MA 105 D2 Lecture 2

Ravi Raghunathan

Department of Mathematics

August 10, 2023

Recap

Bounded Sequences

Cauchy sequences: the definition

Limits of functions
The rigourous definition of a limit

Definition: A sequence an tends to a limit l/converges to a limit l,

or that limn→∞ an = l, if for any  > 0, there exists N ∈ N such
that
|an − l| < 
whenever n > N.
We also stated rules for limits and the Sandwich Theorems and
proved the product rule. Once we have these rules, we can handle
more complicated limits.
Lemma: Every convergent sequence is bounded.
Theorem 3: A montonically increasing (resp. decreasing) sequence
which is bounded above (resp. below) converges to its supremum
(resp. infimum).
Bounded Sequences

The formulæ and theorems stated above can be easily proved

starting from the definitions. We will prove the second formula and
leave the other proofs as exercises.
Definition: A sequence an is said to be bounded if there is a real
number M > 0 such that |an | ≤ M for every n ∈ N. A sequence
that is not bounded is called unbounded.
In our list of examples, Example 1 (an = n) is an example of an
unbounded sequence, while Examples 2 - 5
(an = 1/n, sin(1/n), n!/nn , n1/n ) are examples of bounded
sequences.
Bounded sequences don’t necessarily converge - for instance
an = (−1)n . However,
Convergent sequences are bounded

Lemma: Every convergent sequence is bounded.

Proof: Suppose an converges to l. Choose  = 1. There exists
N ∈ N such that |an − l| < 1 for all n > N. In other words,
l − 1 < an < l + 1, for all n > N, which gives |an | < |l| + 1 for all
n > N. Let

M1 = max{|a1 |, |a2 |, . . . , |aN |}

and let M = max{M1 , |l| + 1}. Then |an | ≤ M for all n ∈ N.
In the slides presented in class, I had forgotten to put absolute
value signs in many places in the proof above and in the next slide.
This has now been corrected.
We will use this Lemma to prove the product rule for limits.
The proof of the product rule (if l2 6= 0)
We wish to prove that limn→∞ an bn = limn→∞ an · limn→∞ bn .
Suppose limn→∞ an = l1 and limn→∞ bn = l2 . We need to show
that limn→∞ an bn = l1 l2 .
Fix  > 0. We need to show that we can find N ∈ N such that
|an bn − l1 l2 | < , whenever n > N. Notice that

|an bn − l1 l2 | = |an bn − an l2 + an l2 − l1 l2 |
= |an (bn − l2 ) + (an − l1 )l2 |
≤ |an ||bn − l2 | + |an − l1 ||l2 |,

where the last inequality follows from the triangle inequality. So in

order to guarantee that the left hand side is small, we must ensure
that the two terms on the right hand side together add up to less
than . In fact, we make sure that each term is less than /2.
The proof of the product rule, continued
Since an is convergent, it is bounded by the lemma we have just
proved. Hence, there is an M such that |an | < M for all n ∈ N.
Given the quantities /2l2 and /2M, there exist N1 and N2 such
that
|an − l1 | < /2l2 and |bn − l2 | < /2M.
Let N = max{N1 , N2 }. If n > N, then both the inequalities above
hold. Hence, we have
   
|an ||bn − l2 | ≤ M · = and |an − l1 ||l2 | ≤ l2 · = .
2M 2 2l2 2
Now it follows that

|an bn − l1 l2 | ≤ |an ||bn − l2 | + |an − l1 ||l2 | < ,

for all n > N, which is what we needed to prove.

The proofs of the other rules for limits are similar to the one we
proved above. Try them as exercises.
Example
Let us look at Exercise 1.5.(i) which considers the sequence
 
1 2
a1 = 3/2 and an+1 = an + .
2 an

 
1 2
an+1 < an ⇐⇒ 2 an + an < an
√
⇐⇒ 2 < an .

On the other hand,

√
 
1 2
an + ≥ 2, (Why is this true?)
2 an
√ √
so an+1 ≥ 2 for all n ≥ 1 and a1 > 2 is given.
Hence, {a√ ∞
n }n=1 is a monotonically decreasing sequence, bounded
below by 2. By Theorem 3, it converges.
Exercise 1. What do you think is the limit of the above sequence
(Refer to the supplement to Tutorial 1)?
Cauchy sequences

As we have seen, it is not easy to tell whether a sequence

converges or not because we have to first guess what the limit
might be, and then try and prove that the sequence actually
converges to this limit. For a monotonic sequence, we have a
criterion, but what about more general sequences?
There is another very useful notion which allows us to decide
whether the sequence converges by looking only at the elements of
the sequence itself. We describe this below.
Definition: A sequence an in R is said to be a Cauchy sequence if
for every  > 0, there exists N ∈ N such that

|an − am | < ,

for all m, n > N.

Cauchy sequences: the theorem
Theorem 4: Every Cauchy sequence in R converges to a limit in R.
Remark 1: One can now check the convergence of a sequence just
by looking at the sequence itself!
Remark 2: One can easily check the converse:
Theorem 5: Every convergent sequence is Cauchy.
Remark 3: Remember that when we defined sequences we defined
them to be functions from N to X , for any set X . So far we have
only considered X = R, but as we said earlier we can take other
sets, for instance, susbets of R. For instance, if we take X = R \ 0,
Theorem 4 is not valid. The sequence 1/n is a Cauchy sequence in
this X but obviously does not converge in X . If we take X = Q,
the example given in 1.5.(i) (an+1 = (an + 2/an )/2) is a Cauchy
sequence in Q which does not converge in Q. Thus Theorem 4 is
really a theorem about real numbers.
The completeness of R

A set in which every Cauchy sequence converges (to a limit which

is also in the set) is called a a complete set. Thus Theorem 4 is
sometimes rewritten as
Theorem 4’: The real numbers are complete.
You will see other examples of complete sets in MA 110.
Sequences in R2 and R3

Most of our definitions for sequences in R are actually valid for

sequences in R2 and R3 . Indeed, the only thing we really need to
define the limit is the notion of distance. Thus if we replace the
modulus function | | on R by the distance functions in R2 and R3
all the definitions of convergent sequences and Cauchy sequences
remain the same.
For instance, a sequence a(n) = (xn , yn ) in R2 is said to converge
to a point l = (l1 , l2 ) (in R2 ) if for all  > 0, there exists N ∈ N
such that q
(xn − l1 )2 + (yn − l2 )2 < 

whenver n > N. A similar defintion can be made in R3 using the

distance function on R3 . Some of our earlier rules will make sense
and some won’t.
Infinite series - a more rigourous treatment
Let us recall what we mean when we write
a
a + ar + ar 2 + . . . = .
1−r
Another way of writing the same expression is
∞
X a
ar k = .
1−r
k=0

The precise meaning is the following. Form the partial sums

n
X
sn = ar k .
k=0

These
P∞ partial sums s1 , s2 , . . . sn , . . . form a sequence and by
k
k=0 ar = a/(1 − r ), we mean limn→∞ sn = a/1 − r .
So when we speak of the sum of an infinite series, what we really
mean is the limit of its partial sums.
Convergence of the geometric series
So to justify our formula we should show that limn→∞ sn = a/1 − r ,
that is, given  > 0, there exists N ∈ N such that
a
sn − < ,
1−r
for all n > N.
In other words we need to show that
a(1 − r n+1 ) a ar n+1
− = <
1−r 1−r 1−r
if n is chosen large enough.
But limn→∞ r n = 0, so there exists N such that r n+1 < (1 − r )/a
for all n > N, so for this N, if n > N,
a
sn − < .
1−r
This shows that the geometric series converges to the given
expression.
Limits of functions of a real variable
Since we have already defined the limit of a sequence rigourously,
it will not be hard to define the limit of a real valued function
f : (a, b) → R.
Definition: A function f : (a, b) → R is said to tend to (or
converge to) a limit l at a point x0 ∈ [a, b] if for all  > 0 there
exists δ > 0 such that
|f (x) − l| < 
for all x ∈ (a, b) such that 0 < |x − x0 | < δ.In this case, we write

lim f (x) = l,
x→x0

or f (x) → l as x → x0 which we read as “f (x)” tends to l as x

tends to x0 ”. This is just the rigourous way of saying that the
distance between f (x) and l can be made as small as one pleases
by making the distance between x and x0 sufficiently small.
A subtle point and the rules for limits

Notice that in the definition above, the point x0 can be one of the
end points a or b.
Thus the limit of a function may exist even if the function is not
defined at that point.
The rules and formulæ for limits of functions are the same as those
for sequence and can be proved in almost exactly the same way. If
limx→x0 f (x) = l1 and limx→x0 g (x) = l2 , then
1. limx→x0 f (x) ± g (x) = l1 ± l2 .
2. limx→x0 f (x)g (x) = l1 l2 .
3. limx→x0 f (x)/g (x) = l1 /l2 . provided l2 6= 0

As before, implicit in the formulæ is the fact that if the limits on

the left hand side exist. We prove the first rule below.
The proof of the addition formula for limits
Proof: We first show that limx→x0 f (x) + g (x) = l1 + l2 . Let  > 0
be arbitrary.
Since limx→x0 f (x) = l1 and limx→x0 g (x) = l2 , there exist δ1 , δ2
such that
 
|f (x) − l1 | < and |g (x) − l2 | <
2 2
whenever 0 < |x − x0 | < δ1 and 0 < |x − x0 | < δ2 . If we choose
δ = min{δ1 , δ2 } and if 0 < |x − x0 | < δ then both the above
inequalities hold. Thus, if |x − x0 | < δ, then

|f (x) + g (x) − (l1 + l2 )| = |f (x) − l1 + g (x) − l2 |

 
≤ |f (x) − l1 | + |g (x) − l2 | < + = ,
2 2
which is what we needed to prove. If we replace g (x) by −g (x) we
get the second part of the first rule.
The Sandwich Theorem(s) for limits of functions
Theorem 5: As x → x0 , if f (x) → l1 , g (x) → l2 and h(x) → l3 for
functions f , g , h on some interval (a, b) such that
f (x) ≤ g (x) ≤ h(x) for all x ∈ (a, b), then

l1 ≤ l2 ≤ l3 .

As before, we have a second version.

Theorem 6: Suppose limx→x0 f (x) = limx→x0 h(x) = l and If g (x)
is a function satisfying f (x) ≤ g (x) ≤ h(x) for all x ∈ (a, b), then
g (x) converges to a limit as x → x0 and

lim g (x) = l
x→x0

Once again, note that we do not assume that g (x) converges to a

limit in this version of the theorem - we get the convergence of
g (x) for free .
Some examples
Let us look at Exercise 1.11. We will use this exercise to explore a
few subtle points.
Let c ∈ [a, b] and f , g : (a, b) → R be such that limx→c f (x) = 0.
Prove or disprove the following statements.
(i) limx→c [f (x)g (x)] = 0.
(ii) limx→c [f (x)g (x)] = 0, if g is bounded.(g (x) is said to be
bounded on (a, b) if there exists M > 0 such that |g (x)| < M
for all x ∈ (a, b).
(iii) limx→c [f (x)g (x)] = 0, if limx→c g (x) exists.
Before getting into proofs, let us guess whether the statements
above are true or false.
(i) false
(ii) true
(iii) true.
(i) Notice that g (x) is not given to be bounded - if this was not
obvious before, you should suspect that such a condition is needed
after looking at part (ii). So the most natural thing to do is to
look for a counter-example to this statement by taking g (x) to be
an unbounded function. What is the simplest example of an
unbounded function g (x) on an open interval?
How about g (x) = x1 on (0, 1)?
What would a candidate for f (x) be - what is the simplest example
of a function f (x) which tends to 0 for some value of c in [0, 1].
f (x) = x, and c = 0 is a pretty simple candidate.
Clearly limx→0 f (x)g (x) = limx→0 1 = 1 6= 0, which shows that (i)
is not true in general.
Exercise 1: Can you find a counter-example to (i) with c in (a, b)
(that is, c should not be one of the end points)? (Hint: Can you
find an unbounded function on a closed interval [a, b]?)
Let us move to part (ii).
Suppose g (x) is bounded on (a, b). This means that there is some
real number M > 0 such that |g (x)| < M. Let  > 0. We would
like to show that

|f (x)g (x) − 0| = |f (x)g (x)| < ,

if 0 < |x − c| < δ for some δ > 0.
Since limx→c f (x) = 0, there exists δ > 0 such that |f (x)| < /M
for all |x − c| < δ. It follows that

|f (x)g (x)| = |f (x)||g (x)| < ·M =
M
for all 0 < |x − c| < δ, and this is what we had to show.
Part (iii) follows immediately from the product rule, but can one
deduce part (iii) from (ii) instead?
Hint: Think back to the lemma on convergent sequences that we
proved in Lecture 1: Every convergent sequence is bounded. What
is the analogue for functions which converge to a limit at some
point? Indeed, you can easily show the following
Lemma 7: Let f : (a, b) → R be a function such that limx→c f (x)
exists for some c ∈ [a, b]. If c ∈ (a, b), there exists an (open)
interval I = (c − η, c + η) ⊂ (a, b) such that f (x) is bounded on I .
If c = a, then there is an open interval I1 = (a, a + η) such that
f (x) is bounded on I1 . Similarly if c = b, there exists an open
interval I2 = (b − η, b) such that f (x) is bounded on I2 .
The proof of the lemma above is almost the same as the the
lemma for convergent sequences. Basically, replace “N” by “δ” in
the proof.
If one applies the Lemma above to g (x), we see that g (x) is
bounded in some (possibly) smaller interval (0, η). Now apply part
(ii) to this interval to deduce that (iii) is true.