Programme 12

SERIES
PART 2
 Programme 12

1 Power series
Introduction: In the first programme (No. 11) on series, we saw how
important it is to know something of the convergence properties of any
infinite series we may wish to use and to appreciate the conditions in
which the series is valid.
This is very important, since it is often convenient to represent a
function as a series of ascending powers of the variable. This, in fact, is
just how a computer fmds the value of the sine of a given angle. Instead
of storing the whole of the mathematical tables, it sums up the terms of
a series representing the sine of an angle.
That is just one example. There are many occasions when we have
need to express a function of x as an infinite series of powers of x.
It is not at all difficult to express a function in this way, as you will soon
see in this programme.

So make a start and turn on to frame 2.

2
Suppose we wish to express sine x as a series of ascending powers of x.
The series will be of the form
sin x =a + bx + ex 2 + dx 3 + ex 4 + ...
where a, b, c, etc., are constant coefficients, i.e. numerical factors of
some kind. Notice that we have used the 'equivalent' sign and not the
usual 'equals' sign. The statement is not an equation: it is an identity.
The right-hand side does not equal the left-hand side: the R.H.S. is the
L.H.S. expressed in a different form and the expression is therefore true
for any value of x that we like to substitute.
Can you pick out an identity from these?
(x + 4) 2 = 3x 2 - 2x + 1
(2x + 1) 2 = 4x 2 + 4x - 3
(x + 2) 2 = x 2 + 4x + 4
When you have decided, move on to frame 3.

327
Series 2

I(x + 2) 2 =x 2 + 4x + 41 3
Co"ect. This is the only identity of the three, since it is the only one in
which the R.H.S. is the L.H.S. written in a different form. Right. Now
back to our series:
sin x =a + bx + cx 2 + dx 3 + ex 4 + ...
To establish the series, we have to find the values of the constant coeffi-
cients a, b, c, d, etc.
Suppose we substitute x =0 on both sides.
Then sin 0 = a + 0 + 0 + 0 + 0 + ...
and since sin 0 = 0, we immediately get the value of a.
a= ................... .

4
Nowcan we substitute some other value forx, which will make all the
terms disappear except the second? If we could, we 'should then find the
value of b. Unfortunately, we cannot find any such substitution, so what
is the next step?
Here is the series once again:
sin x =a + bx + ex 2 + dx 3 + ex4 + ...
and so far we know that a= 0.
The key to the whole business is simply this:
Differentiate both sides with respect to x.
On the left, we get cos x.
On the right the terms are simply powers of x, so we get
cosx =................... .

328
 Programme 12

5 Icos x =b + c .2x +d. 3x 2 + e.4x 3 + ...

This is still an identity, so we can substitute in it any value for x we
like.
Notice that the a has now disappeared from the scene and that the
constant term at the beginning of the expression is now b.
So what do you suggest that we substitute in the identity as it now
stands, in order that all the terms except the first shall vanish?

We substitute x = .................... again.

6
ISubstitute x = 0 again I
Right: for then all the terms will disappear except the first and we shall
be able to find b.
cosx = b + c.2x + d.3x 2 + e.4x 3 + ...
Putx =0
cos 0 = I =b + 0 + 0 + 0 + 0 + ...
:. b =I
So far, so good. We have found the values of a and b. To find c and d
and all the rest, we merely repeat the process over and over again at each
successive stage.

i.e. Differentiate both sides with respect to x

and substitute ....................... .

329
Series 2

l substitute x =0 I 7
So we now get this, from the beginning:
sin x =a+ bx + cx 2 + dx 3 + ex 4 + fx 5 + ...
Put x =0. :. sin 0 =0 =a + 0 + 0 + 0 + . . . :. a =0

{
Diff. cos x = b + c.2x + d.3x 2 + e.4x 3 + [.5x 4 •••

Put X =0. .". COS 0 = 1 = b + 0 + 0 + 0 + . . . .". b = 1

Diff. -sinx = c.2 +d.3.2x + e.4.3x 2 + f.5.4x 3 •••
{
Put x = 0. :. -sin 0 = 0 =· c .2 + o + 0 + . . . :. c = o
f Diff. -::cos x = d.3.2.1 + e.4.3.2x + f.5.4.3x 2 •••

, Putx=O. :. -cos0=-1= d.3!+0+0+ ... :. d= -t!

And again. sin x = e.4.3.2.1 + f.5.4.3.2x + ...

(
Put x =0. :. sin 0 =0 = e.4! + 0 + 0 + :. e =0
Once more. cos x = {5.4.3.2.1 + ...
(
Put X = 0. ·:. COS 0 =1 = /.5!+0+ :.f=!!
etc. etc.

All that now remains is to put these values for the constant coeffi-
cients back into the original series.

sin x = 0 + l.x + O.x 2 + ~~! x 3 + O.x4 + ~! x 5 +

. . x3 xs
1.e. sm x =x - 3! + 5 ! - . . . . ..

Now we have obtained the first few terms of an infinite series represent-
ing the function sin x, and you can see how the terms are likely to
proceed.
Write down the first six terms of the series for sin x.

When you have done so, tum on to frame 8.

330
 Programme 12

8 . x 3 5 x 7 x 9 x 11
smx=x--x+---+--- +
3! 5! 7! 9! 11!
Provided we can differentiate a given function over and over again,
and find the values of the derivatives when we put x = 0, then this
method would enable us to express any function as a series of ascending
powers ofx.
However, it entails a considerable amount of writing, so we now
establish a general form of such a series, which can be applied to most
functions with very much less effort. This general series is known as
Maclaurin's series.

So turn on to frame 9 and we will find out all about it.

9 Maclaurin's series: To establish the series, we repeat the process of

the previous example, but work with a general function,f(x), instead of
sin x. The fir~t differential coefficient of f(x) will be denoted by f'(x );
the second by f"(x); the third by f"'(x); and so on. Here it is then:
Let f(x) = a + bx + cx 2 + dx 3 + ex 4 + fx 5 + ...
Put x = 0. ThenftO) =a+ 0 + 0 + 0 +... :. a= ftO).
i.e. a = the value of the function with x put equal to 0.
Diff. f'(x) = b + c .2x + d .3x 2 + e .4x 3 + f.5x 4 + ...
Put X = 0 .". f'(O) = b + 0 + 0 +. . . .". b =f'(O)
Diff. f"(x) = c.2.1 + d.3.2x + e.4.3x 2 + f.5.4x 3 •••

- 0
Put x- ... f"(O)-- c..
2 1 + 0 + 0 + . ..... c-~
_f"(O)

Now go on and fmd d and e, remembering that we denote

d~ {t!'Cx)}. by f"'(x)

and :X {t"'(x)} by [iV(x), etc.

So, d = .................... and e = ................... .

331
Series 2

[!"(0) [!V(O)
d =3!- ·, e=-,-
10
4.

Here it is. We had:

/''(x) =c.2.1 + d.3.2x + e.4.3x 2 + f.5.4x 3 + .. .
Diff. :. f!"(x) = d.3.2.1 + e.4.3.2x + f.5.4.3x 2 + .. .
[
l"'(O)
·Putx =0 :. f"'(O) = d.3! + 0 + 0... :. d.= ""'3!
Diff. :. pv(x) = e.4.3.2.1 + f.5.4.3.2x + ...
(
r . - pv(o)
Put X = 0 :. F(O) = e.4. +0+0+ ..... e- 7
etc. etc.

So a= ~(o)·, b =f'(O)·, c =f!I(O)

J 2! ,
· d =f!"(O) · e _Jiv(o) ·
3! , - ~· ...

Now, in just the same way as we did with our series for sin x, we put
the expressions for a, b, c, . . . etc., back into the original series and get:
f(x) =................... .

_
f(x) - f(O)
, f"{O)
+I (O).x + 2! .x2 + 3 ! .x3
f"'(O)
+ ... 11
and this is usually written as
2 3
f(x) = f(O) + x.f'(O) -r ~! ./"(0) + ~! ./"'(0) .... I

This is Maclaurin's series and important!

Notice how tidy each term is.
The term in x 2 is divided by 2! and multiplied by /"(0)
" " x3 " " " 3! " " " I '"(0)
" " " " 4! " "
Copy the series into your record book for future reference.

Then on to frame 12.

332
 Programme 12

12 Maclaurin's series
2 3
f(x) = f(O) + x. f'(O) + ~! . f"(O) + ~! .f"'(O) + ...

ODDODODDODOnDDODODDDDODODDDDDDDD DODDDu

Now we will use Maclaurin's series to find a series for sinh x. We have
to find the successive differential coefficients of sinh x and put x = 0 in
each. Here goes, then:
f(x) =sinhx f(O) = sinh 0 = 0
f'(x) = cosh x f'(O) =cosh 0 = 1
{"''(x) =sinhx =
f"(O) sinh 0 0 =
f"'(x) =cosh x f'"(O) .=cosh 0 = 1
fiV(x) = sinhx =
f!V(O) sinh 0 0 =
fY(x) =coshx fY(O) = cosh 0 = 1 etc.

13 Now let us find a series for ln(1 + x) in just the same way.
f(x) =ln(1 + x) :. f(O) = ................... .
1
f'(x)=-=(1 +x} 1
1+x
:. !'(O) =
-1
l"(x) =- (1 + x} 2 = - -2 :. /"(0) =
(1 + x)
f"'(x) = 2( 1 + x) 3
·
=(1- +2 x)
- 3 ... f"'(O) -- ....................
! W(x) = -3.2(1 + x)-4 =- (1.2d__
+x) 4
· fiv(o) --
•• •·••••••••••••••••••

- (14!+ x)s
fY(x)-- 4.3.2(1 + XJ 5- :. fY(O) = ................... .
You complete the work. Evaluate the differentials when x =0,
remembering that ln 1 = 0, and substitute back into Maclaurin's series to
obtain the series for ln(1 +x).
So, ln(l +x)= ....................

333
Series 2

f(O) = ln 1 = 0; f'(O) = 1; f'~(O) = -1; F"(O) = 2; 14

JiV(O) = -3!; fY(O) = 4!; ...
2 3
Also f(x) = /(0) + x.f'(O) + ~! f"(O) + ~! f!"(O) + ...
2 3 4
ln(l +x)=O+x.l +11(-1)+~!(2)+~!(-3!)+ ...

x x x x 2
ln(l +x)=x-::..._+---+--
3 4 5
2 3 4 5

Note that in this series, the denominators are the natural numbers, not
factorials!

Another example in frame 15.

15
Example: Expand sin 2x as a series of ascending powers of x.
Maclaurin's series.:
2 3
f(x) = f(O) + x.f'(O) + ~! .f'!{O) + ~! ./"'(0) + ...

:. f(x) = sin 2x f(O) = ··················

f'(x) = 2 sin x cos x =sin 2x f'(O) =
f"(x) = 2 cos 2x /"(0) =
f!"(x) = - 4 sin 2x F";(O) =
Jiv(x) =............... . [iV(O) = ................. .
There we are! Finish it off: find the first three non-vanishing terms of the
series.

Then move on to frame 16.

334
 Programme 12

16
For f(x) = sin 2x :. f(O) = 0
f'(x) = 2 sin x cos x =sin 2x :. f'(O) = 0
f"(x) = 2 cos 2x :. /';(0) =2
f"'(x) = -4 sm 2x :. /'"{0) : ;: 0
pv(x) = -8 cos 2x :. fiv(O) = -8
[V(x) = 16 sin 2x :. [V(O) = 0
[Vi(x) =32 cos 2x :. [Vi(O) =32 etc.
2 3
f(x) = f(O) + x .f'(O) + ~! .f"(O) + ~! .["!{0) + ...

• . 2 _ x2 x3 x4 x5 x6
.. sm x- 0 + x(O) + 2 ! (2) +3! (O) + 4 ! (-8) + S! (O) + 6 ! (32)

Jx 2x 6
• • 2 _ 2 4
. . Slfi X -X - + 4S ...
Next we will find the series for tan x. This is a little heavier but the method
is always the same.
Move to frame 17.

17 Series for tan x

f(x) =tan x :. f(O) = 0
:. f'(x) = sec 2 x :. f'(O) = 1
:. f "(x) = 2 sec 2x tan x :. f"(O) = 0
:. f"'(x) = 2 sec4 x + 4 sec 2x tan 2x :. f"'(O) = 2
= 2 sec4 x + 4(1 + tan 2x) tan 2x
= 2 sec 4 x + 4 tan 2 x + 4 tan4 x
:. fiv(x) = 8 sec4 x tan x + 8 tan x sec 2x + 16 tan 3x sec 2x
= 8(1 + t 2 ) 2 t + 8t(l + t 2 ) + 16t\1 + t 2 )
= 8(1 + 2t 2 + t 4 )t + 8t + 8t 3 + 16t3 + 16t 5
= 16t + 40t 3 + 24t 5 .'. fiV(O) = 0

:. [V(x) = 16 sec 2x + 120t 2 .sec 2x + 120t4 sec 2x

:. fY(O) = 16
:. tan x = ............. .

335
Series 2

Standard series
18
By Maclaurin's series. we can build up a list of series representing many
of the common function.> - we have already found series for sin x, sinh x
and ln(l + x).
To find a series for cos x, we could apply the same technique all over
again. However, let us be crafty about it. Suppose we take the series for
sin x and differentiate both sides with respect to x just once, we get
. x3 x 5 x7
Sill X =X - 3 ! + S! - ? ! +

Diff. cos x
3x
=1 - 3f Sx 7! . . .
+ Sf - 7x
2 4 6
etc.

In the same way, we can obtain the series for cosh x. We already know
that

so if we differentiate both sides we shall establish a series for cosh x.

What do we get?

19
We get:

Diff.

giving:

Let us pause at this point and take stock of the series we have obtained.
We will make a list of them, so turn on to frame 20.

336
 Programme 12

20 Summary
Here are the standard series that we have established so far.
x3 xs x7 x9
sinx =x-3!+5!-7!+9! ·· · II

x2 x4 x6 xs
cosx = 1 - 2! + 4!- 6! + 8! ... III

x3 2xs
tanx =x+-+-+ IV
3 15
. x3 xs x7
smhx =x+ 3 ! + 5 ! + 7 ! + v
x2 x4 x6 xs
cosh x =1 + 2 ! + 4 ! + 6 ! + 8 ! ... VI

x2 x3 x4 xs
ln(1 +x) =x -2 +3-4+5... VII
Make a note of these six series in your record book.
Then tum on to frame 21.

21 The binomial series

By the same method, we can apply Maclaurin's series to obtain a power
series for (1 + x)". Here it is:
f(x) =(1 + x)" f(O) = 1
f'(x) = n .(1 + x)n-1 f'(O) = n
f"(x) = n(n- 1).(1 + x)r2-2 f"(O) =n(n- 1)
f"'(x) = n(n- l)(n- 2).(1 + xyr--3 l";(O) =n(n -1) (n- 2)
fiV(x)=n(n-1)(n-2)(n-3).(l +x)"-4 fiv(o) =n(n-1) (n- 2) (n-3)
etc. etc.
General Maclaurin's series:
2 3
f(x) = f(O) + x.f'(O) + ~! f"(O) + ~! f"'(O)
Therefore, in this case,
(1 + x'f =1 + xn + x2 x3
2 ! n(n- 1) + J! n(n- 1)(n- 2) ...
n(n- 1) 2 n(n- 1)(n- 2) 3
(1 + xY' = 1 + nx + 2! x + 3! x . . . VIII
Add this result to your list of series in your record book. Then, by
replacing x wherever it occurs by (-x), determine the series for (1 - x)".
When finished, tum to frame 22.

337
Series 2

I(
I-x,\11 = I - nx +
n(n- I) 2
2! x -
n(n- I) (n- 2) 3
3! x +· · ·
22
DDODDODDDDODDDDDDDDODDDDODDDDDDDDDDODO

Now we will work through another example. Here it is:

Example: To find a series for tan- 1x.
As before, we need to know the successive differential coefficients in order
to insert them in Maclaurin's series.

f(x) = tan- 1x and f'(x) =I ! x2

If we differentiate again, we get f"(x) =- (1 .;: 2 ) 2 , after which the work-

ing becomes rather heavy, so let us be crafty and see if we can avoid
unnecessary work.
We havef(x) = tan-•x andf'(x) = I__!_2 = (1 + x 2 } 1• If we now expand
+x
(I + x 2 } 1 as a binomial series, ·Ne shall have a series of powers of x from
which we can easily find the higher differential coefficients.
So see how it works out in the next frame.

To find a series for tan- 1 x

f(x)-= tan-• x :. f(O) = 0 23
I
:. f'(x) = I--2 =(I + X 2J 1
+x
_ 2 (-I) (-2) 4 (-I) (-2) (-3) 6
- I -X + 1.2 X + 1.2.3 X + ...

= l-x 2 +x 4 -x 6 +x 8 - •.. f'(O) =I

:. f"(x) =- 2x + 4x 3 - 6x 5+ 8x 7 - • • • f''(O) = 0
:. f'''(x) =- 2 + 12x 2 - 30x4 + 56x 6 - • • • f"'(O) = -2
:. fiv(x) = 24x- 120x 3 + 336x 5 - • • • fiY(O) = 0
:. [V(x) = 24- 360x 2 + I680x4 - • • • [V(O) = 24 etc.
2 3
tan- 1 x = f(O) + x .f'(O) + ~! !'1{0) + ~! f"'(O) + ...

Substituting the values for the derivatives, gives us that tan- 1x = .............. .
Then on to frame 24.

338
 Programme 12

24 - x2 x3 x4 xs
tan 1x = 0 +x{l) +2! (0) +3! (-2) +41 (0) + S! (24) ...

ltan-lx=x-f +f-f+ ... l X

This is also a useful series, so make a note of it.

DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD DDDDDD
Another series which you already know quite well is the series for ex.
Do you remember how it goes? Here it is anyway.
x2 x3 x4
ex= 1 + x + 2 , + 3 ! + 4 , +... XI
and if we simply replacex by (-x), we obtain the series for e-x.
- x2 x3 x4
eX= 1- X+ 2!- 3! + 4! ... XII
So now we have quite a few. Add the last two to your list.
And then on to the next frame.

Examples: Once we have established these standard series, we can of

25 course, cqmbine them as necessary.
Example 1. Find the first three terms of the series for ex .ln(l + x).
We know that

and that

x2 x3 x4 }{ x2 x3 }
eX Jn ( 1 + X) = {1 + X + 2 ! + 3 ! + 4 ! . . . X - 2 + J - ...

Now we have to multiply these series together. There is no constant

term in the second series, so the lowest power of x in the product will be
x itself. This can only be formed by multiplying the 1 in the first series
by the x in the second.
Thex 2 term is found by multiplying 1 x( -f))x 2 _x2 =x2
and X XX 2 2
3
The x 3 term is found by multiplying 1 X ~

x( - f ) -x3 - -x~2+ x-2 =-

x
3 3 3
and x 3
2
and :!. X X and so on.
2

339
Series 2

26
It is not at all difficult, provided you are careful to avoid missing any of
the products of the terms.
OODODODOODDODOOO OQODDOODDDODDODD DDOOOO

Here is one for you to do in the same way:

Example 2. Find the first four terms of the series for ex sinh x.

Take your time over it: then check your working with that in frame 27.

Here is the solution. Look through it carefully to see if you agree with
the result.
27

ex. sinh x = {1 + x + ~; + ~~ ... } {x + ~~ + ~; + ... J

Lowest power is x
Term in x = 1. x =x

, J_ x 3 +x 2
1 "3! _ 3(1+1)_2x 3
X - 2! .X -X 6 2 - 3

There we are. Now tum on to frame 28.

340
 Programme 12

28 Approximate values
This is a very obvious application of series and you will surely have done
some examples on this topic some time in the past. Here is just an example
or two to refresh your memory.
Example 1. Evaluate v'1·02 correct to 5 decimal places.

1-02 =I+ 0-02

V1·02 = (1 + 0·02) 112

= 1 + 1co-o2) + 2 I.;
lc-l)
co-o2) 2
lc-lH-~)
+2 ~ _2 _3 2 co·02) 2 .•.

= 1 + 0·0 I -l (0-0004) + / 6 (0-000008)- ...

= I + 0-0I- 0·00005 + 0-0000005 ...
= I-OI0001-0-000050
= 1·009951 :. v'1·02 = 1·00995
Note that whenever we substitute a value for x in any one of the
standard series, we must be satisfied that the substitution value for xis
within the range of values of x for which the series is valid.
The present series for ( 1 + x yz is valid for IxI< 1, so we are safe
enough on this occasion.
Here is one for you to do.
Example 2. Evaluate tan- 1 0·1 correct to 4 decimal places.
Complete the working and then check with the next frame.

29 I tan- 1 0-1 = 0-0997 I

tan-tx=x-.:!:3 +~s -~7 +
3 5 7 ...

tan-1 0 . 1 = 0 . 1 _ 0-001 + 0-00001 _ 0-0000001

3 5 7
= 0·1- 0·00033 + 0·000002- ...
= 0·0997
We will now consider a further use for series, so turn now to frame 30.

341
Series 2

Limiting values - Indeterminate forms 30

In Part I of this programme on series, we had occasion to find the
. ..
IliDtttng al fUn+l
v ue o -;;;;- as n -+ oo. Sometimes, we have to find the limiting
value of a function of x when x -+ 0, or perhaps when x -+a.

e.. Lim {x 2 + Sx- 14} = 0 + 0- 14 = _14 = _7,

g x -+o x 2 - Sx + 8 0- 0+ 8 8 4
That is easy enough, but suppose we have to find

Lim { x~ + Sx- 14}

x-+2 x -Sx+ 6
.
Putt mgx .
=2.m the f unction, . 4 _+ 10- 14
gtves 4 10 + 6 %and what is the value
0
of-?
o·
Is it zero? Is it 1? Is it indeterminate?
When you have decided, turn on to frame 31.

~'as it stands, is Iindeterminate!

We can sometimes, however, use our knowledge of series to help us out
31
of the difficulty. Let us consider an example or two.

Example 1. Find the Lim {tan x 3- x}

x-+0 X

If we just substitute x =0 in the function, we get the result ~which is

indeterminate. So how do we proceed?
x3 2xs
Well, we already know that tan x =x + 3 + IS + . . . So if we replace
tan x by its series in the given function, we get

.
Ltm
x-+0
{.tan x
x
- x}
3
.
t../ x3 2xs
(Y'
+ 3 + IT- -+-...
= L1m ---=---=.:;...
x-+o
)
--"---
x
3
-.1)
. { 1 2x 2 } 1
=~0 3+15 + ... =3

... L'tm {tan X-

3
X} =-31 -and t h eJO. b ts· done.1
x-+0 X
Move on to frame 32 for another example.

342
 Programme 12

32 . {sinh
Example 2. To find Lim --
x~o
x}
X

.
Duect SU b stltUtiOn
· · 0 f X = 0 gtves
· - 0- 0 W h"lC h lS
sinh · 0 agam.
0 · S0 we Wl"II
express sinh x by its series, which is
sinhx = ....................... .
(If you do not remember, you will find it in your list of standard series
which you have been compiling. Look it up.)
Then on to frame 33.

33 x3 xs x1
sinh x = x + 31 + 51 + 7f + ...

So Lim {sinh x} = Lim { x + ~; + ~ + ~ ... )

x~o x x~o x

= Lim {1 + x3> xs~ + ... }

x~o . .
=1+0+0+ ... =1
... L"tm {sinh
-- x} = 1
x~o X
Now, in very much the same way, you find Lim {sin:x}
x~o x
Work it through: then check your result with that in the next frame.

34 Lim {sin?}= 1
x~o x

Here is one more for you to do in like manner.

Find Lim {sinh\- x}

Then on to frame 35. x~O X

343
Series 2

35
Here is the working in detail:
. x3 xs x7
smhx =x +- + 5- , + 7 , +
3! . .

. { -1 , + xsl2 + x 4, + ... }
. {sinh x3 - x } = L1m
... L1m 3 7
x-+O X x-+Q · · ·
1 1
=3! =6

... Lim {sinhx-x}=.!.

3 6
x-+o x

So there you are: they are all done the same way.
(i) Express the given function in terms of power series
(ii) Simplify the function as far as possible
(iii) Then determine the limiting value - which should now be possible.
DDDDOOOOOOOOOOOOOODODDDDODOOODOD DDDDDD

Of course, there may well be occasions when direct substitution gives

the indeterminate form ~ and when we do not know the series expansion
of the function concerned. What are we going to do then?
All is not lost! -for we do in fact have another method of finding
limiting values which, in many cases, is quicker than the series method.
It all depends upon the application of a rule which we must first
establish, so turn to the next frame for details thereof.

344
 Programme 12

36 L 'Hopital's rule for finding limiting values.

Suppose we have to find the limiting value of a function F(x) =~;~
at x =a, when direct substitution of x =a gives the indeterminate form
g, i.e. at x =a, f(x) = 0 andg(x) = 0.
If we represent the circumstances graphically, the diagram would look
like this:- Y

Note that at x =a, both of the

graphs y = f(x) andy = g(x) cross
the x-axis, so that at x =a, f(x) = 0
y•g(:c) and g(x) = 0

Y,
At a point K, i.e. x =(a+ h), KP = f(a +h) and KQ = g(a +h)
f(a +h)_ KP
g(a+h) -KQ
Now divide top and bottom by AK
f(a +h)_ KP/AK _tan PAK
g(a +h)- KQ/AK- tan QAK
Now Limf(x) =Lim f(a +h)= Lim tan PAK =~'(a)
x-+ag(x) h-+og(a+h) h-+O tanQAK g(a)

i.e. the limiting value of~~j as x-+ a (at which the function value by
direct substitution gives8) is given by the ratio of the differential coeffi-
cients of numerator and denominator at x =a (provided, of course, that
bothf'(a) and g'(a) are not zero themselves)!
:. Lim { f(x) }=f:(a) =Lim { ~'(x)}
x-+a g(x) g(a) x-+a g(x)

.. . {f(x)}-
. Lim . {f'(x)}
g(-) - L1m ,.-()
x-+a X x-+a g X
This is known as l 'Hopital 1s rule and is extremely useful for fmding
limiting values when the differential coefficients of the numerator and
denominator can easily be found.
Copy the rule into your record book. Now we will use it.

345
Series 2

{ f(x)} _ L" (f'(x)}

Lim--tm~
37
x~a g(x) x~a lg(x)

x3 +x 2 -x-I}
Example 1. To find;~ { x 2 + 2x _ 3

Note first that if we substitute x = I, we get the indeterminate form &.

Therefore we will apply l'Hopital's rule.
We therefore differentiate numerator and denominator separately
(not as a quotient).
. {x 3 +2
Ltm x
2 - .
x-I} = Ltm {3x 2x+2
2 + 2x- 1}
x~l X +lx-3 x~l

3+2-I 4
= -=1
2+2 4
. . {x
.. Ltm
3 + x 2 - x- I } -I
2 2x 3
_
x~l X - -
and that is all there is to it!
Let us do another example, so, on to the next frame.

cosh x- ex}
Example 2. Determine Lim {
x~O X 38
We first of all try direct substitution, but we find that this leads us to
the result I ~ I, i.e. ~which is indeterminate. Therefore, apply l'Hopital's
rule
Lim{ f(x)}= Lim{®}
x~a g(x) x~a g(x)

i.e. differentiate top and bottom separately and substitute the given value
of x in the differential coefficients.
:. Lim{coshx-ex}= Lim{sinhx-ex}
x~o x x~o I
=0-1=-1
I
... L"tm {cosh x - ex } =-I
x~o x
Now you can do this one:
Ll·m { x -2 sin 3x}
2
Determine
x~O X +4x

346
 Programme 12

39 Lim{ x 2 ~sin 3x} = _l_

x-+0 x + 4x 4
The working is simply this:
Direct substitution gives g, so we apply l'Hopital's rule which gives

. {x
L1m
x-+O
2
- sin 3x}
2
x +4x
L'
=1m
x-+o
{2x- 3 cos 3x}
2x+4

=0-3
- - =--3
0+4 ~
WARNING: l'Hopital's rule applies only when the indeterminate form
arises. If the limiting value can be found by direct substitution, the rule
will not work. An example will soon show this .

Consider . {x 2 +4x-3}
Ltm 5 - 2x
x-+2
. t su b stl'tutwn,
By dtree . t h e 1·1m1hng
. . value = 4 +5 _8 -4 3 = 9 . By l'H op1t. al' s
rule;~ {x ~x; 3 } =;~2 {2x~24 }= -4. As you will see, these results
2
;

do not agree.
Before using l 'Hopital 's rule, therefore, you must satisfy yourself that
direct substitution gives the indeterminate form g. If it does, you may
use the rule, but not otherwise.

40
Let us look at another example
. L'1m {X
Examp l e: Determme 2 X}
- sin
x-+0 X
.
By duect su b shtubon,
. . 1·lml't'mg vaIue 0 O =o·
= O- O
Apply l'Hopital's rule:
. {x-sinx}
L1m 2
L' {1-cosx}
=tm
x-+0X x-+0 2x
We now find, with some horror, that substitutingx = 0 in the differen-

g. So what do you
:s x}·
tial coefficients, again produces the indeterminate form

suggest we do now to find ;~ { 1 - (without bringing in the use of

series)? Any ideas?

We ....................... .

347
Series 2

We apply the rule a second time. 41

Correct, for our immediate problem now is to fit!d Lim { 1 - ~s x}. If we
do that, we get: x -+o

Lim { x - s;n x }= Lim { 1 - cos x }= Lim {sin x }= Q =0

x-+0 x x-+o 2x x-+o 2 2
'-·······--·-·/ '\. ...... _____ /
First stage Second stage
. L"lm {
..
x- sin x} =0
2
x-+0 X
So now we have the rule complete:
For limiting values when the indeterminate form (i.e. ~) exists, apply
l'Hopital's rule
Lim f f(x~} = x-+a
x -+a tg(x;
Lim (<(x)}
g (x)

and continue to do so until a stage is reached where either the numerator

and/or the denominator is not zero.
Next frame.

Just one more example to illustrate the point.

. {sinh x- sin x}
. L1m
Example: Determme 3
42
x-+O X

. t
D1rec l U t·10n
SU b St"t · 0Q,
g1ves · 0Q'
- 0 I.e. c·Ind etermmate
· )

.. . L"1m {sinhx-3 sinx} __ L.1m {coshx-

3X 2
.
cosx} , gtves - -1 = 0-
1-
x-+0 X x-+0 0 0
. {sinhx + sinx} . 0-+-0 =00-
=x-+o
L1m
6x '
gtves
0
= Lim {cosh x+ cos x} = 1...!_! =_!_
x-+0 6 6 3
. L" {sinhx-sinx}=!.
3 3
.. lm
x-+0 X

Note that we apply l'Hopital's rule again and again until we reach the
stage where the numerator or the denominator (or both) is not zero. We
shall then arrive at a definite limiting value of the function.
Turn on to frame 43.

348
 Programme 12

43 Here are three Revision Examples for you to do. Work through all of
them and then check your working with the results set out in the next
frame. They are all straightforward and easy, so do not peep at the
official solutions before you have done them all.

Determine (i) Lim {x 3 - ~ 2 + 4x- 3 }

x~l 4x - 5x + 1

(ii) Lim { t~n x - x }

x~o sm x-x
(iii) Lim
x~o
{x cosxx sin x}
3-

Solutions:
44 (i) ~!i {
x3 - 2x 2 + 4x-
4x2 _ Sx + 1
3} (Substitution gives 0)
0
-_ L"Im {3x - 4x + 4 }-
2
3_ 1
---
x~l 8x-5 3
. . {x
.. x~l
Lim
3 - 2x 2 + 4x- 3 } _
4X 2 - 5X+ 1 -1

L" {tanx-x)
(ii)
x~~ sinx-xf (Substitution gives~)
. {sec 2x-1}
=Lim (still gives ~)
x~o cosx-1

= Lim {2 sec2~ tan

x~o -smx
x) (and again!)

= Lim {2 sec 2x sec 2x + 4 sec 2x tan 2x} =2 + 0 = _ 2

x~o -cosx -1
Lim{t~nx-x}=-2
:.
x~o x-x sm

(iii) Lim {x cos x- sin x} (Substitution gives 2.

x~O X
0)
= Lim{-xsinx+cosx-cosx}
x~o 3x

-_ L"Im {-sin
--- x }-
- L.Im {-cos
--- x}-
---1
x~o 3x x~o 3 3

... L"Im {xcosx-sinx} -__ -1

x~O X 3 Next frame.

349
Series 2

Let us look at another useful series: Taylor's series.

2
45
Maclaurin's seriesf(x) = f(O) + x.f'(O) + ~! f''(O) + ... t:Xpresses a
y
y=f(xl function in terms of its
differential coefficients
I
at x =0, i.e. at the point K.
I
I
I
:f(h}
I
I
I
f(Ol I
I

At P,f(h) = f(O) + h.f'(O) + ~; f"(O) + ~; f'"(O) ...

y

If we now move they-axis a

units to the left, the equation
of the curve relative to the
new axes now becomes
y = F(a + x) and the value at
K is now F(a)

At P, F(a +h)= F(a) +h. F'(a) + ~; F"(a) + ~; F"'(a) + ...

This is, in fact, a general series and holds good when a and h are both
variables. If we write a= x in this result, we obtain

f(x +h)= f(x) + h.f'(x) + ~; f"(x) + ~; f"'(x) + ...

which is the usual form of Taylor's series.

Maclaurin's series and Taylor's series are very much alike in some
respects. In fact, Maclaurin's series is really a special case of Taylor's.
46
Maclaurin's x2 x3
series: f(x) = f(O) + x.f'(O) + 2 ! /"(0) + 3 ! f"'(O) + ...
Taylor's h2 h3
series: f(x +h)= f(x) + h.f'(x) + 2! f"(x) + 31 f"'(x) + ...
Copy the two series down together: it will help you learn them.

350
 Programme 12

47 Example 1. Show that, if h is small, then

-1r h)- -1 h xh2
tan \x + -tan x + 1 + x 2 - ( 1 + x 2 ) 2
. 1
approxunate y.
DODDDDDDDDDDDDDDDDDDDDDDDDODDDDO DDDDDD
Taylor's series states
f(x +h)= f(x) + h.f'(x) + ~; f"(x) + ~; f"'(x) ...
where f(x) is the function obtained by putting h = 0 in the function
f(x +h).
In this case then, f(x) = tan- 1x.
1 2x
:. f'(x) = 1 + x2 and f"(x) =- (l + x2)2
Putting these expressions back into the series, we have
-1 { h) - -1 h _1- - h2 ~ +
tan \x + -tan x + . 1 + x2 2 , "(1 + x2)2
_ _1 h _ xh 2
-tan x + f+? (}:i:"X2)2 approx.
Why are we justified in omitting the terms that follow?

The following terms contain higher powers of h which, by

48 defmition, is small. These terms will therefore be very small.
Example 2. Express sin (x +h) as a series of powers of hand evaluate
sin 44° correct to 5 decimal places.
h2 h3
sin(x +h)= f(x) + h.fi(x) + 2 , f"(x) + 3 ! f"'(x) + ...
f(x) =sin x; f'(x) =cos x; f"(x) = -sin x;
f"'(x) = -cos x; Jiv(x) =sin x; etc.
. (
... sm .
x + h) = sm x + h cos x- h2 . h3
2 ! sm x- 3 ! cos x + ...
sin44°=sin(45°-lj=sin(.E:-o.01745) and sin!!:=cosE=
4 4 4 y2
1
:. sin 44°= v'~ {1 +h-r -~ + ... } h=-0·01745

= v'~ {1- 0·01745- 0·000;045 + 0-00~0053 + ... }

= v'~{l- 0·01745- 0-0001523 + 0-0000009 ... }

= 0-7071 (0-982399) = 0-69466

351
Series 2

You have now reached the end of the programme, except for the test 49
exercise which follows. The questions are all straightforward and you will
have no trouble with them. Work through all the questions at your own
speed. There is no need to hurry.

Test Exercise- XII

1. State Maclaurin's series.

2. Find the first 4 non-zero terms in the expansion of cos 2 x.
3. Find the first 3 non-zero terms in the series for sec x.
x3 x 5 x 1
4. Show that tan- 1x = x- 3 + 5 -7 + ...
5. Assuming the series for ex and tan x, determine the series for ex. tan x
up to and including the term in x 4 •
6. Evaluate v'l·05 correct to 5 significant figures.

7. Find (i) Lim { 1- 2 sin 2 x 2- cos x}

3

x~o 5x
(ii) Lim {tan x . tan: 1 x- x 2 }
x~O X

(iii) lim {x- sin x}

x~o x-tanx

8. Expand cos(x +h) as a series of powers of hand hence evaluate

cos 31° correct to 5 decimal places.

You are now ready to start the next programme.

352
 Programme 12

Further Problems-XII
x2 x4 x6
I. Prove that cosx = 1-- +---+ and that the series is valid
2! 4! 6!
for all values of x. Deduce the power series for sin 2 x and show that,
if xis small,
sin 2 x-x 2 cosx 1 x 2
x4 = 6+ 360 approximately.

2. Apply Maclaurin's series to establish a series for ln(l + x). If 1 +x =},

show that 2 3
(b 2 -a 2 )/2ab = x -~2 + ~-2 ...

Hence show that, if b is nearly equal to a, then (b 2 - a 2 )/2ab exceeds

m(~) by approximately (b -a ) 3 /6a 3 .

3. Evaluate (i) Lim {11 - 2 Sl~

. 2 ~- 3 }
COS X
x-+0 t X

(ii) Lim {sin

x-+o
x-x~ cos x} (iii) Lim {tan
x-+0
x- sin x}
X
3

(iv) Lim {sin x; x} (v) Lim {tan X:- x}

x-+O X x-+0 smx
x-

4. Write down the expansions of (i) cos x and (ii) 1- 1 -, and hence
+x
show that
cos x x 2 x 3 l3x 4
--=1-x+---+ - -
1 +X 2 2 24

5. State the series forln(l + x) and the range of values of x for which it
is valid. Assuming the series for sin x and for cos x, find the series for

1nei: x) and In( cos x) as far as the term in x 4. Hence show that, if x

is small, tan x is approximately equal to x. ex 213 .

6. Use Maclaurin's series to obtain the expansion of ex and of cos x in

ascending powers of x and hence determine

Lim {
ex+ e-x- 2}
x-+0 2 cos 2x- 2

353
 Series 2

7. Find t~e first four terms in the expansion of ( 1 _ x~;C: + x 2 ) in

ascendmg powers of x.
8. Write down the series for ln(l + x) in ascending powers of x and
state the conditions for convergence.
If a and b are small compared with x, show that

ln (x + a) - ln x = L(1 + b ~xa ){ln (x + b) - ln x}

9. Find the value of k for which the expansion of

(1 + kx) (1 + ~r 1 ln(l + x)

contains no term in x 2 •
10. Evaluate (i) lim (sinh x - 3 tanh x}
x~o\ x

(ii) Lim (
x~l X
~- x1} (iii) lim { x : sin x}
x~o x +x

11. If u 7 and u 7 • 1 indicate the r 1 h term and the (r -l)th term respectively
of the expansion of (1 + x)n, determine an expression, in its simplest

form, for the ratio Ur . Hence show that in the binomial expansion
Ur-1

of (1 + 0·03) 12 , the r 1h term is less than one-tenth of the (r -1) 1h

term if r > 4. Use the expansion to evaluate (1·03) 12 correct to three
places of decimals.
12. By the use of Maclaurin's series, show that
.- x 3 3x 5
sm 1 x = x +- + - +
6 40
Assuming the series for ex, obtain the expansion of ex sin- 1x, up to
and including the term in x 4 • Hence show that, when x is small, the
graph of y =ex sin- 1x approximates to the parabola y =x 2 + x.

13. By application of Maclaurin's series, determine the first two non-

vanishing terms of a series for ln cos x. Express (1 +cos 0) in terms
of cos 0/2 and show that, if 0 is small,
02 04
ln(l +cos 0) = ln 2-4- 96 approximately.

354
 Programme 12

14. If x is small, show that

(i) j (1-x
1 + ):!): 1 +
X X + X2
2
(11.. ) y'{(l + 3x 2 )eX} _,_ 3x 25x 2
- 1 + -+-
' 1 -x 2 8
15. Prove that

16. Find (i) um{sinh-t;-x}, . . . {esinX-1-x}

(u) Ltm 2 •
x-+-0 X x-+-0 X

17. Find the first three terms in the expansion of

sinh X/.-ln(l + x)
x2(1 +x)J
18. The field strength of a magnet (H) at a point on the axis, distance x
from its centre, is given by

H= ~ ((x2l)2- (x~/)2}
where 21 = length of magnet and M = moment. Show that, if lis
very small compared with x, then H :!): 2 ~.
X

19. Expand [ln(1 + x)) 2 in powers of x up to and including the term in

x 4 . Hence determine whether cos 2x + [ln(l + x)] 2 has a maximum
value, minimum value, or point of inflexion atx =0.

20. If lis the length of a circular arc, a is the length of the chord of the
whole arc, and b is the length of the chord of half the arc, show that

(i) a = 2r sin ir and (ii) b = 2r sin :r, where r is the radius of the
. 1e. By expan d"Illg sill
cue . Ir an d Sill
2
. Ir as senes,
4 - 3-- a
. sh ow that I =Bb

approximat.ely.

355