Math 111

Calculus II
OPEN EDUCATIONAL RESOURCE

Robert Petry, Fotini Labropulu, and Iqbal Husain

LUTHER
COLLEGE
UNIVERSITY OF REGINA

published by Campion College and Luther College

3rd Edition Copyright © 2024 Robert G. Petry, Fotini Labropulu, Iqbal Husain.
2nd Edition Copyright © 2021 Robert G. Petry, Fotini Labropulu, Iqbal Husain.
1st Edition Copyright © 2020 Robert G. Petry, Fotini Labropulu.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free
Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no
Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license is included in the
section entitled “GNU Free Documentation License”.
Permission is granted to retain (if desired) the original title of this document on modified copies.

History

• 1st Edition produced in 2021 entitled “Calculus II” written by Robert G. Petry, Fotini Labropulu, and
Iqbal Husain. Published by Campion College and Luther College.
• 2nd Edition produced in 2020 entitled “Calculus II” written by authors Robert G. Petry and Fotini
Labropulu. Published by Campion College and Luther College.
• 3rd Edition produced in 2024 entitled “Calculus II” written by authors Robert G. Petry, Fotini Labropulu,
and Iqbal Husain. Published by Campion College and Luther College. The authors gratefully acknowledge
funding provided by the University of Regina Open Education and Publishing (OEP) program that
assisted in the typesetting of this edition.

The source of this document (i.e. a transparent copy) is available via

http://campioncollege.ca/resources/dr-robert-petry

About the cover: The cover line fractal drawing is in the public domain and available from
http://openclipart.org.
 2.5.2 Inverse Cosine . . . . . . . 47
2.5.3 Inverse Tangent . . . . . . 48
2.5.4 Other Trigonometric Inverses 49
2.6 L’Hôpital’s Rule . . . . . . . . . . 54
2.6.1 Indeterminate Forms of
Contents type 0 · ∞ and ∞ − ∞ . .
2.6.2 Exponential Indeterminate
56

Forms . . . . . . . . . . . 58
Review Exercises . . . . . . . . . . . . . . 61

3 Integration Methods 63
1 Integration Review 1 3.1 Integration by Parts . . . . . . . . 64
1.1 The Meaning of the Definite Integral 2 3.2 Trigonometric Integrals . . . . . . 68
1.2 The Fundamental Theorem of 3.3 Trigonometric Substitution . . . . 74
Calculus . . . . . . . . . . . . . . 3 3.4 Partial Fraction Decomposition . 78
1.3 Indefinite Integrals . . . . . . . . 3 3.5 General Strategies for Integration 89
1.4 Integration by Substitution . . . . 4 3.6 Improper Integrals . . . . . . . . 92
1.5 Integration Examples . . . . . . . 4 3.6.1 Improper Integrals of the
Review Exercises . . . . . . . . . . . . . 7 First Kind . . . . . . . . . 92
3.6.2 Improper Integrals of the
2 Inverses and Other Functions 9 Second Kind . . . . . . . .
95
2.1 Inverse Functions . . . . . . . . . 10 Review Exercises . . . . . . . . . . . . .
100
2.1.1 Horizontal Line Test . . . 10
2.1.2 Finding Inverse Functions 13 4 Sequences and Series 101
2.1.3 Graphs of Inverse Functions 15 4.1 Sequences . . . . . . . . . . . . . 102
2.1.4 Derivative of an Inverse 4.2 Series . . . . . . . . . . . . . . . . . 111
Function . . . . . . . . . . 15 4.3 Testing Series with Positive Terms 118
2.1.5 Creating Invertible Functions 18 4.3.1 The Integral Test . . . . . 118
2.2 Exponential Functions . . . . . . 19 4.3.2 The Basic Comparison Test 123
2.2.1 The Natural Exponential 4.3.3 The Limit Comparison Test 125
Function . . . . . . . . . . 20 4.4 The Alternating Series Test . . . 129
2.2.2 Derivative of ex . . . . . . 22 4.5 Tests of Absolute Convergence . . 132
2.2.3 Integral of ex . . . . . . . 23 4.5.1 Absolute Convergence . . 132
2.2.4 Simplifying Exponential 4.5.2 The Ratio Test . . . . . . 134
Expressions . . . . . . . . 24 4.5.3 The Root Test . . . . . . . 135
2.3 Logarithmic Functions . . . . . . 26 4.5.4 Rearrangement of Series . 137
2.3.1 Logarithmic Function 4.6 Procedure for Testing Series . . . 138
Properties . . . . . . . . . 27 4.7 Power Series . . . . . . . . . . . . 143
2.3.2 The Natural Logarithmic
4.8 Representing Functions with
Function . . . . . . . . . . 28
Power Series . . . . . . . . . . . . . 151
2.3.3 Solving Exponential and
4.9 Maclaurin Series . . . . . . . . . . 157
Logarithmic Equations . . 29
4.10 Taylor Series . . . . . . . . . . . . . 161
2.3.4 Derivative of the Natural
Review Exercises . . . . . . . . . . . . . 166
Logarithmic Function . . . 32
2.3.5 Derivatives Using Answers 169
Arbitrary Bases . . . . . . 33
2.3.6 Logarithmic Differentiation 34 Summary 183
2.3.7 Integral of x1 and ax . . . 36
2.4 Exponential Growth and Decay . 40 Index 185
2.5 Inverse Trigonometric Functions . 45
2.5.1 Inverse Sine . . . . . . . . 45 GNU Free Documentation License 187

iii
iv
Chapter 1: Integration Review

1
2 1.1 The Meaning of the Definite Integral

1.1 The Meaning of the Definite Integral

The definite integral of the function f (x) between x = a and x = b is written:

Z b
f (x) dx
a

Geometrically it equals the area A between the curve y = f (x) and the x-axis between the vertical
lines x = a and x = b:

y y = f (x)

x
a b

More precisely, assuming a < b, the definite integral is the net sum of the signed areas between the
curve y = f (x) and the x-axis where areas below the x-axis (i.e. where f (x) dips below the x-axis) are
counted negatively.
Rb
The notation used for the definite integral, a f (x) dx, is elegant and intuitive. We are umming ( dA)
R R

the (infinitesimally) small differential rectangular areas dA = f (x) · dx of height f (x) and width dx at
each value x between x = a and x = b:

y y = f (x)

dx

dA
A f (x)

x
a x b

We will see soon how viewing integrals as sums of differentials can be used to come up with formulas
for calculations aside from just area.
Integration Review 3

1.2 The Fundamental Theorem of Calculus

As seen in a previous calculus course, the definite integral can be written as a limiting sum (Riemann
Sum) of N rectangles of finite width ∆x = (b − a)/N where we let the number of rectangles (N ) go to
infinity (and consequently the width ∆x → 0). This method of evaluating a definite integral is hard or
impossible to compute exactly yfor most functions. An easy way to evaluate definite integrals is due to
the Fundamental Theorem of Calculus which relates the calculation of a definite integral with the
evaluation of the antiderivative F (x) of f (x):

Theorem 1-1: The Fundamental Theorem of Calculus:

If f is continuous on [a, b] then
Z b
f (x) dx = F (b) − F (a)
a

for any F an antiderivative of f , i.e. F ′ (x) = f (x).

b
Notationally we write F (b) − F (a) with the shorthand F (x)|a , i.e.
b
F (x)|a = F (b) − F (a) ,

where, unlike the integral sign, the bar is placed on the right.

1.3 Indefinite Integrals

Because of the intimate relationship between the antiderivative and the definite integral, we define the
indefinite integral of f (x) (with no limits a or b) to just be the antiderivative, i.e.
Z
f (x) dx = F (x) + C

where F (x) is an antiderivate of f (x) (so F ′ (x) = f (x)) and C is an arbitrary constant. The latter is
required since the antiderivative of a function is not unique as dx d
C = 0 implies we can always add a
constant to an antiderivative to get another antiderivative of the same function.
Using our notation for indefinite integrals and our knowledge of derivatives gives the following.

Table of Indefinite Integrals

1
Z Z
1. xn dx = xn+1 + C (n ̸= −1) 6. csc2 x dx = − cot x + C
n+1
Z Z
2. cos x dx = sin x + C 7. csc x cot x dx = − csc x + C
Z Z Z
3. sin x dx = − cos x + C 8. cf (x) dx = c f (x) dx
Z
4. sec2 x dx = tan x + C
Z Z Z
9. [f (x) ± g(x)] dx = f (x) dx ± g(x) dx
Z
5. sec x tan x dx = sec x + C
4 1.4 Integration by Substitution

In the last two integration formulae f (x) and g(x) are functions while c is a constant. For indefinite
integrals we say, for example, that n+1
1
xn+1 + C is the (indefinite) integral of xn where xn is the
integrand. The process of finding the integral is called integration. Each of these indefinite integrals
may be verified by differentiating the right hand side and verifying that the integrand is the result.

1.4 Integration by Substitution

The last two general integral results allow us to break up an integral of sums or differences into integrals
of the individual pieces and to pull out any constant multipliers. Another useful way of solving an
integral is to use the Substitution Rule which arises by working the differentiation Chain Rule in
reverse.
Theorem 1-2: Substitution Rule (Indefinite Integrals): Suppose u = g(x) is a differentiable
function whose range of values is an interval I upon which a further function f is continuous, then
Z Z
f (g(x))g (x) dx = f (u) du .
′

where the right hand integral is to be evaluated at u = g(x) after integration.

Here the du appearing on the right side is the differential:

du = g ′ (x)dx

which, recall, can be remembered by thinking du

dx = g ′ (x) and multiplying both sides by dx.
When using the Substitution Rule with definite integrals we can avoid the final back-substitution of
u = g(x) of the indefinite case by instead just changing the limits of the integral appropriately to the
u-values corresponding to the x-limits:
Theorem 1-3: Substitution Rule (Definite Integrals): Suppose u = g(x) is a differentiable
function whose derivative g ′ is continuous on [a, b] and a further function f is continuous on the range
of u = g(x) (evaluated on [a, b]), then
Z b Z g(b)
f (g(x))g ′ (x) dx = f (u) du .
a g(a)

1.5 Integration Examples

Example 1-1
Evaluate the following integrals:

1
Z
x
Z  
1. x + 2 − 3 cos x dx
2 5. √ dx
x x2 +1
Z π/2 Z
√ 6. t2 sec(t3 + 5) tan(t3 + 5) dt
2. x + sin x dx


0
Z √ √
Z
7. 1 + tan θ sec2 θ dθ

3. + sec x tan x dx
3
x2

0 √
(x + 2)2
Z Z
4. √ dx 8. x 3x + 4 dx
x −1
Integration Review 5

Solution:
1
1. Using basic integration formula and recalling that = x−n one has:
xn
1
Z   Z
x + 2 − 3 cos x dx =
2
x2 + x−2 − 3 cos x dx


x
1 1 1
= x3 − x−1 − 3 sin x + C = x3 − − 3 sin x + C
3 3 x
√ 1
2. Recalling that n x = x n we evaluate the definite integral to get:
π
2 3
Z π/2 Z π/2 
√   2
1
x + sin x dx = x 2 + sin x dx = x 2 − cos x


0 0 3 0
 √ 3
2 π 2 2 2π 2
  3  
 π  3
= − cos − (0) − cos(0) =
2 +1
3 2 2 3 6
n
3. Recalling that and (xm ) = xmn gives:
Z √  Z
x + sec x tan x dx = (x2/3 + sec x tan x)dx
3
2

3 5/3
= x + sec x + C
5
xm
4. Expanding the numerator, dividing through by the denominator and simplifying n = xm−n
x
before integrating gives:
(x + 2)2 x + 4x + 4
Z Z 2
√ dx = √ dx
x x
Z  
= x3/2 + 4x1/2 + 4x−1/2 dx

2 5/2 2
 
= x +4 x3/2 + 4 (2) x1/2 + C
5 3
2 8
= x5/2 + x3/2 + 8x1/2 + C
5 3

5. Using the substitution u = x2 + 1 so du = 2x dx =⇒ 21 du = x dx one has:

1 1 1
Z Z Z
x
√ dx = · √ du = u−1/2 du
x2 + 1 2 u 2
p
= u1/2 + C = x2 + 1 + C

6. Using the substitution u = t3 + 5 so du = 3t2 dt =⇒ 31 du = t2 dt one has:

1
Z Z
t2 sec(t3 + 5) tan(t3 + 5) dt = sec u tan u du
3
1 1
= sec u + C = sec(t3 + 5) + C
3 3
7. Using the substitution u = 1 + tan θ so du = sec2 θ dθ one has:
√ √
Z Z Z
1 + tan θ sec2 θ dθ = u du = u1/2 du
2 2/3 2
= u + C = (1 + tan θ)3/2 + C
3 3
6 1.5 Integration Examples

8. Make the following substitution u = 3x + 4 so du = 3dx =⇒ 13 du = dx. Solving for x gives

x = 31 (u − 4) allowing its replacement in the integral. Finally we change to limits in u:

x = 0 =⇒ u = 0 + 4 = 4
x = −1 =⇒ u = −3 + 4 = 1

0 √ 4
1 √ 1
Z Z
x 3x + 4 dx = (u − 4) u du
−1 1 3 3
4
1 4 3/2 1 2 5/2 2
Z   
= (u − 4u ) du =
1/2
u −4 u3/2
9 1 9 5 3
 1
1 2 5/2 8 3/2 1 2 5/2 8 3/2
  
= (4) − (4) − (1) − (1)
9 5 3 9 5 3
1 64 64 1 2 8 1 64 64 2 8
     
= − − − = − − +
9 5 3 9 5 3 9 5 3 5 3
1 62 56 1 94
   
= − = −
9 5 3 9 15
94
=−
135

Further Questions:

Evaluate the following integrals:

√ 2
Z   Z
1. t + t − 2 dt
2 7. sin(5θ) dθ
t
3
3x2 − 1
Z 1 Z

2
2. t2 + 1 dt 8. 2 dx
0 2 (x3 − x)
Z Z

13
3. x2 x3 + 2 9. t2 sin 1 − t3 dt


dx
π √
x − 3x
Z 4
Z
4. (sec x − tan x) sec x dx 10. √ dx
0 2x
√
cos x
Z Z 4
5.
3
√ dx 11. (4x + 9) 2 dx
x 0
Z 2 √ Z
6. x x − 1 dx 12.
4
(cos θ + sin θ) (cos θ − sin θ) dθ
1
Integration Review 7

Answers:
Chapter 1 Review Exercises
Page 169

1-12: Evaluate the given integral.

1 5 √
Z   Z
1. x +
3
√ + dx 7. 2x + 5 dx
x x3
Z 3 √ 3
x + x+1
Z

3/2
2. √ dx 8. x x2 + 16 dx
x 0
√ √
tan x sec x sin(3x)
Z Z
3. √ dx 9. 2 dx
x (2 + cos(3x))
12
1
Z 1 Z
4. x2 − x3 dx 10.


dx
0 5 (x − 4)1/3
1
Z π/2
Z
5.
3
cos x (1 + sin x) dx 11. √ 4 dx
0 (1 + x)
Z Z
5
6. x2 sin x3 + 2 dx 12. x (x + 4) dx


8
Chapter 2: Inverses and Other Functions

9
10 2.1 Inverse Functions

2.1 Inverse Functions

Example 2-1
1
The inverse function of the function f (x) = x3 is g(x) = x 3 .

1
Intuitively g(x) = x 3 is the inverse of is f (x) = x3 because g undoes the action of f . So if f acts on
1
the value 2 so f (2) = 23 = 8 and we act g on the result, g(8) = 8 3 = 2 we are returned to the original
value.
One may wonder whether all functions have inverses. The answer is no. A necessary and sufficient
condition for a function to have an inverse is that the function be one-to-one.

Definition: A function f with domain A and range B is said to be one-to-one if whenever f (x1 ) =
f (x2 ) (in B) one has that x1 = x2 (in A).

A logically equivalent condition is that if x1 ̸= x2 then f (x1 ) ̸= f (x2 ). In words, no two elements in
the domain A have the same image in the range B.

2.1.1 Horizontal Line Test

The Horizontal Line Test says that a function f (x) will be one-to-one if and only if every horizontal
line intersects the graph of y = f (x) at most once.
Example 2-2
The horizontal line test shows that the function y = x2 is not one-to-one while y = x3 is one-to-one.

y
f (x) = x3
f (x) = x2

x x
Inverses and Other Functions 11

The following theorem is intuitively true when one considers the Horizontal Line Test.
Theorem 2-1: Suppose a function f has a domain D consisting of an interval. If the function is
increasing everywhere or decreasing everywhere on D then f is one-to-one.
Example 2-3
Determine whether the given functions are one-to-one:
x √
1. f (x) = 2x3 + 5 2. f (x) = x+1
x−4 3. f (x) = √
x

Solution:

1. f (x) = 2x3 + 5
• Method 1: Using the definition:

f (x1 ) = f (x2 ) =⇒ 2x31 + 5 = 2x32 + 5

=⇒ 2x31 = 2x32
=⇒ x31 = x32 (Take the cubic root of both sides.)
=⇒ x1 = x2

• Method 2: Using the derivative:

f ′ (x) = 6x2 > 0 for all x in D = (−∞, ∞)

Therefore, f (x) is always increasing on an interval which implies that f (x) is one-to-one.
x
2. f (x) =
x−4
We use the definition to prove f (x) is one-to-one:
x1 x2
f (x1 ) = f (x2 ) =⇒ =
x1 − 4 x2 − 4
=⇒ x1 (x2 − 4) = x2 (x1 − 4)
=⇒ x1 x2 − 4x1 = x2 x1 − 4x2
=⇒ −4x1 = −4x2
=⇒ x1 = x2

Note the derivative of f (x) is negative everywhere it is defined:

(1)(x − 4) − x(1) −4
f ′ (x) = = <0
(x − 4)2 (x − 4)2

Therefore f (x) is decreasing for all x in the domain of f . However the domain D = R − {4} =
(−∞, 4) ∪ (4, ∞) is broken into two intervals so this is not sufficient to prove f (x) is one-to-one.
Consider, as a counter-example, f (x) = tan x which increases everywhere on its domain but
fails hopelessly to be one-to-one as is seen easily by the horizontal line test! Care must be
taken to inspect the domain when using the derivative approach. One would need to show
that the range of the function on each interval it contained was disjoint (did not overlap) with
the range of the function on the other intervals. This occurs in this example where f maps
(−∞, 4) to (−∞, 1) and maps (4, ∞) to (1, ∞).
12 2.1 Inverse Functions

√
x+1
3. f (x) = √
x
The square roots require x ≥ −1 and x ≥ 0. Furthermore we cannot divide by zero. Therefore
the domain of the function is composed of a single open interval, D = (0, ∞). We use the
method of the derivative:
√ √
2 (x + 1) x − x + 1 21 x−1/2
1 −1/2


f (x) =
′

√ √
x
√ x − √ x+1 √ √
x+1 x x x+1 x − (x + 1)
= ·√ √ = 3/2 √
2x x x+1 2x x+1
−1
= 3/2 √ < 0 for all x > 0
2x x+1

Therefore, f (x) is decreasing on interval (0, ∞) which implies that f (x) is one-to-one.

Further Questions:

Determine whether the given functions are one-to-one:

1. f (x) = 2x3 + 5
3−x
2. f (x) =
x+1

Definition: Suppose f is a one-to-one function defined on domain A with range B. The inverse
function of f denoted by f −1 is defined on domain B with range A and satisfies

f −1 (y) = x ⇐⇒ f (x) = y

for any y in B.

Here the symbol ⇐⇒ means “if and only if”. “This if and only if that” itself means that both the
following hold

• “If this then that.”

• “If that then this.”

Also observe that the notation f −1 makes it clear that f is the function for which this is the inverse.
Notes:

1
1. f −1 ̸= We call 1
f the reciprocal of f .
f
Inverses and Other Functions 13

2. The definition says that if f maps x to y, then f −1 maps y back to x.

x y

f −1

A B
3. The domain of f −1 is the range of f while the range of f −1 is the domain of f .
4. Reversing the roles of x and y gives

f −1 (x) = y ⇐⇒ f (y) = x

or equivalently
f (y) = x ⇐⇒ f −1 (x) = y
This implies that f itself is the inverse function of f −1 .
5. The following hold (see last diagram)

f −1 (f (x)) = x for every x in A

f f −1 (y) = y for every y in B

The first relationship highlights the utility of the inverse function in solving equations for if we
have, say, f (x) = 3 for some one-to-one function f for which we know the inverse f −1 (x), it
follows, applying f −1 to both sides that f −1 (f (x)) = x = f −1 (3). We are applying inverse
functions all the time when we isolate variables in equations.
This explains why “cube-rooting both sides” of x3 = 64 is a safe way to find the solution to this
equation while “square-rooting both sides” of x2 = 64 is not. The latter finds only one of the two
solutions. (Applying a function in this way can only produce one number.)

2.1.2 Finding Inverse Functions

To find the inverse function of a one-to-one function f proceed with the following steps:

1. Write y = f (x)
2. Solve the equation for x in terms of y (if possible).
3. Interchange the roles of x and y. The resulting equation is y = f −1 (x).

Example 2-4
Find the inverse function of the given function:
x √
1. f (x) = 2x3 + 5 2. f (x) = x+1
x−4 3. f (x) = √
x
14 2.1 Inverse Functions

Solution:

These are the same functions from Example 2-3. As they were all shown to be one-to-one we know
they are invertible.

1. To find the inverse function of f (x) = 2x3 + 5, we solve y = f (x) for x in terms of y.

y = f (x) =⇒ y = 2x3 + 5 =⇒ y − 5 = 2x3

y−5 3 y − 5
r
=⇒ = x =⇒ x =
3
2 2
x−5
r
Exchange x and y =⇒ y =
3

2
3 x − 5
r
Therefore f −1 (x) =
2
x
2. To find the inverse function of f (x) = , we solve y = f (x) for x in terms of y.
x−4
x
y = f (x) =⇒ y = =⇒ y(x − 4) = x =⇒ yx − 4y = x
x−4
=⇒ yx − x = 4y =⇒ x(y − 1) = 4y
4y
=⇒ x =
y−1
4x
Exchange x and y =⇒ y =
x−1
4x
Therefore f −1 (x) =
x−1

√
x+1
3. To find the inverse function of f (x) = √ , we solve y = f (x) for x in terms of y.
x
√
x+1 x+1
r
y = f (x) =⇒ y = √ =
x x
x+1
=⇒ y 2 = =⇒ xy 2 = x + 1 =⇒ xy 2 − x = 1
x
1
=⇒ x(y 2 − 1) = 1 =⇒ x = 2
y −1
1
Exchange x and y =⇒ y = 2
x −1
1
Therefore f −1 (x) = 2
x −1

Further Questions:

Find the inverse function of the given function:

1. f (x) = 2x3 + 5 3−x 3. f (x) = x2 − 9

2. f (x) =
x+1
Inverses and Other Functions 15

Note that if you are able to solve your expression uniquely for x in terms of y in the second step it
follows that the function is one-to-one since, given any y value in the range B there can only be a single
value x in A which maps to it, namely the value which results from evaluating your solved expression
with y.

2.1.3 Graphs of Inverse Functions

The definition of the inverse function implies that if (x, y) lies on the graph of y = f (x) then (y, x) will
lie on the graph of f −1 . Geometrically this means that the graph of f −1 may be obtained by reflecting
the graph of f about the line y = x.

y f (x) = x3

y=x

1
f −1 (x) = x 3

(1/8, 1/2)

(1/2, 1/8) x

Graphically a discontinuity in f would imply a discontinuity in f −1 and vice versa. We have the
following theorem.
Theorem 2-2: Suppose f is a one-to-one continuous function defined on an interval then its inverse
f −1 is also continuous.

2.1.4 Derivative of an Inverse Function

If we let g(x) be the inverse of f then our earlier relationship x = f f −1 (x) = f (g(x)). Differentiating

the left side with respect to x just gives 1. Differentiating the right side of the equation with respect to
x can be done with the Chain Rule. Solving for the derivative g ′ (a) gives the following result.
Theorem 2-3: Suppose f is a one-to-one differentiable function with inverse g = f −1 . If f ′ (g(a)) ̸= 0
then the inverse function is differentiable at a with
1
g ′ (a) = ′
f (g(a))
More generally the derivative of the inverse function is
1
g ′ (x) = .
f ′ (g(x))
16 2.1 Inverse Functions

Example 2-5
For the function f (x) = x2 + 1, x ≥ 0:

1. Show that f is one-to-one.

2. Calculate g = f −1 and find its domain and range.
3. Calculate g ′ (5) using your result from part 2.
1
4. Find g ′ (5) from the formula g ′ (x) = .
f ′ (g(x))

Solution:

1. Using the definition of one-to-one:

f (x1 ) = f (x2 ) =⇒ x21 + 1 = x22 + 1

=⇒ x21 − x22 = 0
=⇒ (x1 − x2 )(x1 + x2 ) = 0
=⇒ x1 = x2 or x1 = −x2

Since the domain of f is restricted to x ≥ 0, x1 = −x2 is not possible as one of the numbers
would necessarily be negative, except in the case of x1 = x2 = 0. We conclude f is one-to-one.
2. Solve y = f (x) for x:

y = f (x) =⇒ y = x2 + 1 =⇒ x2 = y − 1
=⇒ x = ± y − 1 (← Reject negative due to domain of f.)
p

=⇒ x = y − 1
p
√
Exchange x and y =⇒ y = x − 1
√
Therefore g(x) = f −1 (x) = x − 1

Domain of g requires x − 1 ≥ 0 =⇒ x ≥ 1. So Dg = [1, ∞). Range of g equals the domain of

f and so Rg = [0, ∞).
3. Differentiating using the Chain Rule gives:
1 1 1
g ′ (x) =
(x − 1)− 2 (1 + 0) = √
2 2 x−1
1 1
and so g (5) = √
′
=
2 5−1 4

4. Since f ′ (x) = 2x we have using the inverse derivative formula:

1 1 1 1 1
g ′ (5) = = ′ √ = ′ = =
f ′ (g(5)) f ( 5 − 1) f (2) 2(2) 4

Further Questions:
1
For the function f (x) = :
x−1

1. Show that f is one-to-one.

Inverses and Other Functions 17

2. Calculate g = f −1 and find its domain and range.

3. Calculate g ′ (2) using your result from part 2.

1
4. Find g ′ (2) from the formula g ′ (x) = .
f ′ (g(x))

Example 2-6
Find (f −1 )′ (a) for the given a:

1. f (x) = x3 + 4x + 3, a = −2 x3 1
2. f (x) = ,a=−
x2 +4 5

Solution:

1. f ′ (x) = 3x2 + 4 > 0 on (−∞, ∞) =⇒ f (x) has an inverse

1
(f −1 )′ (−2) =
f ′ (f −1 (2))

We need to find f −1 (−2). To do so, call the unknown value w, use the properties of the inverse
to get an equation for it, and then solve for w as follows:

f −1 (−2) = w =⇒ f (w) = −2 =⇒ w3 + 4w + 3 = −2
=⇒ w = −1 (By inspection or direct solution.)
f (f
′
(−2)) = f ′ (−1) = 3(−1)2 + 4 = 3 + 4 = 7
−1

1 1
Therefore (f −1 )′ (−2) = ′ −1 =
f (f (−2)) 7

3x2 (x2 + 4) − x3 (2x) x4 + 12x2

2. f ′ (x) = =
(x + 4)
2 2 (x2 + 4)2

1 1
 
−1 ′
= ′ −1 1


f −
5 f (f (− 5 ))
1 1 w3 1
 
f −1 − = w =⇒ f (w) = − =⇒ 2 =−
5 5 w +4 5
=⇒ 5w3 + w2 − 4 = 0 =⇒ w = −1
(−1)4 + 12(−1)2 1 + 12 13
f ′ (−1) = = =
((−1)2 + 4)2 52 25
1 1 1 1 25
 
Therefore (f −1 )′ − = ′ −1 1 = ′ = 13 =
5 f (f (− 5 )) f (−1) 25
13

Further Questions:

Find the following derivatives:

′
1. f −1 (1) if f (x) = x3 + x + 1 . 2. g ′ (−1) if f (x) = 3x − cos x and g = f −1 .
 18 2.1 Inverse Functions

2.1.5 Creating Invertible Functions

So far one-to-one (and hence invertible) functions seem uncommon. However this is only because
we only considered functions defined on their natural domains, i.e. the set of numbers for which the
function may be evaluated. We can choose to define a function with a smaller domain and by suitable
restriction we can create a function that is one-to-one and hence invertible.

Example 2-7
Define the function f (x) to have the value f (x) = x2 but only be defined on the domain A = [0, ∞).
Since √
f is increasing everywhere on this interval it is one-to-one and hence has an inverse, f −1
p(x) =
1
x = x. If we restricted the domain to be A = (−∞, 0] the inverse would be f −1 (x) = − |x| !
2

Answers:
Exercise 2-1
Page 169

1-4: Show that the given function is one-to-one and find its inverse.
√
1. f (x) = 2x3 + 5 3. f (x) = 2x − 3
2x
2. f (x) =

7
3x − 1 4. f (x) = x3 + 2

5-7: In the following problems

(a) Show that f is one-to-one.

(b) Find f −1 (x) and state the domain and range of f −1 .
√
5. f (x) = x5 + 4 x+1
7. f (x) = √
x+3 x+2
6. f (x) =
2x + 1

′
8-10: Find f −1 (a) for the given a.
8. f (x) = 2x3 + 5, a = 7
9. f (x) = x5 + sin x + 2x + 2, a = 2
√
10. f (x) = 3 x + x + 1, a = 11

11. Suppose f and g are invertible functions. Consider the composite function h = f ◦ g defined
by h(x) = f (g(x)) .

(a) Show that h is one-to-one.

(b) Find h−1 (x).
Inverses and Other Functions 19

2.2 Exponential Functions

If we write the number 25 , then this, recall, means 2 × 2 × 2 × 2 × 2. We call 2 the base and 5 the
exponent. We have already seen that one way to create a function is to replace the base with a variable.
This produces power functions like
1 √
f (x) = x2 y = x2 = x y = x−1 = x1
In general, a power function is of the form y = xr where r is any real constant.
If, on the other hand, we let the exponent be a variable and the base a constant, like:
x
f (x) = 2x , y = (1/2)

we get exponential functions.

Definition: Let a > 0. The function

f (x) = ax
is an exponential function.

The graph of an exponential function has the following form depending on whether a is greater than or
less than 1. Two typical values of a are shown.

y
 x
1
y= 2 y = 2x

Notes:

1. If x = 0 then ax = a0 = 1. Therefore all exponential functions go through the point (0, 1) .

2. If x = n, a positive integer then ax = an = a
| · a ·{z. . . · a} .
n times
3. If x = −n, n a positive integer, then ax = a−n = a1n .
1 √
4. If x = n1 , n a positive integer, then ax = a n = n a . (Hence a < 0 is excluded.)
p 1 √
5. If x is rational, x = pq , then ax = a q = (ap ) q = q
ap .

6. If a ̸= 1 (and a > 0) then f (x) = ax is a continuous function with domain R and range (0, ∞).
20 2.2 Exponential Functions

7. If 0 < a < 1 then f (x) = ax is a decreasing function.

8. If a > 1 then f (x) = ax is an increasing function.

x
9. If a, b > 0 and x, y ∈ R then (a) ax ay = ax+y (b) aay = ax−y (c) (ax )y = axy (d) (ab)x = ax bx .
These relations are readily apparent when one considers x and y as positive integers.

For two bases greater than one the base which is larger is the steeper curve while for two bases less
than one the base which is smaller is steeper.
 x  x y
1 1
y= 2
y= 4 y = 4x y = 2x

As depicted in the previous graphs, we have the following limits:

Theorem 2-4: For exponential functions we have the following limits at infinity : If a > 1, then
lim ax = 0 and lim ax = ∞.
x→−∞ x→∞
If 0 < a < 1, then lim ax = ∞ and lim ax = 0.
x→−∞ x→∞

So the x-axis is a horizontal asymptote for ax provided a > 0, a ̸= 1.

2.2.1 The Natural Exponential Function

Consider the derivative of f (x) = ax :

ax+h − ax ax ah − ax ah − 1
 
f ′ (x) = lim = lim = lim ax
h→0 h h→0 h h→0 h

where here we are able to pull out the ax from the limit because ax does not involve the limit variable
h. The result shows the derivative is proportional to the function f (x) = ax itself with constant of
proportionality c given by the evaluation of the limit:

ah − 1
c = lim
h→0 h
Inverses and Other Functions 21

Due to the presence of the constant a in the limit, one anticipates correctly that the constant c depends
on the choice of base a. Interestingly, one can ask the question if there is some choice of base a for
which the constant is c = 1. The answer is yes, the base is given by Euler’s Number :

e = 2.71828 . . .

for which we have that c = 1 in the above limit:

eh − 1
lim =1.
h→0 h
More constructively, as opposed to e being the solution of such a limit equation, it will be shown that e
may be written as the following limit:
1
e = lim (1 + h) h ,
h→0

or, setting h = 1/n, n

1

e = lim 1+ .
n→∞ n

Definition: If a = e = 2.71828 . . ., then f (x) = ex is the natural exponential function .

Since e = 2.71 . . . > 1 the natural exponential function shares all the aforementioned properties of
f (x) = ax where a > 1. (i.e. continuous, increasing function with domain R, range (0, ∞), limits, etc.)

y = ex

You should identify the natural exponential key ex on your calculator.

22 2.2 Exponential Functions

2.2.2 Derivative of ex

Furthermore from the preceding discussion we have the important result:

Theorem 2-5: The derivative of the natural exponential function is:
d x
e = ex .
dx

Proof is, as above,

ex+h − ex ex eh − ex eh − 1
 
d x
e = lim = lim = lim ex = (1)ex = ex .
dx h→0 h h→0 h h→0 h

A corollary of this theorem, applying the Chain Rule to the function eu with u = g(x) is:
d u du d h g(x) i
Theorem 2-6: e = eu or e = eg(x) g ′ (x)
dx dx dx
Example 2-8
Differentiate the following functions:
2
3
1. f (x) = ex +5
3. f (t) = t2 + 5 et

2. y = cos e2x 4. 2xy 2 + xey = e3x

Solution:
2
1. f (x) = ex +5
Using the Chain Rule:
2
′ 2
f ′ (x) = ex +5
x2 + 5 = ex +5
(2x + 0)
2
= 2x e x +5

2. y = cos e2x

y ′ = − sin e2x (e2x )′ = − sin e2x e2x (2)

= −2e2x sin e2x

3
3. f (t) = t2 + 5 et
Using Product and Chain Rules:
3
3
f ′ (t) = 2tet + t2 + 5 et (3t2 )
3
3
= 2tet + 3t4 + 15t2 et

4. 2xy 2 + xey = e3x

Using implicit differentiation we have:

2y 2 + 4xyy ′ + ey + xey y ′ = 3e3x

=⇒ (4xy + xey ) y ′ = 3e3x − 2y 2 − ey
3e3x − 2y 2 − ey
=⇒ y ′ =
4xy + xey
Inverses and Other Functions 23

Further Questions:

Find dy
dx for the following:
4
1. y = e2x 4. y = ex sin x2 + 1

2. y = ex
3
+x 5. xy + ex = 2xy 2

3. y = esec x + sec (ex )

2.2.3 Integral of ex

Since d x
dx e = ex , we have the following:
Theorem 2-7: The indefinite integral of ex is
Z
ex dx = ex + C

Example 2-9
Evaluate the following integrals:
Z Z 1
1. e6x dx
2
3. xex +1
dx
0

e4x
Z Z
2. ex cos (ex ) dx 4. √ dx
2 + e4x

Solution:

1. Using substitution u = 6x so du = 6 dx =⇒ 61 du = dx one has:

1 u 1 1 6x
Z Z
e6x dx = e du = eu + C = e +C
6 6 6
2. Using substitution u = ex so du = ex dx:
Z Z
ex cos(ex ) dx = cos u du = sin u + C = sin (ex ) + C

3. Let u = x2 + 1 so du = 2x dx =⇒ 1
2 du = x dx and change limits:
x = 0 =⇒ u = 1
x = 1 =⇒ u = 2
2
1
1 1 1 2
1 2
Z Z 2
2
xex +1 dx = eu du = eu = e − e1 =


e −e
0 1 2 2 1 2 2

4. Let u = 2 + e4x so du = 4e4x dx =⇒ 41 du = e4x dx:

e4x 1 1 1 1
Z Z Z
dx = √ du = u−1/2 du = (2) u1/2 + C
2+e 4x 4 u 4 4
1p
= 2 + e4x + C
2
24 2.2 Exponential Functions

Further Questions:

Evaluate the following integrals:

2ex
Z Z
1. e2x dx 4. 2 dx
(1 + ex )
2 + 3ex
Z Z
4
2. x3 ex +1
dx 5. dx
ex
1
1 − 4e3x
Z Z
3. e −x
dx 6. dx
0 e

2.2.4 Simplifying Exponential Expressions

Using the rules of exponents we are often able to consolidate expressions involving several exponents
into an expression involving one exponent.

Example 2-10
√
e2 ex
The expression 3 may be simplified as follows:
(2ex )
√ 1
e2 ex e2 (ex ) 2  √ 1

3 = 3 since n
a = a n , (ab)x = ax bx
(2ex ) 23 (ex )
1
e2 e2x
= ( since (ax )y = axy )
23 e3x
1
e2+ 2 x
= since ax ay = ax+y


8e 3x

1 2+ 1 x−3x ax
 
= e 2 since y = a x−y
8 a
1 2− 5 x
= e 2
8

The usefulness in consolidating exponents in this manner is clear when solving equations.

Example 2-11
Solving the equation √
e2 ex 1
=
(2ex )
3 2
Is equivalent, by using our previous result and multiplying both sides by 8, to
5
e2− 2 x = 4

Now if we could apply an inverse to the natural exponential function on both sides we could solve
for x.
Inverses and Other Functions 25

Answers:
Exercise 2-2
Page 170

1-3: Evaluate the given limit.

6 −x
1. lim e−x 2. lim e−x 3. lim 5e
x→−∞ x→∞ x→∞

4-10: Differentiate the given function.

4. f (x) = x4 + 1 e−2x 7. f (x) = cos e2x + x

8. g(x) = tan (ex ) + etan x

2t2 + e3t
5. g(t) = √
t2 e3t 9. y = sin x2 + e2x
ex + 1 ex + e−x
6. y = 10. f (x) =
ex + 3 ex + 3e−x

11-12: Find the equation of the tangent line to the curve at the given point.
11. y = x − e−x , (0, −1)
12. x − xy + ey + ex = 2e, (1, 1)

13-20: Evaluate the given integral.

Z 1
e3x

2 Z
13. e2x + e−x dx 17. dx
2
0 (e3x + 2)
Z √
2 4
14. xex dx e x
Z
18. √ dx
1 x
Z Z
15. e sin (e ) dx
x x
19. 1 + etan x sec2 x dx

1

2
e2x + 1
Z Z
16. dx 20. 2 dx
ex (ex + e−x )
26 2.3 Logarithmic Functions

2.3 Logarithmic Functions

We finished the last section by suggesting that inverses of exponential functions would be useful for,
among other things, solving equations involving exponentials. Since the exponential function f (x) = ax
with constant a > 0 and a ̸= 1 is either everywhere decreasing (0 < a < 1) or increasing (1 < a) on open
interval R = (−∞, ∞), the exponential function is one-to-one and hence has an inverse function f −1 .

Definition: Given constant a > 0, a =

̸ 0, the logarithmic function of base a, written loga x is
defined by
loga x = y ⇐⇒ ay = x
That is, it is the inverse of the exponential function f (x) = ax .

In words, the logarithm of a value x to a base a is the exponent to which you must take a to get x.
For the case a > 1 , which, as you recall, is the case for a = e = 2.71 . . ., a representative graph of
y = ax and its inverse y = loga x are as follows:

y y = ax

y=x

(0, 1) y = loga x

(1, 0) x

For base a > 1 we saw that larger values of a led to steeper y = ax curves, it follows that larger values
of a will make the logarithmic curves more horizontal in this case:

y y = log2 x
y = loge x
y = log4 x
y = log10 x

(1, 0) x
Inverses and Other Functions 27

2.3.1 Logarithmic Function Properties

Because of their relationship to exponentials as inverses the following are true for logarithmic functions:

1. y = loga x has domain (0, ∞) and range R.

2. y = loga x is continuous on its domain.

3. y = f (x) = loga x is one-to-one with inverse function f −1 (x) = ax .
4. loga (1) = 0

5. The following limits hold (see graph for a > 1 case):

• If 0 < a < 1 then f (x) = loga (x) is a decreasing function with
lim+ loga x = +∞ lim loga x = −∞
x→0 x→∞

• If a > 1, then f (x) = loga (x) is an increasing function with

lim+ loga x = −∞ lim loga x = ∞
x→0 x→∞

Note the y-axis is a vertical asymptote in either case.

6. The following inverse relations hold:

loga (ax ) = x for any x in R

aloga x
= x for any x > 0

The special multiplication, division, and power laws of exponents induce the following important
logarithmic results.
Theorem 2-8: For x > 0 and y > 0 and any real number r the following hold:

1. loga (xy) = loga x + loga y

 
x
2. loga = loga x − loga y
y
3. loga (xr ) = r loga x

To prove the theorem note that if x > 0 and y > 0 then m = loga x and n = loga y exist and,
exponentiating both sides, it follows that x = am and y = an . Evaluating the first equation’s left hand
side we have:
loga (xy) = loga (am an ) = loga am+n = m + n = loga x + loga y

The other conclusions are similarly proven.

These results can be used to simplify logarithmic expressions:

Example 2-12
Simplify the following:
√   
1. log5 50 + log5 2 − log5 4 2. 5 log 104 − 6 log 10 + 2 log 103/2


28 2.3 Logarithmic Functions

Solution:

(50)(2)
 
1. log5 50 + log5 2 − log5 4 = log5 = log5 (25) = log5 52 = 2


4
2. Recalling that by convention log means log10 we have:
√    1 3
5 log 104 − 6 log 10 + 2 log 103/2 = (5)(4) log(10) − 6 log 10 + 2 log 10


2 2
= 20(1) − 3(1) + 3(1) = 20 − 3 + 3 = 20

Further Questions:

Simplify the following:

1. log2 4 + log2 10 − log2 5 2. log5 3 + log5 34 + log5 1

2.3.2 The Natural Logarithmic Function

Definition: The logarithmic function with base a equal to e = 2.71 . . . is called the natural logarith-
mic function and is denoted by ln x. In symbols:

ln x = loge x

All the properties for a logarithm with base a > 1 apply to the natural logarithm. In terms of the
notation for natural logarithms and exponentials we have the definition:

ln x = y ⇐⇒ ey = x

and the properties :

ln e = 1
ln e x
= x (x ∈ R)
eln x = x (x > 0)
ln(xy) = ln x + ln y (x, y > 0)
 
x
ln = ln x − ln y (x, y > 0)
y
ln (xr ) = r ln x (x > 0, r ∈ R)

Note the following:

loga (x + y) ̸= loga x + loga y
loga (x − y) ̸= loga x − loga y
In the specific case of natural logarithms (a = e):

ln(x + y) ̸= ln x + ln y

ln(x − y) ̸= ln x − ln y
Inverses and Other Functions 29

You should identify the natural logarithm key ln on your calculator. Note that the key log on the
calculator means base 10 logarithm log10 x.1
Example 2-13
Simplify the following:
√

1
 
1. ln e + 2 ln(3e) − ln e5


2. ln
e3/2

Solution:

√
1 1 5
1. ln e + 2 ln(3e) − ln e5 = ln(e 2 ) + 2(ln 3 + ln e) − 5 = + 2(ln 3 + 1) − 5 = − + 2 ln 3


2 2
1 3 3 3
   
2. ln 3/2 = ln 1 − ln e3/2 = 0 − ln e = − (1) = −
e 2 2 2

Further Questions:

Simplify the following:

1. ln 5 + 2 ln 3 + ln 1 1 3. eln(x
2
+1)
+ 3x2 − 5
2. ln(4t) − ln(t2 + 1)
2

2.3.3 Solving Exponential and Logarithmic Equations

Solving equations involving logarithmic or exponential functions typically involves using properties of
these functions to simplify those expressions involving the variable and then applying the appropriate
inverse function to undo the exponential or logarithm. Finally one may solve for the variable.2
Example 2-14
We saw that the equation √
1e2 ex
3 = 2
(2ex )
could be written, using properties of exponentials, as
5
e2− 2 x = 4 .
Applying ln, the inverse of the exponential ex , to both sides of the equation, gives
5
2 − x = ln 4 .
2
Solving for x gives
2
x= (2 − ln 4) .
5
1 However in other areas (some computer applications) the symbol log will often refer to a natural logarithm so one

needs to be careful.
2 More complicated equations may allow themselves to be written as a product of factors equal to zero:

(factor1 )(factor2 ) · . . . · (factorn ) = 0

where the factors themselves involve logarithms or exponentials. Note that a strictly exponential factor equalling zero
will provide no solution as ax ̸= 0 for all x. A strictly logarithmic factor equalling zero will be equivalent to the argument
of the logarithm equalling 1.
30 2.3 Logarithmic Functions

Example 2-15
Solve the following equations for x.

1. e2x+5 = 3 2. ln (ln x) = 2 3. ln (x − 2) = 2 + ln (x − 3)

Solution:

1. e2x+5 = 3
Take the natural logarithm of both sides.

ln e2x+5 = ln 3

=⇒ 2x + 5 = ln 3
=⇒ 2x = ln 3 − 5
1
=⇒ x = (ln 3 − 5)
2

2. ln (ln x) = 2
Exponentiate both sides.

eln(ln x) = e2 =⇒ ln x = e2

Exponentiate both sides again.

2 2
eln x = e(e )
=⇒ x = e(e )

2 2
Note that by convention we can write e(e ) = ee (without parentheses). This is because (ee )2
would just be written e2e using our rules of exponents. In general, remember parentheses when
you ladder exponents if you do not remember the convention.
3. ln (x − 2) = 2 + ln (x − 3)
First simplify the expression.

ln(x − 2) − ln(x − 3) = 2
x−2
 
ln =2
x−3

Next exponentiate both sides.

x−2
eln( x−3 ) = e2 =⇒
x−2
= e2 =⇒ x − 2 = e2 x − 3e2
x−3
=⇒ x − e2 x = 2 − 3e2 =⇒ x(1 − e2 ) = 2 − 3e2
2 − 3e2
=⇒ x =
1 − e2
Inverses and Other Functions 31

Further Questions:

Solve the following equations for x:

2
1. 5ex−3 = 4 5. ex −5x+6
=1

2. ln x2 − 3 = 0

6. 3e2x−4 = 10
x−2 x−3
   
7. ln = 1 + ln
3. 4ex e−2x = 6 x−1 x−1

4. ln(2 ln x − 5) = 0

The following relates logarithms in other bases to the natural logarithm .

Theorem 2-9: For a > 0, a ̸= 1 we have:
ln x
loga x =
ln a

Proof comes from observing that since aloga x = x we can take the natural logarithm of both sides and
then use the power rule for the natural logarithm to get

(loga x)(ln a) = ln x .

Solving for loga x gives our result.

This theorem is useful for evaluating an arbitrary base a logarithm on a calculator.

Example 2-16
Write in terms of the natural logarithm (ln):

1. log12 5 2. log2 (e5x ) 3. logx (x2 + 2x + 1)

Solution:

ln 5
1. log12 5 =
ln 12
ln(e5x ) 5x
2. log2 (e5x ) = =
ln 2 ln 2
ln(x2 + 2x + 1) ln (x + 1)2 2 ln(x + 1)


3. logx (x + 2x + 1) =
2
= =
ln x ln x ln x

Further Questions:

Write in terms of the natural logarithm (ln):

1. log5 7 2. log20 x2 + 1 3. log10 e2x

32 2.3 Logarithmic Functions

2.3.4 Derivative of the Natural Logarithmic Function

Theorem 2-10: The derivative of the natural logarithmic function is

d 1
(ln x) =
dx x

To prove the theorem note that if y = ln x then by definition of the logarithm as an inverse we have
ey = x .
Differentiating this implicit equation with respect to x on both sides gives
ey y ′ = 1 ,
and so
dy 1 1
y′ = = y = .
dx e x
A corollary of this result, applying the Chain Rule to the function ln u with u = g(x) is
d 1 du d g ′ (x)
Theorem 2-11: ln u = or ln [g(x)] =
dx u dx dx g(x)
Example 2-17
Find the derivative of the given functions.

1. y = ln(x3 + ex ) 2. f (x) = ln(tan x) 3. f (t) = e2t+ln t

Solution:

In each of the following we use the Chain Rule.

1. y = ln(x3 + ex )

1 x ′ 3x2 + ex
y′ = 3
+ =


x e
x3 + ex x3 + ex

2. f (x) = ln(tan x)

(tan x)′ sec2 x

f ′ (x) = =
tan x tan x

3. f (t) = e2t+ln t

1 2t+ln t
 
′
f (t) = e
′ 2t+ln t
(2t + ln t) = 2 + e
t

Further Questions:

Differentiate the following functions:

x+1
 
1. y = ln x2 − 3x + 1


3. y = ln √
x+2

2. y = ln (x + ln x) 4. y = e(2+x ln x)
Inverses and Other Functions 33

2.3.5 Derivatives Using Arbitrary Bases

Theorem 2-12: The derivative of the logarithm function to base a > 0 (a ̸= 1) is

d 1 d g ′ (x)
(loga x) = [loga g(x)] =
dx x ln a dx g(x) ln a
Proof of the former derivative follows from the identity loga x = ln
ln a :
x

d ln x 1 1 1 1 1
   
d d d
(loga x) = = · ln x = · (ln x) = · =
dx dx ln a dx ln a ln a dx ln a x x ln a
Here note that we used that ln1a is constant since a is constant. The latter derivative in the theorem
follows from the Chain Rule applied to this former result.
Theorem 2-13: The derivative of an exponential function with base a > 0, a ̸= 1 is
d x d h g(x) i
a = ax ln a a = ag(x) g ′ (x) ln a
dx dx
Proof of the former derivative follows by the observation that by our inverse identies the base a may be
written a = eln a and using the Chain Rule:
d x d ln a
x d x ln a d
x
(a ) = e = e = ex ln a (x ln a) = ex ln a (ln a) = eln a (ln a) = ax ln a
dx dx dx dx
Once again the latter derivative given in the theorem is just the result arising from using the Chain
Rule with the former result.
Example 2-18
Differentiate the following functions.

2
1. y = 5x −3x

2. f (θ) = log3 (sin θ)

3. f (x) = log5 (x2 + 3x) + 10ln x

Solution:
2
1. y = 5x −3x

2
y ′ = 5x −3x
(ln 5)(x2 − 3x)′
2
= (2x − 3) ln 5 · 5x −3x

2. f (θ) = log3 (sin θ)

1 cos θ
f ′ (θ) = · (sin θ)′ =
sin(θ) ln 3 (ln 3) sin θ

3. f (x) = log5 (x2 + 3x) + 10ln x

We differentiate term by term to get:
1
f ′ (x) = · (x2 + 3x)′ + ·10ln x ln(10)(ln x)′
(x2 + 3x) ln 5
2x + 3 1 1
= 2 · + ln 10 · 10ln x
x + 3x ln 5 x
34 2.3 Logarithmic Functions

Further Questions:

Differentiate the following functions:

1. y = log10 3x2 + ex 3. y = a3x log4 x

x
2. y = 52e +3x
4. y = 4cos x

2.3.6 Logarithmic Differentiation

Using the properties of logarithms makes taking derivatives of logarithms of products, quotients, and
powers easy.

Example 2-19
To differentiate y = ln[x(x2 + 1)(x − 3)] is easily done if we expand the logarithm first and then
differentiate:
dy d
= ln[x(x2 + 1)(x − 3)]
dx dx
d 
= ln x + ln(x2 + 1) + ln(x − 3)

dx
1 2x 1
= + +
x x2 + 1 x − 3

Wouldn’t it be nice if when working with products, etc., we were always differentiating their logarithm?
In logarithmic differentiation we take the logarithm of both sides of an equation before differentiating.

Example 2-20

3 1
1
To differentiate y = x 2x3 + 1 (x + 5) 2 x2 + 3x − 1 3 one could use the (generalized) Product
Rule. Instead, try taking the logarithm of both sides of the equation to get:

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


2 3
Next differentiate both sides of the equation with respect to x to get:

1 ′ 1 6x2 1 1 1 2x + 3
y = +3 3 + +
y x 2x + 1 2 x + 5 3 x2 + 3x − 1
Multiplying both sides by y and substituting in its value gives our derivative:

1 18x2 1 2x + 3
 

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

Note that implicit differentiation is used to differentiate the ln y that shows up on the left hand side of
the equation with respect to x. This gives the y1 y ′ which is why we need to multiply both sides by y
(for which we have the function).
Inverses and Other Functions 35

Steps in Logarithmic Differentiation

1. Take logarithms of both sides of the equation y = f (x) .

2. Differentiate with respect to x on both sides, remembering to use implicit differentiation on ln y
to get y1 y ′ .
3. Solve for y ′ and substitute f (x) for y.
Example 2-21
Differentiate the following using logarithmic differentiation:
x √
1. y = (x + 2)e (2x + 5)9 x2 + 3x 3. f (x) = (sin x + x)x
2. y =
esin x

Solution:
x
1. First take the logarithm of both sides of y = (x + 2)e .
h x
i
ln y = ln (x + 2)e
=⇒ ln y = ex ln(x + 2)
Next differentiate both sides with respect to x.
1 ′ 1
y = ex ln(x + 2) + ex (1 + 0)
y x+2
ex
 
x
=⇒ y ′ (x) = ex ln(x + 2) + (x + 2)e
x + 2 | {z }
=y

√
(2x + 5)9 x2 + 3x
2. Take the logarithm of y = to get:
esin x
h
1 i
ln y = ln (2x + 5)9 + ln x2 + 3x 2 − ln esin x
   

1
ln y = 9 ln (2x + 5) + ln x2 + 3x − sin x


2
Differentiate both sides with respect to x.
y′ 2+0 1 2x + 3
=9 + − cos x
y 2x + 5 2 x2 + 3x
√
18 x + 3/2 (2x + 5)9 x2 + 3x
 
y′ = + 2 − cos x
2x + 5 x + 3x esin x

3. Taking the logarithms of both sides of f (x) = (sin x + x)x we have:

ln f (x) = ln [(sin x + x)x ] = x ln(sin x + x)
Differentiate both sides with respect to x.
f ′ (x) cos x + 1
= (1) ln(sin x + x) + x
f (x) sin x + x
cos x+1
 
=⇒ f (x) = ln(sin x + x) + x
′
f (x)
sin x + x
cos x + 1
 
=⇒ f ′ (x) = ln(sin x + x) (sin x + x)x
sin x + x
36 2.3 Logarithmic Functions

Further Questions:

Differentiate the following using logarithmic differentiation:

4 ln x
x3 + 5 x2 − 3x 3. f (x) = (ex + 1)


1. y =
x−2
√
x3 x
2. y = x2 + 3 4. y = (2x + 1)

We note that logarithmic differentiation allows, as shown in the previous example, the ability to
differentiate functions of the form y = [f (x)]g(x) which up until now we had no method to differentiate.
A function like y = xx for instance is neither a power function (xa ) nor an exponential function (ax ).
Neither of the rules for differentiating those give the correct answer of dx x = xx ln x + xx obtained by
d x

logarithmic differentiation (show this!). One approach, equivalent to logarithmic differentiation, is to

rewrite the expression as xx = (eln x )x = ex ln x , so the base is now constant, and then use the rule for
exponential derivatives along with the Chain Rule:

1
 
d x d x ln x
x = e = ex ln x (1) ln x + x = xx (ln x + 1)
dx dx x

where we returned ex ln x to xx in the last step. More generally one can write [f (x)]g(x) = eg(x) ln[f (x)]
and differentiate that directly as well. We prefer the logarithmic differentiation approach above as it
avoids pushing the base into the exponent and returning again when you are done.
The approach of pushing the base into the exponent may be useful when the function one wishes to
differentiate involves a sum of expressions only one of which requires logarithmic differentiation. In that
case one would need to evaluate the derivative of that single term on the side with logarithmic differen-
tiation and insert the result in the final derivative. Pushing the base function into the exponent, on the
other hand, can be done inline with the rest of the derivatives. For example if f (x) = ee + ex + xe + xx
then
d e d e
f ′ (x) = (e + ex + xe + xx ) = e + ex + xe + ex ln x


dx dx
= 0 + ex + (e − 1)xe−1 + xx (ln x + 1)

where we differentiated the last term as above. If we wanted to use logarithmic differentiation we would
need to assign y = xx , perform the logarithmic differentiation on the side to find the derivative y ′ , and
then insert that answer into the derivative f ′ (x).

1
2.3.7 Integral of x
and ax

The Power Rule for integration is

1
Z
xn dx = xn+1 (n ̸= −1)
n+1
The answer for the indefinite integral clearly indicates that it cannot work for n = −1 as one would be
dividing by zero. Since dx
d
ln x = x1 however, we now have an antiderivative for x−1 = x1 , namely ln x.
This will only work for values of x > 0 since the domain of ln x is only positive numbers. However a
second antiderivative of x1 that will work when x < 0 is ln(−x), since dx d
ln(−x) = −x
1
· (−1) = x1 by
the Chain Rule. We can combine the results using absolute value bars in the following theorem.
Inverses and Other Functions 37

Theorem 2-14: The indefinite integral of x−1 is

1
Z
dx = ln |x| + C
x

A useful corollary of this result is that one can now integrate the tangent function. Using the substitution
u = cos x (so du = − sin x dx) one has

sin x
Z Z Z
du
tan x dx = dx = − = − ln |u|+C = − ln | cos x|+C = ln | cos x|−1 +C = ln | sec x|+C


cos x u

d ax
Since dx a = ax ln a it follows, dividing both sides of the equation by the constant ln a, that
d x
dx ln a = ax
and so we also have the result:
ax
Z
Theorem 2-15: ax dx = +C
ln a
Example 2-22
Evaluate the following integrals:

2 4x − 1
Z Z
1. dx 3. dx
x ln x 4x2 − 2x + 10
Z
dx
Z
3
2. x2 10x +1 dx 4.
(3x − 2) ln (9x − 6)

Solution:

1
1. Let u = ln x so du = dx.
x
2 1
Z Z
dx = 2 du = 2 ln |u| + C = 2 ln | ln x| + C
x ln x u

2. Let u = x3 + 1, so du = 3x2 dx =⇒ 1
3 du = x2 dx
3
1 1 10u 10x +1
Z Z
x3 +1
x 10
2
dx = 10 du = ·
u
+C = +C
3 3 ln 10 3 ln 10

3. Let u = 4x2 − 2x + 10 so du = (8x − 2)dx =⇒ 12 du = (4x − 1)dx

4x − 1 1 1 1
Z Z
dx = du = ln |u| + C = ln |4x2 − 2x + 10| + C
4x2 − 2x + 10 2 u 2

4. Let u = 9x − 6 so du = 9dx =⇒ 91 du = dx. Also solve for the remaining factor in terms of u
to get u = 3(3x − 2) =⇒ 3x − 2 = u3 . Then

1 1 1 1
Z Z Z
dx
= du = du
(3x − 2) ln(9x − 6) 9 u
3 ln u 3 u ln u

Next let w = ln u so dw = u1 du. The integral becomes

1 1 1 1 1
Z
= dw = ln |w| + C = ln | ln u| + C = ln | ln(9x − 6)| + C
3 w 3 3 3
 38 2.3 Logarithmic Functions

Further Questions:

Evaluate the following integrals:

3
Z Z
x
1. dx 5. ex 5e dx
2x
x2
Z
sec x tan x
Z
2. dx 6. dx
x3 + 5 3 + 5 sec x
ln x
Z
3.
Z
dx
x 7. cot x dx
Z 4 √x
10
4. √ dx
1 x

Answers:
Exercise 2-3
Page 170

1-3: Use the properties of logarithms to expand the given expression.

√
(x + 1)2 x + 4 e2x (2x + 1)3
4
x+1

1. log5 2. ln √ 3. ln √ x
2x + 3 3
x+2 e +1

4-6: Express the given quantity as a single logarithm.

1
4. ln(2x) − 3 ln(ex + 2) − ln(x + 4)
2

1 1
5. 3 log10 x2 + 1 + log10 (x + 4) − log10 3x2 + 5 )


3 4

1
6. 3 ln x + 2 log3 x3 + 2 − ln(3x + 1)
2

7-11: Solve for x.

7. ln x2 + 3 = 4 10. ln (ex − 2) + ln ex = ln 8

8. e2x − 5ex + 6 = 0 11. ex 2x = 5

x
9. ln (2x + 1) + ln x = ln x2 + 2 (Hint: Convert 2x = eln 2 = ex ln 2 )


Inverses and Other Functions 39

12-15: For the following functions find

(a) The domain and the range of f , and
(b) f −1 (x) and its domain.
√ ex − 2
12. f (x) = 2 + 3ex 14. f (x) =
ex + 3
ln x + 1
13. f (x) = ln(3x + 2) 15. f (x) =
ln x + 2

16-22: Differentiate the given function.

√
16. f (x) = ln x2 + ex e2x x2 + 5


20. y = √ 3
x+1
17. y = log4 (sin x + 3x)

ln x
18. g(t) = 10t+2 ln (ln t + 5) 21. f (x) = x2 + ex
 √
19. f (x) = ln x2 x2 + 3 e3x + 1 22. y = xsin x



23-31: Evaluate the given integral.

√
e−2x 4
10
Z Z x
23. √ dx 28. √ dx
e−2x + 3 1 x
ex
Z
24. dx sin 2x
ex + 5
Z
29.
3 + cos 2x
2
1
Z
25. dt
t+9
0
3
Z
Z
2 30. 4 dx
26. dx x (ln x)
3
x [2 + ln x]
log4 x
Z Z
2
27. (2x + 1) 5x +x
dx 31. dx
x

32. (a) Show the (remarkable) result that aln x = xln a .

Z
(b) Using the identity from (a) find the integral 5ln x dx.
Z
(c) Integrate 5ln x dx using the substitution u = ln x.

(d) Is f (x) = 5ln x a power function, an exponential function, or neither?

40 2.4 Exponential Growth and Decay

2.4 Exponential Growth and Decay

Many quantities, such as the number of cells being cultured in a lab dish or the number of radioactive
nucleii of a particular isotope in a radioactive sample remaining undecayed, have a population y(t) that
satisfies the differential equation
dy
= ky .
dt
Here the constant k, called the relative growth rate, characterizes the population under consideration.
It will be positive (k > 0) if the population y(t) is increasing in time and negative (k < 0) if it is
decreasing. The preceding equation is called the law of natural growth or law of natural decay
respectively. The constant k is called relative since if we solve for it, k = y1 dy
dt , we see the rate dy/dt is
constant only relative to the population size y at an given time.
The differential dy = dy
dt dt satisfies
dy = ky dt .
Over a fixed time interval ∆t we have the analogous relation

∆y = ky∆t ,

where y, by the Mean Value Theorem, is evaluated at some time t in the interval. Assuming ∆t is
small enough this can be effectively any time t in the interval as y will be approximately constant over
such an interval. The change ∆y in the population y(t) over a fixed small time interval ∆t is therefore
proportional to the population itself
∆y ∝ y .
which is expected for a population whose growth (or loss) depends on the current size of the population.
Additionally the relation shows the change will also be approximately proportional to the length of the
time interval ∆t considered,
∆y ∝ ∆t ,
assuming again that ∆t is sufficiently small, a result that is also reasonable.
To understand how y changes in time we need to find the function y(t) that satisfies (solves) the
differential equation
dy
= ky .
dt
If the right hand side of the
R equation just involved t explicitly, like dy
dt = t , the answer would just be
2

the antiderivative y(t) = t = 3 t + C. Our differential equation is not of this form, however, as
2 1 3

it has the dependent variable y on the right hand side. Solving such a differential equation such as
ours can be done by the process of separation of variables . Inspired by the Leibniz notation, one
formally proceeds by isolating, if possible, a function of the dependent variable y and its differential dy
on one side of the equation and a function of the independent variable t and its differential dt on the
other to get
dy
= k dt .
y
One then integrates both sides:
Z Z
dy
= k dt
y
⇒ ln y = kt + D ,
Inverses and Other Functions 41

where we have combined the integration constants C1 and C2 arising from both sides of the integral
into D = C2 − C1 . Finally we can solve for y by taking the natural exponential of both sides:

eln y = ekt+D
⇒ y = ekt eD

Calling a new (positive) constant C = eD we have the final solution of the differential equation

y(t) = Cekt .

Despite the lack of rigour in our separation of variable approach, one may readily confirm that
y(t) = Cekt does satisfy the original differential equation as required.
The constant of integration, C can be determined by providing an additional piece of information
regarding the system. If, for instance, one knows the initial size of the population is y(0) = y0 , then
the solution of the resulting initial value problem gives

y0 = y(0) = Cek(0) = Ce0 = C(1) = C .

Placing this value for C = y0 back in y(t) gives

y(t) = y0 ekt .

As such the population at arbitrary time, assuming it is undergoing exponential growth or decay, is
characterized completely by the growth constant k and its initial size y0 . The graphs of the cases
where k > 0 (growth) and k < 0 (decay) are shown below.

k>0 y

k<0

y = y0 ekt y0 y = y0 ekt
y0

t t
42 2.4 Exponential Growth and Decay

Example 2-23
A quantity P (t) measuring the number of bacteria growing in a controlled laboratory setting increases
exponentially. Suppose it doubles in 8 hours.

1. Find the relative growth rate.

2. After how many hours has P increased by 50%?

Solution:

1. P (8 hr) = 2P0 where P0 is the initial population.

=⇒ P0 ek(8 hr) = 2P0 =⇒ e(8 hr)k = 2

 
ln e(8 hr)k = ln 2
(8 hr)k = ln 2
ln 2
k= (1/hr)
8
ln 2
Therefore P (t) = P0 e 8 hr t .
2. An increase of 50% means that at the unknown time t the population has grown to
P (t) = P0 + (0.5)P0 = 1.5P0 .
ln 2 ln 2
=⇒ P0 e 8 hr t = 1.5P0 =⇒ e 8 hr t = 1.5
ln 2
t = ln 1.5
8 hr
8 ln 1.5
t= hr ≈ 4.68 hr
ln 2

Further Questions:

Fox squirrels introduced into a city see their population increase from 50 to 12000 in 4 years.
Assuming the growth was exponential over this time period,

1. Find the relative growth rate k.

2. When will the squirrel population exceed 1 million?

3. Is the latter likely? Explain.

When k < 0 we have a decay formula and the amount y decreases over time. Then the positive constant
λ = |k| = −k, called the decay constant , may be introduced and our formula becomes

y(t) = y0 e−λt .

Rather than the decay constant, one often uses the half-life constant T for a radioactive sample. It is
defined to be the time required for half of the initial decaying substance to disappear (i.e. decay into a
new form), and so y(T ) = 21 y0 . This can be used to determine the decay constant λ .
Inverses and Other Functions 43

Example 2-24
The mass of a certain radioactive material is given by m(t) = m0 ekt . If the half-life of this material
is 1600 years, find the decay constant λ = −k .
Solution:
1 1 1
m(t) = m0 ekt =⇒ m(1600 yr) = m0 =⇒ m0 ek(1600 yr) = m0 =⇒ ek(1600 yr) =
2 2 2
1
   
=⇒ ln e k(1600 yr)
= ln = ln 1 − ln 2 = 0 − ln 2 = − ln 2
2
ln 2
=⇒ k(1600 yr) = − ln 2 =⇒ k = − (1/yr)
1600
ln 2
=⇒ λ = (1/yr) ≈ 0.000433 (1/yr)
1600

Further Questions:

Cobalt-60 is a radioactive isotope used in early radiotherapy and other applications. Sixty is the
mass number of the nucleus, the number of nucleons (protons and neutrons) it contains. A sample
of Cobalt-60 undergoes exponential decay with a half-life of 5.2714 years.

1. Find the decay constant λ = −k of Cobalt-60.

2. How long would it take for a sample containing 40 grams of the isotope to decay to a sample
containing only 10 grams of it?

Finally we note that there are many examples of quantities besides population counts which satisfy the
differential equation dy
dt = ky with solution y = y0 e . As an example, the voltage across a discharging
kt

capacitor in an electronic circuit containing only a resistor and a capacitor (an RC circuit) undergoes
exponential decay from an initial voltage V0 .
 44 2.4 Exponential Growth and Decay

Answers:
Exercise 2-4
Page 172

1. A certain type of bacteria grows exponentially and doubles every 5 hours.

(a) Find the growth constant.

(b) How many bacteria will there be after 20 hours assuming the initial population is 200
bacteria?

2. The initial deer population in a forest was 25 deers. After one year the number of deers had
increased to 75. If the number of deers grows exponentially, how many deers will there be at
the end of two years?
3. In 2016, the population of a small island was estimated to be 30,000 people with an annual
rate of increase of 2.5%.

(a) Find the growth constant.

(b) Estimate the population of the island in 2025.

4. Magnesium-27 decays exponentially and has a half-life of 9.45 minutes.

(a) Find the decay constant for Magnesium-27.
(b) When will an initiall mass of 20 mg be reduced to 5 mg?

5. A patient is given a 400 mg dose of medicine that degrades by 20% every 3 hours. What is
the remaining drug concentration after a day?
6. A certain radiactive isotope decays exponentially according to the formula m = m0 e−0.3t
where m is the mass (in grams) of the isotope at the end of t days and m0 is the initial mass.
Find the half-life of this isotope.
Inverses and Other Functions 45

2.5 Inverse Trigonometric Functions

2.5.1 Inverse Sine

The sine function y = sin x on its natural domain (−∞, ∞) is not a one-to-one function. It clearly fails
the horizontal line test as the intersection with the line y = 12 clearly shows:

y
1 y = sin x
y= 2 1

−2π − 3π
2
−π − π2 π
2 π 3π
2
2π x
−1

However the function y = sin x on domain − π2 , π2 is a one-to-one function:

 

y
y = sin x
1

−2π − 3π
2
−π − π2 π
2 π 3π
2
2π x
−1

Definition: The inverse function of y = sin x; − π2 , π2 is called the inverse sine function or arcsine
 

function and is denoted by y = sin−1 x or y = arcsin x. It satisfies

 π π
y = sin−1 x ⇐⇒ x = sin y − ≤y≤
2 2
y
π
y = sin−1 x
2

x
−1 1

− π2

The domain of inverse sine is [−1, 1] and range is − π2 , π2 . The usual inverse identities apply:
 

π π
sin−1 (sin x) = x for − ≤ x ≤
2 2
sin sin−1 x = x for − 1 ≤ x ≤ 1


46 2.5 Inverse Trigonometric Functions

Example 2-25
Evaluate the following:
  h  i
1. sin−1 √12 2. sec sin−1 √12

Solution:
 
1. Since inverse trigonometric functions return angles, call θ = sin−1 √1
2
. Then by definition of
the inverse we have:
1 π π
sin θ = √ with − ≤ θ ≤
2 2 2
π
=⇒ θ =
4
√
The latter result is by inspection of a 45◦ − 45◦ − 90◦ triangle with side lengths 1, 1, 2 and
recalling sine is the opposite over the hypotenuse.
2. Evaluate the inverse trigonometric function as in part 1 and then insert it to find

1 1 1 √
   π
sec sin −1
√ = sec = π =
√ = 2,
2 2 cos 2 1/ 2

where here we used the 45◦ − 45◦ − 90◦ triangle again to evaluate the cosine.

Further Questions:

Evaluate the following:

√ 
3 1
  
3. sin 2 sin−1 x


1. sin−1 2. tan sin−1
2 2

Theorem 2-16: The derivative of inverse sine is

d 1
sin−1 x = √


,
dx 1 − x2
where −1 < x < 1 .
To prove the theorem note that if y = sin−1 x then by definition of the inverse:

sin y = x

Implicit differentiation of both sides with respect to x yields

(cos y) y ′ = 1

and so dy
dx = cos y .
1
By the trigonometric identity cos2 y + sin2 y = 1 it follows that
q p
cos y = 1 − sin2 y = 1 − x2

where here the positive solution was taken since − π2 < y < π
2 implies cos y > 0. Inserting this in the
formula for dx
dy
gives the result.
Inverses and Other Functions 47

Note that the Chain Rule result is, as expected:

d g ′ (x)
sin−1 g(x) = p


.
dx 1 − g 2 (x)
Example 2-26
Differentiate the following function:
y = sin−1 (ex + 2)
Solution:

By the Chain Rule:

′ 1 ex
y ′ = sin−1 (ex + 2) = p (ex + 0) = p
1 − (ex + 2)2 1 − (ex + 2)2

Further Questions:

Differentiate the following functions:

−1
1. y = sin−1 (ln x + 3) 2. y = esin x
+ sin−1 (ex )

2.5.2 Inverse Cosine

The function y = cos x on domain [0, π] is a one-to-one function:

y
y = cos x
1

−2π − 3π
2
−π − π2 π
2 π 3π
2
2π x
−1

Definition: The inverse function of y = cos x; [0, π] is called the inverse cosine function or arccosine
function and is denoted by y = cos−1 x or y = arccos x. It satisfies
y = cos−1 x ⇐⇒ x = cos y (0 ≤ y ≤ π)

y = cos−1 x y
π

x
−1 1
48 2.5 Inverse Trigonometric Functions

The domain of inverse cosine is [−1, 1] and range is [0, π]. The inverse identities are:

cos−1 (cos x) = x for 0 ≤ x ≤ π

cos cos−1 x = x for −1≤x≤1

Theorem 2-17: The derivative of inverse cosine is

d 1
cos−1 x = − √


,
dx 1 − x2
where −1 < x < 1 .

2.5.3 Inverse Tangent

The function y = tan x on domain − π2 , π2 is a one-to-one function:

y
2
y = tan x
1

−2π − 3π
2
−π − π2 π
2 π 3π
2
2π x
−1

−2

Definition: The inverse function of y = tan x; − π2 , π2 is called the inverse tangent function and

is denoted by y = tan−1 x or y = arctan x. It satisfies

 π π
y = tan−1 x ⇐⇒ x = tan y − ≤y≤
2 2

y
π
2

y = tan−1 x

x
−3 −2 −1 1 2

− π2

The domain of inverse tangent is (−∞, ∞) and range is − π2 , π2 . The usual inverse identities apply:

π π
tan−1 (tan x) = x for −
<x<
2 2
tan tan−1 x = for x ∈ R


x
Inverses and Other Functions 49

as well as the limits:

π π
lim tan−1 x = − lim tan−1 x =
x→−∞ 2 x→∞ 2
So y = ± π2 are horizontal asymptotes of the function.
Theorem 2-18: The derivative of inverse tangent is
d 1
tan−1 x =


.
dx 1 + x2

2.5.4 Other Trigonometric Inverses

Similarly one defines y = csc−1 x, y = sec−1 x, and y = cot−1 x.3

Notes:

• Since trigonometric functions are functions of angles, inverse trigonometric functions return angles.
All our angles above are in radians. On your calculator you must have it set to radian mode to
get these inverse trigonometric function results. If you have your calculator set to degree mode
you will get your answers in degrees.

• The −1 in sin−1 x means inverse not taking to the power of −1 (reciprocal) like the 2 in sin2 x
means. If you mean take to the power of −1, i.e. sin1 x then you must write (sin x)−1 or simply
use the reciprocal trigonometric function csc x.
• It is because none of the trig functions are one-to-one and hence not invertible on their natural
domains that solving a trigonometric equation trig(x) = # requires more than just “applying the
inverse” to both sides (unlike, say comparable logarithmic or exponential equations). So to solve
sin x = 12 the result x = sin−1 (1/2) = π/6 is only one of many solutions. (See the intersections
between y = sin x and y = 1/2 in our initial graph in this section.)

A complete table of the inverse trigonometric derivatives is as follows :

d 1 d 1
sin−1 x = √ cos−1 x = − √



dx 1 − x2 dx 1 − x2
d 1 d 1
tan−1 x = cot−1 x = −



dx 1 + x2 dx 1 + x2
d 1 d 1
sec−1 x = √ csc−1 x = − √



dx x x2 − 1 dx x x2 − 1
Here the derivatives exist on the domains of the inverse trigonometric function except at those values
where the expression is undefined. For the Chain Rule formulae simply replace x by g(x) in each
formula and multiply the result by g ′ (x).

3 Note that the convention for the inverse secant and inverse cosecant functions used here is that the domain of secant

(cosecant), and hence the range of inverse secant (cosecant) is [0, π/2) ∪ [π, 3π/2) ( (0, π/2] ∪ (π, 3π/2] ). One can also
choose the more intuitive interval [0, π/2) ∪ (π/2, π] for secant and [−π/2, 0) ∪ (0, π/2] for cosecant but then absolute
value bars are required about the x outside the radical in the derivative formulae.
50 2.5 Inverse Trigonometric Functions

Example 2-27
Differentiate the following functions:

1. f (x) = cos−1 (ln x) 3. tan−1 y + sin−1 x = ey

tan−1 (x2 + 3)
2. y = 4. y = sec−1 (x2 )
x

Solution:

1 1
 
d −1
1. f ′ (x) = cos−1 (ln x) = p =− p
dx 1 − (ln x)2 x x 1 − (ln x)2
2. We can use the Product Rule:
 ′
1 2x + 0 1 1
    
y = tan (x + 3) ·
′ −1 2
= + tan (x + 3) − 2
−1 2
x 1 + (x2 + 3)2 x x
2 tan−1 (x2 + 3)
= −
1 + (x2 + 3)2 x2

3. Using implicit differentiation differentiate both sides of tan−1 y + sin−1 x = ey to get:

1 1
=⇒ (y ′ ) + √ = ey y ′
1 + y2 1 − x2
1 1
 
=⇒ y ′
− e = −√
y
1 + y2 1 − x2
1
− √1−x 2

1 + y2

=⇒ y ′ = ← Multiply by 1 =
1
1+y 2 − ey 1 + y2
1 + y2
=⇒ y ′ = − √
1− x2 (1 − ey (1 + y 2 ))

d 1 2
4. y ′ = sec−1 (x2 ) = p (2x) = √
dx x (x ) − 1
2 2 2 x x4 − 1

Further Questions:

Differentiate the following functions:

2
1. y = sin−1 (2x − 1) 5. y = tan−1 (ln x) ex +3

6. y = cos−1 e2x − 5


x−3
x r
2. y = tan−1 + ln
3 x+3 7. y = tan−1 x2 + 3 − tan cos−1 x + 1



p
3. y = x cos−1 x − 1 − x2 8. f (t) = sec−1 et + ln t

4. y = sin−1 tan−1 x 9. sin−1 y = x2 + y 2 + ey

Inverses and Other Functions 51

The derivatives of the inverse trigonometric functions give the following results:

Theorem 2-19: We have the following indefinite integrals:

1 1 1
Z Z Z
√ dx = sin−1 x + C, dx = tan−1 x + C, √ dx = sec−1 x + C
1 − x2 1 + x2 x x2 − 1
If one considers the more general integral
1
Z
√ dx ,
a2 − x2
where a > 0 is constant this can be solved by first noting that
s   √ r r r
p x2 x2 x2 x2
a − x = a 1 − 2 = a 1 − 2 = |a| 1 − 2 = a 1 − 2
2 2 2 2
a a a a

and then using the substitution u = xa (so du = dx

a ) to get:

1 1 1
Z Z Z
x
√ dx = dx = √ du = sin−1 u + C = sin−1 + C .
1
q
a2 − x2 a 1− 2 x2 − u2 a
a

Similar generalization can be done to the inverse tangent integral. We thus have:
Theorem 2-20: For constant a > 0,
1 1 1 1 1
Z Z Z
x −1 x x
√ dx = sin−1 + C, dx = tan + C, √ dx = sec−1 + C
2
a −x 2 a a +x
2 2 a a 2
x x −a2 a a

Note on Remembering Formulas

Integral formulas appearing in Theorem 2-20 with literal constants (i.e. a) are what one finds typically in
integral tables. The placement of the a’s need not be memorized as they can be inferred by dimensional
analysis. Suppose x has units of length. Then from the form of the integrand a would also have units of
length since a2 is added to or subtracted from x2 . The resulting integrals involve inverse trigonometric
functions whose arguments must be dimensionless since the result of any trigonometric function is a
dimensionless ratio. To achieve that we divide x in our original integrals of Theorem 2-19 by a. Next to
remember whether we need a a1 in front or not we consider the integrand. For the first one we have dx in
the numerator which will have the same length dimension as x since it is just an (infinitessimal) interval
of x. In the denominator we have the square root of a length-squared which is just a length. As such
for the first integral we are summing a dimensionless quantity since it is a length divided by a length
and the result must be dimensionless. The resulting arcsine function, which returns a dimensionless
angle, therefore cannot have an a out front as that would introduce a length dimension. The latter
two integrals, however, are sums of quantities of dimension length/length2 = 1/length. To achieve that
dimension in our result we must multiply the dimensionless inverse trigonometric function by a1 . Once
again one notes how the differential dx in our notation makes this work. Thank-you Leibniz!
52 2.5 Inverse Trigonometric Functions

Example 2-28
Evaluate the given integrals:

1
Z Z 2
1. √ dx
dt 2. dx
4 − t2 0 4 + x2

Solution:

1. To get the leading 1 we factor out 4:

1 1 1 1
Z Z Z
√ dt =  dt = 2 dt
4−t
r  q
2 t 2
t 2 1−


4 1−


2 2

t 1
Then let u = so du = dt =⇒ dt = 2du . The integral is then:
2 2
1 1
Z
= √ · 2 du = sin−1 u + C
2 1−u 2
 
t
= sin −1
+C
2

2. Again we factor out 4 and to get the desired 1:

1 1
Z 2 Z 2
dx
dx = · dx
0 4 + x 2
0 4 1 + ( x2 )2

x 1
Substitute: u = so du = dx =⇒ 2du = dx.
2 2
Change limits: x = 0 =⇒ u = 0, x = 2 =⇒ u = 1
The integral becomes:
u=1
1 1
1 1
Z
= · · 2 du = tan−1 u
0 4 1+u 2
2
u=0
1  −1 1 hπ i π
= tan (1) − tan−1 (0) = −0 =

2 2 4 8

Further Questions:

Evaluate the following integrals:

3 ex
Z Z
1. √ dx 4. √ dx
4 − 2x2 1 − 8e2x
tan−1 x 3x + 4
Z Z
2. dx 5. dx
1 + x2 2x2 + 3
4
Z
3.
π
i dt cos x
Z 2
6.
h
2 dx
t 9 + (ln t) 1 + sin2 x
0
Inverses and Other Functions 53

Answers:
Exercise 2-5
Page 173

1-4: Find the exact value of the following expressions.

√  √ 
2 3

1. sin −1
3. sin 2 cos−1
2 2
 π 
2. tan−1 sin 4. tan cos−1 x


2

5-11: Differentiate the given functions.

5. f (x) = sin−1 (ex + 2) 9. tan−1 y + ey + xy = 5

6. y = cos−1 (ln x + 5)
10. sin−1 x + sin−1 y = x2 + y 2
7. g(x) = tan−1 (sin x)

ln x
8. y = sin−1 x 11. f (x) = sec−1 (2x)

12-20: Evaluate the given integral.

ex sec2 x
Z Z
12. dx 17. dx
3 + 2e2x
p
1 − tan2 x
2
Z
13. dx
x [5 + (ln x)2 ] 1
Z
18. √ dx
x(2 + x)
3x
Z
14. √ dx
1 − x4
1
Z
Z
2x + 3 19. √ dx
15. dx e2x −1
x2 + 5
√
sin−1 x 2/2
2x
Z Z
16. √ dx 20. √ dx
1 − x2 0 1 − x4
54 2.6 L’Hôpital’s Rule

2.6 L’Hôpital’s Rule

We have already evaluated limits that are indeterminate forms of the type 0
0 and ∞.
∞

Example 2-29
Evaluate the following limits:

x2 + 3x − 10 7x2 + 3x + 1
1. lim 2. lim
x→2 2x2 − 8 x→∞ 9x3 + 4

Solution:

1. Substitution of 2 in the rational function confirms a 00 indeterminate form. The polynomials

evaluating to zero at x = 2 suggest the presence of a factor of (x − 2) is the cause. Factoring
confirms this and allows evaluation of the limit:
x2 + 3x − 10 (x − 2)(x + 5) x+5 7 7
lim = lim = lim = =
x→2 2x2 − 8 x→2 2(x + 2)(x − 2) x→2 2(x + 2) 2(4) 8

2. As x → ∞ the polynomials in the rational function both get large and we have an ∞ ∞
indeterminate form. Factoring out the dominant term in each of the numerator and the
denominator, in this case the highest power of x, allows the limit to be evaluated:

7x2 + 3x + 1 x2 7 + x3 + x12 1 7 + x3 + x12 7+0+0

lim = lim · = lim · =0· =0
x→∞ 9x3 + 4 x→∞ x3 9 + x3
4 x→∞ x 9 + x43 9+0

Further Questions:

Evaluate the following limits:

x2 − 1 3x2 − 5x + 2
1. lim 2. lim
x→1 x2 − 3x + 2 x→∞ 4x2 + 3x − 10

In general:

f (x)
• If lim f (x) = 0 and lim g(x) = 0, then lim is called the indeterminate form of type 0
0
.
x→a x→a x→a g(x)
f (x)
• If lim f (x) = ±∞ and lim g(x) = ±∞, then lim is called the indeterminate form of
x→a x→a x→a g(x)
type ∞
∞
.

Our techniques used above will not work for evaluating all limits of this type:

Example 2-30
2x − 1 ln x
The limit lim is an indeterminate form of type 00 while lim is of type ∞.
∞
Neither limit
x→0 x x→∞ x
may be resolved using the methods of the previous example.
Inverses and Other Functions 55

Theorem 2-21: If f and g are differentiable functions with g ′ (x) ̸= 0 on an open interval containing
f (x)
the value a and lim is an indeterminate form of type 00 or ∞
∞ , (i.e. x→a
lim f (x) = 0 and lim g(x) = 0
x→a g(x) x→a
or lim f (x) = ±∞ and lim g(x) = ±∞) then
x→a x→a

f (x) f ′ (x)
lim = lim ′ ,
x→a g(x) x→a g (x)

provided the limit on the right hand side either exists or is ±∞. This is L’Hôpital’s Rule.

Example 2-31
Evaluate the previous limits using L’Hôpital’s Rule:

2x − 1 2x ln 2 − 0
1. lim = lim = 20 ln 2 = (1)(ln 2) = ln 2
x→0 x x→0 1
ln x 1
1
2. lim = lim x = lim =0
x→∞ x x→∞ 1 x→∞ x

Note that when applying L’Hôpital’s Rule one is not using the Quotient Rule! The derivatives in the
numerator and denominator are taken separately.

Example 2-32
Evaluate the given limits:

x3 − 8 e2x − e2 x2 + 1
1. lim 2. lim 3. lim
x→2 x − 2 x→1 ln x x→∞ ex

Solution:

1. The functions f (x) = x3 − 8 and g(x) = x − 2 are differentiable and

f (2) = 23 − 8 = 8 − 8 = 0
g(2) = 2 − 2 = 0 .

Thus the quotient is of 0

0 indeterminate form. Therefore we can apply L’Hôpital’s Rule:

x3 − 8 3x2
lim = lim = 3(2)2 = 12
x→2 x − 2 x→2 1
LH

2. The function f (x) = e2x − e2 and g(x) = ln x are differentiable and

f (1) = e2 − e2 = 0
g(1) = ln 1 = 0 .

Thus the quotient is of 0

0 indeterminate form. Therefore, we can apply L’Hôpital’s Rule:

e2x − e2 2e2x
lim = lim 1 = lim 2xe2x = 2(1)e2(1) = 2e2
x→1 ln x LH
x→1
x
x→1
56 2.6 L’Hôpital’s Rule

3. The function f (x) = x2 + 1 and g(x) = ex are differentiable and

lim f (x) = lim x2 + 1 = ∞

x→∞ x→∞
lim g(x) = lim (ex ) = ∞
x→∞ x→∞

Thus the quotient is of ∞

∞ indeterminate form. Therefore we can apply L’Hôpital’s Rule:

x2 + 1 2x 2 2
lim x
= lim x = lim x = =0
x→∞ e LH
x→∞ e LH x→∞ e ∞

2x
Here we applied L’Hôpital’s Rule a second time because was also of ∞
∞ determinate form.
ex

Further Questions:

Evaluate the following limits:

ex − 1 ex
1. lim 7. lim
x→0 x x→∞ ln x

sin x cos x
2. lim 8. lim 2
x→0 x x→0 x − 1
√
cos x + 2x − 1 x−1−2
3. lim 9. lim
x→0 3x x→5 x2 − 25
x2 + 3x + 5 sin x
4. lim 10. lim
x→2 x2 − 4 x→0 x − tan x
ln x 3x − 1
5. lim √ 11. lim
x→∞ x x→0 x
x − sin x 4e2x − 4
6. lim 12. lim
x→0 x3 x→0 ex − 1

2.6.1 Indeterminate Forms of type 0 · ∞ and ∞ − ∞

Consider the following indeterminate forms:

• If lim f (x) = 0 and lim g(x) = ±∞ then lim f (x)g(x) is called the indeterminate form of
x→a x→a x→a
type 0 · ∞ .
• If lim f (x) = ∞ and lim g(x) = ∞ then lim [f (x) − g(x)] is called the indeterminate form of
x→a x→a x→a
type ∞ − ∞ .

To solve an indeterminate form of type 0 · ∞, write the product f · g as either

f g
f ·g = or f ·g =
1/g 1/f
This will convert the indeterminate form into a form of type 0
0 or ∞
∞ which can then potentially
evaluated using L’Hôpital’s Rule.
Inverses and Other Functions 57

For indeterminate forms of type ∞ − ∞ try to convert the difference into a quotient (by using a common
denominator or factoring out common terms or rationalization) to once again reduce the limit to type 00
or ∞
∞.

Example 2-33
Evaluate the given limits:

4 1
 
1. limπ sec x sin(2x) 2. lim −
x→ 2
x→2 x2 − 4 x − 2

Solution:

π
1. As x → we have sec x → ∞, and sin(2x) → 0. Thus the product is of 0 · ∞ indeterminate
2
0
form. We write it as a form noting that cosine is the reciprocal of secant.
0
sin(2x) 2 cos(2x)
lim sec x sin(2x) = limπ = lim
x→ π
2 x→ 2 cos x LH x→ π2 − sin x

2 cos 2 π2 2 cos π


=− =− = −2(−1) = 2
sin π2 (1)

2. As x → 2+ one has x24−4 → ∞ and x−2 1

→ ∞. As x → 2− one has x24−4 → −∞ and
x−2 → −∞. Thus, the difference is of ∞ − ∞ (or −∞ + ∞) indeterminate form. Convert the
1

difference into a quotient using a common denominator.

4 1 4 1 4 − (x + 2)
   
lim − = lim − = lim
x→2 x2 − 4 x−2 x→2 (x + 2)(x − 2) x−2 x→2 x2 − 4
2−x 0 1
 
−1 −1
= lim 2 form = lim = =−
x→2 x − 4 0 x→2 2x 2(2) 4
LH

Further Questions:

Evaluate the following limits:

1 1
 
1. lim+ x2 ln x 4. lim −
x→1 x2 − 1 x−1
x→0

2. limπ (2x − π) sec x p 

x→ 2
5. lim x2 + x − x
x→∞
1 1
 
3. lim − 6. lim (x − ln x)
x→0 ex − 1 x x→∞
58 2.6 L’Hôpital’s Rule

2.6.2 Exponential Indeterminate Forms

g(x)
Several indeterminate forms arise from lim [f (x)] .
x→a

• If lim f (x) = 0 and lim g(x) = 0 then indeterminate form of type 00 .

x→a x→a

• If lim f (x) = ∞ and lim g(x) = 0 then indeterminate form of type ∞0 .

x→a x→a

• If lim f (x) = 1 and lim g(x) = ±∞ then indeterminate form of type 1∞ .

x→a x→a

Each of these can be evaluated either by taking the natural logarithm:

y = [f (x)]g(x) ⇒ ln y = g(x) ln[f (x)] ,

or by writing the function as an exponential:

[f (x)]g(x) = eg(x) ln[f (x)] .

In either case an indeterminate form of type 0 · ∞ will result. This resulting indeterminacy is the reason
these three exponential forms are themselves indeterminate. We will use the former approach, as we
did with logarithmic differentiation, to evaluate these limits.

Example 2-34
1
As a practical example prove our limit formula for e by evaluating lim (1+x) x , a limit of indeterminate
x→0
form 1∞ .
1
Let y = (1 + x) x . Taking the natural logarithm of both sides results in
1
ln y = ln(1 + x)
x
Taking the limit as x → 0 of the righthand side gives an indeterminate form of type 0
0 readily
evaluated using L’Hôpital’s Rule:
1
ln(1 + x) 1 1
lim ln y = lim = lim 1+x = lim = =1
x→0 x→0 x x→0 1 x→0 1 + x 1+0
So  
1 lim ln y
lim (1 + x) = lim y = lim eln y = e
x x→0 = e1 = e
x→0 x→0 x→0

This was the limit stated for e given before.

Example 2-35
Evaluate the following limits:
1 1
1. lim (sin x)x 2. lim (ln x) x 3. lim (2x − 1) ln x
x→0 x→∞ x→1
Inverses and Other Functions 59

Solution:

1. As x → 0, (sin x)x → 00 indeterminate form.

Let y = (sin x)x =⇒ ln y = ln [(sin x)x ] = x ln(sin x)

Then lim ln y = lim x ln(sin x) (0 · ∞ form)

x→0 x→0
ln(sin x)  ∞ 
= lim form
x→0 x−1 ∞
cos x
−x2 cos x 0
 
= lim sin x
= lim form
LH
x→0 − 12
x
x→0 sin x 0

−2x cos x + x2 sin x 0

= lim = =0
LH
x→0 cos x 1

Finally exponentiate the result lim ln y = 0 to get the desired limit:

x→0

lim (sin x)x = lim y = lim eln y = elimx→0 ln y = e0 = 1

x→0 x→0 x→0

1
2. As x → ∞, (ln x) x → ∞0 indeterminate form.
1
h 1
i 1 ln(ln x)
Let y = (ln x) x =⇒ ln y = ln (ln x) x = ln(ln x) =
x x
ln(ln x)  ∞ 
Therefore lim ln y = lim = form
x→∞ x→∞ x ∞
1
· 1
1 1
= lim ln x x = lim = =0
LH
x→∞ 1 x→∞ x ln x ∞

Exponentiating we have that lim ln y = 0 implies:

x→∞

1
lim (ln x) x = lim y = lim eln y = elimx→∞ ln y = e0 = 1
x→∞ x→∞ x→∞

1
3. As x → 1, (2x − 1) ln x → 1∞ indeterminate form.
1
h 1
i ln(2x − 1)
Let y = (2x − 1) ln x =⇒ ln y = ln (2x − 1) ln x =
ln x
ln(2x − 1) 0
 
Therefore lim ln y = lim form
x→1 x→1 ln x 0
2
2x−1 2x
= lim = lim
LH
x→1 1
x
x→1 2x − 1

2(1) 2 2
= = = =2
2(1) − 1 2−1 1

Then lim ln y = 2 implies:

x→1

1
lim (2x − 1) ln x = lim y = lim eln y = elimx→1 ln y = e2
x→1 x→1 x→1
 60 2.6 L’Hôpital’s Rule

Further Questions:

Evaluate the following limits:

e−x ln x
1. lim (1 + ex ) 4. lim− (1 − x)
x→∞ x→1
x
2. lim+ (ex − 1) 5. lim (tan x)cos x
x→0
x→ π
2
−
x
3 5

3. lim 1 + + 2
x→∞ x x

Answers:
Exercise 2-6
Page 174

1-16: Evaluate the given limit if it exists. If it does not exist but has an infinite trend (+∞ or
−∞) indicate the trend.

2x2 − x − 6 2 3
 
1. lim 2 9. lim −
x→2 x + x − 6 x→0 ex − 1 x
sin 3x  π 
2. lim 10. lim x + tan−1 x
x→0 5x x→−∞ 2
1 − sin x 
1 1

3. limπ 11. lim −
x→ 2 2 cos x x→1 x − 1 x2 − 1
tan−1 x 12. lim [ln x − ln(2x + 3)]
4. lim x→∞
x→0 5x
2x
ln x 5

5. lim 2 13. lim 1−
x→∞ x x→∞ x
3ex + ln x
3x
6. lim 14. lim e2x − 1
x→∞ ex + x x→0

3 e−x
7. lim (x − 2)e−x 15. lim (1 + ln x)
x→∞ x→∞
x
8. lim e−2x ln x 16. lim (sin 2x)
x→∞ x→0
Inverses and Other Functions 61

Answers:
Chapter 2 Review Exercises
Page 174
1
1. For the function f (x) = 3 + :
x3
(a) Show the function is one-to-one.
(b) Find the inverse of the function, g(x) = f −1 (x) .
(c) Calculate the derivative g ′ (11) .

2-13: Find the indicated derivative.

2 ln x + 5 ′ t
8. f (t) = 10e , f ′ (t)
2. f (x) = , f (x)
e3x + 4
3. f (t) = ln t4 + e2t + 1 , f ′ (t)

9. y = (ln x)cos x , y ′

x+1 ′
r
10. ln x + ln y = exy , y ′
4. g(x) = ln 3 , g (x)
2x + 4
p 11. f (x) = tan−1 (ex + ln x) , f ′ (x)
5. F (y) = e4y + e−4y + 2ey , F ′ (0)
2
6. f (x) = x2 e−x , f ′′ (x) 12. g(t) = sin−1 (tan t + 3), g ′ (t)
4x
7. y = (2x + 3) , y′ 13. h(x) = log cos−1 x + 1 , h′ (x)

14-23: Evaluate the given integral.

√
cos x
Z
x
e4
Z
14. √ dx 19. p dx
x 3 − sin2 x
4x2 tan(ln x)
Z Z
15. dx 20. dx
2x + 1 x
x+2
Z Z
16. 4 e x 2x
dx 21. √ dx
16 − x2
2
Z Z
17. 4x (4 + cos 4x ) dx 22. √ dx
3x x4 − 1
1
e2x π/2
sin x
Z Z
18. dx 23. dx
0 1 + e4x 0 1 + cos2 x

24. Find the points on the graph of y = tan−1 (2x) at which the tangent line is parallel to the
line x − 2y − 8 = 0 .
62 Review Exercises

25-31: Evaluate the given limit.

x3 + 4x + 2 1 1
 
25. lim 28. lim −
x→∞ ln (x + 2) x→0 ex − 1 ex − e−x
1/x
29. lim (ex + x)
sin−1 (2x) x→∞
26. lim
x→0 tan−1 (2x) 1
30. lim (1 + 2 ln x) x−1
x→1
x2 −1
27. lim (ln x) 31. lim [ln(5x + 2) − ln(2x + 5)]
x→1 x→∞

32. A certain bacteria culture doubles in size every 5 minutes. If there are 1000 bacteria initially
present, how many will there be in 30 minutes?

33. A drug used for sedation decays exponentially. It is observed that the amount decays by 25%
after each hour. A patient receiving this drug should not drive until there is only 30 mg left
in their system. If the initial dose is 400 mg, how long will it be until it is safe to drive?
Chapter 3: Integration Methods

63
64 3.1 Integration by Parts

3.1 Integration by Parts

Just as the Chain Rule for differentiation leads to the useful Method of Substitution for solving integrals,
so too does the Product Rule result in a useful method for solving integrals. Starting with the Product
Rule
d
[f (x)g(x)] = f ′ (x)g(x) + f (x)g ′ (x) ,
dx
one can integrate both sides of the equation to get:
Z Z Z
d
[f (x)g(x)] dx = f ′ (x)g(x) dx + f (x)g ′ (x) dx
dx
An antiderivative of the derivative of a function is just the function itself so the left-hand side becomes
f (x)g(x) + C. The constant C may be absorbed into the indefinite integrals on the right and so one
has: Z Z
f (x)g(x) = f ′ (x)g(x) dx + f (x)g ′ (x) dx .

Reordering the terms gives

Z Z
f (x)g (x) dx = f (x)g(x) −
′
f ′ (x)g(x) dx .

The formula suggests that a useful strategy for evaluating an integral is to consider an integrand as
a product of two terms, one of which may be differentiated (f (x)) and one which may be integrated
(g ′ (x)) to produce a new integral that is perhaps more easy to evaluate than the original. It is customary
to define u = f (x) and v = g(x). One then has the corresponding differentials du = f ′ (x)dx and
dv = g ′ (x)dx. The formula becomes:
Z Z
u dv = uv − v du

This is the Integration by Parts formula.

Example 3-1
Integrate x2 ln x dx.
R

The fact that ln x is easily differentiated and x2 easily integrated suggests we reorder the terms and
identify u and dv as follows: Z
ln x |x2{zdx}
|{z}
=u =dv

Then u = ln x implies (differentiating) that du =R x1 dx. The differential dv = x2 dx is integrated to

give v = 13 x3 . The Integration by Parts formula u dv = uv − v du implies:
R

1 3 1 3 1
Z   Z    
(ln x) · x2 dx = (ln x) ·


x − x · dx
3 3 x
1 3 1
Z
= x ln x − x2 dx
3 3
1 3 1 1
= x ln x − · x3 + C
3 3 3
1 3 1
= x ln x − x3 + C
3 9
Integration Methods 65

In general, to apply Integration by Parts select u and dv so that

1. The product u dv is equal to the original integrand.

2. dv can be integrated.

3. The new integral v du is easier than the original integral.

R

4. For integrals involving xp eax try u = xp , dv = eax dx .

5. For integrals involving xp (ln x)q try u = (ln x)q , dv = xp dx .

As with assigning u to ln x in the last suggestion, we often will assign u to an inverse function because
its derivative is known. Integration by Parts can then (hopefully) convert the expression into something
that can be integrated.
An integral may have several ways it can be broken into u and dv that can be tried when applying
Integration by Parts. Only one of these, or indeed none of these, may actually work to allow integration
of the function.

Example 3-2
Evaluate the following integrals:
Z Z Z 1
1. (x2 + x)e−x dx 2. e2x sin x dx 3. sin−1 x dx
0

Solution:

1. Let u = x2 + x so du = (2x + 1)dx and dv = e−x dx. Integrating the latter (using the
substitution w = −x if needed) gives v = −e−x . Integration by Parts gives:
Z Z
(x2 + x)e−x dx = (x2 + x) −e−x − −e−x (2x + 1)dx

Z
= −(x + x)e + (2x + 1)e−x dx
2 −x

Use Integration by Parts again. Let u = (2x + 1) so du = 2dx and dv = e−x so v = −e−x
again to get:
Z Z
(x2 + x)e−x dx = −(x2 + x)e−x + (2x + 1) −e−x − −e−x 2 dx

Z
= −(x + x)e − (2x + 1)e + 2 e−x dx
2 −x −x

= −(x2 + x)e−x − (2x + 1)e−x − 2e−x + C

= (−x2 − 3x − 3)e−x + C

2. Let u = e2x so du = 2e2x dx and dv = sin x dx so (integrating) v = − cos x. Using Integration

by Parts gives:
Z Z Z
e2x sin x dx = e2x (− cos x) − (− cos x)e2x dx = −e2x cos x + 2 e2x cos x dx
66 3.1 Integration by Parts

Use Integration by Parts on the final integral with u = e2x so du = 2e2x and dv = cos x dx so
v = sin x to get
Z  Z 
e sin x dx = −e cos x + 2 e sin x − (sin x) 2e
2x 2x 2x 2x


dx
Z
= −e2x cos x + 2e2x sin x − 4 e2x sin x dx

This is a wrap-around integral in that we have produced the same integral with which we have
started. Solve for the integral as we would any variable:
Z Z
e sin x dx + 4 e2x sin x dx = −e2x cos x + 2e2x sin x
2x

Z
=⇒ 5 e2x sin x dx = −e2x cos x + 2e2x sin x
1 2
Z
=⇒ e2x sin x dx = − e2x cos x + e2x sin x + C
5 5
Here a constant of integration must be included, as with any indefinite integral, to provide the
most general solution.
1
3. Let u = sin−1 x so du = √ dx and dv = dx so v = x:
1 − x2
Z 1 1 Z 1
x
sin −1
x dx = x sin −1
x − √ dx
0 0 0 1 − x2
Note for Integration by Parts how the limits carry over to the new expression. They do not
change as they would for a substitution. To evaluate the integral on the right-hand side we do
however require a substitution. Let w = 1 − x2 so dw = −2x dx =⇒ 21 dw = −x dx. The new
limits are x = 1 =⇒ w = 1 − 12 = 0 and x = 0 =⇒ w = 1 − 02 = 1 and we have:
1
1 dw
Z Z 0
sin x dx = x sin x +
−1 −1
√
0 1 w 2
1 √ 0
= x sin−1 x 0 + w 1
√ √ π
= (1) sin−1 (1) − (0) sin−1 (0) + 0 − 1 = − 1
2
Alternatively, to use the same limits all the way through, evaluate the indefinite integral on
the side to get the antiderivative:
1 dw √
Z Z
x p
− √ dx = √ = w + C = 1 − x2 + C
1 − x2 w 2

The original integral becomes:

Z 1 1 Z 1 1
x h p
sin−1 x dx = x sin−1 x − √ dx = x sin−1 x + 1 − x2
0 0 0 1 − x2 0
h p i h p i π
= (1) sin−1 (1) + 1 − 12 − (0) sin−1 (0) − 1 − 02 = − 1
2
as before.
Integration Methods 67

Further Questions:

Evaluate the following integrals:

Z Z
1. ln x dx 7. sin(ln x) dx
Z
2. xex dx
Z
8. θ sec2 θ dθ
Z
3. x2 e−x dx Z
3
Z 9. x5 e−x dx
4. ex cos x dx
Z
1
10. x sin(x2 ) dx
Z
5. tan−1 x dx
0
Z Z
6. x3 (ln x)2 dx 11. cos2 x dx

Answers:
Exercise 3-1
Page 175

1-10: Evaluate the given integral.

Z Z
1. x2 e−5x dx 6. e2x sin 4x dx

Z Z π/3
2. x2 cos−1 x dx 7. x sec2 x dx
π/4
Z √ √
Z
3. t e2 t
dt 8. sin(3 ln x) dx
Z Z
4. x 10
ln x dx 9. x2 5x dx
√
Z Z 3
5. x sin x dx
5 3
10. tan−1 x dx
0
68 3.2 Trigonometric Integrals

3.2 Trigonometric Integrals

Z
Strategy for Evaluating sinm x cosn x dx

1. For an odd power of sine (m = 2k + 1), save one sine factor and express the remaining sine factors
in terms of cosine using the identity sin2 x = 1 − cos2 x:
Z Z Z

k
k
sin2k+1 x cosn x dx = sin2 x cosn x sin x dx = 1 − cos2 x cosn x sin x dx

Then substitute u = cos x .

2. For an odd power of cosine (n = 2k + 1), save one cosine factor and express the remaining cosine
factors in terms of sine using the identity cos2 x = 1 − sin2 x:
Z Z Z

k
k
sinm x cos2k+1 x dx = sinm x cos2 x cos x dx = sinm x 1 − sin2 x cos x dx

Then substitute u = sin x .

3. If the powers of both sine and cosine are even, use the trigonometric identities:
1 1
sin2 x = (1 − cos 2x) cos2 x = (1 + cos 2x)
2 2
These may need to be used repeatedly. The identity sin x cos x = 1
2 sin 2x may also be useful.

Either 1 or 2 can be used if the powers of sine and cosine are both odd.

Example 3-3
Evaluate the following integrals:
Z Z
1. sin3 x cos6 x dx 2. tan2 x cos5 x dx

Solution:

1. Take one factor of sin x out of the odd power of sine to become part of the differential and
write the remaining powers in terms of cosine:
Z Z Z
sin3 x cos6 x dx = sin2 x cos6 x sin x dx = (1 − cos2 x) cos6 x sin x dx

Then substitute u = cos x so du = − sin x dx =⇒ −du = sin x dx. The integral becomes:
Z Z
= (1 − u2 )u6 (−du) = (u8 − u6 )du
1 9 1 7 1 1
= u − u + C = cos9 x − cos7 x + C
9 7 9 7

2. All trigonometric expressions can be written in terms of sine and cosine. We try that here.

sin2 x
Z Z Z
tan x cos x dx =
2 5
cos x dx = sin2 x cos3 x dx
5
cos2 x
Integration Methods 69

Next take out cos x from the odd power of cosine to be part of the differential:
Z Z
= sin2 x cos2 x cos x dx = sin2 x(1 − sin2 x) cos x dx

Finally let u = sin x so du = cos x dx. The integral becomes

1 1 1 1
Z Z
= u (1 − u ) du = (u2 − u4 ) du = u3 − u5 + C = sin3 x − sin5 x + C
2 2
3 5 3 5

Further Questions:

Evaluate the following integrals:

Z Z
1. sin4 x cos3 x dx 4. sin2 x dx
Z Z
2. sin x dx
3
5. sin4 x dx
Z Z
3. cot5 x sin2 x dx 6. cos2 x sin2 x dx

Z
Strategy for Evaluating tanm x secn x dx

1. For an odd power of tangent (m = 2k + 1), save a factor of sec x tan x and express the remaining
factors of tangent in terms of sec x using the identity tan2 x = sec2 x − 1:
Z Z Z

k
k
tan2k+1 x secn x dx = tan2 x secn−1 x sec x tan x dx = sec2 x − 1 secn−1 x sec x tan x dx

Then substitute u = sec x .

2. For an even power of secant (n = 2k), save a factor of sec2 x and express the remaining secant
factors in terms of tan x using the identity sec2 x = 1 + tan2 x:
Z Z Z

k−1
k−1
tan x sec x dx = tan x sec x
m 2k m 2
sec x dx = tanm x 1 + tan2 x
2
sec2 x dx

Then substitute u = tan x .

3. If m is even and n = 0 (i.e. no factors of secant), convert a single factor of tan2 x using
tan2 x = sec2 x − 1. The first term will then be integrable and the procedure may be repeated on
the second integral of now lower power.

Z
Strategy for Evaluating cotm x cscn x dx

1. For an odd power of cotangent (m = 2k + 1), save a factor of csc x cot x and express the remaining
factors of cotangent in terms of csc x using the identity cot2 x = csc2 x − 1:
Z Z Z

k
k
cot2k+1
x csc x dx =
n
cot x csc
2 n−1
x csc x cot x dx = csc2 x − 1 cscn−1 x csc x cot x dx

Then substitute u = csc x .

70 3.2 Trigonometric Integrals

2. For an even power of cosecant (n = 2k), save a factor of csc2 x and express the remaining factors
of cosecant in terms of cot x using the identity csc2 x = 1 + cot2 x :
Z Z Z

k−1
k−1
cotm x csc2k x dx = cotm x csc2 x csc2 x dx = cotm x 1 + cot2 x csc2 x dx

Then substitute u = cot x .

3. If m is even and n = 0 (i.e. no factors of cosecant), convert a single factor of cot2 x using
cot2 x = csc2 x − 1. The first term will then be integrable and the procedure may be repeated on
the second integral of now lower power.

Note: This strategy is identical for that of tangents and secants with the identification tan ⇒ cot and
sec ⇒ csc .

Example 3-4
Evaluate the following integrals:
Z Z Z π/4
1. tan5 x sec4 x dx 2. cot4 x csc4 x dx 3. tan x sec5 x dx
0

Solution:

1. Imagining a tangent substitution we pull out a sec2 x to be part of the differential. The
remaining even power of secant can be converted to tangent as needed.
Z Z Z
tan x sec x dx = tan x sec x sec x dx = tan5 x(1 + tan2 x) sec2 x dx
5 4 5 2 2

Let u = tan x so du = sec2 x :

Z Z
= u5 (1 + u2 ) du = (u5 + u7 ) du
1 6 1 8 1 1
= u + u + C = tan6 x + tan8 x + C
6 8 6 8
It is to be noted here that a secant substitution will also work here as pulling out sec x tan x
to be part of the differential leaves an even power of tangent which can be converted to secant.
The final answer, in terms of secant, will differ at most by a constant from the answer above.

2. Envisioning a cotangent substitution we try pulling out csc2 x to be part of the differential.
The remaining even power of cosecant can be converted to cotangent:
Z Z Z
cot4 x csc4 x dx = cot4 x csc2 x csc2 x dx = cot4 x(1 + cot2 x) csc2 x dx

Let u = cot x so du = − csc2 x dx:

Z Z
= u (1 + u )(−du) = (−u4 − u6 ) du
4 2

1 1 1 1
= − u5 − u7 + C = − cot5 x − cot7 x + C
5 7 5 7
Note that removing csc x cot x for the differential for a csc x substitution will not work here as
the remaining power of cotangent would be odd.
Integration Methods 71

3. Anticipating a secant substitution we separate sec x tan x for the differential:

Z π/4 Z π/4
tan x sec5 x dx = sec4 x sec x tan x dx
0 0
Let u = sec x so du = sec x tan x dx. The limits become:
1 1 √
x = π/4 =⇒ u = sec(π/4) = = √ = 2
cos(π/4) 1/ 2
1 1
x = 0 =⇒ u = sec(0) = = =1
cos(0) 1
√
Z √2
1
2
1h √ 5 i 4 √2 − 1
= u4 du = u5 = ( 2) − 15 =
1 5 1 5 5

Further Questions:

Evaluate the following integrals:

Z Z
1. tan3 x sec3 x dx 6. tan4 x dx
Z Z
2. tan x sec x dx
2 4 7. cot3 x csc4 x dx

Z Z
3. tan x dx
3 8. cot3 x csc3 x dx
Z
9. csc x dx
Z
4. sec x dx
Z 3π
Z 4

5. sec x dx
3 10. csc4 x dx
π
4

Z Z Z
Strategy for Evaluating sin mx cos nx dx, sin mx sin nx dx, cos mx cos nx dx

Apply the corresponding trigonometric identity:

1
• sin a cos b = [sin(a − b) + sin(a + b)]
2
1
• sin a sin b = [cos(a − b) − cos(a + b)]
2
1
• cos a cos b = [cos(a − b) + cos(a + b)]
2

with a = mx and b = nx .

Example 3-5
Evaluate the following integrals:
Z Z π
2
1. sin 3x cos 4x dx 2. sin 2x sin 3x dx
0
72 3.2 Trigonometric Integrals

Solution:

1. Letting a = 3x and b = 4x in the appropriate trigonometric identity gives:

1
Z Z
sin 3x cos 4x dx = [sin(3x − 4x) + sin(3x + 4x)] dx
2
1 1
Z Z
= [sin(−x) + sin(7x)] dx = (− sin x + sin 7x) dx
2 2
1 1
 
= cos x − cos 7x + C
2 7

Note here we used that sine is an odd function to write sin(−x) = − sin x before integrating.
We then integrated term by term doing the substitution u = 7x in the second integral.
2. Letting a = 2x and b = 3x in the appropriate trigonometric identity gives:
π π
1
Z 2
Z 2
sin 2x sin 3x dx = [cos(2x − 3x) − cos(2x + 3x)] dx
0 2 0
π Z π
1 1 2
Z 2
= [cos(−x) − cos(5x)] dx = (cos x − cos 5x) dx
2 0 2 0
π
1 1 1 π 1 5π 1 1
 2
   
= sin x − sin 5x = sin − sin − sin 0 − sin 0
2 5 0 2 2 5 2 2 5
1 1 1 5−1 1 4 2
 
= 1 − (1) = · = · =
2 5 2 5 2 5 5

Further Questions:

Evaluate the following integrals:

Z Z π
4
1. sin 4x cos 5x dx 3. cos 2x cos 4x dx
0
Z Z
2. sin 2x sin 6x dx 4. sin 2x sin 6x cos 2x dx

A Note on Trigonometric Identities

Note that the various trigonometric identities require in this section follow readily from the three basic
identities:
a) sin2 x + cos2 x = 1
b) sin(x ± y) = sin x cos y ± cos x sin y
c) cos(x ± y) = cos x cos y ∓ sin x sin y
The half-angle identities follow from c) setting y = x and then replacing alternately sin2 x or cos2 x
using a). Dividing a) by cos2 x gives the identity involving tangent and secant, while dividing a) by
sin2 x gives the identity involving cotangent and cosecant. The last three identities on this page follow
by solving for the various products using the + and − equations from the appropriate angle addition
formula b) or c).
Integration Methods 73

Answers:
Exercise 3-2
Page 176

1-10: Evaluate the given integral.

Z Z
1. cos10 x sin5 x dx 6. cot2 x csc4 x dx
Z Z
2. sin4 x cos3 x dx 7. tan4 x dx
Z Z
3. sin4 x dx 8. sin 9x cos 7x dx
Z Z
4. tan x sec x dx
4 4
9. sin 3x sin 5x dx
Z Z
5. tan5 x sec5 x dx 10. cos 4x cos 5x dx
74 3.3 Trigonometric Substitution

3.3 Trigonometric Substitution

Some integrals, typically involving roots, may be resolved by using the Substitution Method where the
old variable is defined in terms of a new variable via a trigonometric function.

Example 3-6
Z p
Find the indefinite integral 4 − x2 dx . We consider the substitution θ(x) defined via

x = 2 sin θ

(and so dx = 2 cos θ dθ). Unlike our usual application of the substitution method here we have
defined θ implicitly. To make θ(x) unique as required we add the additional constraint − π2 ≤ θ ≤ π2
.
(Equivalently we recognize that the explicit substitution which has been done is just θ = sin−1 x2
which, recall, is defined with this range.) The integral becomes
Z p Z p
4 − x2 dx = 4 − 4 sin2 θ · 2 cos θ dθ
Z √ p
= 4 1 − sin2 θ · 2 cos θ dθ
Z Z
= 4 cos θ cos θ dθ = 4 cos2 θ dθ
1
Z Z Z
= 4 (1 + cos 2θ) dθ = 2 dθ + 2 cos 2θ dθ
2
= 2θ + sin 2θ + C = 2θ + 2 sin θ cos θ + C
x x 1p
= 2 sin−1 +2 4 − x2 + C
2 2 2
x 1 p
= 2 sin−1 + x 4 − x2 + C
2 2

Note that when we solved the identity 1 − sin2 θ = cos2 θ for cos θ we used that − π2 ≤ θ ≤ π2 to get
p
cos θ = 1 − sin2 θ since cos θ is indeed positive on the interval. This choice of positive sign was also
used in our final step where again cos θ was represented by a positive value:
p p 1p
cos θ = 1 − sin2 θ = 1 − x2 /4 = (4 − x2 )/4 = 4 − x2 1/4 = 4 − x2
p p p
2
The restriction of the domain for θ here means that other trigonometric functions
appearing in the expression may always be evaluated using an acute triangle. Here, 2
for example, since sin θ = x/2, we can draw a right triangle
√
with angle θ and length x

x opposite and hypotenuse of 2 to work out cos θ = 2 . 4−x2

θ
√
4 − x2

This
√ method √ is called Trigonometric
√ Substitution. More generally if an integrand contains one of
a2 − x2 , a2 + x2 , or x2 − a2 (where a > 0 is constant) then the radical sign can be removed via
the appropriate substitution:

Expression Substitution Identity

√
x = a sin θ − π2 ≤ θ ≤ π2 1 − sin2 θ = cos2 θ


a2 − x2
√
a2 + x2 x = a tan θ − π2 < θ < π2 1 + tan2 θ = sec2 θ


√
x = a sec θ 0≤θ< π
or π ≤ θ < 3π
sec2 θ − 1 = tan2 θ


x2 − a2 2 2
Integration Methods 75

Note that all the ranges of θ have been chosen so that θ will equal the relevant inverse trigonometric
function with argument x/a.

Example 3-7
Prove the Archimedian result that the area of a circle of radius R is A = πR2 .
√
The area of a semi-circle of radius R is the area under the curve y = R2 − x2 between x = −R
and x = R and so the area of a circle is
Z R p
A=2 R2 − x2 dx
−R

Using substitution x = R sin θ, with − π2 ≤ θ ≤ π2 gives dx = R cos θ dθ. So θ = sin−1 (x/R) and the
limits become, for x = R, θ = sin−1 (R/R) = sin−1 (1) = π/2 and for x = −R, θ = sin−1 (−R/R) =
sin−1 (−1) = −π/2. The solution of the integral follows, similar to the last example,
π
Z R p Z 2 p
A=2 R2 − x2 dx = R2 − R2 sin2 θ · R cos θ dθ
−R −π
2
π
Z 2 √ p
= 2 R2 1 − sin2 θ · R cos θ dθ
−π
2
Z π Z π
2 2
= 2R2 cos θ cos θ dθ = 2R2 cos2 θ dθ
−π
2 −π
2
π π
1
Z 2
Z 2
= 2R2 (1 + cos 2θ) dθ = R2 (1 + cos 2θ) dθ
−π
2
2 −π
2
π
1
 2
= R θ + sin 2θ
2
2 −π
2 
1 π 1
 
π
= R 2
+ sin π − − + sin(−π)
2 2 2 2
= πR2

Example 3-8
Evaluate the following integrals:

1 1
Z Z
x3
Z
1. 3 dx 2. √ dx 3. √ dx
(9 − x2 ) 2 4 + x2 x2 − 4x − 1

Solution:
√
1. Recognizing the form a2 − x2 , let x = 3 sin θ so dx = 3 cos θ dθ.
1 1 3 cos θ
Z Z Z
3 dx = 3 3 cos θ dθ =  3 dθ
(9 − x )
2 (9 − 9 sin θ)
2
9(1 − sin2 θ) 2
2 2


3 cos θ 3 cos θ 1 1
Z Z Z
= 3 dθ = dθ = dθ
[9 cos θ]
2 2 27 cos3θ 9 cos 2θ

1 1 1
Z
x
= sec2 θ dθ = tan θ + C = √ +C
9 9 9 9 − x2
76 3.3 Trigonometric Substitution

In the final step tan θ is returned to x by noting that, by the substitu-

tion sin θ = x3 , so a triangle can be drawn with opposite side length of x
3
and hypotenuse length of 3. The remaining side is solved for using the x
Pythagorean Theorem and then tan θ evaluated as the opposite side
length over the adjacent. Any other trigonometric function of θ may θ
√
be evaluated in this manner. If θ is required note that θ = sin −1 x
3 .


9 − x2
√
2. Recognizing the form a2 + x2 , let x = 2 tan θ so dx = 2 sec2 θ dθ.
x3 8 tan3 θ 16 tan3 θ sec2 θ
Z Z
√ dx = p 2 sec2 θ dθ = dθ
4+x 4(1 + tan2 θ)
p
2
4 + 4 tan2 θ
16 tan3 θ sec2 θ 16 tan3 θ sec2 θ 16 tan3 θ sec2 θ
Z Z Z
= √ dθ = √ dθ = dθ
4 sec2 θ 4 sec2 θ 2 sec θ
Z Z Z
= 8 tan3 θ sec θ dθ = 8 tan2 θ sec θ tan θ dθ = 8 (sec2 θ − 1) sec θ tan θ dθ

In the last step we recognized a trigonometric integral and reorganized it anticipating the
substitution u = sec θ so du = sec θ tan θ dθ. The integral becomes:
1 8
Z  
= 8 (u2 − 1) du = 8 u3 − u + C = sec3 θ − 8 sec θ + C
3 3
√ !3 √
8 4 + x2 4 + x2 1 3
p
= −8 + C = (4 + x2 ) 2 − 4 4 + x2 + C
3 2 2 3

Here sec θ is returned to x by noting that, by the substitution tan θ = x2 , √

so a triangle can be drawn with opposite side of length x and adjacent 4 + x2
x
side of length 2. The remaining side is solved for using the Pythagorean
Theorem and then sec θ = 1/ cos θ is evaluated as the hypotenuse side
θ
length over that of the adjacent.
2
3. For this integral we must first remove the linear (x) term by completing the square. The
polynomial x2 − 4x − 1 has b = −4 so dividing that by 2, adding it to x, and squaring gives:
2
(x + (−4/2)) = (x−2)2 = x2 −4x+4 =⇒ x2 −4x = (x−2)2 −4 =⇒ x2 −4x−1 = (x−2)2 −5 .
The integral becomes
1 1 1
Z Z Z
√ dx = dx = √ du
x − 4x − 1 (x − 2) − 5 u −5
p
2 2 2

where here we did the substitution u = x − 2, so du = dx, to√get the integral into √ a form
ready√for a trigonometric substitution. Recognizing the form u2 − a2 . Let u = 5 sec θ so
du = 5 sec θ tan θ dθ. The integral becomes:
1 √ 1 √
Z Z
= √ 5 sec θ tan θ dθ = 5 sec θ tan θ dθ
5 sec θ − 5 5(sec θ − 1)
p
2 2
Z √ Z √
5 sec θ tan θ 5 sec θ tan θ
Z
= √ dθ = √ dθ = sec θ dθ (← Use the known result.)
5 tan2 θ 5 tan θ
√ √
u u2 − 5 x−2 x2 − 4x − 1
= ln |sec θ + tan θ| + C = ln √ + √ + C = ln √ + √ +C
5 5 5 5
To return to u note that the substitution implies sec θ = √u5 . To
find tan θ a triangle can be drawn, since
√ secant is the reciprocal of
cosine, with adjacent side of length 5 and hypotenuse of length u. u √
u2 − 5
The remaining side is solved for using the Pythagorean Theorem and
then the result follows by evaluating the opposite side length over the θ
√
adjacent side length. Finally we return to x using u = x − 2 . 5
Integration Methods 77

Further Questions:

Evaluate the following integrals:

1 1
Z 2
Z
1. √ dx 5. √ dx
x2 1 16 − x2 x3 x2 − 25
Z √ 2 Z
x2
x −9 6. dx
2. dx 3
x4 (2 − 9x2 ) 2

1 1
Z Z
3. 7. dx
2 dx 5
(x + 2x + 2)
2 (5 − 4x − x2 ) 2
Z √
2x − 3 x−4
Z
4. dx 8. dx
x2 − 4x + 8 x

Answers:
Exercise 3-3
Page 176

1-10: Evaluate the given integral.

Z p
x2
Z
1. 16 − x2 dx 6. dx
3/2
(x2 + 4)
Z √
x2 + 9 Z
x2
2. dx 7. √ dx
x 5 − 3x2
1 1
Z Z
3. √ dx 8. dx
x4 x2 − 1 x2 − 4x + 6
1
Z
x
Z
4. dx 9. 5/2
dx
(16 − u2 )
5/2
(x2 − 6x + 15)
3x + 2
Z
1
Z
5. √ dx 10. 3/2
dx
4x2 − 9 (4 − 4x − x2 )
78 3.4 Partial Fraction Decomposition

3.4 Partial Fraction Decomposition

x+2 3 2 3
The rational function can be shown to be equal to − − 2 + . Therefore the integral
3
x −x 2 x x x−1
of the former rational function is:
x+2 3 2 3 2
Z Z  
dx = − − + dx = −3 ln |x| + + 3 ln |x − 1| + C
x3 − x2 x x2 x−1 x

This example suggests that determining a technique to decompose a rational function in this way would
provide a method for its integration.
A polynomial P (x) = a0 + a1 x + a2 x2 + . . . + an xn , an ̸= 0 is said to have degree n. A function
P (x)
f (x) = Q(x) where P (x) and Q(x) are polynomials is a rational function.

The rational number 74 is called improper because the numerator is larger than the denominator.
Through division of 4 into 7 one can write 74 as 1 34 where the fractional part, 34 is a proper fraction.
Analagous definitions are made for rational functions.

P (x)
Definition: A rational function f (x) = Q(x) is called proper if the degree of P is less than the degree
of Q. Otherwise f (x) is called improper if deg(P ) ≥ deg(Q).

Note:

1. If f (x) = P (x)/Q(x) is proper then it is possible to express it as a sum of simpler fractional

functions called partial fractions which are integrable.
2. If f (x) is improper, then use long division to divide P by Q until a remainder R(x) is obtained
such that deg R < deg Q. Then

P (x) R(x)
f (x) = = S(x) +
Q(x) Q(x)

where S(x) and R(x) are polynomials. S(x) is then integrable as it is a polynomial while the
proper rational function R(x)/Q(x) can in turn be integrated by the method of partial fractions
thereby making f (x) integrable.

Example 3-9
For the following rational functions determine if they are proper or improper. For those that are
improper write them as a polynomial plus a proper rational function.

x3 + 2 3x4 + 2x2 + x + 7
1. 2.
(x2 + 4)2 x2 + 2

Solution:

1. The numerator P (x) = x3 + 2 so deg P = 3. Expanding the denominator gives

Q(x) = (x2 + 4)2 = x4 + 8x2 + 16

showing that deg Q = 4 so deg P < deg Q and the rational function is proper.
Integration Methods 79

2. Since P (x) = 3x4 + 2x2 + x + 7 and Q(x) = x2 + 2 we have 4 = deg P ≥ deg Q = 2 so the
rational function P (x)/Q(x) is improper. Polynomial long division of P (x) by Q(x) gives

3x2 −4
x +2 3x + 2x + x + 7
2

4 2

− 3x4 − 6x2
− 4x2 + x + 7
4x2 +8
x + 15

and it follows that

3x4 + 2x2 + x + 7 x + 15
= 3x2 − 4 + 2 .
x +2
2 | {z } x +2
S(x) | {z }
R(x)/Q(x)

Further Questions:

For the following rational functions determine if they are proper or improper. For those that are
improper write them as a polynomial plus a proper rational function.

x+1 x2 + 1 x4 + 5x2 + 1
1. f (x) = 2. f (x) = 3. f (x) =
x3 − 3x2 + 2 x2 + 3x x2 + 2

Definition: Let g(x) = ax2 + bx + c be a quadratic function with real coefficients. If b2 − 4ac ≥ 0
then g(x) is called reducible because it can be written as a product of linear factors with real
coefficients. If b2 − 4ac < 0 then g(x) is called irreducible because it cannot be written as a
product of linear factors with real coefficients.
Example 3-10

1. The function g(x) = x2 + 5x + 6 has b2 − 4ac = 25 − 24 = 1 > 0 and so is reducible. It clearly

factors as g(x) = (x + 2)(x + 3).

2. The function g(x) = 2x2 + 4x + 5 has b2 − 4ac = 16 − 40 = −24 < 0 and is irreducible.

Note: It can be shown, as a consequence of the Fundamental Theorem of Algebra, that any polynomial
Q(x) with real coefficients can be factored as a product of linear factors of the form (ax + b) and/or
quadratic irreducible factors of the form ax2 + bx + c, where a, b, and c are real numbers.
Theorem 3-1: If P (x) and Q(x) are polynomials and deg P < deg Q the it follows that
P (x)
= F1 + F2 + . . . + Fn
Q(x)
where each Fi has one of the forms
A Ax + B
or
(ax + b)i (ax2 + bx + c)j
for some nonnegative integers i and j. The sum F1 + F2 + . . . + Fn is called the partial fraction
P (x)
decomposition of Q(x) and each Fi is called a partial fraction. The denominator polynomials are
real linear functions and irreducible quadratics respectively.
80 3.4 Partial Fraction Decomposition

Steps for finding Partial Fraction Decomposition

P (x)
To decompose f (x) = Q(x) into partial fractions do the following:

1. If deg P ≥ deg Q then use long division to get

P (x) R(x)
= S(x) +
Q(x) Q(x)

2. Express Q(x) as a product of linear and/or quadratic irreducible factors.

3. Express the proper rational function ( P (x)/Q(x) or R(x)/Q(x) ) as a sum of partial fractions of
the form
A Ax + B
and/or
(ax + b) i (ax + bx + c)j
2

4. Evaluate the constants.

Once the partial fraction decomposition has been accomplished the necessary integration may be
completed.
Upon factoring Q(x) there are four cases that are logically possible.

Case I: Q(x) contains a nonrepeated linear factor.

If Q(x) has a nonrepeated linear factor ax + b then the partial fraction decomposition will have the
following term due to that factor:
A
ax + b
where A is constant.
For example, suppose that
Q(x) = (a1 x + b1 )(a2 x + b2 ) . . . (ak x + bk )
where no linear factor is repeated. Then there exist constants A1 , A2 , . . . Ak such that
P (x)
 
R(x) A1 A2 Ak
or = + + ... +
Q(x) Q(x) a1 x + b1 a2 x + b2 ak x + bk
Example 3-11
Evaluate the following integrals:

x+9 x2 + 3x + 4
Z Z
1. dx 2. dx
x − 3x − 10
2
x(x + 1)(x − 2)

Solution:

1. The denominator of the rational function factors as x2 − 3x − 10 = (x − 5)(x + 2), two linear
distinct factors, so the partial fraction decomposition will have the form:
x+9 A B x+9 A x+2 B x−5
= + =⇒ = · + ·
(x − 5)(x + 2) x−5 x+2 (x − 5)(x + 2) x−5 x+2 x+2 x−5
x+9 A(x + 2) + B(x − 5)
=⇒ =
(x − 5)(x + 2) (x − 5)(x + 2)
Integration Methods 81

Here we combined the two fractions on the right by getting a common denominator. Equating
the numerators we have:
x + 9 = A(x + 2) + B(x − 5)
which must be true for all x. (Given the rational function is not defined at x = 5 or x = −2
one might argue these values should be excluded, but in this case the limits as x → −2 and
x → 5 on both sides of the equation would need to be equal, and, as these are polynomials,
this amounts to the functions themselves being equal at these two values.) Two methods can
be used to find A and B.
• Method 1: Choose x values and solve for constants.
Since the numerator equation is true for all x we can choose any x and solve for the
constants. The form of the equation suggests using x = −2 and x = 5 since then only
one constant remains:

x = −2 =⇒ −2 + 9 = A(−2 + 2) + B(−2 − 5)
=⇒ 7 = −7B =⇒ B = −1
x = 5 =⇒ 5 + 9 = A(5 + 2) + B(5 − 5)
=⇒ 14 = 7A =⇒ A = 2

• Method 2: Equate polynomial coefficients.

Since x + 9 = A(x + 2) + B(x − 5) is true for all x the polynomials on both sides must be
equal. Expanding the right-hand side and collecting like powers of x gives:

1x + 9 = (A + B)x + (2A − 5B)

Equating coefficients of like powers of x implies:

x0 : 9 = 2A − 5B
x1 : 1 = A + B

This is a linear system of two equations in two unknowns for which many strategies can be
used for solution. Multiplying the second equation by 2 gives 2 = 2A + 2B. Subtracting
that on both sides from the first equation gives

9 − 2 = 2A − 5B − (2A + 2B) =⇒ 7 = −7B =⇒ B = −1

Inserting B = −1 in the first equation gives for A:

9 = 2A − 5(−1) =⇒ 4 = 2A =⇒ A = 2

x+9 2 −1
Having A = 2 and B = −1 our decomposition is then = + and the
x2 − 3x − 10 x−5 x+2
integral is easily solved:

x+9 2 1
Z Z  
dx = − dx = 2 ln |x − 5| − ln |x + 2| + C ,
x2 − 3x − 10 x−5 x+2

where here we integrated term by term with substitution u = x − 5 and u = x + 2 respectively.

82 3.4 Partial Fraction Decomposition

2. Here we have three linear factors and the partial fraction decomposition will have the form:

x2 + 3x + 4 A B C
= + +
x(x + 1)(x − 2) x x+1 x−2

Get a common denominator on the right-hand side and equate numerators:

x2 + 3x + 4 A(x + 1)(x − 2) + Bx(x − 2) + Cx(x + 1)

=⇒ =
x(x + 1)(x − 2) x(x + 1)(x − 2)
=⇒ x2 + 3x + 4 = A(x + 1)(x − 2) + Bx(x − 2) + Cx(x + 1)

Evaluate the constants A, B, and C by evaluating the numerator equation at different values
of x:

x = 0 =⇒ 4 = −2A =⇒ A = −2
2
x = −1 =⇒ 1 − 3 + 4 = B(−1)(−3) =⇒ 2 = 3B =⇒ B =
3
14 7
x = 2 =⇒ 4 + 6 + 4 = C(2)(3) =⇒ 14 = 16C =⇒ C = =⇒ C =
6 3
Therefore:
x2 + 3x + 4 −2 2 1 7 1
Z Z  
dx = + · + · dx
x(x + 1)(x − 2) x 3 x+1 3 x−2
2 7
= −2 ln |x| + ln |x + 1| + ln |x − 2| + C
3 3

Further Questions:

Evaluate the following integrals:

1 4x2 + 13x − 9 4x2 + 3x + 1

Z Z Z
1. dx 2. dx 3. dx
x + 2x − 3
2
x3 + 2x2 − 3x x2 − 1

Case II: Q(x) contains a repeated linear factor.

If Q(x) has a factor (ax + b)r then the partial fraction decomposition will have the following terms due
to that factor:
A1 A2 Ar
+ + ... +
ax + b (ax + b) 2 (ax + b)r
where A1 , A2 , . . . Ar are constants. Use distinct constants (i.e. A, B, C, etc.) for each such factor.

Example 3-12
x4 + 1
The (proper) rational function decomposes into
x(3x + 2)3 (2x − 1)2

x4 + 1 A B1 B2 B3 C1 C2
= + + + + +
x(3x + 2) (2x − 1)
3 2 x 3x + 2 (3x + 2) 2 (3x + 2) 3 2x − 1 (2x − 1)2

where the constants A, B1 , B2 , B3 , C1 , and C2 would then have to be determined.

Integration Methods 83

Example 3-13
Evaluate the integral:
x2 + 4
Z
dx
x2 (2x + 1)
Solution:
Since the denominator has linear, repeated factors, the partial fraction decomposition will have the
following form:

x2 + 4 A1 A2 B
= + 2 +
x2 (2x
+ 1) x x 2x + 1
x +4
2
A1 x(2x + 1) + A2 (2x + 1) + Bx2
=⇒ 2 =
x (2x + 1) x2 (2x + 1)
=⇒ x2 + 4 = A1 x(2x + 1) + A2 (2x + 1) + Bx2

Evaluating at x = 0 and 2x + 1 = 0 =⇒ x = −1/2 are obvious choices for finding two of our
constants. Due to the repeated x factor we need to choose another arbitrary value to find the
remaining constant. We can use the previous results as shown below to find the last constant.

x = 0 =⇒ 0 + 4 = A1 (0) + A2 (0 + 1) + B(0)
=⇒ 4 = A2 =⇒ A2 = 4
1 1 1
x = − =⇒ + 4 = A1 (0) + A2 (0) + B
2 4 4
17 1
=⇒ = B =⇒ B = 17
4 4
x = 1 =⇒ 1 + 4 = A1 (1)(3) + A2 (3) + B
=⇒ 5 = 3A1 + 3(4) + 17
=⇒ 5 = 3A1 + 29 =⇒ 3A1 = −24 =⇒ A1 = −8

Note here we could have also used our second method of expanding the right-hand side of our
numerator equation and equating polynomial coefficients to get a linear system of equations:

A2 = 4


x2 + 4 = (2A1 + B)x2 + (A1 + 2A2 )x + A2 =⇒ A1 + 2A2 = 0
2A1 + B = 1


and solved to get A1 = −8, A2 = 4, B = 17 as before. Therefore

x2 + 4 4 17
Z  
−8
Z
dx = + + dx
x2 (2x + 1) x x2 2x + 1
4 17
= −8 ln |x| − + ln |2x + 1| + C
x 2
where we integrated term by term and used the substitution u = 2x + 1, so du = 2 dx, in the last
integral.
Further Questions:

Evaluate the following integrals:

x3 − 4x − 1 3x2 + 5x − 10
Z Z
1. dx 2. dx
x(x − 1)3 x2 (3x − 5)
84 3.4 Partial Fraction Decomposition

Case III: Q(x) contains a nonrepeated irreducible quadratic factor.

If Q(x) has a nonrepeated irreducible factor ax2 + bx + c (so b2 − 4ac < 0), then the partial fraction
decomposition will have the following term due to that factor:
Ax + B
ax2 + bx + c
where A and B are constants.
Example 3-14
Evaluate the integral:
2x4 + 7x2 + 6
Z
dx
x3 + 3x
Solution:

Since the degree of the numerator (4) is greater than or equal to the degree of the denominator (3),
this is an improper rational function. Performing polynomial long division one has
2x
x + 3x 2x + 7x + 6
3

4 2

− 2x4 − 6x2
x2 + 6
and the integral can be written as:
2x4 + 7x2 + 6 x2 + 6
Z Z  
dx = 2x + 3 dx
x3 + 3x x + 3x
The denominator can be factorized as x(x2 + 3). Here x2 + 3 = 1x2 + 0x + 3 is an irreducible
quadratic since b2 − 4ac = −12 < 0. Therefore, the partial fraction decomposition takes the form:
x2 + 6 A Bx + C
= + 2 =⇒ x2 + 6 = A(x2 + 3) + Bx2 + Cx
x(x2 + 3) x x +3
Because the factor of x is repeated and the irreducible quadratic vanishes for no real value of x the
second method of finding constants is preferred. Expanding the polynomial gives:
 A+B =1


=⇒ x2 + 6 = (A + B)x2 + Cx + 3A =⇒ C =0
3A = 6


The solution is straightforward:

3A = 6 =⇒ A = 2
C=0
A + B = 1 =⇒ B = 1 − A =⇒ B = 1 − 2 =⇒ B = −1
Therefore
2x4 + 7x2 + 6 2 1
Z Z  
x
dx = 2x + − 2 dx = x2 + 2 ln |x| − ln |x2 + 3| + C,
x3 + 3x x x +3 2
where substitution u = x2 + 3, so du = 2x dx, was used on the final term of the integral.
Further Question:

Evaluate the following integral:

x3 − 4x2 + 2
Z
dx
(x2 + 1) (x2 + 2)
Integration Methods 85

Case IV: Q(x) contains a repeated irreducible quadratic factor.

r
If Q(x) has an irreducible factor ax2 + bx + c then the partial fraction decomposition will have the
following terms due to that factor:
A1 x + B 1 A2 x + B2 Ar x + B r
+ + ... + r
ax + bx + c (ax + bx + c)
2 2 2
(ax2 + bx + c)

where A1 , A2 , . . . , Ar , and B1 , B2 , . . . , Br are constants.

Example 3-15
Evaluate the integral:
2x + 8
Z
dx
x(x2 + 4)2
Solution:

The repeated quadratic x2 + 4 is irreducible since 02 − 4(1)(4) = −16 < 0. The partial fraction
decomposition will therefore have the following form:
2x + 8 A B1 x + C1 B2 x + C2
= + + 2
x(x + 4)
2 2 x x +4
2 (x + 4)2
=⇒ 2x + 8 = A(x2 + 4)2 + (B1 x + C1 )x(x2 + 4) + B2 x2 + C2 x
=⇒ 2x + 8 = A(x4 + 8x2 + 16) + B1 x4 + 4B1 x2 + C1 x3 + 4C1 x + B2 x2 + C2 x
=⇒ 2x + 8 = (A + B1 )x4 + C1 x3 + (8A + 4B1 + B2 )x2 + (4C1 + C2 )x + 16A

Solving the system of equations generated by equating coefficients of xn gives:

1
16A = 8 =⇒ A =
2
1
A + B1 = 0 =⇒ B1 = −A =⇒ B1 = −
2
8A + 4B1 + B2 = 0 =⇒ 4 − 2 + B2 = 0 =⇒ B2 = −2
C1 = 0
4C1 + C2 = 2 =⇒ 4(0) + C2 = 0 =⇒ C2 = 2

Therefore
2x + 8 1 1 1 2x 2
Z Z  
x
dx = · − · − + 2 dx
x(x2 + 4)2 2 x 2 x2 + 4 (x2 + 4)2 (x + 4)2
1 1 2x 2
Z Z Z
x
= ln |x| − − dx + dx
2 2 x2 + 4 (x2 + 4)2 (x2 + 4)2

Note that the last partial fraction has been broken into two pieces anticipating the two different
integration techniques required. To evaluate the second and third integrals use the substitution
u = x2 + 4 so du = 2x dx :
1 1 1 1 1
Z Z
x
− dx = − du = − ln |u| + C = − ln |x2 + 4| + C
2 x2 + 4 4 u 4 4
2x 1 1 1
Z Z
− dx = − du = + C = 2 +C
(x2 + 4)2 u2 u x +4
86 3.4 Partial Fraction Decomposition

2
Z
To evaluate dx use the trigonometric substitution x = 2 tan θ so dx = 2 sec2 θ dθ :
(x2 + 4)2

2 1 sec2 θ 1 1
Z Z Z Z
dx = 2 · 2 sec2
θ dθ = 4 dθ = dθ
(x + 4)
2 2 (4 tan θ + 4)
2 2 16 sec θ
4 4 sec2 θ
1 1
Z Z
= cos2 θ dθ = (1 + cos 2θ) dθ
4 8
1 1 1 1
= θ+ sin 2θ + C = θ + sin θ cos θ + C
8 16 8 8
Note the identity for sin 2θ is used to get functions of only the angle θ.
Return to x by noting that, by the substitution tan θ = x2 , a triangle √
can be drawn with opposite side length of x and adjacent length x2 + 4
of 2. The remaining side is solved for using the Pythagorean Theorem x
and then sin θ and cos θ are evaluated using their definitions and
θ
the triangle.
In this integral θ itself is required and we note that
2
θ = tan−1 x2 by the original substitution. The integral becomes:

1 x 1 x 2
= tan−1 + ·√ ·√ +C
8 2 8 x2 + 4 x2 + 4
1 x 1 x
= tan−1 + · 2 +C
8 2 4 x +4
Putting all the components together gives for the original integral:
2x + 8 1 1 1 1   1
Z
−1 x x
dx = ln |x| − ln |x 2
+ 4| + + tan + · 2 +C
x(x + 4)
2 2 2 4 x +4 8
2 2 4 x +4
1 1 1 x+4 1 x
= ln |x| − ln |x2 + 4| + · 2 + tan−1 +C
2 4 4 x +4 8 2

Further Question:

Evaluate the following integral:

2x6 + 5x4 + 2x2 + 1
Z
2 dx
x (x2 + 1)

Note the following:

• Use distinct constants (i.e. A, B, C, etc.) in each partial fraction.

• Factors in Q(x) are considered the same if they differ only by a multiplicative constant. For
example (3x − 1) and (x − 1/3) are repeated linear factors.
Integration Methods 87

Having looked at all the cases we can write down the partial fraction decomposition of arbitrary rational
functions.

Example 3-16
After factoring the denominator Q(x) suppose a (proper) rational function equals

x2 + 5x + 1
3 2 .
x (x − 1) (3x + 2) (x2 + 2x + 4) (x2 + 9)

Then it will have the following partial fraction decomposition

A B C1 C2 C3 Dx + E F1 x + G1 F2 x + G2
= + + + + + 2 + + 2 .
x x − 1 3x + 2 (3x + 2)2 (3x + 2)3 x + 2x + 4 x2 + 9 (x + 9)2

Here we had two unrepeated linear factors, x and (x − 1), a repeated linear factor (3x + 2)3 , an
unrepeated irreducible quadratic factor (x2 + 2x + 4), and a repeated irreducible quadratic factor
(x2 + 9)2 .
We would now proceed to solve for the constants (A, B, C1 , C2 , C3 , D, E, F1 , G1 , F2 , and G2 ) by
combining the partial fractions and equating numerators as discussed above. Once this was done
the original rational function could be integrated by integrating the partial fraction decomposition
term by term.
Further Questions:

Write down the form of the partial fraction decomposition of the following rational functions. Do
not evaluate the constants.

x2 − x − 21 x6 + 5x3 + x − 1
1. 3.
2x3 − x2 + 8x − 4 x4 + 5x2 + 4
x3 + 2 x+1
2. 2 4. 2
(x2 + 4) (x2 − 4) (x2 + 3)

Rationalizing Substitutions

Some nonrational functions can be changed into rational functions by means of a substitution.

Example 3-17
Evaluate the integral:
1
Z
dx
e2x − 1
Solution:
2 1
Noting e2x = (ex ) we try the rationalizing substitution u = ex so du = ex dx =⇒ du = dx :
u
1 1 1 1
Z Z Z
dx = · 2 du = du
e −1
2x u u −1 u(u + 1)(u − 1)
 88 3.4 Partial Fraction Decomposition

The partial fraction decomposition takes the form:

1 A B C
= + +
u(u + 1)(u − 1) u u+1 u−1
=⇒ 1 = A(u + 1)(u − 1) + Bu(u − 1) + Cu(u + 1)

Evaluate coefficients by choosing u as follows:

u = 0 =⇒ 1 = −A =⇒ A = −1
1
u = 1 =⇒ 1 = C(1)(1 + 1) =⇒ C =
2
1
u = −1 =⇒ 1 = B(−1)(−1 − 1) =⇒ B =
2
Therefore:
1 1 1 1 1 1 1
Z Z Z  
dx = du = − + · + · du
e2x − 1 u(u + 1)(u − 1) u 2 u+1 2 u−1
1 1
= − ln |u| + ln |u + 1| + ln |u − 1| + C
2 2
1 1
= − ln |e | + ln |e + 1| + ln |ex − 1| + C
x x
2 2
1 1
= −x + ln |e + 1| + ln |ex − 1| + C
x
2 2

Further Question:

Evaluate the following integral: √

4 x
Z
dx
x−2

Answers: Exercise 3-4

Page 177
1-10: Evaluate the given integral.

x4 + x2 + 1
Z
x
Z
1. dx 6. dx
x − 5x + 6
2
(x − 1)
3

x +1
Z 3
10
Z
2. dx 7. dx
x2 − 9 x2 (x2 + 5)
3x − 1
Z
x6 + x3 − 6
Z
3. 2 dx 8. dx
(x − 1) (x + 2) x4 + 3x2
2x + 4 x2 + 2x + 5
Z Z
4. dx 9. dx
(x + 1) (x2 + 1) x4 + 4x2 + 3
x−9
Z
2x2 − 4x + 4
Z
5. 2 dx 10. dx
x (x2 + 3) x3− x2 + x − 1
Integration Methods 89

3.5 General Strategies for Integration

Unlike differentiation which is largely a deterministic application of rules, integration is an art, with
many indefinite integrals not even having an antiderivative that may be written in terms of known
functions.
The following basic strategies have been seen

1. Basic Formulas of Integration

2. Substitution
3. Integration by Parts
4. Trigonometric Integrals

5. Trigonometric Substitution
6. Partial Fraction Decomposition
7. Rationalizing Substitution

One or more of these strategies along with using functional identities to rewrite the integrand may
need to be applied to evaluate an integral.

Example 3-18
Evaluate the following integrals:

1 1 ex
Z Z Z
1. dx 2. 1 dx 3. √ dx
x2 + 6x + 16 3
x 2 − 4x 2 e2x + 2ex − 5

Solution:

1. We observe that 62 − 4(1)(16) = −28 < 0 indicates that x2 + 6x + 16 is an irreducible quadratic

so this rational function cannot be decomposed further. To integrate it, first complete the
square to remove the x term.
1 1 1
Z Z Z
dx = dx = dx
x2 + 6x + 16 (x + 3)2 − 9 + 16 (x + 3)2 + 7

Next substitute u = x + 3 so du = dx and the integral becomes:

1 1 1 1 1
Z Z Z
= du = du = du
u +7
2 7 u
7 +1
2
7 ( 7) + 1
√u 2

u du √
Let w = √ so dw = √ =⇒ du = 7dw to get our integrable form:
7 7
√ Z
7 1 1 1 1 −1 x + 3
   
u
= dw = √ tan −1
(w) + C = √ tan −1
√ + C = √ tan √ +C
7 w2 + 1 7 7 7 7 7

√
Alternatively we could have used the general integral 1
= 1
tan−1 ( xa ) with a = 7 to
R
u2 +a2 a
avoid the w substitution.
90 3.5 General Strategies for Integration

1 1
2. Perform the rationalizing substitution u = x 2 so du = 12 x− 2 dx =⇒ 2u du = dx:

1 1 2 2
Z Z Z Z
dx = · 2u du = du = du
3 1
x 2 − 4x 2 u3 − 4u u2 − 4 (u + 2)(u − 2)

The partial fraction decomposition is

2 A B 2 A(u − 2) + B(u + 2)
= + =⇒ =
(u + 2)(u − 2) u+2 u−2 (u + 2)(u − 2) (u + 2)(u − 2)
=⇒ 2 = A(u − 2) + B(u + 2)

Evaluate the constants:

1
u = 2 =⇒ 2 = 4B =⇒ B =
2
1
u = −2 =⇒ 2 = −4A =⇒ A = −
2
Therefore
1 2 1 1 1 1
Z Z Z  
1 dx = du = − · + · du
3
x − 4x 2
2 (u + 2)(u − 2) 2 u+2 2 u−2
1 1 1 √ 1 √
= − ln |u + 2| + ln |u − 2| + C = − ln | x + 2| + ln | x − 2| + C
2 2 2 2

3. Let u = ex so du = ex dx and then complete the square:

ex 1 1 1
Z Z Z Z
√ dx = √ du = dx = dx
e + 2e − 5 u + 2u − 5 (u + 1) − 1 − 5 (u + 1)2 − 6
p p
2x x 2 2

Let w = u + 1 so dw = du :
1
Z
= √ dw
w2 − 6
√ √ √
For w2 − a2 do the trigonometric substitution w = 6 sec θ so dw = 6 sec θ tan θ dθ :
Z √
1 √ 6 tan θ sec θ
Z Z
= √ · 6 sec θ tan θ dθ = √ dθ = sec θ dθ = ln | sec θ + tan θ| + C
6 sec2 θ − 6 6 tan θ
To return to w note that the substitution implies sec θ = √w6 . To
find tan θ a triangle can be drawn, √ since secant is the reciprocal of
cosine, with adjacent side of length 6 and hypotenuse of length w.
√
The remaining side is solved for using the Pythagorean Theorem and w w2 − 6
then the result follows by evaluating the opposite side length over the
θ
adjacent side length. Finally we return to u using w = u + 1 and to x √
6
using u = e :
x

√
w2 − 6 u+1 (u + 1)2 − 6
p
ex
Z
w
√ dx = ln √ + √ + C = ln √ + √ +C
e + 2e − 5
2x x 6 6 6 6
√
ex + 1 e2x + 2ex − 5 p
= ln √ + √ + C = ln ex + 1 + e2x + 2ex − 5 + D
6 6
√ √
In the final simplification we used that ln |a/ 6| = ln |a| − ln( 6) and then combined the
latter constant with C to get a new arbitrary constant D.
Integration Methods 91

Further Questions:

Evaluate the following integrals:

e3t e3x
Z Z
1. dt 6. dx
1 + e6t 1 + ex
cos3 x
Z Z
x
2. ex+e dx 7. √ dx
1 + sin x
1 + ex
Z Z
x
3. dx 8. dx
1 − ex csc (5x2 )
Z Z
4. x ln(1 + x) dx
2
9. (2x + 2x + 2π ) dx

7x2 + 20x + 65
Z Z
5. tan x sec x dx
6
10. dx
x4 + 4x3 + 13x2

Answers:
Exercise 3-5
Page 177

1-10: Evaluate the given integral.

√
sin 2x
Z
e x
Z
1. √ dx 6. dx
x sin2 x − sin x − 6
x3 + 5x
Z
√
Z
2. sin x + 5 dx 7. 2 dx
(x2 + 1)
ex
Z
e2x
Z
3. dx 8. √ dx
3 + ex e2x + 4ex + 6
e3x 1
x+2
Z Z
4. dx 9. dx
5 + e6x 0 x2 + 2x + 3
4
1 4x
Z Z
5. √ dx 10. 2 dx
1 3+ x (x2 + 2x + 9)
92 3.6 Improper Integrals

3.6 Improper Integrals

The definite integral due to Riemann which we use involves functions integrated over a closed interval
[a, b]. Functions which are piecewise continuous where there are only a finite number of jump disconti-
nuities are integrable. We now consider improper integrals where these restrictions do not hold. We
consider two cases:

Improper Integrals of the First Kind : The interval of integration is infinite.

Improper Integrals of the Second Kind : The interval of integration contains an infinite discon-
tinuity.

We can define definite integrals under these circumstances by considering suitable limits of integrals
over closed intervals.

3.6.1 Improper Integrals of the First Kind

Suppose we wish to find the area under the curve y = 1

x3 over the interval [1, ∞) shaded in the following
diagram.

1
y= x3

a=1 x

Intuitively one would find the area by evaluating the area under the curve (the definite integral) over
the closed interval [1, t], and then consider the limit of that as t → ∞ :

1
y= x3

a=1 t x

Should such a (finite) limit exist we would define that to be the area under the curve over the open
interval [1, ∞).
The previous discussion prompts the following definition for the improper integral over an infinite
interval [a, ∞) and, similarly, over intervals (−∞, b], and (−∞, ∞).
Integration Methods 93

Definition: Define the following improper integrals of the first kind :

Z ∞ Z t
a) f (x) dx = lim f (x) dx (Where the latter integrals must exist for every t ≥ a.)
a t→∞ a
Z b Z b
b) f (x) dx = lim f (x) dx (Where the latter integrals must exist for every t ≤ b.)
−∞ t→−∞ t
Z ∞ Z a Z ∞
c) f (x) dx = f (x) dx + f (x) dx (Where a is any real number.)
−∞ −∞ a

The improper integrals in a) and b) are convergent if the limit exists (i.e. is finite) and divergent
otherwise. For c) the integral is convergent if and only if both integrals on the right side are
convergent.
Z ∞ Z t
Note that f (x) dx is not defined to be lim f (x) dx. The integral over (−∞, ∞) by definition
−∞ t→∞ −t
must be broken into two pieces for which independent limits must be taken.

Example 3-19
Determine whether the following integrals are convergent or divergent and evaluate those that are
convergent.
Z ∞ Z ∞
1 ex
Z 0
1. e−4x dx 2. dx 3. 2x + 1
dx
0 −∞ 3 − 2x −∞ e

Solution:
These are all improper integrals of the first kind since the x values are approaching ∞, −∞, or both.
Z ∞ Z t
1. e−4x dx = lim e−4x dx
0 t→∞ 0

1
Substitute u = −4x so du = −4dx =⇒ − du = dx
4
Limits: x = 0 =⇒ u = −4(0) = 0, x = t =⇒ u = −4t

The integral becomes:

Z −4t 
1 1 1

−4t
= lim eu − du = − lim eu =− lim e−4t − e0
  
t→∞ 0 4 4 t→∞ 0 4 t→∞
1 1 1 1
 
= − lim − 1 = − [0 − 1] =
4 t→∞ e 4t 4 4

Here we used the known properties of the exponential function, that lim ex = ∞, to evaluate
x→∞
the limit. Therefore the improper integral is convergent with value 14 .

1 1
Z 0 Z 0
2. dx = lim dx
−∞ 3 − 2x t→−∞ t 3 − 2x

1
Substitute u = 3 − 2x so du = −2dx =⇒ − du = dx
2
Limits: x = t =⇒ u = 3 − 2t, x = 0 =⇒ u = 3 − 2(0) = 3
94 3.6 Improper Integrals

The integral becomes:

3
13
1 1
Z  
= lim · − du = − lim ln |u|
t→−∞ 3−2t u 2 2 t→−∞ 3−2t
1
= − lim [ln |3| − ln |3 − 2t|]
2 t→−∞
1
= − [ln 3 − ∞] = ∞
2
Here we used the property of the logarithm, that lim ln x = +∞, to evaluate the limit.
x→∞
Therefore the integral is divergent.
3. For this integral we must first break the infinite domain into two separate integrals:
Z ∞ Z 0 Z ∞
ex ex ex
2x + 1
dx = 2x + 1
dx + dx
−∞ e −∞ e 0 e2x + 1
Since the integrand is the same for both integrals, let us first evaluate the following indefinite
integral with the substitution u = ex so du = ex dx :
ex 1
Z Z
dx = du = tan−1 u + C = tan−1 (ex ) + C
e2x + 1 u2 + 1
The integrals become:
Z 0 Z 0
ex ex h 0
dx = lim dx = lim tan−1 (ex )
−∞ e
2x +1 t→−∞ t e 2x +1 t→−∞ t

= lim tan (e ) − tan (e ) = lim tan (1) − tan−1 (et )

 −1 0 −1 t
  −1 
t→−∞ t→−∞
π π
= − tan−1 (0) =
4 4
∞ Z t
ex ex
Z h t
dx = lim dx = lim tan−1 (ex )
0 e +1
2x t→∞ 0 e 2x +1 t→∞ 0

= lim tan−1 (et ) − tan−1 (e0 ) = lim tan−1 (et ) − tan−1 (1)
 
t→∞ t→∞
π π π
= − = ,
2 4 4
π
where we used that lim tan x = . The original integral therefore is
−1
x→∞ 2
Z ∞
ex π π π
dx = + =
−∞ e
2x +1 4 4 2
and hence convergent.

Further Questions:

Determine whether the following integrals converge or diverge. Find the value of any convergent
integral.
Z ∞ Z ∞
1 1
1. dx 4. dx
−∞ 1 + x
x 3 2
1
Z ∞
1 Z 0
2.
2

x−1
dx 5. xe−x dx
2 −∞
Z 0 Z ∞
ln x
3. xex dx 6. dx
−∞ 1 x
Integration Methods 95

3.6.2 Improper Integrals of the Second Kind

In the second case we consider those situations where the function being integrated has an infinite
discontinuity at some point over which we want to integrate. Consider the area under the curve
1
y=√ between x = 1 and x = 5. The situation is shown in the following diagram.
5−x

y= √1
5−x

a=1 b=5 x

The function has an infinite discontinuity at the right endpoint (b = 5). Intuitively we can imagine
finding the area under the curve over the closed interval [1, t] with t < b and then consider the limit as
t→b:

y= √1
5−x

a=1 t b=5 x

This discussion suggests the following definition for improper integrals involving infinite integrands.
Our example illustrated an integral where the right endpoint had the discontinuity. Similarly integrals
with a discontinuity at the left endpoint or within the interval are defined.

Definition: Define the following improper integrals of the second kind :

a) Suppose f (x) is continuous on [a, b) but discontinuous at x = b then:
Z b Z t
f (x) dx = lim f (x) dx
a t→b− a
b) Suppose f (x) is continuous on (a, b] but discontinuous at x = a then:
Z b Z b
f (x) dx = lim+ f (x) dx
a t→a t
96 3.6 Improper Integrals

c) Suppose f (x) is continuous on [a, b] except at a value c in (a, b) then:

Z b Z c Z b
f (x) dx = f (x) dx + f (x) dx
a a c

The improper integrals in a) and b) are convergent if the limit exists (i.e. is finite) and divergent
otherwise. For c) the integral is convergent if and only if both integrals on the right side are
convergent.
Example 3-20
Determine whether the following integrals are convergent or divergent. Find the value of any
convergent integral.
1
ln x 0 2
Z Z Z
x x
1. dx 2. √ dx 3. dx
0 x2 −4 x+4 0 x−1

Solution: These are improper integrals of the second kind and we need first to identify where the
integrand is discontinuous.

1. Here the integrand is undefined at 0 both because ln x is undefined there and also because we
cannot divide by zero. The improper integral is therefore defined by

ln x ln x
Z 1 Z 1
2
dx = lim+ dx
0 x t→0 t x2
Use Integration by Parts to evaluate the integral:

u = ln x dv = x−2 dx
1 1
=⇒ du = dx v = −
x x
The integral then equals
" #
1
1
ln x ln x
1
ln x 1
Z Z Z 1
dx = lim+ dx = lim+ − + dx
0 x2 t→0 t x2 t→0 x t t x
2

1
ln x 1 − ln 1 1 ln t 1
  
= lim+ − − = lim+ − + +
t→0 x x t→0 1 1 t t
 t
ln t 1

= −0 − 1 + lim+ + (−∞ + ∞ form)
t→0 t t
ln(t) + 1  ∞ 
= −1 + lim+ form
t→0 t ∞
1/t
= −1 + lim+ =∞
LH
t→0 1

Therefore the integral is divergent.

2. The integrand is discontinuous at x = −4 so the improper integral is defined by
Z 0 Z 0
x x
√ dx = lim + √ dx
−4 x + 4 t→−4 t x +4

Substitute u = x + 4 so du = dx, x = u − 4
Limits : x = 0 =⇒ u = 0 + 4 = 4, x = t =⇒ u = t + 4
Integration Methods 97

The integral becomes:

u−4
Z 0 Z 0 Z 4 Z 4 
x x √ − 21

√ dx = lim + √ dx = lim + √ du = lim u − 4u du
−4 x+4 t→−4 t x+4 t→−4 t+4 u t→−4+ t+4
4
2 3 2 2
  
1 3 1 3 1
= lim + u 2 − 8u 2 = lim + (4) 2 − 8(4) 2 − (t + 4) 2 + 8(t + 4) 2
t→−4 3 t+4 t→−4 3 3
2 2 3 1 16 16 − 48
= (8) − 8(2) − (−4 + 4) 2 + 8(−4 + 4) 2 = − 16 − 0 + 0 =
3 3 3 3
32
=−
3
Therefore it is convergent.
3. The integrand is undefined at x = 1 so we must break the improper integral into two integrals:
Z 2 Z 1 Z 2
x x x
dx = dx + dx
0 x − 1 0 x − 1 1 x − 1

Evaluate the first integral:

Z 1 Z t
x x
dx = lim dx (← an improper rational function)
0 x − 1 t→1 −
0 x − 1
x−1+1 x−1 1
Z t Z t 
= lim dx = lim + dx
t→1− 0 x−1 t→1− 0 x−1 x−1
1
Z t  h t
= lim− 1+ dx = lim− x + ln |x − 1|
t→1 x − 1 t→1 0
h0 i
= lim− t + ln |t − 1| − 0 − ln | − 1|
t→1
= 1 − ∞ − 0 − 0 = −∞ ,

where here we used that lim+ ln x = −∞. Since the fist integral is divergent the given integral
x→0
Z 2
x
dx is also divergent. (There is no need to evaluate the second composite integral.)
0 x−1

Further Questions:

Determine whether the following integrals converge or diverge. Find the value of any convergent
integral.
5
1 7
1
Z Z
1. √ dx 3. 2 dx
1 5−x −2 (x + 1) 3
1 2
1
Z Z
2. x ln x dx 4. dx
0 0 x2 − 4x + 3
98 3.6 Improper Integrals

A consideration of the areas represented by improper integrals in the following diagram makes the
following theorem plausible:

y y = f (x)

y = g(x)

a x

Theorem
R∞ 3-2: Let f and g beR ∞continuous functions satisfying
R∞ f (x) ≥ g(x) ≥ 0 for Rall x ≥ a. If
∞
a
f (x) dx is convergent then a
g(x) dx is convergent. If a
g(x) dx is divergent then a f (x) dx is
divergent.
Analagous theorems for the infinite intervals (−∞, b] and (−∞, ∞) as well as for improper integrals of
the second kind may also be written. The theorems are useful for determining convergence or divergence
of functions that are difficult to integrate.

Example 3-21
Determine whether the following integrals are convergent or divergent.
Z ∞ Z 1 x
dx e
1. √ 2. dx
1 x4 + 5 0 x 2

Solution:
Z ∞
dx
1. The integrand in the improper integral √ has no obvious antiderivative. However,
1 x4 + 5
for x ≥ 1 we have
p √ 1 1
x4 + 5 ≥ x4 = x2 > 0 =⇒ 0 < √ ≤ 2.
x4 + 5 x

The following integral

∞ t
1 t
1 1 1 1
Z Z   
dx = lim dx = lim − = lim − + =0+1=1
1 x2 t→∞ 1 x2 t→∞ x 1
t→∞ t 1

is convergent. Thus, since our desired positive integrand lies below a function whose integral
converges we have, by the Comparison Test for Integrals, that the given integral is convergent.

1
ex
Z
2. The integral 2
dx is improper since the integrand is undefined at zero. Since the integrand
0 x
has no obvious antiderivative we try a comparison with an integrable function. For 0 ≤ x ≤ 1
we have
ex 1
ex ≥ 1 > 0 =⇒ 2 ≥ 2 > 0 .
x x
Integration Methods 99

The following integral

1
1
1 1
1 1 1
Z Z   
dx = lim+ dx = lim − = lim+ −1 + = −1 + ∞ = ∞
0 x2 t→0 t x 2 t→0 + x t t→0 t

is divergent. Since our desired integrand lies above a positive function whose integral diverges,
we have, by the Comparison Test for Integrals, that the given integral is also divergent.

Further Questions:

Determine whether the following integrals are convergent or divergent.

Z ∞
1
Z 1 −x
e
1. √ dx 2. 2 dx
1 x3 + 1 0 x3

Answers:
Exercise 3-6
Page 178

1-10: Determine whether the given integral is convergent or divergent.

Z ∞
1 1
Z 5
1. dx 6. √ dx
3 (x − 2)2
0 5−x
Z ∞
1 Z 2
1
2. dx 7. dx
5 x−4 1 (x − 1)2/3
Z 0
1
4
Z 2
3. x3 e−x dx 8. dx
0 x − 4x + 3
−∞ 2
Z ∞
1
Z e
x
4. 4 + 16
dx 9. dx
x 1 x (ln x)
2
−∞

1
Z 0 Z π/2
5. 2 − 4x + 3
dx 10. tan2 x dx
−∞ x 0
 100 Review Exercises

Answers:
Chapter 3 Review Exercises
Page 178

1-12: Evaluate the given integral.

1
Z Z
1. x tan−1 x dx 7. √ dx
x2 + 6x + 12
Z
ln 2
e3x
Z
2. tan x sec3 x dx 8. dx
0 1 + ex
2
4x2 − 8x − 6
Z Z
3. ln(2 + x) dx 9. dx
0 (x − 1)(x − 2)(x − 3)
1
Z
6x3 − 4x2 + 5
Z
4. 5/2
dx 10. dx
(x2 + 36) x4 + 5x2 + 4
x−2
Z Z
5. 5 dx 11. x5/2 ln x dx
(x + 1)
cos x
Z
4
Z
6. dx 12. p dx
x3 + 2x 1 + sin2 x

13-15: Determine whether the given integral is convergent or divergent.

Z ∞ Z ∞
1
13. xe −3x
dx 15. dx
0 −∞ 5 + x2
sin x
Z π
14. √ dx
0 1 + cos x
Chapter 4: Sequences and Series

101
102 4.1 Sequences

4.1 Sequences

Definition: An ordered list of numbers:

{a1 , a2 , a3 , . . . , an , . . .}

is called a sequence. The numbers are called terms with a1 here being the first term, and, more
generally, an being the nth term in the sequence.

The above sequence may represented by the compact notation {an } or sometimes with the index limits
∞
made explicit as {an }n=1 . An explicit index is useful if we start enumerating the sequence from a
value other than 1.
Some texts will distinguish finite and infinite sequences depending on whether the sequence terminates
or not. For our purposes we will be assuming infinite sequences unless otherwise noted.
An equivalent way of thinking of a sequence is as a function f whose domain is the positive integers. In
this case an = f (n). Writing an as just such a function of the index is a convenient way of representing
a sequence.

Example 4-1
The following are several ways to represent the same sequences.

∞
1. 1, 2 , 3 , . . . , n1 , . . . = n1 = n1 , an =
 1 1   1
n=1 n n o∞
Note this sequence could also have been represented by 1
n+1
n=0
n n+1 2
o n n+1 2
o∞ n+1
(−1) n n n2
2. 1 −4 9
2, 5 , 8, . . . , 3n−1 , . . . = (−1)
3n−1 , an = (−1)
3n−1
n=1
∞
3. {4, 4, 4, . . . , 4, . . .} = {4}n=1 , an = 4

A sequence may not have a simple defining function in terms of the index n.

Example 4-2
The sequence generated by the digits of π = 3.14159 . . .

{3, 1, 4, 1, 5, 9, . . .}

is not representable by a simple function f (n).

Since theoretically any sequence is a function an = f (n) on the set of positive integers we can graphically
represent it by plotting the coordinate points (n, an ).

Example 4-3
cos(nπ) + n2
 
A graph of the sequence is as follows:
2n2
Sequences and Series 103

an 0.8

0.6
1
L= 2

0.4

0.2

0
0 1 2 3 4 5 6 7 8 9 10
n

The above graph clearly approaches the value 1/2 as n gets large. In symbols we would write
cos(nπ) + n2 1
lim =
n→∞ 2n 2 2
This limit is analagous to the limit of a function f (x) as x → ∞ with the only difference being that n
is restricted to positive integers. This discussion motivates the following definition for the limit of a
sequence.1

Definition: If the terms an of sequence {an } get arbitrarily close to the value L for sufficiently large n
then we say the sequence converges to limit L or is convergent with limit L. Symbolically
an → L as n → ∞ or
lim an = L .
n→∞
If a sequence is not convergent (i.e. it has no limit) then the sequence diverges or is divergent.

A divergent sequence may have a trend to infinity.2

Definition: If the terms an of sequence {an } get arbitrarily large (positively) for sufficiently large n
we say that the sequence {an } diverges to infinity and we write
lim an = ∞
n→∞

An analagous definition holds for a sequence to diverge to −∞.

Example 4-4
The Fibonacci Sequence satisfies a1 = 1, a2 = 1 and an = an−2 + an−1 for n > 2, i.e.

{1, 1, 2, 3, 5, 8, 13, 21, . . .}

 
The sequence diverges to ∞ lim an = ∞ .
n→∞

1 A more rigorous definition of the limit of a sequence is that a → L as n → ∞ if and only if for any ϵ > 0 there
n
exists m > 0 such that n > m implies |an − L| < ϵ.
2 A more rigorous definition for lim a = ∞ is that for any M > 0 there exists an index m > 0 such that n > m
n
n→∞
implies an > M .
104 4.1 Sequences

The limit of a sequence with an = f (n) is essentially the limit of f (x) as x → ∞ with x restricted to
the positive integers (instead of the continuous real axis).
Example 4-5
cos(xπ) + x2
If we plot y = f (x) = over our earlier sequence we have:
2x2

y 0.8

0.6
1
L= 2

0.4

0.2

0
0 1 2 3 4 5 6 7 8 9 10
x

The limit of f (n), with n an integer, clearly cannot differ from that of f (x) if the latter exists, thereby
leading to the following theorem.
Theorem 4-1: If lim f (x) = L then the limit of sequence {an } with an = f (n) is also L,
x→∞

lim an = L .
n→∞

(Note the converse of this theorem is not true, lim an = L ̸⇒ lim f (x) = L . )
n→∞ x→∞

For a sequence which converges to L = 0 we have the following result:

Theorem 4-2: lim an = 0 if and only if lim |an | = 0.
n→∞ n→∞

These theorems are convenient for evaluating the limits of certain sequences.
Example 4-6
Find the limit of the following sequences:

n2 + 4 n + ln n ln n
   2   
1. 2. 3. (−1) · 2
n
3n2 + 5n + 1 4n2 n

Solution:

1. To evaluate the limit pull out the highest power of n (here n2 ) from each of the numerator
and denominator.
n2
n2 + 4 n2 + 4
1 + n42 1+0 1
lim = lim 2 · n2 n2
= lim 1 · = =
n→∞ 3n + 5n + 1
2 n→∞ n 3n2
+ 5n
+ 1 n→∞ 3 + n5 + n12 3+0+0 3
n2 n2 n2
Sequences and Series 105

n2 + ln n
2. Here the nth term an = f (n) where f (n) = . Considering the function as a function
4n2
x + ln x
2
of a continuous variable x, so f (x) = , we have:
4x2
x2 + ln x  ∞ 
lim f (x) = lim form
x→∞ x→∞ 4x2 ∞
2x + x1  ∞ 
= lim form
LH
x→∞ 8x ∞

2 − x12 2−0 1
= lim = =
LH
x→∞ 8 8 4

n2 + ln n 1
Therefore lim = by Theorem 4-1.
n→∞ 4n2 4
(−1)n ln n
3. Consider the limit of the absolute value of an = :
n2
ln n ln n  ∞ 
lim |an | = lim (−1)n · = lim form
n→∞ n→∞ n2 n→∞ n2 ∞

ln x
Letting f (x) = , a continuous function of x, we have
x2

ln x  ∞  1
1
lim f (x) = lim form = lim x
= lim =0
x→∞ x→∞ x 2 ∞ x→∞ 2x x→∞ 2x2
LH

ln n
=⇒ lim |an | = lim = 0 (by Theorem 4-1)
n→∞ n2
n→∞
ln n
=⇒ lim an = lim (−1)n 2 = 0 (by Theorem 4-2)
n→∞ n→∞ n

Further Questions:

Find the limits of the following sequences:

2n 5n (−1)n (3n + 1)
     
1. 2. 3.
5n − 3 e2n n2 + 5

Theorem 4-3: Given sequences {an } and {bn } are convergent, c is a constant, and f a function defined
at an and continuous at L = lim an , then the following hold:
n→∞

1. lim c = c an lim an
n→∞ 5. lim = n→∞
n→∞ bn lim bn
2. lim (an ± bn ) = lim an ± lim bn n→∞
n→∞ n→∞ n→∞

3. lim can = c lim an (Here we require lim bn ̸= 0.)

n→∞
n→∞ n→∞
     
4. lim (an · bn ) = lim an · lim bn 6. lim f (an ) = f lim an
n→∞ n→∞ n→∞ n→∞ n→∞
106 4.1 Sequences

By the last item of the previous theorem applied to f (x) = xk one has the corollary
Theorem 4-4: Given non-negative sequence {an } (i.e. an ≥ 0) and power k > 0 one has
 k
lim (an )k = lim an .
n→∞ n→∞

Here the sequence must have an ≥ 0 for (an )k to be defined and the power k cannot be negative to
accommodate sequences for which lim an = 0.
n→∞

∞
Definition: A sequence is geometric if it has the form {arn }n=0 = a, ar, ar2 , ar3 , . . . , arn , . . .


(a ̸= 0). Here a is the first term in the sequence and r is called the common ratio since
an+1
an = r (constant) for a geometric sequence.

∞
One notes that a geometric sequence is of the form arn−1 if we start with index n = 1.

n=1

The following theorem results from consideration of the behaviour of the limit of the exponential function
limx→∞ rx when r ≥ 0 in Theorem 4-1 and use of Theorem 4-2 noting that limn→∞ |rn | = limn→∞ |r|n
when −1 < r < 0.
∞
Theorem 4-5: The geometric sequence {arn }n=0 = a, ar, ar2 , ar3 , . . . , arn , . . . (a =
̸ 0) is convergent


when −1 < r ≤ 1 with

0 if −1 < r < 1

lim arn = .
n→∞ a if r = 1
It is divergent for all other values of r, diverging to infinity (∞) for r > 1.
Example 4-7
Determine whether the following sequences are divergent or convergent. For convergent sequences
determine the limit.
  n ∞
1 2n + ln n
 
2. {3(−5)n } 3.
1. 5
10 n=0
5n + 3 ln n

3 Solution:

1 1
1. This is a geometric sequence with a = 5 and r = . Since |r| = < 1 it is convergent with
10 10
 n
1
lim 5 = 0.
n→∞ 10

∞
2. Without explicit index labels we infer {3(−5)n } = {3(−5)n }n=1 . We can put it in standard
form for a geometric sequence starting at n = 1 by rewriting it:

∞ ∞ ∞
{3(−5)n }n=1 = 3(−5)(−5)n−1 = −15(−5)n−1
 
n=1 n=1
.
This is a geometric sequence with a = −15 and r = −5. Since r = −5 < −1 it is divergent.
Alternatively we can show the sequence is geometric by evaluating

an+1 3(−5)n+1
= = −5
an 3(−5)n

proving it has a constant (common) ratio of r = −5. The first term is a = 3(−5)1 = −15 .
Sequences and Series 107

2x + ln x
3. Let f (x) =
5x + 3 ln x

2x + ln x 2 + x1 2+0 2
lim f (x) = lim = lim = =
x→∞ x→∞ 5x + 3 ln x x→∞ 5 + 3 5+0 5
LH ln x

2n + ln n 2
=⇒ lim =
n→∞ 5n + 3 ln n 5
Therefore the sequence is convergent.

Further Questions:

Determine whether the following sequences are divergent or convergent. For convergent sequences
determine the limit.
( 1 )
n+1 3 3. {2n }
1.
8n n 
1

4. −
  n  3
1
2. 5 5. {(−3)n }
2

A special class of sequences are those that are monotonic.

Definition: A sequence is monotonic if it is either

increasing : an < an+1 for all n, or
decreasing : an > an+1 for all n, or
nondecreasing : an ≤ an+1 for all n, or
nonincreasing : an ≥ an+1 for all n.

Note that increasing sequences are, by definition, nondecreasing as are decreasing sequences nonincreas-
ing. An example of a nondecreasing sequence that is not an increasing sequence is

{1, 1, 2, 2, 3, 3, 4, 4, . . .}
108 4.1 Sequences

Example 4-8
Classify the monotonicity of the following sequences.
√
n +3
 2 
2. n+2− n
√
1.
n+5

Solution:

1. We illustrate two approaches:

• Method 1: Using the Definition
an < an+1 (n ≥ 1)
n +3 2
(n + 1)2 + 3
⇐⇒ < (n ≥ 1)
n+5 (n + 1) + 5
n2 + 3 n2 + 2n + 4
⇐⇒ < (n ≥ 1)
n+5 n+6
⇐⇒ (n2 + 3)(n + 6) < (n2 + 2n + 4)(n + 5) (n ≥ 1)
⇐⇒ n3 + 6n2 + 3n + 18 < n3 + 5n2 + 2n2 + 10n + 4n + 20 (n ≥ 1)
⇐⇒ 0 < n2 + 11n + 2 (n ≥ 1)

n2 + 3
 
Since the final statement is true we have proven that an < an+1 and thus is
n+5
an increasing sequence.
• Method 2: Using the Derivative
x2 + 3
Consider f (x) = . Then
x+5
2x(x + 5) − (x2 + 3)(1) 2x2 + 10x − x2 − 3 x2 + 10x − 3
f ′ (x) = = = > 0 for x ≥ 1 .
(x + 5)2 (x + 5)2 (x + 5)2
n +3
 2 
Therefore, f (x) is an increasing function for x ≥ 1. Thus is an increasing
n+5
sequence.
Since the sequence is increasing it is also nondecreasing.
√ √
2. Considering f (x) = x + 2 − x we have
1 1 1 1
f ′ (x) =
(x + 2)− 2 (1) − x− 2
2 2
1 1
= √ − √ < 0 for x ≥ 1 .
2 x+2 2 x
√
Therefore f (x) is a decreasing function for x ≥ 1. Thus n+2− n is a decreasing
√

sequence, and also, therefore, a nonincreasing sequence.

Further Questions:

Classify the monotonicity of the following sequences.

2 2n + 3 3n + 2
     
1. 2. 3.
n+3 3n + 5 2n + 1
Sequences and Series 109

Definition: Sequence {an } is bounded below if there exists some N such that N ≤ an for all n.
The sequence {an } is bounded above if there exists some M such that an ≤ M for all n. The
sequence {an } is bounded if it is bounded both below and above.

Example 4-9
Determine whether the following sequences are bounded above, bounded below, or bounded.

2n + 1
 2 
1. {en } 2. {−n4 } 3.
n2 + 3

Solution:

1. Since en > 0 for all n ≥ 1 the sequence is bounded below. lim en = ∞ shows that it is not
n→∞
bounded above.
2. Since −n4 < 0 for all n ≥ 1 the sequence is bounded above. lim −n4 = −∞ shows that it is
n→∞
not bounded below.
2n2 + 1
3. Clearly > 0 for all n ≥ 1 so the sequence is bounded below. Considering the limit of
n2 + 3
the sequence
2n2
2n2 + 1 n2 + n2 2+ 2+0
1 1
n2
lim 2 = lim 2 · = lim 1 · n2
= =2
n→∞ n + 3 1+ 1+0
n2 + n2
n→∞ n n 2 3 n→∞ 1
n2

suggests it may have an upper bound around 2. The inequality

2n2 + 1
< 2 ⇐⇒ 2n2 + 1 < 2(n2 + 3) ⇐⇒ 0 < 5
n2 + 3
2n2 + 1
confirms 2 is an upper bound for the sequence. Thus 0 < < 2 and the sequence is
n2 + 3
bounded.

Further Questions:

Determine whether the following sequences are bounded above, bounded below, or bounded.
 
1. n2 n

2.
n+1

The following theorem can be used to determine whether a monotonic sequence converges or not.
Theorem 4-6: A monotonic sequence converges if and only if it is bounded.
 110 4.1 Sequences

Answers:
Exercise 4-1
Page 179

1-3: Find the first four terms of the infinite sequence and evaluate lim an if it exists.
n→∞

n + 3n
3 2 2e
n
n2 + 3
1. an = 2. an = 3. an = (−1)n
2n3 + 5 3en + 1 n3 + 2n + 4

4-7: Evaluate lim an if it exists and determine whether the infinite sequence is convergent or
n→∞
divergent.

en ln n
4. an = 6. an =
2en + n 3n
tan−1 n n+1
5. an = 7. an = (−1)n
n n2 + 5

8-10: Determine whether the infinite sequence is increasing, decreasing or not monotonic.

8. an = ne−n
2
n2 + 3 en
9. an = 10. an =
n2 + 5 en + 2
Sequences and Series 111

4.2 Series

Definition: Given a sequence {ak }, the summation of its terms,

a1 + a2 + a3 + · · · + ak + · · ·
∞
X
is called an (infinite) series. A series is abbreviated in sigma notation as ak or sometimes
k=1
without explicit index limits as ak .
P

Due to the sum being over an infinite number of terms it need not exist.
Example 4-10
The series
∞
X
1 = 1 + 1 + ··· + 1 + ···
k=1

clearly cannot approach a number when added.

To rigorously define what we mean by the value of the sum of a series we introduce the following.
∞
X
Definition: Given the series ak define the sum of the first n terms of the series to be the nth partial
k=1
sum Sn :
n
X
Sn = ak = a1 + a2 + · · · + an
k=1

Example 4-11
For the series 1 + 1 + 1 + · · · + 1 + · · · the nth partial sums are

S1 = 1
S2 = 1+1=2
S3 = 1+1+1=3
..
.
Sn = 1 + 1 + 1 + ··· + 1 = n
| {z }
n times

The partial sums Sn for a series ak themselves form a sequence {Sn } the limit of which we will
P
consider the sum of the series.

Definition: Let series ak have nth partial sums Sn . If the sequence {Sn } is convergent, so
P

lim Sn = S ,
n→∞

then we say that the series ak is convergent and call S the sum of the series,
P

∞
X
ak = a1 + a2 + a3 + · · · + ak + · · · = S .
k=1

If a series is not convergent then it is divergent .

112 4.2 Series

Example 4-12
∞
1
X
The series can be written using partial fraction decomposition as
(k + 1)(k + 2)
k=1
∞ 
1 1
X 
− . The nth partial sum is therefore
k+1 k+2
k=1

1 1 1 1 1 1 1 1
     
− Sn =
+ − + ··· + − = −
2 3 3 4 n+1 n+2 2 n+2

1 1 1
 
Since lim Sn = lim − = the series is convergent with sum 1/2, i.e.
n→∞ n→∞ 2 n+2 2
∞
X 1 1
= .
(k + 1)(k + 2) 2
k=1

Because of the cancellation arising in the partial sum the series is called a telescoping series. The
series collapses like a spyglass in which the tubes disappear into each other to shorten the telescope.

Example 4-13
∞
X
We saw the series 1 has partial sum Sn = n. Therefore lim Sn = lim n = ∞ and so the
n→∞ n→∞
k=1
∞
X
sequence {Sn } and hence the series 1 are divergent (as expected).
k=1

Theorem 4-7: The geometric series

∞
X
ark = a + ar + ar2 + · · · + ark + · · · = a 1 + r + r2 + · · · + rk + · · ·

k=0

a
is convergent if −1 < r < 1 with sum and is otherwise divergent.3
1−r
∞
X
One notes that the geometric series also has the form ark−1 if the lower limit is taken to be k = 1
k=1
rather than k = 0 .
Example 4-14
Determine whether the following series are convergent and, if so, find the sum.

1. 2 + 6 + 18 + 54 + · · ·
∞
X (−1)k 6k+1
2.
5k
k=0
∞  n
X 1
3.
n=1
4
3 Proof follows by noting that if Sn = 1 + ar + · · · + arn−1 then


(1 − r)Sn = a + ar + ar2 + · · · + arn−1 − ar + ar2 + ar3 + · · · + arn = a − arn ,
a(1−r n ) a
and so Sn = 1−r
. Then S = lim Sn = since lim rn = 0 for −1 < r < 1 .
n→∞ 1−r n→∞
Sequences and Series 113

Solution:

1. The series has first term a = 2 and the ratios of subsequent terms
6 18 54
= = = ···
2 6 18
all equal to 3 shows the series is geometric with common ratio r = 3. Since |r| = 3 > 1 the
geometric series is divergent. Alternatively we note that we can write the series in sigma
∞
X ∞
X
notation as 2(3)k or 2(3)k−1 to prove it is geometric and to identify a = 2 and r = 3.
k=0 k=1

∞
X
2. The series starts at k = 0 and so the standard form for that geometric series is ark .
k=0
Rewriting the series
∞ ∞ ∞ k
(−1)k 6k+1 (−1)k 6k 6
X X X 
= 6 = 6 −
5k 5k 5
k=0 k=0 k=0

proves the series is geometric with first term a = 6 and common ratio r = − 65 . Since |r| = 6
5 >1
the geometric series is divergent.
∞
X
3. The series starts at n = 1 so put it in the standard form arn−1 :
n=1

∞  n ∞    n−1
X 1 X 1 1 1 1
= =⇒ a = ,r=
n=1
4 n=1
4 4 4 4

Then |r| < 1 implies the series converges to sum

a 1 1
1 4 1
S= = 4 = 4
= · =
1−r 1− 1
4
3
4
4 3 3

Further Questions:

Determine whether the following series are convergent and, if so, find the sum.

 n−1
2 2 1
1. 2 + + + · · · + 2 + ···
2 4 2
∞
X
2. (3)n−1
n=1

3. 1 + x + x2 + x3 + · · · (for |x| < 1)

114 4.2 Series

A necessary (but not sufficient) requirement for a series to converge is that its terms must approach
zero as n → ∞.
X∞
Theorem 4-8: If series ak is convergent then lim ak = 0 .
k→∞
k=1
∞
X
Note the converse of the last theorem, namely that if lim ak = 0 then ak is convergent, is not
k→∞
k=1
true, as demonstrated in the following example.
Example 4-15
The harmonic series,
∞
X 1 1 1 1
= 1 + + + ··· + + ···
k 2 3 k
k=1

is divergent. To see this note that the harmonic series

1 1 1 1 1 1 1 1
1 + + + + + + + + + ···
2 3 4 5 6 7 8 9
is strictly greater than
1 1 1 1 1 1 1 1 1
+ + + + + + + + + ··· ,
1 2
|{z} |{z} 4 4
| {z } 8
| 8 {z 8 8
} 16
> 12 = 21 = 12 = 12

which diverges to infinity as one can exceed any multiple of 1/2 one wants by taking enough
terms. Specifically, the partial sums of the harmonic series form an increasing sequence in which
S2n > 12 (n + 1) and therefore the sequence {Sn } is unbounded (and hence divergent).

The contrapositive of Theorem 4-8 (which logically must be true) provides a useful method to test if
some series are divergent:
Theorem 4-9: The Term Test for Divergence:  
If the terms ak of series ak approach a non-zero limit lim ak = L ̸= 0 or lim ak does not exist,
P
k→∞ k→∞
then ak is divergent.
P

We note that by Theorem 4-2 it follows that if lim |ak | =

̸ 0 or does not exist then lim ak must either
k→∞ k→∞
be non-zero or not exist and vice versa. As such, when applying the Term Test for Divergence, it is
sufficient to test lim |ak | which may be easier to evaluate.
k→∞
Example 4-16
Determine whether the series
∞
1
X  
cos
k
k=1
converges or diverges.
Solution:

Let ak = cos 1
be the k th term in the series. Consider cos x1 , a continuous function of x. Then



k

1 1
   
lim cos = lim = cos(0) = 1
x→∞ x x→∞ x
Sequences and Series 115

1
 
implies, by Theorem 4-1, lim ak = lim cos = 1 ̸= 0. By the Term Test for Divergence the
k→∞ k→∞ k
series ak is divergent.
P

Further Question:

Determine whether the series

1 2 3 k
+ + + ··· + + ···
3 5 7 2k + 1
converges or diverges.

Note that if we find lim ak = 0 we know nothing about the convergence or divergence of series ak .
P
n→∞
∞
X ∞
X
Theorem 4-10: If c is any constant and ak , bk are convergent series then the following series
k=1 k=1
are convergent with the given results:
∞
X ∞
X
1. cak = c ak
k=1 k=1
∞
X ∞
X ∞
X
2. (ak ± bk ) = ak ± bk
k=1 k=1 k=1

∞
X ∞
X ∞
X ∞
X
If ak is divergent and c =
̸ 0 then cak is divergent. If one of ak , bk is convergent and one
k=1 k=1 k=1 k=1
∞
X
is divergent then (ak ± bk ) is divergent.
k=1
Example 4-17
Prove that the following series converges and find its sum.
∞ 
1 3
X 
+
i=1
5i 10i

Solution:
∞  ∞ ∞
1 3 1 3
X  X X
+ i = +
i=1
5i 10 i=1
5i
i=1
10 i

∞  i−1 ∞  i−1
X 1 1 X 3 1
= · + ·
i=1
5 5 i=1
10 10
| {z } | {z }
geometric series geometric series
a = 15 , r = 51 a = 10
3
, r = 10
1

|r| < 1 (convergent) |r| < 1 (convergent)

1 3 1 3
1 5 3 10 1 1 3+4 7
= 5
+ 10
= 5
+ 10
= · + · = + = =
1− 1
5 1− 1
10
4
5
9
10
5 4 10 9 4 3 12 12

7
Thus the given series is convergent with sum .
12
116 4.2 Series

Further Question:

Prove that the following series converges and find its sum.
∞ 
2 7
X 
+
3k−1 k(k + 1)
k=1

Example 4-18
Determine whether each series is convergent or divergent. For convergent series, find its sum.
∞ ∞ ∞  n
2n2 + 5 4 3

X X X n
1. 2. 3. +
n=1
6n + 3n + 1
2
n=1
(n + 1)(n + 3) n=2
4 ln n

Solution:

1. Let an be the nth term in the series and consider lim an :

n→∞
2
2n2 + 5 n2 + n2 2 + n52
2n 5
n2
lim = · lim = lim 1 ·
n→∞ 6n2 + 3n + 1 6 + n3 + n12
n2 + n2 + n2
2
n2 n→∞ 6n 3n 1 n→∞

2+0 2 1
= = = ̸= 0
6+0+0 6 3

Therefore the given series an is divergent by the Term Test for Divergence.
P

2. Using partial fraction decomposition, the series can be written as

∞ ∞ 
4 2 2
X X 
= − .
n=1
(n + 1)(n + 3) n=1 n + 1 n + 3

The nth partial sum may be evaluated due to the telescoping terms:
!
2 2 2 2 2 2 2 2 2 2
       
Sn = − + − + − + − + ... + −
2 4 3 5 4 6 5 7 n+1 n+3
2 2 2 2 2 5 2
= + − =1+ − = −
2 3 n+3 3 n+3 3 n+3
Taking the limit gives
5 2 5 5
 
S = lim Sn = lim − = −0= ,
n→∞ n→∞ 3 n+3 3 3
5
so the given series is convergent with sum .
3
Note that the series and are both separately divergent since they are, up to a
P 2 P 2
n+1 n+3
scalar multiplier and a few missing initial terms, both harmonic series. So separating a series
into two divergent components does not imply the original series is divergent.
3. Consider the series as a sum of two series:
∞  n ∞  n ∞
3 3
 X
X n X n
+ = + .
n=2
4 ln n n=2
4 n=2
ln n
Sequences and Series 117

The first series is geometric with r = 4.

3
Since |r| < 1 that series converges. However the
second series diverges since
x 1
lim = lim = lim x = ∞ ,
x→∞ ln x LH x→∞ 1/x x→∞

n
which implies lim = ∞ and the second series diverges by the Term Test for Divergence.
n→∞ ln n
Since the original series is the sum of a convergent and divergent series, it is divergent.

Further Questions:

Determine whether each series is convergent or divergent. For convergent series, find the sum.
∞ ∞
"   k #
X 3k 3
k
2
1.
X
5. +
5k − 1 2 3
k=1 k=1
∞
X
2. k! ∞
X 6k
k=1 6.
∞  7k+1
1 1 k=0
X 
3. −
3k 4k
k=1
∞
2n
 
1 2
X
4. + + · · · +
n
+ ··· 7. ln
2 3 n+1 n=3
3n − 7

Here the factorial k! = k · (k − 1) · · · (1) for k ≥ 1 (and 0! is defined to be 1).

Exercise 4-2 Answers:

Page 179
1-11: Determine whether the infinite series converges or diverges. If it converges, find its sum.
3 3 3 ∞ 
1 5

1. 3 + + 2 + · · · + n−1 + · · ·
X
2 2 2 7. +
n=1
2n n(n + 1)
 2  n−1
3 3 3
2. 1 + + +···+ +··· ∞ 
e e e 1 2
X 
8. −
X∞
n=1
n n(n + 1)
3. 2−n 3n−1
n=1 ∞
3n + 4
X  2 
∞ 9. ln
X 4n 2n2 + 1
4. n=1
n=1
5n−1
∞
∞ X
X 10. tan−1 n
5. (−3)1−n
n=1
n=1
∞ ∞ 
en 1 1
X X    
6. 11. sin − sin
n k k+1
n=1 k=1
118 4.3 Testing Series with Positive Terms

4.3 Testing Series with Positive Terms

It is often difficult to find an exact sum of a series. In most cases a simple formula for the partial
sum Sn cannot be found. It is therefore of interest toP develop techniques to test whether a series is
convergent or divergent. We start by considering series ak with positive (ak > 0) terms. Because the
terms are positive the sequence of partial sums, {Sn }, is increasing.

4.3.1 The Integral Test

P∞
Suppose a series k=1 ak has terms ak = f (k) written
Rn in terms of a function f (x) that is continuous,
positive, and decreasing for x ≥ 1. The integral 1 f (x) dx will be smaller than the partial sum Sn−1 ,

n−1
X
Sn−1 = ak = a1 + a2 + · · · + an−1 = (a1 )(1) + (a2 )(1) + · · · + (an−1 )(1) ,
k=1

since the latter can be considered the total area of rectangles of height ak and width 1 for k = 1 to
k = n − 1 as shown in the following diagram:

y y = f (x)
a1

a2

a3
..
.
an−1

1 2 3 ··· n −1 x

Here we areR ∞considering the ak to be the height on the left R t side of the rectangles. Consider the
case that 1 f (x) dx is divergent. Since f (x) is positive, 1 f (x) dx is an increasing function of t
R∞
and 1 f (x) dx = +∞. Suppose that increasing sequence {Sn } were bounded with upper bound M .
Rn Rn
Then the relationship 1 f (x) dx < Sn−1 implies that 1 f (x) dx < M for any integer n and therefore
Rt R∞
1
f (x) dx < M for any real t ≥ 1, a contradiction to the divergence of 1 f (x) dx. Hence {Sn } must
R∞
be
P an unbounded monotonic sequence and therefore is divergent. Thus if 1 f (x) dx is divergent then
ak is divergent.
Alternatively if we consider rectangles with height being ak on the right (so k = 2 to k = n) we have
the following diagram:
Sequences and Series 119

y y = f (x)

a2

a3
..
.
an

1 2 3 ··· n x
Rn
In this case it follows that the integral 1
f (x) dx must be greater than:

(a2 )(1) + (a3 )(1) + · · · + (an )(1) = a2 + a3 + · · · + an = Sn − a1

R∞
Consider the case where 1 f (x) dx is convergent. Let M be the value of the integral. Since f (x) is
Rn Rn
positive, 1 f (x) dx < M . From Sn − a1 < 1 f (x) dx it follows that for any n, Sn < M + a1 and thus
R∞
monotonic sequence {Sn } is bounded and therefore convergent. Thus if 1 f (x) is convergent, then
ak is convergent.
P

We summarize our result in the following theorem.

Theorem 4-11: The Integral Test:
∞
X
Let f (x) be a continuous positive decreasing function for x ≥ 1 and ak be a series with ak = f (k).
k=1

∞
R∞ X
1. If 1
f (x) dx is convergent then ak is also convergent.
k=1
∞
R∞ X
2. If 1
f (x) dx is divergent then ak is also divergent.
k=1

Example 4-19
Determine whether the series converges or diverges.
∞ ∞
X 1 X (ln n)2
1. 2.
n=1
n + 16
2
n=1
n

Solution:

1
1. Setting f (x) = we have that f (x) is continuous and positive for x ≥ 1. Also
x2 + 16
−2x
f ′ (x) = < 0 for x > 0
(x2 + 16)2
120 4.3 Testing Series with Positive Terms

shows f (x) is decreasing for x ≥ 1 .

Z ∞
1
Z t
1

1  t
−1 x
dx = lim dx = lim tan
1 x2 + 16 t→∞ 1 x2 + 16 t→∞ 4 4 1
1 1 1 π 1
       
t
= lim tan −1
− tan −1
= − tan −1
4 t→∞ 4 4 4 2 4
Z ∞ ∞
1 X 1
Integral 2 + 16
dx is convergent and therefore the series 2 + 16
is convergent by
1 x n=1
n
the Integral Test.
(ln x)2
2. Let f (x) = . Then f (x) is continuous and positive for x ≥ 1. Consider the derivative:
x
2(ln x) x1 · x − (ln x)2 (1) 2(ln x) − (ln x)2 (ln x)(2 − ln x)
f ′ (x) = 2
= =
x x2 x2
For x ≥ 2 both ln x and x2 are positive. Solve the inequality for the remaining factor to see
where it decreases:
2 − ln x < 0 =⇒ 2 < ln x =⇒ e2 < eln x =⇒ e2 < x =⇒ x > e2 ≈ 7.3891
(Note here we used that ex is an increasing function so that the inequality was preserved under
exponentiation.) Thus f (x) is decreasing for x > e2 . Evaluating the improper integral from
x = 8 onward gives:
Z ∞
(ln x)2 (ln x)2 1
Z t
dx = lim dx (← u = ln x so du = dx)
8 x t→∞ 8 x x
ln t
1 3 1 1
Z ln t   
= lim u du = lim
2
u = lim (ln t) − (ln 8) = ∞
3 3
t→∞ ln 8 t→∞ 3 t→∞ 3 3
ln 8
P∞ 2
Therefore the improper integral is divergent and the series n=8 (lnnn) must also diverge by
P∞ 2
the Integral Test. The original series n=1 (lnnn) differs from the latter by a finite sum of
seven terms and so it also diverges.
In practice, when applying the Integral Test and, indeed, other series convergence tests in this
chapter, we need only apply the test to the infinite tail of the sequence starting at any finite
index (like 8 here). The requirements of the theorem have to only be met from that index
forward. The convergence (or not) of a series can only depend on the tail of the series as the
initial sum of a finite number of terms is necessarily finite and cannot, therefore, impact the
convergence properties of the series as a whole.

Further Questions:

Determine whether each of the following series is convergent or divergent.

∞ 1 1 1
2
4. 1 + + 2 + ··· + 2 + ···
X
1. ne−n 22 3 n
n=1
∞
∞
1
X 5
5. √
X
2. (The harmonic series) n
k n=1
k=1
∞ ∞
X 1 X ln n
3. (The hyperharmonic or p-series) 6.
kp n2
k=1 n=1
Sequences and Series 121

We summarize our results for the p-series from the Further Questions of the last example in the following
theorem.
∞
X 1
Theorem 4-12: The p-series is convergent for p > 1 and divergent otherwise.
kp
k=1

Notes on the Integral Test:

1. The Integral Test can be relaxed to consider a function R ∞f (x) that is continuous positive and
decreasing only on xP≥ n with the corresponding integral n f (x) dx. This determines convergence
∞ P∞
or not of the series k=n ak but this in turn determines convergence of the entire series k=1 ak
since these two series differ only by a finite number of terms having a finite sum.
2. It follows from the Integral Test that the improper integral and the series either are both convergent
or both divergent. This means that determining the convergence or not of a series can be used to
determine the convergence properties of the improper integral of a continuous positive decreasing
function f (x) if that were desired.

Estimating the Series Sum

Even if a series is convergent it may be impossible to sum due to the impossibility of finding a closed
form for the partial sum Sn for which we can take the limit. In that case one may resort to numerically
calculating the partial sum Sn itself, for “large” n as an approximation for the sum of the series, S ≈ Sn .
The error in the approximation is the remainder Rn = S − Sn which is the sum of the terms that
were not included:
∞
X
S= ak = a1 + a2 + · · · + an + an+1 + an+2 + · · · = Sn + Rn
| {z } | {z }
k=1 n ∞
X
Sn =
X
ak Rn = ak
k=1 k=n+1

For a convergent series ak with ak = f (k) where f (x) is a continuous positive decreasing function
P
one can place bounds on the size of the remainder, thereby estimating the error in the numerical
approximation. In our previous discussion we found that the nth partial sum satisfied
Z n
Sn − a1 < f (x) dx < Sn−1
1

In the case of convergence these inequalities imply

Z ∞
S − a1 ≤ f (x) dx ≤ S .
1

If we start summing at the n term instead of the first this generalizes to

th

Z ∞
(an + an+1 + an+2 + · · · ) − an ≤ f (x) dx ≤ an + an+1 + an+2 + · · · ,
n

from which it follows that Z ∞

Rn ≤ f (x) dx ≤ Rn−1 .
n
R∞
Thus n f (x) dx is an upper bound for the error Rn . The substitution n − 1 → n for the inequality on
R∞
the right implies n+1 f (x) dx ≤ Rn , thereby providing a lower bound on the remainder (error) as well.
We summarize the result in the following theorem:
122 4.3 Testing Series with Positive Terms

∞
X
Theorem 4-13: For convergent series ak with ak = f (x) where f (x) is a continuous positive
k=1
∞
X
decreasing function, the remainder Rn = ak = S − Sn satisfies:
k=n+1
Z ∞ Z ∞
f (x) dx ≤ Rn ≤ f (x) dx ,
n+1 n

and therefore Z ∞ Z ∞
Sn + f (x) dx ≤ S ≤ Sn + f (x) dx .
n+1 n
Example 4-20

∞
X 1
1. Estimate the sum of the series using the third partial sum S3 .
n=1
n2 + 16

2. Find bounds on the error (the remainder R3 ) that you are making by using that approximation.

Solution:

1. From Example 4-19 Problem 1 we know that the series is convergent. The sum of the series S
is approximately:
3
X 1 1 1 1 1 1 1 253
S3 = 2 + 16
= + + = + + = ≈ 0.1488
n=1
n (1) 2 + 16 (2)2 + 16 (3) 2 + 16 17 20 25 1700

2. S = S3 + R3 where R3 satisfies
Z ∞ ∞
1 1
Z
≤ R3 ≤ .
3+1 x + 16
2
3 x2 + 16

Evaluate the integrals as we did in Example 4-19 .

Z ∞
1 1 π 4 1 hπ πi
  
π
dx = . . . = − tan −1
= − = ≈ 0.1963
4 x + 16
2 4 2 4 4 2 4 16
Z ∞
1 1 π 3
  
dx = . . . = − tan−1 ≈ 0.2318
3 x + 16
2 4 2 4
Thus 0.1963 < R3 < 0.2318. Our estimate of 0.1488 is therefore too low by somewhere between 0.1963
and 0.2318. In other words, the actual sum S of the series must lie between 0.1488 + 0.1963 = 0.3451
and 0.1488 + 0.2318 = 0.3806 .
Further Questions:

Leonhard Euler was able to show the sum of the p-series with p = 2 is
∞
X 1 1 1 1 π2
= 1 + + + + · · · = = 1.644934 . . .
k2 4 9 25 6
k=1

1. Find S4 .
2. Find the remainder R4 .
3. Show R4 falls within the bounds of the last theorem.
4. What partial sum Sn is required to be in error less than 0.01?
Sequences and Series 123

4.3.2 The Basic Comparison Test

We have seen several examples of series with their associated convergence properties:

∞
X
geometric: ark−1 is convergent for |r| < 1, divergent for |r| ≥ 1
k=1
∞
X 1
telescoping: (for example) is convergent
k(k + 1)
k=1
∞
X 1
harmonic: is divergent
k
k=1
∞
X 1
p-series: is convergent for p > 1, divergent for p ≤ 1
kp
k=1

We now develop some series convergence tests that use the known convergence properties of one series
to determine that of another. The first test is a discrete analogue to our improper integral test found
in Theorem 3-2.
Theorem 4-14:
P The Basic Comparison Test:
Let ak and bk be series with positive terms satisfying ak ≤ bk for all k.
P

1. If bk is convergent then ak is convergent.

P P

2. If ak is divergent then bk is divergent.

P P

Pn Pn
Proof: Let Sn = k=1 ak and Tn = k=1 bk denote the partial sums of the series ak and
P P
bk
respectively. Since ak ≤ bk it follows that Sn ≤ Tn for any n. Furthermore sequences {Sn } and {Tn }
are both increasing (and hence monotonic) since terms ak and bk are positive.
P∞
Part 1 of the theorem follows from noting that monotonic sequence {Tn } has upper bound T = k=1 bk
since the series bk , and hence the sequence {Tk } converges. The sequence {Sn } must also then have
P
this upper boundP since Sn ≤ Tn ≤ T . Thus {Sn } is a monotonic bounded sequence and hence converges
to S. Therefore ak is convergent.
Part 2 follows from noting that if ak is divergent then monotonic sequence {Sn } is unbounded, which
P
implies it has no upper bound as it is bounded below by 0. Now Sn ≤ Tn implies P that monotonic
sequence {Tn } has no upper bound and hence does not converge, thereby proving bk is divergent.
Note that the Basic Comparison Test is also known as the Direct Comparison Test.

Example 4-21
Determine whether the series converges or diverges.
∞ ∞
X 4n X 4n (n + 2)2
1. 2.
n=1
5 +n
n
n=1
n2 − 3

Solution:
∞
X 4n
1. The given series is an where an = and is therefore a positive series.
n=1
5n + n
124 4.3 Testing Series with Positive Terms

∞  n
X 4n 4
Consider bn where bn = = . Then
n=1
5n 5

∞ ∞  n ∞    n−1
X X 4 X 4 4
bn = =
n=1 n=1
5 n=1
5 5

4
shows bn is a geometric series with r = . Since |r| < 1 it is a convergent series. We now
P
5
show that an ≤ bn for n ≥ 1 :
4n 4n
an ≤ bn ⇐⇒ ≤ n
5n
+n 5
⇐⇒ (4n )(5n ) ≤ (4n )(5n + n)
⇐⇒ 0 ≤ n(4n )
(which is true for n ≥ 1)

Therefore, by the Basic Comparison Test, the given positive series an is convergent because
P
it lies below a convergent series.
∞
X 4n (n + 2)2
2. The given series is an where an = and is therefore a positive series for n ≥ 2.
n=1
n2 − 3
Noting that for large n both the polynomials in the numerator and denominator will be
4 n · n2
dominated by their quadratic (n2 ) term, consider bn = = 4n . Then
n2
∞
X ∞
X ∞
X
bn = 4 =
n
4(4n−1 )
n=1 n=1 n=1

shows bn is a geometric series with a = 4 and r = 4 > 1 and therefore is divergent. Now we
P
must show that an ≥ bn for n ≥ 2 :

4n (n + 2)2 (n + 2)2
an ≥ bn ⇐⇒ ≥ 4n ⇐⇒ ≥ 1 ⇐⇒ (n + 2)2 ≥ n2 − 3
n −3
2 n2 − 3
⇐⇒ n2 + 4n + 4 ≥ n2 − 3 ⇐⇒ 4n + 7 ≥ 0
7
⇐⇒ n ≥ − (which is true for n ≥ 2)
4
Therefore, by the Basic Comparison Test, the given series an is divergent because it lies
P
above a divergent series that is ultimately positive (for n ≥ 2). Once again one notes that
the conditions of the test need only apply for the infinite tail of the given series and one can
ignore a finite number of initial terms (here the first term a1 = −18) as these do not affect the
convergence of the series.

Further Questions:

Determine whether the series converges or diverges.

∞ ∞ ∞
X 1 X 3 X 1
1. 2. √ 3.
2 + 5k n=2
n−1 n=1
n3 n
k=1
Sequences and Series 125

Remainder Estimate

Note that if one uses convergent series bk to showP ak is convergent by the Basic Comparison Test,
P P
∞
it follows, since ak ≤ bk , that the remainder Rn = k=n+1 ak is less than or equal to the remainder
P∞
R̃n = k=n+1 bk . Hence P if we have an estimate for the size of the error R̃n ≤ ϵ for series bk , this
P
implies Rn ≤ ϵ for series ak .

4.3.3 The Limit Comparison Test

Our next test involves taking the limit of the ratio of terms of two series, one of whose convergence
properties are presumed known.
Theorem 4-15: The Limit Comparison Test:
Let ak and bk be series with positive terms.
P P

ak
1. If lim = L > 0 then both series are convergent or both divergent.
k→∞ bk

ak
2. If lim = 0 and bk is convergent then ak is convergent.
P P
k→∞ bk

ak
3. If lim = ∞ and bk is divergent then ak is divergent.
P P
k→∞ bk

The convergence conclusions of the Limit Comparison Test can be remembered by noting that the limit
condition effectively suggests that the tail of the series satisfies ak = Lbk , in other words the series
tail is effectively a multiple of that of the other series by a constant c = L. For c = L =
̸ 0 we saw in
Theorem 4-10 that the new series has the same convergence properties as the original.
ak
Proof: Consider the case where lim = L > 0. Then there exists M̃ > 0 and Ñ > 0 such that
k→∞ bk

ak
M̃ < < Ñ for k > n
bk
n o
Let M be the minimum of the finite set of numbers abkk |k ≤ n and the number M̃ . Similarly let N
n o
be the maximum of the finite set of numbers abkk |k ≤ n and the number Ñ . It follows that for all k:

ak
M< <N
bk
Since bk > 0 we have, for all k,
M bk < ak < N bk .
If bk converges then N bk convergesP
and ak < N bk implies ak converges by P
the Basic Comparison
P P P
Test. Similarly if bk diverges then M bk diverges and M bk < ak implies ak diverges by the
P
Basic Comparison Test.
ak
In the case lim = 0 we can only argue that ak /bk < N and so only bk convergent implies
P P
ak
k→∞ bk
ak
convergent. In the case lim = ∞ we can only argue that M < ak /bk and so only bk divergent
P
k→∞ bk
implies ak divergent.
P
126 4.3 Testing Series with Positive Terms

Example 4-22
Determine whether the following series are convergent or divergent.

∞ ∞
X 4n X n2 + 2 ln n
1. 2.
n=1
5n − 6n n=1
n4 + 3n2 + 5

Solution:

∞
X 4n
1. The given series is an where an = . This is a positive series for n ≥ 2 since the
n=1
− 6n 5n
denominator becomes positive at n = 2 (equal to 13) and only increases (so remains positive)
after that since dx
d
(5x − 6x) = 5x ln 5 − 6 > 0 when x ≥ 2. Since convergence of a series
depends only on its infinite tail we can identify a suitable comparison series by looking at the
large n properties of an . WePnote that the denominator
is dominated by the exponential 5n
∞ n n
which suggests considering n=1 bn where bn = 54n = 45 . Then
∞ ∞  n ∞  n−1
X X 4 X 4 4
bn = =
n=1 n=1
5 n=1
5 5

is a geometric series with a = 4

5 and |r| = 4
5 = 4
5 < 1 and therefore convergent.

4n
an 5n −6n 4n 5n 5n
lim = lim = lim · n = lim n
n→∞ bn n→∞ 4n n→∞ 5n − 6n 4 n→∞ 5 − 6n
5n
5n
5 n
1 1
= lim · 5n
= lim = lim =1
n→∞ 5n 5n
− 6n n→∞ 1 − n6n n→∞ 1 − 0
5n 5n 5

Here we used L’Hôpital’s Rule in evaluating the last step of the limit:

6x 6 6
lim = lim = =0
x→∞ 5x LH x→∞ 5x ln 5 ∞

The latter limit shows that 5n dominates 6n as claimed above and indicates why we factored out
an
that term (and not 6n) when evaluating our original limit. Therefore, since lim = 1 > 0,
n→∞ bn
by the Limit Comparison Test, the given series an is also convergent.
P

As an aside, note that this series is very similar to Problem 1 in Example 4-21. However
attempting
P to use the Basic Comparison Test in the current example using the same comparison
series bn will fail to be conclusive since here, as the reader may confirm, an > bn for n ≥ 2.
In general the Limit Comparison Test will often work to establish convergence of a series where
the Basic Comparison Test will not.
∞
X n2 + 2 ln n
2. The given series is an where an = and therefore positive. Consideration of
n=1
n4 + 3n2 + 5
the limiting behaviour of the terms at large n suggests a comparison with the series
P
bn
2
where bn = n4 . Then
n

∞ ∞ ∞
X X n2 X 1
bn = 4
=
n=1 n=1
n n=1
n2
Sequences and Series 127

is a p-series with p = 2 > 1 and therefore convergent. Taking the limit of the ratio of terms
gives:
n2 +2 ln n
an n4 +3n2 +5 (n2 + 2 ln n)n2 n4 + 2n2 ln n
lim = lim = lim = lim 4
n→∞ bn n→∞ 1 n→∞ n + 3n + 5
4 2 n→∞ n + 3n2 + 5
n2
n4 2n2 ln n
n4 4 + 1 + 2 nln2 n 1 + limn→∞ 2 nln2 n
= lim 4 · nn4 n4
= lim =
n→∞ n
n4 + n4 + n4
3n2 5 n→∞ 1 + 2 + 4
3
n n
5 1+0+0
2 ln n
= 1 + lim =1
n→∞ n2
Here we evaluated the limit at the last step using L’Hôpital’s Rule on the corresponding
2 ln x
continuous function f (x) = :
x2

2 ln x 2 · x1 1
lim f (x) = lim = lim = lim 2 = 0
x→∞ x→∞ x 2 x→∞ 2x x→∞ x

This limit confirms that the power function dominates the logarithm at large n as claimed
above and why it is the power, and not the logarithm, which we factor out of the above limit.
an
Therefore lim = 1 > 0 and the given series an is also convergent by the Limit
P
n→∞ bn
Comparison Test.
As a final note, the Limit Comparison Test is seen to be superior to the Basic Comparison
Test here again as the presence of the logarithm and fourth order polynomial in an would
make evaluation of the required inequality in the latter test difficult.

Further Questions:

Determine whether the following series are convergent or divergent.

∞ ∞
X 1 X 1 + 2k
1. √ 3.
3
k2 + 1 1 + 3k
k=1 k=1

∞ ∞
X 3n2 + 5n X 1
2. 4.
n=1
2n (n2 + 1) n=1
n 2 ln n

Note that since convergence is entirely determined by the infinite tail of a series, we can relax the Basic
and Limit Comparison Tests to require that they only have positive terms for k > n for some fixed n
and, in the case of the Basic Comparison Test, that additionally ak ≤ bk for k > n.
 128 4.3 Testing Series with Positive Terms

Answers:
Exercise 4-3
Page 179

1-10: Determine whether the infinite series is convergent or divergent.

∞ ∞
X ln(n + 1) X ln n
1. 6.
n=1
n+1 n=1
n6
∞ ∞
X 5 X 2 + 4n
2. 7.
n=1
2 + 3n2 n=1
3 + 5n
∞ ∞
X 1 X 1
3. √ 8. √
n=1 n ln n
5
n=1
4
2n3 + 5
∞ ∞
X 2n2 + 5n X 5 + cos 2n
4. 9.
n=1
5n3 + 3 n=1
n3
∞ ∞
X 1 X 1
5. √ 10.
9n4 + 5n n(n + 5)
p
n=1 n=1
Sequences and Series 129

4.4 The Alternating Series Test

We now consider series where all the terms are not positive. An alternating series is a special case of
such a series.

Definition: An alternating series is a series of either of the forms

∞
X
a1 − a2 + a3 − · · · + (−1)k−1 ak + · · · = (−1)k−1 ak ,
k=1

∞
X
−a1 + a2 − a3 + · · · + (−1)k ak + · · · = (−1)k ak ,
k=1
where ak is positive for all k.

Theorem 4-16: The Alternating Series Test:

∞
X ∞
X
If an alternating series of the form (−1)k−1 ak or (−1)k ak with ak > 0 satisfies
k=1 k=1

1. ak+1 ≤ ak for all k ,

2. lim ak = 0 ,
k→∞

then the alternating series is convergent.

ak+1
Note that ak+1 ≤ ak is equivalent to ak+1 − ak ≤ 0 and ≤ 1.
ak
P∞
Proof: Suppose we have an alternating series of the form k=1 (−1)k−1 ak . Consider the even partial
sums S2 , S4 ,. . . , where, in general the (2n)th partial sum is S2n for n a positive integer given by:

S2n = a1 − a2 + a3 − a4 + a5 − a6 · · · + a2n−1 − a2n .

Then grouping the terms in pairs one has

S2n = (a1 − a2 ) + (a3 − a4 ) + (a5 − a6 ) · · · + (a2n−1 − a2n ) ,

where each term in parentheses is nonnegative since ak ≥ ak+1 . This implies {S2n } is a nondecreasing
sequence (S2 ≤ S4 ≤ S6 ≤ . . . ≤ S2n ≤ . . .). The terms of the even partial sum S2n may be regrouped
as
S2n = a1 − (a2 − a3 ) − (a4 − a5 ) − · · · − (a2n−2 − a2n−1 ) − a2n ,
where, once again, the terms in parentheses are positive. This shows S2n < a1 for all n. The monotonic
sequence of even partial sums {Sn } is bounded and hence has limit S.
The odd partial sums are S2n+1 for n a positive integer and may be written

S2n+1 = S2n + a2n+1 .

Taking the limit of the odd partial sum sequence {S2n+1 } gives

lim S2n+1 = lim (S2n + a2n+1 ) = lim S2n + lim a2n+1 = S + 0 = S

n→∞ n→∞ n→∞ n→∞

Since both even and odd partial sum sequencesPapproach the same limit S, the sequence {Sn }
∞
approaches
P∞S as well and alternating
P∞ sequence k=1 (−1)
k−1
ak is convergent. Since alternating
sequence k=1 (−1) ak = (−1) k=1 (−1)
k k−1
ak this completes the proof for the other possible case.
130 4.4 The Alternating Series Test

Example 4-23
The alternating harmonic series
∞
X (−1)k−1 1 1
=1− + − ···
k 2 3
k=1

is convergent since ak = 1
k satisfies ak+1 = 1
k+1 ≤ 1
k = ak and limk→∞ ak = limk→∞ 1
k = 0.

Example 4-24
Determine whether the following alternating series converge or diverge.
∞ ∞
X n2 + 2 X n
1. (−1)n−1 2. (−1)n
n=1
n4 + 1 n=2
ln n

Solution:
∞ ∞
X X n2 + 2
1. The given series is cn = (−1)n−1 an where cn = (−1)n−1 an and an = > 0. The
n=1 n=1
n4 + 1
x2 + 2
series is therefore an alternating series. Consider f (x) = . Then
x4 + 1
2x(x4 + 1) − (x2 + 2)(4x3 ) 2x5 + 2x − 4x5 − 8x3 −2x5 − 8x3 + 2x
f ′ (x) = = =
(x + 1)
4 2 (x + 1)
4 2 (x4 + 1)2

implies f ′ (x) < 0 for x ≥ 1 and hence f (x) decreases there. This implies f (n + 1) ≤ f (n) for
n ≥ 1 and hence an+1 ≤ an . Next consider the limit

x2 + 2 2x 1
lim f (x) = lim = lim = lim =0
x→∞ x→∞ x4 + 1 x→∞ 4x3 x→∞ 2x2
LH

where we used L’Hôpital’s Rule on the indeterminate form. Thus lim an = lim f (n) = 0 .
∞
∞ n→∞ n→∞
Since both conditions are met, the series cn is convergent by the Alternating Series Test.
P

∞ ∞
X X n
2. The given series is cn = (−1)n an where cn = (−1)n an and an = > 0. The series is
n=2 n=2
ln n
x
therefore an alternating series. Consider f (x) = . Then
ln x
x 1
lim f (x) = lim = lim 1 = lim x = ∞
x→∞ x→∞ ln x x→∞ x→∞
LH x

where we used L’Hôpital’s Rule on the ∞

∞ indeterminate form. Thus n→∞
lim an = lim f (n) = ∞.
n→∞
Since lim an = ∞ ≠ 0 we cannot draw any conclusions about the convergence or divergence of
n→∞
the given series using the Alternating Series Test. However since an = |cn | wePcan conclude by
Theorem 4-2 that lim cn ̸= 0 or does not exist. Therefore, the given series cn is divergent
n→∞
by the Term Test for Divergence.
As this examplePillustrates, in practice it makes sense, when applying the Alternating Series
Test to series cn , to first determine whether the requirement that the limit of an = |cn |
Sequences and Series 131

goes to zero holds. If it does not, then, as shown in this example, the series must diverge by
the Term Test for Divergence. If the limit does go to zero, proceed to determine whether the
other condition, that an decreases, is met or not. If it is met, then the series converges by the
Alternating Series Test. If it is not met, then that test is inconclusive.

Further Questions:

Determine whether the following alternating series converge or diverge.

∞ ∞
X 2n X ln n
1. (−1)n−1 3. (−1)n−1
n=1
4n2 − 3 n=1
n
∞ ∞
X 2k X 3k 2 + 2
2. (−1)k 4. cos (kπ)
4k − 3 2k 2 + 1
k=1 k=1

Remainder Estimate

An estimate of the error made when approximating the sum S of an alternating series with the nth
partial sum Sn is given by the following theorem.

Theorem 4-17: For an alternating series with terms of absolute value ak > 0 the remainder Rn = S−Sn
satisfies |Rn | < an+1 .

Answers:
Exercise 4-4
Page 179

1-10: Determine whether the infinite series is convergent or divergent.

∞ ∞
X n+1 X 3n + 2
1. (−1)n 6. (−1)n−1 √
n=1
n2 + 5 n=1
n+1
∞ ∞ √
1 3n + 10
X   X
2. (−1) ln 1 +
n
7. (−1) n+1

n=1
n n=1
5n + 7
∞ ∞
5n2 + 2 (−3)n
X   X
3. (−1)n−1 ln 8.
n=1
4n2 + 3 n=1
n3
∞ ∞
X e2n X 1
4. (−1)n 9. (−1)n 2
n=1
n2 n=1 (ln n)
∞ ∞
X n+5 X cos(nπ)
5. (−1)n+1 10.
n=1
4n n=1
n
132 4.5 Tests of Absolute Convergence

4.5 Tests of Absolute Convergence

For
P series ak whose terms have mixed sign, one can consider the convergence properties of the series
P
|ak | with nonnegative terms generated by taking the absolute value of the terms of the original series.

4.5.1 Absolute Convergence

Definition: A series ak is absolutely convergent if the series |ak | is convergent.

P P

A series may be convergent that is not absolutely convergent, prompting the following definition.

Definition: A series ak that is convergent but not absolutely convergent is called conditionally
P
convergent .
Example 4-25
The alternating harmonic series 1 − 12 + 13 − 41 + · · · is conditionally convergent because it converges
but the harmonic series, which is the series of the absolute values of its terms, does not.

The following theorem shows a convergent series can only be absolutely or conditionally convergent.
Theorem 4-18: If a series ak is absolutely convergent then it is convergent.
P

The theorem also

P shows that convergence of some series ak may be determined by considering
P
convergence of |ak |.
We remind the reader of some properties of absolute value that will be frequently encountered:
x |x|
|xy| = |x||y| = |x + y| ≤ |x| + |y|
y |y|
and for a > 0 :
|x| < a ⇐⇒ −a < x < a
|x| = a ⇐⇒ x = ±a
|x| > a ⇐⇒ x < −a or x > a
|x| = 0 ⇐⇒ x = 0

Example 4-26
Determine whether the following series are absolutely convergent or conditionally convergent.
∞ ∞ √
X X n
1. (−1)n e−n 2. (−1)n−1
n=1 n=1
n+1

Solution:
∞
X
1. Consider the series of absolute values of the given series (−1)n e−n , that is
n=1

∞ ∞ ∞  n ∞  n−1
X X X 1 X 1 1
(−1)n e−n = e−n = = .
n=1 n=1 n=1
e n=1
e e
Sequences and Series 133

1 1 1
This is a geometric series with a = and |r| = = < 1 and therefore convergent. Thus
e e e
the given series is absolutely convergent and therefore also convergent.
∞ √
X n
2. The given series is cn where cn = (−1) n−1
. Consider the series of absolute values
n=1
n + 1
√
n
an where an = |cn | = . Consideration of the large-n behaviour of an suggests a
P
P n+1
comparison with series bn where
√
n 1
bn = =√ .
n n
∞ ∞
X X 1
Then bn = √ is a p-series with p = 1
2 < 1 and therefore divergent. Taking the limit
n=1 n=1
n
√
n
an n+1 n n n
1 1
lim = lim = lim = lim · n
= lim = =1>0
n→∞ bn n→∞ √1 n→∞ n + 1 n→∞ n n
+ 1 n→∞ 1+ 1 1+0
n n n n

∞
X
shows the series of absolute values, an , is also divergent by the Limit Comparison Test.
n=1
∞ ∞ √
X X n
Returning to the original series we see that it is cn = (−1)n−1 an with an = and
n=1 n=1
n+1
so is an alternating series. Then
√ √
n n 1 1 1 1
lim an = lim = lim · =√ · =0· = 0.
n→∞ n→∞ n + 1 n→∞ n 1+ 1 n 1+ 1 1+0
n n

Also for n ≥ 1 one has

√ √
n+1 n
an+1 ≤ an ⇐⇒ ≤
n+2 n+1
3 √
⇐⇒ (n + 1) 2 ≤ n(n + 2) (⇐ Cross-multiplying by positive terms)
⇐⇒ (n + 1)3 ≤ n(n + 2)2 (⇐ Squaring of positive sides preserves inequality)
⇐⇒ n3 + 3n2 + 3n + 1 ≤ n3 + 4n2 + 4n
⇐⇒ −n2 − n + 1 ≤ 0 (which is true for all n ≥ 1)

Therefore, the given series cnPis convergent by the Alternating Series

P Test. Since the series
P
of its absolute values |cn | = an is divergent, the convergence of cn is conditional.
P

Further Questions:

Determine whether the following series are absolutely convergent or conditionally convergent.
∞ ∞ ∞
X 1 X sin n X 1
1. (−1)k−1 2. 3. (−1)k+1 √
k2 n=1
n2 k
k=1 k=1
134 4.5 Tests of Absolute Convergence

4.5.2 The Ratio Test

The following convergence test considers the limit of the ratio of terms within a series.
Theorem 4-19: The Ratio Test:
∞
X
Suppose the ratio of consecutive terms of series ak has limit
k=1

ak+1
lim =L,
k→∞ ak
then

1. If L < 1 the series ak is absolutely convergent (and hence convergent).

P

2. If L = 1 the test is inconclusive.

3. If L > 1 the series ak is divergent.
P

ak+1
If lim = ∞ then ak is also divergent.
P
k→∞ ak
Note that if the Ratio Test is inconclusive (L = 1) this means that ak is potentially absolutely
P
convergent, conditionally convergent, or divergent.
The convergence conclusions of the Ratio Test can be remembered by noting that the limit condition
effectively suggests that the tail of the series satisfies ak+1 = Lak , in other words it behaves like a
geometric series with r = L. From this it follows r = L < 1 should converge and r = L > 1 should
diverge.
Example 4-27
Determine whether the following series are absolutely convergent, conditionally convergent or
divergent.
∞ ∞ ∞
X 5n X 2n X 1
1. (−1)n−1 2. (−1)n 3. √
n=1
n!(n + 4) n=1
n4 n=1
n+1

Solution:
∞
X 5n
1. The series is an where an = (−1)n−1 . The presence of the factorial suggests
n=1
n!(n + 4)
trying the Ratio Test:
(−1)n ·5n+1
an+1 (n+1)!(n+5) 5n+1 n!(n + 4)
lim = lim = lim ·
n→∞ an n→∞ (−1)n−1 5n n→∞ (n + 1)!(n + 5) 5n
n!(n+4)

5n · 5 n!(n + 4)
= lim · (⇐ (n + 1)! = (n + 1)(n) · · · (1) = (n + 1)n!)
n→∞ (n + 1)n!(n + 5) 5n
5(n + 4) 5n + 20 n 5n
+ 20
= lim = lim 2 = lim 2 · n2 n 6n n 5
n→∞ (n + 1)(n + 5) n→∞ n + 6n + 5
n2 + n2 + n2
n→∞ n

1 5 + 20 5+0
= lim · n
=0· =0<1
n→∞ n 1 + 6 + 52
n n
1 + 0+0

Therefore the given series is absolutely convergent by the Ratio Test.

Sequences and Series 135

∞
X 2n
2. The series is an where an = (−1)n . Applying the Ratio Test gives:
n=1
n4

n+1
an+1 (−1)n+1 (n+1)
2
4 2n+1 n4 2n4
lim = lim = lim · n = lim
(−1) n4 n→∞ (n + 1) 2 n→∞ (n + 1)4
2 n
n→∞ an n→∞ n 4

 4  4  n 4
n n n
= 2 lim = 2 lim = 2 lim · n n 1
n→∞ n + 1 n→∞ n + 1
n + n
n→∞ n
4 4
1 1
 
= 2 lim =2 =2>1
n→∞ 1 + 1
n
1+0

Therefore the given series is divergent by the Ratio Test.

∞
X 1
3. The series is an where an = √ .
n=1
n+1
√
1 n+1 n+1 n+1
r r
an+1
lim =√ · = lim = lim
n→∞ an n+2 1 n→∞ n + 2 n→∞ n +2
s
n 1 + n1 1+0
r
= lim · = =1
n→∞ n 1 + 2
n
1+0

Therefore the Ratio Test is inconclusive and a different test is required. Since an = √1
n
·√1 1
1+ n
shows that an ≈ bn = for large n, using the Limit Comparison Test with the divergent
√1
n
(p = 1/2) p-series bn will show an is also divergent. Quite generally the Ratio Test (and
P P
also the Root Test introduced below) will always fail to be conclusive for any series that
behaves like a p-series at large n.

Further Questions:

Determine whether the following series are absolutely convergent, conditionally convergent, or
divergent.
∞ ∞
X 3k X
1. (−1)k 3. e−k k!
k!
k=1 k=1

∞ ∞
X 4n X 2n
2. 4.
n=1
n2 n=1
2n2 + 1

4.5.3 The Root Test

The next convergence test considers the limit of the k th root of |ak |.

Theorem 4-20: The Root Test:

∞
X
Suppose the terms of series ak satisfy
k=1

lim |ak | = L ,
p
k

k→∞
136 4.5 Tests of Absolute Convergence

then

1. If L < 1 the series ak is absolutely convergent (and hence convergent).

P

2. If L = 1 the test is inconclusive.

3. If L > 1 the series ak is divergent.

P

If lim |ak | = ∞ then ak is also divergent.

p
k
P
k→∞

The convergence conclusions of The Root Test can be remembered by noting that the limit condition
effectively suggests that the tail of the series behaves like ak = Lk , in other words like a geometric
series with r = L. From this it follows that r = L < 1 should converge and r = L > 1 should diverge.

Example 4-28
Determine the convergence or divergence of the following series.
∞  n ∞
X n+1 X (n + 1)n
1. 2. (−1)n+1
n=1
2n + 3 n=1
e2n

Solution:
∞ n
n+1
X 
1. The given series is an with an = . The presence of the power n in the term
n=1
2n + 3
suggests trying the Root Test:
s n
n+1 n+1
lim |an | = lim = lim
p
n n

n→∞ n→∞ 2n + 3 n→∞ 2n + 3

n + n 1+ 1+0 1
n 1 1
n
= lim · = lim n
= = <1
n→∞ n
n + n
2n 3 n→∞ 2 + 3
n
2+0 2

Therefore the given series is absolutely convergent by the Root Test.

∞
X (n + 1)n
2. The given series is an with an = (−1)n+1 . Applying the Root Test gives:
n=1
e2n
s r  n1
(n + 1)n (n + 1)n (n + 1)n

lim |an | = lim (−1)n+1 = lim = lim
p
n n n

n→∞ n→∞ e2n n→∞ e2n n→∞ e2n

n
(n + 1) n n+1
= lim 2n = lim =∞
n→∞ en n→∞ e2
Therefore the given series is divergent by the Root Test.

Further Questions:

Determine the convergence or divergence of the following series.

∞ ∞
X 23k+1 X (−1)n
1. 2.
kk n=2
(ln n)n
k=1
Sequences and Series 137

4.5.4 Rearrangement of Series

For a finite summation of terms the order in which we add the numbers does not matter, i.e. 1 + 4 + 2 =
4 + 1 + 2, or, in symbols a1 + a2 + a3 = a2 + a1 + a3 . The second summation is a rearrangement
of the first. To make the idea precise we note that the indices on the second summation, (2, 1, 3)
are a permutation of those on the first (1, 2, 3). A permutation on the infinite set of positive
indices (1, 2, 3, . . .) of a series can similarly be defined thereby making the intuitive definition of a
rearrangement of a series precise.
If a series is absolutely convergent then we get the same sum regardless of the order in which the terms
are added (as expected), as summarized in the following theorem.
Theorem 4-21: Any rearrangment of absolutely convergent series ak has the same sum as the
P
original series.
However for a series that is only conditionally convergent the order in which we add the terms does
matter. Indeed we get the following remarkable result:
Theorem 4-22: Riemann Rearrangement Theorem: Let ak be a conditionally convergent
P
series with sum S. Then for any real number R therePexists a rearrangement of series ak having sum
P
R. Additionally there exist rearrangements of series ak which diverge to +∞, −∞, and which fail to
approach any limit, finite or infinite.
Example 4-29
By the latter theorem it follows that the (conditionally convergent) alternating harmonic series can
be rearranged to sum to any number or to diverge.

Exercise 4-5 Answers:

Page 180
1-6: Determine whether the infinite series is convergent or divergent.
∞ ∞ ∞
X (n + 1)! X 5n+1 X (3n)n
1. 3. 5.
n
n=1
e2n n=1
nn n=1
4n + n5
∞ ∞ ∞
X 2n + 25 X n23n X (ln n)n
2. (−1)n 4. (−1)n−1 6.
n=1
3n n=1
7n−1 n=1
nn

7-12: Determine whether the infinite series is absolutely convergent, conditionally convergent, or
divergent.
∞ ∞
n n
20 − n n+1 n + 3
2
X X e
7. 10. (−1)
n=1
n! n=1
(n + 2) n

∞ ∞
X 5n X n!
8. (−1)n+1 11. (−1)n+1
n=1
n10n n=1
(3n − 4)!
∞ n ∞
3n2 + 2 5n
X  X
9. (−1)n 12. (−1)n
n=1
2n4 + 5 n=1
n!
138 4.6 Procedure for Testing Series

4.6 Procedure for Testing Series

We have seen several methods for testing for the convergence and divergence of series. The form of the
series should suggest the type of test to be used. The following steps will be helpful for determining
convergence and divergence of series.

1. Recognize known series with associated convergence and divergence properties:

n
X
telesoping series: Series for which nth partial sum Sn = ak can be directly evaluated and
k=1
lim Sn taken to determine convergence.
n→∞
∞
X ∞
X
geometric series: ark−1 = ark is convergent for |r| < 1 and otherwise divergent.
k=1 k=0
∞
X 1
p-series: is convergent for p > 1 and otherwise divergent.
kp
k=1
(Note that p = 1 is k the (divergent) harmonic series.)
P1

2. If lim ak ̸= 0 or that limit does not exist then the series is divergent by the Term Test for
k→∞
Divergence.
3. If lim ak = 0 then proceed as follows:
k→∞

(a) If the terms of the series are positive, use one of the following tests.
Basic Comparison Test: Useful when ak is a rational or algebraic function of k (i.e.
involving roots of polynomials). Consider a suitable geometric or p-series for comparison.
Remember any comparison series must be positive.
Limit Comparison Test: Same criteria as the Basic Comparison Test. Choose this one if
evaluating the limit of the ratio of comparing terms is easier than proving an inequality
between them as required in the Basic Comparison Test.
Ratio Test: Useful for series involving factorials or other products (including a constant
raised to power k). Do not use this test for rational or algebraic functions of k as these
result in inconclusive (L = 1) results.
Root Test: Useful if ak may be written ak = (bk )k .
R∞
Integral Test: Useful if ak = f (k) for positive, continuous, decreasing f (x) and 1 f (x) dx
is easily evaluated.
(b) If the series is alternating (either (−1)k ak or (−1)k−1 ak for ak > 0) either:
P P

i. Use the Alternating Series Test .

ii. Apply a positive
P series test from
P 3(a)Pabove k−1 to the absolute value of the alternating
series (either (−1)k ak = ak or (−1) ak = ak ) since the convergence of
P
the latter implies the convergence of the alternating series.
(c) If the terms of the series ak are neitherPpositive nor
P alternating, apply a test P from 3(a)
P
above to the absolute value of the series, |ak |. If |ak | is convergent then ak is also
convergent.
(d) If the series only satisfies theorem criteria (positivity, decreasing, alternating, etc.)
Pafter a
∞
certain point (i.e. for k ≥ n for some n) apply the above steps to the tail series k=n ak .
The convergence or divergence of the entire series will be the same as that of the tail series
since they only differ by a finite number of terms of finite sum.
Sequences and Series 139

Example 4-30
Determine the convergence (absolute or conditional) or divergence of the following series.

∞ ∞
X (−1)n n! X (n + 2)3
1. 4.
n=1
(2n + 1)5 n=1
(n3 + 5)2
∞ ∞  n
X n 2n + 1
2. (−1)n
X
5.
n=1
en
n=1
2n + 5
∞ 
1 5
X 
3. + 3
n=1
10n n

Solution:

∞
X (−1)n n!
1. The given series is an with an = . The presence of the factorial suggests trying
n=1
(2n + 1)5
the Ratio Test:

(−1)n+1 (n+1)!
an+1 (2n+3)5 (n + 1)!(2n + 1)5
lim = lim = lim
n→∞ an n→∞ (−1)n n! n→∞ n!(2n + 3)5
(2n+1)5

(n + 1)n!(2n + 1)5 (n + 1)(2n + 1)5

= lim = lim
n→∞ n!(2n + 3) 5 n→∞ (2n + 3)5
5 5
2n + 1 2n + 1
 
= lim (n + 1) lim = lim (n + 1) lim
n→∞ n→∞ 2n + 3 n→∞ n→∞ 2n + 3
5 5
n 2nn + n 2 + n1
 1  
= lim (n + 1) lim · 2n = lim (n + 1) lim
n + n
n→∞ n→∞ n 3 n→∞ n→∞ 2 + 3
n
=∞·1=∞

Therefore the given series is divergent by the Ratio Test.

∞ ∞
X X n
2. The given series is cn = (−1)n an where cn = (−1)n an and an = > 0. So the series
n=1 n=1
en
x
is an alternating series. Let f (x) = , then
ex

x 1 1
lim f (x) = lim x
= lim x = =0
x→∞ x→∞ e LH x→∞ e ∞

=⇒ lim an = lim f (n) = 0

n→∞ n→∞
(1)ex − xex ex (1 − x) 1−x
f ′ (x) = = = < 0 for x > 1
e2x e2x ex
=⇒ an+1 = f (n + 1) ≤ f (n) = an for n ≥ 1

Therefore the given series cn is convergent by the Alternating Series Test.

P
140 4.6 Procedure for Testing Series

∞ ∞
X X n
Now, consider the series of absolute values: an = :
n=1 n=1
en

an+1 n+1
(n + 1)en (n + 1)en n+1
lim = lim en+1
n = lim n+1
= lim n
= lim
n→∞ an n→∞
en
n→∞ ne n→∞ ne e n→∞ ne
1 n +1 n
1 1 + n1 1 1+0 1
=lim · n n = n
lim = · = <1
e n→∞ n n e n→∞ 1 e 1 e

Therefore the series an = |cn | is convergent by the Ratio Test. Thus, the series
P P
∞
X X n
cn = (−1)n n is absolutely convergent.
n=1
e
The astute reader will notice that because the series turned out to be absolutely convergent it
was
P unnecessary to apply the Alternating Series Test. Had we just applied the Ratio Test to
cn directly we could have concluded that series absolutely converged (and hence converged).
As a general rule, if you must determine not only the convergence of a series but the type of
convergence (conditional or absolute) always consider the convergence of the absolute series
first. If that converges the original series is absolutely convergent and no further consideration
of the series is required.
∞
X
3. The given series is an with The given series can be written as:
n=1

∞  ∞ ∞
1 5 1 5
X  X X
+ = +
i=1
10 n n 3
n=1
10 n
n=1
n3
∞  n−1 ∞
X 1 1 X 1
= · + 5
n=1
10 10 n=1
n 3
| {z } | {z }
geometric series p-series
a = 10
1
, r = 10
1
p=3
|r| < 1 (convergent) p > 1 (convergent)

Thus the given series is convergent as it is the sum of two convergent series. The convergence
is absolute as the series is already positive.
∞
X (n + 2)3
4. The given series is an with an = . Consideration of the order of the polynomials
n=1
(n3 + 5)2
∞ ∞
n3 n3 1 X X 1
suggests a comparison with bn where bn = 3 2 = 6 = 3 . Then bn = is a
P
(n ) n n n=1 n=1
n 3

convergent p-series since p = 3 > 1. Also,

(n+2)3
an (n3 +5)2 n3 (n + 2)3 n3 (n3 + 6n2 + 12n + 8)
lim = lim = lim = lim
n→∞ bn n→∞ 1 n→∞ (n3 + 5)2 n→∞ n6 + 10n3 + 25
n3
n6 6n5 12n4
n6 + 6n5 + 12n4 + 8 n6 + + + 8
= lim = lim · n6 n6 n6 n6
n→∞ n6 + 10n3 + 25 n→∞ n6 n6
n6 + 10n3
n6 + 25
n6
1 + n6 + n122 + n86 1+0+0+0
= lim = =1>0
n→∞ 1 + n103 + n256 1+0+0

Therefore the given series is also convergent by the Limit Comparison Test. Since the series is
positive it is absolutely convergent.
Sequences and Series 141

∞
X  n
5. The given series is an with an = 2n+1
2n+5 . The power in the term suggests trying the
n=1
Root Test:
s n
2n + 1 2n + 1 n 2n
+ 1
lim |an | = lim = lim = lim
p
n n n n
·
n→∞ n→∞ 2n + 5 n→∞ 2n + 5 n→∞ n 2n
n + 5
n
2+ 1
2+0
= lim n
= =1
n→∞ 2 + 5
n
2+0
Therefore the Root Test is inconclusive. Next
x consider trying the Term Test for Divergence.
2x + 1

To evaluate the limit let f (x) = . Then lim f (x) is the indeterminate form 1∞ .
2x + 5  x→∞
2x + 1
Taking the logarithm gives ln f (x) = x ln and so
2x + 5
2x + 1
 
lim ln f (x) = lim x ln (∞ · 0 form)
x→∞ x→∞ 2x + 5
ln( 2x+5
2x+1
) 0
 
= lim form
x→∞ x−1 0
1 2(2x+5)−2(2x+1)
2x+1 · (2x+5)2
4x+10−4x−2
(2x+5)(2x+1)
= lim 2x+5
= lim
LH
x→∞ −x−2 x→∞ −x−2

8x2 8x2
= − lim = − lim
x→∞ (2x + 5)(2x + 1) x→∞ 4x2 + 12x + 5
16x 16
= − lim = − lim = −2 .
x→∞ 8x + 12 x→∞ 8
LH LH

Thus lim f (x) = e−2 and therefore lim an = lim f (n) = e−2 ̸= 0 .
x→∞ n→∞ n→∞

Thus the given series is divergent by the Term Test for Divergence.

Further Questions:

Determine the convergence (absolute or conditional) or divergence of the following series.

∞ ∞
2k 2 1
 
X X π
1. 6. sin +
k +1
2 2 k
k=1 k=1
∞
n ∞
kk
X
2. (−1)n
X
n2 +1 7.
n=1 10k
k=1
∞
X 1000 − n √
3. ∞
X n−1 n
n! 8. (−1)
n=1
n=1
n2 + 1
∞
X
4. e−2k ∞ 
1 5
X 
k=1 9. +√
3n n
∞ n=1
X 5n
5.
n=1
n6 3n+1
 142 4.6 Procedure for Testing Series

Answers:
Exercise 4-6
Page 180

1-14: Determine whether the infinite series is absolutely convergent, conditionally convergent, or
divergent.
∞ ∞
X 5n3 + 4n + 2 X 50
1. 8. √
n=1
7n3 + 2n2 + 6 n=1
n + 10
∞ ∞
cos n1


X n2 X
2. (−1) n+1
9.
n=1
5n3 + 6 n=1
n2
∞ ∞
X nn X 2
3. (−1)n−1 10.
n=1
(n!)n n=1
n(n + 2)
∞ ∞
X
n−1 e 2n n X
4. (−1) 11. [ln(2n + 3) − ln(n + 2)]
n=1
n3n n=1

∞ ∞ 
4 3
X X 
5. tan−1 n 12. +
n=1 n=1
3n 4n
∞ ∞
nπ + 5 cos nπ
X   X
6. tan 13.
n=1
4n + 7 n=1
n
∞ ∞
X 2n3 + 4 X 5n + 7
7. 14. (−1)n+1
n=1
3n5 + 6 n=1
2n3 + 25
Sequences and Series 143

4.7 Power Series

We have seen that in many cases the terms of the series ak may be writtenPas a function of the
P
summation index, namely ak = f (k), for some function f . However if the series ak is convergent its
value is a number, namely its sum. The final sum does not depend on the index k in the same way
Rb
that a definite integral a f (t) dt results in a number independent of the dummy variable t.
Consider, however, the situation where the terms ak depend additionally on an actual variable, say x,
different from the summation index, present in the series (i.e. ak = ak (x)). In this case the sum of the
series (and indeed its convergence) depends on the value of x.
Example 4-31
The geometric series with a = 1 and r = x is given by

∞
X
xk = 1 + x + x2 + · · · + xk + · · ·
k=0

1
Here ak (x) = xk . The sum is now a function of x, namely , and is valid for |x| < 1 for which
1−x
the series is convergent.

We could introduce the variable x into the terms of a series in many ways, for instance
∞
X sin(kx)
.
k!
k=1

If we choose to introduce it as in our geometric series above, namely so that the terms of the series
look like terms in a polynomial, ck xk , we have a power series.

Definition: Let x be a variable. A series of the form

∞
X
ck xk = c0 + c1 x + c2 x2 + · · · + ck xk + · · · ,
k=0

where ck are real constants (for k = 0, 1, 2 . . .) is a power series in x . The constants ck are
called the coefficients of the series.

Example 4-32
The geometric series with r = x above, 1 + x + x2 + · · · + xk + · · · , is a power series in x with
coefficients ck = 1 for all k.

If we choose to make the k th term of the series have the more general form ck (x − a)k we get the
following.

Definition: Given real constant coefficients ck and real constant a the power series in (x − a) is
∞
X
ck (x − a)k = c0 + c1 (x − a) + c2 (x − a)2 + · · · + ck (x − a)k + · · · .
k=0

The series is also known as the power series about a or centred on a .

144 4.7 Power Series

A power series in x is just a special case of this last definition with a = 0.

Since the convergence of a power series (and indeed its sum should it converge) will, in general, depend
on the value of the variable x, an obvious question is to find the values of x for which the power series
is convergent.
Example 4-33
Find the values of x for which the following power series are convergent.
∞ ∞
X (−1)n n X xn
1. x 3.
n=1
(n + 1)! n=2
ln n
∞ ∞
X 2n + 1 n X
2. x 4. nn (x + 3)n
n=1
3n n=1

Solution:
∞
X (−1)n n
1. The power series is an where an = x . Consider the following Ratio Test limit:
n=1
(n + 1)!

(−1)n+1 n+1
an+1 (n+2)! x xn+1 (n + 1)!
lim = lim = lim
n→∞ an n→∞ (−1)n n n→∞ (n + 2)! xn
(n+1)! x

(n + 1)! |x|
= lim |x| = lim =0
n→∞ (n + 2)(n + 1)! n→∞ n + 1

Note that here, since the limit is in n, the value x is treated as a constant for the purpose of
evaluating the limit. Since the limit of 0 is less than 1 we have, by the Ratio Test, that the
power series converges for all real values of x.
∞
X 2n + 1 n
2. The power series is an where an = x . The Ratio Test limit is:
n=1
3n

an+1 2n+3 n+1

3n+1 x (2n + 3)xn+1 3n 2n + 3
lim = lim = lim = lim |x|
n→∞ an n→∞ 2n+1 n
3 n x n→∞ 3n+1 (2n + 1)x n n→∞ 3(2n + 1)

n 2n
+3 2 + n3 2+0 1
= lim · n2n n1
|x| = lim
|x| = |x| = |x|
n→∞ n 3
n + n
n→∞ 3+ 2+ n 1 3(2 + 0) 3

Thus, by the Ratio Test, the series is convergent if the limit satisfies
1
|x| < 1 =⇒ |x| < 3 =⇒ −3 < x < 3 .
3
1
The Ratio Test is inconclusive if |x| = 1 =⇒ |x| = 3 (so x = 3 or x = −3). We must study
3
these two values separately.
• If x = 3 then the series becomes:
∞ ∞
X 2n + 1 n X
3 = (2n + 1) .
n=1
3n n=1

Then lim (2n + 1) = ∞ shows that when x = 3 the series diverges by the Term Test for
n→∞
Divergence.
Sequences and Series 145

• If x = −3 then the series becomes:

∞ ∞
X 2n + 1 X
(−3) n
= (−1)n (2n + 1) .
n=1
3n
n=1

Then the fact that lim (−1)n (2n + 1) does not exist shows that when x = −3 the series
n→∞
is divergent by the Term Test for Divergence.

Therefore the given power series is convergent if −3 < x < 3 and divergent otherwise.
∞
X xn
3. The power series is an where an = . The Ratio Test limit is:
n=2
ln n

xn+1
an+1 ln(n+1) xn+1 ln n ln n
lim = lim = lim · n = lim |x|
n→∞ an n→∞ xn
ln n
n→∞ ln(n + 1) x n→∞ ln(n + 1)

ln n
Again |x| here is just a constant with respect to the limit and we need to evaluate lim .
n→∞ ln(n + 1)
ln y
If f (y) = , then
ln(y + 1)

ln y ∞ 
lim f (y) = lim form
y→∞ y→∞ ln(y + 1) ∞
1
y+1 1
 
y
= lim = lim = lim 1 + = 1 + 0 = 1.
y→∞ 1 y→∞ y y→∞ y
LH y+1

Thus,

an+1 ln n
lim = lim |x| = (1) |x| = |x|
x→∞ an n→∞ ln(n + 1)

Therefore, the series is convergent if |x| < 1 =⇒ −1 < x < 1 by the Ratio Test. The Ratio
Test is inconclusive when |x| = 1 (so x = 1 and x = −1) and we must study these two values
separately.

• If x = 1 then the series becomes:

∞ ∞
X 1n X 1
= .
n=2
ln n n=1 ln n

∞ ∞
1 1 X X 1
Let an = , bn = . Then bn = is a divergent harmonic series.
ln n n n=2 n=2
n
Also for n ≥ 2 we have
1 1
0 < ln n ≤ n =⇒ ≥ for n ≥ 2
ln n n
Thus when x = 1 the series is divergent by the Basic Comparison Test.
• If x = −1 then the series becomes an alternating series:
∞ ∞
X 1 X 1
(−1)n = (−1)n an where an = >0
n=2
ln n n=2
ln n
146 4.7 Power Series

Then
1 1
lim an = lim = =0
n→∞ n→∞ ln n ∞
1 1
If f (y) = , then f ′ (y) = − < 0 for y ≥ 2 so an = f (n) is a decreasing
ln y y(ln y)2
sequence. Thus when x = −1 the series is convergent by the Alternating Series Test.

Therefore the given power series is convergent if −1 ≤ x < 1 .

∞
X
4. The power series is an where an = nn (x + 3)n . Since (x + 3) = (x − (−3)) this is a power
n=1
series centred at a = −3. Consider the following Root Test limit:
√ p
lim n |an | = lim n |nn (x + 3)n | = lim n nn n |x + 3|n = lim n|x + 3|
p p
n→∞ n→∞ n→∞ n→∞
0 if x = −3

=
∞ if x ̸= −3

Notice that while x and hence x + 3 are effectively constants when evaluating the limit, the
case where |x + 3| = 0 (so x = −3) does result in a unique limit of zero since an = 0 when
x = −3. By the Root Test only in that case is the limit less than 1. Hence the power series
converges only for x = −3.
As an aside, while the Root Test can be used to analyze power series convergence, the Ratio
Test is almost always preferred. The Root Test usually results in limits with indeterminate
forms requiring evaluation. This example was an exception.

Further Questions:

Find the values of x for which the following power series are convergent.
∞ ∞ ∞
X xk X n2 n X ln n
1. 2. x 3. (x − e)n
k! n=0
2n n=1
en
k=0

As suggested by the previous example a power series about a will converge on an interval centred on a
as detailed in the following theorem.
∞
X
Theorem 4-23: The power series ck (x − a)k will either:
k=0

1. Converge only at x = a .
2. Converge for |x − a| < R and diverge for |x − a| > R for some positive real number R .

3. Converge for all x .

P∞
Definition: The radius of convergence R for a power series k=0 ck (x − a)k is the value R if
Part 2 of Theorem 4-23 applies. For Part 1 the radius of convergence is defined to be R = 0 and
for Part 3 the radius of convergence is defined to be R = ∞.
Sequences and Series 147

P∞
Definition: The interval of convergence I of a power series k=0 ck (x − a)k is the set of values
of x for which the series converges. For the three possibilities of convergence one has:
1. The set containing the single value x = a (i.e. {a}) for R = 0 .
2. One of [a − R, a + R], [a − R, a + R), (a − R, a + R], or (a − R, a + R) for R > 0 finite.
3. (−∞, ∞) for R = ∞ .

The choice of interval in the second case depends upon the convergence or not of the series at the
interval endpoint values x = a ± R .

Example 4-34
For Example 4-33 find the radii and intervals of convergence of each series.
Solution:

Consideration of the solutions of Example 4-33 shows the radii and intervals of convergence are:

1. R = ∞, I = (−∞, ∞) 3. R = 1, I = [−1, 1)
2. R = 3, I = (−3, 3) 4. R = 0, I = {−3}

Further Questions:

For Example 4-33 find the radii and intervals of convergence of each series in the Further Questions.

Example 4-35
Find the radius and interval of convergence of each of the following series.
∞ ∞
X xn 3n X (−5)n (3x − 2)n
1. 3.
n=1
(n!)2 n=1
n2
∞ ∞
X n!(x − 2)n X (x − 4)n
2. 4. (−1)n+1 √
n=1
5n n=1
n+1

Solution:
∞
X xn 3n
1. The power series is an where an = . The Ratio Test limit is
n=1
(n!)2

xn+1 3n+1
an+1 [(n+1)!]2 (n!)2 (3) 3n!n!
lim = lim = lim |x| = lim |x|
n→∞ an n→∞ x n 3n n→∞ [(n + 1)!] 2 n→∞ [(n + 1)n!]2
(n!)2

3
= lim |x| = 0 · |x| = 0 < 1 for all x .
n→∞ (n + 1)2

Therefore, the power series converges for all x. Thus the radius of convergence is R = ∞ and
the interval of convergence is I = (−∞, ∞).
148 4.7 Power Series

∞
X n!(x − 2)n
2. The power series is an where an = . The Ratio Test limit is
n=1
5n
(n+1)!(x−2)n+1
an+1 5n (n + 1)n! n+1
lim = lim 5n+1
= lim |x − 2| = lim |x − 2|
n→∞ an n→∞ n!(x−2)n n→∞ 5n+1 n! n→∞ 5
5n
0 if x = 2

=
∞ if x ̸= 2
This power series is convergent only if x = 2. Therefore, the radius of convergence is R = 0
and the interval of convergence is I = {2}. Note that the interval is centred at 2 as expected
for a power series about a = 2 .
∞
X (−5)n (3x − 2)n
3. The power series is an where an = . The Ratio Test limit is
n=1
n2
(5)n+1 (3x−2)n+1 2
5n2

an+1 (n+1)2 n
lim = lim = lim |3x − 2| = 5 lim |3x − 2|
n→∞ an n→∞ (−5)n (3x−2)n n→∞ (n + 1)2 n→∞ n+1
n2
2 2
n
1
 
n
= 5 lim · n
|3x − 2| = 5 lim |3x − 2|
n→∞ n n
n + 1
n
n→∞ 1+ 1
n
2
1

=5 |3x − 2| = 5|3x − 2|
1+0
The series is convergent by the Ratio Test if:
2 2 1
5|3x − 2| < 1 =⇒ 15 x − < 1 =⇒ x − <
3 3 15
1 2 1 1 2 1 2
=⇒ − <x− < =⇒ − + < x < +
15 3 15 15 3 15 3
3 11
=⇒ <x<
5 15
3 11
The Ratio Test is inconclusive (limit equals 1) at the endpoints x = and x = . We must
5 15
test these two values separately.
3
• If x = then the series becomes:
5
∞
n ∞ ∞
X (−5)n 9 − 2 X (−5)n (− 51 )n X 1
5
= =
n=1
n2 n=1
n 2
n=1
n2

This is a convergent p-series since p = 2 > 1.

11
• If x = then the series becomes:
15
∞
n ∞ ∞
X (−5)n 33 − 2 X (−5)n ( 1 )n X (−1)n
15 5
=
n=1
n2 n=1
n2 n=1
n2
1 1
Consider an = 2
and define f (y) = 2 then:
n y
1
lim f (y) = lim = 0 =⇒ lim an = lim f (n) = 0 .
y→∞ y→∞ y2 n→∞ n→∞
2
f ′ (y) = − < 0 for y > 0 =⇒ an = f (n) decreases .
y3
Sequences and Series 149

∞
X (−1)n
Thus the series is convergent by the Alternating Series Test.
n=1
n2
1
Therefore the radius of convergence of the power series is R = and the interval of convergence
15
is I = 5 , 15 . Notice that the interval is centred at a = 23 . That the original power series is
 3 11 

centred at this value can be seen by writing (3x − 2)n = 3n (x − 2/3)n in the original series.
∞
X (x − 4)n
4. The power series is an where an = (−1)n+1 √ . The Ratio Test limit is
n=1
n+1
(−1)n+2 (x−4)n+1 √
√
n+1 n+1
r
an+1 n+2
lim = lim = lim √ |x − 4| = lim |x − 4|
n→∞ an n→∞ (−1) n+1
√
(x−4)n n→∞ n+2 n→∞ n+2
n+1
s s
n n
+ n1 1 + n1
= lim · n
|x − 4| = lim |x − 4| = |x − 4|
n→∞ n n
n + n2 n→∞ 1 + n2
The series is convergent by the Ratio Test if |x − 4| < 1 = R. Then
|x − 4| < 1 =⇒ −1 < x − 4 < 1 =⇒ 3 < x < 5
which is as expected since the series is centred at a = 4 with radius 1. The Ratio Test is
inconclusive if x = 3 or x = 5 and we test these values separately.
• If x = 3 then the series becomes
∞ ∞ ∞
X (−1)n+1 (−1)n X (−1)2n+1 X 1
√ = √ =− √
n=1
n + 1 n=1
n + 1 n=1
n +1
∞ ∞
1 1 X X 1 1
Let an = √ and bn = √ . Then bn = √ is a p-series with p = ≤ 1 and
n+1 n n=1 n=1
n 2
hence divergent. Taking the limit
√1 √ r r
an n+1 n n n
lim = lim = lim √ = lim = lim
n→∞ bn n→∞ √1
n
n→∞ n + 1 n→∞ n + 1 n→∞ n + 1
s s
n
1 1
r
n
= lim · n n
1 = lim = =1>0
n→∞ n
n + n
n→∞ 1 + 1
n
1 + 0
∞
X 1
shows the series √ is divergent by the Limit Comparison Test.
n=1
n+1
• If x = 5 then the series becomes
∞ ∞
X (−1)n+1 (1)n X (−1)n+1
√ = √
n=1
n+1 n=1
n+1
1 1 1
Then lim √ = = 0. Let f (y) = √ then:
n→∞ n+1 ∞ y+1
1
f ′ (y) = − 3 < 0 for x ≥ 1
2(y + 1) 2
∞
X (−1)n+1
shows the magnitude of the terms, f (n), is decreasing. Therefore the series √
n=1
n+1
converges by the Alternating Series Test.
 150 4.7 Power Series

Therefore the radius of convergence of the power series is R = 1 and the interval of convergence
is I = (3, 5] .

Further Questions:

Find the radius and interval of convergence of each of the following series.
∞ ∞
X 1 X
1. (−1)k (x − 3)k 3. k!(2x − 1)k
k+1
k=0 k=0

∞ ∞
X X (2x − 3)n
2. n3 (x − 5)n 4.
n=0 n=1
n 3n

Answers:
Exercise 4-7
Page 180

1-5: Find the interval and the radius of convergence of the power series.
∞
X n2 + 4 n
1. x
n=0
2n3 + 5
∞
X ln n
2. 2
(x − 1)n
n=2
n
∞
X 3n 3n
3. x
n=0
(3n)!
∞
X 1
4. (−1)n (4x − 1)n
n=1
n3n
∞
X n!
5. (x − 1)n
n=0
10 n
Sequences and Series 151

4.8 Representing Functions with Power Series

The sum of a power series for a given x from its interval of convergence I results in a number. As such
it is natural to consider the power series as defining a function f (x) of x on I with the value of the
function being the sum. We write
X∞
f (x) = ck xk (x ∈ I)
k=0

If the sum of the power series can be written in a closed form, then the power series can be considered
a representation of that function valid on I.

Example 4-36
P∞
The power series k=0 xk = 1 + x + x2 + · · · converges for |x| < 1, and therefore is a function of x on
I = (−1, 1). On that interval the sum for given x is 1−x
1
. The power series thus is a representation
of the function 1−x on this restricted domain:
1

∞
1 X
= 1 + x + x2 + · · · + xk + · · · = xk for |x| < 1
1−x
k=0

One can find power series representations of other functions using known power series.

Example 4-37
Find representations of the following functions with power series. Indicate the values of x for which
the representatives are valid.

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

Solution:

1. The given functions can be written as:

1 1
= .
2 − 5x 2(1 − 25 x)

1
The power series representation of is:
1−x
∞
1 X
= xn , |x| < 1 .
1 − x n=0

Thus
∞  n ∞  n ∞
1 1 1X 5 1X 5 X 5n n
= = x = x n
= x ,
2 − 5x 2(1 − 2 x)
5 2 n=0 2 2 n=0 2 n=0
2n+1

which converges when

5 2 2 2
x < 1 =⇒ |x| < =⇒ − < x < .
2 5 5 5
152 4.8 Representing Functions with Power Series

2. This function can be written as

∞
x 1 X
n
=x· =x −4x2 , | − 4x2 | < 1
1 + 4x2 1 − (−4x2 ) n=0
∞
X
= (−1)n 4n x2n+1 .
n=0

Note here that the variable x is effectively a constant with respect to the series in n and can
therefore be brought in and out of the summation without affecting its convergence properties.
The series converges if
1 1
| − 4x2 | < 1 =⇒ |4x2 | < 1 =⇒ 4x2 < 1 =⇒ x2 < =⇒ |x| < .
4 2

3. This function can be written as

x2 x2 x2 1
= = ·
4 + 3x 4(1 + 4 x2 )
3 4 1 − − 43 x2
2


∞  n X ∞
x2 X 3 3n
= − x2 = (−1)n n+1 x2n+2 .
4 n=0 4 n=0
4

The power series converges if

3 3 4 2
− x2 < 1 =⇒ x2 < 1 =⇒ x2 < =⇒ |x| < √ .
4 4 3 3

Further Questions:

Find representations of the following functions with power series. Indicate the values of x for which
the representations are valid.

1 1 3.
x
1. 2. 2 − 3x
1+x 1 − x3

As functions power series behave like polynomials. They are continuous and can be termwise differenti-
ated and integrated to produce new functions of x as detailed in the following theorem.
Theorem 4-24: Suppose f (x) is defined by the power series
∞
X
f (x) = ck (x − a)k = c0 + c1 (x − a) + c2 (x − a)2 + · · · + ck (x − a)k + · · ·
k=0

with radius of convergence R > 0 (finite or infinite). Then on the interval (a − R, a + R) :

1. f (x) is continuous.
2. f (x) is differentiable with derivative
∞
X
f ′ (x) = kck (x − a)k−1 = c1 + 2c2 (x − a) + 3c3 (x − a)2 + · · · + kck (x − a)k−1 + · · ·
k=0
Sequences and Series 153

3. f (x) is integrable with integral

Z ∞
X ck c1 c2 ck
f (x) dx = C+ (x−a)k+1 = C+c0 (x−a)+ (x−a)2 + (x−a)3 +· · ·+ (x−a)k+1 +· · ·
k+1 2 3 k+1
k=0

The series resulting from differentiation and integration both have radius of convergence R .
Note that the above theorem shows that for power series the calculus operations of differentiation and
integration can be exchanged with summation, just as occurs with finite sums.
d X X d 
ck (x − a)k = ck (x − a)k

dx dx
Z hX i X Z 
ck (x − a)k dx = ck (x − a)k dx

If the power series is a representation of a function with a closed form, differentiation and integration
can be used to find power series representations of other functions.
Example 4-38
Find a power series representation for each of the following functions. Indicate the interval for
which the representation is valid.

1 2. ln(1 − x)
1.
(4 − x)2

Solution:

1
1. To find the power series representation of we note that it is just the derivative of the
(4 − x)2
1
function which has the representation
4−x
∞ ∞
1 1 1 X  x n X xn
= = =
4−x 4(1 − x4 ) 4 n=0 4 n=0
4n+1
∞
1 X xn x
=⇒ = , < 1.
4 − x n=0 4n+1 4

Differentiating both sides with respect to x we have

∞ ∞
1 d X xn
   n 
d X d x
= =
dx 4−x dx n=0 4 n+1
n=0
dx 4n+1
∞ ∞
1 X 1 X n n−1
= nx n−1
= x ,
(4 − x)2 n=0
4n+1
n=1
4n+1

where we changed the starting index to n = 1 since the n = 0 term equals zero. The power
series converges if
x
< 1 =⇒ |x| < 4 .
4
2. For the geometric series we have
∞
1 X
= xn , |x| < 1 .
1 − x n=0
154 4.8 Representing Functions with Power Series

We desire a power series representation of ln |1 − x| which is, up to a factor of −1, just the
integral of this function. Integrating both sides with respect to x we obtain
∞ ∞ Z
!
1
Z Z X X
dx = x n
dx = xn dx
1−x n=0 n=0
| {z }
u=1−x,du=−dx
∞
X xn+1
=⇒ − ln |1 − x| = +C
n=0
n+1
∞
X xn+1
=⇒ ln |1 − x| = − −C
n=0
n+1

For fixed x the function on the left is a number and so the series on the right cannot contain
an arbitrary constant. To determine what it equals evaluate both sides at x = 0 to get

ln |1| = −0 − C =⇒ C = 0 .

Thus
∞
X xn+1
ln(1 − x) = − ,
n=0
n+1

which converges for |x| < 1 as follows by Theorem 4-24. Since 1 − x > 0 in that interval the
absolute value bars have been dropped.
The series representation can be simplified by changing to a new index. Similar to introducing
a new variable in an integral, here let m = n + 1. The new limits of the series then become

n = ∞ =⇒ m = ∞ + 1 = ∞
n = 0 =⇒ m = 0 + 1 = 1

and our series representation simplifies to

∞
X xm
ln(1 − x) = − , |x| < 1 .
m=1
m

Just like an integration variable we can change the index name back to n now if we so desire.
Reindexing in this manner has several uses. In addition to simplifying a series, one can also
compare two series whose lower limits are not the same by reindexing one to have the same
limits as the other. So for example one can show the following two forms of the geometric
series are equivalent,
X∞ X∞
arn = arm−1 ,
n=0 m=1

by letting m = n + 1 in the original sequence, changing the limits as above, and noting that
m = n + 1 =⇒ n = m − 1 when rewriting the general term of the first series in terms of
the new index m. Reindexing also allows one to align the limits of two series to allow their
addition. Alternatively one may reindex power series so that they all have the same power of
xn so that they may be added and simplified.
Our new version of the representation ln(1 − x) shows that at x = 1 the series representation
diverges as it reduces to the harmonic series. This is unsurprising as ln(1 − 1) = ln(0) is
undefined as well. At x = −1 the series becomes the alternating harmonic series and therefore
converges conditionally at that value. The representation is, in fact, valid at x = −1 as its
sum may be shown to be ln(1 − (−1)) = ln 2. As such the series representation of ln(1 − x) is
valid on [−1, 1) .
Sequences and Series 155

Further Questions:

Find a power series representation for each of the following functions. Indicate the interval for which
the representation is valid.

1 2. ln(1 + x) 3. tan−1 x
1.
(1 + x)2

Because power series representations are easy to integrate, they may be used as a means to integrate
difficult functions. The integral will only be valid for x lying within (a − R, a + R).

Example 4-39
Integrate each of the following using a power series.

1 ln(1 − x)
Z Z
1. dx 2. dx
1 + x6 x

Solution:
∞
1 X
1. Starting with the geometric series = xn , |x| < 1 first get a representation for the
1 − x n=0
integrand:
∞ ∞
1 1 X X
= = (−x ) =
6 n
(−1)n x6n ,
1 + x6 1 − (−x6 ) n=0 n=0

which converges for | − x6 | < 1 =⇒ |x| < 1 . Thus

∞ ∞ ∞
1 x6n+1 (−1)n 6n+1
Z Z X X X
dx = (−1)n x6n dx = (−1)n +C = x +C,
1 + x6 n=0 n=0
6n + 1 n=0
6n + 1

which is valid for |x| < 1 .

∞
X xn
2. In Example 4-38 Problem 2 we found ln(1 − x) = − , |x| < 1 . Thus
n=1
n

∞ ∞ ∞
ln(1 − x) 1 X xn 1 1 x(n−1)+1
Z Z X Z X
dx = − dx = − xn−1 dx = − +C
x x n=1 n n=1
n n=1
n (n − 1) + 1
∞
X 1 n
=− x +C,
n=1
n2

which is valid for |x| < 1 .

These indefinite integrals could be used to approximate a definite integral by suitably truncating
the series at large n and subtracting its evaluation at the limits of the integral as we would for
any antiderivative. The integration limits would need to be within the interval (−1, 1) due to the
representation of these antiderivatives only being valid there.
 156 4.8 Representing Functions with Power Series

Further Questions:

Integrate each of the following using a power series.

1
Z Z
x
1. dx 2. dx
1+x 3 1 − x4

Answers:
Exercise 4-8
Page 180

1-5: Find a power series representation for f (x) and specify the interval of convergence.
1
1. f (x) =
1 + x2
x
2. f (x) =
5 − 4x
x
3. f (x) =
2 + x2
1
4. f (x) =
(2 + 3x)2
5. f (x) = ln(3 + 2x)
Sequences and Series 157

4.9 Maclaurin Series

P∞
We expect some functions f (x) can be represented by an infinite series k=0 ck x for a certain,
k

potentially restricted, domain. While we verified that 1−x could be represented by the geometric series,
1

we would like a general mechanism for determining the coefficients of the power series that correspond
to an arbitrary f (x). Suppose f (x) has a power series expansion:

f (x) = c0 + c1 x + c2 x2 + c3 x3 + c4 x4 + · · · .

One observes that if we set x = 0 that f (0) = c0 . In other words, we can determine c0 by evaluating f
at x = 0. We can differentiate the power series above term by term to get:

f ′ (x) = c1 + 2c2 x + 3c3 x2 + 4c4 x3 + · · ·

If we evaluate this at x = 0 we get c1 = f ′ (0). The next derivative is:

f ′′ (x) = 2c2 + 6c3 x + 12c4 x2 + · · ·

′′
and so c2 = f 2(0) . Repeated differentiation and evaluation relates the k th coefficient ck to the derivative
f (k) evaluated at x = 0 as follows:
f (k) (0)
ck =
k!
where recall the factorial is defined by k! = k · (k − 1)· · ·2 · 1. The formula is true for k = 0 as well
with the convention f (0) (x) = f (x) and noting that 0! = 1 by definition. Plugging our ck into the
original power series we get the following definition.

Definition: Given function f (x) differentiable to all orders at x = 0 the power series in x given by
∞
X f (k) (0) f ′′ (0) 2 f ′′′ (0) 3 f (k) (0) k
f (x) = xk = f (0) + f ′ (0)x + x + x + ··· x + ···
k! 2! 3! k!
k=0

is the Maclaurin series for f (x).

As a power series, the Maclaurin series must converge for |x| < R for some radius of convergence R
dependent upon the function.4

Example 4-40
Find the Maclaurin series and its integral of convergence for each of the following functions.

1. f (x) = e2x 2. f (x) = cos x

Solution:
∞
X f (n) (0) n
1. Since the Maclaurin series is f (x) = x we compute the following derivatives and
n=0
n!

4We will discuss shortly whether the Maclaurin series of f (x) actually is a valid representation of the function on this

interval.
158 4.9 Maclaurin Series

evaluate them at zero:

f (x) = e2x , f (0) (0) = f (0) = e0 = 1

f ′ (x) = 2e2x , f ′ (0) = 2e0 = 2
f ′′ (x) = 4e2x , f ′′ (0) = 4e0 = 4
f ′′′ (x) = 8e2x , f ′′′ (0) = 8e0 = 8
f (4) (x) = 16e2x , f (4) (0) = 16e0 = 16
f (5) (x) = 32e2x , f (5) (0) = 32e0 = 32

Since 0! = 1 and 1! = 1 the Maclaurin series is given by:

1 ′′ 1 1 1
f (x) = f (0) + f ′ (0)x +f (0)x2 + f ′′′ (0)x3 + f (4) (0)x4 + f (5) (0)x5 + . . .
2! 3! 4! 5!
1 1 1 1
=⇒ e2x = 1 + 2x + (4)x + (8)x + (16)x + (32)x + . . .
2 3 4 5
2! 3! 4! 5!
22 2 23 3 24 4 25 5
=⇒ e2x = 1 + 2x + x + x + x + x + . . .
2! 3! 4! 5!
∞
X 2n n
=⇒ e2x = x
n=0
n!

Alternatively if one only wants the series in sigma notation one can observe that f (n) (0) = 2n
and substitute that immediately into our Maclaurin formula.
To find the interval of convergence we will use the Ratio Test.

2n+1 xn+1
an+1 (n+1)! 2n+1 xn+1 n! 1
lim = lim = lim = 2 lim |x|
n→∞ an n→∞ 2n xn
n!
n→∞ (n + 1)n! 2n xn n→∞ n + 1

= 2(0)|x| = 0 < 1 for all x

Thus the integral of convergence of the Maclaurin series is I = (−∞, ∞) .

2. We compute the following derivatives and evaluate them:

f (x) = cos x , f (0) = cos 0 = 1

f (x) = − sin x ,
′
f ′ (0) = − sin 0 = 0
f (x) = − cos x ,
′′
f ′′ (0) = − cos 0 = −1
f ′′′ (x) = sin x , f ′′′ (0) = sin 0 = 0
f (4) (x) = cos x , f (4) (0) = cos 0 = 1
.. ..
. . (pattern repeats)

Thus the Macluarin series is

1 1 1
cos x = 1 + (−1)x2 + (1)x4 + (−1)x6 + . . .
2! 4! 6!
1 2 1 4 1 6
=⇒ cos x = 1 − x + x − x + . . .
2! 4! 6!
∞ 2n
X x
=⇒ cos x = (−1)n .
n=0
(2n)!
Sequences and Series 159

To find the interval of convergence, we will use the Ratio Test.

(−1)n+1 x2n+2
an+1 (2n+2)! (2n)! (2n)!
lim = lim = lim x2 = lim x2
n→∞ an n→∞ (−1)n x2n n→∞ (2n + 2)! n→∞ (2n + 2)(2n + 1)(2n)!
(2n)!
1
= lim x2 = 0 · x2 = 0 < 1 for all x
n→∞ (2n + 2)(2n + 1)

Therefore the interval of convergence is I = (−∞, ∞).

One notes that the fact that cosine is an even function is reflected in the disappearance of odd
powers of x in its Maclaurin series which make it look like an even polynomial with an infinite
number of terms.
Also note that the n index in our found series is not the same as the n that appears in the
Maclaurin formula. The series found skips the terms with zero coefficient. This is not merely
for convenience. It was required so that the Ratio Test could be applied to the series to find
convergence.

Further Questions:

Find the Maclaurin series and its interval of convergence for each of the following functions.

1 3. f (x) = sin x
1. f (x) =
1−x
2. f (x) = ex 4. f (x) = ln(1 + x)

Some important Maclaurin series and their domains of validity are:

∞
1 X
= xk = 1 + x + x2 + x3 + · · · ; (−1, 1)
1−x
k=0
∞
X xk x x2 x3
ex = =1+ + + + ··· ; (−∞, ∞)
k! 1! 2! 3!
k=0
∞
X x2k+1 x3 x5 x7
sin x = (−1)k =x− + − + ··· ; (−∞, ∞)
(2k + 1)! 3! 5! 7!
k=0
∞
X x2k x2 x4 x6
cos x = (−1)k =1− + − + ··· ; (−∞, ∞)
(2k)! 2! 4! 6!
k=0
∞
X x2k+1 x3 x5 x7
tan−1 x = =x− + − + ··· ; [−1, 1]
2k + 1 3 5 7
k=0
∞
X xk+1 x2 x3 x4
ln(1 + x) = (−1)k =x− + − + ··· ; (−1, 1]
k+1 2 3 4
k=0

With these series representations we now have a means of numerically calculating these functions for
any value of x in the interval of convergence to arbitrary precision!
 160 4.9 Maclaurin Series

As we saw with the more general power series, we can derive series of new functions from known
Maclaurin series.

Example 4-41
For the following functions we have the power series:

∞
k ∞
3 X x3 X x3k
1. ex = =
k! k!
k=0 k=0

Series is valid for x3 in (−∞, ∞) which implies for x in (−∞, ∞).

2.
∞ ∞
1 1 1X (2x)2k 1 X 22k−1 x2k
sin2 x = (1 − cos 2x) = − (−1)k = + (−1)k+1
2 2 2 (2k)! 2 (2k)!
k=0 k=0
∞
X 22k−1 x2k
= (−1)k+1
(2k)!
k=1

(Here we simplified our expression in the last line by noting the k = 0 term in the series equals
−1/2 and so cancels the 1/2.) Series is valid for 2x in (−∞, ∞) and so for x in (−∞, ∞).

These series can be confirmed to be the Maclaurin series of the given functions by direct computation.

Answers:
Exercise 4-9
Page 181

1-5: Find the Maclaurin series for f (x) and state the radius of convergence.
1. f (x) = e−2x
2. f (x) = x2 e2x
3. f (x) = cos2 x

4. f (x) = x2 sin 4x
5. f (x) = cos x3


Sequences and Series 161

4.10 Taylor Series

Maclaurin series can be generalized by expanding in powers of (x − a)k rather than xk for some
constant a. A similar argument to that before gives the following definition.

Definition: Given function f (x) differentiable to all orders at x = a the power series in (x − a) given
by
∞
X f (k) (a) f ′′ (a) f ′′′ (a) f (k) (a)
f (x) = (x−a)k = f (a)+f ′ (a)(x−a)+ (x−a)2 + (x−a)3 +· · · (x−a)k +· · ·
k! 2! 3! k!
k=0

is the Taylor Series for the function f (x) at a (or about a or centred on a).

The Taylor series expansion at a will converge within some radius R about a, i.e. for |x − a| < R.
We note that Maclaurin series is just a special case of the Taylor series when a = 0.
Example 4-42
Find the Taylor series of the given function at the specified value and determine the interval of
convergence.

1. f (x) = cos x at a =
π
2. f (x) = e−2x at a = −1 1
2 3. f (x) = at a = 1
x

Solution:
∞
X f (n) (a)
1. Since the Taylor series about x = a is f (x) = (x − a)n we compute the following
n=0
n!
π
derivatives and evaluate them at x = a = :
2
π π
f (x) = cos x , f = cos = 0
2 2
′ π π
 
f ′ (x) = − sin x , f = − sin = −1
2
π 2
π
f ′′ (x) = − cos x , f ′′ = − cos = 0
2 2
′′′ π π
 
f ′′′ (x) = sin x , f = sin = 1
2
π 2
π
f (4) (x) = cos x , f (4)
= cos = 0
2 2
.. ..
. . (pattern repeats)
Thus the Taylor series is given by:
π π  π  1 ′′  π   π 2 1 π  π 3
f (x) = f + f′ x− + f x− + f ′′′ x−
2 2 2 2! 2 2 3! 2 2
1 (4)  π   π 4
+ f x− + ...
4! 2 2
 π  1  π 3 1  π 5
=⇒ cos x = (−1) x − + (1) x − + (−1) x − + ...
2 3! 2 5! 2
 π  1  π 3
 1  π 5

=⇒ cos x = − x − + x− − x− + ...
2 3! 2 5! 2
∞
X (−1)n+1  π 2n+1
=⇒ cos x = x−
n=0
(2n + 1)! 2
162 4.10 Taylor Series

To find the interval of convergence we will use the Ratio Test.

(−1)n+2 (x− π
2)
2n+3

an+1 (2n+3)! 1  π 2
lim = lim = lim x−
(−1)n+1 (x− π
2)
n→∞ (2n + 3)(2n + 2) 2
2n+1
n→∞ an n→∞
(2n+1)!
 π 2
=0· x− = 0 < 1 for all x
2
Thus the interval of convergence is I = (−∞, ∞).

2. We compute the following derivatives and evaluate them at x = a = −1:

f (x) = e−2x , f (−1) = e2

f ′ (x) = −2e−2x , f ′ (−1) = −2e2
f ′′ (x) = 4e−2x , f ′′ (−1) = 4e2
f ′′′ (x) = −8e−2x , f ′′′ (−1) = −8e2

Therefore the Taylor series is

1 ′′ 2 1 3
f (x) = f (−1) + f ′ (−1) (x − (−1)) + f (−1) (x − (−1)) + f ′′′ (−1) (x − (−1)) + . . .
2! 3!
1 1
=⇒ e−2x = e2 + −2e2 (x + 1) + 4e2 (x + 1)2 + −8e2 (x + 1)3 + . . .




2! 3!
22 23
 
=⇒ e−2x = e2 1 − 2(x + 1) + (x + 1)2 − (x + 1)3 + . . .
2! 3!
∞
X 2n
=⇒ e−2x = e2 (−1)n (x + 1)n .
n=0
n!

Alternatively, if only the sigma form of the series is desired, one can substitute the result
f (n) (−1) = (−1)n (2)n e2 obtained from analyzing the pattern of the derivative evaluation
directly into the Taylor series formula.
To find the interval of convergence, we will use the Ratio Test.

(−1)n+1 2n+1 (x+1)n+1

an+1 (n+1)! 2
lim = lim = lim |x + 1|
n→∞ an n→∞ (−1)n 2n (x+1)n n→∞ n+1
n!

= 0 · |x + 1| = 0 < 1 for all x

Therefore the interval of convergence is I = (−∞, ∞) .

3. We compute the following derivatives and evaluate them at x = a = 1:

1 1
f (x) = , f (1) = =1
x 1
1
f ′ (x) = − 2 , f ′ (1) = −1
x
2
f ′′ (x) = 3 , f ′′ (1) = 2
x
6
f ′′′ (x) = − 4 , f ′′′ (1) = −6
x
24
f (4) (x) = 5 , f (4) (1) = 24
x
Sequences and Series 163

Thus the Taylor series is

1 ′′ 2 1 3
f (x) = f (1) + f ′ (1) (x − 1) + f (1) (x − 1) + f ′′′ (1) (x − 1)
2! 3!
1 (4) 4
+f (1) (x − 1) + . . .
4!
1 1 1 1
=⇒ = 1 + (−1)(x − 1) + (2)(x − 1)2 − (−6)(x − 1)3 + (24)(x − 1)4 + . . .
x 2! 3! 4!
1
=⇒ = 1 − (x − 1) + (x − 1) − (x − 1) + (x − 1) + . . .
2 3 4
x
∞
1 X
=⇒ = (−1)n (x − 1)n
x n=0

Alternatively one can observe that f (n) (1) = (−1)n n! directly and place that in the Taylor
series formula. Apply the Ratio Test to find the interval of convergence.

an+1 (−1)n+1 (x − 1)n+1

lim = lim = lim |x − 1| = |x − 1|
n→∞ an n→∞ (−1)n (x − 1)n n→∞

The Taylor series converges if

|x − 1| < 1 =⇒ −1 < x − 1 < 1 =⇒ 0 < x < 2 .

At x = 0 the function x1 is undefined and the series evaluates to the divergent series 1. At
P
x = 2 the series evaluates to (−1)n which also diverges by the Term Test for Divergence.
P
Therefore the interval of convergence of the Taylor series is I = (0, 2).

Further Questions:

Find the Taylor Series of the given function at the specified value and determine the interval of
convergence.
π
1. f (x) = sin x at a = 2. f (x) = ln x at a = 1 3. f (x) = ex at a = 2
2

Assuming convergence to f (x), Taylor series gives us a mechanism for calculating trigonometric and
other functions, namely by evaluating the first n terms of the series at x. How many terms of the
series are required for a good approximation will depend on the function, the value a about which it is
expanded, and x.
Tayor series allows expansion of the function f about values other than a = 0 which is useful for
functions that are not defined at 0. Also in general fewer terms of the expansion will be required for a
good approximation if the Taylor series is generated about a value a near the x of interest. Indeed by
truncating the Taylor series at the k = 1 term
f (x) ≈ f (a) + f ′ (a)(x − a) ,
or at the k = 2 term,
f ′′ (a)
f (x) ≈ f (a) + f ′ (a)(x − a) + (x − a)2 ,
2!
we are just reproducing the linear and quadratic approximations of a function at x = a arrived at our
previous course. In general truncating at the k = nth term,
f ′′ (a) f ′′′ (a) f (n) (a)
f (x) ≈ f (a) + f ′ (a)(x − a) + (x − a)2 + (x − a)3 + · · · + (x − a)n
2! 3! n!
164 4.10 Taylor Series

yields increasingly better approximations. The right hand side is called the nth Taylor polynomial of
the function f (x) at a. The following diagram shows the first four Taylor polynomials of sin x at a = 0
(i.e. the Maclaurin series):

y n=1

Taylor polynomials (a = 0)

n=5

y = sin(x)
n=7
n=3

Here are the first four polynomials for the Taylor series of sin x at a = π/2:

y
Taylor polynomials (a = π/2)

n=0
n=4

π x
2

y = sin(x)
n=2
n=6

One observes both the greater accuracy of the higher order approximations as well as the utility of
expanding the Taylor series at a value a near the x at which you wish to approximate the function.
One useful result of Theorem 4-24 is that it justifies our original proof for the coefficients for the
Maclaurin series where (recall) we required the power series to be differentiable. Generalizing this
result to Taylor series we have the following result:
Theorem 4-25: If function f (x) is represented by the power series
∞
X
f (x) = ck (x − a)k = c0 + c1 (x − a) + c2 (x − a)2 + · · ·
k=0

on an open interval containing a then the coefficients are the Taylor series coefficients ck = f (k) (a)/k!,
′′
(i.e. f (x) = f (a) + f ′ (a)(x − a) + f 2!(a) (x − a)2 + · · · ).
Sequences and Series 165

In other words, if a function f (x) has a power series that converges to f (x) it will be the Taylor series.
The theorem does not say, however, that the Taylor series at a for a given function f will necessarily
converge to f . For instance define the continuous piecewise function:

 0 if x < −π/2


f (x) = cos x if −π/2 ≤ x ≤ π/2

0 if x > π/2


Then f will have the same Taylor series at a = 0 (i.e. Maclaurin series) as the function cos x. However
that series clearly cannot represent both functions. One must determine that the Taylor series for f at
a really does converge to the function.
Theorem 4-26: If a function f has derivatives to all orders in an interval centred on a, then
∞
X f (k) (a)
f (x) = (x − a)k
k!
k=0

will hold on the interval if and only if

lim Rn (x) = 0
n→∞

for all x in the interval, where

n
X f (k) (a)
Rn (x) = f (x) − (x − a)k .
k!
k=0

Exercise 4-10 Answers:

Page 181
1-5: Find the Taylor series for f (x) centered at the given value of a.

1. f (x) = e−x , a = ln 3
1
2. f (x) = sin πx, a =
2
3. f (x) = ln x, a = e

4. f (x) = 2x , a = 1
5. f (x) = ln(3 + x), a = 1
 166 Review Exercises

Answers: Chapter 4 Review Exercises

Page 181
1-3: Determine whether the given sequence is convergent or divergent.
2n
1. an =
3n + 5
3
 
2. an = ln
2n + 1
p
3. an = 4n2 + 5n − 2n

4-5: Determine whether the infinite sequence is increasing, decreasing or not monotonic.
n+1
4. an =
5n + 3
5. an = n2 e−6n

∞ ∞
k+1
X X  
6. Consider the series ak = ln .
k
k=1 k=1

(a) Find the limit lim ak . What does this say about the convergence of the series ak ?
P
k→∞
n
X
(b) Find the nth partial sum Sn = ak of the series.
k=1
(Hint: Expand the logarithm to get a telescoping sum.)
∞
X
(c) Evaluate lim Sn . Does the series ak converge or diverge?
n→∞
k=1

7-21: Determine whether the infinite series is convergent or divergent.

∞ ∞ ∞
X 4n X 4 X 2n
7. 4+7
12. n3 e−2n 17. (−1)n−1
n=1
n n=1 n=1
n3 + 4
∞ ∞ ∞
X 2n X 5n2 + 6 X 200 − n
8. 2 + 10
13. 18.
n=1
n n=1
n2 en n=1
n!
∞ ∞ √ ∞
X 7 + 6n X 2n + 3 X 3
9. 14. (−1) n
19. (−1)n
n=1
4n n=1
6n + 7 n=1 n2 (ln n)
4

∞  ∞ ∞
5 (−3)n (n + 2)!
X  X X
10. − 2−n 15. 20.
n=1
n(n + 1) n=1
n3 n=1
e n2
∞ ∞ ∞  n
X10 X n+3 X 3n − 5
11. 16. (−1) n
21.
n=1
4n + 3 n=1
(n + 1)(n + 2) n=1
5n + 6
Sequences and Series 167

22-25: Find the interval and the radius of convergence of the power series.
∞ ∞
X (x − 4)3n X 1
22. (−1)n 24. √ (x − 10)n
n=1
(3n + 1)! n=0
3
n
∞ ∞
X 1 X 1
23. (x − 7)n 25. (2x + 9)n
n=0
n2 n=0
n2 +3

26-27: Find the Maclaurin series for f (x) and state the radius of convergence.

26. f (x) = sin2 x

27. f (x) = xe−5x

28-29: Find the Taylor series for f (x) centered at the given value of a.

28. f (x) = e−5x , a = 2

29. f (x) = ln(x + 2), a = 0
168
Answers

Chapter 1 Review Exercises (page 7)

1 √ 5 1 3/2
1. x4 + 2 x − 2 + C 7. (2x + 5) + C
4 2x 3
2 7/2 2101
2. x + x + 2x1/2 + C 8.
7 5
√ 1 1
3. 2 sec x + C 9. +C
3 2 + cos 3x
1 9
4. 10.
12 2
15 1 2
5. 11. − √ 2 + √ 3 +C
4 (1 + x) 3 (1 + x)
1 1 7 2 6
6. − cos x3 + 2 + C 12. (x + 4) − (x + 4) + C


3 7 3
Exercise 2-1 (page 18)
3 x − 5
r
1. f (x) =
−1
2
x
2. f (x) =
−1
3x − 2
3. f (x) = 2 x2 + 3
−1 1


√
4. f −1 (x) = x1/7 − 2
3

√
5. (b) f −1 (x) = 5 x − 4, Domain=(−∞, ∞), Range=(−∞, ∞)
3−x
6. (b) f −1 (x) = , Domain=R − 12 , Range=R − − 12
 
2x − 1
2
1 − 2x

7. (b) f −1 (x) = , Domain= 12 , 1 , Range=[0, ∞)


x−1
√ √
Note that for the domain when solving x = f (y) for y you get x = 1−2y
y−1 . Since x ≥ 0 this
implies y−1 ≥ 0. Exchanging variables means x−1 ≥ 0 which restricts the domain of f −1 .
1−2y 1−2x

′
8. f −1 (7) = 16

′
9. f −1 (2) = 31

′
10. f −1 (11) = 12 13

11. (a) h(x1 ) = h(x2 ) =⇒ f (g(x1 )) = f (g(x2 )) =⇒ f −1 [f (g(x1 ))] = f −1 [f (g(x2 ))] =⇒
g(x1 ) = g(x2 ) =⇒ x1 = x2
(b) h−1 (x) = g −1 f −1 (x) so h−1 = g −1 ◦ f −1

169
170 Answers

Exercise 2-2 (page 25)

1. ∞
2. 0
3. 1

4. f ′ (x) = 4x3 − 2x4 − 2 e−2x

2
5. g ′ (t) = −6e−3t −
t3
2ex
6. y ′ = 2
(ex + 3)
7. f ′ (x) = − 2e2x + 1 sin e2x + x

8. g ′ (x) = ex sec2 (ex ) + sec2 x etan x

−1/2 √
9. y ′ = x + e2x x2 + e2x cos x2 + e2x

4
10. f ′ (x) = 2
(ex + 3e−x )
11. y = 2x − 1
e 2e − 1
12. y = − x+
e−1 e−1
13. 1 4x
4e + 2ex − 12 e−2x + C
1 x2
14. 2e +C
15. − cos (ex ) + C
1 3x
16. e + 2ex − e−x + C
3
1 1
17. −
9 3 (e + 2)
3

18. 2e2 − 2e
19. tan x + etan x + C
1
20. − +C
2 (e2x + 1)

Exercise 2-3 (page 38)

1. 4 [log5 (x + 1) − log5 (2x + 3)]

1 1
2. 2 ln(x + 1) + ln(x + 4) − ln(x + 2)
2 3
1
3. 2x + 3 ln(2x + 1) − ln(ex + 1)
2
2x
4. ln √
(ex + 2)3 x + 4
Chapter 2 Exercises 171

3 √
x2 + 1 3 x + 4
5. log10 √
4
3x2 + 5

2/ ln 3
x3 x3 + 2
6. ln √
3x + 1
√
7. x = ± e4 − 3
8. x = ln 3, x = ln 2
9. x = 1

10. x = ln 4
ln 5
11. x =
1 + ln 2
√
12. (a) Domain=(−∞, ∞), Range= 2, ∞

x2 − 2 √
(b) f −1 (x) = ln , Domain= 2, ∞


3
13. (a) Domain= − 3 , ∞ , Range=(−∞, ∞)
2

ex − 2
(b) f −1 (x) = , Domain=(−∞, ∞)
3
14. (a) Domain=(−∞, ∞), Range= − 32 , 1

3x + 2
 
(b) f (x) = ln
−1
, Domain= − 23 , 1


1−x
15. (a) Domain=R − {e−2 }, Range=R − {1}
1−2x
(b) f −1 (x) = e x−1 , Domain=R − {1}
2x + ex
16. f ′ (x) =
x2 + ex
cos x + 3
17. y ′ =
(ln 4) (sin x + 3x)
10t+2
18. g ′ (t) = 10t+2 (ln 10) ln (ln t + 5) +
t (ln t + 5)
2 x 3e3x
19. f ′ (x) = + 2 + 3x
x x +3 e +1
√
e2x x2 + 5 1
 
x
20. y ′ = √ 2 + −
3
x+1 x2 + 5 3(x + 1)
" #
x ln x ln x + e (2x + ex ) ln x
2 x


21. f (x) = x + e
′ 2
+


x x2 + ex

sin x
 
22. y ′ = xsin x cos x ln x +
x
p
23. − e−2x + 3 + C
24. ln |ex + 5| + C
172 Answers

11
25. ln
9
1
26. − 2 +C
(2 + ln x)
2
5x +x
27. +C
ln 5
180
28.
ln 10
1
29. − ln |3 + cos 2x| + C
2
1
30. − 3 +C
(ln x)
2
(ln x)
31. +C
2 ln 4
ln x ln a
32. (a) aln x = eln a = e(ln a)(ln x) = eln x = xln a
 

1
Z Z
(b) 5ln x dx = xln 5 dx = x1+ln 5 + C (Integration Power Rule with n = ln 5)
1 + ln 5
(c) u = ln x =⇒ du = dxx . But solving gives x = e so dx = e du.
u u

Z Z Z Z Z
5ln x dx = 5u eu du = [eln 5 ]u eu du = eu ln 5 eu du = e(1+ln 5)u du

1+ln 5
ew e(1+ln 5)u e(1+ln 5) ln x eln x
Z
dw
= ew = +C = +C = +C = +C
1 + ln 5 1 + ln 5 1 + ln 5 1 + ln 5 1 + ln 5
1
= x1+ln 5 + C
1 + ln 5
(d) Since 5ln x = xln 5 it is a power function because the base is the variable x and the constant
n = ln 5 is in the exponent.

Exercise 2-4 (page 44)

ln 2
1. (a) k =
5
(b) P (20) = 3200 bacteria

2. P (2) = 225 deer

3. (a) k = 0.02469
(b) P (9) = 73465
ln 2
4. (a) k = − = −0.07334890
9.45
(b) t = 18.9 minutes
5. k = −0.0743811, m(24) = 67.10 mg

6. The half-life is 2.31 days.

Chapter 2 Exercises 173

Exercise 2-5 (page 53)

π
1.
4
π
2.
4
√
3
3.
2
√
1 − x2
4.
x
ex
5. f ′ (x) = q
2
1 − (ex + 2)

1
6. y ′ = − p
x 1 − (ln x + 5)2
cos x
7. g ′ (x) =
1 + sin2 x
" #
ln sin −1
ln


ln x x x
8. y ′ = sin−1 x +


√
x sin−1 x 1 − x2

y 1 + y2


9. y = −
′
1 + (x + ey ) (1 + y 2 )
2x − √ 1
1−x2
10. y ′ =
√ 1
− 2y
1−y 2

1
11. f ′ (x) = √
x 4x2 − 1
!
1 2 x
r
12. √ tan −1
e +C
6 3

2 ln x
 
13. √ tan−1 √ +C
5 5
3
14. sin−1 x2 + C


2
3
 
x
15. ln x + 5 + √ tan−1
2
+C


√
5 5
1
2
16. sin−1 x + C
2
17. sin−1 (tan x) + C
√ 
√ x
18. 2 tan−1 √ +C
2
19. sec−1 (ex ) + C
π
20.
6
174 Answers

Exercise 2-6 (page 60)

7 5. 0 9. −∞ 13. e−10
1.
5
3 6. 3 10. −1 14. 1
2.
5
3. 0 7. 0 11. ∞ 15. 1
1 1
4. 8. 0 12. ln 16. 1
5 2

Chapter 2 Review Exercises (page 61)

1 1 1 1
1. (a) f (x1 ) = f (x2 ) =⇒ 3 + = 3 + 3 =⇒ 3 = 3 =⇒ x31 = x32 =⇒ x1 = x2
x31 x2 x1 x2
1
r
(b) g(x) = f −1
(x) = 3

x−3
1 1 1 1
(c) g ′ (11) = = ′ = =−
f ′ (g(11)) f (1/2) −3(1/2)−4 48
2 e3x + 4 − 3xe3x (2 ln x + 5)


2. f ′ (x) = 2
x (e3x + 4)
4t3 + 2e2t
3. f ′ (t) =
t4 + e2t + 1
1 1 2
 
4. g ′ (x) = −
3 x + 1 2x + 4
5. F ′ (0) = 2
2 2 2
6. f ′′ (x) = 2e−x − 10x2 e−x + 4x4 e−x
8x
 
4x
7. y = (2x + 3)
′
4 ln(2x + 3) +
2x + 3
t
8. f ′ (t) = (ln 10)(et )10e
h cos x i
9. y ′ = (ln x)cos x − sin x ln(ln x) +
x ln x
xy 2 exy − y
10. y ′ =
x − x2 yexy
xex + 1
11. f ′ (x) = 2
x + x (ex + ln x)
sec2 t
12. g ′ (t) = p
1 − (tan t + 3)2
1 1
13. h′ (x) = − √
ln 10 (cos−1 x + 1) 1 − x2
1 4√x
14. e +C
2
1 1
15. (2x + 1)2 − (2x + 1) + ln |2x + 1| + C
4 2
Chapter 3 Exercises 175

4x e2x
16. +C
ln (4e2 )
1
17. [4 (4x ) + sin 4x ] + C
ln 4
1 h −1 2
π i
18. tan e −
2 4
sin x
 
19. sin−1 √ +C
3
20. ln | sec(ln x)| + C
p x
21. − 16 − x2 + 2 sin−1 +C
4
1
22. sec−1 x2 + C


3
π
23.
4
√   √
3 π 3 π

24. , , − ,−
2 3 2 3
25. ∞
26. 1
27. 1
28. ∞

29. e
30. e2
5
31. ln
2
32. 64000 bacteria
33. λ = −k = − ln(3/4) 1
hr =⇒ t = 1
k ln(30/400) ≈ 9.0039 hours

Exercise 3-1 (page 67)

1 2 2 −5x
1. − x2 e−5x − xe−5x − e +C
5 25 125
1 1
1/2 1
3/2
2. x3 cos−1 x − 1 − x2 + 1 − x2 +C
3 3 9
√ √ √ 1 √
3. te2 t − t e2 t + e2 t + C
2
1 11 1 11
4. x ln x − x +C
11 121
1
1
5. − x3 cos x3 + sin x3 + C


3 3
1 1 2x
6. − e2x cos 4x + e sin 4x + C
5 10
176 Answers

√
3π π 1
7. − − ln 2
3 4 2
1 3
8. x sin(3 ln x) − x cos(3 ln x) + C
10 10
1 2 x 2 2
9. x 5 − x 5x + 5x + C
ln 5 (ln 5)2 (ln 5)3
√
3π
10. − ln 2
3
Exercise 3-2 (page 73)

1 2 1
1. − cos11 x + cos13 x − cos15 x + C
11 13 15
1 1
2. sin5 x − sin7 x + C
5 7
3 1 1
3. x − sin 2x + sin 4x + C
8 4 32
1 1
4. tan5 x + tan7 x + C
5 7
1 2 1
5. sec9 x − sec7 x + sec5 x + C
9 7 5
1 1
6. − cot3 x − cot5 x + C
3 5
1
7. tan3 x − tan x + x + C
3
1 1
8. − cos 2x − cos 16x + C
4 32
1 1
9. sin 2x − sin 8x + C
4 16
1 1
10. sin x + sin 9x + C
2 18
Exercise 3-3 (page 77)
x 1 p
1. 8 sin−1 +x 16 − x2 + C
4 2
p p
2. 9 + x2 + 3 ln | 9 + x2 − 3| − 3 ln |x| + C
√ √ !3
x2 − 1 1 x2 − 1
3. − +C
x 3 x
" 3 #
1 1

u u
4. √ + √ +C
256 16 − u2 3 16 − u2
1 p
5. ln 2x + 4x2 − 9 + C
2
p x
6. ln 4 + x2 + x − √ +C
4 + x2
Chapter 3 Exercises 177

r !
5 3 1 p
7. √ sin −1
x − x 5 − 3x2 + C
6 3 5 6

1 x−2
 
8. √ tan−1 √ +C
2 2
3
1 1 x−3 1 x−3


−3/2
9. − x2 − 6x + 15 + √ − √ +C
3 12 x − 6x + 15 36
2 x − 6x + 15
2

4−x
10. √ +C
2 4 − 4x − x2
Exercise 3-4 (page 88)

1. −2 ln |x − 2| + 3 ln |x − 3| + C
1 2 13 14
2. x + ln |x + 3| + ln |x − 3| + C
2 3 3
21 2 1 7
3. ln |x − 1| − − ln |x + 2| + C
27 3x−1 9
1
4. ln |x + 1| − ln |x2 + 1| + 3 tan−1 x + C
2
√
1 3 1 3 x 1 x
5. − ln |x| + ln |x + 3| −
2
+ tan−1 √ + +C
2 2 x2 + 3 18 3 6 x2 + 3
x2 6 3 1
6. + 3x + 7 ln |x − 1| − − +C
2 x − 1 2 (x − 1)2
2 2
 
x
7. − − √ tan −1
√ +C
x 5 5
x3 2 1 11
 
x
8. − 3x + + ln |x2 + 3| + √ tan−1 √ +C
3 x 2 3 3
1 1 1
 
x
9. − ln |x2 + 3| − √ tan−1 √ + ln |x2 + 1| + 2 tan−1 x + C
2 3 3 2
1
10. ln |x − 1| + ln |x2 + 1| − 3 tan−1 x + C
2
Exercise 3-5 (page 91)
√
1. 2 e +C x

√ √ √
2. −2 x + 5 cos x + 5 + 2 sin x + 5 + C
3. 3 + ex − 3 ln |3 + ex | + C
1
 3x 
e
4. √ tan−1 √ +C
3 5 5
5. 2 − 6 ln 5 + 6 ln 4
6 4
6. ln | sin x − 3| + ln | sin x + 2| + C
5 5
1 2
7. ln |x2 + 1| − 2 +C
2 x +1
178 Answers

√
e2x + 4ex + 6 ex + 2
8. ln √ + √ +C
2 2

ln 6 1 √ 1 1 1
 
9. + √ tan−1 2 − ln 3 − √ tan−1 √
2 2 2 2 2
√
2 8 x+1 1 x+1
 
10. − − tan−1 √ − +C
(x + 1)2 + 8 32 8 4 (x + 1)2 + 8

Exercise 3-6 (page 99)

1. convergent 6. convergent
2. divergent 7. convergent

3. convergent 8. divergent
4. convergent 9. divergent
5. convergent 10. divergent

Chapter 3 Review Exercises (page 100)

1 2 1 1
1. x tan−1 x − x + tan−1 x + C
2 2 2
1
2. sec3 x + C
3
3. 6 ln 2 − 2
" 3 #
1 1

x x
4. √ − √ +C
1296 x2 + 36 3 x2 + 36
1 1 3 1
5. − + +C
3 (x + 1)3 4 (x + 1)4
6. 2 ln |x| − ln |x2 + 2| + C
√
x2 + 6x + 12 + x + 3
7. ln √ +C
3
1 3
8. + ln
2 2
9. −5 ln |x − 1| + 6 ln |x − 2| + 3 ln |x − 3| + C
7 x
10. − ln |x2 + 1| + 3 tan−1 x + 4 ln |x2 + 4| − tan−1 +C
2 2
2 7/2 4 7/2
11. x ln x − x +C
7 49
p
12. ln | 1 + sin2 x + sin x| + C
13. convergent
14. convergent
15. convergent
Chapter 4 Exercises 179

Exercise 4-1 (page 110)

4 20 54 112 1
1. a1 = , a2 = , a3 = , a4 = , lim an =
7 21 59 133 n→∞ 2
2e 2e2 2e3 2e4 2
2. a1 = , a2 = 2 , a3 = 3 , a4 = 4 , lim an =
3e + 1 3e + 1 3e + 1 3e + 1 n→∞ 3
4 7 12 19
3. a1 = − , a2 = , a3 = − , a4 = , lim an = 0
7 16 37 76 n→∞
1
4. lim an = , convergent
n→∞ 2
5. lim an = 0, convergent
n→∞

6. lim an = 0, convergent
n→∞

7. lim an = 0, convergent
n→∞

8. decreasing

9. increasing
10. increasing

Exercise 4-2 (page 117)

1. convergent, Sum=6 7. convergent, Sum=6

2. divergent
8. divergent
3. divergent
9. divergent
4. convergent, Sum=20
3 10. divergent
5. convergent, Sum=
4
6. divergent 11. convergent, Sum=sin(1)

Exercise 4-3 (page 128)

1. divergent 6. convergent

2. convergent 7. convergent
3. divergent 8. divergent
4. divergent 9. convergent
5. convergent 10. divergent

Exercise 4-4 (page 131)

1. convergent 4. divergent
2. convergent 5. convergent
3. divergent 6. divergent
180 Answers

7. convergent 9. convergent

8. divergent 10. convergent

Exercise 4-5 (page 137)

1. divergent 7. absolutely convergent

2. convergent 8. absolutely convergent
3. convergent 9. absolutely convergent

4. divergent 10. divergent

5. convergent 11. absolutely convergent
6. convergent 12. absolutely convergent

Exercise 4-6 (page 142)

1. divergent 8. divergent

2. conditionally convergent 9. absolutely convergent

3. absolutely convergent 10. absolutely convergent
4. divergent 11. divergent
5. divergent 12. absolutely convergent

6. divergent 13. conditionally convergent

7. convergent 14. absolutely convergent

Exercise 4-7 (page 150)

1. I = [−1, 1), R = 1
2. I = [0, 2], R = 1
3. I = (−∞, ∞), R = ∞
1 3
 
4. I = − , 1 , R =
2 4
5. I = {1}, R = 0

Exercise 4-8 (page 156)

∞
X
1. (−1)n x2n , I = (−1, 1)
n=0
∞
4n n+1 5 5
X  
2. x , I= − ,
n=0
5n+1 4 4
∞
X 1  √ √ 
3. (−1)n x 2n+1
, I = − 2, 2
n=0
2n+1
Chapter 4 Exercises 181

∞
1X n3n 2 2
 
4. − (−1)n n xn−1 , I = − ,
6 n=0 2 3 3
∞
2n+1 3 3
X  
5. (−1) n
xn+1 + ln 3, I = − ,
n=0
(n + 1)3 n+1 2 2

Exercise 4-9 (page 160)

∞
X 2n n
1. (−1)n x , R=∞
n=0
n!
∞
X 2n n+2
2. x , R=∞
n=0
n!
∞
1 1X 22n 2n
3. + (−1)n x , R=∞
2 2 n=0 (2n)!
∞
X 42n+1 2n+3
4. (−1)n x , R=∞
n=0
(2n + 1)!
∞
X 1
5. (−1)n x6n , R = ∞
n=0
(2n)!

Exercise 4-10 (page 165)

1 1 1 2 1 3
1. − (x − ln 3) + (x − ln 3) − (x − ln 3) + . . .
3 3 6 18
2 4
π2 1 π4 1
 
2. 1 − x− + x− + ...
2! 2 4! 2
1 1 2 1 3
3. 1 + (x − e) − 2 (x − e) + 3 (x − e) + . . .
e 2e 3e
2 2 1 3 3
4. 2 + 2 ln 2 (x − 1) + (ln 2) (x − 1) + (ln 2) (x − 1) + . . .
3
1 1 2 1 3
5. ln 4 + (x − 1) − (x − 1) + (x − 1) + . . .
4 32 192
Chapter 4 Review Exercises (page 166)

1. convergent
2. divergent
3. convergent

4. decreasing
5. decreasing
k+1 k+1
   
6. (a) lim ak = lim ln = ln lim = ln(1) = 0 . Since the terms converge to
k→∞ k→∞ k k→∞ k
zero this tells us nothing about the convergence of the series ak .
P
182 Answers

n n
k+1
X   X
(b) Sn = ln = [ln(k + 1) − ln(k)]
k
k=1 k=1

= [  − ln(1)] + [ln(3)
ln(2)  ln(2)]
 −
  + [ln(4)
 ln(3)]
 −
  + . . . + [ln(n + 1) − ln(n)]

 = ln(n + 1)
| {z }
=0
(c) lim Sn = lim ln(n + 1) = ∞. Series diverges.
n→∞ n→∞

7. convergent 15. divergent

8. divergent 16. convergent
9. divergent
17. divergent
10. convergent
18. convergent
11. divergent
19. convergent
12. convergent
13. convergent 20. convergent

14. convergent 21. convergent

22. I = (−∞, ∞) , R = ∞
23. I = [6, 8] , R = 1
24. I = [9, 11) , R = 1
1
25. I = [−5, −4] , R =
2
∞
1 1X 22n 2n
26. − (−1)n x , R=∞
2 2 n=0 (2n)!
∞
X 5n n+1
27. (−1)n x , R=∞
n=0
n!
∞
X 5n e−10
28. (−1)n (x − 2)n
n=0
n!

1 1 1 3
29. ln 2 + x − x2 + x + ...
2 8 24
Summary 183

Table of Derivatives

d d n
1. (c) = 0 2. (x ) = nxn−1
dx dx

d x d 1
3. (e ) = ex 5. (ln x) =
dx dx x
d x d 1
4. (a ) = ax ln a 6. (loga x) =
dx dx x ln a

d d
7. (sin x) = cos x 10. (sec x) = sec x tan x
dx dx
d d
8. (cos x) = − sin x 11. (csc x) = − csc x cot x
dx dx
d d
9. (tan x) = sec2 x 12. (cot x) = − csc2 x
dx dx

d 1 d 1
13. sin−1 x = √ 16. sec−1 x = √



dx 1 − x2 dx x x2 − 1
d 1 d 1
14. cos−1 x = − √ 17. csc−1 x = − √



dx 1 − x2 dx x x2 − 1
d 1 d 1
15. tan−1 x = 18. cot−1 x = −



dx 1 + x2 dx 1 + x2

d df
19. (cf ) = c
dx dx
d df dg
20. (f ± g) = ±
dx dx dx
d df dg
21. (f g) = g+f (Product Rule)
dx dx dx
df
g − f dg
 
d f
22. = dx 2 dx (Quotient Rule)
dx g g
d n n−1 ′
23. [f (x)] = n [f (x)] f (x) (Generalized Power Rule)
dx
d
24. f (g(x)) = f ′ (g(x)) · g ′ (x) (Chain Rule)
dx

Here c, n, and a > 0 are constants, f and g are functions of x, and primes (f ′ , g ′ ) denote differentiation.
184 Summary

Table of Indefinite Integrals

1 1
Z Z
1. xn dx = xn+1 + C (n ̸= −1) 2. dx = ln |x| + C
n+1 x

1 x
Z Z
3. ex dx = ex + C 4. ax dx = a +C
ln a

Z Z
5. sin x dx = − cos x + C 8. sec x dx = ln | sec x + tan x| + C
Z Z
6. cos x dx = sin x + C 9. csc x dx = ln | csc x − cot x| + C
Z Z
7. tan x dx = ln | sec x| + C 10. cot x dx = ln | sin x| + C

Z Z
11. sec2 x dx = tan x + C 13. csc2 x dx = − cot x + C
Z Z
12. sec x tan x dx = sec x + C 14. csc x cot x dx = − csc x + C

1 1
Z Z
x
15. √ dx = sin−1 x + C 18. √ dx = sin−1 + C
1 − x2 a2 −x2 a
1 1 1
Z Z
x
16. dx = tan−1 x + C 19. dx = tan−1 + C
1 + x2 a 2 + x2 a a
1 1 1
Z Z
x
17. √ dx = sec−1 x + C 20. √ dx = sec−1 + C
x x2 − 1 x x2 − a2 a a

Z Z
21. cf (x) dx = c f (x) dx
Z Z Z
22. [f (x) ± g(x)] dx = f (x) dx ± g(x) dx
Z Z
23. f (g(x))g (x) dx =
′
f (u) du where u = g(x) (Substitution Rule)
Z Z
24. u dv = uv − v du (Integration by Parts)

Here c, n, and a > 0 are constants, f and g are functions of x, and primes (g ′ ) denote differentiation.
Index

absolute convergence, 132 Integral Test, 118, 138

Alternating Series Test, 129, 138 integral(s)
antiderivative, 3 definite, 2
area, 2 improper, 92
indefinite, 3
Basic Comparison Test, 123, 138 integrand, 4
integration, 4
convergent, 93 by partial fraction decomposition, 78
absolutely, 132 by a rationalizing substitution, 87
conditionally, 132 by parts, 64
by substitution, 4
decay constant, 42
general strategies, 89
dimensional analysis, 51
trigonometric, 68
divergent, 93
trigonometric substitution, 74
Euler’s Number, 21 interval of convergence, 147
exponential function(s), 19 inverse cosine function, 47
derivative of, 33 definition of, 47
limits at infinity of, 20 derivative of, 48
properties of, 19 inverse function(s), 10
simplifying, 24 definition of, 10
integral of, 23 derivative of, 15
exponential growth, 40 graphs of, 15
inverse sine function, 45
factorial, 157 definition of, 45
Fundamental Theorem of Calculus, 3 derivative of, 46
inverse tangent function, 48
half-life, 42 definition of, 48
Horizontal Line Test, 10 derivative of, 49
inverse trigonometric functions, 45
improper integrals of the first kind, 92 table of, 49
definition of, 93 invertible functions, 18
improper integrals of the second kind, 95 irreducible, 79
definition of, 95
indefinite integrals l’Hôpital’s Rule, 54
table of, 3 law of natural decay, 40
indeterminate forms of limits, 54 law of natural growth, 40
of type 0 · ∞, 56 Limit Comparison Test, 125, 138
of type 00 , 58 logarithmic differentiation, 34
of type 1∞ , 58 steps , 35
of type 00 , 54 logarithmic function(s), 26
of type ∞∞
, 54 converting to the natural logarithm, 31
of type ∞ − ∞, 56 derivatives using arbitrary bases, 33
of type ∞0 , 58 of base a, 26

185
186 INDEX

properties of, 27 p-, 121, 123, 138

simplifying, 27 absolutely convergent, 132
alternating, 129
Maclaurin Series, 157 changing index, 154
table of, 159 conditionally convergent, 132
convergent, 111
natural exponential function, 21 divergent, 111
properties of, 28 estimating the series sum, 121
derivative of, 22 geometric, 112, 123, 138
natural logarithmic function, 28 harmonic, 114, 123, 138
properties of, 28 infinite, 111
derivative of, 32 Maclaurin, 157
partial sum of, 111
one-to-one function, 10
power, 143
partial fraction decomposition, 78 sum of, 111
steps of, 80 Taylor, 161
partial fractions, 78 telescoping, 112, 123, 138
power function(s), 19 Substitution Rule, 4
power series, 143 for definite integrals, 4
centred on a, 143 for indefinite integrals, 4
coefficients of, 143
Taylor polynomial, 164
differentiation of, 152
Taylor series, 161
in x, 143
Term Test for Divergence, 114, 138
integration of, 153
terms
interval of convergence, 147
of a sequence, 102
radius of convergence, 146
Test for convergence and divergence of a series
representing functions with, 151
Alternating Series Test, 129, 138
radius of convergence, 146 Basic Comparison Test, 123, 138
Ratio Test, 134, 138 Integral Test, 119, 138
rearrangement of a series, 137 Limit Comparison Test, 125
reciprocal, 12 Ratio Test, 134, 138
reducible, 79 Root Test, 135, 138
relative growth rate, 40 Limit Comparison Test, 138
remainder estimates trigonometric integrals, 68
for the Alternating Series, 131 strategies for evaluating, 68, 69, 71
for the Basic Comparison Test, 125 trigonometric substitution, 74
for the Integral Test, 121 table of, 74
Riemann Rearrangement Theorem, 137
Root Test, 135, 138

separation of variables, 40
sequence, 102
bounded, 109
bounded above, 109
bounded below, 109
convergence of, 103
divergence of, 103
monotonic, 107
Fibonacci, 103
finite, 102
infinite, 102
series, 111
 GNU Free Documentation License 2. VERBATIM COPYING

You may copy and distribute the Document in any medium, either commer-
Version 1.3, 3 November 2008 cially or noncommercially, provided that this License, the copyright notices, and
Copyright c 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc. the license notice saying this License applies to the Document are reproduced
in all copies, and that you add no other conditions whatsoever to those of this
License. You may not use technical measures to obstruct or control the reading
<http://fsf.org/> or further copying of the copies you make or distribute. However, you may accept
compensation in exchange for copies. If you distribute a large enough number of
copies you must also follow the conditions in section 3.
Everyone is permitted to copy and distribute verbatim copies of this license You may also lend copies, under the same conditions stated above, and you
document, but changing it is not allowed. may publicly display copies.

Preamble
3. COPYING IN QUANTITY
The purpose of this License is to make a manual, textbook, or other func-
tional and useful document “free” in the sense of freedom: to assure everyone the
effective freedom to copy and redistribute it, with or without modifying it, either If you publish printed copies (or copies in media that commonly have printed
commercially or noncommercially. Secondarily, this License preserves for the au- covers) of the Document, numbering more than 100, and the Document’s license
thor and publisher a way to get credit for their work, while not being considered notice requires Cover Texts, you must enclose the copies in covers that carry,
responsible for modifications made by others. clearly and legibly, all these Cover Texts: Front-Cover Texts on the front cover,
This License is a kind of “copyleft”, which means that derivative works of the and Back-Cover Texts on the back cover. Both covers must also clearly and
document must themselves be free in the same sense. It complements the GNU legibly identify you as the publisher of these copies. The front cover must present
General Public License, which is a copyleft license designed for free software. the full title with all words of the title equally prominent and visible. You may
We have designed this License in order to use it for manuals for free software, add other material on the covers in addition. Copying with changes limited to
because free software needs free documentation: a free program should come with the covers, as long as they preserve the title of the Document and satisfy these
manuals providing the same freedoms that the software does. But this License is conditions, can be treated as verbatim copying in other respects.
not limited to software manuals; it can be used for any textual work, regardless If the required texts for either cover are too voluminous to fit legibly, you
of subject matter or whether it is published as a printed book. We recommend should put the first ones listed (as many as fit reasonably) on the actual cover,
this License principally for works whose purpose is instruction or reference. and continue the rest onto adjacent pages.
If you publish or distribute Opaque copies of the Document numbering more
than 100, you must either include a machine-readable Transparent copy along
1. APPLICABILITY AND DEFINITIONS with each Opaque copy, or state in or with each Opaque copy a computer-network
location from which the general network-using public has access to download using
public-standard network protocols a complete Transparent copy of the Document,
This License applies to any manual or other work, in any medium, that con- free of added material. If you use the latter option, you must take reasonably
tains a notice placed by the copyright holder saying it can be distributed under prudent steps, when you begin distribution of Opaque copies in quantity, to ensure
the terms of this License. Such a notice grants a world-wide, royalty-free license, that this Transparent copy will remain thus accessible at the stated location until
unlimited in duration, to use that work under the conditions stated herein. The at least one year after the last time you distribute an Opaque copy (directly or
“Document”, below, refers to any such manual or work. Any member of the pub- through your agents or retailers) of that edition to the public.
lic is a licensee, and is addressed as “you”. You accept the license if you copy, It is requested, but not required, that you contact the authors of the Docu-
modify or distribute the work in a way requiring permission under copyright law. ment well before redistributing any large number of copies, to give them a chance
A “Modified Version” of the Document means any work containing the to provide you with an updated version of the Document.
Document or a portion of it, either copied verbatim, or with modifications and/or
translated into another language.
A “Secondary Section” is a named appendix or a front-matter section of the
Document that deals exclusively with the relationship of the publishers or authors
of the Document to the Document’s overall subject (or to related matters) and
4. MODIFICATIONS
contains nothing that could fall directly within that overall subject. (Thus, if
the Document is in part a textbook of mathematics, a Secondary Section may
not explain any mathematics.) The relationship could be a matter of historical You may copy and distribute a Modified Version of the Document under the
connection with the subject or with related matters, or of legal, commercial, conditions of sections 2 and 3 above, provided that you release the Modified
philosophical, ethical or political position regarding them. Version under precisely this License, with the Modified Version filling the role
The “Invariant Sections” are certain Secondary Sections whose titles are of the Document, thus licensing distribution and modification of the Modified
designated, as being those of Invariant Sections, in the notice that says that the Version to whoever possesses a copy of it. In addition, you must do these things
Document is released under this License. If a section does not fit the above in the Modified Version:
definition of Secondary then it is not allowed to be designated as Invariant. The A. Use in the Title Page (and on the covers, if any) a title distinct from that
Document may contain zero Invariant Sections. If the Document does not identify of the Document, and from those of previous versions (which should, if
any Invariant Sections then there are none. there were any, be listed in the History section of the Document). You
The “Cover Texts” are certain short passages of text that are listed, as Front- may use the same title as a previous version if the original publisher of
Cover Texts or Back-Cover Texts, in the notice that says that the Document is that version gives permission.
released under this License. A Front-Cover Text may be at most 5 words, and a B. List on the Title Page, as authors, one or more persons or entities respon-
Back-Cover Text may be at most 25 words. sible for authorship of the modifications in the Modified Version, together
A “Transparent” copy of the Document means a machine-readable copy, with at least five of the principal authors of the Document (all of its prin-
represented in a format whose specification is available to the general public, that cipal authors, if it has fewer than five), unless they release you from this
is suitable for revising the document straightforwardly with generic text editors requirement.
or (for images composed of pixels) generic paint programs or (for drawings) some C. State on the Title page the name of the publisher of the Modified Version,
widely available drawing editor, and that is suitable for input to text formatters as the publisher.
or for automatic translation to a variety of formats suitable for input to text D. Preserve all the copyright notices of the Document.
formatters. A copy made in an otherwise Transparent file format whose markup, E. Add an appropriate copyright notice for your modifications adjacent to
or absence of markup, has been arranged to thwart or discourage subsequent the other copyright notices.
modification by readers is not Transparent. An image format is not Transparent F. Include, immediately after the copyright notices, a license notice giving
if used for any substantial amount of text. A copy that is not “Transparent” is the public permission to use the Modified Version under the terms of this
called “Opaque”. License, in the form shown in the Addendum below.
Examples of suitable formats for Transparent copies include plain ASCII with- G. Preserve in that license notice the full lists of Invariant Sections and re-
out markup, Texinfo input format, LaTeX input format, SGML or XML using a quired Cover Texts given in the Document’s license notice.
publicly available DTD, and standard-conforming simple HTML, PostScript or H. Include an unaltered copy of this License.
PDF designed for human modification. Examples of transparent image formats I. Preserve the section Entitled “History”, Preserve its Title, and add to it
include PNG, XCF and JPG. Opaque formats include proprietary formats that an item stating at least the title, year, new authors, and publisher of the
can be read and edited only by proprietary word processors, SGML or XML Modified Version as given on the Title Page. If there is no section Entitled
for which the DTD and/or processing tools are not generally available, and the “History” in the Document, create one stating the title, year, authors, and
machine-generated HTML, PostScript or PDF produced by some word processors publisher of the Document as given on its Title Page, then add an item
for output purposes only. describing the Modified Version as stated in the previous sentence.
The “Title Page” means, for a printed book, the title page itself, plus such J. Preserve the network location, if any, given in the Document for public
following pages as are needed to hold, legibly, the material this License requires access to a Transparent copy of the Document, and likewise the network
to appear in the title page. For works in formats which do not have any title page locations given in the Document for previous versions it was based on.
as such, “Title Page” means the text near the most prominent appearance of the These may be placed in the “History” section. You may omit a network
work’s title, preceding the beginning of the body of the text. location for a work that was published at least four years before the Doc-
The “publisher” means any person or entity that distributes copies of the ument itself, or if the original publisher of the version it refers to gives
Document to the public. permission.
A section “Entitled XYZ” means a named subunit of the Document whose K. For any section Entitled “Acknowledgements” or “Dedications”, Preserve
title either is precisely XYZ or contains XYZ in parentheses following text that the Title of the section, and preserve in the section all the substance and
translates XYZ in another language. (Here XYZ stands for a specific sec- tone of each of the contributor acknowledgements and/or dedications given
tion name mentioned below, such as “Acknowledgements”, “Dedications”, therein.
“Endorsements”, or “History”.) To “Preserve the Title” of such a section L. Preserve all the Invariant Sections of the Document, unaltered in their text
when you modify the Document means that it remains a section “Entitled XYZ” and in their titles. Section numbers or the equivalent are not considered
according to this definition. part of the section titles.
The Document may include Warranty Disclaimers next to the notice which M. Delete any section Entitled “Endorsements”. Such a section may not be
states that this License applies to the Document. These Warranty Disclaimers included in the Modified Version.
are considered to be included by reference in this License, but only as regards N. Do not retitle any existing section to be Entitled “Endorsements” or to
disclaiming warranties: any other implication that these Warranty Disclaimers conflict in title with any Invariant Section.
may have is void and has no effect on the meaning of this License. O. Preserve any Warranty Disclaimers.
 If the Modified Version includes new front-matter sections or appendices that 9. TERMINATION
qualify as Secondary Sections and contain no material copied from the Document,
you may at your option designate some or all of these sections as invariant. To
You may not copy, modify, sublicense, or distribute the Document except as
do this, add their titles to the list of Invariant Sections in the Modified Version’s
expressly provided under this License. Any attempt otherwise to copy, modify,
license notice. These titles must be distinct from any other section titles.
sublicense, or distribute it is void, and will automatically terminate your rights
You may add a section Entitled “Endorsements”, provided it contains noth- under this License.
ing but endorsements of your Modified Version by various parties—for example, However, if you cease all violation of this License, then your license from a par-
statements of peer review or that the text has been approved by an organization ticular copyright holder is reinstated (a) provisionally, unless and until the copy-
as the authoritative definition of a standard. right holder explicitly and finally terminates your license, and (b) permanently, if
You may add a passage of up to five words as a Front-Cover Text, and a the copyright holder fails to notify you of the violation by some reasonable means
passage of up to 25 words as a Back-Cover Text, to the end of the list of Cover prior to 60 days after the cessation.
Texts in the Modified Version. Only one passage of Front-Cover Text and one Moreover, your license from a particular copyright holder is reinstated per-
of Back-Cover Text may be added by (or through arrangements made by) any manently if the copyright holder notifies you of the violation by some reasonable
one entity. If the Document already includes a cover text for the same cover, means, this is the first time you have received notice of violation of this License
previously added by you or by arrangement made by the same entity you are (for any work) from that copyright holder, and you cure the violation prior to 30
acting on behalf of, you may not add another; but you may replace the old one, days after your receipt of the notice.
on explicit permission from the previous publisher that added the old one. Termination of your rights under this section does not terminate the licenses
The author(s) and publisher(s) of the Document do not by this License give of parties who have received copies or rights from you under this License. If your
permission to use their names for publicity for or to assert or imply endorsement rights have been terminated and not permanently reinstated, receipt of a copy of
of any Modified Version. some or all of the same material does not give you any rights to use it.

10. FUTURE REVISIONS OF THIS

5. COMBINING DOCUMENTS LICENSE
The Free Software Foundation may publish new, revised versions of the GNU
You may combine the Document with other documents released under this Free Documentation License from time to time. Such new versions will be similar
License, under the terms defined in section 4 above for modified versions, provided in spirit to the present version, but may differ in detail to address new problems
that you include in the combination all of the Invariant Sections of all of the or concerns. See http://www.gnu.org/copyleft/.
original documents, unmodified, and list them all as Invariant Sections of your Each version of the License is given a distinguishing version number. If the
combined work in its license notice, and that you preserve all their Warranty Document specifies that a particular numbered version of this License “or any
Disclaimers. later version” applies to it, you have the option of following the terms and condi-
The combined work need only contain one copy of this License, and multiple tions either of that specified version or of any later version that has been published
identical Invariant Sections may be replaced with a single copy. If there are (not as a draft) by the Free Software Foundation. If the Document does not spec-
multiple Invariant Sections with the same name but different contents, make the ify a version number of this License, you may choose any version ever published
title of each such section unique by adding at the end of it, in parentheses, the (not as a draft) by the Free Software Foundation. If the Document specifies that
name of the original author or publisher of that section if known, or else a unique a proxy can decide which future versions of this License can be used, that proxy’s
number. Make the same adjustment to the section titles in the list of Invariant public statement of acceptance of a version permanently authorizes you to choose
Sections in the license notice of the combined work. that version for the Document.
In the combination, you must combine any sections Entitled “History” in
the various original documents, forming one section Entitled “History”; likewise
combine any sections Entitled “Acknowledgements”, and any sections Entitled
“Dedications”. You must delete all sections Entitled “Endorsements”.
11. RELICENSING
“Massive Multiauthor Collaboration Site” (or “MMC Site”) means any World
Wide Web server that publishes copyrightable works and also provides prominent
facilities for anybody to edit those works. A public wiki that anybody can edit is
6. COLLECTIONS OF DOCUMENTS an example of such a server. A “Massive Multiauthor Collaboration” (or “MMC”)
contained in the site means any set of copyrightable works thus published on the
MMC site.
“CC-BY-SA” means the Creative Commons Attribution-Share Alike 3.0 li-
You may make a collection consisting of the Document and other documents cense published by Creative Commons Corporation, a not-for-profit corporation
released under this License, and replace the individual copies of this License in the with a principal place of business in San Francisco, California, as well as future
various documents with a single copy that is included in the collection, provided copyleft versions of that license published by that same organization.
that you follow the rules of this License for verbatim copying of each of the “Incorporate” means to publish or republish a Document, in whole or in part,
documents in all other respects. as part of another Document.
You may extract a single document from such a collection, and distribute it An MMC is “eligible for relicensing” if it is licensed under this License, and
individually under this License, provided you insert a copy of this License into if all works that were first published under this License somewhere other than
the extracted document, and follow this License in all other respects regarding this MMC, and subsequently incorporated in whole or in part into the MMC, (1)
verbatim copying of that document. had no cover texts or invariant sections, and (2) were thus incorporated prior to
November 1, 2008.
The operator of an MMC Site may republish an MMC contained in the site
under CC-BY-SA on the same site at any time before August 1, 2009, provided
the MMC is eligible for relicensing.
7. AGGREGATION WITH INDEPENDENT
WORKS ADDENDUM: How to use this License for
your documents
A compilation of the Document or its derivatives with other separate and
independent documents or works, in or on a volume of a storage or distribution To use this License in a document you have written, include a copy of the
medium, is called an “aggregate” if the copyright resulting from the compila- License in the document and put the following copyright and license notices just
tion is not used to limit the legal rights of the compilation’s users beyond what after the title page:
the individual works permit. When the Document is included in an aggregate,
this License does not apply to the other works in the aggregate which are not
themselves derivative works of the Document.
If the Cover Text requirement of section 3 is applicable to these copies of the Copyright c YEAR YOUR NAME. Permission is granted to copy,
Document, then if the Document is less than one half of the entire aggregate, distribute and/or modify this document under the terms of the
the Document’s Cover Texts may be placed on covers that bracket the Document GNU Free Documentation License, Version 1.3 or any later ver-
within the aggregate, or the electronic equivalent of covers if the Document is in sion published by the Free Software Foundation; with no Invariant
electronic form. Otherwise they must appear on printed covers that bracket the Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy
whole aggregate. of the license is included in the section entitled “GNU Free Docu-
mentation License”.

8. TRANSLATION If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts, re-
place the “with . . . Texts.” line with this:

Translation is considered a kind of modification, so you may distribute trans-

lations of the Document under the terms of section 4. Replacing Invariant Sec- with the Invariant Sections being LIST THEIR TITLES, with the
tions with translations requires special permission from their copyright holders, Front-Cover Texts being LIST, and with the Back-Cover Texts be-
but you may include translations of some or all Invariant Sections in addition to ing LIST.
the original versions of these Invariant Sections. You may include a translation
of this License, and all the license notices in the Document, and any Warranty
Disclaimers, provided that you also include the original English version of this
License and the original versions of those notices and disclaimers. In case of a If you have Invariant Sections without Cover Texts, or some other combination
disagreement between the translation and the original version of this License or of the three, merge those two alternatives to suit the situation.
a notice or disclaimer, the original version will prevail. If your document contains nontrivial examples of program code, we recom-
If a section in the Document is Entitled “Acknowledgements”, “Dedications”, mend releasing these examples in parallel under your choice of free software li-
or “History”, the requirement (section 4) to Preserve its Title (section 1) will cense, such as the GNU General Public License, to permit their use in free soft-
typically require changing the actual title. ware.