 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of

William R. Wade, the author of this book.

Written by Gi-uk Kim
The Solutions of Analysis

Writer : Gi-uk Kim.

Belongs To : Department of Mathematics,
SOGANG University,
1 Shinsoo-dong, Mapo-gu, Seoul,
in the Republic of Korea.

No part of this document may be reproduced, in any form or by any

means, without permission in writing from this writer.
And if you find errors, I would appreciate your contacting me at the
e-mail address below.
My E-mail Address : [email protected]
My Cellular Phone Number : 010-2289-5204
These solutions might be inexact, yet do you rely on them?
Copyright this writer. All rights reserved.

Chapter 6 Infinite Series of Real Numbers

6.1 Introduction

p.158
Exercises
1. Show that
∞ n
x
∑ xk = 1 - x
k =n

for ∣x∣< 1 and n = 0, 1, ⋯.

Proof) This proof is by induction on n.
If n = 0, by Geometric Series,
∞ ∞
x 1 x0
∑ xk = x0 + ∑ xk = 1 + 1 - x = 1-x
=
1-x
.
k =0 k=1

Otherwise, suppose that the formula holds for n = t.

Then, if n = t + 1,
∞ ∞ t t +1
x x
∑ x k = ∑ xk - x t = 1 - x - xt = 1 - x .
k = t +1 k =t

That is, the formula is true for t + 1.

Thus, the formula holds for all t ∈ N.

2. Prove that each of the following series converges and find its value.

These solutions might be inexact, yet do you rely on them?

- 1 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
(-1) k + 1
(a) ∑ .
k =1 πk
∞
(-1 ) k +1 ∞ k
Solution) Since ∑
k =1 π k = (-1 ) ∑ -
k =1
1
π ( ) and ∣- π1 ∣< 1,
by Geometric Series, the series converges.
And, its value is
1
-
∞
(-1 ) k +1 ∞ k
∑
k =1 π k = (-1 ) ∑ -
k =1
1
π ( ) = (-1 )
1+
1
π
=
1
π+1
.
π
∞
(-1 ) k + 4
(b) ∑ .
k =1 5k
∞ k ∞ k
Solution) Since ∑
k =1
( - 15 ) and ∑
k =1
( 15 ) converge,

∞
(-1 ) k + 4 ∞ k ∞ k
then ∑
k =1 5k
= ∑
k =1
( - 15 ) +4 ∑
k =1
( 15 ) also converges.

And, its value is

(-1 ) k + 4 k k
(- 15 ) ( 15 )
∞ ∞ ∞
∑ =∑ +4 ∑
k =1 5k k =1 k =1

1 1
-
5 5
= + 4⋅
1 1
1+ 1-
5 5
5
=
6
∞
3k
(c) ∑ .
k=1 7k -1
∞
3k ∞ k
Solution) Since ∑
k =1 7 k -1 = 7∑
k =1
( 37 ) and ∣ 37 ∣< 1,
by Geometric Series, the series converges.
And, its value is
3
3k k
( )
∞ ∞
3 7 21
∑ = 7∑ = 7⋅ = .
k =1 7k-1 k =1 7
1-
3 4
7
∞
(d) ∑ 2 k e - k.
k=0

k
( 2e ) ∣ 2e ∣< 1,
∞ ∞
Solution) Since ∑ 2k e-k = 1 + ∑ and
k =0 k=1

by Geometric Series, the series converges.

And, its value is
2
∞ ∞ k
∑ 2 e
k=0
k -k
=1 + ∑
k =1
( ) 2
e
= 1+
1-
2
e
=
e
e-2
.
e

These solutions might be inexact, yet do you rely on them?

- 2 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

3. Represent each of the following series as a telescopic series and find its value.
∞
1
(a) ∑ k ( k + 1)
.
k =1

( ) and
∞ ∞
1 1 1
Solution) Since ∑ k ( k +1)
= ∑
k
-
k+1
k =1 k =1

{ 1k } is a convergent sequence,
by Telescopic Series,
∞ ∞
∑
k =1
1
k ( k +1)
= ∑
k =1
( 1
k
-
1
k+1 ) = 1- lim
k →∞
1
k+1
= 1.

( )
∞
k ( k +2)
(b) ∑ log .
k =2 ( k + 1 )2

( ) { log ( ) - log ( )}
∞ ∞
k ( k +2) k k+1
Solution) Since ∑ log = ∑
k +1 k+2
k =2 ( k + 1 )2 k =2

and { log ( k
k +1 ) } is a convergent sequence,
by Telescopic Series,

( )
∞ ∞
∑ log
k =2
k ( k +2)
( k + 1 )2
= ∑
k =2
{ log ( k
k+1 ) - log ( kk ++ 12 )}
lim log (
k+1 )
2 k
= log -
3 k →∞

2
= log
3
jk

( 1 - ( π4 ) ), where
∞
k π 1
(c) ∑ 4
jk = -
k ( k +1)
for k ∈ N.
k =1

jk

( 1 - ( π4 ) ) = ∑ ( ) and
∞ ∞
k π k π ( k +1) π
Solution) Since ∑ 4 4
-
4
k =1 k =1

{ k π
4 } is a convergent sequence,
by Telescopic Series,
jk

( )
∞ ∞
∑
k =1
k π
4 ( 1- ( ) )
π
4
=∑
k =1
k π
4
- ( k +1) π
4
π k π
= - lim
4 k →∞ 4
π
= -1
4

4. Find all x ∈ R for which

∞
∑ 3 ( x k - x k -1 ) ( x k + x k -1 )
k =1

converges. For each such x, find the value of this series.

These solutions might be inexact, yet do you rely on them?

- 3 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

5. Prove that each of the following series diverges.

∞
1
(a) ∑ cos .
k =1 k2
1
Proof) Let f ( x ) = cos for x ∈ ( 0, ∞).
x2
Since f is continuous on ( 0, ∞), by Theorem 3.24,

lim f ( x ) = lim cos

x →∞ x →∞
1
x2
= cos ( lim
x →∞
1
x2 )= cos 0 = 1.

1
By Sequential Characterization of Limits, lim cos = 1.
k →∞ k2
∞
1
By Divergence Test, ∑ cos diverges.
k=1 k2
k
( ).
∞
1
(b) ∑ 1-
k
k =1

x
Proof) Assume that e : = lim
x →∞
( 1+
1
x ).
x
Let f ( x ) = ( 1 - 1x ) for x ∈ (-∞, 0 ).

By Theorem 3.24,
x -x -1
lim
x → -∞
f (x ) = lim
x → -∞
( 1 - 1x ) = { lim
x → -∞
( 1 - 1x ) } =
1
e
.

k
By Sequential Characterization of Limits, lim
k →∞
( 1 - 1k ) =
1
e
.

∞ k
By Divergence Test, ∑
k =1
( 1-
1
k ) diverges.

∞
k +1
(c) ∑ .
k =1 k2
∞
k+1
Proof) Suppose to the contrary that ∑ converges.
k =1 k2
∞ ∞ ∞
k+1 1 1
Since ∑ = ∑
k
+∑ 2 , then
k =1 k2 k=1 k =1 k

∞ ∞
1 1
both ∑ k and ∑ converge, also.
k =1 k =1 k2
∞
1
By Harmonic Series, ∑ k diverges, a contradiction.
k =1

∞
6. (a) Prove that if ∑ a k converges, then its partial sums s n are bounded.
k=1

n ∞
Proof) Set s n : = ∑ a k and L : = ∑ a k .
k=1 k=1

These solutions might be inexact, yet do you rely on them?

- 4 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
Since ∑ a k converges, for given ε = 1,
k=1

there is an N ∈ N such that n ≥ N implies ∣s n - L∣< 1.

By Triangle Inequality, ∣s n∣≤ 1 + ∣L∣ for all n ≥ N.
Set M : = max { ∣s 1∣, ∣s 2∣, ⋯, ∣s N - 1∣, 1 + ∣L∣}.

Then, ∣s n∣≤ M for all n ∈ N.

∞
(b) Show that the converse of part (a) is false. Namely, show that a series ∑ ak
k =1

may have bounded partial sums and still diverge.

Showing) Set a n : = ( -1 ) n for all n ∈ N. Then,

n
sn = ∑ ak =
k =1 { -1,

0,
n ’s are odd numbers

n ’s are even numbers

Thus, ∣s n∣≤ 1 for all n ∈ N.
∞
However, by Example 6.3, ∑ a k still diverge.
k=1

7. Let { bk } be a real sequence and b ∈ R.

(a) Suppose that there is an N ∈ N such that ∣b - b k∣≤ M for all k ≥ N.
Prove that

∣ n b - ∑ b ∣≤ ∑ ∣b - b ∣ + M ( n - N )
n N
k k
k=1 k=1

for all n > N.

Proof) For n > N,

∣n b - ∑ b ∣ = ∣ ∑ ∣
n n n
k b - ∑ bk
k=1 k=1 k=1
n
≤ ∑ ∣b - b k∣
k=1
N n
= ∑ ∣b - b k∣ + ∑ ∣b - b k∣
k=1 k = N +1
N
≤ ∑ ∣b - b k∣ + M ( n - N )
k=1

(b) Prove that if b k → b as k → ∞, then

b1 + b 2 + ⋯ + b n
→ b
n
as n → ∞.
Proof) Let ε > 0 be given.
If M is replaced by ε in part (a), then

These solutions might be inexact, yet do you rely on them?

- 5 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
n n

∣ ∑ bk
k=1
n
-b ∣ ∣ =
n b - ∑ bk
n
k=1
∣
N
∑ ∣b - b k∣ ε
k=1
≤ + (n -N )
n n
Taking the limit of this inequality as n → ∞, we conclude that
n

∣ ∑ bk
k =1
n
-b ∣
≤ ε.

(c) Show that the converse of (b) is false.

Showing) Let b n = (-1 ) n. Then,
b 1 + b2 + ⋯ + bn -1
if n is odd, = → 0 as n → ∞,
n n
b 1 + b 2 + ⋯ + bn 0
and if n is even, = → 0 as n → ∞.
n n
However, { bn } does not converge.

∞
8. A series ∑ a k is said to be Cesaro summable to an L ∈ R if and only if
k=0

n -1
σn : = ∑
k=0
(1- k
n )a k

converges to L as n → ∞.
n -1
(a) Let s n = ∑ a k. Prove that
k=0

s1 + s2 + ⋯ + sn
σn =
n
for each n ∈ N.
Proof) The proof is by induction on n.
For n = 1,
0 0
σ 1 = ∑ ( 1 - k ) a k = ∑ a k = s 1. It is obvious for n = 1.
k =0 k =0

s1 + s2 + ⋯ + sn
Suppose that σ n = for some n ∈ N.
n
Then, by inductive hypothesis,

These solutions might be inexact, yet do you rely on them?

- 6 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
n
σn +1 = ∑
k=0
( 1-
k
n+1 )a k

n -1
= ∑
k=0
(1- k
n+1 ) a +(1-
k
n
n+1 )a n

n -1 n-1
n-k 1 1
= ∑ ak + ∑ a + a
k=0 n+1 k=0 n+1 k n+1 n
n -1 n -1
=
n
∑
n + 1 k=0 ( 1 - kn ) a k +
1
∑
n + 1 k=0 k
a +
1
a
n+1 n
n 1 1
= σ + s + ( s n +1 - s n )
n+1 n n +1 n n+1
s 1 + s 2 + ⋯ + s n + s n +1
=
n +1
That is, the formula is true for n + 1.
We conclude by induction that the formula holds for all n ∈ N.
∞ ∞
(b) Prove that if a k ∈ R and ∑ a k = L converges, then ∑ a k is Cesaro
k=0 k=0

summable to L .
Proof) i ) The first elucidation.
n -1
Set s n : = ∑ a k.
k=0

∞
Suppose that ∑ a k = lim s n = L converges.
k=0 n →∞

That is, there is an N 1 ∈ N such that

ε
n ≥ N 1 implies ∣s n - L∣< for every ε > 0.
2
Set M : = max { ∣s 1 - L∣, ∣s 2 - L∣, ⋯, ∣s N 1 -1 - L∣}.
By the Archimedean Principle,
2 M N1
there is an N 2 ∈ N such that N2 > .
ε
If n ≥ max { N 1, N 2 }, then
1
∣L - σ n∣ = ∣ n L - ( s 1 + s 2 + ⋯ + s n )∣
n
1
≤ { ∣s 1 - L∣+ ⋯ + ∣s N 1 -1 - L∣+ ⋯ +∣s n - L∣}
n

≤ 1
n {(N 1 - 1 ) M + ( n - N1 + 1 )
ε
2 }
( N1 - 1 ) M ( n - N1 + 1 ) ε
= +
n n 2
N1 M ε
< +
N2 2
ε ε
< + = ε
2 2

These solutions might be inexact, yet do you rely on them?

- 7 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
ii ) The second elucidation.
By part (a) and 7 (b), since s n → L as n → ∞, then
s1 + s2 + ⋯ + sn
σn = converges to L.
n
∞
1
(c) Prove that ∑ (-1 ) k is Cesaro summable to 2
; hence the converse of (b)
k =0

is false.
n -1
Proof) Set s n : = ∑ (-1 ) k.
k =0

For n = 2 k - 1 ( k ∈ N), s n = 1,
and for n = 2 k ( k ∈ N), s n = 0. Then,
s 1 + s 2 + ⋯+ s n
σn =
n
k

{
, n = 2 k -1
2 k -1
=
k
, n = 2k
2k
1
Thus, σ n → as n → ∞.
2
∞
But, ∑ (-1 ) k does not converge.
k =0

∞
(d) [Tauber]. Prove that if a k ≥ 0 for k ∈ N and ∑ a k is Cesaro summable
k=0

∞
to L, then ∑ a k = L.
k=0

n -1 n -1
Proof) Let s n = ∑ a k and σ n = ∑
k=0 k =0
(1- k
n )a.
k

∞
Suppose that ∑ a k is Cesaro summable to L.
k=0

n-1
Then lim σ n = lim ∑
n →∞ n → ∞ k =0
(1- k
n )a k exists and equals to L.

Since a k ≥ 0, { sn } and { σn } are increasing sequences.

Fix m. We consider that
m-1 n -1
∑
k =0
( 1 - kn ) a k ≤∑
k =0
( 1 - kn ) a k ≤ L for all n ≥ m.

Taking the limit of this inequality as n → ∞, we obtain

m-1
s m = ∑ a k ≤ L.
k=0

m-1 m-1
Moreover, σ m = ∑
k=0
(1- k
m )a k ≤ ∑ a k = s m.
k =0

That is, σ m ≤ s m ≤ L.

These solutions might be inexact, yet do you rely on them?

- 8 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
By the Squeeze Theorem, we have the desired result like the following.
∞
∑ a k = lim s m = L.
k=0 m→∞

9. (a) Suppose that { ak } is a decreasing sequence of real numbers. Prove that if

∞
∑ a k converges, then k a k → 0 as k → ∞.
k=0

∞
Proof) Since { ak } is decreasing and ∑ a k converges, a k ↓ 0 as k → ∞.
k=0

Thus, a k ≥ 0 for all k ∈ N.

And, it holds that
a 2n ≤ a n and a 2n ≤ a n + 1, a 2n ≤ a n + 2, ⋯⋯, a 2n ≤ a 2n.
Adding up each of both sides of the inequalities,
2n ∞
0 ≤ n a 2n ≤ ∑ a k ≤ ∑ a k.
k=n k =n

∞
Since ∑ a k converges, by Cauchy Criterion,
k=0

for given ε > 0, there is an N ∈ N such that

∣ ∣
∞
n ≥ N implies that ∑ a k < ε.
k=n

Thus, ∣n a 2n ∣< ε for all n ≥ N.

Let k = 2 n, ∣ 2k a ∣< ε for all

k k ≥ 2 N.

Hence, ∣k a k ∣< ε for all k ≥ 2 N.

n
(-1 ) k +1
(b) Let sn = ∑ for n ∈ N. Prove that s 2n is strictly increasing,
k =1 k
s 2n + 1 is strictly decreasing, and s 2n +1 - s 2n → 0 as n → ∞.
n
(-1 ) k + 1
Proof) Since s n = ∑ ,
k=1 k
2n
(-1 ) k + 1 2n + 1
(-1 ) k + 1
s 2n = ∑ and s 2n + 1 = ∑ .
k=1 k k =1 k
Then,
2n + 2
(-1 ) k +1 2n
(-1 ) k +1
s 2( n + 1) -s 2n = ∑ -∑
k =1 k k =1 k
(-1 ) 2n + 2 (-1 ) 2n +3
= +
2 n +1 2 n +2
1 1
= - > 0
2n +1 2n +2
and

These solutions might be inexact, yet do you rely on them?

- 9 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
2n + 3
(-1 ) k + 1 2n + 1
(-1 ) k + 1
s 2( n + 1) +1 -s 2n + 1 = ∑ - ∑
k=1 k k=1 k
(-1 ) 2n +3 (-1 ) 2n +4
= +
2 n+2 2 n +3
1 1
= - + < 0
2 n +2 2n+3
(-1 ) 2n +2 1
And, s 2n +1 - s 2n = = → 0 as n → ∞.
2 n +1 2n +1
(c) Prove that part (a) is false if "decreasing" is removed.
(-1 ) k
Proof) Set a k = .
k
∞
Then, ∑ a k converges,
k=0

but k a k = (-1 ) k →
/ 0 as k → ∞.

∞ ak
10. Suppose that a k ≥ 0 for k large and ∑ k
converges. Prove that
k =1

∞ a
lim ∑ j +kk = 0.
j →∞ k =1

Proof) Since a k ≥ 0 for k large, then there is an N 1 ∈ N such that

n ≥ N 1 implies a n ≥ 0.
∞ ak
Since ∑ k
converges, by Cauchy Criterion,
k =1

for given ε > 0, there is an N2 > N1 such that

∣∑ ak
∣<
∞
ε
n ≥ N 2 implies .
k =n k 2
By Archimedean Principle,
N2
2
there is an N 3 ∈ N such that N 3 >
ε
∑ ∣a k∣.
k =1

Set N = max { N 1, N 2, N 3 }.
If j ≥ N, then

∣∑ ∣ ≤∣ ∑ ∣+ ∣ ∣
N2
∞ ak ak ∞ ak
j +k j +k
∑ j +k
k=1 k =1 k = N 2 +1
N2 ∣a k∣ ∞ ak
≤ ∑ + ∑
k =1 j+k k = N 2 +1 j+k
N2
∣a k∣ ∞ ak
< ∑ j
+ ∑ k
k =1 k = N 2 +1

N3 ε ε
< +
2N 2
≤ ε

These solutions might be inexact, yet do you rely on them?

- 10 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

 A FEW CRUXES MADE BY THE WRITER AT FIRSTHAND FOR ALL TESTEES 

( ) for some
∞
1
No. 1. Suppose that S = ∑ tan -1 S ∈ R. What is the value of
k =1 2k2
tan S.
-1 -1
Solution) Let α = tan a and β = tan b.
Then a = tan α and b = tan β.
tan α - tan β a -b
Thus tan ( α - β ) = = .
1 + tan α tan β 1 +a b

We conclude that α - β = tan - 1 ( a -b

1+a b )
That is, tan -1
a - tan -1
b = tan -1
( a -b
1+a b ).
We can find the value of S. Since
n n
∑ tan -1
k =1
( 1
2k2 ) = ∑ tan - 1
k =1
{ ( 2 k +1 ) - ( 2 k -1 )
1+ ( 2k+1 ) (2k -1 ) }
n
= ∑ tan - 1 ( 2 k + 1 ) - tan -1 ( 2 k - 1 )
k =1

= tan -1 ( 2 n + 1 ) - tan -1 1

= tan -1 { ( 2n+1 ) -1
1+ (2 n+1 ) }
= tan -1 ( n
n +1 )
n
S = lim
n →∞
∑ tan -1
k =1
( 1
2k2 )= lim tan - 1
n →∞
( n
n+1 )= tan - 1 1 =
π
4
.

Hence tan S = 1.

6.2 Series With Nonnegative Terms

p.164
Exercises
1. Prove that each of the following series converges.
∞
k-3
(a) ∑ .
k=1 k3 + k + 1
k -3 k 1
Proof) First, 0 ≤ < = for all k ≥ 3.
k3 + k + 1 k3 k2
Then, by p-Series Test and Comparison Test,
∞ ∞
1 k-3
∑ converges and it implies that ∑ converges, too.
k =3 k2 k =3 k3 + k + 1
Hence,

These solutions might be inexact, yet do you rely on them?

- 11 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞ 2 ∞
k-3 k -3 k-3
∑ = ∑ +∑
k=1 k3 + k + 1 k =1
3
k +k +1 k=3 k3 + k + 1
∞
25 k-3
= - +∑ 3
33 k =3 k +k+1
also converges.
∞
k-1
(b) ∑ .
k =1 k 2k
k -1 1
Proof) First, 0 ≤ < for all k ∈ N.
k 2k 2k
∞
1
Then, by Geometric Series, ∑ converges.
k =1 2k
∞
k-1
And, by Comparison Test, ∑ also converges.
k =1 k 2k
∞
log k
(c) ∑ , p > 1.
k =1 kp
p -1
Proof) If p > 1, then > 0.
2
By 4.3Exercises 4 (b), there exists a constant B p -1 > 1 such that
2

p-1
2
0 ≤ log k ≤ k for all k ≥ B p -1 .
2
Set N : = [ B p -1 ]
2

where [ x ] is the biggest integer which is not bigger than x.

Then,
p -1
∞ 2 ∞
k 1
since ∑ p = ∑ p +1 converges (by p-Series Test) and
k = N+1 k k = N+1 2
k
∞ N ∞
log k log k log k
∑ =∑ + ∑
k =1 kp k =1 k p
k = N +1 kp
N ∞
≤∑ log k 1
p + ∑ p +1
k=1 k k = N +1 2
k
by Comparison Test,
∞
log k
∑ converges, also.
k =1 kp
∞
1
(d) ∑ .
k =1 k 3 k -1
1 1
Proof) First, 0 ≤ ≤ for all k ∈ N.
k 3k -1 3k -1
∞
1
By Geometric Series, ∑ converges.
k =1 3 k -1

These solutions might be inexact, yet do you rely on them?

- 12 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
1
And, by Comparison Test, ∑ converges, too.
k =1 k 3 k -1
∞
(e) ∑
k =1
( 10 + 1k ) k . -e

Proof) First, ( 10 +
k )
1 -e 11
k ≤ for all k ∈ N.
ke
By p-Series Test and Comparison Test,
∞
∑
k =1
( 10 + 1k ) k -e
converges, too.

∞
3 k2 - k
(f) ∑ .
k=1 k4 - k2 + 1
3 k2 - k 1
Proof) Set a k : = and b k : = 2 .
k4 - k2 + 1 k
an 3 n 4 -n 2 n
Then, L : = lim = lim = 3.
n→∞ bn n →∞ n 4 - n2 + 1
By Limit Comparison Test,
∞ ∞
3 k2 - k
since ∑ b k converges, then so does ∑ .
k =1 k=1 k4 - k2 + 1

2. Prove that each of the following series diverges.

∞ k
k
(a) ∑ k
.
k =1

k
k 1
≥ 1 for all k ∈ N.
Proof) First, =
k 1-
1
k
k
k
∞ k
k
Suppose to contrary that ∑ k
converges.
k =1

∞
1
Then, by Comparison Test, ∑ k also converges.
k =1

∞
1
But, by p-Series Test, ∑ k rightly diverges, a contradiction.
k =1

∞
1
(b) ∑ p , p > 0.
k =1 log ( k + 1 )
∞
1
Proof) First, we need to prove that ∑ k+1
diverges.
k =1

1 1
Set a k : = and b k : = .
k+1 k
an k
Since lim = lim = 1,
n →∞ bn n →∞ k +1
by contraposition of the Limit Comparison Test,
∞
1
∑ k+1
diverges.
k =1

These solutions might be inexact, yet do you rely on them?

- 13 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
On the other hand, by 4.3Exercises 4 (b),
there exists a constant B 1 > 1 such that
p

1
p
0 ≤ log ( k + 1 ) ≤ ( k + 1 ) for all k + 1 ≥ B 1.
p
1 1
That is, ≥ for all k ≥ B 1 - 1.
p
log ( k + 1 ) k +1 p
∞ ∞
1 1
Since ∑ k +1
= ∞, by Comparison Test, ∑ p = ∞.
k =1 k =1 log ( k + 1 )
∞
1
Hence, ∑ diverges.
k =1 log p ( k + 1 )

∞
k2 + 2 k + 3
(c) ∑ .
k =1 k - 2 k2 + 2
3

k2 + 2 k + 3 1
Proof) Set a k : = and b k : = .
k - 2 k2 + 2
3 k
an n3 + 2 n2 + 3 n
Then, L : = lim = lim = 1.
n→∞ bn n→∞ n3 - 2 n2 + 2
By Limit Comparison Test,
∞
1 ∞
k2 + 2 k + 3
since ∑ k diverges, then so does ∑ .
k =1 k =1 k - 2 k2 + 2
3

∞
1
(d) ∑ , p ≤ 1.
k =2 k log p k
∞
1
Proof) To find out whether ∑ converges or diverges,
k =2 k log p k
1
set f ( x ) = and observe that f is positive on [ 2, ∞).
x log p x

Since f ' ( x ) = -
1
x2 ( 1
log p x
+
1
log p +1 x ) is negative on [ 2, ∞),

it is also clear that f is decreasing.

For p < 1,
⌠ f ( x ) dx = lim ⌠ b
∞
1
⌡2 dx
b → ∞ ⌡2 x log p x
b
= lim
b →∞
[ 1
1 -p
( log x ) 1 - p ] 2

= ∞
and for p = 1,
∞ b
⌠ f ( x ) dx = lim ⌠ 1
⌡2 dx
b → ∞ ⌡2 x log x
= lim [ log ( log x ) ]b2
b →∞

= ∞

These solutions might be inexact, yet do you rely on them?

- 14 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
it follows from the contraposition of Integral Test that
∞
1
∑ diverges for every p ≤ 1.
k =2 k log p k

3. Find all p ≥ 0 such that the following series converges.

∞
1
∑ p .
k =1 k log ( k + 1 )
∞
1
Solution) To find out whether ∑ p converges or diverges,
k =1 k log ( k + 1 )
we will use 6.2Exercises 2 (b).
First,
∞ ∞
1 1 1
∑ p = +∑
k =1 k log ( k + 1 ) log p 2 k =2
p
k log ( k + 1 )
∞
1 1
< +∑
log p 2 k =2 k log p k
∞ ∞
1 1
∑ p > ∑
k =1 k log ( k + 1 ) k =1 ( k + 1 ) log p ( k + 1 )
∞
1
=∑
k =2 k log p k
∞ ∞ ∞
1 1 1 1
Thus, ∑ <∑ < +∑ .
k =2 k log p k k =1
p
k log ( k + 1 ) log p 2 k =2 k log p k
∞
1
It implies that ∑ converges
k =2 k log p k
∞
1
if and only if ∑ converges,
k =1 k log p ( k + 1 )
∞
1
or ∑ diverges
k =2 k log p k
∞
1
if and only if ∑ diverges.
k =1 k log p ( k + 1 )
1
Set f ( x ) = and observe that f is positive on [ 2, ∞).
x log p x

Since f ' ( x ) = -
1
x2 ( 1
log p x
+
1
log p +1 x ) is negative on [ 2, ∞),

it is also clear that f is decreasing.

For 0 ≤ p < 1,
⌠ f ( x ) dx = lim ⌠ b
∞
1
⌡2 dx
b → ∞ ⌡2 x log p x
b
= lim
b →∞
[ 1
1 -p
( log x ) 1 - p ] 2

= ∞
and for p = 1,

These solutions might be inexact, yet do you rely on them?

- 15 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
⌠ f ( x ) dx = lim ⌠ b
∞
1
⌡2 dx
b → ∞ ⌡2 x log x
= lim [ log ( log x ) ]b2
b →∞

= ∞
and for p > 1,
⌠ f ( x ) dx = lim ⌠ b
∞
1
⌡2 dx
b → ∞ ⌡2 x log p x
b
= lim
b →∞
[ 1
1 -p
( log x ) 1 - p ] 2

1
= ( log 2 ) 1 - p
p-1
∞
1
Hence, for all p > 1, ∑ converges.
k =1 k log p ( k + 1 )

4. If a k ≥ 0 is a bounded sequence, prove that

∞ ak
∑
k =1 ( k + 1 )p
converges for all p > 1.
Proof) Since a k ≥ 0 is a bounded sequence, then

there is an M ∈ R such that ∣a n∣≤ M for all n ∈ N.

an M M
We have 0 ≤ ≤ < for all n ∈ N.
( n +1 ) p ( n + 1 )p np
By p-Series Test and Comparison Test,
∞ ak
∑ converges for all p > 1.
k =1 ( k + 1 )p

∞
5. Suppose that a k ∈ [ 0, 1 ) and a k → ∞ as k → ∞. Prove that ∑ arcsin a k
k=1

∞
converges if and only if ∑ a k converges.
k=1

∞
Proof) (⇒) Suppose that ∑ arcsin a k converges.
k=1

Let b k = arcsin a k. Then, a k = sin b k.

By 4.1Exercises 4 (vi ),
π
for 0 ≤ b k < , it holds that 0 ≤ sin b k ≤ b k.
2
That is, 0 ≤ a k ≤ b k for all k ∈ N.
Thus, 0 ≤ a k ≤ arcsin a k for all k ∈ N.

These solutions might be inexact, yet do you rely on them?

- 16 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
By Comparison Theorem, ∑ a k converges, also.
k =1

∞
(⇐) Suppose that ∑ a k converges.
k=1

Then by the contraposition of Theorem 6.5, lim a n = 0.

n→∞

And, lim b n = lim arcsin a n = arcsin ( lim a n ) = 0.

n →∞ n →∞ n →∞

Using the Limit Comparison Test,

an sin b n
L = lim = lim = 1.
n →∞ arcsin a n n →∞ bn
∞ ∞
Hence, ∑ arcsin a k converges if and only if ∑ a k converges.
k=1 k =1

∞
6. If ∑ ∣a k∣ converges, prove that
k=1

∞ ∣a k∣
∑
k =1 kp
converges for all p ≥ 0. What happens if p < 0?
∣a k∣
Proof) Since 0 < ≤∣a k∣ for all p ≥ 0 and k ∈ N,
kp
∞ ∣a k∣
by Comparison Test, ∑ converges for all p ≥ 0.
k=1 kp
Nevertheless, if p < 0,
∞ ∣a k∣
we do not find out whether ∑ converges or not.
k=1 kp
1
For instance, if a k = and p = - 0.5,
k2
1
∞ ∣a k∣ ∞
k2 ∞
1
∑ = ∑ = ∑ converges,
k=1 kp k=1 k
- 0.5
k=1 k 1.5

but if p = -1,
1
∞ ∣a k∣ ∞
k2 ∞
1
∑ = ∑ = ∑
k
diverges.
k=1 kp k=1 k
-1
k=1

7. Suppose that a k and b k are nonnegative for all k ∈ N.

∞ ∞ ∞
(a) Prove that if ∑ a k and ∑ b k converge, then ∑ a k b k also converges.
k=1 k =1 k=1

Proof) i ) The first elucidation.

∞ ∞
Set L 1 = ∑ a k and L 2 = ∑ b k .
k=1 k=1

These solutions might be inexact, yet do you rely on them?

- 17 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
2 2

∣ ∑ a ∣ and L = ∣ ∑ b ∣ .
∞ ∞
Then, L 1 2 = k 2
2
k
k=1 k=1

2
≤∣ ∑ a ∣ ≤ L
n n
2 2
And, 0 ≤ ∑ a and k k 1
k=1 k=1

2
≤∣ ∑ b ∣ ≤ L .
n n
2 2
0 ≤∑ b k k 2
k =1 k=1

n n
Let s n = ∑ a k 2 and t n = ∑ b k 2.
k=1 k=1

Since a k 2 ≥ 0 and b k 2 ≥ 0, { sn } and { tn } are increasing sequences.

Moreover, { sn } and { tn } are bounded above by L 1 2 + L 2 2 .

By the Monotone Convergence Theorem,
∞ ∞
∑ a k 2 = lim s n and ∑ b k 2 = lim t n exist.
k=1 →∞ n k =1 →∞ n

∞ ∞ ∞
Let s = lim s n = ∑ a k 2, t = lim t n = ∑ b k 2 and r = ∑ ( a k + b k ) 2.
n →∞ k =1 n →∞ k =1 k=1

( ak + b k )2 - ak 2 - bk 2
We consider that a k b k = .
2
∞
r -s -t
Hence, ∑ ak bk = 2
, it converges.
k=1

ii ) The second elucidation.

∞
Since ∑ b k converges, we can choose an N ∈ N such that
k =1

n ≥ N implies ∣b n∣< 1.
Thus a n b n = a n∣b n∣≤ a n for all n ≥ N.
Moreover,
∞ N -1 ∞
∑ ak bk = ∑ ak b k + ∑ ak bk
k=1 k=1 k=N
N-1 ∞
≤ ∑ ak bk + ∑ ak
k=1 k=N

∞ ∞
Since ∑ a k converges, so does ∑ a k.
k=1 k=N

∞
By the Comparison Test, ∑ a k b k also converges.
k=1

iii ) A wrong elucidation.

The following proof is wrong.
∞
Since ∑ a k converges, then
k=1

an bn
L = lim = lim a n = 0.
n →∞ bn n →∞

These solutions might be inexact, yet do you rely on them?

- 18 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
By the second of the Limit Comparison Test, ∑ a k b k converges.
k=1

However, this proof is NOT closely exact

because there might be n’ s such that b n = 0.
(b) Improve this result by replacing convergence of one of the series by
something else.

8. Prove Theorem 6.16 (ii ) and (iii ).

an
Proof of ii ) By hypothesis, L = lim = 0.
n →∞ bn

Then, let ε = 1, there is an N ∈ N such that

n ≥ N implies ∣ ba ∣< 1.
n

Since a n ≥ 0 and b n ≥ 0, then, a n < b n for large n .

∞ ∞
By Comparison Test, if ∑ b k converges, so does ∑ a k.
k =1 k=1

an
Proof of iii ) By hypothesis, L = lim = ∞.
n →∞ bn

Then, for every M ∈ R, there is an N ∈ N such that

an
n ≥ N implies > M.
bn
Since a n ≥ 0 and b n ≥ 0, then a n > M b n for large n.
∞ ∞
If ∑ b k diverges, then so does ∑ a k.
k =1 k=1

b
9. Suppose that a, b ∈ R satisfy ∈ R \ Z. Find all q > 0 such that
a
∞
1
∑
k =1 ( a k + b ) qk
converges.
Proof) Case 1 : q > 1.
∞
1
By Geometric Series, ∑ converges.
k =1 qk
Using the Limit Comparison Test,
1
( a n + b ) qn 1
L = lim = lim = 0.
n →∞ 1 n →∞ a n +b
qn
∞
1
Then, ∑ converges.
k =1 ( a k + b ) qk

These solutions might be inexact, yet do you rely on them?

- 19 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
Case 2 : 0 < q ≤ 1.
1
∞
1 n
Since ∑ k diverges and L 1 = lim 1
= a,
k =1 n→∞
an +b
∞
1
by Limit Comparison Test, ∑ a k+b
diverges.
k =1

1
a n+b
And, L 2 = lim = lim q n exists as 0 or 1.
n →∞ 1 n →∞
( a n + b ) qn
∞
1
If L 2 = 1, then we have that ∑ diverges promptly.
k =1 ( a k + b ) qk
Otherwise, that is L 2 = 0,
∞
1
suppose to contrary that ∑ converges.
k =1 ( a k + b ) qk
∞
1
By Limit Comparison Test, then ∑ a k +b
converges, also.
k =1

∞
1
It contradicts that ∑ a k+b
diverges.
k =1

∞
10. Suppose that a k → 0. Prove that ∑ a k converges if and only if the series
k=1

∞
∑ ( a 2k + a 2k + 1 ) converges.
k=1

Proof) (⇒) Let ε > 0 be given.

∞
Suppose that ∑ a k converges.
k=1

n n
Set s n = ∑ a k, L 1 = lim s n and t n = ∑ ( a 2k + a 2k + 1 ).
k=1 n→∞ k=1

Then
n
t n = ∑ ( a 2k + a 2k + 1 )
k=1

= ( a 2 + a 3 ) + ( a 4 + a 5 ) + ⋯ + ( a 2n + a 2n + 1 )
2n + 1
= - a1 + ∑ ak
k=1

= s 2n + 1 - a 1

Since L 1 = lim s n, there is an N ∈ N such that

n →∞

n ≥ N implies ∣s n - L 1∣< ε.
If n ≥ N, then

These solutions might be inexact, yet do you rely on them?

- 20 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∣t n - ( L 1 - a 1 )∣ = ∣t n - s 2n + 1 + a 1 + s 2n + 1 - L 1∣

≤ ∣t n - s 2n + 1 + a 1∣+∣s 2n + 1 - L 1∣

= ∣s 2n + 1 - L 1∣

< ε
Hence lim t n = L 1 - a 1.
n →∞

(⇐) This proof is similar to (⇒).

Since L 2 : = lim t n, there is an N ∈ N such that
n→∞

n ≥ N implies ∣t n - L 2∣< ε.
If n ≥ N, then
∣s 2n + 1 - ( L 2 + a 1 )∣ = ∣s 2n + 1 - t n - a 1 + t n - L 2∣

≤ ∣s 2n + 1 - t n - a 1∣+∣t n - L 2∣

= ∣t n - L 2∣

< ε
Thus lim s 2n + 1 = L 2 + a 1.
n→∞

Since a n → 0 as n → ∞ and a 2n + 1 = s 2n + 1 - s 2n,

it follows that
lim s 2n = lim ( s 2n + 1 - a 2n + 1 )
n →∞ n→∞

= lim s 2n + 1 - lim a 2n + 1
n →∞ n→∞

= L2 + a 1
We conclude from 2.1Exercises 8 that lim s n = L 2 + a 1.
n →∞

∞
Hence ∑ a k converges.
k=1

6.3 Absolute Convergence

p.171
Exercises
1. Prove that each of following series converges.
∞
1
(a) ∑ k!
.
k=1

These solutions might be inexact, yet do you rely on them?

- 21 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
1
Proof) Let a k = / 0 for all k ∈ N.
=
k!
∣a k + 1∣ 1
Then, r = lim = lim = 0.
k →∞ ∣a k∣ k →∞ ∣k + 1∣
∞
1
By Ratio Test, ∑ k!
converges absolutely.
k =1

∞
1
(b) ∑ .
k =1 kk
1
Proof) Let a k = for all k ∈ N.
kk
1
k 1
Then, r = limsup k→∞ ∣a k∣ = limsup k→∞ = 0.
∣k∣
∞
1
By Root Test, ∑ converges absolutely.
k =1 kk
∞
2k
(c) ∑ k!
.
k=1

2k
Proof) Let a k = / 0 for all k ∈ N.
=
k!
∣a k + 1∣ 2
Then, r = lim = lim = 0.
k →∞ ∣a k∣ k →∞ ∣k + 1∣
∞
2k
By Ratio Test, ∑ k!
converges absolutely.
k=1
2
k
( )
∞
k
(d) ∑ k +1
.
k =1

k2
Proof) Let a k = ( k
k +1 ) for all k ∈ N.

Then,
1
k
r = limsup k →∞ ∣a k∣
k
= limsup k →∞ ( k
k+1 )
1
= limsup k →∞ k
( 1 + 1k )
1
= < 1
e
∞ k2
By Root Test, ∑
k =1
( k
k +1 ) converges absolutely.

2. Decide, using results covered so far in this chapter, which of the following
series converge and which diverge.

These solutions might be inexact, yet do you rely on them?

- 22 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
k2
(a) ∑ k.
k =1 π

k2
Solution) Let a k = / 0 for all k ∈ N.
=
πk
∣a k +1∣ 2
Then, r = lim
k →∞ ∣a k∣
= lim
k → ∞
1
π ( 1 + 1k ) =
1
π
< 1.

∞
k2
By Ratio Test, ∑ converges absolutely.
k =1 πk
∞
k!
(b) ∑ .
k =1 2k
k!
Solution) Let a k = / 0 for all k ∈ N.
=
2k
∣a k +1∣ k
Then, r = lim = lim = ∞ > 1.
k →∞ ∣a k∣ k → ∞ 2
∞
k!
By Ratio Test, ∑ diverges.
k =1 2k
∞ k
(c) ∑
k =1
( k +1
2 k +3 ).
k
Solution) Let a k = ( k+1
2 k +3 ) for all k ∈ N.

1
k k +1 1
Then, r = limsup k →∞ ∣a k∣ = limsup k →∞ = < 1.
2k+3 2
k
( )
∞
k +1
By Root Test, ∑ 2 k +3
converges absolutely.
k =1

∞
(d) ∑
k =1
( π - 1k ) k -1
.

Solution) Let a k = ( π - 1k ) k -1
and b k =
1
k
for all k ∈ N.

an
Then, L = lim
n →∞ bn
= lim
n →∞
( π - 1n ) = π.
∞
By Limit Comparison Test, ∑
k =1
( π - 1k ) k -1
diverges.

k2
( )
∞
k!
(e) ∑ ( k + 2 )!
.
k =1

k2
Solution) Let a k = ( k!
( k + 2 )! ) for all k ∈ N.

1
k2 k
If r = limsup k →∞ {( k!
( k + 2 )! ) } exists, then

These solutions might be inexact, yet do you rely on them?

- 23 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
log r = limsup k →∞ k log ( k!
( k + 2 )! )
= -limsup k →∞ k ( log ( k + 1 ) + log ( k + 2 ) )

= -∞
-∞
Therefore, r = e = 0.
∞ k2
By Root Test, ∑
k =1
( k!
( k + 2 )! ) converges absolutely.

( 3 + (-1 ) k
).
∞
(f) ∑ 3
k =1

k
Solution) Let a k = ( 3 + (-1 ) k
3 ) for all k ∈ N.

1
k 3 + (-1 ) k 4
Then, r = limsup k→∞ ∣a k∣ = limsup k →∞ = > 1.
3 3
k

( 3 + (-1 ) k
)
∞
By Root Test, ∑ 3
diverges.
k =1

∞
( 1 + (-1 ) k ) k
(g) ∑ .
k =1 ek
( 1 + (-1 ) k ) k
Solution) Let a k = for all k ∈ N.
ek
1
k 1 + (-1 ) k 2
Then, r = limsup k→∞ ∣a k∣ = limsup k →∞ = < 1.
e e
∞
( 1 + (-1 ) k ) k
By Root Test, ∑ converges absolutely.
k =1 ek

3. Using Exercise 9, p. 135, prove that

∞
(-1 ) k x 2k +1 ∞
(-1 ) k x 2k
sin x = ∑ and cos x = ∑
k =0 ( 2 k + 1 )! k =0 ( 2 k )!

for all x ∈ [ 0, π2 ].
Proof) By Exercise 9, p. 135,
for each nonnegative integer m, set
m
(-1) k x 2k +1 m
(-1) k x 2k
sm (x ) : = ∑ and c m ( x ) : = ∑ .
k =0 ( 2 k + 1 )! k =0 ( 2 k )!
Then it follows that
s 2n +1 ( x ) ≤ sin x ≤ s 2n ( x ), s 2n +1 ( x ) ≤ sin x ≤ s 2n +2 ( x ),
c 2n +1 ( x ) ≤ cos x ≤ c 2n ( x ), and c 2n +1 ( x ) ≤ cos x ≤ c 2n +2 (x )
hold for n = 0, 1, 2, ⋯.
2n
(-1) k x 2k +1 2n + 1
(-1) k x 2k + 1
Thus s 2n ( x ) = ∑ and s 2n +1 ( x ) = ∑ .
k =0 ( 2 k + 1 )! k =0 ( 2 k + 1 )!
Using the Ratio Test,

These solutions might be inexact, yet do you rely on them?

- 24 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

∣ (-1) k +1 x 2k +3
∣
∣ ( 2 k + 2 )x ( 2 k + 3 ) ∣= 0.
2
( 2 k + 3 )!
r = lim = lim
k →∞
∣ (-1) k x 2k +1
( 2 k + 1 )! ∣ k →∞

Therefore s2n (x ) and s 2n +1 ( x ) converge for all x ∈ R.

Say it s ( x ) : = lim s 2n ( x ) = lim s 2n +1 ( x ).
n →∞ n →∞

By the Squeeze Theorem, we conclude that

∞
(-1 ) k x 2k +1
sin x = s ( x ) = lim s 2n ( x ) = ∑ for all x ∈ R.
n →∞ k =0 ( 2 k + 1 )!
2n
(-1) k x 2k 2n + 1
(-1) k x 2k
Similarly, c 2n ( x ) = ∑ and c 2n +1 ( x ) = ∑ .
k =0 ( 2 k )! k =0 ( 2 k )!
Using the Ratio Test,

∣ (-1) k +1 x 2k +2
∣
r = lim
( 2 k + 2 )!
= lim ∣ x2
∣= 0.
k →∞
∣ (-1) k x 2k
( 2 k )! ∣ k →∞ ( 2 k +1 ) ( 2 k +2 )

Therefore c 2n ( x ) and c 2n +1 ( x ) converge for all x ∈ R.

Say it c ( x ) : = lim c 2n ( x ) = lim c 2n +1 ( x ).
n →∞ n →∞

By the Squeeze Theorem, we conclude that

∞
(-1 ) k x 2k
cos x = c ( x ) = lim c 2n ( x ) = ∑ for all x ∈ R.
n →∞ k =0 ( 2 k )!

4. Define a k recursively by a 1 = 1 and

-1
a k = (-1 ) k ( 1 + k sin ( 1k )) a k - 1, k > 1.
∞
Prove that ∑ a k converges absolutely.
k=1

Proof) Since a 1 = 1, then a k is not zero.

Then,
∣a k +1∣
r = lim
k →∞ ∣a k∣
= lim
k →∞ ∣ (-1 ) k +1
1 + ( k + 1 ) sin ( 1
k+1 )∣
1
=
2
∞
By Ratio Test, ∑ a k converges absolutely.
k =1

1 ∞
5. Suppose that a k ≥ 0 and ak k
→ a as k → ∞. Prove that ∑ a k xk
k =1

These solutions might be inexact, yet do you rely on them?

- 25 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
1
converges absolutely for all ∣x∣< if a =
/ 0 and for all x ∈ R if a = 0.
a
Proof) Case 1 : a = 0.
For every x ∈ R, there is an M ∈ R such that ∣x∣≤ M.
1
Given ε = , there is an N ∈ N such that
2
1
k ≥ N implies ∣a ∣<k
k ε
M
.

1
That is, ∣a ∣<
k
k 1
2M
for large k.

Then,
1
k k
r = limsup k→∞ ∣a k x ∣
1
= limsup k→∞ ∣a k
k
∣∣x∣
1 ⋅
< limsup k→∞ M
2M
1
=
2
∞
By the Root Test, ∑ a k x k converges absolutely.
k =1

Actually, we can grasp r = 0.

Case 2 : a =
/ 0.
1
k
Since a k ≥ 0, then a k ≥ 0 for all k ∈ N.
1
n
Thus a = lim a n is nonnegative, too.
n→∞

1
For all ∣x∣< ,
a
1
k k
r = limsup k→∞ ∣a k x ∣
1
= limsup k→∞ ∣a k
k
∣∣x∣
= a ∣x∣

< 1
∞
By the Root Test, ∑ a k x k converges absolutely.
k =1

6. For each of the following, find all values of p ∈ R for which the given series
converges absolutely.

These solutions might be inexact, yet do you rely on them?

- 26 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
1
(a) ∑ .
k =2 k log p k
Solution) Case 1 : p ≥ 0 and p =
/ 1.
To find out whether
∞
1
∑ converges or diverges for some p ∈ R,
k =2 k log p k
1
set f ( x ) = and observe that f is positive on [ 2, ∞).
x log p x
p x + log x
Since f ' ( x ) = - < 0 on [ 2, ∞) for p ≥ 0, then
x 2 log p +1 x
f is decreasing.
b
[ ]
∞
⌠ 1 1
⌡2 dx = lim log 1 -p x
x log p x b →∞ p -1 2

1
= { p -1
∞,
log 1 - p 2, p > 1

0 ≤ p < 1
Case 2 : p = 1.
Similarly,
∞
⌠ 1
⌡2 dx = lim [ log ( log x ) ]b2
x log x b →∞

= ∞
Case 3 : p < 0.
1 1
A factor > for all k > e.
k log p k k
∞
1
Since ∑ k diverges by the p-Series Test,
k =3

it follows from the Comparison Test that

∞ ∞
1 1 1
∑ = +∑
k =2 k log p k 2 log p 2 k =3 k log p k
∞
1 1
> +∑
2 log p 2 k =3 k

= ∞
∞
1
We conclude that ∑ converges for all p > 1.
k =2 k log p k
∞
1
(b) ∑ .
k =2 log p k
Solution) Case 1 : p = 0.
∞ ∞
1
∑ = ∑ 1 = ∞.
k =2 log p k k =2

Case 2 : p > 0.

These solutions might be inexact, yet do you rely on them?

- 27 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞ ∞
1 1
By 6.2Exercises 2 (b), ∑ = ∑ diverges.
k =2 log p k k =1 log p ( k + 1 )
Case 3 : p < 0.
1
Since does not converge to zero for p < 0,
log p k
∞
1
by Divergence Test, ∑ diverges.
k =2 log p k
We conclude that
∞
1
there are not any values p ∈ R for which ∑ converges.
k =2 log p k
∞
kp
(c) ∑ .
k =1 pk
Solution) Case 1 : p = 1.
∞
kp ∞
∑ k = ∑ k =
∞.
k =1 p k =1

Case 2 : p = -1.

∣∑ (-1 ) k
∣= ∣ (-1 ) ( )∣
∞
n 1 1 1
- + -⋯
k=n k n n +1 n+2
1 1 1 1 1
= - + - + -⋯
n n +1 n+2 n+3 n+4

=
1
n
+ - ( 1
n+1
+
1
n+2 ) + (- 1
n+3
+
1
n +4 )+⋯
1
<
n
∞
(-1 ) k
Thus, by Corollary 6.9, ∑ k
converges.
k =1

∣ (-1 ) k
∣=∑
∞ ∞
1
But, by p-Series Test, ∑ k k
diverges.
k =1 k =1

Case 3 : 0 <∣p∣< 1.
kp
Let a k = / 0 for all k ∈ N.
=
pk
∣a k + 1∣ p
Then, r = lim
k →∞ ∣a k∣
= lim
k →∞
1
∣p∣ ∣1 + 1k ∣ = 1
∣p∣
> 1.

∞
kp
By Ratio Test, ∑ diverges.
k =1 pk
Case 4 : ∣p∣> 1.
kp
Similarly with Case 3, let a k = / 0 for all k ∈ N.
=
pk
∣a k + 1∣ p
Then, r = lim
k →∞ ∣a k∣
= lim
k →∞
1
∣p∣ ∣ 1+
1
k ∣= 1
∣p∣
< 1.

These solutions might be inexact, yet do you rely on them?

- 28 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
kp
By Ratio Test, ∑ k converges absolutely.
k =1 p

We conclude from Case 1, 2, 3 and 4 that

∞
kp
∑ converges absolutely for ∣p∣> 1,
k =1 pk
and converges conditionally for p = -1,
and diverges for -1 < p ≤ 1.
∞
1
(d) ∑ .
k =2 k ( kp - 1 )
1
Solution) Case 1 : p ≤ .
2
1
Then p + ≤ 1.
2
1 1 1
Since > = for all k > 2,
k ( kp - 1 ) k kp p+
1
2
k
it follows from the p-Series Test and the Comparison Test that
∞
1
∑ diverges.
k =2 k ( kp - 1 )
1
Case 2 : p > .
2
1
Then p + > 1.
2
1 1
Let a n = and b n = for all n = 2, 3, 4, ⋯.
p+
1
2 n ( np - 1 )
n
an
Then L = lim
n →∞ bn
= lim
n →∞
( 1 - n1 ) = 1.
p

∞
1
Since ∑ p+
1 converges by the p-Series Test,
k=2 2
k
it follows from the Limit Comparison Test that
∞
1
∑ converges, also.
k =2 k ( kp - 1 )

∞
2 kp k!
(e) ∑ .
k=1 kk
Solution) Case 1 : p = log 2 e.
∞
2 kp k! ∞
e k k!
Then ∑ k = ∑ .
k=1 k k=1 kk
∞
e k k!
We do not know whether ∑ converges or not.
k=1 kk
Thus, we should make good use of Stirling’s Formula (p.447).

These solutions might be inexact, yet do you rely on them?

- 29 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
k!
According to the theorem, for large k, ≈ 1.
2π k k k e-k
Then
e k k! k!
= ⋅ 2π k
kk 2π k k k e -k
≈ 2π k
∞
e k k!
Thus by the Divergence Test, ∑ diverges.
k =1 kk
Case 2 : p =
/ log 2 e.

2 kp k!
Let a k = for all k ∈ N. Then
kk
∣a k + 1∣ k
r = lim
k →∞ ∣a k∣
= lim
k →∞
∣2 ( p k
k+1 )∣
= lim
k →∞ ∣( 1+
2p
1
k )
k
∣
2p
=
e
∞
2 kp k!
By the Ratio Test, ∑ converges absolutely for p < log 2 e
k=1 kk
and diverges for p > log 2 e.
We conclude from Case 1 and 2 that
∞
2 kp k!
∑ converges absolutely for p ∈ (-∞, log 2 e )
k=1 kk
and diverges for p ∈ [ log 2 e, ∞ ).
∞
(f) ∑ ( k 2p + 1 - k p ).
k =1

Solution) Case 1 : p > 1.

First, we have

k 2p + 1 - k p =
( k 2p + 1 - k p ) ( k 2p + 1 + k p ) = 1
.
2p p
k +1 +k k + 1 + kp
2p

1 1 1
Since < = for all k ∈ N,
k + 1 + kp
2p
k 2p
+k p
2 kp
it follows from the p-Series Test and the Comparison Test that
∞
∑ ( k 2p + 1 - k p ) converges.
k =1

Case 2 : 0 < p ≤ 1.
1 1
Let a n = and b n = .
np n 2p + 1 + n p

These solutions might be inexact, yet do you rely on them?

- 30 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
an
Then L = lim
n →∞ bn
= lim
n →∞
( 1
1 + 2p
k
+1 ) = 2.
∞
1
Since ∑ diverges by the p-Series Test,
k =1 kp
it follows from the Limit Comparison Test that
∞
1
∑ 2p
diverges, also.
k =1 k + 1 + kp
Case 3 : p ≤ 0.
Since k p → 0 as k → ∞, it satisfies that

lim ( k 2p + 1 - k p ) = 1.
k →∞

∞
It follows from the Divergence Test that ∑ ( k 2p + 1 - k p ) diverges.
k =1

7. Suppose that a kj ≥ 0 for k, j ∈ N. Set

∞
A k = ∑ a kj
j=1

∞
for each k ∈ N, and suppose that ∑ A k converges.
k=1

(a) Prove that

(∑ a ) ( ∑ a ).
∞ ∞ ∞ ∞
∑ kj ≤∑ kj
j =1 k =1 k =1 j =1

(b) Show that

( ∑ a ) = ∑ ( ∑ a ).
∞ ∞ ∞ ∞
∑ kj kj
j =1 k =1 k =1 j =1

(c) Prove that (b) may not hold if a kj has both positive and negative values.
Hint : Consider
 1, j =k


akj =  -1, j = k +1


 0, otherwise

∞ ∞
8. (a) Suppose that ∑ a k converges absolutely. Prove that ∑ ∣a k∣ p converges
k=1 k=1

for all p ≥ 1.
Proof) By the contraposition of Divergence Test, a k → 0 as k → ∞.

That is, for given ε = 1, there is an N ∈ N such that

k ≥ N implies ∣a k∣< ε.
Then, ∣a k∣< 1 for all k ≥ N.

These solutions might be inexact, yet do you rely on them?

- 31 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
It is clear that ∣a k∣ p ≤∣a k∣ for all k ≥ N and p ≥ 1.
∞
By Comparison Test, ∑ ∣a k∣ p converges for all p ≥ 1.
k=1

∞ ∞
(b) Suppose that ∑ a k converges conditionally. Prove that ∑ k p a k diverges for
k =1 k=1

all p > 1.
∞
Proof) Suppose that ∑ ∣a k∣ diverges.
k=1

∣k p a k∣
Then, L = lim = lim ∣k p∣= ∞.
k →∞ ∣a k∣ k→∞

∞
By Limit Comparison Test, ∑ ∣k p a k∣ diverges.
k=1

∞
Then, ∑ k p a k diverges or converges conditionally.
k=1

∞
Suppose to contrary that ∑ k p a k converges.
k=1

Given ε = 1, there is an N ∈ N such that

k ≥ N implies ∣k p a k∣< ε.
1
Then, ∣a k∣< for all k ≥ N.
kp
∞
By p-Series Test and Comparison Test, ∑ ∣a k∣ converges.
k=1

∞
It contradicts that ∑ ∣a k∣ diverges.
k=1

a2
9. (a) Let an > 0 for n ∈ N. Set b 1 = 0, b 2 = log , and
a1
ak ak -1
b k = log - log , k = 3, 4, ⋯.
ak -1 ak -2
Prove that if
an
r = lim
n→∞ a n -1
exists and is positive, then
1 n ∞
lim log a nn = lim
n →∞ n →∞
∑
k=1
(1- k-1
n )b k = ∑ b k = log r.
k=1

Proof)
By the Telescopic Series,

These solutions might be inexact, yet do you rely on them?

- 32 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
n n
∑ bk = b1 + b2 + ∑ bk
k =1 k=3

( log a a - log aa )
n
k k -1
= b1 + b2 + ∑
k=3 k -1 k -2

+ ( log
a )
a2 a a n 2
= 0 + log - log
a1 a n -1 1

an
= log
an -1
∞ n a
Thus ∑ b k = lim ∑ b k = lim log a n = log r.
k=1 n →∞ k=1 n →∞ n -1

By 6.1Exercises 8 (b),
n
(1- )b
∞
k -1
that ∑ b k exists implies that lim ∑ n k exists.
k =1 n →∞ k =1

∞ n
Moreover, it holds that ∑ b k = lim
k =1 →∞
∑
n k =1
(1- k -1
n )b . k

Since
1
1
log a nn = log a n
n

=
1
n ( log an -1
an
+ log
an -1
a n -2
+ ⋯ + log
a2
a1
+ log a 1 )
( ∑ b + ∑ b +⋯+ ∑ b + )
n n-1 2 1
1
=
n k k k ∑ b k + log a 1
k=1 k=1 k=1 k=1

( ∑ ( n - k + 1 ) b + log a )
n
1
= k
n k=1
1

n log a 1
= ∑
k=1
(1- k -1
n )b k +
n
1 n
lim
n →∞
log a nn = lim
n →∞
∑
k =1
(1- k -1
n )b k

We conclude that
1 n ∞
lim log a nn = lim
n →∞ n →∞
∑
k=1
(1- k-1
n )b k = ∑ b k = log r.
k=1

(b) Prove that if a n ∈ R \ { 0 } and ∣ a a ∣→ n +1

n
r as n → ∞, for some r > 0,

1
n
then ∣a n ∣ → r as n → ∞.

∞
10. Let x ≤ y be any pair of extended real numbers. Prove that if ∑ a k is
k=1

∞ ∞
conditionally convergent, then there is a rearrangement ∑ b j of ∑ a k whose
j =1 k=1

partial sums sn satisfy

These solutions might be inexact, yet do you rely on them?

- 33 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
liminf n →∞ s n = x and limsup n →∞ s n = y.

6.4 Alternating Series

p.176
Exercises
1. Prove that each of following series converges.

( )
∞
π
(a) ∑ (-1 ) k 2
- arctan k .
k =1

π π
Proof) Since arctan k ↑ as k → ∞, - arctan k ↓ 0 as k → ∞.
2 2
This follows immediately from the Alternating Series Test.
∞
Hence, ∑ (-1 ) k
k =1
( π
2
- arctan k ) converges.
∞
(-1 ) k k 2
(b) ∑ .
k =1 2k
k2
Proof) Let a k = . If k > 1 + 2 , then
2k
( k + 1 )2 k2 k2 - 2 k - 1
a k +1 - ak = k +1 - k = - < 0.
2 2 2k + 1
Thus, { ak } is decreasing for large k.

k2 ↓
And, by L’hospital’s Rule, 0 as k → ∞.
2k
∞
(-1 ) k k 2
Thus, by Alternating Series Test, ∑ converges.
k =1 2k
∞
(-1 ) k
(c) ∑ , p > 0.
k =1 kp
1 ↓
Proof) Since 0 as k → ∞,
kp
this follows immediately from the Alternating Series Test.
∞
(-1 ) k
Hence, ∑ converges for p > 0.
k =1 kp
∞
sin ( k x )
(d) ∑ , x ∈ R, p > 0.
k =1 kp
n
Proof) By Example 6.34, ∑ sin ( k x ) is a bounded sequence for x ∈ R.
k =1

1 ↓
And, since 0 as k → ∞,
kp
this follows from the Alternating Series Test.

These solutions might be inexact, yet do you rely on them?

- 34 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
sin ( k x )
Hence, ∑ converges for p > 0.
k =1 kp
∞
(-1 ) k 2⋅4⋅⋯⋅( 2 k )
(e) ∑ 1⋅3⋅⋯⋅( 2 k - 1 )
.
k =1 k2
Proof) Since
2⋅4⋅⋯⋅( 2 k ) 2 ⋅ 4 ⋅⋯⋅ 2 k - 2 ⋅(
= 1⋅ 2k )
1⋅3⋅⋯⋅( 2 k - 1 ) 3 5 2k-1
< 2k
we conclude that
1 2⋅4⋅⋯⋅( 2 k ) 1 2
0 < < 2 (2k ) = .
k2 1⋅3⋅⋯⋅( 2 k - 1 ) k k
1 2⋅4⋅⋯⋅( 2 k )
Thus, → 0 as k → ∞.
k2 1⋅3⋅⋯⋅( 2 k - 1 )
By Alternating Series Test,
∞
(-1 ) k 2⋅4⋅⋯⋅( 2 k )
∑ 1⋅3⋅⋯⋅( 2 k - 1 )
converges.
k =1 k2

2. For each of the following, find all values x ∈ R for which the given series
converges.
∞
xk
(a) ∑ k
.
k =1

∞
1 ↓
Solution) Since
k
0 as k → ∞, if ∑ x k is bounded,
k =1

∞
xk
by Dirichlet’s Test, ∑ k
converges.
k =1

∞
Since ∑ x k is bounded for x ∈ [-1, 1 ),
k =1

∞ k
x
∑ k
converges for -1 ≤ x < 1.
k =1

Strictly speaking, we must examine that

∞
xk
∑ k
diverges for x ∉ [-1, 1 ).
k =1

xk 1
If x ≥ 1, then ≥ .
k k
∞
xk
By p-Series Test and Comparison Test, ∑ k
diverges.
k =1

Otherwise, if x < -1,

∣ kx+ 1 ∣
k +1

r = lim = lim ∣ k
∣∣x∣ = ∣x∣> 1,
∣ xk ∣ k+1
k
k →∞ k →∞

These solutions might be inexact, yet do you rely on them?

- 35 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
xk
by Ratio Test, ∑ k
diverges.
k =1

∞
x 3k
(b) ∑ .
k =1 2k
k
x 3k
( x3
),
∞ ∞
Solution) Since ∑ k = ∑
2
k =1 2 k =1

∞
x 3k
by Geometric Series, ∑ converges for - 3 2 < x < 3
2.
k =1 2k

3 3 x 3k
Otherwise, if x ≥ 2 or x ≤ - 2 , then lim =
/ 0.
k →∞ 2k
∞ 3k
x
By Divergence Test, ∑ diverges.
k =1 2k
∞
(-1 ) k x k
(c) ∑ .
k =1 k2 + 1
Solution) Use the Ratio Test, then

r = lim
k→∞
∣a k + 1∣
∣a k∣
= lim
k→∞
∣x 2
k2 + 1
k +2 k+2 ∣ = ∣x∣.
∞
(-1 ) k x k
Thus ∑ converges absolutely for ∣x∣< 1
k =1 k2 + 1
and diverges for ∣x∣> 1.
∞
(-1 ) k x k ∞
1
If x = -1, then ∑ 2
= ∑ .
2
k =1 k +1 k =1 k +1
∞
1
Since ∑ k diverges and
k =1

1
k k2 + 1
L = lim = lim = 1,
k →∞ 1 k →∞ k
k2 + 1
∞
1
by the Limit Comparison Test, ∑ diverges.
k =1 k2 + 1
∞
(-1 ) k x k ∞
(-1 ) k
And if x = 1, then ∑ 2
= ∑ .
k =1 k +1 k =1 k2 + 1
1
Since ↓ 0 as k → ∞, by the Alternating Series Test,
k2 + 1
∞
(-1 ) k
∑ converges.
k =1 k2 + 1
∞
( x + 2 )k
(d) ∑ .
k =1 k k +1
∞
Solution) If -3 ≤ x < -1, since ∑ ( x + 2 ) k is bounded,
k =1

These solutions might be inexact, yet do you rely on them?

- 36 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
( x + 2 )k
by Dirichlet’s Test, ∑ converges.
k =1 k k+1
Also, if x = -1, since
1
k k k +1
L = lim = lim = 1,
k →∞ 1 k →∞ k
k k +1
by p-Series Test and Limit Comparison Test,
∞
1
∑ converges.
k =1 k k+1
( x + 2 )k
However, if x > -1 or x < -3, then lim =
/ 0.
k →∞ k k +1
∞
1
Thus, ∑ diverges for x > -1 or x < -3.
k =1 k k +1

3. Using any test covered in this chapter, find out which of the following series
converge absolutely, which converge conditionally, and which diverge.
∞
(-1 ) k k 3
(a) ∑ ( k + 1 )!
.
k =1

(-1 ) k k 3
Solution) Let a k = .
( k + 1 )!
∣a k +1∣ 3
Then, r = lim
k →∞ ∣a k∣
= lim
k →∞
( 1 + 1k ) 1
k+1
= 0.

∞
(-1 ) k k 3
By Ratio Test, ∑ ( k + 1 )!
converges absolutely.
k =1

∞
(-1 )⋅(-3 )⋅⋯⋅( 1 - 2 k )
(b) ∑ 1⋅4⋅⋯⋅( 3 k - 2 )
.
k=1

(-1 )⋅(-3 )⋅⋯⋅( 1 - 2 k )

Solution) Let a k = .
1⋅4⋅⋯⋅( 3 k - 2 )
1⋅3⋅⋯⋅( 2 k - 1 )
That is, a k = (-1 ) k .
1⋅4⋅⋯⋅( 3 k - 2 )
∣a k +1∣
Then, r = lim
k →∞ ∣a k∣
= lim
k →∞
∣ (-1 ) 2k +1
3k +1 ∣= 2
3
< 1.

∞
(-1 ) k k 3
By Ratio Test, ∑ ( k + 1 )!
converges absolutely.
k =1

∞
( k + 1 )k
(c) ∑ , p > e.
k =1 p k k!
( k + 1 )k
Solution) Let a k = .
p k k!
∣a k + 1∣ k +1
Then, r = lim
k →∞ ∣a k∣
= lim
k →∞
∣ 1p ( 1 + 1
k +1 ) ∣= e
p
< 1.

These solutions might be inexact, yet do you rely on them?

- 37 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
( k + 1 )k
By Ratio Test, ∑ converges absolutely.
k =1 p k k!
∞
(-1 ) k +1 k
(d) ∑ k+1
.
k =1

1
k k +1
Solution) Since L = lim = lim = 1,
k →∞ k k→∞ k
k+1
∞
k
by p-Series Test and Limit Comparison Test, ∑ k+1
diverges.
k =1

∣ (-1 ) k +1
∣ diverges.
∞
k
That is, ∑ k+1
k =1

k ↓
Since 0 as k → ∞, by Alternating Series Test,
k+1
∞
(-1 ) k +1 k
∑ k +1
converges.
k =1

∞
(-1 ) k +1 k
Hence, ∑ k+1
converges conditionally.
k =1

∞
(-1 ) k k + 1
(e) ∑ .
k =1 k kk
(-1 ) k k + 1
Solution) Let a k = .
k kk
∣a k +1∣ k2 + k k
Then, r = lim
k →∞ ∣a k∣
= lim
k →∞ ( k + 1 )2 ( k
k +1 ) = 0⋅
1
e
= 0.

∞
(-1 ) k k + 1
By Ratio Test, ∑ converges absolutely.
k =1 k kk

∞
4. [Abel’s Test] Suppose that ∑ a k converges and b k ↓ b as k → ∞. Prove
k =1

∞
that ∑ a k b k converges.
k =1

Proof) Let 0 < ε < 1 be given.

By hypothesis, there is an N 1 ∈ N such that
n ≥ N 1 implies ∣b n - b∣< ε.
That is, ∣b n∣ < ∣b∣+ 1 for large m.

In addition, there is an N 2 ∈ N such that

∣ ∑ a ∣<
n
ε
n ≥ m ≥ N2 imply ∣A n, m∣ = k .
k=m 4∣b∣+ 4
Set N = max { N 1, N 2 }.
By Abel’s Formula, if n ≥ m ≥ N, then

These solutions might be inexact, yet do you rely on them?

- 38 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

∣∑ ∣ = ∣A ∣
n n -1
ak bk n, m bn - ∑ A k, m ( b k +1 - b k )
k=m k =m

b ∣+ ∣ ∑ ∣
n -1
≤∣A n, m n A k, m ( bk +1 - bk )
k=m

4∣b∣+ 4 ∣ ∑ ∣
n -1
ε ε
< (∣b∣+ 1 ) + ( bk +1 - bk )
4∣b∣+ 4 k =m

ε ε
< + ∣b n - b m ∣
4 4∣b∣+ 4
ε ε
≤ + ⋅2∣b ∣
4 4∣b∣+ 4
ε ε 3
< + = ε < ε
4 2 4
∞
By Cauchy Criterion, ∑ a k b k converges.
k =1

5. Prove that
∞
∑ a k cos ( k x )
k =1

converges for every x ∈ ( 0, 2 π ) and every a k ↓ 0. What happens when x = 0?

Proof) By the Telescopic Series,
n n
x x
sin
2
∑ cos k x = ∑ cos k x sin 2
k =1 k =1
n
=
1
2 ∑
k =1
( sin ( ( k + 12 ) x ) - sin ( ( k - 12 ) x ))
=
1
2 { sin ( ( n + 12 ) x ) - sin ( 12 x )}
And

∣ ( ( n + 12 ) x ) - sin ( )
∣
1
sin x
∣∑ ∣=
n
2 1
cos k x ≤ .
k =1
2 sin
x
2 ∣ sin
x
2 ∣
n
Since ∑ cos k x is bounded for x ∈ ( 0, 2 π ) and a k ↓ 0 as k → ∞,
k =1

∞
by Dirichlet’s Test, ∑ a k cos ( k x ) converges.
k =1

∞ ∞
Even if x = 0, then ∑ a k cos ( k⋅0 ) = ∑ a k.
k=1 k =1

∞
Thus, we do not know whether ∑ a k cos ( k x ) converges or not.
k =1

∞
1 1
For instance, if a k =
k
or , then a k ↓ 0 but ∑ a k diverges.
k k=1

∞
1
On the other hand, if a k = , then a k ↓ 0 and ∑ a k converges.
k2 k=1

(We had yet to learn its value,

These solutions might be inexact, yet do you rely on them?

- 39 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
1 π2
but Leonhard Euler (1707~83) had proven that ∑ 2 is equal to 6 .)
k =1 k

6. Suppose that a k ↓ 0 as k → ∞. Prove that

∞
∑ a k sin ( ( 2 k + 1 ) x )
k =1

converges for all x ∈ R.

Proof) By the Telescopic Series,
n n
2 sin x ∑ sin ( ( 2 k + 1 ) x ) = ∑ sin ( ( 2 k + 1 ) x ) sin x
k =1 k =1
n
= ∑ ( cos ( 2 k x ) - cos ( ( 2 k + 2 ) x ) )
k =1

= cos ( 2 x ) - cos ( ( 2 n + 2 ) x )
And

∣∑ ∣ =∣
n

k =1
sin ( ( 2 k + 1 ) x )
cos ( 2 x ) - cos ( ( 2 n + 2 ) x )
2 sin x ∣
1
≤
∣ sin x∣
n
Thus ∑ sin ( ( 2 k + 1 ) x ) is bounded for x ∈ R satisfying sin x =
/ 0.
k =1

n
That is, ∑ sin ( ( 2 k + 1 ) x ) is bounded for all x =
/ m π and m ∈ Z.
k =1

Even if x = m π,
∞ ∞ ∞
∑ a k sin ( ( 2 k + 1 ) x ) = ∑ a k sin ( ( 2 k + 1 ) m π ) = ∑ a k ⋅ 0 = 0.
k =1 k =1 k =1

n
Therefore ∑ sin ( ( 2 k + 1 ) x ) is bounded for all x ∈ R.
k =1

∞
By the Dirichlet’s Test, ∑ a k sin ( ( 2 k + 1 ) x ) converges.
k =1

7. Show that under the hypotheses of Dirichlet’s Test,

∞ ∞
∑ a k b k = ∑ s k ( b k - b k +1 ).
k=1 k =1

k n
Showing) Let s k = ∑ a j and A n, m = ∑ a k.
j =1 k=m

Then, A n, 1 = s n and A k, 1 = s k.
By Abel’s Formula,

These solutions might be inexact, yet do you rely on them?

- 40 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
n n -1
∑ ak bk = A n, 1 bn - ∑ A k, 1 ( bk+1 - bk )
k=1 k=1
n-1
= s n bn - ∑ s k ( bk +1 - bk )
k=1
n-1
= sn bn + ∑ sk ( bk - bk+1 )
k=1

Since b n ↓ 0 as n → ∞, taking the limit of this equality, we have

∞ ∞
∑ a k b k = ∑ s k ( b k - b k +1 ).
k=1 k =1

8. Suppose that { ak } and { bk } are real sequences such that

∣ ∑ b ∣≤
∞ n
lim a k = 0, ∑ ∣a k + 1 - a k∣ < ∞, and k M, n ∈ N.
k→∞ k=1 k=1

∞
Prove that ∑ a k b k converges.
k =1

n
Proof) Let A n, m = ∑ b k.
k=m

Then,

∣∑ b∣
n
∣A n, m∣ = k
k=m

=∣ ∑ b - ∑ b ∣
n m-1
k k
k=1 k =1

≤ ∣ ∑ b ∣+∣ ∑ b ∣
n m-1
k k
k=1 k=1

≤ 2M
And, since lim a k = 0, for given ε > 0, there is an N 1 ∈ N such that
k →∞

ε
n ≥ N 1 implies ∣a n∣< .
4M
∞
And, since ∑ ∣a k + 1 - a k∣< ∞, this implies that
k=1

the series is bounded and is increasing.

∞
Thus, ∑ ∣a k + 1 - a k∣ converges,
k=1

∞
that is, ∑ ( a k +1 - a k ) converges absolutely.
k =1

By Cauchy Criterion, for every ε > 0, there is an N 2 ∈ N such that

∣ ∑ (a ∣
n -1
ε
n - 1 ≥ m ≥ N 2 imply k +1 - ak ) < .
k=m 4M

∣∑ - a )∣ = ∣a
n -1
ε
In detail, ( ak +1 k n - a m∣< .
k=m 4M
If n > m ≥ max { N 1, N2 }, then

These solutions might be inexact, yet do you rely on them?

- 41 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

∣∑ ∣ = ∣A ∣
n n -1
ak bk n, m an - ∑ A k, m ( ak +1 - a k )
k=m k=m

∣ ∣
n-1
≤ ∣A n, m a n ∣+ ∑ A k, m ( ak+1 - ak )
k=m

∣ - a )∣
n -1
ε
< 2 M⋅
4M
+2M ∑ ( a k +1 k
k=m

ε
= + 2 M ∣a n - a m∣
2
ε ε
< + = ε
2 2
∞
By Abel’s Test, ∑ a k b k converges.
k=1

∞ ∞
9. Suppose that ∑ a k converges. Prove that if b k ↑ ∞ and ∑ a k b k converges,
k=1 k=1

then
∞
bm ∑ ak → 0
k =m

as m → ∞.

e
6.5 Estimation of Series

p.181
Exercises
1. For each of the following series, let s n represent its partial sums and s its
value. Prove that s is finite and find an n so large that s n approximates s to an

accuracy of 10 - 2.

( )
∞
π
(a) ∑ (-1 ) k 2
- arctan k .
k =1

n
π
Proof and Solution) Let a k = - arctan k, s n = ∑ (-1 ) k a k,
2 k=1

π
s = lim s n and f ( x ) = - arctan x.
n →∞ 2
π
Since f ( x ) = - arctan x → 0 as x → ∞ and
2
1
f '( x ) = - < 0 is negative for all x ∈ R,
1 + x2
it follows that a k ↓ 0 as k → ∞.
Thus the given series converges by the Alternating Series Test.
By Theorem 6.38 if f (n ) < 10-2, then

These solutions might be inexact, yet do you rely on them?

- 42 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
sn will estimate s to an accuracy of 10 - 2.

That is, n > tan ( π

2
- 10 - 2 ) ≈ 99.996666644444232802116380730872.
Hence it takes at most 100 terms
to approximate the value of this series to an accuracy of 10 - 2 .
∞
(-1 ) k k 2
(b) ∑ .
k =1 2k
k2 n
Proof and Solution) Let a k = , s n = ∑ (-1 ) k a k,
2k k=1

x2
s = lim s n and f (x ) = .
n →∞ 2x
Then by the L’hospital’s Rule,
x2 2x 2
lim f ( x ) = lim = lim = lim = 0.
x→∞ x →∞ 2x x
x → ∞ 2 log 2 x →∞ 2 x ( log 2 ) 2
x2
Since f (x ) = → 0 as x → ∞ and
2x
2x x2 log 2 x ( 2 - x log 2 )
f '( x ) = x - =
2 2x 2x
2
is negative for x > ,
log 2
it follows that a k ↓ 0 as k → ∞.
Thus the given series converges by the Alternating Series Test.
By Theorem 6.38 if f (n ) < 10-2, then
sn will estimate s to an accuracy of 10 - 2.
14 2
Since f (14 ) = = 0.011962890625
2 14
15 2
and f ( 15 ) = = 0.006866455078125,
2 15
it takes at most 15 terms
to approximate the value of this series to an accuracy of 10 - 2 .
∞
(-1 ) k 2⋅4⋅⋯⋅( 2 k )
(c) ∑ 1⋅3⋅⋯⋅( 2 k - 1 )
.
k=1 k2
n
1 2⋅4⋅⋯⋅( 2 k )
Proof and Solution) Let a k = , s n = ∑ (-1 ) k a k
k2 1⋅3⋅⋯⋅( 2 k - 1 ) k =1

s = lim s n and f ( n ) = a n.
n →∞

Since

These solutions might be inexact, yet do you rely on them?

- 43 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
1 2⋅4⋅⋯⋅( 2 k )
ak = 2
k 1⋅3⋅⋯⋅( 2 k - 1 )
1 2 4 ⋯ 2 k -2
= 2k
k2 3 5 2 k -1
2
<
k
and
a k +1 k2 2k+2 2 k2
= = < 1,
ak ( k + 1 )2 2k+1 2
2 k +3 k +1
it follows that a k ↓ 0 as k → ∞.
Thus the given series converges by the Alternating Series Test.
By Theorem 6.38 if f (n ) < 10-2, then
sn will estimate s to an accuracy of 10 - 2.
1 2⋅4⋅⋯⋅( 2 n )
f (n ) =
n2 1⋅3⋅⋯⋅( 2 n - 1 )
1 2n n !
=
n2 1⋅2⋅⋯⋅( 2 n )
2⋅4⋅⋯⋅( 2 n )
1 2n n !
=
n2 (2n )!
2n n !
2n
1 2 ( n ! )2
=
n2 (2n )!
Since f ( 31) = 0.0103105897534055 and f ( 32) = 0.00982983804863363,
it takes at most 32 terms
to approximate the value of this series to an accuracy of 10 - 2 .

2. (a) Find all p ≥ 0 such that the following series converges.

∞
1
∑ p .
k =1 k log ( k + 1 )
Proof) See 6.2Exercises 3.
Hence p > 1.
(b) For each such p, prove that the partial sums of this series s n and its value s
satisfy

∣s - s n∣≤
n+p -1
n ( p - 1) ( 1
log p -1 ( n ) )
for all n ≥ 2.
1
Proof) Let f ( x ) = .
x log p ( x + 1 )
1 p
Since f ' ( x ) = - p - ≤ 0
2
x log ( x + 1 ) x ( x + 1 ) log p +1 ( x + 1 )
for x > 1, f is positive and decreasing on [ 1, ∞ ).

These solutions might be inexact, yet do you rely on them?

- 44 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
n ∞
Fix p > 1. Set s n = ∑ f ( k ) and s = ∑ f ( k ).
k =1 k =1

By the second of Theorem 6.35,

∞
since ∑ f ( k ) converges for p > 1,
k =1

n ∞ ∞
0 ≤ ∑ f (k ) - ∑ f ( k ) ≤ f (n ) - ⌠
⌡n f ( x ) dx
≤ f (n ).
k =1 k =1

We have that ∣s - s n∣≤ f ( n ).

If n = 2, to prove that ∣s - s n∣≤ f ( n ) ≤

n +p -1
n ( p -1) ( 1
log p -1 ( n ) ),
we need to verify that

f (2) ≤
2+p -1
2 ( p -1) ( 1
log p -1 2 ).
Since

f (2) -
2+p -1
2 ( p -1) ( 1
log p -1 2 )
=
1
p
2 log ( 2 + 1 )
-
2+p -1
2 ( p -1) ( 1
log p - 1 2 )
( p - 1 ) log p -1 2 - ( p + 1 ) log p 3
=
2 ( p - 1 ) log p 3 log p -1 2
p ( log p -1 2 - log p 3 ) - ( log p -1 2 + log p 3 )
=
2 ( p - 1 ) log p 3 log p -1 2
≤ 0

it is true that ∣s - s n∣≤

n +p-1
n ( p -1) ( 1
log p -1 ( n ) ) for n = 2.

And prove this theorem for n ≥ 3.

If n ≥ 3 and p > 1, then log n > 1 and p - 1 > 0.
1
≤ 1 < 1+
n n +p -1
Thus = .
log n p -1 p -1
Then
1
f (n ) = p
n log ( n + 1 )
≤ 1
n log p n
1
=
n ( log n ) ( log p -1 n )
≤ n +p -1
n ( p - 1) ( 1
log p -1 ( n ) )
We conclude from Theorem 6.35 that

∣s - s n∣≤ f ( n ) ≤
n +p -1
n ( p - 1) ( log
1
p -1
(n ) ) for all n ≥ 2.

These solutions might be inexact, yet do you rely on them?

- 45 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
3. For each of the following series, let s n represent its partial sums, s represent
its value. Prove that s is finite and find an n so large that sn approximates s to

an accuracy of 10 - 2 .
∞
1
(a) ∑ k!
.
k=1

∞
1
(b) ∑ .
k =1 kk
∞ k
2
(c) ∑ k! .
k =1

∞ k2
(d) ∑
k =1
( k
k +1 ) .

4. Prove Theorem 6.40 (ii ).

6.6 Additional Tests

p.182
Exercises
1. Using any test covered in this chapter, find out which of the following series
converge absolutely, which converge conditionally, and which diverge.
∞
3⋅5⋅⋯⋅( 2 k + 1 )
(a) ∑ 2⋅4⋅⋯⋅2 k
.
k=1

3⋅5⋅⋯⋅( 2 k + 1 ) 2 k +1
Solution) Since > > 1 for k ∈ N,
2⋅4⋅⋯⋅2 k 2k
∞
3⋅5⋅⋯⋅( 2 k + 1 )
by Comparison Test, ∑ 2⋅4⋅⋯⋅2 k
diverges.
k=1

∞
1⋅3⋅⋯⋅( 2 k - 1 )
(b) ∑ 5⋅7⋅⋯⋅( 2 k + 3 )
.
k=1

1⋅3⋅⋯⋅( 2 k - 1 ) 3
Solution) First, = .
5⋅7⋅⋯⋅( 2 k + 3 ) ( 2 k + 1) ( 2 k +3)
∞
1
Since ∑ converges, by Limit Comparison Test,
k =1 k2
∞
1⋅3⋅⋯⋅( 2 k - 1 )
∑ 5⋅7⋅⋯⋅( 2 k + 3 )
converges.
k=1

∞
1
(c) ∑ log log k .
k =2 ( log k )
1
Solution) Let a k = log log k .
( log k )

These solutions might be inexact, yet do you rely on them?

- 46 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

Set p = lim
log ( 1
∣a k∣ ).
k →∞ log k
( log ( log k )) 2
Then, p = lim .
k →∞ log k
( log ( log x ))2
Let f ( x ) = .
log x
By L’hospital’s Rule,
( log ( log x )) 2
lim f ( x ) = lim
x →∞ x →∞ log x
2 ( log ( log x ))
= lim
x →∞ log x
2
= lim
x →∞ log x
= 0
By Sequential Characterization of Limits,
( log ( log k )) 2
p = lim = 0.
k →∞ log k
∞
1
By Logarithmic Test, ∑ log log k diverges.
k =2 ( log k )
k

( ).
∞
k -1
(d) ∑
k =1 k
k
= ( ).
k -1
Solution) Let a k
k

log (
∣a ∣ )
1
k
Set p = lim .
k →∞ log k

Then, p = lim
-k log ( k -1
k ).
k →∞ log k

Let f (x ) =
-x log ( x -1
x ).
log x
By the L’hospital’s Rule,

These solutions might be inexact, yet do you rely on them?

- 47 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim

lim f ( x ) = lim
-x log
x(
x -1
)
x →∞ x →∞ log x

= lim
-x log (1- 1
x )
x →∞ log x
3
1 -
2
x
- log (1- 1
x ) -x ⋅
2
-
1
2
1-x
= lim
x →∞ 1
x
3 3
1 -
2 1 -
2
x - x
2 4
- 1 - 1 2

= lim
1-x
-
2
( 1-x -
2
)
x →∞ 1
-
x2
1
1 2 1
x -
4 2
= lim 1 2
x →∞
( 1 -x ) -
2

= ∞
By Sequential Characterization of Limits,

p = lim
-k log ( k -1
k ) = ∞.
k →∞ log k
k

( )
∞
k -1
By Logarithmic Test, ∑ converges absolutely.
k =1 k

2. For each of the following, find all values of p ∈ R for which the given series
converges absolutely, for which it converges conditionally, and for which it
diverges.
∞
- kp
(a) ∑ ke .
k=1

- kp
Solution) Let a k = k e . Then,
∣a k +1∣
r = lim
k →∞ ∣a k∣

= lim
k →∞
∣ ( k + 1 ) e -( k + 1 ) p
k e -k p ∣
= lim
k →∞
∣ k +k 1 e -p ∣
= e -p
∞
If e -p < 1, that is p > 0, then ∑ ke - kp
converges absolutely.
k=1

These solutions might be inexact, yet do you rely on them?

- 48 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
∞
-p - kp
And, if e > 1, that is p < 0, then ∑ ke diverges.
k =1

∞ ∞
- k ⋅0
Even if p = 0, ∑ ke = ∑ k = ∞.
k=1 k=1

∞
1
(b) ∑ p log k .
k =2 ( log k )
1
Solution) Let a k = p log k . Then,
( log k )

p ' = lim
log ( 1
∣a k∣ )
k →∞ log k
p ( log k ) log ( log k )
= lim
k →∞ log k
= lim p log ( log k )
k →∞

∞
1
If p > 0, then p ' = ∞, ∑ p log k converges absolutely.
k =2 ( log k )
∞
1
And, if p < 0, then p ' = -∞, ∑ p log k diverges.
k =2 ( log k )
∞ ∞
1
Even if p = 0, then ∑ 0⋅log k = ∑ 1 = ∞.
k =2 ( log k ) k =2

∞
1
Hence, ∑ p log k converges for p > 0.
k =2 ( log k )
∞
( p k )k
(c) ∑ k!
.
k =1

( p k )k
Solution) Let a k = . Then,
k!
∣a k +1∣
r = lim
k→∞ ∣a k∣
k
= lim
k→∞
∣( k +1
k ) p∣
= e∣p∣
1 ∞
( p k )k
If ∣p∣<
e
, then ∑ k!
converges.
k =1

1 ∞
( p k )k
And, ∣p∣>
e
, then ∑ k!
diverges.
k =1

∞
( p k )k 1
We do not know whether ∑ k!
converges or not for ∣p∣ =
e
.
k =1

Thus, we should make good use of Stirling’s Formula (p.447).

k!
According to the theorem, for large k, ≈ 1.
2π k k k e-k
1
If p = , then
e

These solutions might be inexact, yet do you rely on them?

- 49 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
k
( 1
e
k ) =
kk
k! ek k!
1 1
= ⋅
k! 2π k
2π k k k e -k
1
≈
2π k
∞
( p k )k 1
Thus, ∑ k!
diverges for p =
e
.
k =1

1 ∞
( p k )k ∞
(-1 ) k k k
Even if p = -
e
, then ∑ k!
=∑ .
k =1 k =1 e k k!
By Stirling’s Formula,
∞
(-1 ) k k k ≈ ∞ (-1 ) k
∑ ∑ .
k =1 e k k! k =1 2π k
∞
( p k )k
By Alternating Series Test, ∑ k!
converges conditionally.
k =1

In strict secrecy, this proof must not be completely exact.

3. (a) Prove that the Root Test applied to the series

∞
1
∑ log k
k =2 ( log k )
yields r = 1. Use the Logarithmic Test to Prove that this series converges.
1
Proof) Let a k = log k .
( log k )
1
k
If r = limsup k →∞ { 1
( log k ) log k } = limsup k →∞
1
log k
k
( log k )
log k
exists, then so does log r = -limsup k→∞ log ( log k ).
k
log x
Let f ( x ) = log ( log x ).
x
By the L’hospital’s Rule,
log x
lim f ( x ) = lim log ( log x )
x →∞ x →∞ x
log ( log x ) + 1
= lim
x →∞ x
1
= lim
x →∞ x log x
= 0
By the Sequential Characterization of Limits,
log k
log r = -limsup k→∞ log ( log k ) = 0.
k

These solutions might be inexact, yet do you rely on them?

- 50 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
0
Therefore, r = e = 1.
Since the Root Test is inconclusive when r = 1,
∞
1
we do not know whether ∑ log k converges or not.
k =2 ( log k )
Thus we should use the Logarithmic Test.

Set p = lim
log ( 1
∣a k∣ ).
k →∞ log k
log k
log ( log k )
Then p = lim = lim log ( log k ) = ∞.
k →∞ log k k →∞

∞
1
By Logarithmic Test, ∑ log k converges.
k =2 ( log k )
(b) Prove that the Ratio Test applied to the series
∞
1⋅3⋅⋯⋅( 2 k - 1 )
∑ 4⋅6⋅⋯⋅( 2 k + 2 )
k=1

yields r = 1. Use Raabe’s Test to Prove that this series converges.

4. Suppose that f : R → (0, ∞) is differentiable, f (x ) → 0 as x → ∞, and

x f '(x )
α : = lim
x →∞ f (x )
∞
exists. If α < -1, prove that ∑ f (k ) converges.
k =1

Proof) By the L’hospital’s Rule,

lim
log ( 1
∣f ( x )∣ ) = lim
- log f ( x )
x →∞ log x x →∞ log x
f '( x )
-
f (x )
= lim
x →∞ 1
x
x f '(x )
= - lim
x →∞ f (x )
= -α

> 1
By the Sequential Characterization of Limits,

lim
log ( 1
∣f ( k )∣ ) > 1.
k →∞ log k
∞
By the Logarithmic Test, ∑ f (k ) converges.
k =1

5. Suppose that { ak } is a sequence of nonzero real numbers and

These solutions might be inexact, yet do you rely on them?

- 51 -  Last Revised : March 30th, 2008.
 The Exercise Solutions of 『An Introduction To ANALYSIS Third Edition』of
William R. Wade, the author of this book.
Written by Gi-uk Kim
p = lim k
k →∞
( 1 -∣ ak +1
ak ∣)
∞
exists as an extended real number. Prove that ∑ a k converges absolutely when
k=1

p > 1.
Proof) Suppose that p > 1.
By the Density of Rationals and Irrationals,
there is a q ∈ R such that p > q > 1.

Since p = lim k
k →∞ ( 1 -∣ aa ∣),
k+1

there is an N ∈ N such that k ≥ N implies q < k ( 1 -∣ aa ∣).

k +1

Then ∣ aa ∣< 1 - qk
k +1

k
for large k.

∞
Thus by the Raabe’s Test, ∑ a k converges absolutely.
k=1

These solutions might be inexact, yet do you rely on them?

- 52 -  Last Revised : March 30th, 2008.