MATH2201/MATH2241

Introduction to Mathematical Analysis

Dr. Yat-Ming Chan
Department of Mathematics
The University of Hong Kong
Second Semester 2013-14
Chapter Outline
Chapter 1. The Real Numbers
Chapter 2. Sequences and Series
Chapter 3. Continuity
Chapter 4. Dierentiation
Chapter 5. The Riemann Integral
Reference Books
Main Reference
1. Class lecture notes This course will follow closely with the lecture notes which can be down-
loaded from the course website/moodle. You are expected to have the relevant notes during the
lectures.
2. Kenneth A. Ross, Elementary Analysis: The Theory of Calculus, Springer-Verlag
3. Robert G. Bartle, Donald R. Sherbert, Introduction to Real Analysis, Wiley
Suggested Readings
1. Patrick M. Fitzpatrick, Advanced Calculus: A Course in Mathematical Analysis, Boston: PWS
Pub. Co.
2. Walter Rudin, Principles of Mathematical Analysis, McGraw-Hill International
3. Michael Spivak, Calculus, Cambridge University Press
Ch1/MATH2201/2241/YMC/2013-14/2nd 1
Chapter 1. The Real Numbers
The foundation of calculus rests upon the concept of real numbers, so let us begin with a brief
description of the properties of the real numbers. Since you are probably familiar with most of such
properties, the aim of this chapter is to review the important features and persuade you that all the
properties could be traced back to only a few basic ones, namely, the eld axioms, the order axioms
and the completeness axiom.
1.1. The Field Axioms
On the set R of real numbers there are two binary operations, denoted by + and , called addition
and multiplication respectively. By a binary operation on a set S we mean for every ordered pair of
elements (a, b) S S we can associate a unique element a b in S. This suggests that for every
ordered pair (a, b) R R, we can associate a unique real number a + b, called the sum of a and b,
and a unique real number a b, called the product of a and b. We will assume that these two binary
operations on R satisfy the axioms listed in this section.
We begin with the axioms for addition:
(A1) a + b = b + a for all a, b R (Commutative Law for Addition)
(A2) a + (b + c) = (a + b) + c for all a, b, c R (Associative Law for Addition)
(A3) There exists an element of R, called an additive identity and denoted by 0, such that a+0 = a
for any a R (Existence of an Additive Identity)
(A4) For any a R, there exists an element a R, called an additive inverse or a negative of
a, such that a + (a) = 0 (Existence of an Additive Inverse/Negative)
We will rst establish some basic facts which are easy consequences of (A1)-(A4):
Proposition 1.1 (Uniqueness of Additive Identity) If x is an element of R such that a+x = a for
any a R, then x = 0.
Proof.
x = x + 0 by (A3)
= 0 + x by (A1)
= 0 by assumption with a = 0
2
Ch1/MATH2201/2241/YMC/2013-14/2nd 2
Proposition 1.2 (Uniqueness of Additive Inverse) Let a R. If x is an element of R such that
a + x = 0, then x = a.
Proof.
x = x + 0 by (A3)
= x + (a + (a)) by (A4)
= (x + a) + (a) by (A2)
= (a + x) + (a) by (A1)
= 0 + (a) by assumption
= (a) + 0 by (A1)
= a by (A3)
2
Proposition 1.1 and 1.2 allow us to say the additive identity of R and the additive inverse (or
the negative) of an element of R.
Proposition 1.3 Let a R. We have (a) = a.
Proof. Since
(a) + a = a + (a) = 0 by (A1) and (A4)
and
(a) + ((a)) = 0 by (A4),
we conclude from Proposition 1.2 that (a) = a. 2
Exercise 1.4
(a) Let a R. Prove that if x, y are elements of R such that a + x = a + y, then x = y.
(b) Let a, b R. Prove that (a + b) = (a) + (b).
(c) Prove that 0 = 0.
Next we list the axioms for multiplication:
(M1) a b = b a for all a, b R (Commutative Law for Multiplication)
(M2) a (b c) = (a b) c for all a, b, c R (Associative Law for Multiplication)
(M3) There exists an element of R, called a multiplicative identity and denoted by 1, such that
a 1 = a for any a R (Existence of a Multiplicative Identity)
(M4) For any a = 0 in R, there exists an element a
1
R, called a multiplicative inverse or a
reciprocal of a, such that a a
1
= 1 (Existence of a Multiplicative Inverse/Reciprocal)
Ch1/MATH2201/2241/YMC/2013-14/2nd 3
Here are some simple consequences of (M1)-(M4):
Proposition 1.5 (Uniqueness of Multiplicative Identity) If x is an element of R such that ax = a
for any a R, then x = 1.
Proposition 1.6 (Uniqueness of Multiplicative Inverse) Let a = 0 in R. If x is an element of R
such that a x = 1, then x = a
1
.
Proposition 1.7 Let a = 0 in R. We have (a
1
)
1
= a.
We omit the proofs for these propositions which can be easily translated from those for Proposition
1.1, 1.2 and 1.3. It follows from Proposition 1.5 and 1.6 that we can talk about the multiplicative
identity of R and the multiplicative inverse (or the reciprocal) of a nonzero element of R.
Exercise 1.8
(a) Let a = 0 in R. Prove that if x, y are elements of R such that a x = a y, then x = y.
(b) Let a = 0, b = 0 and a b = 0 in R. Prove that (a b)
1
= a
1
b
1
.
(Remark: In fact there is no need to assume a b = 0 as we shall see later in Proposition 1.12)
(c) Prove that 1
1
= 1.
The following axiom, known as the distributive law, connects the two operations:
(D) For any a, b, c R, a (b + c) = a b + a c (Distributive Law)
Here are more consequences involving the distributive law:
Proposition 1.9 For any a, b, c R, (a + b) c = a c + b c.
Proof.
(a + b) c = c (a + b) by (M1)
= c a + c b by (D)
= a c + b c by (M1)
2
Ch1/MATH2201/2241/YMC/2013-14/2nd 4
Proposition 1.10 For any a R, a 0 = 0.
Proof. Since
a + a 0 = a 1 + a 0 by (M3)
= a (1 + 0) by (D)
= a 1 by (A3)
= a by (M3),
we conclude from Proposition 1.1 that a 0 = 0. 2
Proposition 1.11 We have (1) (1) = 1.
Proof. Since
(1) + (1) (1) = (1) 1 + (1) (1) by (M3)
= (1) (1 + (1)) by (D)
= (1) 0 by (A4)
= 0 by Proposition 1.10,
we conclude from Proposition 1.2 with a = 1 that (1) (1) = (1), which is equal to 1 by
Proposition 1.3 with a = 1. 2
The following shows an important property of multiplication of real numbers the product of two
real numbers is zero only when one (or both) of these numbers is zero:
Proposition 1.12 Let a, b R. If a b = 0, then a = 0 or b = 0.
Proof. It suces to assume a = 0 and prove that b = 0. If a = 0, then there exist a unique
multiplicative inverse a
1
in R. Multiplying a b by a
1
gives
a
1
(a b) = (a
1
a) b by (M2)
= 1 b by (M4)
= b by (M3).
On the other hand, since we have assumed that a b = 0,
a
1
(a b) = a
1
0 = 0
by Proposition 1.10. Hence we have b = 0. 2
Exercise 1.13
(a) Let a, b R. Prove that a (b) = (a b).
(b) Let a R. Prove that (1) a = a.
Ch1/MATH2201/2241/YMC/2013-14/2nd 5
Remarks
(a) The additive identity 0 and the multiplicative identity 1 are normally required to be distinct, that
is, we shall assume 0 = 1. (What would happen if 0 = 1?)
(b) Subtraction and Division are dened respectively by
a b := a + (b)
and
a/b := a b
1
provided that b = 0.
Note that division by 0 is excluded by (M4).
(c) From now on we will drop the use of the dot to indicate multiplication and write ab for a b.
(d) We will write a
1
= a, a
2
= aa, a
3
= (a
2
)a, and in general, a
n+1
= (a
n
)a for any n N.
Moreover, if a = 0, then we write a
0
= 1 and a
n
= 1/(a
n
) for any n N.
(e) Let S be a set with two binary operations + and . Substitute S in place of R in (A1)-(A4),
(M1)-(M4) and (D), and assume that + and satisfy all these axioms together with 0 = 1.
Then S is called a eld. Hence the set R of real numbers is an example of a eld. This is why
the axioms (A1)-(A4), (M1)-(M4) and (D) are known as the eld axioms.
1.2. The Order Axioms
Next we would like to capture the idea that the set of real numbers has a natural order with some
axioms. We assume that there is a nonempty subset P of R, called the set of positive real numbers,
satisfying
(P1) If a, b P, then a + b P.
(P2) If a, b P, then ab P.
(P3) If a R, then exactly one of the following holds:
a P, a = 0, a P.
The above axioms are known as the order axioms. Note that 0 / P by axiom (P3).
We introduce the following notations:
a > b means a b P a < b means b a P
a b means a b P {0} a b means b a P {0}
With these notations the notion of inequality between two real numbers is dened in terms of the
set P. Moreover, the order axioms can now be rewritten as
Ch1/MATH2201/2241/YMC/2013-14/2nd 6
(P1) If a > 0 and b > 0, then a + b > 0.
(P2) If a > 0 and b > 0, then ab > 0.
(P3) If a R, then exactly one of the following holds:
a > 0, a = 0, a < 0.
From this version of the order axioms, we see that (P3) divides R into three distinct types of
elements and hence it is usually called the Trichotomy Law. The Trichotomy Law also implies that for
a, b R exactly one of the following holds:
a > b, a = b, a < b.
The following are some consequences of the order axioms:
Proposition 1.14 We have 1 > 0.
Proof. Suppose it is not true that 1 > 0. Since we have assumed 1 = 0, it follows from (P3) that
1 > 0. Then
1 > 0 = (1)(1) > 0 by (P2)
= 1 > 0 by Proposition 1.11.
But 1 > 0 and 1 > 0 cannot be both true by (P3) and hence we arrive at a contradiction. 2
Proposition 1.15 Let a, b R. Then
a > b if and only if a < b.
Proof.
a > b a b > 0 by denition
(a) b > 0 by Proposition 1.3
b (a) > 0 by (A1)
a < b by denition.
2
Proposition 1.16 Let a, b, c R. Then
(a) a b or b a
(b) a b and b a = a = b
(c) a b and b c = a c
Ch1/MATH2201/2241/YMC/2013-14/2nd 7
Proof. The proofs for (a) and (b) are left as exercise. For (c), if a = b or b = c, then there is nothing
to prove. Hence we need only consider the case where a < b and b < c. Applying (A1)-(A4) we see
that
c a = (c b) + (b a)
and it follows from (P1) that c a > 0 as required. 2
Proposition 1.17 Let a, b, c R.
(a) If a < b, then a + c < b + c.
(b) If a < b and 0 < c, then ca < cb.
(c) If a < b and c < 0, then ca > cb.
Proof.
(a) If a < b, then (b + c) (a + c) = b a > 0. Hence a + c < b + c as required.
(b) If a < b and 0 < c, then c(b a) > 0 by (P2). But from (D) and Exercise 1.13 (a) we have
c(b a) = cb + c(a) = cb ca. Hence ca < cb as required.
(c) Similar to (b). 2
Exercise 1.18
(a) Let a R. Prove that a
2
0.
(b) Prove that if a b and c d, then a + c b + d.
(c) Prove that if a > 0, then a
1
> 0.
(d) Prove that if a, b > 0 and a < b, then b
1
< a
1
.
By the Trichotomy Law (P3), given any a R, exactly one of a > 0, a = 0 or a < 0 will hold.
Thus the function | | : R R given by
|a| =

a if a > 0
0 if a = 0
a if a < 0
is well-dened. This is known as the absolute value function. We leave some basic properties of the
absolute value function as exercise:
Exercise 1.19
(a) Prove that | a| = |a| for any a R.
Ch1/MATH2201/2241/YMC/2013-14/2nd 8
(b) Prove that |ab| = |a||b| for any a, b R.
(c) Let c 0. Prove that |a| c if and only if c a c.
(d) Prove that |a| a |a| for any a R.
The following inequality will be of particular importance:
Theorem 1.20 (Triangle Inequality) Let a, b R. Then |a + b| |a| +|b|.
Proof. One can prove the inequality case by case. Alternatively, from Exercise 1.19 (d), we have
|a| a |a| and |b| b |b|.
From Exercise 1.18 (b) above, inequalities can be added and hence we have
(|a| +|b|) a + b |a| +|b|.
It follows from Exercise 1.19 (c) that |a + b| |a| +|b|. 2
Exercise 1.21 Let a, b R.
(a) Prove that |a b| |a| +|b|.
(b) Prove that ||a| |b|| |a b|.
1.3. The Completeness Axiom
Recall that a set with two binary operations satisfying the eld axioms is called a eld. Any eld
that satises the order axioms is called an ordered eld. Thus R is an ordered eld. Another example
of an ordered eld is given by the set Q of rational numbers note that Q also satises the eld and
order axioms. However, the set Q is not a nice ordered eld to work with in the sense that it has many
missing points or gaps for example, there is no r Q such that r
2
= 2. Thus Q is lacking in some
ways and the goal of this section is to look at an essential property of R a property that is not shared
by the rational numbers. In order to describe this special property and to handle the notion of gaps
better, we begin with the following denitions:
Denition 1.22 Let A be a nonempty subset of R.
1. The set A is bounded above if there exists u R such that a u for all a A. Each such u is
called an upper bound of A.
2. The set A is bounded below if there exists l R such that l a for all a A. Each such l is
called a lower bound of A.
3. The set A is bounded if it is bounded above and below, otherwise it is said to be unbounded.
Ch1/MATH2201/2241/YMC/2013-14/2nd 9
Example 1.23
1. Let A = {a R : 1 a < 5}. Then A is bounded above and 5, 9, 20.123 are examples of upper
bounds; it is also bounded below and 1, 3 are examples of lower bounds. It follows that the set
A is bounded.
2. The set B = {b R : b < } is bounded above but not below, and hence it is unbounded.
3. The set N of natural numbers is bounded below. The obvious fact that N is not bounded
above will be discussed later in Section 1.4.
Denition 1.24 Let A be a nonempty subset of R.
1. The number u R is a supremum (or a least upper bound) of A if
(a) u is an upper bound of A, and
(b) u u

for any upper bound u

of A.
2. The number l R is an inmum (or a greatest lower bound) of A if
(a) l is an lower bound of A, and
(b) l

l for any lower bound l

of A.
Although a set can have many upper bounds (or lower bounds), it can have only one supremum (or
inmum):
Theorem 1.25 Let A be a nonempty subset of R. If A has a supremum or an inmum, then these
numbers are unique.
Proof. If u
1
and u
2
are both suprema of A, then by 1(b) of Denition 1.24 we have u
1
u
2
and
u
2
u
1
. Hence u
1
= u
2
. One can use a similar argument to show the uniqueness for inmum. 2
Theorem 1.25 suggests that we can refer to the supremum and the inmum of a set A. We shall
write these numbers as sup A and inf A respectively.
Example 1.26 Let A = {a R : 0 a < 1}. Clearly 1 is an upper bound of A. Now suppose that
there is a smaller upper bound b. Then b < 1, and as it is an upper bound, we have b 0. Hence
1
2

b+1
2
< 1, which means
b+1
2
A. But
b+1
2
is greater than b, contradicting the fact that b is an
upper bound. We conclude that sup A = 1. Similar argument shows that inf A = 0. 2
Observe that sup A and inf A may or may not be elements of A. If it happens that the supremum
belongs to the set A, then it is called the maximum of A, and is denoted by max A. Similarly, if the
inmum belongs to A, then it is called the minimum of A, and is denoted by min A. Thus the supremum
(or inmum) of a set can exist and not be the maximum (or minimum), but when the maximum (or
minimum) exists then it must also be the supremum (or inmum). For example, the set A in Example
1.26 has no maximum, but 0 is its minimum (which is also its inmum).
Ch1/MATH2201/2241/YMC/2013-14/2nd 10
Remark If A is a nite subset of R, then both the maximum and minimum of A exist (one may use
induction to prove this fact).
By denition, to show that u = sup A, one has to show that u is an upper bound of A and that any
other upper bound of A must be greater than or equal to u. The following provides an equivalent way
of characterizing suprema no number smaller than the supremum can be an upper bound of the set.
Theorem 1.27 Let A be a nonempty subset of R and let u R be an upper bound of A. Then u =
sup A if and only if for any > 0, there exists a A such that u < a.
Proof. Suppose u = sup A. For any > 0, consider the number u . Since u < u, part 1(b)
of Denition 1.24 implies that u is NOT an upper bound of A. It follows that there must be some
element a A such that u < a.
Conversely, assume u is an upper bound of A satisfying the condition that for any > 0, there exists
a A such that u < a. If v is any number less than u, then we let = uv > 0. Hence there exists
a A such that v = u < a. This implies that v cannot be an upper bound of A. Thus if u

is an up-
per bound, then we must have u u

. This proves part 1(b) of Denition 1.24 and hence u = sup A. 2

The analogous characterization for inma is the following:
Theorem 1.28 Let A be a nonempty subset of R and let l R be a lower bound of A. Then l =
inf A if and only if for any > 0, there exists a A such that a < l + .
Proof. Exercise. 2
Example 1.29 Let A be a nonempty subset of R. Dene
A = {a : a A} = {b : b = a for some a A}.
Prove that
(a) A is bounded above by u if and only if A is bounded below by u.
(b) sup A exists in R if and only if inf (A) exists in R. When one of these numbers exists, we have
inf (A) = sup A.
Proof.
(a)
A is bounded above by u a A, a u
a A, a u
b A, b u
A is bounded below by u.
Ch1/MATH2201/2241/YMC/2013-14/2nd 11
(b) Suppose sup A exists in R. Since sup A is an upper bound of A, sup A is a lower bound of A
by (a). To show that inf (A) exists and equals sup A, we are going to apply Theorem 1.28,
that is, we will show that for any > 0, there exists b A such that b < sup A + . But
from Theorem 1.27, we have for any > 0, there exists a A such that sup A < a, or
equivalently, a < sup A + . Thus we are done if we take b = a. Similar argument shows
the other direction. 2
Let us turn our attention back to the essential property which distinguishes Q and R. The following
axiom is our nal axiom about R concerning the existence of suprema:
Completeness Axiom
Let A R be a nonempty set which is bounded above. Then sup A exists in R.
In other words, the Completeness Axiom asserts that every nonempty subset of R that has an upper
bound also has a supremum in R. The analogous statement for the existence of inma can be deduced
from the Completeness Axiom (there is no need to introduce another axiom):
Theorem 1.30 Let A R be a nonempty set which is bounded below. Then inf A exists in R.
Proof. Suppose that A is a nonempty subset of R that is bounded below. Example 1.29 (a) suggests
that the nonempty set A is bounded above. By the Completeness Axiom, sup A exists in R, and it
follows from Example 1.29 (b) that inf A exists in R (and is equal to sup (A)). 2
We mentioned earlier that Q is missing some numbers and is therefore not a nice ordered eld to
work with. We can now say more about these missing numbers in terms of the suprema of some subsets
of Q there are subsets of Q that are bounded above but for which no suprema exist in Q. It is exactly
these suprema that are missing from Q. As an example, we shall show later that the set
{x Q : x
2
< 2}
is one such subset. It follows that the ordered eld Q does not satisfy the Completeness Axiom, and
hence the completeness property is a property that we can use to distinguish Q from R. We say that
R is a complete ordered eld but Q is not.
Ch1/MATH2201/2241/YMC/2013-14/2nd 12
1.4. Consequences of Completeness
We will discuss in this section some important applications of the Completeness Axiom on R. The
rst one is a result that guarantees the existence of square roots.
The Existence of Square Roots
Theorem 1.31 There exists a unique positive real number such that
2
= 2.
Proof. Consider the set
A = {x R : x
2
< 2}.
Since 1
2
< 2, so 1 A and A is nonempty. Moreover, A is bounded above by 2 since if x > 2, then
x
2
> 4 and hence x / A. It then follows from the Completeness Axiom that A has a supremum in R
and we write = sup A. We remark that 1 > 0 is positive.
We are going to prove
2
= 2 by ruling out the possibilities
2
< 2 and
2
> 2.
Assume that
2
< 2:
The idea here is to nd a positive real number h which is small enough so that ( +h)
2
< 2. This
will then suggest that +h A and hence contradict the fact that is an upper bound of A. To see
how to choose h, we note that
( + h)
2
=
2
+ 2h + h
2
<
2
+ 2h + h =
2
+ 3h
if we require that 0 < h < . Furthermore,

2
+ 3h <
2
+ (2
2
) = 2
if we require that 0 < h <
2
2
3
(note that we are assuming
2
< 2 and hence
2
2
3
is positive).
Therefore, if we choose
h =
1
2
min

,
2
2
3

,
then we have h > 0, h < and h <
2
2
3
. It then follows from our construction of h that
( + h)
2
< 2,
which means we have constructed a number +h A, contradicting the fact that is an upper bound
of A. Hence the case
2
< 2 is not possible.
Assume that
2
> 2:
This time the idea is to nd a positive real number h which is small enough so that ( h)
2
> 2.
This will then suggest that there exists a A such that h < a (by Theorem 1.27) and hence
( h)
2
< a
2
< 2, contradicting our choice of h.
Ch1/MATH2201/2241/YMC/2013-14/2nd 13
To see how to choose h, we note that
( h)
2
=
2
2h + h
2
> 2 + h
2
> 2
if we require that
2
2h > 2, that is, h <

2
2
2
(note that we are assuming
2
> 2 and hence

2
2
2
is positive). Therefore, if we choose
h =
1
2

2
2
2

,
then we have 0 < h <

2
2
2
. It then follows from our construction of h that
( h)
2
> 2.
Now since h > 0 and = sup A, by Theorem 1.27, there exists a A such that h < a. But then
this implies (h)
2
< a
2
< 2, contradicting (h)
2
> 2. Hence the case
2
> 2 is also not possible.
Since it is not possible to have
2
< 2 and
2
> 2, by the Trichotomy Law, we must have
2
= 2.
The uniqueness part follows from Proposition 1.12: suppose is a positive real number such that

2
= 2, then

2
=
2
=
2

2
= 0 = ( + )( ) = 0.
We then conclude that = or = . But > 0 and so < 0, meaning that is the unique
positive number satisfying
2
= 2. 2
Remarks
(a) We write

2 or 2
1/2
for this unique .
(b) A small modication of the above proof can be applied to show that for any a 0, there exists
a unique real number, denoted by

a or a
1/2
, whose square is a. The number

a is known as
the positive square root of a.
(c) Using the binomial theorem one can also show that for any a 0, there exists a unique positive
n-th root of a, denoted by
n

a or a
1/n
, for each n N.
What happens if we replace the set A in Theorem 1.31 by the set of rational numbers
B = {x Q : x
2
< 2} ?
The set B consists only of rational numbers whose squares are all less than 2. It is certainly nonempty
and bounded above, and hence = sup B exists by the Completeness Axiom. If is rational, then the
same argument as in Theorem 1.31 suggests that
2
= 2. But this is impossible and hence the set B
does not have a supremum which belongs to Q. Thus we have found a nonempty subset of Q which is
bounded above but for which the supremum does not exist in Q. In other words, the ordered eld Q
does not satisfy the Completeness Axiom, as we have mentioned in the last part of the previous section.
Ch1/MATH2201/2241/YMC/2013-14/2nd 14
Archimedean Property
Another important consequence of the Completeness Axiom is known as the Archimedean Property
which basically states that the set N of natural numbers is not bounded above in R. Although this
property may seem obvious, it depends on the Completeness Axiom (together with the property that N
is closed under addition). After studying the Archimedean Property, we will also look at how the set Q
is sitting inside R the density of Q in R.
Theorem 1.32 (Archimedean Property) For any x R, there exists an n N such that x < n,
i.e. N is not bounded above.
Proof. Assume to the contrary that N is bounded above. By the Completeness Axiom, N should have
a supremum. We set = sup N. Now consider the number 1, which is smaller than the supremum
and so by Theorem 1.27, there exists an n N such that 1 < n. But then we have < n + 1.
Since n + 1 N, we have a contradiction to the fact that is an upper bound for N. 2
Corollary 1.33 For any real number > 0, there exists an n N such that
1
n
< .
Proof. This follows easily from Theorem 1.32 by letting x =
1

. 2
Remark: Corollary 1.33 is often refered to as the Archimedean Property as well.
Example 1.34 Let
A = {
1
n
: n N} = {a : a =
1
n
for some n N}.
Prove that inf A = 0.
Proof. Clearly, A is a nonempty subset of R bounded below by 0. We will use Theorem 1.28 to show
that 0 is the inmum of A. For any > 0, the Archimedean Property (Corollary 1.33) implies that
there exists n N such that
1
n
< . Thus for any > 0, there exists a A such that a < . It then
follows from Theorem 1.28 that inf A = 0. 2
Exercise 1.35
(a) Let A N be a nonempty set. Prove that if A is bounded above, then sup A A (and hence
max A exists and is equal to sup A). Similarly, prove that if A is bounded below, then inf A A
(and hence min A exists and is equal to inf A).
(b) For any x 0, prove that there exists m N such that m1 x < m.
(Hint: Consider the set A = {n N : x < n}. Use the Archimedean Property (Theorem 1.32)
to show that A is nonempty and then use (a) to construct m.)
The next result will concern how the set Q ts inside R:
Ch1/MATH2201/2241/YMC/2013-14/2nd 15
Theorem 1.36 (Density of Q in R) For any x, y R with x < y, there exists a rational number
r Q such that x < r < y.
Proof. Without loss of generality, we may assume that x 0. The case where x < 0 follows quickly
from this proof. The main idea here is to nd m, n N so that
x <
m
n
< y.
Since x < y, we have y x > 0 and it follows from the Archimedean Property (Corollary 1.33) that
there exists an n N such that
1
n
< y x, or equivalently, there exists an n N such that
nx + 1 < ny.
Next we apply Exercise 1.35 (b) to nx and so there exists m N such that m 1 nx < m. It
follows that
nx < m nx + 1 < ny,
so that nx < m < ny and therefore
x <
m
n
< y,
as desired. 2
Theorem 1.36 shows the betweeness property for Q in the sense that given any two real numbers
there is a rational number between them. We say that Q is dense in R. Without working too hard, we
can use this result to show that the set of irrational numbers is also dense in R:
Corollary 1.37 For any x, y R with x < y, there exists an irrational number s such that x < s < y.
Proof. Exercise (one may use the fact that

2 is irrational). 2
Nested Intervals Property
There is a natural class of subsets of R given by intervals. If a, b R with a < b, then we have
(a, b) = {x R : a < x < b} , [a, b] = {x R : a x b} ,
[a, b) = {x R : a x < b} , (a, b] = {x R : a < x b} .
Intervals of the type (a, b) are called open intervals with the endpoints being excluded in the set. In-
tervals of the type [a, b] are called closed intervals with the endpoints being included in the set. Intervals
using both square and round brackets are called half-closed intervals or half-open intervals. In the spe-
cial case a = b, then [a, b] = [a, a] = {a} is a singleton set, whereas the others are all equal to empty set.
So far these intervals are all bounded (both above and below). If a R, then
(a, ) = {x R : a < x} , [a, ) = {x R : a x} ,
(, a) = {x R : x < a} , (, a] = {x R : x a} .
These are all unbounded intervals: (a, ) and [a, ) have no upper bounds, and (, a) and (, a]
have no lower bounds. We sometimes write (, ) for R. Note that and are just convenient
Ch1/MATH2201/2241/YMC/2013-14/2nd 16
symbols, and they are NOT real numbers. Do not apply any theorem or property that is stated for real
numbers to and .
Now suppose we have a collection of intervals I
1
, I
2
, I
3
, . . . such that
I
1
I
2
I
3
I
k
I
k+1

that is, I
n+1
is a subset of I
n
for each n N. Such a collection of intervals is called a nested sequence
of intervals.
For example, if we let I
n
= [0, 1/n] for each n N. Then we have I
n
I
n+1
for each n N
and hence this is a nested sequence of intervals. We note that 0 is a point which belongs to every I
n
.
Moreover, 0 is the only common point to all I
n
(this follows from the Archimedean Property). We can
write this as

n=1
I
n
= {0}.
Thus the intersection of all the I
n
is nonempty (and is a single point in the above example). However,
the situation is completely dierent if the intervals in the sequence are not closed. For example, if
I
n
= (0, 1/n) for each n N. Then the intersection of all the I
n
in this nested sequence will be empty
(prove this!)
Now we would like to know under what assumptions on the intervals will the intersection of all
intervals be nonempty? The following result will provide us the answer, and the Completeness Axiom
plays an important role in the proof of it:
Theorem 1.38 (Nested Intervals Property) For each n N, if I
n
= [a
n
, b
n
] is a nested sequence
of closed and bounded intervals, then there exists x R such that x I
n
for all n N, that is, we
have

n=1
I
n
= .
Proof. Since I
n
= [a
n
, b
n
] is a nested sequence of intervals, we have
a
1
a
2
a
3
a
n
b
n
b
3
b
2
b
1
.
Consider the set A = {a
n
: n N} of left endpoints of the intervals: Thus each b
n
is an upper bound
for A. By the Completeness Axiom, the supremum x = sup A exists. We now claim that x I
n
for
all n N. Consider a particular I
n
= [a
n
, b
n
]. Since x is an upper bound for A, we have a
n
x. But
Ch1/MATH2201/2241/YMC/2013-14/2nd 17
b
n
is an upper bound for A and x is the least upper bound. Thus we have x b
n
. Hence x I
n
and since this is true for every n N, we conclude that x

n=1
I
n
and the intersection is nonempty. 2
Remark (and Exercise): We have seen an example earlier that if the nested intervals are not closed,
then the intersection is empty. Thus the closedness assumption cannot be removed from the theorem.
Similarly, the theorem is not true if the boundedness assumption is removed. Construct an example of
a nested sequence of unbounded intervals such that the intersection of all of them is empty.
Exercise 1.39
Construct an example of a nested sequence of closed and bounded intervals such that the intersection
of all of them is not a single point.