π] and the integrals of sin(θ)2 and cos(θ)2 are π/4 over each of
the θ-intervals [0, π/2], [π/2, π], [π, 3π/2] and [3π/2, 2π].

1. Figure for Exercise 1:

The area A is given by

∫ 2π
1 2
A = (1 + sin(θ)) dθ
2 0
∫
1 2π ( )
= 1 + 2 sin(θ) + sin(θ)2 dθ
2 0
1 ( π)
= 2π + 0 + 4 ·
2 4
3π
= .
2

2. Figure for Exercise 2:

227
228

The area A is given by

∫ π
1 π3
A= θ2 dθ = .
2 0 6

3. Figure for Exercise 3:

Using symmetry, we see that the requested area A is

( ∫ ) ∫ ( )π/2
1 π/2 (√ )2 π/2
1 1
A=2 sin(2 θ) dθ = sin(2 θ) dθ = − cos(2 θ) = .
2 0 0 2 0 2

4. Figure for Exercise 4:

229

Using symmetry, and making the substitution u = 2θ, du = 2 dθ with limits of integration u = 0
corresponding to θ = 0 and u = π corresponding to θ = π/2, we see that the requested area A is
( ∫ ) ( ∫ )
π/2 π
1 1 1 π π
A=4 2
sin(2 θ) dθ =4 sin(u)2 du =4· ·2· = .
2 0 4 0 4 4 2

5. Figure for Exercise 5:

The requested area A is given by

∫ ( √ )2 ∫ π
−1
π
1 sin(θ) 1 π sin(θ) 1 1
A= dθ = dθ = − (5 − 4 cos(θ)) = .
2 0 5 − 4 cos(θ) 2 0 (5 − 4 cos(θ))2 8 0 9

6. Figure for Exercise 6:

230

A polar curve passes through the origin when r = 0. The solutions of r = cos(3 θ) = 0 in the first and
fourth quadrants are θ = π/6 and θ = −π/6. The polar graphs of these equations are the black rays
in the figure. (They are tangent to the rose.) Making the change of variable u = 3 θ, du = 3 dθ, we see
that the requested area is

∫ π/6 ∫ π/2
1 1 1 π π
A= 2
cos(3 θ) dθ = cos(u)2 du = ·2· = .
2 −π/6 6 −π/2 6 4 12

7. Figure for Exercise 7:

√
The slope of the given line segment is −1/ 3, or tan(−π/6). The graph of the given line segment
231

therefore has polar equation θ = −π/6, 0 ≤ r ≤ 1 − sin(−π/6). The requested area A is given by
∫ π/2
1 2
A = (1 − sin(θ)) dθ
2 −π/6
∫
1 π/2 ( )
= 1 − 2 sin(θ) + sin2 (θ) dθ
2 −π/6
∫ ( π/2 )
1 1
= 1 − 2 sin(θ) + (1 − cos(2θ)) dθ
2 −π/6 2
∫ π/2 ( )
3 1
= − sin(θ) − cos(2θ) dθ
−π/6 4 4
( )π/2
3 1
= θ + cos(θ) − sin(2θ)
4 8 −π/6
( ) ( √ √ )
3π π 3 1 3
= + 0 + 0) − − + −
8 8 2 8 2
π
=
2

8. Figure for Exercise 8:

The points of intersection occur where 2 sin(θ) = r = 1, or sin(θ) = 1/2, or θ = π/6 and θ = 5 π/6: The
rays corresponding to these two polar equations have been drawn in black. The requested area is
∫ (( ) √
1 5 π/6 )2 π
A= 2 sin(θ) − 12 dθ = 3 − .
2 π/6 3

9. Figure for Exercise 9:

232

The points of intersection occur where 1 − sin(θ) = r = 1/2, or sin(θ) = 1/2, or θ = π/6 and θ = 5 π/6:
In setting up the integral, it will be convenient to use the angle θ = π/6 + 2 π instead of θ = π/6. The
rays corresponding to these two polar equations have been drawn in black. The requested area is
∫ ( ( )2 ) √
1 π/6+2 π ( )2 1 7 3 5π
A= 1 − sin(θ) − dθ = + .
2 5 π/6 2 8 6

10. Figure for Exercise 10:

Equate the two formulas for r to find the points of intersection: 3 cos(θ) = 1 + cos(θ), or cos(θ) = 1/2,
or θ = ±π/3. The requested area is
∫ (( )
1 π/3 )2 2
A= 3 cos(θ) − (1 + cos(θ)) dθ = π.
2 −π/3
233

11. Figure for Exercise 11:

(√ )2
Setting the two formulas for r2 equal to each other, we have 2 sin(θ) = sin(2 θ), or 2 sin(θ)2 =
2 sin(θ) cos(θ). For the nonzero solution divide each side by sin(θ) to obtain sin(θ) = cos(θ). This yields
θ = π/4, which is the equation for the line drawn in black in the figure. The area is
∫ ∫ (√ )2
1 π/4 (√ )2 1 π/2
π
2 sin(θ) dθ + sin(2 θ) dθ = .
2 0 2 π/4 8

12. Figure for Exercise 12:

√ √ √
3−1)/(1− 3), or -1. The Cartesian equation of the line is therefore y = −(x−1)+ 3,
a) The slope is ( √
or y = −x + 1 + 3. √
b) From part (a), we have√ x + y = 1 + √3. Replacing x with r cos(θ) and y with r sin(θ) results in
r cos(θ) + r sin(θ) = 1 + 3, or r = √
(1 + 3)/(cos(θ) + sin(θ).
c) Because cos(π/4) = sin(π/4) = 1/ 2, and in view of the addition formula sin(A+B) = sin(A) cos(B)+
234

cos(A) sin(B), we have

√
1+ 3
r =
sin(θ) + cos(θ)
√
1+ 3
= √
2 (sin(θ) cos(π/4) + cos(θ) sin(π/4))
√
1+ 3
= √
2 sin(θ + π/4)
√ (
1+ 3 π)
= √ csc θ + .
2 4

d) The requested area A is given by

 ( ) 
∫ π/3
√ ( ) 2
1 22 − 1 + 3 π  dθ
A = √ csc θ +
2 π/6 2 4
( √ )π/3
(1 + 3)2 ( π )
= 2θ + cot θ +
4 4
π/6
( )
π √ 5π
= − (2 + 3) cot .
3 12