Session 1.

1 - The Limiting Value of a Function

In our Calculus 1 course we attempted the definition of a limiting value in the following words :

'A sequence of values u1, u2, u3, …, un ,… would tend to a limit L if the difference between u n and L can
be made smaller than any arbitrarily small value of ε that we care to state, i.e. | un- L | < ε, where ε →
0.'

The time has now come, however, for a more rigorous definition, which we will state as follows :

In other words, we can make the positive difference between f(x) and L as small as we like by taking x
sufficiently close to a.
Consider, for example, the function 3x + 1 as x → 2 (from above or below).

Approaching 2 'from above', we could obtain, e.g., 3 x 2.1 + 1, 3 x 2.01 + 1, 3 x 2.001 + 1, 3x 2.0001 + 1,
… , i.e. 7.3, 7.03, 7.003, 7.0003,… , whilst approaching 2 'from below' we could obtain 3 x 1.9 + 1, 3 x
1.99 + 1, 3 x 1.999 + 1, … , i.e. 6.7, 6.97, 6.997, … In either case, the limiting value would appear to be
7.

Note carefully that we are trying to find the limiting value of 3x + 1 as x approaches the value of 2.
Even though, in this case, the limiting value can be found by simply substituting x = 2, this is not the
object of the exercise ! Can we be sure that either of the above sequences would approach so close to 7
that the difference between a member of the sequence and '7' can be made as small as possible ?

Suppose we wish to make the difference less than 10-60 , i.e. make | 3x + 1 − 7 | < 10-60 .
This gives |3x − 6| < 10-60 , i.e. 3 | x − 2 | < 10-60 and | x − 2 | < 1/3 x 10-60 . Therefore a value of δ (i.e.
1/3 x 10 -60) can be found which makes | f(x) − 7 | < 10-60 . It is not difficult to see that a corresponding
value of δ can be found which makes | f(x) − 7 | less than any arbitrary small value of ε.

Taking | f(x) − 7 | < ε, then | 3x − 6 | < ε and | x − 2 | < 1/3 ε. Therefore δ = 1/3 ε and can obviously be
found for any stated value of ε.
Example: Prove that lim(3x − 7) = 5 , using the formal proof.
x →4

The definition of a limit requires that for any number ε > 0 , we need to find a δ > 0 so that the
following is true.

(3 x − 7) − 5 < ε whenever x − 4 < δ

The first inequality can be simplified a little however:

3x − 7 − 5 < ε
3 x − 12 < ε
3( x − 4) < ε
3 x − 4 <ε
3 x − 4 <ε
ε
x−4 <
3

So, the definition of the limit requires that,

ε
x−4< whenever x − 4 < δ
3

ε ε
Therefore, if we choose δ = , we will get: (3 x − 7) − 5 < ε , whenever x − 4 <
4 3

Hence, lim(3 x − 7) = 5
x →4
Theorems/Calculation of limits
As we shall now discover, however, the calculation of limits becomes much easier when we establish
certain theorems on limits. The following theorems relate to the results

Theorem 1

To prove this result we need to find a value δ such that, when | x − a | < δ, then

| f(x) + g(x) − P − Q | < δ for any arbitrarily small δ.

If | f(x) + g(x) − P − Q | < δ, then | f(x) − P + g(x) − Q | < δ.

But | f(x) − P + g(x) − Q | ≤ | f(x) − P | + | g(x) − Q |. Therefore, as the limiting values

(as x → a) of f(x) and g(x) are P and Q respectively, then | f(x) − P | < ½ δ for

| x − a | < δ1 and | g(x) − Q | < ½ δ for | x − a | < δ 2 . [ Remember that we can make

| f(x) − P | < any arbitrarily chosen small value ].

Therefore | f(x) + g(x) − P − Q | < ½ δ + ½ δ, i.e. | f(x) + g(x) − P − Q | < δ for

| x − a | < δ, where δ is the smaller of δ 1 , δ 2 , i.e. the result is proved.

Therefore
Theorem 2

As | f(x) − P | < | ε/k | when | x − a | < δ, then | k | | f(x) − P | < ε, (ε > 0), when | x − a | < δ,

i.e. | kf(x) − kP | < ε when | x − a | < δ. Therefore

Theorem 3

Let f(x) = P + F(x), g(x) = Q + G(x). Therefore, as the limits of f and g (as x→ a) are P

and Q, F(x) → 0 and G(x) → 0 as x → a.

Now, f(x).g(x) = (P + F(x)).(Q + G(x)) = PQ + P.G(x) + Q.F(x) + F(x).G(x).

Theorem 4

As | g(x) − Q | is less than any arbitrarily small value ε for | x − a | < δ, then

| g(x) − Q | < ε | Q g(x) | when | x − a | < δ and ε is small.

Theorem 5

Examples: Find the following limits , given that lim(mx + c ) = ma + c and using any of the Theorems 1
x →a
to 6:

lim (3x − 2) 3
2
(i)
x→−2

 x 3 + 3x 
lim 
x →0
(ii)  x 

( )
1

(iii) lim x 2 + 3x − 6 3

x →−2

Solutions:
lim (3 x − 2 )3
2

x → −2
(i)
[( ) ]
2
= 3 lim x − lim 2 3
x →−2 x → −2
2
= [3(− 2 ) − 2] 3

( )
2
= [− 8] =
2
3 3
−8 =4
  x 3 + 3x 
lim 
x →0
(ii)  x 
x( x 2 + 3)
= lim
x →0 x
= lim( x + 3)
2

( )
x →0

= lim x 2 + lim 3
x →0 x →0

=0 +3=3
2

( )
1

(iii) lim x 2 + 3x − 6 3

x →−2

[( ) + 3(lim x)− lim 6]

1
= lim x
2 3
x → −2 x →−2 x →−2

= [(− 2 ) + 3(− 2 ) − lim 6]

1
2 3
x →−2

= [4 − 6 − 6] = [− 8] 3 = −2
1 1
3
Session 1.2 - Convergent and Divergent Series
Have you ever wondered how a calculator gives the value of a function such as sin 20º or e2 so quickly
? Certainly the calculator cannot store all the billions of possible values of sin θº or ex , wasting
precious memory space ! No, the calculator is programmed to calculate the values of sin 20º, e2, etc.
How does it do this (and do it so quickly) ? The answer, in all probability, is that it uses a power series
method. As we will find out in the next unit, ex (for example), can be written as a series of ascending
powers of x in the form

Is this answer correct ? Check your calculator. How can we then obtain so accurate an answer by
taking only the first 12 terms of what is, in fact, an infinite series ? I'm sure you've noticed that the
terms become smaller and smaller with higher powers of x, but is this a sufficient reason ? Will all the
millions of terms that come after

has a limiting value L , i.e. a value which the sum approaches (however closely), but cannot reach !
The limiting value L in this case cannot be given as an exact decimal − let us just say that the sum
cannot even reach 7.3891. The series is therefore said to converge to L, and this is the reason why we
can find its sum to any given degree of accuracy.
A Necessary Condition for Convergence
If the sum Sn of a series of n terms has a limiting value L as n →∞, then Sn and Sn-1

will both tend to L as n → ∞, i.e. Sn-1 → Sn and S - Sn-1 → 0. But Sn - Sn-1 = un ,

where un is the nth term of the series. Therefore un → 0.

A necessary condition for convergence is, therefore, that

un → 0 as n → ∞ .

This is clearly what we would expect for a series to have a limiting value. It does not follow,
however, that if un → 0, then the series must converge , i.e. the condition that un → 0 is
necessary but not sufficient.

therefore clearly does not converge (even though un → 0 as n → ∞).

The major question is then, "Given that un → 0 as n → ∞, how can we determine whether

or not a series converges ? The answer to this question will now be discussed at length.

Formula is known for the sum of the first n terms

One obvious means of determining whether or not a series converges is to find a formula for the sum
of the first n terms (i.e. the nth partial sum). Consider, for example, the geometric series

1 + ½ + ¼ + ? + … + (½)n-1 + …

If we denote the sum of the first n terms by Sn , i.e. S1 = 1, S2 = 1½, S3 = 1¾, etc, the sequence of
partial sums 1, 1½, 1¾, 1?, … is obtained. It certainly looks like the sequence converges to a value of 2.
Re-writing the sequence makes it more obvious,

i.e. 1, 2 − ½, 2 − (½)², 2 − (½)³, 2 − (½)4 , …

We can now see that the nth term of the sequence (i.e. Sn ) would be 2 − (½)n-1 , and this clearly tends
to a limiting value of 2 as n → 8.

Example 1

Will the series with the following nth partial sums converge or diverge ? If the series converges, find
its limiting value.

You might now be saying to yourself, "This isn't much of a problem, is it ?"

However, I'm afraid you would be jumping to conclusions a bit ! In practice, it is not

usual to be given a formula for the nth partial sum and, if it is not given, it is quite

difficult to figure it out. We would then have to resort to other means to determine

whether a particular series converges or diverges.

Comparison Test 1
This test (in part) states that

A series of positive terms converges if its terms are less

than (or equal to) the corresponding terms of some
known convergent series.

Let Σ ur and Σ vr be two series of positive terms and suppose that Σ vr is convergent

with a limiting value of V. Therefore, if ur≤ vr for all values of r, Σ ur must be ≤ V and

hence the series Σ ur is convergent.

This result, simple though it sounds, is a very useful one indeed. For example, we have

There is also a corresponding result regarding divergent series, i.e.

A series of positive terms diverges if its terms are

greater than (or equal to) the corresponding terms of a
known divergent series.
Limit Comparison Test
This test states that

Therefore vn (L − ε) < un< vn (L + ε).

Therefore, as un< vn (L + ε), if Σ vr converges, so does Σ ur .

Similarly, as un > vn (L − ε), if Σ vr diverges, so does Σ ur . The result is therefore proved.

Example 2
An Important Series

The Ratio Test

The test states that
You may be thinking at this time, "What happens when p = 1 ?" Well, I'm afraid that the

Ratio Test does not cover this (quite popular) situation and each test where the limiting

value = 1 needs to be considered on its merits.

Proof of the Ratio Test Result

Therefore, for sufficiently large values of n, un+1 > un , i.e. un cannot → 0 as n → ∞.

Therefore the series does not converge.

[ Remember, as we have said before, that this test does not cover the situation where

p = 1. Unfortunately, however, this particular situation occurs quite regularly !]

Example 3

Will the following series, with the given nth terms, converge or diverge ?

Therefore, as u n does not →0 as n → ∞, the series must diverge.

We will complete our discussion of the convergence or divergence of a series in the next

lesson. We will also determine the range of values of x for which the power series

expansion of a function has a limiting value (i.e. converges).

Session 1.3 - Convergence Tests (continued)
Alternate Positive and Negative Terms
If Σ ur is a series of positive terms such that ur+1 < ur for all r = 1, and ur → 0 as r → ∞,

then the series u1 – u2 + u3 - u4 + … + u2n-1 – u2n + … must be convergent .

This is not difficult to prove as

(i) if the series is written in the form (u1 – u2 ) + (u3 – u4 ) + … + (u2n-1 – u2n ) + …, we can

see that the sum of the series must be positive (as u1 > u2 , u3 > u4 , etc), and

(ii) writing the series in the form u1 – (u2 – u3 ) – (u4 – u5 ) - … , we can see that the

sum must be less than u1 .

The condition that ur → 0 as r → ∞ is also clearly necessary for convergence to be possible.

It therefore follows that the series u1 – u2 + u3 – u4 + … is convergent with a limiting value < u1.

Absolute and Conditional Convergence

If we let Σ ur be a convergent series of positive terms, and then change the signs of some of these
terms to give a new series Σ vr , we can easily show that the new series Σ vr must also be convergent. In
fact, the series Σ vr is said to be absolutely convergent .

The result follows quite easily as, if the limiting value of the series Σ ur is L, then - L < Σ vr< L , making
the series Σ vr convergent also.

Example 1
Example 2

Explain why the series 1 – ½ + 1/3 - ¼ + … is not absolutely convergent.

The series would be convergent as the terms are alternately positive and negative,

| un+1 | < | un | for all n = 1, and un → 0 as n → 8.

However, the series is not absolutely convergent as the series obtained by taking the

absolute (or positive) values, i.e. 1 + ½ + 1/3 + ¼ + … is not convergent.

A series such as 1 – ½ + 1/3 - ¼ + … is then said to be conditionally convergent .

Example 3
 Summary of Unit 1
Limit Theorems
Assume lim f ( x ) and lim g ( x ) both exist and c is any arbitrary constant, then:
x→a x→a

1. lim c f ( x ) = c lim f ( x )
x→a x→a

2. lim{ f ( x ) + g ( x) } = lim f ( x ) + lim g ( x )

x→a x→a x→a

3. lim{ f ( x ) g ( x) } = lim f ( x ) × lim g ( x)

x→a x→a x→a

 f ( x)  lim
x→a
f ( x)
4. lim   = , provided lim g ( x) ≠ 0
x→a x →a
 g ( x)  lim
x→a
g ( x)

n
lim[ f ( x)] = lim f ( x)
n
5.
x→a  x → a 
Convergence/Divergence Tests
Divergence Test

If lim an ≠ 0 then
n→∞
∑a n will diverge.

Comparison Test 1

Suppose that we have two series ∑a n and ∑b

n with an , bn ≥ 0 for all n and an ≥ bn for

all n. Then,

1. If ∑b n is convergent then so is∑a , n

2. If ∑ a n is divergent then so is ∑ b . n

Limit Comparison Test

Suppose that we have two series ∑a n and ∑b

n with an , bn ≥ 0 for all n and an ≥ bn for

all n. Then,

a 
c = lim n 
x→a b
 n 

If c is positive (i.e. c > 0) and is finite (i.e. c < ∞ ) then either both series converges or both diverge.

Ratio Test

a n +1
Suppose we have the series ∑a n . Define: L = lim
n→∞ an

Then ,

1. if L < 1, the series is absolutely convergent ( and hence convergent)

2. if L > 1, the series is divergent.
3. if L = 1, inconclusive