CALCULUS:

A FIRST COURSE

DAVID J. MOORE
CALCULUS:

a first course
 2gau0
=~
4 ; oe
* My 7 ~

e 5 ‘ p :
S 7

gagn0) tea & -

Calculus: a first course

DAVID J. MOORE

Department of Mathematics
University of Glasgow

ARKLAY PUBLISHERS
 © David J. Moore, 1995

ISBN 0 9507126 9 8

First Edition 1992.

Second Edition (with minor corrections) 1995.
Reprinted 1997.

All rights reserved. No part of this work may be reproduced,

stored in any information storage and retrieval system, or transmitted
in any form or by any means, electronic or mechanical, including
photocopying or recording, without the prior permission in writing of
the Publisher.

Published by Arklay Publishers,

64 Murray Place, Stirling, Scotland. FK8 2BX.

Printed by Bell and Bain Ltd, Glasgow.

 Preface
This book, as its title suggests, provides a first course in Calculus.
The aim of the book is to develop a facility in the basic operations of
Calculus, differentiation and integration, and to show how these may
be used in practical applications. We also examine how curve sketching
can help in understanding the behaviour of functions.
This book presents the course which is currently offered to students
attending Mathematics class 1B here at Glasgow University. Such stu-
dents are primarily interested in studying one or more years of Mathe-
matics, usually as a back-up to their main course of study. Their interest
is in the application of Mathematics, rather than in the abstraction of
Mathematics. This book reflects that interest.
In writing the book, I have adopted a non-rigorous approach. I
have always strongly believed that the initial step for students learning
Calculus is to learn to carry through the basic operations quickly and
reliably, rather than becoming bogged down in the technicalities con-
cerning limits, etc. The time for getting involved with such matters,
vital as they are, is after the Calculus has been successfully mastered.
In adopting a non-rigorous approach, I have tried to write in an
informal manner, trying to explain as we go along the ideas that we are
using. I have also included rather more detail of the algebraic manip-
ulations than is common with books of this level. This is an attempt
to make it easier for readers to follow the working in the book. I hope
that the many students for whom such detail is not necessary will skip
over it, and forgive its inclusion.
This book owes a lot to the many fine teachers who have devel-
oped Calculus courses here at Glasgow University over the years. In
particular I would like to thank my colleagues, Dr. J. B. Hickey and
Dr. D. B. Webber, for their sterling efforts in reading the text in detail,
and offering numerous suggestions for improvement. Inevitably, errors
will remain in the text, but these have been minimized as far as possible,
and of course such errors are my responsibility and not the responsibility
of the checkers.
I would also like to record my thanks to Dr. I. S. Murphy for his
considerable contribution to this project.
Finally, I must also thank my wife, my children and my parents for
their contributions to the book, none of which have been mathematical.

David J. Moore

University of Glasgow
July, 1992
 ~~ —— ee

sos
eHTSD cieaivos geri cabot jrommaade sity otf wet dost het
4! aang udmsty ‘evel, nel ii wbiftost a diesel of a dood si 1a tie adT
ya, SeSii vitiT west od Bias ittrele sim bia eptia(teaastth @alivalsD
BAT Jndle lit West witiiiayxs, als acl wren tasking iaxitem@tg of bor ad
sap sere, %o'wotvedid ot? gabanioniiay af ghd war
“shasbalie >! Peera iad YOLerrise a ibid a Ser ail}, reeng dood et'T
nea etki vdiato “ uy Wry Re )? ta et til. awala cmsiteM Bhibns die
heath wis: Pu TTS ay NT BAe: Cet ii Sydipesederi inant ang:oa alosh
atetat wel. hase ls Sania didi ‘id estdir-dowe) «an giieeae oitecn
Ms aGitoetterla bri} ai oe ste aiianretisit to aottwaliqgas ail! af ei
= rani jad ajuties doad cit T ealdame:linM
T abemaee esuwun on « be gobe oy 1 gleed ah gative ci
Bitte! airline te ret it ngi wild gait) fewachad|Ti
e
eg aie ator smad sey agree) ; ¢bag OFiF
iy ead Salt oF unis Bolan pti iy
. bs Aric’ (forte iLsiw tr4Nviovie ‘Pte "i? hit het
2 io beennt 72 mah KORNAY 1k aby “a ae he
et ‘Saw bert? oved )slokerqge anion ec - BMidqndis al
eit aw dad? ceohi aiid mnie oy ow te nisiqze of utter rages: Lanraotrsi
nen vendysln s i o Viateh som wader Bibobsat Gee wired 1 ania
{mAMde ue 2) dT ..oxy ehdhgae eel dit stgnnenoy st Bod t arorieln
meek. abc) col) ohio ad Wee pahares 408 +i solast03
ciate lie Vicon gon = Meith dae coma You | nm odd jad)
cto atte bad ot 370
ler see ode dros) oie! you oldie wwe Adod «i'l_
- gal Meme? tli ow vliwwerint): wy wesi ie) 4a! foil wilinlad: beat
baw: woot. capellon gm aivneld)chai Mlwew 1 saluted
_pldiiehaitre) 201)oe @) efsalts silos ad — & @ ag
here orldstivoo: _iensrea ver separ) anal ani ah Ade ne ,
«bias an itl an Lasatainicn aoad eum semeie a
ioe ad? don bie aati! a

a | ns Dew “uarbliss Peg i Aa

cordvest tied — doickor we ah jae

%inl iL bidvth
Je
 Contents
Functions and Graphs
The Trigonometric Functions
The Inverse Trigonometric Functions
Differentiation
Examples involving Differentiation
Higher Derivatives
Implicit Differentiation
Sketching Rational Functions without using Calculus 64
Se Concavity and Points of Inflection
Oe
ge
Local Maxima and Local Minima
Absolute Maxima and Absolute Minima
Introduction to Integration
Integration by Change of Variable 105
Integration by Parts 114
Trigonometric Integrals
Trigonometric Substitutions 128
Partial Fractions 133
Two Methods using Complex Numbers 145
Tables of Integrals 149
The Definite Integral 154
The Area under a Curve 163
Volumes of Revolution 169
Arc Length and Surface Area of Revolution 173
Numerical Integration 178
Newton’s Method 186
Maclaurin Series 191
Differential Equations 199
Applications of Differential Equations 216
EO
WN
AAP
SHDN
WNP
COMNAACRF
NYO
NON
KH
DY
WN
Be
KY
YH
HY
KP
HB
HP
NN
NNN
BR Basic Techniques 225
Numerical Answers to the Examples 234
Some Standard Integrals 242
Index 243
 wo eve

| alaotae® a ; | |
_ ailqnr
<2 d inn eat di
ei avin crag ay |S
es sitar aismooght? « ns vals &
at mits es ; i»
$a) coitsimns hid yaivlots ale a 3
f° rahe ta. a
no noid ptitnt nin cas x
re eoluole 3 giiey dowdihe eanolideiell larte:tAfl-,alls 8
iy (ietoetiol to. ais Brey oes
: a
i mpiald ics bmi. xe he
pt wT otaleed/. baie pike,
- 8¢ “torseceaeed Gt
oor sicdoheV Yo egos) ql nena
af eer.
ers forme
at
Eki
aot waaay said aut ao
abt wo
on —-
na! asad
eer ac Ht Be
aj oll CUTie Rear ich
ers natneshiesst Yosails poets
8 mid
wt '
Mi
fe!
i
 CHAPTER 1
Functions and Graphs.
A function f is a rule, often given by a formula, that associates
with each permitted value of the independent variable (which is fre-
quently denoted by xz) a unique value of the dependent variable (often
y).
For example, y = 2”, y = 22° +5, and y = 9— 2? all express y as a
function of z, since, in each case, for each value of x there corresponds
a unique value of y.
If we use the notation f(r) to denote a function of z, then f(a) will
denote the value of the function when z is replaced by the specific value
a. So, for example, if f(z) = 273 + 5, then f(1) =2 x 13+5 =7, while
f(2) = 2 x 2° 452121 and. f(—3)\=2.%(43)
4+ 6.2 6-21)
a5 =
—54+5 = —49.
It is frequently useful to be able
to represent a function by means of a positive part
graph. To do this we use the rectan- of the y-axis
gular (or Cartesian) coordinate sys-
tem. negative part positive part
In theory, to represent a func- of the z-axis of the z-axis
tion f graphically, we take each pos- O
sible value of z, calculate the corre-
sponding value f(z), and then mark
on the graph the point with coordi-
negative part
nates (x, f(x)). So the point marked °"8!
of the y-axis
on the graph is the point which is a
distance z along the z-axis (negative values of z being marked to the
left of the origin) and a distance f(z) above the z-axis (negative values
of f(z) being marked below the z-axis).
Since the above procedure is not practicable as it would take in-
finitely long to perform, in practice we choose a few points that we hope
will prove to be typical, and hope to see a pattern emerging. It is some-
times difficult to decide which points to plot, and when sufficient have
been plotted to give a good indication of the shape of the curve. Only
experience can really help us in many cases. However one of our aims
in this course is to obtain some feel for the behaviour of some typical
functions.
Example 1. Sketch y = 2? and y = (x—2)? on the same diagram.
Solution. To obtain an idea of the behaviour of these two graphs,
we draw up a table of alae
 2 Chit

Notice that the graphs have the same shape, but that the graph
of y = (x — 2)? reaches its minimum
value at x = 2, whereas y = z? has
its minimum at z = 0. So the effect
of replacing x by x — 2 is to move the
graph 2 units to the right. Note that
it 1s what is subtracted from x that
tells us how far to move the graph to
the right.
Each,of the above graphs has an
axis of symmetry. In the case of
y = 2” the graph is symmetric about
the y-axis, while the line z = 2 is
the line of symmetry of the graph of
y =(x—2)?.
Example 2. Sketch the graph of y= (x +1)’.
Solution. We need to write
xz +1 as x —(-1), in order to see
that we are subtracting —1 from z.
As we saw above, this tells us that
the graph of y = (z + 1)? will be the
same as the graph of y = 2”, but
moved —1 units to the right [which
is the same as moving the graph 1
unit to the left]. Hence we see that
the graph has its minimum value at
xz = —1, and that z = —1 is the axis
of symmetry of the graph.

Example 3. Sketch y = x? and y = 32? on the same diagram.

Solution. In this example,
yly = 327
we see that every y-value for y = zx?
is multiplied by 3 to give the cor- 2
responding y-value on the graph of
y = 3z?. The result of this is to cre-
ate a new curve, with a basic shape
similar to the shape of y = x”, and
still symmetric about the y-axis, but
rather steeper than was the case for
y= 2".

In preparation for the next set of

examples, we consider the technique
known as completing the square. Suppose we start with a typical
 Ch: 1 3

quadratic expression az? + bz + c, where we shall assume that a is not

zero. The routine for completing the square involves taking the terms
involving z, and re-writing them as the square of a single z-term, of the
form (r+?)?. So we write
az’? + br +c =ar’? 4+ brte

Silesi@)alts
ees

on taking out as a factor a, the coefficient of z?. To deal with the

[x? + (b/a)z]-term, we need to take z+ one-half of the b/a-term, and
square this. [The reason is that when we square z+? then we obtain
x? +2x?+??, and so we need to have 2x? equalling b/a, or ? = b/2a]]
We also need to subtract the term (b/2a)’, to compensate for having
just added in this extra term. So we have
b
az? +brte=als?+ (2)a]+0
a

mA 3b * _ 0 — 4ac
me 2a son a

You will notice that we have collected the {—b?/4a}-term and the c-
term over the common denominator 4a, and taken a minus sign out,
which leads to the compensatory — before the 4ac-term.
We shall now do a numerical example, which will help to clarify the
above ideas.
Example 4. Complete the square in 32? + 127 + 5.
Solution. Following the above routine, we see that

32? + 122 +5 = SSS

32? + 122 +5
= 3[x? + 42] +5
= 3[(z + 2)? — (2)?] +5
= 3|(c +2)? -4] +5
= 3(z +2)? —124+5
= 3(z + 2)? —7.
 4 Ghisl

Now we shall apply the idea of completing the square to help us to

sketch the graphs of other quadratic functions.

Example 5. Sketch y = 2x? + 122 — 14.

Solution. If we apply the technique of completing the square,
we have
2x? + 122 — 14 = er
22” + 122-14
= 2[r? + 62] —14
= 2[(2 + 3)? —37] -14
= 2[(x +3)? —9] —14
= 2(2 + 3)? -18-14
= 2(x + 3)? — 32.
So we require to plot the graph of
y = 2(x + 3)? — 32. As we have al-
ready seen in similar examples, the
graph will be the same in shape as
the graph of y = 22”, but centred
around the line z = —3, and moved
down by 32 units. Hence we obtain
the graph shown.
To find where the graph cuts the
x-axis, we need to find the values of (—3, 32)
xz which make y = 0. We do this by eee
factorising 2x? + 127 — 14. Now Yea

22? + 122 ~ 14 = 22? + 62 — 7]

= 2((2
+ 7)(z —1)).
If you need further explanation of how we obtained the above factori-
sation, please see the chapter on Basic Techniques at the end of the
book.
It follows that 27? + 122 — 14 = 0 when 2(z + 7)(z — 1) = 0, which
can only happen if one of the factors is zero. Hence we need either 2 = 0
(which is impossible), or z + 7 = 0 (which happens when z = —7), or
xz —1=0 (which means that z = 1). From this, we conclude that the
graph cuts the z-axis at the values z = —7 and z = 1, as marked on the
above diagram.
To find where the graph hits the y-axis is rather simpler. For this,
we merely need to calculate the value of y that corresponds to z = 0.
Obviously, if y = 2x? + 122 — 14, then when z = 0 we have y = —14,
again as marked on the diagram.
 Example 6. Sketch y = 8z — 7 — 2z?.
Solution. Completing the square, we have
82 — 7 — 22? = —22?
+ 82 -7
Sa eee

= —2[r? — 42] —7
= =2((@ — 2)? — (-2)"]} -7
= -2[(x — 2)?
—4] -7
= —2(r — 2)? +8-7
= —2(r — 2)? +1.
This time we have to consider the graph of y = —2(z — 2)? +1. We
have already looked at graphs like y = x* 2 on several occasions. The
graph of y = —2z? will be the same
as the graph of y = 2z? but with ev-
ery y-value multiplied by —1. The
effect of this is to give the reflection
of y = 22? in the z-axis. Next, the
graph of y = —2(z — 2)? will be sim-
ilar to the graph of y = —2z?, but
displaced 2 units to the right. The
graph of y= —2(x—2)?+1 will be the
same as for y = —2(z — 2)? but with
all y-values increased by 1 unit. The
result of this is to move the whole
graph upwards 1 unit. Hence we ob-
tain the graph shown in the diagram.
As in the last example, we see that the graph meets the y-axis at
the point z = 0, y = —7.
From looking at the graph, we can see that the curve will meet the
z-axis. However it is not at all easy to see how we can factorise the
expression 8z — 7 — 2x”. In these circumstances, the best approach is to
use the well-known formula for the solution of quadratic equations:
—b+/v b? — 4ac
az? +bzr+c=0 when x=
2a :
[In this formula, a, b, and c are assumed to be constants, with a # 0.]
Applying this formula with a = —2, b = 8, and c = —7, we see that
the roots are given by
—8 + ,/ 8* — 4(-—2)(-7)
2(=2)
i
e—

so that
—8 + / 64 — 56
4 9
or j

ee
—8+/8

If we rewrite V8 as ./4 x 2, then we have /4 x 2 = V4x /2=2x 2:

Hence, on dividing by —4, we easily see that the roots of the equation
8a —7—22? =Oarer =2F Se Hence we deduce that the graph meets
the z-axis at the points =2+ ¥%,y=0 andr =2- ¥, y=0.
The expression ‘b? — 4ac’ in the quadratic formula given above is
known as the discriminant, and is often denoted by A.
If A is positive, then the equation ar? + br +c = 0 has two distinct
real roots. When A is zero, the equation has just one (repeated) real
root. Finally, if A is negative, then the equation az? + br + c = 0 has
no real roots (although there will be two distinct complex roots).
This means that the graph of y = az? + br +c will meet the z-axis
in two distinct points if A > 0, will touch the z-axis at just one point if
A =0, and will not meet the z-axis at all if A < 0.

Having considered the graphs of

quadratic functions for some time, let
us now look at the graphs of y = x”
when n is a positive integer larger
than 1.
Firstly we consider the graphs of
y = r” when nis an even integer. We
find that the graphs are all symmet-
ric about the y-axis (as for y = 2”)
and that as n increases the graph of
y = x” becomes flatter near the ori-
gin but is steeper further away from
the origin. All of the graphs pass
through the points (0,0), (1,1) and (—1,1), and have the z-axis as
a tangent at the origin.
These conclusions can readily be deduced from drawing up a table
of values for the equations y = z?, y = x* and y = 2°.
re [oO], #ia |ei, oa [si, a2, 23] 2a
re [ol aye | 1s [one |1
ret[0|a/a56 |isis |81/256 | 1
[a*[0|1/4006 |1/64
Now we consider the graphs of y = x” when n is an odd integer.
We find that these graphs are all symmetric about the origin, and, when
n > 1, they all have similar basic shapes. Again for n > 1, the graphs
are flatter near the origin the larger that n is, while they are steeper
further away from the origin the larger n is. The z-axis is a tangent to
 Ch: 1 7

the graph of y = x” at the origin for allodd n > ap The pattern can be
seen by tabulating the values of y = x? and y = x° against values of z.
Ea cd LG ee
ads oe |eee e278] >aye" TOR
a
See ee eee aa
reo] aye | 1/s |27/ea [18 |27|
e® [0|a/toaa [1/32 |2431024 [1 [32 |245|

The graphs of cubic functions (that is functions where the highest

power of z occuring is z*) can also have the shapes exemplified in the
following:
(a) y=a2*—2?+2-—1=(x—-1)(z? +1), which meets the z-axis
at the one point rz = 1 and is a constantly rising curve as we look at it
from left to right. We describe the function y = r* —2?+2—lasa
strictly increasing function, because as z increases, so y increases.
(b) y=2?—27r?-52+6 = (r+2)(zr—1)(z —3), which meets the
x-axis at the three points r = —2, r = 1 and z = 3. In this case there are
portions of the curve which rise as we
look from left to right, but also a por-
tion of curve which is falling. We de-
scribe y = 2° — 27? —5z +6 as be-
ing a strictly decreasing function
on a certain set of z-values, (from
xz =atoz = 6b in the picture) be-
cause, on these z-values, y decreases
as x increases.

Besides the three shapes that we

have seen above, there are also the re-
flections of these shapes in the z-axis.
These can be distinguished from the
 8 Chl

above because the coefficient of z° in the equation of the curve would

be negative.
We shall see shortly how Calculus helps us to distinguish between
the possible shapes of cubic curves by telling us how many “turning
points” a curve has, and also about the nature of the turning points.
In the meantime however we have to rely on the fact that a cubic
curve can meet the z-axis at at most three points, and can have at most
two turning points.

Example 7. Sketch the curve y = x? — 122 + 16 by noting that

ao? — 122 +16 = (x — 2)*(r +4).
Solution. From the equation of the curve in factorised form,
we can see that y = 0 when z = 2 and when z = —4. We can also
easily calculate that when z = 0, y = 16. To decide on the behaviour
of the curve, and in particular where it is above the z-axis and where
below, we shall use a table of sign for y. In this table we include those
values of x for which y becomes zero. These are the z-values r = 2 and
x= —4.

(c +.4)(2 — 2) ca beSR
In filling in the table of sign, it is a good idea in each row to fill in
firstly the point for which the term in that row is zero (if there is one).
For example, in the row labelled x + 4 we would start by putting the 0
under z = —4. After this, all we need to do is to decide whether the
term is positive or negative to each side of its zero. Remember that the
square of any real number is positive unless it is zero.
In interpreting the information from the table of sign, we must be
clear that it is telling us about whether y is positive, negative or zero.
In other words, it is telling us about whether the graph of the curve lies
above, below or on the z-axis. It can be helpful to include an extra row
in the table of sign for y to bring this out, as shown below.
a ee ee Be eee
EY Ma NS ET
fc eae
ane 2) ae aio bata
(x +4)(x — 2)?

As we see, the graph of y = x3 — 122 + 16 = (x — 2)"(zx + 4) will

cross the x-axis (from below it to above it) at the zero at z = —4, while
 Ch: 1 9

it will lie above the z-axis to the left and to the right of the zero at
z=2.
From this we see that the graph
y = 2 — 122+ 16
of y = z* —12r +16 is as shown in
the diagram.
Let f(z) = x? — 122 + 16. We
can factorise the function f in the
form (z — 2)*(z +4), from which we
see that f vanishes (i.e. it has the
value zero) when z is —4 and 2. We
can also use the table of sign to see
that the function f(z) is (strictly)
positive (f(x) > 0) whenever z lies
between —4 and 2 and also when z is
larger than 2. We can express this in
interval notation by saying that
f(x) >0 precisely when z € (—4,2)U(2, 00).
Here (—4,2) denotes the set of all real numbers which lie strictly
between —4 and 2, not including the two end-points x = —4 and zr = 2.
The set of all real numbers which are strictly larger than 2 is denoted by
(2,00). The U (union) symbol that connects these two sets of numbers
indicates that we should take all the real numbers that belong either to
the interval (—4, 2) or to the interval (2, 00) (or to both of these intervals
if there were any such points). The above intervals are both examples
of open intervals.
The interval [—4, 2] is an example of a closed interval. It denotes
the set of all real numbers between —4 and 2, including these two end-
points (in contrast to an open interval, from which the end-points are
excluded).
From the graph we obtained above, it is easily seen that, when
f(z) = 2* — 127 + 16, then
f(z) <0 precisely when sz € (—o0, —4].
(The interval (—oo, —4] denotes the set of all real numbers that are less
than, or equal to, —4.)
There are a couple of conventions to be observed when using interval
notation. Firstly, we usually write the numbers in an interval with the
smaller number first (so for example (3,—2) would not be allowed).
Secondly we always use a round bracket at the end of an interval where
there is either co or —oo. This is because there is no such real number
as oo, but rather we are using the oo-symbol in an expression such as
(—oo, —4] purely as a convenient notation.
A function that is sometimes useful is the modulus, or absolute
value, function.
 10 Ch: 1

For a real number x, we define the modulus of z, denoted |x|, by

fit ze.0:
|x|= —z, otherwise.

[In words, the modulus of a positive real number is that positive number,
the modulus of 0 is 0, while the mod-
ulus of a negative number is the posi-
tive number obtained by omitting the
minus sign.]
For example, since 4 and 6/7 are
positive, |4| = 4, and |6/7| = 6/7,
whereas —8 and —7 are negative, so
that |— 8| = 8, and also |— x|=7. —30 —20 —10 O
By considering the four different
possible cases, with r > 0, z < 0,
y > 0 and y < 0, it can be shown that |ry| = |z| x |y|, for any real
numbers 2g, y.
The graph of f(r) = |z| is the same as the graph of y= z for x > 0,
and the same as the graph of y = —z for z < 0.
Example 8. Sketch the graph of f(r) = |x —3|—2, and determine
the values of x for which f(x) > 0.
Solution. The graph of y = |r — 3| would be the same as
the graph of y = |z|, but centred on
x = 3 rather than on the origin. The
graph of y = |z — 3| — 2 will be the
same as the graph of y = |x — 3], but
moved down by 2 units. So the graph
is made up of two half-lines having y= la—3]-2
(3, —2) as their point of intersection.
Since the two half-lines have gra-
dients +1, the lines will meet the z- ~”
axis 2 units in either direction from -
the point z = 3. Hence the lines meet
the z-axis at zc= 1 and at z = 5. It
follows that

f(z) >0 precisely when sz € (—00, 1] U[5, 00).

EXAMPLES 1
1. Sketch the graph of
(a) y=(z-1), (b) y=(2+2)?,
(c)) Wie —2)2 55: (d) y=(e+1)?-1.
 Ch:1 11

2. Sketch on the same diagram the graphs of

(a) y=a2? and y=2z?,
(‘b) ye—r? and y= =dr’.

3. Complete the square in each of the following quadratics:

(a) 2?-—62+4+7; (b) 224427 —4,
(c) 32? -—62+7; (d) 9442 —2z2?.

4. For each of the following curves, use the technique of complet-

ing the square to find the point on the curve where y reaches its min-
imum value, and sketch the curve, showing clearly the point(s) where
the curve cuts the coordinate axes:
(ayy rea?
= 42: (b) =y=a
= 27 — 3:

5. By considering discriminants, say whether each of the following

quadratic equations has two distinct real roots, exactly one real root,
or no real roots, and solve those equations which fall into either of the
first two categories:
(a) 2?-—8r+20=0; (b) 3274112 —-4=0;
(c) 227 —S2=-3
= 0; (d) 227+4+8r+9=0.

6. For each of the following curves, use the technique of com-

pleting the square to find the point on the curve where y reaches its
maximum or minimum value, and sketch the curve, showing clearly the
point(s) where the curve cuts the coordinate axes:
(a) y=2?
— 42+ 3; (b) y=3-22 —2?;
(c) y= 122 —32?; (d) y=12+482 —42?.

7. Sketch on one diagram the graphs of

(a) y=a* andy=2",
(b) 9 Siand
4 ='2°,
(cr yar rand isn”:

8. By noticing that x? — 2x? — 52 + 6 = (x + 2)(x — 1)(x — 3),

sketch the graph of y = x? — 2x? — 5z + 6, and express as a union of
intervals {x € R: 2* — 22? — 52+ 6 < O}.
9. By noticing that 2° + 27? — 11x —-12 = (2+ 4)(x# + 1)(a ~— 3),
sketch the graph of y = z* + 2x? — 112 — 12, and express as a union of
intervals {z €R: 2? + 22? — 11z — 12 > 0}.
 12 Ch: 1

10. By noticing that 2° — 42? — 3z + 18 = (x + 2)(z — 3), sketch

the graph of y = r* — 4x” — 3x + 18, and express as a union of intervals
{z ER: 2 — 42? — 3z + 18 > 0}.
11. Sketch the graph of y = |x — 2| — 3, and express as a union of
intervals {z € R: |x — 2| —3 > O}.
12. Sketch the graph of y = |32 — 6| +2. [Hint: You may find it
helpful to draw up a table of the values of y for integer values of x from
z= —2 tor =6,]
 13

CHAPTER 2
The Trigonometric Functions

In this chapter we shall introduce the trigonometric functions and

meet some of the formulae that relate them, as well as applying the
trigonometric functions to various problems.
The fact that underpins trigonometry is that if we have two similar
triangles, that is, triangles in which the corresponding angles are equal,
then the ratio of the lengths of each pair of sides of the triangles is
independent of the triangle being considered and just depends on the
angles involved in the triangle.
In the diagram shown, if the lengths of the sides of the one triangle
are a, b and c, then there is a constant k such that the corresponding
sides of the other triangle will have lengths ka, kb and kc.

G. ; Al4

Our first consideration relates to how we should measure angles.

Just as in measuring lengths, where there are two separate units in
common use, namely feet and metres, and in the measurement of tem-
peratures, where again there are two separate units in common use,
this time Fahrenheit and Celsius, so P
there are two distinct ways of mea-
suring angles. For applications such 5
as surveying and navigation, angles
are usually measured in degrees. In
advanced mathematics, in contrast, JL
it is usual to use radian measure for A
angles.
Suppose that we are given an an-
gle 6 which we have to measure. We
draw a circle C, of radius r and cen-
tre O. We also choose any point A on
the circumference of the circle. Let P
 14 Ch: 2

be the point on the circle such that the angle AOP is the angle 8 we were
given. We then measure the distance from A to P along the circum-
ference of the circle. If A to P along the circle is in the anti-clockwise
direction we let s denote this distance. However, if A to P along the
circle is in the clockwise direction we let s denote the negative of this
distance. We then define the measure of the angle (in radians) to be
given by

C= 3/5.

Angles in the anti-clockwise direction are therefore positive, while clock-

wise is the negative direction.
We know that the circumference of a circle of radius r is 27r, so
that it turns out that the radian measure of one complete revolution
is 2mr/r = 27. Equally we know that one complete revolution mea-
sures 360°. Hence we have the important relationship between the two
methods of measurement:

360° = 27 radians and hence mwradians = 180°.

From this we can see that 1 radian = (180/7)° = 57°. [Remembering of

course that 7 is approximately 3.14159... .] Equally we can calculate
that 1° = (7/180) radians + 0.01745... radians.
From the above relations, we can see that to change degree measure
to radian measure you multiply by 7/180, while in the reverse direction
to change radian measure to degree measure you multiply by 180/7.
In mathematics, some angles seem to turn up with great regularity.
One reason for this is that they are convenient fractions of one complete
revolution. It is useful to know the less familiar radian measure of these
angles.
For example, 30° = 30 x 1/180 = 7/6, while we can easily see that
45° = 45 x 7/180 = 7/4, and 60° = 60 x 7/180 = 7/3. Also 90° = 7/2
and 180° = z.
You will have noticed that we
have given up writing ‘radians’ to de-
note when we are using radian mea-
sure for angles. This is because in
Mathematics the convention is that
we always use radian measure unless
we indicate that we are using degree
measure by including the ° symbol.
From the way we defined the ra-
dian measure of an angle, it follows
that if we have an arc of a circle of
radius r, subtending an angle of 0 ra-
dians at the centre of the circle, then the length of the arc will be r.
 Ch: 2 15

Also, since the area of a circle of radius r is mr”, then the area of a
circular sector, subtending an angle 6 radians at the centre, will be

6
(5) xm ye 50
eee 6.

Now we have come to the heart of this chapter—the definition of the

trigonometric functions. There are
y
six such functions, the sine, cosine,
tangent, cosecant, secant and cotan- P(z,y)
gent.
Let 6 be a given real number,
and let C be a circle of radius r and /
centre O. Draw a pair of perpendicu- Te
lar axes meeting at the point O. Let
the positive part of the z-axis cut the
circle C at the point A. Along the cir-
cumference of the circle, measure the
distance r@, starting from A, in the
anti-clockwise direction if @ is posi-
tive. Suppose we arrive at the point P, as shown in the diagram, so
that by our earlier definition the angle AOP is 6 radians. Let the co-
ordinates of the point P be (z,y). Then we define the trigonometric
functions by

sing =%r (= 2PP

hyp
Eafe Bd)
Seb i ie=
A Ula Ghe Nite ae
ee L c =F) fcnt)

cosec
9= — |= —
y opp aero) Raea) =-
sin6

RR RTOS |S
ee ~ ¢ \ adj ~ cos6
z ( = ( =) ( 1 )
cot@ =—| =— = Page th
y opp sin 6 tan6
provided that the ratios in question exist. For example, we do not define
tan (7/2), since from the construction given above, if the angle is 7/2,
then we would have P at the point where the positive y-axis meets the
circle, so that y = r and x = 0. Hence the ratio y/z would not exist in
this case.
It is convenient to remember the values of the trigonometric func-
tions for some of the more commonly used angles. These values are
given in the following table, and should be memorised. The values can
 16 Ch:?2

In this table the letters ND are used to denote that tan@ is not
defined when 6 = 7/2. For a description of where the diagrams come
from, please see the chapter on Basic Techniques.

The coordinate axes divide the coordinate plane into four quad-
rants. From the definitions, we have the following pattern of signs for
the main trigonometric functions.
A useful way to remember this
pattern is to use the word ‘CAST’,
which stands for Cos, All, Sin, Tan. Sine, positive All —
; c E positive
This tells us which of the trigono- €osine, tangent
metric functions is positive in which negative
quadrant, listing them in the anti-
clockwise direction, starting from the
fourth quadrant.
The signs of sec@, cosec@ and
cot 9 are the same as those of cos, Tangent positive} Cosine positive

sin8 and tan 9, respectively. sine, cosine sine, tangent

Pythagoras’ theorem, applied to Nee aie negative
the right-angled triangle shown, gives
2 2 2
RS SY KI

from which, on dividing by r?, we obtain

 Ch: 2 17

In terms of the trigonometric functions this gives

cos? 6 + sin? @ = 1.
Similarly, if y # 0, then divid-
ing the above equation by y? and re-

(ya)
arranging yields
oe? r\?

y y
which translates into

cosec? § = 1 + cot? 6.

Similarly we also have

sec? @= 1+ tan? 6.

The above formulae hold for all

values of 8, provided only that the
trigonometric functions are defined.
In the above formulae, we have adopted a common convention. This
is that it is normal to write sin? 6, when we really should be writing
(sin 9). Similarly we write tan? @ rather than (tan @)°. The reason for
this convention is to reduce the number of brackets that would otherwise
be necessary to avoid ambiguity.
There is one exception to this convention. As we shall see later,
tan~! 6 has a completely different meaning to (tan 0)". So if we mean
(tan 9)? then we must write this, or 1/tan 6, or cot 0, but not tan! 0.
The same remark applies to the other trigonometric functions.
If we imagine a point P travelling around a circle of radius 1, start-
ing from the point A where the circle cuts the z-axis, then the sine of
the angle subtended by AP at the centre of the circle is given by the
perpendicular distance from P to the z-axis, with a positive sign if P
is above the axis, and negative if P is below. If we draw the graph of
y = sin@ then we see that it is as given below.
 18 Ch2

Similarly the graphs of y = cos@ and y = tan@ are as sketched

below. For the cosine graph, the value of cos@ is given by the perpen-
dicular distance of P from the y-axis, this time with a positive sign if P
is to the right of the y-axis and negative to the left. The value of tan 6
is given by the slope of the line OP.

y : Yi—tanio tes

ty Thee TAN, O ; 7 ; 2m : 0

The sine and cosine functions are periodic with period 27 (which
means that if we increase a real number 6 by 27 to 6+ 27
then the sine
and cosine are unchanged). You can see this from the way in which
the
sine and cosine are defined. If, using the real number 6 in the above
construction, we arrive at the point P, then using the real number
0+27
we will also arrive at the same point P, but after one further
revolution
 Ch: 2 19

of the circle. So we have the formulae

sin (@ + 27) = sin 8,
cos (6 + 27) = cos 8.
The tangent function is also periodic, but this time the period is 7.
Hence we have
tan(@ +7) = tané.
Furthermore we notice that the graphs of the sine and cosine functions
are the same, except that they start at different points in the cycles.
From what we saw in the last chapter, this corresponds to a shift in the
@-value. Hence the graph of y = cos (6 — =) would be identical to the
graph of y = cos 6, but centred on the line 6 = $. So we can easily see
that the graph of y = cos (6 — z) is the same as the graph of y = sin 8.
In other words, we have

cos (3= =) = sin @.

The above are all examples of the addition formulae. These

formulae, which should all be learnt, are

sin(z + y) = sinzcosy + coszsiny,

cos(z + y) = coszcosy — sinzsiny,
tanz + tany
ee 1—tanztany
The proofs of the first two of these formulae can be deduced as
shown in the chapter on Basic Techniques. The third formula is deduced
from the first two as follows:
sin(z + y)
tan(z + y) = Eset

sin z cos y + cosz sin y

cos cosy — sinzsiny
sinzcosy | coszsiny
cosTcosy coszrcosy
costcosy singsiny’
COSTCOSy cosz cosy
(on dividing each term by cos z cos y,)
_ tanz+tany
~ l—tanztany’
as required.
If we take z = y in these formulae then we obtain the formulae
sin 27 = 2sinzcosz,
cos 2r = cos? z — sin’ z,
 20 Ch: 2

Replacing cos? x by 1 — sin? ’z, or sin? z by 1 — cos” z, in the second of

these formulae, we have the alternative forms

cos? « — sin? x
cos2z = ¢{ 2cos*z —1
1 —2sin? z.
These formulae are also very useful in the re-arranged form

cos’ z = 3(1 + cos2z)

sin? x = $(1 —cos2z).
All of the above formulae should be memorized. They are known
as the double angle formulae.

Example 1. Use the double angle formulae to express cos 6z in

zr
terms of 3x, and to express cosxz in terms of on

Solution. Applying the double angle formulae with x replaced

by 3z we see that

cos? 3x — sin? 3z
cos 6z = 4 2cos* 3z — 1
1 — 2sin? 32.

Again, if we use the double angle formulae, this time with z replaced
by > we have
x SEL
cos? — = sin?
2 2
Be
cosz = <« 2cos* ——1
2 x
1 —2sin? —.
2

It is difficult to over-emphasise how important it is to know the

trigonometric formulae. It helps us to see where some of the ideas
we use are coming from, and also it helps us to know what formulae
are available, and therefore what possible options we have in trying to
simplify trigonometric expressions.
If we refer back to the graphs of the trigonometric functions, we see
that the graph of the cosine function is symmetric about the y-axis.
This means that the y-values corresponding to any two z-values which
are symmetric with respect to the origin will themselves be equal. So,
for any real z,
cos(—2z) = cosz.
Equally, the graphs of the sine and tangent functions are symmetric
about the origin. The meaning of this is that the origin is the mid-
point of the line joining the points on the curves corresponding to the
 Ch: 2 21

z-values zr and —z. Hence, for any real x for which the functions are
defined, we have,
sin(—z) = —singz,
tan(—z) = —tanz.
If we use these facts, and replace y by —y in the addition formulae,
then we obtain the formulae
sin(z — y) = sinzcosy — coszsiny,
cos (xz — y) = cosrcosy + sinzsiny,
t —t
tana Sai ee
1+ tanztany
As before, it is important to know the above trigonometric formu-
lae, which will frequently be used without comment.
Example 2. Use the above formulae to show that
cos
(a — 0) = —cos8,
and
sin(a — 8) = sin@.

Solution. Using the formula given above, with z replaced by 7

and y replaced by 0, we have
cos (a — 9) = cosmcos@ + sin7m sin 6.
Since cos 7 = —1 and sin zx = 0, it follows that cos(a — 6) = —cos8,
as required.
We obtain the other result we require by making the same substi-
tutions in the formula
sin(x — y) = sinz cosy — coszsiny.

Example 3. If @ is an acute angle for which sin@ = 5/13, find

cos 6, tan 6, cosec 9, sec@ and cot @.
Solution. One representa-
tion of the angle @ is shown in the
right-angled triangle ABC, in which C
we have ZBAC = 6, |AC| = 13 and é
|BC| = 5. By Pythagoras’ theorem, 13 5
we know that
5? + |AB|’ = 13? A = B
so that
|AB|? = 13? — 5? = 169 — 25 = 144.
Hence |AB| = 12. From the definitions therefore, we have cos @ = 12/13,
tan @ = 5/12, cosec @ = 13/5, sec@ = 13/12 and cot 6 = 12/5.
 22 Ch: 2

Example 4. Sketch the graph of y = cos 2z and also the graph

of y = 3cos (22 + z),
Solution. As z varies from 0 to $, 2z will increase from 0 to 7.
Hence the graph of y = cos 2z will start at 1 and reduce to the value
—1 as x varies from 0 to $. The graph of y = cos 2z will continue like
this, tracing out the same set of y-values as y = cos z, but at twice the
rate. It follows that the graph of y = cos 2z will have the typical shape
of the cosine curve, but will trace out two cycles for every one on the
curve y = cosz. Hence the graph is as shown in the diagram.

To sketch y = 3cos (Oz + z), we note first of all that the basic
curve will be similar to the graph of y = cos2z. The effect of the 3
is to multiply all the y-values by 3. This means that the graph will
oscillate between the values 3 and —3, rather than between 1 and —1.
We describe this situation by saying that the graph has amplitude 3
rather than 1. .
Now we have to see what the effect is of the }-term. To see this,
we rewrite the function in the form

y = 3cos (2(2+ ay) ;

which is the same as

_
y= 00 (2(e~ (-3)))
Hence the curve y = 3cos (22 + m) will be the same as the curve
y = 3cos2z, but translated so that it ‘starts’ from z = —%. We call
—% the displacement or phase shift of the curve. [Please note that
there are several alternative definitions in use of the words displacement
and phase shift. It is important, when using these words, to be certain
that other people are using them with the same meaning as you.] The
period of the curve will be 2* = 7, since the graph repeats itself after
TT.
 Ch: 2 23

y = 3cos
(2z + £)

J3 / /

—3

Example 5. Express the oscillation y = —\V/3sin 3z + cos3z as a

single sine curve, and determine the amplitude, displacement and period
of this oscillation.
Solution. [When faced with an expression which has the form
asin px + bcos pz, (where a, b and p are constants), we start by taking
out as a factor the value of a? + b?. We then try to find a real num-
ber @ such that cos@ = a/Va? + b? and sin@ = b/Va? + b?. We can
then rewrite asin pr + bcos pz as Va? + b?(sin pr cos 6 + cos pz sin 6], or
as Va? + b? [sin(pz + 6)], or as Va? + 6? [sin (p(x + (8/p)))], which has
amplitude Va? + b?, displacement —6/p and period 27/p.]
If y = —V3sin3z
+ cos3z, then A ENE +12 = /3+4+1 = 2,
and so y = 2 [(—V3/2) sin 3z + (1/2)cos3z]. Let the real number 0
be such that cos@ = —V/3/2 and sin@ = 1/2. Since sin@ is positive
and cos@ is negative, we need to have @ in the second quadrant. By
comparing the situation we have now with our earlier diagrams, we can
see that the real number @ is 57/6. Hence we have
y= 2[sin 3z cos (57/6) + cos 3z sin (57 /6)|
= 2[sin (32 + 57/6)]
= 2[(sin (3(z + 57/18))]
= 2[sin (3 (x — (—57/18)))]
Hence the sinusoidal oscillation has amplitude 2, period 27/3 and phase
 24 Ch: 2

shift —57/18.

Example 6. Use the addition formulae to find tan ts

Solution. Since = = tax 180)* = 7 x 15° = 105°, which we

can break up as 105° = 60° + 45° = = + 4, we can use the addition
formula. We have
Gs , 0%
tan 1 = tan (eaa >)

_ tan} +tan a
1—tan ¢ tan 7
ven
“hens

_ (V3 +1) x (1+ V3)

~ (1 = v3) x (1+ v3)
(multiplying top and bottom by 1 + V3 to rationalise the denominator)
_ V3 +1434 V3
1-V3+ 3-3
4423
FWietaD
=—2~— V3.

Example 7. Simplify se Jet WBS oy

sin z cos £
Solution. Using the appropriate trigonometric formulae we have

cos3z sin3z cos 3x cosz — sin3zsinz

sin £ COs £ sin © cos £
cos (3x + x)
sin © COs
cos 42
~ sinz cosa
_ 2cos4z
~ Qsinzcosz
_ 2cos4zr
sin2r —

Notice that it is very important that we should be able to use for-

mulae such as cos 3z cos x— sin 3z sinz = cos(3z + 2) in this “reverse”
form. All formulae should be remembered in this form as well as in the
more usual form cos(3z + 2) = cos 3zcosz — sin3z sin z.
 Ch: 2 25

Example 8. Establish the trigonometric identity

sin 82 = 4sin 22 cos 2z — 8sin® 27 cos 2z.

Solution. [When we require to establish a formula such as the

above, we must start by trying to simplify one side of the equation
(usually the more complicated side) into the form of the other side. The
thing you should not do is to assume the identity holds and use it in the
course of the proof.]

The right hand side is = 4sin 2z cos 2z — 8 sin* 2x cos 2z

= 4sin 2z cos 2z(1 — 2sin? 22)
= 2 x (2sin 2z cos 2z)(cos 4r)
= 2sin 42 cos 4z
= SInS7,
= the left hand side,
as required.
Again, notice the use of the formula
1—2sin?6=cos20 with 6=22

and the double use of

2sinOcos@=sin26 with @=2zr andwith @=4z.

Example 9. In the diagram shown, which represents a kite, cal-

culate the lengths (to the nearest centimetre) of the pieces of wood
required to form the diagonals of the kite.
Solution. From the symmetry of the kite, ZAX zBwill be a right-
angle. Hence in the triangle AX B we
have sin2BAX= |BX|/|AB|, and
cos 2BAX = |AX|/|AB|. Hence we
find that |BX| = 71 x sin25°, and
also that |AX| = 71 x cos25°. We
now need to calculate the distance
|ICX|.
We find this distance by consid-
ering the right-angled triangle BXC.
Now tanZBCX = |BX|/|CX|, so
we have |CX| = |BX|/tan ZBCX =
71 x sin 25°/ tan ZBCX which equals
71 x sin 25°/ tan 40°.
We can now use our calculator to find the distances BDI and |AC|.
By symmetry, |BD| = 2x|BX|= 2x71xsin 25°= 60cm, to the nearest
 26 Ch? 2

centimetre. Also we can see that |AC| = |AX|+|CX| = 71 x cos 25° +

71 x sin25°/ tan 40° = 71 x [cos 25° + (sin 25°/
tan 40°)] = 100cm, to
the nearest centimetre.
If you know the Sine Rule, you will realise that it provides a much
easier way to find the distance |AC|. If you don’t know the Sine Rule,
you might like to look in the chapter on Basic Techniques where it is
explained.
From the Sine Rule, we have |AC|/sin 115° = |AB|/sin 40°. Hence
|AC| = |AB| x sin115°/ sin 40°, which gives the value we obtained be-
fore.
We could show that cos 25° + (sin 25°/ tan 40°) = sin 115°/ sin 40°,
which is what is required to show that the above answers are the same.
This provides a useful exercise in the use of the trigonometric formulae.
EXAMPLES 2
1. For each of the following angles 6, write down sin 6, cos 6, tan 8,
cosec 8, sec @, cot 9, if these are defined. Write N.D. for any which are
not defined.
(Se Deas (ec) 3, (4) 3, (antame:
(f) an (g) 2, (h) TT, (i) —T, (j) ar

(k) 3. Uae (nm) Giges = (eee a =(ce). =

F-
2. Ifsin? = ; and tan @ is negative, find the values of cos 0, tan 6,
sec 9, cosec 6 and cot 8.

3 if tang — 3 and @ is not in the first quadrant, find the values

of sin 6, cos 8, sec 6, cosec 8 and cot 6.

4. Sketch the graphs of

(a) y=2sin(x— 2%), (b) y=3cos3z,
(c) y=-—2tan2z, (d) y=3sin (32 — 3).
Also give the period, phase-shift and (except for (c)) the amplitude of
each of the oscillations.

5. A flag-pole C’P is held in a vertical position by two guy-lines,

AP and BP, which are fastened to the ground at A and B so that the
base C of the flag-pole is directly between the points A and B, and so
that the distance between A and B is 30m. Given that the angle CAP
is 55°, that the angle CBP is 46° and that the ground between A and
B is horizontal, find the height of the flag-pole. |
6. A ship starts from a position A and travels a distance of 5.9km
due East to a point B. It then alters its course through an angle of 45°
so that it is travelling in a North-Easterly direction. It continues on
this course for a distance of 7km until it reaches the point C. The ship
again alters course, this time through an angle of 115°, so that now the
 Ch: 2 21

ship is heading slightly to the north of due West. The ship continues on
this course until it reaches the point D, which is due North of the point
A. Find how far north of the point A the ship is when it is at D.
7. Twelve equally spaced marks are made on the circumference of
a circular fly-wheel. Given that the straight-line distance between two
adjacent marks is 3.52 cm, find the radius of the fly-wheel.

8. A surveyor is standing at a point P and wishes to determine the

height of a tall building ABC. The point A is at the base of the building
and is at the same level as the surveyor. The point B is 20ft vertically
above the point A, while C is at the
top of the face of the building. Given
that the surveyor measures that the
angles of elevation of B and C are 36°
and 71° respectively, find the height
AC of the buiding.
9. Part of the pattern of roads
in a particular town is as shown in
the diagram. Given that from A to
C is 340m, that CD is 160m and
that AE is 81.6m, find the distance
DF.
10. Use the fact that 5 = [ — ] to find sin75, cos 75, and tan 5.

11. Use the fact that & = {+ to find sin G cos u, and tanae
12. Use the fact that | = $ + | to find sin F, cos F, and tan F

13. By writing 30 = 20 + 6, show that

sin 39 = 3sin6 — 4sin® 6,
and
cos 36 = 4cos® 6 — 3cos 6.

14. By writing 40 = 20 + 20, show that

sin 40 = 4sin 6 cos* @ — 4sin® 6 cos 8,
and
cos 49 = 8cos*
6 — 8cos?
6 +1.

15. Show that

sin @ cos @
(a) = cosa, (b) = sina,
tana cot a
(c) sinfcotB = cosP, (d) tan #cosec B = sec B.
 28 Ch: 2

16. Show that

(a) sin @(cosec 0 — sin @) = cos? 8,
(b) sec @cosec @ — cot 6 = tan 9,
l—cos¢? __— sing
(c) sing 1+cos¢’
sec ¢ + cosec
(d) cosecd
=1+tand¢.

17. Show that

(a) sin(x+y)sin(r# —y) = sin? x — sin? y,
(b) cos(a+)+sin(a— f) = (cosa + sina)(cos# — sin f),
1 — tan? +
(c) oe
cos2y nar
(4) sin 3y __ cos 3Y ao
sin ¥ cos

18. In an investigation concerning the properties of light, the equa-

tion j ‘
a( =f -1) = 3 (SF +1)
tan @ tan @

was found. Show that this can also be expressed as

Bsin(B+a) = Asin(B—-a).

19. The displacements, y, and y2, of two waves travelling through

the same medium are related to position z and time t by the equations
y, = Asin ((t/T) — (x/X)) and y2 = Asin ((t/T) + (z/X)). Find an
expression in the simplest possible form for the displacement y; + y2 of
the combination of the waves.
20. In Calculus, the expression
tan(z+h)—tanz
h
is discovered when using one method to find the derivative of tanz.
Show that this can be expressed as
sinh
sec x sec(z + h).
h
 29

CHAPTER 3
The Inverse Trigonometric Functions
In this chapter we shall introduce the inverse trigonometric func-
tions, which are introduced to provide an answer to questions such as
if sinz = 1/2, what is z?
If we look at the graph of y = sinz, we see that there are a large
number of values of x for which sinz = 1/2. These are the z-coordinates
of the points of intersection between the graph of the curve y = sinz
and the line y = 1/2.

One of the points of intersection is given by x = 7/6. Then from

the fact that the curve repeats itself exactly every 27, there will be roots
at 2x + 7/6, 4x + 17/6, 627 + 7/6, ... . Also from the fact that from
xz = 1/2 to x = 7m the curve descends exactly in the reverse of the way
it ascended between x = 0 and x = 7/2, we see that there will be a
second series of solutions, given by — 7/6, 27+ -—7/6, 4r+7—7/6,
67 + x -- 1/6, etc. Equally, there will be further roots to the left of the
y-axis. These are given by —27 + 7/6, —4n + 1/6, —6%7 + 7/6, ... and
—2r+n—7/6, -4n +7 —-—7/6, —67 + 7 — 7/6, etc.
The set of all solutions of the equation sinz = 1/2 can therefore be
expressed as |
{rxER:2=2kr4+7/6
ort =2r+n-—7/6:kE ZZ}.
We can see that all of these solutions are related to the basic solution
oma 0,
Recall that our definition of a function required that, for every value
of the independent variable, we obtain a unique value of the dependent
variable. So in defining the inverse sine function we want to have a
unique answer to the question ‘if sinz = 1/2, what is 2?’ We agree
 30 Che s

to concentrate our attention on the region of the graph of y = sinz for

which —1/2 < x < 1/2. [The reason for choosing this section of graph is
that it is the region closest to the origin for which the function y = sinz
takes on all the y-values between —1 and 1, exactly once each.]
We now come to the definition of the inverse sine function. For
—1 <2 <1 we define sin! z or arcsinz to be the unique real number
y with —17/2 < y < 7/2 such that siny = z.
We can think of sin~! z as giving us the real number y in the range
[—2/2, 7/2] whose sine is z.
Since sin0 = 0 and 0 € [—7/2,7/2], we have sin~'0 = 0, while
since sinw/2 = 1 and 1/2 € [-71/2,7/2], we have sin7’1 = 7/2.
Again sin(—7/3) = —V3/2 and —7/3 € [—71/2,7/2], so that we have
sin7!(—/3/2) = —7/3. In each of the above examples, the angle chosen
belongs to the interval [—1/2, 7/2], and so sin’ z took the “obvious”
value in each case. We shall later see an example where this is not the
case.
By a similar argument to the one we used earlier, we see that, for
any real number / with —1 < / < 1, the set of all solutions of the
equation sin x = | is given by
{reER:2=2kr+sin-'lorgc=2kr+n-sin l:kEZ}.

The curly brackets we are using here denote that we are specifying
a set. c € R denotes that we are specifying z to be a subset of the set of
all real numbers, while k € ZZ denotes that k is allowed to vary through
the set of all integers (including 0). The colon stands for ‘such that’.
Finally the condition —1 < 1 < 1 specifies all those real numbers that
simultaneously satisfy the two conditions | > —1 and / <1. In interval
notation, we could write the condition —1 <1 <1 as/€ [—1,1].

The definition of the inverse cosine function follows the same

pattern as for the inverse sine function.
For —1 < xz < 1, cos~'g, or arccosz, is defined to be the unique
real number y, with 0 < y < 7, such that cosy = z. [The reason for
imposing the restriction 0 < y < 7 is to obtain a set of values of.ysuch
that cos y takes on all the values between —1 and +1 precisely once each
as y varies through the set of values we have chosen.]
Since cos0 = 1 and 0 € [0,7], we have cos~’ 1 = 0. Since cos 7/2 =
0 and 7/2 € [0, x], we have cos~1 0 = 1/2, while since cos 31/4 = —1/./2
and 37/4 € [0,7], we have cos!(—1//2) = 37/4. The deductions we
have made are valid here because the real numbers of which we are
taking the cosines are all in the interval [0,7].
By examining the graph of y = cosz, we can see that for any real
number /, such that —1 < 1 < 1, solutions of the equation cosz = / are
given by x = cos! 1 (the value between 0 and 7), and by z = —cos~? 1.
Then, since the graph repeats itself every 27, there are the solutions
 Ch: 3 31

2m —cos7!1 x

given by 27+cos~! 1, 2r—cos—! 1, 4a + cos! 1, 4a —cos—! 1, 64+cos~! 1,

Gx,— cos VI, ...%
In set notation, the set of all solutions of the equation cosz = ! is
given by
{zER:¢=2krtcos'l:kEZ}.

The inverse tangent function is defined in the following manner:

for zc € R, tan! z, or arctan z, is defined to be the unique real number
y, with —1/2 < y < 1/2, for which tany = z.
Since tan (1/3) = V3 and 7/3 € (—1/2, 7/2), we can deduce that
tan-! /3 = 7/3. Since tan(—7/4) = —1 and —7/4 € (—72/2,7/2), we
similarly see that tan~1(—1) = —7/4.
y So fiStiane

Given | € R, the set of all solutions of tanz = | is given by

{rER:c=kr+tanIl:keZ},
 32 Ch: 3

since there is the unique solution tan! / lying between —7/2 and 7/2,
and then the other solutions arising from the fact that the graph of
y = tanz repeats itself with period 7.
Now let us consider the graph of y = sin™’ z. We start by drawing
up a table of values to see what is happening.

You will see from the table that

the values of y = sin™! z are obtained
by interchanging the z- and y-values
for y = sinz. It follows from this
that the graph of y = sin™}z will
be the reflection in the line y = 2
of the graph of y = sing, or, to be us
—|
more precise, the reflection in the line
o)

y = z of that part of the graph of y=sing

y =sinz for which z € [—7/2, 7/2].

In a similar way, the graph of
y = cos! z will be the reflection in
the line y = z of the part of y = cosz
for which 0 < z < 7, while the graph of y = tan7! z will be the reflec-
tion in the line y = z of the part of the graph of y = tanz for which
—nt/2<a4< 7/2. 7

oer ew woe ee

Mienie 0),¢. eh 6)‘ele 6 6

y=tanz
 Ch: 3 33

From the definitions of sin~! z, cos~! z and tan~! z, it follows that

sin(sin~'z) =z (Vz € [-1,1))
cos(cos-'z)=2 (Vz € [-1,1])
tan(tan-}z)=2 (Vz ER).
On the other hand the equation sin~'(sinz) = z only holds when
z € [—%, 3], while similarly cos~!(cos x) = z only holds when z € [0,7]
and also tan~'(tanx) = z only holds when z € (—3, 3).
To show that sin~1(sinz) = z is not always true, consider the value
xz = 2radians. Then using a calculator we have
sin~'(sin 2) = sin~1(0.9092974) = 1.1415927 # 2.

Example 1. For z € (—1,1), express tan(sin7! z) in a form not

involving the trigonometric functions.
Solution. We shall start by cosidering the case where 0 < z < 1.
Let 9 = sin! z, so that we need to find tan§. Now since sin@ = z,
we need to consider the right-angled
triangle shown. The lengths of the
sides of this triangle have been cho-
sen so as to give sin@ = z. In this
diagram, we know that the third side
is of length V1 — 2?, by Pythagoras’
theorem. Hence tan@ = z/V1 — z?, ;
which can be rewritten in the form
tan (sin7! r) = 2/V1 — 2?.
The case in which —1 < « < 0 Tose
is dealt with in a similar manner. In
this case, the angle 6 will be in the
fourth quadrant and we have a similar, but not identical, diagram to
consider. The final result is however the same. You might like to work
through the details of this case yourself.
In both cases we finish up with the same answer,

tan (sin x) = era

Vila?

Example 2. Solve the equation

sin (32 + >) = mess

Solution. | The routine that we are about to use in the following

solution should be carefully noted. We shall find the general solution
of the equation by using the formula for the general solution to obtain
3z + {. We shall then solve for z to obtain the general solution for z.
 34 Chas

If we were to start by trying to find a particular solution for z and then

trying to get the general solution by just adding in multiples of 27, we
would find some, but not all, of the solutions.
Now

when

3
32 + ; = 2kn + sin (-2)

3
or 3a + 7 = 2ke +1 — sin” (-2) (k € Z)

which is equivalent to
™ T T us
8a + —7 = 2kr
kr -— quid
— st +7—-= ka +a + =5 (kkeEZ
€ Z)

which holds when

he 40 TK. 5%
or == 2kr -=>--—
yaks 3x = ka +a +3=--—Z (kkeZ
€ Z)

which is the same as

2k Oh OOK Qka ini ger Ae
=—---— = —+-+--— ke Z
BA gag OS gmail gaky1a ts eon,
which we can express in the form
Zhe 2kr.. -13a
Be walks ee = ke Z).
Se ie IGS” isa ba
Hence the general solution of the equation is that

sin (32 + *) — a

precisely when
Qkxn 36 Tx Qkx 137
e{—ea
—-—:k euhu{ sees
eeeWee) = 1b }

Example 3. Solve the trigonometric equation

2sin 2z cos 3x = sin 2z.

Solution. Rearranging the equation, we see that

2sin 2z cos 3z = sin2z

holds when
2sin 2z cos 3x — sin2z = 0
 Chi3 35

which is true when

sin 22(2cos3z — 1) = 0

which is equivalent to

sin2z7 =0 or 2cos3r—1=0

which holds when

sin27 =0 or cos3r= ;

which is true when

Qe =2ker4+sin7'0 or 22 =2kr4+7-—sin—!0

or. 32 = 2kx +cos7' (5) (k € Z)

which can be expressed as

20 = 2kR ZOU p28 =k Or 32 = kr + = (k € Z)

or as

2=kr or a= kr+ = or = tt (k € Z).

Hence the general solution of the equation

2 sin 2z cos 3z = sin 2z

is that z satisfies the equation precisely when
k
refin:keabul erst ikem}ul at irenh,
There are a couple of important points to make about the above
solution.
(i) Firstly we should note that we could not start the solution by
dividing both sides of the original equation by sin2z. Had we done this
then we would have thrown away a number of solutions of the equation.
You may only divide an equation by something if you know that the
something ts non-zero.
(ii) Secondly we should note that we used the fact that a product
is zero precisely when one (or more) of the terms in the product is zero.
While this is valid, note that it is not true that if, for example,
(c — 3)(x + 1) = 5 then either x -3 = 5 or r+ 1 = 5. The way to
solve an equation such as (z — 3)(4 + 1) = 5 is to expand the brackets,
subtract the 5 from both sides of the equation, and then try to solve
the resulting quadratic equation. If we do this, we would find that the
solutions are z = 4 and xz = —2.
 36 Ch: 3

Example 4. A hemispherical bow] of radius 7.58 cm is filled with

water to a depth of 5.52cm. Assuming that the bowl starts from a level
position, find the angle through which the bowl can be tilted before
water begins to spill out.
Solution. Let a be the angle we require to find. Then if we con-
sider the diagram, we see that sina =
2.06/7.58, so that a = sin’ (298),
and from the calculator we see that
a = sin~’ (0.27176781) = 0.2752, to
4 significant figures.
Notice that the final answer is
given as a number of radians. Had we
used the calculator in degree mode,
we would have obtained an answer
of 15.77°, correct to 4 significant fig-
ures. )
EXAMPLES 3
1. If y = sin4z + cos4z, express the oscillation as a single sine
curve, and determine the amplitude, displacement and period of the
oscillation.

2. If y = 8sin2z — 15cos2z, express the oscillation as a single

sine curve, and determine the amplitude, displacement and period of
the oscillation. [You may assume that the angle @ in the first quadrant
with sin@ = + is 1.081 radians.]
3. The current 7 in an electric circuit is expressed in terms of
time t by i = 4e~** (sin 3t — V3cos3t). By re-expressing the term in
the brackets as a single sine term, find the time when the current is first
zero.

4. Express cos (tan7! z) in a form which does not involve the

trigonometric functions.

5. Express cos(sin~’z) in a form which does not involve the

trigonometric functions, valid for z > 0.

6. Ifsin@ = (x + 2) /3, find cos6 and tan@ in terms of z.

7. If tan@ = (a — 3) /2, find sin@ and cos@ in terms of z.
8. Evaluate
(a) cos~? (sin Z), (b) sin~’ (cos 22),
(c) sin? (sin 5), (d) tan! (cot 3).
 Ch: 3 37

9. Evaluate
(a) sin (sin? (-2)), (b) cos (cos? (-¥5)),

10. Find the general solution of each of the following equations

(a) cos zr = ya) (b) sin 2f = ,,
(c) sin3z = cos 47, (d) sin(4r+%) = FR:

11. Find the general solution of each of the following equations

(a) 2sin2zrcos3r = —V3cos3z, (b)) sin 47 = sind,
(¢) “sin ai-'cos'e = 15 (d) V3sin2z —cos2zr = 1.
(Hint. In (c) and (d), start by expressing the left hand side of the
equation in terms of a single sine term.]
12. A mathematics lecturer is trying to fasten together four pieces
of wood of equal lengths to form the sides of a square drawer. The
mathematician fastens together the pieces of the drawer so as to form
a parallelogram. He checks whether it is a square by measuring the
diagonals, and finds to his dismay that the diagonals measure 74cm
and 72.5cm. Find the larger angle of the parallelogram, and hence find
the angle through which the one side should be rotated in order to form
the desired square drawer.
13. A rope of length 55 feet is fastened to a harbour wall at a height
of 20 feet above sea-level. The other end of the rope is fastened to a
rowing-boat, at a point which is 3 feet above sea-level. Assuming that
a strong wind is blowing the rowing-boat away from the harbour wall,
so that the rope remains taut, find the angle between the rope and the
horizontal.
 38

CHAPTER 4
Differentiation

Calculus has held pride of place for the past three hundred years as
the most important branch of Mathematics, particularly with regard to
applications. Calculus consists of the twin operation of differentiation
and integration. These operations were introduced by Newton and Leib-
niz, working independently, at about the same time. There followed a
long and acrimonious dispute between the supporters of these two great
mathematicians, each group claiming that their champion was the true
creator of this important branch of the subject.
In this chapter we shall study differentiation, leaving integration
for later in the course.
Differentiation examines the way in which a change in the value of
the independent variable z will affect the value of the dependent variable
y, when y is given as a function of z.
We start by considering a particular example, in order to try to get
an idea of what happens in the general iceyice Let P be the point
with coordinates (2,4) on the curve y = r*. Let Q be the point with
coordinates (21, y,) on the curve, and let m Mencte the gradient of the
line PQ passing through the points P and Q. It will be convenient to
let h = x; —2 and k = y; — 4, so that h and k represent the increments
(or increases) in z and y respectively.
Consider the following table:
(A ON MEead)
ee
Pew —4 [12 [5| 2.25 |041 [0.0407 |
| Peeaae
SH
{ati fesfonPoor
oe —2

From this table we see that as Q approaches P, so that the value

of h tends towards the value zero, then the gradient m of the chord PQ
will approach very close to the limiting value 4.
The above table has simply suggested this as a deduction from
calculations for the specific values of h we chose, all greater than zero.
Now let us try to establish the same fact in general, for any non-zero
value of h, whether positive or negative.
Let Q be the point on y = zx? with z-coordinate 2 +h. Then the
y-coordinate of Q will be (2+h)? =44+4h+h?. Therefore k = y, —4=
(4+4h+h?)-—4 = 4h +h?. Hence the gradient of the chord PQ
through the points P(2,4) and Q(2 + h,(2+h)?) will be
4h +h?
=4+h.
h
 As we let h > 0, so that Q ap-
proaches close to the fixed point P,
the value of the gradient will become
arbitrarily close to 4. This bears out
the conclusion we drew from the ta-
ble above.
As the point Q approaches very
close to the point P, then the chord
PQ will become the tangent to the
curve y = z? at P. We can think
of the tangent as being a chord that
touches the curve at two coincident
points at P. The tangent to the curve
y = 2? at the point P(2, 4) will have
gradient 4, the value we obtained in the above process in the limit as h
tended to 0.
We could obviously carry out the same calculation for different
points on the curve y = z”, in which case we would expect to obtain
different values for the gradient m of the tangent.
We could also hope to carry through similar calculations in the case
of functions other than y = z?.
Assume that y = f(z) expresses | Q(x +h,f(x+h))
y in terms of z. Then the derivative
of y with respect to zr, denoted by
d
f'(x), or by a! is given by

dy _,. f(z+h)— f(z)

eee h
From the diagram we see that
d
“= gives us the gradient of the tan-
x
gent to the curve y = f(z) at the
point P(z,y). To find the tangent
to the curve, we take the chord PQ,
passing through the points P(z, f(x)) and Q(r+h, f(z+h)), and we let
h become so small that the points P and Q are indistinguishable. Thus
we can think of the tangent as meeting the curve at the two coincident
points P and Q = P.
We describe the process of finding the derivative of a function as
differentiating the function.
 40 Ch: 4

The following table gives the derivatives of the functions which arise
most frequently in mathematics:
f'@)
|
[eked pe né€R,n #0, nfixed

There are various possible ways of combining together functions.

The following rules tell us how to find the derivatives of these combi-
nations. In these rules, f(z) and g(z) will denote any differentiable
functions—that is, functions which have derivatives—and c will denote
an arbitrary constant.

<(ef(2)) = ©
<(fa) t9(2))= Fs.
The Product Licpis
d a df dg
aaf(2) x 9(z)) = g(a) x o- + f(z) x F
The Quotient Rule.

4 (fl) 20% o - j(2)x FB

a(2) War
Example 1. Find the derivative of x* + 42° —7+ =.

Solution. We shall think of Se as being 9z~3, and then we shall

use the above rules for siffereut BURAa sum of functions and a constant
times a function. Then
d 9 d =
x (#?+42°-74 5) = 4, (440° -7 492 )

= 5d (0°) +42d (29) - 2 +92 = (2)

= 327 +4 x 524-049 x (-3)z~4
= 327 + 20zr4 — 272-4
27
 Ch: 4 41

Example 2. Find the derivative of

(has _ 22s) COS Z.

Solution. Since this time we have a product of cosz with an

expression similar to that in Example 1, we must use the Product Rule
to differentiate the product. So we get

< {(7x? — 22~*) cosz}

= (142 + 102~°) cos x+ (72? — 227°) (—sinz).

Example 3. Find the derivative of e” sin zr.

Solution. Again using the Product Rule, we have
d
aes {e7 sinz} = e’ cosz +e’ sinz.

Example 4. Find the derivative of

x? + 32
r344 0

Solution. | This time we must use the Quotient Rule, with f(z) =
z? + 32 and g(x) = x? + 4. So we have
d (zx? +32
dx (z3+4 )

(x? + 4) x Ce + 3x) — (2? + 32) x ass + 4)

2) dx dx
+a
_ (2? +4) x (22
+ 3) — (2? + 32) x 32?
=: TirahWa sda(ciSEs le teen ta

_ (224 + 82 + 3x° + 12) — (32* + 92°)

a Rocanqinpgtepegla
Rem na 20
_ 2r4 + 82 + 32° + 12 — 324 — 92°
= RESP RGTIT Lo Be

—z* — 62° + 82+ 12

STS 4 4p |
We have quoted the Product and Quotient Rules without attempt-
ing to justify them. You might be interested in consulting the chapter
on Basic Techniques, where we give an indication of how to prove these
results. For now, we shall content ourselves with an example to show
that if we ignore the Product Rule, then the rules for differentiation
would not work.
 42 Ch: 4

d
If y = x”, then we can show from the definition that - er.
However if we regard x? as rxz and differentiate by multiplying together
the derivative of each term, then we would obtain the different, and
d
wrong, answer 7 (x) x aa (2) =1x1l=1.

We write —(???) to denote that we are differentiating what is

Fr
contained within the bracket. A common notational mistake is to write
down what is to be differentiated and then write ‘=’ the derivative. If,
for example, asked to differentiate r? please never commit a horror such
as writing z? = 2z, when of course what is intended should have been
written as — (x?) ate
dz
There is one further important way of combining together two func-
tions, namely by composition. Suppose that f(z) and g(x) are two
functions of z. Then the composition (f o g)(x) is defined to be the
function given by
(f og)(z) = f(g9(z)).
What this means is that we take a value of z, use the formula for g
to calculate g(x), and then we apply the formula for f to the value g(x)
to calculate the value of f (g(z)).
Suppose, for example, that f(z) and g(x) are the functions given
by f(z) =sinz and g(x) = 3z° —5, respectively. Then the composition
(f 0 g)(z) is given by

(f og)(z) = f(g(z)),
so that
(f og)(x) = f (32° — 5) =sin (32° — 5).
This is because the function g(x) acts by taking any number, cubing
it, multiplying by 3 and then subtracting 5. The function f(x) acts
by taking any real number and taking the sine of that number. So the
function f takes the real number 32° — 5 and gives us the real number
sin (3z* — 5). Hence the effect of the composition f og is to start with
the value x and to yield sin (32° — 5).
Example 5. Find functions f(x) and g(x) such that each of the
following can be expressed as a composition (f 0 g)(z):
(1) A,y(x) = tan (9z — 4); (2a h(a) = In (tany bz):
(3) hg(x) = sin‘ x.
Solution. (1) hi(zx) is obtained by first calculating 92 -- 4 and
then taking the tangent of this number. We will obtain precisely this if
we take g(r) = 9x — 4 and take f(z) = tanz.
(2) To find h(x) we first need to calculate tan~! z and then
take the natural logarithm of this number. It follows that we need to
 Ch: 4 43

take g(z) = tan~'z and f(z) = Inz. With this choice we see that
(f 0g)(x) = ho(z), as required.
(3) To find h3(z) we need to recall that sin‘ z is a shorthand
for (sinz)*. The calculation of sin‘ z will therefore involve finding sinx
and then raising this to the fourth power. So we let g(x) = sinz and
f(z) = 2%.
Students frequently confuse the product with the composition, and
yet they are very different ways of combining functions. Note for ex-
ample the difference between (sinz)(tanz) and sin(tanz). The first of
these involves calculating sin z and tan z and then taking their product,
while the second involves finding tan z and then taking the sine of this.
So the second of these functions is a composition.
One way of distinguishing between the two is to imagine that you
are using a calculator to evaluate a function. If you use the multiplica-
tion button, then you have a product. If on the other hand you simply
use function buttons, then you will have a composition.
The rule for differentiating a composition is:
The Chain Rule

© (F(a(2))) = F(a(2)) x a'(2)

So to differentiate a composition f og, we firstly differentiate f with
respect to its variable g(z), and evaluate it at g(x), and then multiply
by the derivative of g(x).
The Chain Rule can also be remembered in the form
df df _ du
dx du dz’
where u denotes the function g(r) we used in the earlier version.
Example 6. Differentiate each of the following functions:
(1) In(22 + 5);
(2) sin(2x? — 42 +5);
(3) .sin® z.
Solution. (1) We should think of In(2z + 5) as being the com-
position of the functions g(x) = 22 +5 and f(z) = Inz. Then we
d
have yf = a so that <(in(22 +5)) = tee es= 2, and
dz ce 2r+5 dz
; d 1
hence we have, by 5 nga Rule, ae, (In (2z + 5)) = oF LE x 2, so on

simplifying we have °F (In (2z + 5)) = Eat

(2) To find < (sin (22? — 42 + 5)), we think of this function

x
as the composition of f(r) = sinz and g(r) = 22? — 42 +5. Then
 44 Ch: 4

2 = cos Zz, so that i (sin (22? — 42 +5)) =cos (22? — 42 + 5), while
dg 4x — 4. Hence, using the Chain Rule, we have
rave
KD
ia (sin (22? — 42 + 5)) = (cos (22? — 42 + 5)) x (42 — 4)
dr
= 4(r — 1)cos (2x? —4r + 5).

(3) To find £ (sin’ x), we need to regard the function as the

:
composition of f(z) = z? and g(x) = sinz.
af
Now Ta 3x7, so that

a = 3g? = 3(sinz)”, and us = cosz. Hence, by the Chain Rule, we

9 L
have
dey :
TE (sin? x) = 3sin? z x cosz.
‘Tr
Normally when we have identified that we have to differentiate a
composition, so that we have to use the Chain Rule, we do not bother
formally to separate the component parts in writing out the solutions,
but rather we try to figure out this decomposition mentally.
Example 7. Differentiate each of the following functions:
(1)
27>
(2) zr+7
(3)
(52° — 7x? — 42 + 8).
d
Solution. (1) 7 (eS*) eye" wax oi se

(2) Here we see that

satd peal 1ites 9 |S cekd ;

-1/2
dx ee rt=| dx ((32 DD )
= (—1/2)(32 + 7)~3/? x < (32 + 7)
= (—1/2)(32 + 7)~3/? x 3
= (—3/2)(3z + 7)73/?,

(3) In this case we have

d
ae (52° — 7x? — 42 + 8)4)

= 4(52° — 72? — 42 + 8)° x “.(5c° — 7x? — 42 + 8)

= 4(52° — 7x? — 42 + 8)° x (152? — 142 — 4).
 Ch: 4 45

NOTE. It is important that we avoid writing things that can have

more than one interpretation. To this end, we put in brackets to
avoid any posssible misunderstandings. For example, in differentiat-
ing cos(2z — 3) we would obtain (—sin(2z — 3)) x 2. Had we missed
out the brackets surrounding the — sin (2z — 3)-term, then our answer
could have been misinterpreted as — sin((2z — 3) x 2), which is a very
different answer.
There is a problem though in writing brackets to avoid ambiguity,
and that is that we can finish up with expressions involving large num-
bers of brackets which are quite difficult to decipher. To reduce the
numbers of brackets, we adopt some conventions.
We have already met the convention of writing sin? x rather than
(sinz)*.
Another convention is that when taking a function such as sine,
cosine, In, etc., we interpret it as meaning the sine or whatever of the
symbol immediately following the sine, except that where the symbol is
multiplied by a number (and in multiplication we include division), or
where the symbol is raised to a power, then we include the extra term
or terms before taking the sine.
So for example sin3z? would mean that we should calculate 32?
and then take the sine of this expression. However sin3z? + 4 would
mean that we should again calculate sin3z? and add 4 on to this. In
other words sin 3x? + 4 would be better written as 4+sin 3x. If we had
meant to take first the value of 32? + 4 and then take the sine of this,
then that would have to be written as sin (3x? + 4).
Note that e? for example, is taken to mean el), the exponential
of z*, rather than the other possible interpretation.

Using the rules for differentiation that have been given above, we
can find the derivatives of some more of the elementary functions that
we have met.

Li i xlx3 1
Te

We prove the above results as follows:

 46 €h:i4

(cos z)(cos z) — (sin z)(— sin z)

a [cos x]?
cos? x + sin? x
[cos x]?
1
e [cos x]?
| 2

a (=)

= sec? a,

as required.

d d 1
ae (secr) = me (=)

Ze (cos z)(0) — (1)(—sinz)

[cos x]?

~~ [cos 2]?
(1 ) (sin z )
= x
cos x cos 2
=seczrtanc.

The proofs for the derivatives for cosecz and cot z are similar to
the ones given above.

anos d
aa (4 Mee Gp (exP(z Ina))

= exp(z Ina) x - (clna)

=(a- )ocin a:
as required. Note the use in the above of the formula for expressing a”
in terms of the exponential function. Note also that we have written
exp(z Ina) rather than the equivalent alternative form e*!"*, purely for
the sake of variety.
If we use the formula for the change of base in logarithms to express
log, z in terms of log,, then we can differentiate, and obtain

Fp
d
(ea) = =d (PE
(log, z
log, a
|
silasfARE
~ dz \lna
 Ch: 4 47

The derivatives of the three inverse trigonometric functions are

given by the following table:

ne
The proofs of all these follow a similar pattern. We shall illustrate
the method by considering sin™! z.
Let y = sin7z, so that z = siny. Hence
d
= = cosy= V/1—sin?y = V1-—2z?.
y

pe ed
However, using the Chain Rule, it can be shown that —

d:
so that aL .
Gree n/a

Example 8. Differentiate each of the following functions:

(1) sinffan= ¥r); (2) =_cot*(4e* + 7);
log, a Pt = |
(3) Rane (4) 4* sin z.

Solution. (1) Using the Chain Rule to differentiate, we obtain

qi.
ae (sin(tan~’ -1 z)) = (cos(tan~ -1 z)) x mn
d (tan~’ -1 z)

(cos (tan? z)) x (3)

cos (tan™! z)
P42?"

(2) If we use the Chain Rule twice over, we obtain

£ (cot?(42? + 7)) = 2 x cot (4x? + 7) x < (cot (42? + 7))

 48 Ch: 4

d
= 2 x cot (4x? + 7) x (—cosec?(42? + 7)) x ae (4x? + 7)
= 2 x cot (4x2? + 7) x (—cosec?(42? + 7)) x 8
= —16z cot (4x? + 7) cosec?(4x” + 7).

(3) Using the Quotient Rule we have

aN loese tan z (=) — log; x sec? x

dz (2= )a [tan zc]?

tan ©
— log, x sec? x
= alas
tan? r
_ sinzcosz—zlnz
~ (sin? z)(z1n3) ’
on multiplying top and bottom by cos? z and using one of the laws of
logarithms.

(4) Using the Product Rule, we have

d i
— (4* sin z) = 47 + sin! 2 47 In4.
dz V1—<2?

EXAMPLES 4
1. Differentiate with respect to zr
(a) 72° —424 + 823 — 52? + 22 —7,
(b) 425 —523 + 7x? — 82 + 11,
(c) 62? — 32? +52-8+ = aaé
L
(d) ES lleSe ey ree
pS Gi
2. Use the Product Rule to differentiate with respect to z
(a) (2? — 5x? + 22 —4) (x4 — 22? +7),
(b) (34 + 7x3 + 92 +7) (52? + 32 + 4),
(c) («3 +42)sinz,
(d) (22° — 92°) Inz.
3. Use the Quotient Rule to differentiate with respect to
(a) z3>4+52?-%r+4 (b) z*? —7z +4
z2 49 ; xz? —3r+4’
(c) sin £ Inc
2403” (ioe
 Ch: 4 49

4. Use the Chain Rule to differentiate with respect to z

(a) (x4 — 22° — 52? + 42 — 1?cate
(b) (32° — 423 — 7? + 32 + His
(c) > *tan*z,
(d) sec* z,
(e) cos(52? — 42 +7),
(f) sin (tan (x? — 42 + 2)).
5. Differentiate each of the following with respect to z, simplifying
your answer where possible:
(a) (x-5)°(r+3)°; = (b) (32 +4)" (7x — 2)’;
(c)
c si 3z cos® : z;
sin d) sin’ r cos 4z;
sin“
(e) — sin® zcos? 22; (f) sec 2z cosec? z;
(g) tan? Vz; (h) cos (28);
(i) cosec (<#); (j) (1 —- tan? z)?;
xZcosz — sinz sin’ x
(k) zsinz +cosz’ (1) cos 3z’
sin 5z cos? 3x
ey cos? x’ (2) sin? 2r’
(0) sin7’ 32; (p)— Cos 5z:
(q) sin’ 1(3z + 2); (c) tan. 42:
(s) (sin7? x)’; (@); — tan=* (=:*).

6. Differentiate each of the following with respect to z:

(a) In(3z2—5); (b) In(2+ 42%);
(c) Inv6+2?; (d) In (32? —7) 3.
(e) In(cosz); (f) ae? In 2a:
(g) tan(Inz); (h) In(sin4z — cos 4z);
(i) (2? — 52?) In (2? + 3); (j) (xz? +72) In (22° — 52);

) m(5*%),
zr? +4
() ogre (82? +4);
(m) log, (cos 42); (n) logy) (tan 2z);
(0) (x — 2) log, (z + 4); (p) e"*;
CO) Gos es
(s) esin ¥ (t) eee.

(u) z*et*, (v) zte® ~2245.

 50 Ch: 4

(w) 24; (3) e538Fi

(Vip Seki Sym (em Zale.
7. By first expressing the functions in the form e*!"°, find the
derivative with respect to z of each of the following:
(a) tan t. (b) sin =

(c) (sinz)’; (d) (a? +2245)".

 51

CHAPTER 5
Examples involving Differentiation

Example 1. Let P be the point (1,4) on the curve defined by

the equation y = r3 — 42? +2+6. Find the equation of the tangent and
of the normal to the curve at P, and find the coordinates of the point
where the tangent meets the curve again.
; : d
Solution. Since y = 2? — 427 +z + 6, then = = 32? — 82 +1.
The gradient of the tangent to the curve at the point with z = 1 is
: ' : : d :
found by putting z = 1 into this expression for °Y' Hence the gradient
My
of the tangent to the curve at r = 1 is 3—8+1 = —4. It follows that
the equation of the tangent passing through the point (1,4) is

y—-4= —4(x —1),

from which
y—4=-47 +4,
or
y=8-—4r.

The normal to the curve is the line perpendicular to the tangent.

Since it can be shown that two lines are perpendicular precisely when
the product of their gradients is —1, and since in this case the tangent
has gradient —4, the gradient of the normal will be (—1)/(—4) = 1/4.
Hence the equation of the normal will be

y—4=(1/4)(x — 1),
which gives
4y-—16=2-1,

or
4y=2+165.
The tangent and the curve will meet at those points whose coordi-
nates (z,y) simultaneously satisfy the two equations
y=2° —477 +246

and
y=8—-4z.
Substituting for y, we need
z® —427+2+4+6=8—4z,
which can be expressed as
zr? —477 +54 -—2=0.
 52 Ch: 5

In order to solve this cubic equation, we can simply try to fac-

torise it, using the Remainder Theorem to give us one of the roots.
However, we can make this much easier for ourselves if we realise that
we already know two solutions of this
equation. We know that the tangent
meets the curve at the point z = 1,
and indeed we know that z = 1 is
a point at which the tangent meets
the curve in two coincident points.
Hence we would expect that z = 1
will be a repeated root of the equa-
tion z? — 427 + 52 —-2 = 0. Using
synthetic division, or long division of
polynomials, to find the other factor,
we can easily see that

a? — 47? + 52 —2 = (2 — 1)?(z — 2).

It follows that the other point of intersection of the tangent and

the curve is given by xz— 2 = 0, or z = 2. When z = 2, y = 0 (which is
obtained most simply by substituting
xz = 2 in the equation of the tangent
Y
y= 8—4r). tangent parallel
to 2-axis
A critical point, also known as
a turning point, of a function f isa
value of z for which the derivative f'
is zero. Hence a critical point of f is
a value of z giving a point where the
gradient of the tangent to the curve critical point
is zero, meaning that the tangent is
parallel to the z-axis.

Example 2. Find the critical points for the real function

sin 42
f(z)= cos! r

List explicitly those critical points that lie in the range (0,27).

Solution. We start the solution by calculating the derivative of

 Ch: 5 53

f(x), using the Quotient and Chain Rules in the process. We have

(cos* z) is (sin 4x) — (sin 4z) - (cos* zx)

fi(x) = = [cos* x]?
(cos* x) (4cos 4x) — (sin 4x) (—4 cos? z sin z)
cos§ x
4 cos* z[cos x cos 4z + sin 4z sin z]
cos’ x
4cos (4zr — z)
cos® x
4cos 3x
cos’ x ’
on cancelling cos? z from the numerator and denominator of the above
expression. Note that we are allowed to do this, because at the z-values
that would make cos? z equal to zero, the original function f(z) will be
undefined. Note also the simplification of the above expression by using
the formula cos (41 — z) = cos4x cosz + sin 4z sing.
. 4cos3 ;
Now f'(z) = 0 precisely when — sey jl tts happens exactly
os® x
when 4cos3r = 0. Note that we do not want cos*z = 0, since if
cos’ z = 0 then f'(z) is undefined and does not equal zero. Now
4cos3z = 0
when
cos 3z = 0,

which happens when

Sx = kn + 5 (k € Z)
and therefore when

r=—+— (k € Z).

The critical points in the interval [0,27] are found by substituting

different values of k into this formula to determine those solutions in
: : at Sut lus ate lia llr
the required range. These solutions are f=, 7, 4, |, oe and =.

One of the main uses of differentiation in applications is as a way of

finding the rate of change of one variable (say y) with respect to another
(say zr), where y is expressed as a function of z.
To see why this is so, let us consider the problem of trying to find
the average velocity of a vehicle. Suppose that the displacement s at
time t of the vehicle from its initial position is given by the function
s = f(t). Then at time t = to, the displacement of the vehicle will
be f(to). At time t = to +h, the displacement of the vehicle will be
 54 Ch: 5

f(to +h). Hence in the time period of h, the vehicle will have moved
f(to +h) —f(to). So the average rate of change of the displacement over
the time period to to tp +h will be

f(to +h)- — f(to) :

Allowing the length of the time period to tend to zero, we shall

obtain the (instantaneous) rate of change of the displacement at the
time t = to, given by
r f(to + h) — f(to)
im Sa ee ee OES
h—0 h

which is i at t = to, also denoted by f'(to).

Velocity is the name normally given to the rate of change of dis-
placement with respect to t. Equally, acceleration is the name for the
rate of change of velocity with respect to time.
Example 3. An object is moving through a resisting medium.
Suppose the displacement s at time t of the object from its initial posi-
tion is given by
s = 5t + 16t? —6e7*/8.
Find the velocity and the acceleration of the object at time t = 10.
Solution. The velocity v at time t is given by
ds
ete ;
Fe o-+puzes ze —t/3o

Substituting t = 10 in this, we find that v = 5+320+ 2e—19/3 ~ 325.07

(correct to 5 significant figures). The acceleration a is given by the
derivative of v with respect to t, and so

ee ss sertl3,
Hence the acceleration at time t = 10 is given by a = 32 — 2e-'s ~
31.976 (correct to 5 significant figures).

Example 4. A spherical balloon is being inflated. Find “ the

rate of change of the surface area A of the balloon with respect to the
radius r. In particular, find the value of this rate of change when the
balloon has radius 15 cm.
Solution. The surface area A of a sphere of radius r is given
by A = 4nr*. Hence, on differentiating with respect to r, we have
dA
im 8mr. When r = 15, we have aS 8x x 15 = 1207. Hence the
rate of change of A with respect to r is 1207 cm.
 Ch: 5 55

EXAMPLES 5
1. Find the point(s) (if any) on the curve y = 24 + 227 +102 +5
at which the tangent has gradient 2.

2. Find the values of z which give the point(s) (if any) on the
curve y = 2cos3z at which the tangent has gradient 3.
3. For each of the following curves, find the equations of the tan-
gent and normal to the curve at the given point. Find the point(s) (if
any) where the tangent at the given point cuts the curve again:
(a) y=a?—47?+44r+41 at (1,2);
(b) y=2*—3r*—-227?4+2+4+10 at (2,—4).
4. For each of the following curves, find the gradient of the tangent
to the curve at the point given. Hence find the coordinates of any further
point(s) on the curve such that the tangent to the curve at the point
found is parallel to the tangent to the curve at the given point:
(a) y=a2*—32?4+4r-—2 at (2,2)
(b) y=3r*t — 8x? — 302? +402 —2 at (1,3)
5. Let f(x) denote each of the following functions, in turn. Find
the derivative f'(z), simplifying as far as possible, and hence find the
critical points of f.
(a) sin 2 + cos2z, (b) cos* xcos4z,
3 ©
sin” 1
es, d Hoge,
(c) cos 3 (d) sin? 22 cos 4x
Then find those critical points of f that lie in the interval [0,27].
6. Find the velocity and acceleration at time t = 2 of a particle
whose displacement s from the origin varies with t according to the
equation
s = t° — 23 + 32? — 208.

7. Find the velocity and acceleration at time t = 3 of a particle

whose displacement s from the origin varies with t according to the
equation
s =t? —3t7 +t —2Qsint.

8. A manufacturer believes that the cost C of production of an

article is related to the number n produced by the formula

C = 2n? —5n? +4n—-6.

Find the marginal cost a when 50 units are being produced.

 56 Ch: 5

9. The revenue R when n units of a certain commodity are sold

is given by
R= 244 4n—2ni.
Find the marginal revenue, =, when 100 units are being sold.

10. The number N of bacteria in a colony at time ¢t days is given

by N = 50t + 2e*. At what rate is the colony growing after 2 days?
11. Find a formula for the area A of a square in terms of its diagonal
z. Hence find the value of = the rate of change of the area with respect
to z, when z = 5.
12. The formula V = ‘nr? expresses the volume of a sphere in
terms of its radius r. Hence find the value of tl, the rate of change of
r
the volume with respect to r, when r = 3.
 57

CHAPTER 6
Higher Derivatives
If y is a function of z (for example, if y = sinz) then we can
calculate the derivative of y with respect to rz, denoted by at (If
? d
y = sinz, then = = cos 2, of course.)
<
Now = is again a function of z, and so there is nothing to stop
P
us differentiating it with respect to z. If we do this, we will obtain the
second derivative of y with respect to z, which is denoted by —
e e ° ° . d?

x
2 d
(In our example, ——> = — (cosz) = —sinz.)
We could repeat this process, obtaining the third derivative of
’ ee d® ere :
y with respect to z, which is denoted by 7a (Continuing with our
£
By d
example, we have —> = —-(—sinz) = —cosz.)
We could continue like this, obtaining successively the fourth, fifth,
' ie: Mise da?
sixth, ... derivatives of y with respect to z. The derivatives —, bey
4 ae wae
y
Gyan are known as the higher derivatives of y with respect to z.
L
Please note that it is very important to be careful about notation
d? dy\*
here. It is quite easy to confuse, for example, =e an ($4) . However
z L
2
these are two totally separate functions of z. a is obtained by taking
£
d Lah
“Y and differentiating it, while
Ge ad iek:
“Y \ is obtained by taking <Y and
Bair
dr dz dr
squaring it.
There are various other notations which are quite often used to
denote derivatives. For example, if y = f(z), then we can denote the
derivative of y with respect to x by y’, by f'(z), or by Fs (f(z)). We

can denote the second derivative by y”, f"(x), or by qa? (f(x)). Similar
a
notations for the third derivative are y'”, f'"(r), and a3 (f(y).

Example 1. If y = 424 — 7x? + 82? + 52 — 11, find the first,

second, third, fourth and fifth derivatives of y with respect to z.
Solution. Since y = 424 — 7x? + 82? + 52 — 11, we have

oY ies 162° — 21z* + 16x +5,

dx
 58 Ch: 6

oe = 487? — 422 + 16,

zr

3
ot = 96x — 42,
zr

dty
i
and -
digas
dx®
Notice that the fifth derivative, and all the subsequent derivatives,
will be zero. This is typical behaviour for a polynomial function,
one that just involves non-negative powers of the independent variable
x. If y = f(x), where f(z) is a polynomial function of x of degree n,
for n € IN, then the (n + 1)-th and all subsequent derivatives of y with
respect to x will be zero. Here IN denotes the set of positive integers.
A differential equation (in z, y) is an equation involving the
d
variables +, y and one or more of the derivatives oy a etc. The
study of differential equations, and in particular, fidhis methods for
solving differential equations, forms one of the most important branches
of applicable mathematics.
We shall study some of the methods available for solving the sim-
plest types of differential equations later in the book. For now though we
shall concentrate on a simple example involving a differential equation.
Example 2. Show that
y=z2’cosz —3rsinz+2lnz
satisfies the differential equation
2
d*y
x 2 So (2 2 + 2)2° 2 cost =x dy
=

Solution. Since

y= 2’ cost —3rsinz+alnz
we know that

—dy = 2zrcosz
— zrasinz
.
—3sinz 1
—3rcosr+Inz+2x —
dz x
=Inz+1-—2?si
—3sinz —zcosz
nz
and also
dy —Ss
ap ersine i
A cosz — 3cosz
;
— cosz+zsinz
dz x

= es xrsinz — 2“ 2 cosz — 4cosz.

 Ch: 6 59

Substituting these into the differential equation, we see that

dy
L.H.S. r 2 qe? tytle 2 + 2)x* 2 cos x
=zr—2*sinz—2z 4 cosr —4zr? cosx ae? COS ©

—3rsine+2lnz24
2% cosz + 22? coszr
=2r-—r°sinz —2x* cosx —3zrsinz+2lnz
= 2(1—2’sinz — xcosz — 3sinz + Inz)

T a es wn
as required.
Notice that, as with our verification of trigonometric identities ear-
lier, we must not assume that the given equation holds, but rather we
must take one side of the equation and show that it equals the other
side.
EXAMPLES 6
2 dy dy d*y a: y
1. Find qn’? de?? da and Int where

(a) [w= cos2z, (b) ® «y =e'sin3z,

(c)) hey =alina, (dye ys es.
2. Prove that y = rtanz satisfies the differential equation

dy ody — 2rysec? sr= —2tanz.

ee dz

3. Prove that y = r°e?7+3z cos 3z—sin 3z satisfies the differential

equation
ae d
aap _ fee + Ory = 6x* (32 + 2) e**.
 60

CHAPTER 7
Implicit Differentiation

An equation such as

y = 5tan*zlnz

defines y explicitly in terms of z. For every allowed value of z, we

have a formula that enables us to calculate the corresponding value of
y. We can, in theory at least, plot the graph of the curve, and we can
use differentiation to find the gradient of the tangent to the curve at
any point on the curve.
Contrast the above situation with the curve defined by the equation

32? + 5y” — 6r + 20y — 69 = 0.

This equation defines y implicitly in terms of z, in the sense that if

we substitute a value of z then we still have an equation to solve before
obtaining the corresponding value(s) of y.
This equation defines an ellipse,
which looks rather like a squashed
circle, and it is reasonable that we
should want to find the equation of a
tangent to this curve, just as we want
to find the tangent to the graph of —%
an explicitly defined function. How-
ever we would like to do this without
having to resort to solving the equa-
tion to find y explicitly in terms of z.
There are two reasons for this: one is
3a? + 5y? |-6a + 20y — 69 = 0
that such solving will always involve
unnecessary extra work; the other is
that sometimes it will be impossible to solve a certain equation explicitly,
and yet we would still like to obtain the information that differentiation
can give us.
The technique we use is known as implicit differentiation. More
properly, this should be known as the differentiation of implicitly defined
functions. In this procedure, we shall assume that an equation we are
given defines y as a differentiable function of z, and then we differentiate
each term with respect to z, remembering that, by the Chain Rule, to
differentiate a term involving y, we first differentiate it with respect to
y and then multiply by ze.
L
 Ching 61

Example 1. Find y' and y" at the point (3,2) on the ellipse
3x? + 5y? — 6r + 20y — 69 = 0. Hence find the equation of the tangent
and normal to the ellipse at (3,2).
Solution. On differentiating implicitly with respect to z the
equation 32? + 5y? — 6x + 20y — 69 = 0, we obtain
6r + 10yy’ —6 + 20y' —0=0.
Notice that in carrying out this differentiation, the derivative of 5y?
: ; d Aes oe
with respect to z is 5 x 2 x y x = which is 10yy’, and the derivative
a
of the constant —69 is 0.
The above relationship gives us the connection between zx, y and
y’ at every point on the curve. To find the gradient of the tangent at a
specific point, we need to substitute in the appropriate values for z and
y. So we substitute z = 3, y = 2 into the equation to find the value of
y', and we obtain
18 + 20y' —6 + 20y' = 0,
from which
40y' = —12,

or

y' = —3/10.
It follows that the equation of the tangent to the curve at the point
(3,2), having gradient —3/10, will be

Us Peiey ge RO)

which is
10y — 20 = —3z +9,

or

10y + 32 = 29.
The normal to the curve will be perpendicular to the tangent and
so the normal will have gradient Gay which is ». Hence the equation
10

of the normal will be “ie

Yo 2= a(t es 3) ’

which can be simplified to

3y = 102 — 24.

To find the value of y" at the point (3,2), the simplest way is to
differentiate the equation 6z + 10yy’ — 6 + 20y' = 0 implicitly with re-
spect to z. We are assuming that y is defined for us as a differentiable
 62 Ch: 7

function of z, and we shall assume similarly that y’ will also be a dif-

ferentiable function of z, whose derivative we can find by differentiating
implicitly using the Chain Rule.
Notice that a term such as yy’ involves a product of two functions
of z, and so must be differentiated using the Epied Rule. iden we
do this differentiation for yy’ we will obtain y x fi (y')+y' x Te (y),
ha d
which is y X <5 +y'x or yy" + y'y’.
On differentiating 6z + 10yy’ — 6 + 20y’ = 0 with respect to x we
therefore obtain 6 + 10(y'y’ + yy”) —0+ 20y” = 0. Substituting in
the values for z, y and y’, which are zt= 3, y = 2 and y’ = —3/10,
we have 6 + 10(—3/10)? + 10 x 2 x y+ 20y” = 0, which simplifies to
40y"” = —(69/10), or y” = —(69/400).

Example 2. Find dy in terms of z, y for the variable y defined

L
implicitly in terms of x by the relation

z* — 32°y + Tzy* = 32.

Solution. Differentiating the above equation implicitly with

respect to z, we obtain

3x? — 3 (32?y + c%y') +7 (y* +2 x 4y*y’) =0.

Notice here that the terms 3z7y + x3y' come from differentiating x?y
using the Product Rule, while the terms y* + z x 4y*y' come about on
differentiating ry*, again using the Product Rule. Notice that the term
32 also has to be differentiated, which leads to the 0 on the right hand
side of the equation. Forgetting to differentiate the constant is a very
common source of mistakes in this type of example.
If we re-arrange the terms in the above equation, we can solve it to
d
find = in terms of x and y. We obtain

32? — 92?y — 32°y' + Tyt + 282y3y' = 0,

from which we have

32? — 927y + Ty* = 323y! — 28cy3y’,

or
ee 32? — 9a?y + Ty#
323 — 282y3 ’
provided of course that 323 — 28ry? $ 0.
 Ch:7 63

EXAMPLES 7
1. If 2? + 3y? — 2ry dy in terms of z and y.
+ 52 — 4y = 9, find oi
z

2. If zsiny — ysinz = 0, find ovin terms of z and y.

3. Find y’ and y” at the point (1,—3) on the curve

x? +y? +2ry — 42 + 6y +18 =0.

4. Find y' and y” at the point (2,1) on the curve

x? +y* — 32y = 3.

5. Find the equations of the tangent and normal to the curve

4r° + 22*y? + y* = 16
at the point (—2, 2).
6. Find the equation of the tangent to the curve
sin (xz + 2y) + cos(2z — y) = 1
at the point (3, at),
7. Prove that the curves
z?+y?=10 and 2?+2y?—38r+19=0
intersect orthogonally. [Find the points of intersection of the curves
and show that the tangents to the two curves at each of the points of
intersection are perpendicular.|
 64

CHAPTER 8
Sketching Rational Functions without using Calculus
In this chapter we shall sketch some rational functions, working
simply from the form of the function, without using more advanced
methods such as Calculus. A rational function is one that can be
expressed as a ratio of two polynomials. So a rational function has the
form
_ n(x)
f(z) = d(z)’

where n(x) and d(z) are both polynomials in z, which we shall assume
to have no factors in common.
Example 1. Sketch y = 1/z.
Solution. To get an idea of the behaviour of this function, we
start by drawing up a table of values of y against z.
Se
ese =O

From the table we see that as rz becomes very large then y remains
positive but becomes very small. We write this as y — 0+ as rz — 0.
[The plus sign here indicates that y remains positive while approaching
the value zero.|
Again, if z is large (in modulus) but negative, then y again ap-
proaches the value zero, but this time rel on negative values. We
write this as y > 07 as x — —oo.
Now we consider what happens as x tends towards zero, taking
positive values. When this happens we see that y is positive and that y
increases without bound. We write this as y — oo as t > OT.
Finally we see that y — —oo as x — 0-. This indicates that
as xz tends to zero, taking negative
values, then y becomes very large (in
modulus) and negative.
Notice that it is not permissible
to divide by zero, and so there is no
point on the curve corresponding to
xz = 0. Hence the graph of the func-
tion is as shown.
We call the y-axis the vertical
asymptote to the curve, since it is
a vertical line to which the curve ap-
proaches arbitrarily close as z tends
 Ch: 8 65

towards the value zero. For a similar reason, we can say that the z-axis is
a horizontal asymptote to the curve. We can also describe this latter
situation by saying that the curve y = 1/z behaves asymptotically
like y = 0 (written y = 0 as r — +00).

Example 2. Sketch y = sy —4.

Solution. This curve will have the same general shape as the
curve y = 1/z. As we saw in Chapter 1, the effect of the (x — 2)-
term will be to move the vertical asymptote from the z-axis to the line
xz = 2, while the effect of the (—4)-term will be to move the horizontal
asymptote to y = —4. Hence y —4 :
as © —> -too.
We can determine some points
on the curve by finding where it cuts
the axes. Putting rt = 0, we have
Y = sy —4 = —9/2. So the curve
cuts the y-axis at (0, —9/2).
The curve meets the z-axis when
y = 0. This happens when

so that

or
r—-2=1/4,
from which zt = 9/4. Hence the curve meets the z-axis at the point
(9/4, 0).
1
Example 3. Sketch y = se
Solution. We consider the ta-
ble of values given on the next page.
1 la\e
Notice that y= > = (=) , so that
x x
every y-value of y = — will be the
square of the corresponding y-value
1 4
for y = —. In particular, all y-values
will be positive, so that the curve will
lie wholly above the z-axis. Values
for y = — with y > 1 will be increased on the curve y = while those
2?’
for Bien 0 < y < 1 will be decreased on this curve.
 66 Ch: 8

35: ‘Stet : 1
The y-axis is again a vertical asymptote for y = —, and equally
the z-axis will be a horizontal asymptote to the curve. Notice that
this curve is symmetric about the y-axis, whereas the curve y = a wes
symmetric about the origin.
[= [0 [som [402 [205 [#1 [22 |#8 | 5100|
r= 3 [few |as [+[> [oesoo [oom
We shall now develop a systematic way for sketching the rational
function

where n(x) and d(x) are two polynomial functions in the variable z,
having no common factor. We follow the procedure outlined below.
1) Factorise the numerator n(x) and find where n(x) =0. This
will give the points where the curve meets or crosses the z-axis.
2) Factorise the denominator d(zr) and find where d(z) = 0. This
will give the vertical asymptotes to the curve.
3) Ignore all terms in the numerator and denominator except
those involving the highest powers of x. By finding the ratio of these
terms, determine the asymptotic behaviour of the curve y = f(z) as
£.—> 4-00,
4) By putting z = 0, find where the curve cuts the y-axis.
5) Draw up a table of sign for f(z), including in ascending order
all the z-values found in 1) and 2) above. This table can be used to
decide where the curve lies above the z-axis and where below. From this
we can decide how the curve approaches its zeros and how it approaches
its vertical asymptotes.
6) Sketch the curve, incorporating all of the information found
above. Any inconsistency will indicate that somewhere in the working
lies an error.
We shall now carry through an example of this type, in order to
demonstrate the method in action.
27? — 82 +6
Example 4. Sketch y =
z?—-4
Solution. 1) We can factorise 2r? — 82 +6 as 2(r — 1)(x — 3).
Since this is zero when rz = 1 and z = 3, we see that the curve meets or
crosses the z-axis where x = 1 and where z = 3.
2) «x? —4 is a difference of two squares, and can be readily fac-
torised in the form (zr — 2)(x + 2). Hence rz? — 4 = 0 when z = 2 and
when z = —2. It follows that there are vertical asymptotes to the curve
at z = 2 and at x = —2.
 Ch: 8 67

3) For z large (in modulus) we know that 22? — 82 + 6 = 2z? and

that 2? —4 x 2?. Hence it follows that, as r + too,

So the curve behaves asymptotically like the straight line y = 2. We

write this as y © 2.
6 3
4) When z = 0, we see that y = Srigris Hence the curve cuts
the y-axis at the point (0, —3/2).
5) We now draw up a table of sign for y, including all the z-values
we found in 1) and 2) above. This means that we need to include, in
order, the t-values ae
Pane Te Se

SE~ IRs een? ap es | Ye |FO ee

eng {=} fete dee
eee ie 8) | eg eed 0 |=)
eta Le
ee
Ba erage tele
ieee OE Te po
Ee ete
gg
Above Below a — ae
r-axis
In the final row of the above table, we have used A and B to stand
for Above and Below respectively, denoting whether the curve lies above
or below the z-axis, On to denote when the curve has a point on the
z-axis, and VA where there is a vertical asymptote at the point under
consideration.
We know that there is a vertical asymptote to the curve at r =
—2. The table of sign tells us how the curve approaches its vertical
asymptote.
To the left of the vertical asymptote at r = —2, we know that the
y-values on the curve are either getting very large or are becoming very
large (in modulus) and negative. From the table of sign we see that to
the left of x = —2, y is positive. Hence the curve is approaching the top
end of the asymptote as x — —2, with x < —2.
Similarly, to the right of z = —2, the curve is again approaching
the vertical asymptote, but this time we see from the table that y is
negative, and so this time the curve approaches the asymptote at the
bottom end.
Now let us look at how the curve behaves near its zero at x = 1.
Consulting the table of sign, we see that the value of y is negative
for z close to 1, but with z < 1. Hence the curve is below the z-axis
for values of x close to x = 1, but to the left of s =1. For z slightly to
 68 Ch: 8

the right of r = 1, the table tells us that y is positive. Hence, as shown

in the diagram, the curve rises from below the z-axis to above it as z
increases from being a little bit smaller than z = 1 to being a bit larger
than = I.
The behaviour of the curve near the vertical asymptote at z = 2,
and near the zero at x = 3, is established in a similar way.
It should be clearly realised that we are using the table of sign for
y to tell us about the behaviour of the curve. There are three different
tables of sign that can give us useful information. These are the tables
for y, y’ and y"’. Each tells us different things about the curve, and
so must be given the correct interpretation. This is why I suggest you
include the final line shown in the table of sign for y, to help you to
remember how to interpret what the table is telling you.

We have completed the diagram in the simplest way possible con-

sistent with the information found above. We can show that this is
correct in this case, but the proof of that will have to wait until we have
Calculus fully at our disposal.

Example 5. Sketch y = wit oe

7 j Y © ont = geeeG
Solution. We are asked here to sketch a function which is the
reciprocal of the function we were asked to sketch in Example 4. A few
moments thought will convince us that there is the following relationship
between the two curves.
 Ch: 8 69

1) Zeros on the original curve will correspond to vertical asymp-

totes to the reciprocal curve.
2) Vertical asymptotes to the original curve will correspond to
zeros on the reciprocal curve.
3) Positive values on the original curve will correspond to positive
values on the reciprocal curve.
4) Negative values on the original curve will correspond to nega-
tive values on the reciprocal curve.
2x? — 82 + 6
5) Since y = 52 behaves asymptotically like y = 2, the
reciprocal curve will behave asymptotically like y = 1/2.
6) Ifthere had been any, then non-zero local maxima on the orig-
inal curve would correspond to non-zero local minima, while non-zero
local minima on the original curve would correspond to non-zero local
maxima on the reciprocal curve.
It follows that a sketch of the curve we require is as given.

Notice that in this sketch we have drawn the curve so that as x

increases past the value z = 2 then the curve goes from being above
the z-axis to being below it. We have done this because from the above
we know that there is a zero at zx = 2 corresponding to the vertical
asymptote to the original curve at x = 2, and we know that the original
curve lay above the z-axis for z < 2 (and so the reciprocal curve does
also) while the original curve was below the z-axis for x > 2 and so
 70 Ch: 8

the reciprocal curve is also below the z-axis just to the right of the zero
at z = 2. A similar argument deals with the behaviour near the two
vertical asymptotes and near the other zero on the reciprocal curve.
EXAMPLES 8
1. Sketch the curves
(a) y==,4 (b) y=24—5,1
3
(c) Paseo. (d) adageoiys

2. Without using Calculus, sketch the curves

i pach _ =2)(@
(2 —4) (e+ 2)
Oli ola
gyana)” blued ru\ caign cee
ayenmayT
_ (e-—3)(z-1) A xz? (x — 2)
oO) Urea” wer eee
3. Without using Calculus, sketch the curve
7 z*—427+3
Ys 7 EDmer )

and deduce a sketch of the curve

es r+2
ome Pry

4. Without using Calculus, sketch the curve

ca |
~ 22-28"
and deduce a sketch of the curve

Pesca xz? —24

|as —8
ee
xr-—1l
 71

CHAPTER 9
Concavity and Points of Inflection

We have seen that f'(c) gives the rate of change at x = c of f(z)

for increasing x. Put another way, f'(c) gives the slope, at z = c, on the
curve y = f(x). So f'(c) tells us whether the curve y = f(z) is going
upwards or downwards as we look at it from left to right, and also f'(c)
tells us how rapidly f(x) is increasing or decreasing.
Typical sketch
Large y increases
positive rapidly as
x increases

Small y increases
positive slowly as
Zz increases

Small (in y decreases

modulus) and slowly as
negative x increases

Large (in y decreases

modulus) and rapidly as
negative x increases

(ce, f(c)) ;

Just as the derivative of y tells us about the slope of y = f(x), so

dy y
the derivative of oe will tell us about how the gradient aS changes as
 (2 Ch: 9
thar ad
xz changes. For if f"(c) is positive, then we know that the gradient oe
is increasing as z increases through the value zr = c. ea the other hand,
if f""(c) is negative, then we know that the gradient = is decreasing as
xz increases through the value z = c.
The following table indicates the behaviour of y = f(x) for different
combinations of the signs of f'(c) and f"'(c).

y increases as ave UP
x increases,
with gradient
increasing as
x increases

y increases as
x increases
but with gradient
decreasing as
x increases

y decreases as
zr increases,
with steepness
of gradient
becoming less
as x increases

y decreases as
zr increases, but
with steepness
of gradient
becoming greater
as © increases
curve

We say that y = f(x) is concave up at z = c if the curve y = f(z)

lies above the tangent to the curve at zx= c. This means that, as we
move from left to right along the curve, the curve is bending to the left.
Similarly, we say that y = f(z) is concave down at z = c if the
curve y = f(x) lies below the tangent to the curve at z = c. This means
that, as we move from left to right along the curve, the curve is bending
to the right.
From the table given above, we can see that, at r = c, y = f(z) is
concave up tf f"(c) > 0, and that y = f(z) is concave down if f"(c) < 0.
 Ch: 9 73

A point of inflection on y = f(z) is a point at which the concavity —

changes (whether from up to down, or from down to up). It follows that
if z = c gives a point of inflection on the curve y = f(z), and if f'(z) is
differentiable at z = c, then f"(c) = 0. This is because, to one side of
the point of inflection, we must have f'’(z) > 0, with f(z) < 0 on the
other side of the point of inflection. Hence, at least if we assume that
f"(z) is continuous near z = c, then we must have f"(z) = 0 at the
point of inflection.

concave UP concave UP
(f(a) > 0) (f'"(«) > 0)

point. of inflection point of inflection

on a rising curve on a falling curve
CHE = 0)
(f'"(c) = 0)

concave DOWN concave DO\WN

(f(a) < 0) (Fey 0)

Let us imagine a motor-cyclist who is riding through an S-bend,

and who is leaning first to the one side and then to the other as he takes
the bends. Then a point of inflection corresponds to the point between
the two bends when the motor-cyclist is momentarily upright.

Notice that it is not always true that if z = cis such that f"(c) =0
then xz = c must give a point of inflection. For, if we consider the three
curves given by y = 24, y = 2° and y = —z*, then on each of these
curves we can readily see that the second derivative is zero at x = 0.
 74 Ch: 9

However, if we examine the diagrams given on the previous page, we see

that in the case of y = 24 we have a minimum at z = 0, on y = —2* we
have a maximum at z = 0, and it is only on y = 2° that we actually do
have a point of inflection.
To ensure that we do have a point of inflection at a point z = c on
y = f(z) for which f"(c) = 0, we have to check that the sign of f(x)
does indeed change (whether from positive to negative, or from negative
to positive) as z increases through the value z = c.

Example 1. Sketch the curve

y=a*—62r? +3242

in a neighbourhood of the points (0,2), (2,0) and in a neighbourhood

of its points of inflection.
dy d’y
Solution. We start the solution by finding os and qa
— Since

y = 21 — 62? + 3r 4 2,
we have

ce 473 — 127 +3
dz

and
d*y 2

We need to find the values

of y, uiand at'r = Ovand at 7 = 2:

Substituting z = 0 first, we have y = 2, avy aS ey

From these we see that when the curve is in a oteatied of the point
(0,2), then the curve has a steep positive gradient there (it is sloping at
roughly 71° to the z-axis) and that the curve is concave down, so that
it lies below the tangent to the curve at (0, 2).
Now we turn our attention to the point z = 2. Substituting z = 2,
we have y = 0, dy Stil anda 36. In this case, the curve has an
dz dx?
even steeper positive gradient (this time the curve is at roughly 85° to
the x-axis), and the curve is concave up. This time therefore the curve
lies above its tangent at r = 2.
Possible points of inflection on the curve occur when 12x? — 12 = 0.
This requires that 12(z — 1)(a4 + 1) = 0, or that z = 1 or zg = -1.
We need to check that the concavity does indeed change at both
xz = 1 and at. x = —1. We do this by drawing up a table of sign for y”
near each of these z-values.
 Ch: 9 75

Sab 20ere x0 =*s0 Stale aineittn efaiog afr bla]

peek mabe hon ob 1 He
xz-—1

(Fame-perp_+]o,
-
a
Since the concavity does change at both zr= 1 and zx = —1, we see
that these are both points of inflection on the curve. Substituting these
: d
z-values in turn we find that, when z = 1 we have y = 0 and ieee: —5,
dx
while when z = —1 we see that y = —6 anddy
— A
Near the point of inflection at (—1, —6), oe see that the curve has
a steep positive gradient, and that, as x increases from less than —1 to
more than —1, the curve changes from being above its tangent to being
below its tangent.
Similarly, near the point of inflection at (1,0), the curve has a fairly
steep negative gradient, and this time
the curve changes from lying below
its tangent to lying above it as z in-
creases through z = 1.
The sketch given is not yet com-
plete because we do not have enough
information to be sure how we should
join up the sections of the curve. To
complete the curve, we need to use
Calculus to find where the turning
points are. We also need to see what
happens to the curve when zr — +oo.

EXAMPLES 9
1. Find the set of real numbers for which the function
f(z) =x? — 122? + 252 + 30
is concave down.
2. Express as a union of intervals the set of real numbers for which
the function
f(x) = x* — 22° — 122? + 132 + 20
is concave up.
3. Find the points on the curve y = z+ + 42° — 182? — 40z + 50
7
for which ath = 0, and draw up a table of sign for « Y near each of
these points. Hence determine whether they are points of inflection on
ze . . .

the curve.
 76 Ch: 9

4. Find the points on the curve y = 32°—5r4—102°

+302? +32—23
for which <= = 0, and draw up a table of sign for = near each of
these points. Hence determine whether they are points of inflection on
zr ° e

the curve.

5. For the curve y = 2° — 9x? — 212 + 47, find the gradient of the
tangent to the curve at each of the points (—1,58) and (3,—70), find
the concavity of the curve near each of these points, and hence sketch
the curve in a neighbourhood of each of these points.
ae :
6. Find any points for which oa = 0 on the curve given by the
equation y = x4 — 8224242? — 10x —2. Find the gradient of the tangent
to the curve at each of the points (0, —2) and (1,5) and at any points on
the curve for which 2 = 0. Also find the concavity of the curve near
z
each of these points, and hence sketch the curve in a neighbourhood of
each of these points.
7. Find any points of inflection on the curve defined by the equa-
tion y = 4 — 22° — 3627 +80r+100. Find the gradient of the tangent to
the curve at each of the points (—1, —13) and (4, —28) and at any points
of inflection on the curve. Also find the concavity of the curve near each
of these points, and hence sketch the curve in a neighbourhood of each
of these points.
 17

CHAPTER 10
Local Maxima and Local Minima
In this chapter we shall see the interpretation of the derivative of a
function f(z) in terms of the slope of
the graph of the function. We shall “
assume that f(x) is a function which
is differentiable (and therefore con- a strictly
tinuous) on an interval J. creasing
function
We say that a function f(z) is
strictly increasing on an interval J
if f(z2) > f(z1) for any pair of num-
bers z2, 2, in the interval J for which
Z2 > 23. Clearly if the gradient is
positive on a graph then the func-
tion is increasing, and so, if f'(z) > 0
for all points on an open interval J,
then f(z) is (strictly) increasing on
the open interval I.
Similarly, a function f(z) is said
to be strictly decreasing on an in- a strictly
terval I if f(z2) < f(x1) for any pair decreasing
function
of numbers x2,2; in the interval I
for which zz > 2. Clearly if the
gradient is negative on a graph then
the function is decreasing, and so, if
f'(z) < 0 for all points on an open
interval J, then f(z) is (strictly) de-
creasing on the interval I.
A value of x for which f'(c) =0
is called a critical point, or turning point. As we have already seen,
these are the values of x for which the tangent to the curve y = f(z) is
parallel to the z-axis.
Providing the gradient f'(z) is continuously changing round about
the critical point x = c, there are four possibilities that can occur. These
are tabulated below.
From the table given below, we can see that there are two possible
ways of determining the nature of a point on a curve for which f'(c) = 0.
Method 1. Use a table of sign for f’(x) to find the signs of f'(zr)
on both sides of the value z = c. We can classify the point as being
a local maximum, local minimum or point of inflection (on a rising or
falling curve) according as to which of the four possible arrangements
of signs occurs.
 78 Ch: 10

Method 2. Find the value of the second derivative f"(r) at z =.

If f(c) > 0 the curve is concave up, and so has a local minimum at
z =c. On the other hand, if f"(c) < 0 the curve is concave down, and
so has a local maximum at z = c. If f"(c) = 0, it is best to use Method
1 to decide on the nature of the turning point.

f'(z)
Shape
of curve

Name Local Minimum

Concavity Up
f(z) + +/0

Shape
of curve
Name Horizontal Point Horizontal Point
of Inflection on of Inflection on
a falling curve a rising curve
Concavity Down | Down
f"(z)
Method 1 has the advantage that we do not need to find f(z) in
order to apply the method, but it has the draw-back that we have to use
a table of sign to determine the nature of the turning point. Method 2
has the advantage that we only need to evaluate f'’(r) at the one point
x = c, but it has the disadvantage that if f'"(c) = 0 then Method 2 gives
us no information about the nature of the point zt = c.
Usually it is preferable to use Method 2, provided finding f'(z) is
not too difficult, retaining Method 1 as a fall-back method.

Example 1. Sketch the curve

y = f(z)
= 2° — 52? — 82 + 12,

finding where it crosses the coordinate axes, the coordinates and natures
of the critical points, and the coordinates of the point of inflection. Find
the intervals on which f(z) is strictly increasing, and find the values of
z for which y = f(z) is concave up.
Solution. If we substitute values of z in turn, we readily find
that f(1) = 0, from which it follows that z—1 is a factor of f(z). Using
synthetic division, we see that

f(z) = (x —1)(z? — 42 — 12).

 Ch: 10 79

Continuing to substitute values of x in g(r) = r* — 4x — 12, we see

that g(1) = —15, g(—1) = —7, while g(2) = —16 and g(—2) = 0. From
this it follows that z+ 2 is a factor of
g(x), and so we conclude that g(x) =
1-5 -8 12
(x + 2)(z —6). Hence
xl] 0 1 —4 -12
f(x) = (x —1)(z + 2)(x — 6). +|]|1-4 -12 0
So the graph of y = f(z) cuts
the z-axis at r = 1, r = —2 and also
at zr= 6. The graph intersects the merle)
y-axis at r= 0, y = 12. Now (aD) IeOD aye12
f'(z) = 32? — 102 —8 4Fl ii haem
= (z — 4)(3z + 2),
and so f'(z) = 0 when z = 4 and
when zt = —2/3. Also f(r) = 62 — 10, so that f"(z) = 0 when
£'= 15/3:
When z = 4, y = f(4) = 44-5 x 427-8 x 4412 = —36, and
f"(4) =6 x 4-10 = 14, which is positive. Thus the point (4, —36) is a
local minimum on the curve.
When zx = —2/3,
f(—2/3) = (—2/3)? — 5 x (—2/3)? — 8 x (—2/3) +12
_ —8— 60+ 144 + 324
> 27

Also f'"(—2/3) = 6 x (—2/3) — 10 = —14, which is negative. It follows

from this that the point (—2/3, 400/27) is a local maximum on the curve.
The only possible point of inflection on the curve occurs where
f"(z) = 0, which happens when z = 5/3. When z = 5/3, we find that
y = f(5/3)
= (5/3)° — 5 x (5/3)? —8 x (5/3) + 12
__ 125 — 375 — 360 + 324
tx OF

Diy
To decide whether this point is indeed a point of inflection on the
curve, we need to draw up a table of sign for f(z) in the vicinity of
the value z = 5/3. Since we are also required to discuss the concavity
of the whole curve, we will draw up the table of sign considering all
possible values of z at which the concavity could change. However,
since the second derivative of f(z) is only zero when xz = 5/3, the table
we require is very simple.
 80 Ch: 10

f(a) = 62 — 10
DO
Since the concavity of the curve does indeed change at x = 5/3,
it follows that the point (5/3, —286/27) is a point of inflection on the
curve.
To find where the curve is strictly increasing, we need to find the
intervals on which the gradient is positive, and so we draw up a table of
sign for f'(x), and consider those values of x for which f'(x) > 0. Since
the sign of f'(z) might change where f'(r) = 0, we need to include in
our z-values the values z = 4 and x = —2/3 at which f(z) = 0.
er es on eS ee
ee a Pe ie
z—4 — 0 Sm

f@MH@s4G2.+2) 1 sts] 0 ie OS eee

[InefDec
[Ine [0 | Deo|0-|Ine|
From this we see that the curve y = f(z) is strictly increasing on
the intervals (—oo, —2/3) and (4, 00). Notice that although the function
is strictly increasing on both of these intervals, it is not true that if
we choose a point in the first interval, then the value of the function
at that point will be less than the value at any point in the second
interval. As an example, we easily see
that f(—1) = 14 while f(5) = —28.
The reason for this is of course
that the curve decreases in value be-
tween the two intervals, and so starts
increasing at z = 4 from a smaller
value than the one it reached near
z= —2/3.
Finally notice that from the ta-
ble of sign for the second derivative,
we know that the curve is concave up
precisely when z € (5/3, 00).
We complete the solution of this
example with a sketch of the curve.
Example 2. Consider the rational function
2
iO) ise
(x + 12)(42 — 15)
Find the coordinates and nature of the critical point(s) on the curve
y = f(z), and also find any point(s) of inflection on the curve. Find also
the points where the curve meets the coordinate axes, the asymptotic
behaviour of the curve as r —+ too, and the vertical asymptotes on the
 Ch: 10 81

curve. Hence sketch the curve. Find also those values of x for which
y = f(x) is concave down.
Solution. In solving problems like this one, we follow the routine
we introduced for sketching rational functions, with the addition that we
can now use Calculus to determine the precise positions of the maxima,
minima and any points a inflection.
f(z) = DO Alar 2)98) meets the z-axis where z = 2, and at no
(x + 12)(42 — 15)
other points, since z = 2 is the only value of x which makes f(z) = 0.
The curve has vertical asymptotes at x = —12 and z = 15/4, and
for no other values of x, since for a vertical asymptote we are looking
for the denominator of f(z) to be zero, so that f(z) will be undefined.
2
As z — +00, y & = = 1. So the curve approaches the line y = 1
for positive or negative values of x which are large in modulus.
When zx = 0, y = 4(—2)?/(12)(—15) = —4/45, giving the point
where the curve cuts the y-axis.
We now draw up a table of sign for y against z, including the z-
values z = 2, x = —12 and z = 15/4 found above. This table will tell
us whether the curve lies above or below the z-axis.
a ea EE a
[oP Ce ia On deed oe oe ceed A
ee ea
A 12

Den= Bi+ 12)(42 — 15) pte

ee
aABOVE/ON/BELOW
Eh ay Des Ml HEP LUND
I- axis
iy ena hae
We have used A and B in the final line of the table to indicate
whether the curve lies Above or Below the z-axis, as well as writing
ON to indicate where the curve meets the z-axis and VA to denote a
Vertical Asymptote.
From the information gathered above, we could draw a reasonable
sketch of the curve. However we shall wait before doing that so that we
can use Calculus to determine the precise positions of the critical points
on the curve. # e
We use the Quotient Rule twice in order to calculate oe and a
ie
We have
 82 Ch: 10

dy (x +12)(4x — 15)8(a — 2) — 4(x — 2)?(8z + 33)

dz [(x + 12)(42 — 15))?
_ A(x — 2)[2(a + 12)(4a — 15) — (x — 2)(82 + 33)]
<i (x + 12)2(4z — 15)?
4(x — 2)[8x? + 662 — 360 — (8x? + 172 — 66]
i (x + 12)?(42 — 15)?
4(x — 2)[8x? + 662 — 360 — 8x? — 17z + 66]
i (x + 12)?(42 — 15)?
_ A(x — 2)(492 — 294)
~ (x + 12)2(42 — 15)?
_ 196(z — 2)(z — 6)
~ (x +12)?(42 — 15)?"
Notice that in using the Quotient Rule above, we needed to find the
derivative of (x + 12)(4z — 15). We chose to multiply out and express
this as 42? + 332 — 180, which has derivative 8z + 33, as we stated. An
alternative would have been to differentiate (x + 12)(4z — 15) using the
Product Rule, and then simplify the answer. Either method would be
acceptable. We chose the method we did because it was slightly easier
in this case.
Now we need to use the Quotient Rule again in order to find the
second derivative.
In anticipation of this, and to reduce the complexity of the cal-
culations involved, let us start by working out the derivatives of the
numerator and denominator, and simplifying these as far as possible.
Then

= (196(2 — 2)(x —6)) = 196 ((2 —2)x1+(a#-6) x 1)

= 196 (22 — 8).

Also

<((z + 12)?(42 — 15)*) = <a + 12)(4ax — 15))”

= 2(a + 12)(4r — 15) =(C + 12)(4r — 15))

= 2(a + 12)(4r — 15) (82 + 33),

quoting the result we found above for the derivative of (x +12)(4a —15).
Then
 Ch: 10 83

(x + 12)?(4x2 — 15)?196(2z — 8)
dy —196(z — 2)(x — 6)2(z + 12)(42 — 15) (82 + 33)
dx? [(z + 12)?(42 — 15)?]?
196(z + 12)(42 — 15)[(z + 12)(42 — 15)(2z — 8)
—(x — 2)(x — 6)(16z + 66)]
(a + 12)4(4a2 — 15)4
_ 196 [(z + 12)(8x? — 62x + 120) — (x — 2)(16x? — 302 — 396)|
a (x + 12)3(42 — 15)3
196[823 + 34x? — 6242 + 1440 — 162° + 622? + 3362 — 792]
(x + 12)3(42 — 15)
196[—82* + 962? — 288z + 648]
(x + 12)3(4a2 — 15)3
_ (—8)(196)[z* — 122? + 362 — 81]
vi (x + 12)3(4z2 — 15)3
_ —1568(2 — 9)(x? — 32 + 9)
~ (x + 12)3(4a — 15)3
It should be noted that we had to do a fair amount of quite difficult
algebraic manipulation in this particular example. This is not typical of
examples of this type, where usually things will be arranged to work out
much more nicely. However it must be pointed out how essential it is
that we can perform such algebraic manipulation accurately, if required.
One other point to make is that at the end of the calculation of
i we needed to be able to factorise x? — 127? + 36x — 81. This we did
Lz
using the Remainder Theorem, although as usual we could have used
Synthetic Division as an alternative. Having taken out z — 9 as a factor,
we were then left with +? — 3x + 9, which we find cannot be factorised
further into real factors since it has a negative discriminant.
We are now able to use the values of cu and = found above to
z
determine the coordinates of the turning points on the curve.
Recall that
dy _ 196(z — 2)(z — 6)
dx (x +12)?(4x — 15)?’
from which we see that turning points on the curve will occur when
196(x — 2)(z — 6) = 0, which is when z = 2 or « = 6. ;
dy _ —1568(—7)(7)
When z = 2, we have y = 0 and also , which
dz?—«<43(—-7)8
simplifies to the value —8/98. Since ey is negative, we have a local
maximum at the point (2,0).
 84 Ch: 10

Substituting + = 6, we have y = 4(4)?/18 x 9 = 32/81 and also

d’y — (—1568)(—3)(27)
which simplifies to the value 196/9*. Since
dz? 183(9)3
a is positive, we have a local minimum at the point (6, 32/81).
L
The only possible point of inflection on the curve will occur when
d?
a = 0, which from the factorised form of — can only happen when
z L
xz = 9. To decide whether this does indeed give us a point of inflection,
. y
we d draw up a table
table of of sign forfor —..
>
In anticipation of the final part of the question in which we are
asked about the concavity of the curve, we will include in our table of
d?
sign all those z-values at which a could possibly change in sign. Since

d’y _ —1568(x — 9)(x? — 32 + 9)

dz? ——s (x +: 12)3(42 —15)3)—’
the x-values of interest to us are r = 9, r = —12 and t = 15/4. So the
required table of sign is

“rel pe in |
RRS ee Pee ee ey ee

elim Aor (42 — 15) = 0

|
Tora oa ee
CONCAVITY
[fw] — po
[UP | | DN |toon
.
 Ch: 10 85

Since the curve changes from being concave up to being concave

down as x increases through the value z = 9, it follows that we do have
a point of inflection at zc = 9, y = 4/9.
Finally, we deduce from the above table of sign that
y = f(x) is concave down precisely when zx € (—12, 13) U (9, 00).
We complete the solution of the example with the sketch of the
curve given earlier, incorporating all the information we have obtained.
Before our final example of this Chapter, we need to consider the
relationship between the sizes of the three functions 2", e” and Inz as
z — too. As x — oo, we know that x” (where n is a positive real
number), e” and Inz all become large and positive. However, if we
consider the values of, say, 2°, e? and Inz at x = 1000 we see that
there is a great difference between the values. For 1000? = 10°, while
1000 ~ 2 x 10434, but In1000 = 7. If we took a larger value of 2, then
the difference between the values would be even more pronounced.
What this suggests is that, for large values of z, e” dominates 2”,
while in turn z” dominates In z, where n is any positive real number.
[The restriction that n should be positive is imposed so that z” shall
tend to oo as x increases. If n were negative, then x” would tend to
zero, instead.|
The implication of this domination is that if we were required to
calculate, for example, z*/ exp z, as z — 00, then we would have
a large positive quantity
a far larger positive quantity
so that the ratio would tend to the value zero, through positive values.
In symbols, we would write
at/expr—0t as 2-00.
Again using the idea of domination, we could see that as z — oo,
exp z/Inz would become
a very large positive quantity
a much smaller positive quantity’
and so expz/Inz would tend to +00.
Now let us consider what will happen to x™/expz as x — —oo,
where m is a positive integer. Remember that 1/expz can be rewritten
as exp(—z), and so we are considering what happens to +™ exp(—z)
when x — —oo. Let us put y = —z, so that we are finding what
happens to (—y)™ expy as y > +oo. As y becomes very large, then
exp y will become extremely large, and also (—y)™ will become large (in
modulus) and either positive or negative, depending on whether m is
even or odd. So we have
r™
/ expr — oo as © — —on, if m is even,
 86 GhAl0

and
rz™/expz — —oo as x + —on, if m is odd.

Since Inz is not defined when zx < 0, it does not make sense to ask
what happens to Inz as x — —oo. However we can ask what is the
situation when we have, for example, a power of z multiplied by Inz,
and z — OT.
For example, consider what happens to z°Inz as x — 0+. Notice
that in this case, as — 0*, z° — 0, while Inz — —oo. We resolve the
problem of finding the limit by putting y = 1/z. In this case, we want
to find what happens as y — 00 to the expression (1/y)° In(1/y), which
we can rewrite as (—Iny)/(y°). We already know that both Iny and y®
will tend to oo as y — oo, and we also know that any positive power of y
will dominate In y. Hence it follows that, as y > 00, (—Iny)/(y®) 4 07.
After this preliminary discussion, we are ready to tackle the follow-
ing example, which uses the ideas we met in sketching rational functions,
applied to a different situation.

Example 3. Sketch the curve y = (4 — 2)e~”, finding the critical

point, the point of inflection, and the points where the curve cuts the
coordinate axes.
Solution. Since e~* is always positive, we can only have y = 0
when z = 2. Hence the curve meets the z-axis at the point (2,0) and
at no other point.
The curve cuts the y-axis where x = 0, so that y = (—2)e® = —2.
Hence the curve meets the y-axis at the point with coordinates (0, —2).
As tz — oo, y & 2x/e*, which, by the discussion preceding the
example, will tend to zero through positive values.
On the other hand, as 2 — —oo, x — 2 will tend to —oo while e~7
will tend to oo. Hence the product of these terms will tend to —oo as
r— —o0.
Notice that when we are calculating the behaviour of the function
for values of x which are large (in modulus), we can ignore the effect of
the —2 as this will be swamped by the z-term.
We now need to use Calculus to obtain the coordinates of the critical
point and point of inflection on the curve.
Since y = (x — 2)e~*, we have, on differentiating using the Product
Rule,
d
= (2 DeF pe-F = (1-2 + 2)e™* = (8—z)e™*.
Differentiating again using the Product Rule, we will find that

qa? = (3 —2)eF + (=e? = (1-3 + a)e* = (e— 4)e™*,

 Ch: 10 87

From these derivatives it follows that the curve can only have a
critical point at z = 3 (since e~* is always positive), and also that the
only possible point of inflection on the curve will lie at the point x = 4.
When z = 3, we have y = e~* and dy = —e*. It follows that we
have a local maximum at the point (3, e~%).
When z = 4, we know that y = 2e~*. To decide whether this is a
ot
point
hang. ! d?
of inflection, we need to determine the signs for ra in the vicinity
by
of the z-value x = 4.
Since the expression for oy is fairly simple in this case, we shall
dz p
manage to do this without using a table of sign. For oa =(r—4)e”,
L
where e~* will always be positive. Therefore the sign of —§ will be
the same as the sign of z — 4. It follows that as x increases through the
value x = 4, so the sign of
ay will change from being negative, to zero,
az”
to being positive.
Hence we do indeed have a point
of inflection on the curve at the point
(4, 2e~*).
We are now able to complete this
example by sketching the curve, us-
ing all the information we have found
above. Notice that, unlike the ra-
tional functions we sketched earlier,
there will not be any vertical asymp-
totes to the curve, since these only
occur where we are dividing by an
expression that tends to 0. On this
curve there is no finite value of zx for
which that happens.
EXAMPLES 10
1. Let f(z) = 2r* +32? — 122 —6. Express as a union of intervals
the set of real numbers for which f'(z) > 0.
2. Let f(x) = 2t + 42° — 8x? — 482 + 50. Express as a union of
intervals the set of real numbers for which f(z) < 0. .
3. Let f(x) = (x? — 6x + 7) e?*. Find the intervals on which f(z)
is strictly increasing.
4. Let f(r) = sin? zcos2z. Find the intervals, contained within
[0, x], on which f(z) is strictly decreasing.
 88 Ch: 10

5. For each of the following curves, find where it meets the axes,
find the z-coordinates and nature of each critical point, and the coordi-
nates of any points of inflection, and hence sketch the curve:
(a) y=2*-—32?-2
+3;
(b) y=a2?
—62? +92 —4.

6. (a) Find the critical points on the curve

(x-6)(z+2)
1s Wa(eeae)
Determine the nature of each critical point, the intersections of the curve
with the coordinate axes, and the vertical asymptote to the curve. Find
the asymptotic behaviour of the curve as z — too. Draw up a table of
sign for y, and hence sketch the curve.
(b) Deduce from (a) a sketch of the curve

(x -7)
SENG)
7. (a) Find the critical points on the curve
_2(#-1)(e+7)
(g=s2)e er
Determine the nature of each critical point, the intersections of the curve
with the coordinate axes, and the vertical asymptote to the curve. Find
the asymptotic behaviour of the curve as r — too. Draw up a table of
sign for y, and hence sketch the curve.
(b) Deduce from (a) a sketch of the curve
Bs (x — 2)
u 2(@-N@+)
8. (a) Find the critical point and the point of inflection on the
curve
4 (x — 3)

(x-1)"
Determine the nature of the critical point, the intersections of the curve
with the coordinate axes, and the vertical asymptote to the curve. Find
the asymptotic behaviour of the curve as sr+ too. Draw up a table of
sign for y, and hence sketch the curve.
(b) Deduce from (a) a sketch of the curve

rd (x — We
akGesy)
 Ch: 10 89

9. (a) Find the critical point and the point of inflection on the
curve
= (x + 3)
(x+2)
Determine the nature of the critical point, the intersections of the curve
with the coordinate axes, and the vertical asymptote to the curve. Find
the asymptotic behaviour of the curve as r — too. Draw up a table of
sign for y, and hence sketch the curve.
(b) Deduce from (a) a sketch of the curve
_ (e+2)
Te erry
10. (a) Find the critical point on the curve
fe +4) (@=-4)
~ (2 +2)(2 —2)
Determine the nature of the critical point, the intersections of the curve
with the coordinate axes, and the vertical asymptotes to the curve.
Show that the curve has no point of inflection. Find the asymptotic
behaviour of the curve as x — too. Draw up a table of sign for y, and
hence sketch the curve.
(b) Deduce from (a) a sketch of the curve
fyok(x + 2) (2 — 2)
(x + 4)(r@ — 4)

11. (a) Find the critical point on the curve

go)(tf 1)
~ (x —5)(r—1)°
Determine the nature of the critical point, the intersections of the curve
with the coordinate axes, and the vertical asymptotes to the curve.
Show that the curve has no point of inflection. Find the asymptotic
behaviour of the curve as x — too. Draw up a table of sign for y, and
hence sketch the curve.
(b) Deduce from (a) a sketch of the curve
PT GeOHASSY
CEES (Cos 85)
 90 Ch: 10

12. Do each of the following (1) become very large, (2) become
very large (in modulus) and negative, or (3) tend to zero?
(a) 2%e?* as x becomes very large.
(b) 2%e-3* as x becomes very large.
(c) 2e~* as zx becomes very large (in modulus) and negative.
(d) 27e3? as x becomes very large (in modulus) and negative.
(e) x *e? as x becomes very large.
(f) 2zte?* as x tends to zero.
(g) «2 *Inz_ as z tends to zero (zr > 0).
(h) 2z*Ilnz as z tends to zero (xz > 0).

13. Sketch each of the following curves, after finding any critical
points, any points of inflection, those points where the curve meets the
coordinate axes, and after considering also the behaviour of the curve
as x — too:
(a) y=(e—3)e-*;
(b) y=(e+2)e;
(c) y=(e—-1)';
(a) y=(2+2)%en*.
 CHAPTER 11
Absolute Maxima and Absolute Minima
In this chapter, and in examples based on the ideas developed here,
we shall find it useful to have at our
disposal certain formulae, which are
all quite well known. We shall prove
them all later in the chapters on ap-
plications of integration.
For a right circular cylinder, of
height h, base radius r, the volume
is tr*h, while the curved surface area
(not including the base and the top)
is 2mrh.
For a sphere of radius r, the vol-
ume is Sar’, and the surface area is
Arr?,
For a right circular cone having
height h, base radius r and with slant
height J (so that 1 = Wh? + r?), the
volume is aar7h, while the curved
surface area (not including the base)
is mrl. (Note that the slant height
of a right circular cone is the dis-
tance from the vertex of the cone to
any point on the circumference of the
base of the cone.)
In some circumstances, partic-
ularly when tackling problems hav-
ing a ‘practical’ basis, we sometimes
need to find the greatest or the least
value that a function f(x) can take
when the variable z is restricted to
lie in a certain range of values.
We call f(c) the absolute maximum of f(z) for z € S if we have
f(c) > f(z) for all z € S (with of course c € S). From this definition,
the absolute maximum that f(x) takes on the set S is precisely what it
says, that is, it is the greatest possible value the function takes while z
varies throughout the set S.
We define f(c) to be the absolute minimum of f(z) for z € S if
f(c) < f(z) for all c € S (with of course c € S). Again, the absolute
minimum on a set S is precisely the least value the function takes as zr
is allowed to vary throughout the set S.
From the diagram, which shows a typical situation, you can see
that an absolute maximum and an absolute minimum need not occur
 92 Ch: 11

at local maxima and minima on the curve.

It can be shown though that if S
1s a closed interval then the absolute absolute and
local maximum
mazimum and absolute minimum of
f(z) on S must occur either at end-
points of S or at local mazima or lo-
cal minima of f(x) on S.
Suppose we are given a closed
set S, and asked to find the absolute
maximum and absolute minimum of local minimum absolute
minimum
f(x) on S. Our routine is to find the
values of f(z) at any turning points
on the curve in the set S, and to find O}| €—closed interval S ———}r
the values of the function at the end-
points of S, and then choose the greatest of these to be the absolute
maximum and the least to be the absolute minimum.

Example 1. Find the absolute maximum and absolute minimum

values of the function f(x) = 2x3 — 32? — 362 + 70 for0 <2 <4.
Solution. We start by finding the turning points on the curve.
Now
f'(x) = 62? — 62 — 36
= 6(x? — x -6)
= 6(z — 3)(z + 2).

Hence the turning points on the curve will occur at the points x = 3
(which satisfies the condition 0 < z < 4) and x = —2 (which does not
satisfy the condition and so can be ignored).
Now we evaluate the function at the end-points of the interval [0, 4]
and at the turning point that lies in the interval. We have f(0) = 70,
f(4) = 6 and f(3) = —11. Hence the absolute maximum for f(z) on
the interval [0,4] is 70 which occurs when z = 0, while the absolute
minimum is —11 at z = 3.

In more practical problems, we will sometimes not be given the

restriction on the variable, but instead we will be expected to work
this out from the implications of the practical situation involved. One
frequent condition that is useful is that measurements of many physical
quantities, for example lengths, have to be positive. This condition is
used in the next example.
Example 2. A farmer wishes to enclose a field with a length of
fencing. Because of the use he wishes to put the field to, the farmer
wants to have a rectangular field, one side of which should have a gap of
10 m without a fence. Furthermore, the farmer wishes to have a fence
down the centre of the field, not completely dividing the field into two
 Chaat 93

pieces, but leaving a gap of 10 m at the edge of the field near to the
gap in the perimeter fence. Given that the farmer has a length of 196 m
of fencing to use, find the dimensions of the field having the largest
possible area.
Solution. Let x m be the length of the side of the field that
contains the gap, and let y m be the
length of the sides of the field per- zm
pendicular to that one. Then let A
denote the area of the field so that
A= zy. We need to find the abolute
maximum value that A can have.
We have a restriction imposed
on z and y by the fact that the farmer
has a finite length of fence at his dis-
10 m
posal. The total length of fence he
10 m
will use in the set-up we have de-
scribed is x + (2 — 10) + y+y+(y— 10) = 22 + 3y — 20. We require
2z + 3y — 20 = 196, from which we find that x = $(216 — 3y) or that
y = 4(216 — 2z).
From practical considerations, we need to have z > 10, since we
require a gap of 10 m in the fence on one of the sides of the field of
length z, and also we need x < 93. This requirement arises because we
need to have y > 10, so that, using z = +(216 — 3y), we obtain the given
restriction on z.
Hence, substituting for y in A = ry, we have A = }$2(216—2z). So
we want to find the absolute maximum of the function A = }2(216—2z)
subject to 10 < rt < 93. Now
2D = =(216
— rea - 42)
_ 5(54 ae
so that the turning point on the curve occurs at r = 54. We evaluate
A at each of the values z = 10, zc = 54 and x = 93. We find that
A(10) = 1960/3, A(93) = 930 and A(54) = 1944. Hence the absolute
maximum value occurs at the local maximum of the function, with the
field having dimensions 54 m by 36 m.
Example 3. It has been suggested that there is a relationship
N = 2002(108 — x?) between the number N of birds that can survive
in a breeding colony of sea-birds and the density x of the nests in the
colony. It is also believed that the values x = 2 and x = 8 represent
extreme values for x beyond which the colony would not be viable. Find
the absolute maximum and absolute minimum values for N.
Solution. We simply need to find the absolute maximum and
absolute minimum values for N with z restricted to the range [2, 8].
 94 Ch: 11

Now
N = 200 (1082 — 2°)
so that AN
—ae = 200 Fk (108 — 32?
x’)
or ony
iS 600 (36 — x”) = 600(6 — z)(6 +z).
z
From this we see that the turning points on the curve will occur at z = 6
and at x = —6. However, notice that we are not interested in the value
xz = —6 since this lies outside the range of values we are to consider.
Now we evaluate the function at the three points r = 2, ¢ = 6 and
xz = 8, since these are the end-points of the interval together with the
turning point.
We have N(2) = 41600, N(6) = 86400 and N(8) = 70400. Hence
the absolute maximum for N is 86400 which occurs when z = 6, while
the absolute minimum is N = 41600 which happens when z = 2.
When we are considering the case of a function defined on an in-
terval S which is not a closed interval, we have to treat the problem
of finding the absolute maximum and absolute minimum slightly differ-
ently. In this case, we again consider all the values of the function at
any turning points that lie within S, and the values of the function at
any end-points of S that lie within $. However in this case we also need
to consider the value that the function approaches as x approaches any
end-points of S that do not lie within S. The greatest of these values
gives the absolute maximum for the function, provided it occurs for a
value of z lying within S. Otherwise we say that the absolute maximum
does not exist. We deal in a similar manner with the absolute minimum.
Example 4. Two positive numbers are such that their sum is
100. Find the numbers if the product of the square of the one with the
cube of the other is to be a maximum.
Solution. Let the two numbers be z and y. Then we have
z+y= 100. We need to find the maximum value of T = 273, subject
to the requirement that z, y are positive. We can put these conditions
on z in the form z > 0 and x < 100, since if y> 0 and r+y = 100 then
xz < 100 follows.
We need to find the maximum value of T subject to 0 < z < 100.
We use the equation z + y = 100 in the form y = 100 — z to eliminate
y from the formula for T, so that we can use differentiation to find the
maximum.
Now T = 27(100— 2), and so, using the Product and Chain Rules
to find at we find that
dz
 Ch: 11 95

d fe
dz = 22(100 — z)* + 2? x 3(100 — z)? x (—1)
= z(100 — z)*2 (2(100 — x) — 3z)
= z(100 — x)? (200 — 2x — 3z)
= z(100 — x)? (200 — 5z).

The turning points occur where ~ = 0, from which we require that

z(100 — x)*(200 — 5z) = 0. This happens when z = 0, or x = 100, or
200 — 52 = 0 from which we deduce that x = 40.
Recalling that we are only interested in those values of x for which
0 < xz < 100, we see that the only turning point satisfying this condition
is the turning point at zr = 40. At z = 40, T = 40?.60°.
We also have to consider what happens to T as x tends to the ends
of the open interval to which it belongs. Now as x > 0, T > 0, while
as x — 100, T > 0 again.
It follows that the (absolute) maximum value of T occurs at the
turning point, and is given when the two numbers are 40 and 60.

Example 5. It has been suggested that in an underwater cable

with a core of radius r and an insulating sheath of radius R, the rate
of signalling is given by S = —kz*|nz, where k is a positive constant,
and z = r/R. Find the value of x which gives the most rapid rate of
transmission.
Solution. We need to find the maximum value of S, subject to
the conditions 0 < x < 1. (There clearly must be a core to the cable,
and so we must have r > 0 and soz > 0. On the other hand, it would be
feasible, though rather impracticable, to have a cable consisting entirely
of an uninsulated core, which would mean that r = R, or z = 1.)
Now,
go = —2kelnz — kx*(1/z)
dz
= —-2krlnz —kzr
= —kz(2Inz + 1).

So 2 = 0 when x = 0 and when Inz = —1/2. This happens when

be
z = 0 and when z = exp(-—1/2). The turning point which belongs to
the interval (0, 1] is x = exp(—1/2). When this happens we have

S = —k(e7/?))? In (ey)

= -k(e71)(-1/2)
= k/2e.
 96 Ch:

We also have to consider what happens at the end-point z = 1 and

as xz — 0.
Now, at z = 1, we have S = —k(1)*1In1 =0, since In1 = 0. Finally
as x — 0+, then since, by the methods we used in the last chapter,
z*Inz > 07, it follows that S — OF.
Hence the maximum rate of transmission will occur for the value
x = exp(—1/2) = 0.607. (Although we were not required to find it,
notice that the minimum will occur here at z = 1.)
EXAMPLES 11
1. Find the absolute maximum and absolute minimum values for
the function
f(z) = x? — 327 — 92 + 24
for x in the interval (1, 4].
2. Find the absolute maximum and absolute minimum values for
the function as
f(z)=ar+ a
for x in the interval (4, 8].
3. A rectangle has a perimeter of 100cm. What is the maximum
possible area of the rectangle?
4. An isosceles triangle has a perimeter of 60cm. What is the
maximum possible area of the triangle?
5. The sum of two positive numbers is 24. Find the numbers if the
product of the one with the square of the other is to have its maximum
value.
6. A rectangular sheet of paper has an area 294cm?. The margins
at the top and bottom are 3cm while those at the sides are 2cm. What
is the maximum possible area of printing on the sheet?
7. A cylindrical mug is to hold 54cm? of liquid. The cost of
production of 1cm? of the base is twice the cost of 1cm? for the side.
Find the dimensions of the mug that will hold the required volume for
which the cost is a minimum.
8. A lens of focal length 75cm forms the image of an object on a
screen. If the distance ucm between the object and the lens is related
to the distance v cm between the image and the lens by the equation
el 1
uo 75°
express the distance d (= u + v) between the object and the image as
a function of u alone, and hence find the minimum possible distance
between the object and the image.
 9. A ray of light from the point
A is reflected by a plane mirror to the
point B. Let X and Y be the perpen-
dicular projections from the points A
and B to the mirror. Let |AX| = a,
|BY| = band |XY| = c. Suppose the
ray of light strikes the mirror at the
point P between X and Y, and that
|X P| = x. Express the distance trav-
elled by the ray of light in terms of z,
and show that this distance is a min-
imum when sina = sinf, where a
and # are the angles APQ and BPQ
respectively, where PQ is the line perpendicular to the mirror through
ie
10. Suppose that A and B are points with parameters a and b
respectively on the parabola y? = 4cr. Suppose that we are to find
the point T with parameter t on the A
parabola such that the area of trian-
gle ABT inscribed in the parabola is
a maximum, where T is assumed to T
lie between A and B. It can be shown
that the area of the triangle is
c? (t —a)(b—t)(b—a).
B
Find the value of ¢ in terms of a and
b which gives the maximum for the
above expression, subject to the con-
ditiona <t <b.
 98

CHAPTER 12
Introduction to Integration
Calculus consists of two important techniques—differentiation and
integration—and the application of these techniques to different types
of problems. We have already seen how to differentiate, and some of the
applications that can be made of differentiation. Now we shall follow
through the same procedure for integration.
Differentiation is concerned with the problem of finding the rate
of change of a variable y with respect to +, when we are given y as a
function of z.
Integration is concerned with the reverse problem. If we know that
d ;
y is a function of z, and we are given = the rate of change of y with
L
respect to x, can we find the function which expresses y in terms of z?
We shall start by considering the following simple example.
Example 1. A stone is thrown vertically upwards with an initial
velocity of 50 ft/sec. Assuming that the stone is subject to a steady
downward acceleration, due to gravity, of 32 ft/sec*, find how high the
stone is, and its velocity, 3 sec after it is thrown.
Solution. Let t denote the time (measured in seconds) from the
moment the stone is thrown, and let y denote the height (in feet) of
the stone above its starting position at time t. Also, we shall let v (in
feet per second) be the velocity of the stone at time t. Notice that v
will be given by = and that v will be positive if the stone is travelling
upwards, and that v will be negative if y is decreasing.
By assumption, as the stone moves through space, it is subjected
to a downward acceleration of 32 ft /sec”. Since acceleration is given by
the rate of change of velocity, it follows that we have
dv
aT —32.
Notice that we have a negative sign here because the velocity is
positive in an upward direction, and so the rate of change of velocity
would be positive if the upward velocity were increasing. Clearly this is
not the case, since the upward velocity is eee. as time increases.
v
We want to find what v is, knowing that = —32. One possible
dt
: d
value for v is that v = —32t (since then = = —32). Another possibility
is that we could have

v = —32t+ any constant

and in fact it can be shown that if ee = —32 then

dt
v = —32t+c,
 Ch: 12 99

for some value of the constant c. Furthermore it can be shown that this
is the only possible form of the solution.
From the conditions of the problem, we know that, when t = 0,
then v = 50. Substituting these values into the equation for v, we see
that
50 = —32 x 0+,
from which we have
es faX0).

Notice that we could have also obtained this result graphically, by

plotting the graph of v against t, which must be a straight line of slope
—32, passing through the point v = 50, t = 0.
Next we know that
dz
— =v= 50 - 32t.
dt
We now need to find what z could be. One possibility is that
z = 50t — 16¢?,
d
(because if z = 50t — 16¢? then = = 50 — 32t). Again it can be shown
that all possible solutions for z will have the form
x = 50t — 16t? + d,
for some suitably chosen value of the constant d.
Now under the conditions of this particular example, we have z = 0
when t = 0. We substitute these values into the formula for z and we
obtain
0=50x0-16x 074d,
from which we have d = 0. Hence we have obtained
x = 50t — 16t”.
Finally, we need to find the values of z and v when t = 3. Substi-
tuting t = 3 into the equations for v and z, we have
v = 50 — 32 x 3 = —46
and
z=50x3—16x
3? = 150-144 =6.
From these we see that after 3 seconds the stone is 6 feet above
the point from which it was thrown, and that it is falling at 46 feet per
second. (We obtain the information that the stone is falling from the
fact that the velocity (measured in an upward direction) is negative.)
Given a function f(z) of z, the indefinite integral of f(z), which
is denoted by
‘h
f(z) dz,
 100 Ch: 12

is the collection of functions of the form

F(xz)+¢,

where the derivative with respect to 2 of F(x) +c is f(z).

We call f the integral sign, f(x) the integrand, z the variable
of integration, and c the constant of integration, or arbitrary con-
stant.
Notice that

d
[tea = F(x)+c precisely when a ala ey Mea)

It follows from this that differentiation and integration are inverses of

one another, in the sense that

| EEO)d = Fe)+6
and

= (feer)
=10)
Since differentiation and integration are inverses of one another, it
follows that the rules concerning differentiation will have analogues for
integration. We shall later see the rules for integration that correspond
to the Product and Chain Rules. In the meantime we simply state
the following rules which apply for any functions f and g and for any
constant c:

fee dr = cff(2) de,

[u@+acey dt.= [tear [o(2)ae

The following table enables us to integrate many of the standard

functions. In each case the result is checked by showing that the deriva-
tive of the entry given as the integral in the table is the integrand. It is
important to memorise this table of standard integrals. This should not
be difficult since it is basically our earlier table of standard derivatives
of standard functions read back-to-front.
 Chal? 101

f
k
OX
grtl
n
cae n#0,n4-1SS
1

ee el Ce cet sae ee
aconstant
aFI,a>0
cos sive d Bal are tayston |
i
ie cc i i
secs2 2
cosec’ £
iene
—cotr+c
iv |
secatans
cosec xz cot x —cosecr+c
1 : =]
fron) SEs an SID 011
oa r —1,1

1 ~—1
3 tan” zr+ec
l+a

Only one of the above results is not quite straightforward, and that
; : 1 ;
is the one concerning / —dz. For z > 0 then certainly
z
d d 1
qq (initi+¢) = qq ine to) = i

If c < 0, then we let u = —z, so that u > 0. Also |r| = (—z) = u, so

that
In jal = Ine
By the Chain Rule, differentiating with respect to z, we have
d d du
qq (ln lz! +c)= qin +c) x 3

* x (=1)
I
—x(-l
7
x (-1)
(2)
Se
Hence, provided z # 0,
d 1
qq (in ltl + ¢) = -

From this it follows that

[5 =In|z|+c,
z
 102 Ch:il2

provided that z # 0, as required.

In a number of integration examples, the main difficulty lies in
trying to reduce the integrand to the forms of the standard integrals
given in the above table.
3 8
Example 2. Find i}G — 82° + 72? —7 + ae =) dz.
x £
Solution. Using the rules for integration, we have

[ (2-824 10?-74 5-5) dz

x x

= iB
(ae CE frst Sgde cosirmatahy OP osi AP 82) dz

3
= 50° — fot + io? Tet} Se - Set te

x 7x3 3 8
SeteDy a er pee ee ;
6 “T sh-gt eon = oa oa
Note that the routine for integrating powers of z, other than 27, is to
increase the power by 1 and divide by the new power.
Notice also that when, as in the above example, we have a number
of arbitrary constants arising from integrating the sum of a number of
terms, then we lump them all together as one arbitrary constant.
Finally notice that it is a good idea to check the result of finding
an indefinite integral. The easiest way to do this is to differentiate the
answer you have obtained, and see whether this reduces to the integrand
you started from.
Example 3. Find f cot? rdz.
Solution. Using the trigonometric formula cosec? = 1+ cot? z,
we have
[oot Cag = J(cose? x —1)dz

=-—cotr—zr4+e.

2
Example 4. Find / G + 5) dz.
Ho

Solution. We start the solution by expanding the bracket. We

have
5 2 2 10 Bee —
RE rey dz (23 +4773 +47 )de
Mb ‘

rs 4x3 4r-3
Naar mag a or
zy ye a:
3732
= = +6r$\- ac
 Ch: 12 103

Example 5. A curve passes through the point (0,3), and is such

that its gradient at the typical point (x,y) is 3r7 — 4+ sinz. Find the
equation of the curve.
é ca ¢
Solution. At the point (z,y), the gradient of the curve is os
b
which is given to be 3z? — 4 +sinz. Hence
d
oo = 3r* —4+4sinz.
dz
Therefore, on integrating this with respect to z, we have

y= ‘ (32? — 4+ sinz) dz,

from which we see that

y=2?>—4r-—cosrt+e,

where cis the arbitrary constant of integration. We know that the curve
must pass through the point (0,3), and this enables us to determine the
value of the constant c. For, on substituting ¢ = 0 and y = 3, we have

3=0° —4x0-cos0+¢,

from which c = 4. It follows that the equation of the curve is

y=2?
—42 —coszr +4.

One way of looking at the above solution is that for different values
of the arbitrary constant c we have a family of curves, all parallel to each
other, and each having the required gradient 32? — 4+ sinz. From our
earlier considerations of curve-sketching, we know that different values
of the arbitrary constant c will have the effect of moving the curve up
or down. So we require to choose the appropriate value of c such that
the curve passes through the point (0, 3).
Integration is a subject which developed as a means of solving prob-
lems which arose in practical situations. Before examining some of these
applications, we would like to extend the range of functions that we can
integrate. We shall do this by introducing two important methods of
integration, namely Integration by Change of Variable and Integration
by Parts.
EXAMPLES 12
1. A variable y is expressed in terms of x in such a way that
d
ae 2277,
dx
Given that y = 2 when xz = 1, find y in terms of z and find the value of
y when z = 2.
 104 Ch: 12

2. A particle passes the origin at time t = 0 with velocity 5 units.

The acceleration of the particle at time t is given by
d ‘
oe 2t — sint,
dt
where v gives the velocity of the particle at time t. Find v and z, the
displacement of the particle from the origin, in terms of t, and find
the velocity and acceleration of the particle when t = 7, and also the
distance of the particle from the origin at that time.
3. Find each of the following:
(a) if (323 +82 —-—2+ Jaa) dz;
(b) if (423 — 327? — 82—*) dz;
(c) f (a — 2) (a +4) dz;

(d) f (2? - We dz;

(e) f tan? z dz;
(f) f (secr + tanz)? dz;
(g) J cosec x(cosecx— cot x) dz;
(h) f5e dz.

4. A curve has gradient 3x? +cosz at the point (z,y) and passes
through the point (0,1). Find the equation of the curve.
5. A curve has gradient 2sinz — 3cosz at the point (z,y) and
passes through the point (4,2). Find the equation of the curve.
 105

CHAPTER 13
Integration by Change of Variable
The technique of Change of Variable enables us to reverse the re-
sults obtained by differentiating using the Chain Rule.
For example, if we wish to differentiate (x? — 52? + 7x — 3)°, we
put u = 2° — 52? + 7z — 3, so that on differentiating with respect to z
using the Chain Rule we obtain
3(2° — 52? + 72 — 3)? x (32? — 102 +7).
—————ooeesnsn iO
from diff. u® w.r.t. u ae

Again, if we wish to differentiate sin(5z +4) then we set u = 52 +4,

so that the derivative is
( + 4)
cos(5z es
Xecoey
from diff. sinu w.r.t. u ped

As another example, if we wish to differentiate e 42*—32 with respect

to z, the derivative is (if we let u = 4x? — 3z),
2
Cae x (82 — 3).
——_——— a
from diff. e* w.r.t. u qu
=z

Integration being the reverse process to differentiation, it follows

that

[oe — 52? + 7x — 3)? x (32? — 10z + 7) dz = (x* —52? + 72 —3)* +c,

[costs + 4) x 5dz =sin(5¢ + 4) + ¢,

ficial x (82 — 3) dr = eft’

32 4 c

c being the arbitrary constant of integration in each case.

In each of the above examples, the integrand has the form
(function of u) x (derivative of u w.r.t. x),
where u is a suitably chosen function of z. In each case we determine
the indefinite integral by integrating the function of u ies respect to
u, and then expressing u back in terms of z.
d
For example, if we put u = 2°—52?+72r—3, then = = 3r?-102+7,
and so

[oe — 52? + Tx — 3)? x (3x? — 10z+7) dz = [ow x “ dz

=u'+e

= (2? — 527 + 72 — 3)? +.

 106 Ch: 13

d
Similarly, taking u = 42? — 32, so that — = 8r — 3, we have

d
sfates x (82
— 3) dz safe x ie
=e"+c
= e4t ee
Pa

In both of the above, the connection is obtained by differentiating

the third expression with respect to z, using the Chain Rule. However
this does suggest a possible technique for integration.
Given an integral, try to express the integrand in the form
du
f(u) x ar’

where u is a function of z which is to be chosen from the form of the

integrand we are given. Then we find the integrand of f(u) with respect
to u, and finally we substitute for u in terms of z.
This procedure is known as integration by change of variable,
or integration by substitution.
There are a number of standard types of integrand for which there
are specific changes of variable. However, before tackling these standard
types, we shall look first at some non-standard integrands.

Example 1. Find ii(3x4 — 72? + 20)” (1223 — 14z) dz.

Solution. In this example, we notice that the integrand is in
two parts, and that the derivative of 324 — 7x? + 20 is 12zr° — 14z,
which is also part of the integrand. This suggests that we should put
d
u = 3r* — 7x? + 20, so that = = 127° — 142. Hence
t

/(Gre ES (oe cl. 20)” (haa — 142) dz = fe x ode

x
i
= -y® +c
6
1
= g (32" — 7x? + 20)® +c.
When making a change of variable in an integral, it is normal to
write du instead of “ dz, to denote that we are integrating a function
of u with respect to u.
Notice that after a change of variable in an indefinite integral, we
must always present our answer in terms of the original variable z, rather
than in terms of a variable u that we have chosen to introduce in order
to solve the problem.
 Ch: 13 107

Example 2. Find / (52° — 3)3 x? dz.

Solution. In this example, we shall put u = 52° — 8, since

we notice that the derivative of 5r* — 8 is (except for being multiplied
by a constant) also a term appearing in the integrand. So we have
— = 152”, which we rewrite as r2dz = ydu.
U ° °

dz

[st-abetar= fut x atu

= S103 is
ipl cong Nm
1 4
= 50 (= - 8)* +e.

Example 3. Find J(cos(see z)) secx tana dz.

Solution. Here we notice that secz appears in the integrand,
and also that secz tan z (which is the derivative of secx) is part of the
integrand. So we let u = secz, from which 7 eam ead tan z, which we
£
can think of as du = seczxtanz dz.

J(ces(eccz))secztanzdz = [eos du
=sinu+c
= sin(secr) + ¢.
From these examples, you can see how important it is that we can
recognise that an integrand can be written in the form
(function of u) x (derivative of wu),
where u denotes a suitable function of x. In order to be able to recognise
possible substitutions, it is very important to have the table of standard
derivatives firmly fixed in the memory.

Example 4. Find iSo dr.

Solution. In this example, we will put u = 327 +7, because then

d
= = 62, and so we can replace x dz by < du. So
L

x _ Vd
fgdu
lwere- u
while da
Pe u
1
= —In|u| +c
6
 108 Ch:i13

1
In |3x? + 7| + ¢.
>|

Example 5. Find ioe ® sec? x dz.

Solution. In this example, the key is to notice that the derivative
of tanz is sec? z. This suggests that we should put u = tanz, from
which we have du = sec? x dz.

[etn sect 2 de = eta

e"+c
= e ne

One very common type of change of variable is the linear change

of variable
u=azr+b,
where a, 6 are constants with a # 0. We can use linear changes of
variable in order to extend our earlier table of standard integrals.

faegrand =Fla) |Integral =JFe)

1 17 a a positive constant
copa s s in —+c¢ x € (—a,a)
Va? — x? a
1 ae
Oe Fay —tan —-+c a a non-zero constant
a
Teepe rw nonzero constant
a*‘+a2z a

1
] tant -

Each of the above can easily be checked by differentiating the in-

tegral we have quoted and showing that we obtain the integrand given.
Instead, however, we prefer to establish some of these integrals by means
of change of variable.
Let m be a non-zero constant. Then, if we let u = mz, we have
du = mdz, and so

. 3 i)
sinmzdz= | sinu x — du
m
1
= ~ fsinudu
m
1
= oh te
1
= ——cosmr +c,
m
 Ch: 13 109

while similarly

f
[cosme de = |cosy x —du
m
1
== |cosuds
m
em
i=
SIT th =
m
a
= —sinmzr +c.
m
Now suppose that a is a positive constant. Then this time we shall
x
take u = —, so that du = — dz, or dx = adu. Hence, we have
a a

Making the same substitution, but this time taking a to be any

non-zero constant, we have

dz =/ adu
az+22 | a2+4a2u2
=f adu
< a? (1+u?)
-</ du
af 1+u?

49
—+c.
a
Now we assume that m is a non-zero constant. We shall let u = mz,
from which we have du = mdz, so that

ities: |<* x sae

m
 110 Ch: 13

1
= feta
m
1 u
=—e +c
m

GS1 mr
op (ee
m

For our final verification, we take a to be any constant. Then

putting u = 2 +a, we have du = dz, and so

dx du
[=> UU
=In|u| +c
= In|z + a|+c.
We shall use the above results in the following straightforward ex-
amples.

Example 6. Find fe sin 2x — 6 cos 3z) dr.

Solution.

J(2sin22 - 6.0532) de =2 |sin2x de ~6 fcose dz

=a? (-5) cos 2z — 6 (3)sin3z+c

= —cos2zr — 2sin3z +c.

; dz
Example 7. Find 943

Solution.
/ dz 1 denabeo
Gaeaie 5 aoe
Sometimes we are given an integrand of the form a) ae
J/quadratic
Saas: In order to deal with such integrands, we often need to
complete the square and then make a linear change of variable in order
to convert the integrand to one of the standard forms we have listed in
the table above.
 Ch: 13 111

Example 8. Find ——— ge

V7 — 6x — 2?
Solution. We start by completing the square in the expression
7 — 62 — 2?. Now
7 —62 —2? =7 — (2? + 62)
=7—|(2 +3)? —9]
= 16 —(z +3)?
= 4* —(r +3)’.
Hence, if we put u = 2 +3 so that du = dz, we have

dz
E xamplele 9. Find
in fh
| —————~.
Peer a

Solution. Again we start by completing the square, this time in

the expression xz? + 8z + 41. Now

2? + 82 +41 =(2# +4)? — 16441

= (x +4)? +25
= («+ 4)? +57.
Hence, taking u = z + 4, so that du = dz, we have

/ dz = dx
z?4+82+41 J (x44)? +5?
=f du
Sur +8?
 112, Ch:4i3
EXAMPLES 13
1. By making a suitable change of variable, find each of the fol-
lowing:

eo /o/e four, B= 5 i# (ae* —4)° de

e 43
© fame © faa
(a + 32? — 5). (ae + x) dz;

(x* — 22? yg) (<3 — x) dz;

iS (sin (tan x)) sec? x dz;

(tan® z — tan? z) sec 2 > dz;

e"™* cos
x dz; (j) sfisete sec x dz.

2. Write down the integral with respect to z of each of the follow-

ing:
(a) sin5z + cos3z; (b) cos4z —sin2z;

oO a= Oates
OMe © ao
(GB) ae: (bh) ems
0 5 oO
3. By making a suitable linear change of variable, find each of the
following:

(a) ii(32 +5)" + (32 + 5)*| dz;

(b) / [(22Bee ea OES 3)**] dz:
(c) Je + 7)3 dz; (d) [i - 5)? dz;

© [a> © {ie
(©) fcos(3e—2)de; (hy) fsin( 52 +2)dr;
 Ch: 13 113

(i) /sec” 7x dz; (j) /cosec? 4z dz;

(k) [sects + 4) tan(3z + 4)dz;

(1) Jcosee(2 + 5) cot(2zr + 5)dz;

m /dr @ [>
dr

(0) tofoot des 2oii(porn fredttdey

(q) eae (r) [orae.

4. By completing the square, find each of the following:

dz dr
(a) l= ») /=7
dr dr
(<)
ee
V15 — 22 — x?
d
(d)
SS
V21+ 42 — x?
dr dz
(e) Vea (f) Wise: Sa
 114

CHAPTER 14
Integration By Parts

The method of Integration by Parts corresponds to the Product

Rule for differentiation. Recall that, if f(x) and g(z) are two functions
that can be differentiated, then by the Product Rule, we have

4 (4(2) x o(2)) = f(2) x £ (o(2)) + (2) x & (S@):

From this it follows, since integration is the reverse of differentiation,
that

Fey ene fifle) x & (g(2)) de + /g(2) x = (f(2)) de.

d ;
Writing f', g’ rather than < (f(z); fe (g(x)), respectively, and rear-
ranging the terms, we see that

fax ffolde+
|of'de,
or

|totae=to-
fof de
This last version should be memorized, and is the rule for Integration
by Parts.
In order to use the method of Integration by Parts successfully, we
need to be able to recognise an integrand as a product of two terms
(f and g’), one of which (g') we will need to integrate, the other (f) we
will need to differentiate.
Be warned that not all integrands which are products of two terms
need to be integrated by parts. In fact the majority of examples from
the last chapter were integrands which were products. However the
examples in that chapter had the feature that where there were two
parts to the integrand, one part was the derivative of the other. This
was what led us in those examples to use a Change of Variable.
It is a good idea to consider Change of Variable as the preferred
option, reserving Integration by Parts as an alternative when Change of
Variable methods are unsuccessful.
Having decided to apply the technique of Integration by Parts to
a particular example, the hardest part of the solution is frequently to
decide which part of the integrand is to be f and which is to be g’. Notice
that g' has to be a function that can be integrated, and that f must
be a function that we can differentiate, and which preferably becomes
simpler when we differentiate it. Although there are exceptions, one
good candidate for f is a positive integral power of z.
 Ch: 14 115

Example 1. Find [oc —- 4) sin 4x dz.

Solution. In this example, we put f = 52 — 4, from which
we have f’ = 5, and we put g’ = sin4z, leading to g = fsin4rdzr =
—} cos 4z. The reason for this choice is that 52—4 becomes simpler when
differentiated, while sin4z does not become much more complicated
when integrated. Had we chosen to take f and g’ the other way round,
then when we integrated 52 — 4 we would have obtained 33? —Az,a
polynomial of higher degree than we started with.

[oe — 4)sin42 dr = (52 — 4) (~7cos4] _ [joo x 5 dz

1 4)
lI = qlse -_ 4) cos 4z + i [cosa dx

1 ‘Sab
— 7 (52 — 4) cos4z + a x q sinde +c

Lis 4) cos4a + dy ck
4 : ge.
led ro cos <== ERI Gs

You will notice that in the above example, and in all examples involving
Integration by Parts, we do not include an arbitrary constant when
finding g from g’ by integration. The reason for this is that if we did
include such an arbitrary constant then we would find that it would
cancel out from the final answer.
However we do need to include an arbitrary constant when we have
completed the integration. Notice that in the final lines of the above
example we wrote the solution as
1 ar a
+7 (52 — 4) cos4z + nie q sin de +c

rather than as
1 5 1
— 7 le — 4)cos4z + ries G sin 42 + ) '

which equals
1 2): eee 9)
—4\52 — 4) cos 4z + 4% q sin 4a + 4% ee

These are equivalent, since c and c* are both arbitrary constants, and
’ x (an arbitrary number) is equal to another arbitrary number. We
chose the form we did simply because it looks slightly neater.

Example 2. Find ee cos 32 dr.

Solution. In this example, it will be convenient to let

f— ia cos
32 dz.
 116 Ch: 14

For the same reason as in the last example, we put f = 2”, giving
f' =2z, and we put g! = cos3z, from which g = f cos3z dz = i sin 3z.

h= fe cos
3z dz

1
=z? (jsin32) - [ jsinse x 2x dr

2
= = sin Be — = | sin edz

In order to integrate z sin 3z we use integration by parts again. We

must take f = z, leading to f’ = 1, and g’ = sin3z, from which we have
g = f sin3z dz = —} cos3z. If we took f = sin 3z and g' = z then after
integrating by parts again we would return to the original integral and
need to start the solution all over again.
2
1 S singe — 5f asin 3edz

2
= sin 3a — 5(:x (—5cos
3) - |- 30832 x 1d: )

2 y) 2
= sin 32 + qr eset = 5 |cose dz
2
= = sin Se + oy bse 9e - Pontngeite,
9 27
In the next example, the natural thing to do would be to take
f =2z°+6z and g’ = Inz. That would be no good however since we do
not yet have a way of integrating Inz. Since we can differentiate Inz
and integrate x? +62, we choose to put f =Inz and g' = x? + 6z. This
at least enables us to start on the solution of the problem.

Example 3. Find fe +6z) Ina dz.

Solution. Taking f = Inz and g' = x? + 62, for the reasons

given in the introduction to this example, we will have f’ = — while
z
g = J (2° + 62) dz = 424 + 32?. Hence, we have

Je + 6r) Inzdz = (j= +32") Inz -| a +3:') use:

x
 Chita 107

Example 4. Find /(2 — 3) e3? dr.

Solution. The choice of f and g’ is made here for exactly the
reasons we used in the earlier examples. We put f = 2r — 3 so that
f' = 2 and we put g’ = e®* from which we have g = }e°*.

ic — 3) e°* dr = (22 — 3) (Se) = iise x 2dzr

= 5(22 — 3)e3* — 5 fe dz

= 5(22 — 3)e** — =e +c.

In the following example, we use a trick which sometimes works and

which enables us to use Integration by Parts on integrands which do not
appear to be products. The trick consists of writing the integrand as 1
times itself.

Example 5. Find [tn de.

Solution. Having written Inz as 1 x Inz, we put f = Inz, so

1 ; :
that f' = —, and g' = 1 from which g = zx. There would be no sense in
taking f = 1 which would just lead to f’ = 0, and leave us no further
forward. Then

proede= [1x Inedr

1
=enz- [2(2) dr
x

=elne- [ids

=rlnzr—r+c.

A similar method works for f sin~! x dz and for jitenSc dz:

Example 6. Find ii
sin”! z dz.

1
Solution. Here we shall let f = sin™' 2, so that f' = Fra
and we shall let g' = 1 from which g = z. Then we have

j[snteds = in x sin-} 2dzr

=z2zxsin-!2— | sx ———=dr
 118 Ch: 14

In order to find
la dz, we notice that the denominator

involves 1—2z? and that the numerator is z, which is a constant multiple

of —2z. So we make the substitution u = 1— x?, so that du = —2z dz,
and -4 du = zdz. Hence we see that
5: 51 Ley bd zr
[osm zdzr=zrsin 2£ [a2

250Pes fo =3 du—
Ju
1 du
= TSsin~ 2s)
2 U2

1
=zrsin-t 2+ 5 (2u*) +c
1

=rsin tr+ (1-27)? ta

=rsin tr+VJ/l—z24+c.

EXAMPLES 14
1. Use Integration by Parts to find each of the following:

(a) ze>* dz;

(b) x sin 3z dz;

(c) zr? Inz dz;

xrcos5z dz;

x? sin 42 dz;

(x? — 2) e** dz;

(a4 + 42 — 5) sin 22 dz;

(x? + 2) cos 52 dz;

zt? Ina dz;

sin! 2z dz.
SS
=
 119

CHAPTER 15
Trigonometric Integrals
In this chapter we shall develop methods by which a number of
integrals involving the trigonometric functions can be found. Later on,
we shall be concerned to find methods for integrals involving powers of
sec x and tana, and for integrals involving powers of cosec x and cot z.
Firstly though we deal with the simplest case, which is integrals of
the form
[se rcos™zdz, where m,n>0.

There are different methods for various different cases. These are
as follows:
(1) if m is an odd positive integer, and n is not, substitute
u= sin 7;
(2) if n is an odd positive integer, and m is not, substitute
== COS e:
(3) if m, n are both odd positive integers, put u equal to the
one of sinz and cosz which occurs to the larger power (noticing that if
the powers of sinz and cosz are equal, then it is slightly easier to put
u = sin, but it really does not make much difference in this case);
(4) if m, n are both even positive integers, use the double angle
formulae
1 ul
cos 2 t= 5(1 + cos2z), Silo 2 f= 5 (1 — cos 2z)
to reduce the integrand to a form which can be integrated by the meth-
ods of (1)-(3) above.
In each of the cases (1)-(3) above, it may also be necessary to use
the formula cos? z + sin? 2 = 1 in either of the forms
cos’? z =1-—sin’z
or
2
sin 2>=1-cos*z
in order to reduce the integrand to a suitable form.

Example 1. Find [st xzcosz dz.

Solution. Since this integral involves an odd power of cosz and

an even power of sin z, it is of type (1). So we will substitute u = sin z,
so that du = cosz dz.
[sint x cose dz = futa
 120 Ch: 15

Example 2. Find fsin? x cos® x dz.

Solution. Again, this integral involves an odd power of cosz,

and so we will put u = sinz, and du = cosrdz. This time though, we
will need to use the formula cos? z = 1 — sin? z in order to express the
integrand in terms of sinz and hence in terms of u.

[ow zcos® z dz = [sx z (1 — sin? x) cos x dz

set
bouitiligu +c

i 1
= 3sin'c — Fsin’ a te.

Example 3. Find /sin? x dz.

Solution. This time the power of cosz in the integrand is the

power zero. The expression sinz occurs to an odd power. This means
that this integral is of type (2), and so the appropriate substitution is
u =cosz, with du = —sinzdzr. Again we will need to use the formula
sin? x = 1 —cos? z in order to put the integrand in terms of cos z before
we can integrate it.

[sno ede = |snte sine de

= He(sins x)” sin z dr

aes
(1 — cos? z) sin xzdz

(1 — 2cos? x + cos* 2) sinz dr

(
(substituting u = cos 2)
=
= («
2 3 ive
quit eu’)te

3
me eS aNd3 ee 5 s-r+ec.
eye +
 Ch: 15 121

Example 4. Find [sat zr cos zxdz.

Solution. In this integral we have a fractional power of sin z.

This does not affect the difficulty of the problem, since we have an odd
power of cos z, so that method (1) will apply.
ee 1
[sn coos dz = fubdu

3
= SUE See

No
bo
Cl 3
=-—sin?zr+4+c.
eo

Example 5. Find [se x cos’ x dz.

Solution. In this integral we have both sinz and cosz to even

powers. Therefore we have to use method (4), which involves using the
double angle formulae to try to reduce the integrand to a more amenable
form.
1 il
sin? ¢ cos?x = Ee — cos 20) x Ee az cos2)

1
= 7(1 - cos’ 22)

1 1
ay ¢ - (50 + os 42)))

pal 1 il il c
= 5 mg £

af il il
= 3 (5-542)

1 1
Sizes

Notice the second use of the double angle formula above in order to
write cos? 2z = a(1 +cos42). Asa result of the above trigonometric
manipulation, we have
il 1
[xt x cos’ zdzr = ‘ (5_ 5 cos) dz

ph
Peas
eee,
Pe vitienpias Cc

=
Ait
i
g”
canting
5
Saat a Cc.

Before seeing how to tackle integrals involving secz and tanz, we

must consider how to deal with the simplest integrals of this type. Both
of the following results are standard integrals, which should be learnt.
 122 Ch: 15

[rsnedz = In|seczr|+c

and
[sccxas =In|secr
+ tanz|+c.

In both of these, the simplest proof involves using the fact that we
know the correct answer. For

sec
7 tans
frsncde= [2 as
sec zr
du
Uu

(if we let u = secz so that du = secz tanz dr)

=In|ul
+c
= In|secz|+c.
Similarly, we have
[recede = [ EHR? ae
seczr +tanz
(sec? z + sec zr tan z) q
EEE Ky

secr + tanz
oe jdu
u
(on putting u = secz + tanz and du =(secz tanz + sec? x) dr)
=In|ul| +c
= In|secr + tan2z|+c.
The methods for dealing with integrands of the form sec™ z tan” z,
where m, n are integers (m, n > 0, of course) are as follows:
(1) if there is an even positive power of sec z then put u = tanz;
(2) if there are odd positive powers of both secz and tanz then
put u = secz;
(3) if there is an odd positive power of secz and an even power
of tan z then express the integral purely in terms of sec z, and use Inte-
gration by Parts.
This final case is usually a very unpleasant one to deal with, in-
volving quite difficult manipulations.
In all of the above cases, the formula sec? = 1+ tan? x can be
very important.
 Ch: 15 123

Example 6. Find [se ztan® x dz.

Solution. In this example we have an even power of secz and

so method (1) applies. We shall therefore put u = tanz from which we
have du = sec? x dz.
sec? tan’2 da = [tas

on we
4
=- Sete: Ce
4

Example 7. Find [sc ztan® «dz.

Solution. In this example we have an odd power of secz and an

odd power of tanz, and hence method (2) applies. We shall therefore
put u =secz so that du = secrtanz dz.

[se ztan® 2 dz= [sc! z tan’ z(secztanz) dz

= [sc xz (sec? z — 1)(secz tanz) dz

= few —1)du

i
== 5 +c
7
1 1
= 7 sec! x — z seer +c.

As suggested earlier the hardest type of example is that for which

we have an odd power of sec x combined with an even power of tan z.

Example 8. Find [sect eae.

Solution. It will be helpful in this example to let

i= [sect eae.

We begin the solution of the example by writing the integrand in the

form secxz x sec? z, and we shall use Integration by Parts to find an
expression for J. In the Integration by Parts we shall let f = secz,
which leads to f' = sec z tan z, and we shall take g' = sec? z, from which
we have g = f{g'dz = f sec*rdz = tanz. Any alternative choices for
f and g’ would have led to a more complicated situation after the first
Integration by Parts.
 124 Ch: 15

[sect ede = [sccxsect dz

=secatan— /tanz(sec tan) da

=secxtanz— | secrtan’ 22d

rar

= seczrtanz — [sce (sec? xr 1) dz

(from the formula sec? z = 1 + tan? z)

= secrtanz — [see xz —secxr)dr

=secztane— fsehzde+ |secxdz

= seczrtanz —I+I1n|secz + tanaz| + 2c.

Hence we have
I =secztanz
—I +1n|secr + tanz| + 2c
so that
21 =secrtanz
+ In|secz + tanz|+ 2c
or
1 1
[= 5secxtanz+ 5 In|seca + tan z|+e.

You will have noticed that in the above example, we arrived back
at our original integral after Integrating by Parts, so that we were then
able to obtain an equation involving J which we could solve to find J.
You will also have noticed that when we introduced an arbitrary
constant we called it 2c rather than c. The reason for this was simply
that we knew what was coming, and by taking the arbitrary constant
as 2c we obtained a slightly neater-looking answer. However it would
not have been wrong to have taken the constant as c.
Finally, notice the point at which we did introduce the arbitrary
constant. This was the point at which we were equating the two versions
of f sec’ cdz. The reason for bringing in an arbitrary constant at that
point is that any two indefinite integrals for sec® x will both equal the
same function of x, but they could differ by a constant. So we introduced
the constant at the point we did in order to compensate for this possible
difference.

The rules involved in integrating functions involving cosecaz and

cot x are very similar to those involving sec z and tan z, with the obvious
changes that tanz is replaced by cot z and sec z is replaced by cosec z.
There are also some changes of plus signs to minus signs.
 Ch: 15 125

Before tackling some integrals involving cosec z and cot z, we must

consider how to deal with integrating these functions themselves. Both
of the following results are standard integrals, which should be learnt.

[cot ade =In|sinz|+c

and
[cose de = In|cosec r — cotz| + c.

In the second of these, the simplest proof involves using the fact
that we know the correct answer. For the first, we shall write cot z =
COS £
, and use a Change of Variable. We shall put u = sinz, so that
sin xz
du = cosxdz. Then

peotede = [ ae
sin z
du
UL
= In|u| +c
= In|sinz|+c.
We also have
(cosec z — cot x) cosec r
cosecrdz = | ~—————_——_—_—_ dx
cosec xz — cot x
_ [ (cosec* x — cosec z cot x) d
-. cosec z — cot z 7
_ fadu
bP apa
(if we let u = cosec x — cot x giving du = (cosec 2? x — cosec zcot z) dr)
=In|u] +c
= In| cosec rz— cotz| +c.

Example 9. Find Jcosect z cot? x dz.

Solution. In this example, we will need to make the change of

variable u = cot z, from which du = — cosec? z dz. We need to use the
formula cosec? r = 1 + cot? z to express one of the cosec? z terms in
terms of cot r. Then

Jcoseet z cot? x dz = ik + cot? x) cot? x cosec? x dx

=- fQtw)urdu
 126 Ch: 15

aN
Il | nN
ind ——Q.=
—< +

a
= Ba anes Uu +c

!
=--C ot? x = cs pti y
SS ds 5
3 5

Example 10. Find Jcosectecot x dr.

Solution. In this example, we shall substitute u = cosec z, since

we have an odd power of both cosecz and cotz. Then we know that
du = —cosecz cot x dz.

/cosec® x cot x dz = — /cosec* x (— cosec x cot x dz)

== [uta

Lhe
ps ete

zs ab cosec” z+.

EXAMPLES 15
1. By putting u = sinz or u = cosa, find each of the following:

(a) ih
sin’ x cos x dz; (b) fsin® x cos x dz;

(c) [oot x cos x dz; (d) ih

sin‘ x cos® z dz;

(e) [sx x cos? x dz; (f) | /sin z cos® zx dz;

(g) [sx zcos? 2 dz; (h) /sin’ z cos? z dr.

2. By using the formulae

i 1
sin? z = 5(1 —cos2r) and cos*z = g(t + cos2zr),

find each of the following:

(a) [sn 2 dn; (b) [<* Dar:

(c) [sn' fae (d) icos‘ x dz;

(e) [sn zt cos? x dz; (f) [si’ x dz.

 Ch: 15 127

3. Find each of the following:

(a) [se z tan? z dz; (b) [se ztan? 2 dz;

(c) [sc ztan? z dz; (d) [vs! ga bi

(e) /sec x tan® z dz; (f) isec x tan’ z dz;

(g) [sc ztan® z dz; (h) [sx ztan’ z dz;

(i) if
sec z tan? z dz; (j) Jcoseet z cot x dz;

(k) Jcosec?x cot’ x dz; (1) [cot x dz;

(m) /cosec z cot? x dz; (n) Jcoseet z cot? z dz.

 128

CHAPTER 16
Trigonometric Substitutions
There are certain types of integrals which have substitutions which
lead to the trigonometric integrals we studied in the last chapter. These
integrals are those which involve either Va? — r?, a? + x? or Vz? — a?.
For integrands which involve a? — z?, where a is a positive con-
stant, we put z = asin @.
For integrands which involve a? + x?, where a is a positive con-
stant, we put z = atand@.
For integrands which involve x? — a?, where a is a positive con-
stant, we put z = asec 0.
In each case, we assume that the form of the integrand does not
lead us to make a straightforward substitution of the form u = a? — 2?,
for example.
In some cases, it may be necessary to complete the square in z
and make a linear change of variable before we are able to reduce the
integrand to one of the above forms.
dr
Example 1. Find | ———<—.
E my ls — x?
Solution. Since this integrand involves 1 — x?, and there is no
obvious change of variable, we put r = sin @ so that dr = cos @d6. Notice
also that then 1 — x? = 1 — sin? 6 = cos? 6, so that V/1 — 2? = cos8.

/ dz =| cos 6 d6
zVv1— 22 ~ J sin@
x cos6
=} dé
~ J sind

= [cosec do

= In| cosec
@ — cot 6|+c
1 vVl—<z?
= In|-— —-———|+c
x x

1-—vVJl-—-7z?
= In +c.
x

Notice how in this example we

use a right-angled triangle in order I
to express our answer in terms of the
original variable z instead of the 6
we have introduced. If @ is the an- i
gle shown then we may take the side aa
 Ch: 16 129

opposite 9 to have length z while the hypoteneuse has length 1. By

Pythagoras’ theorem, the third side of the triangle has length /1 — z?.

Example 2. Find [l— oF

Vax? — 6r + 25
Solution. In this example, we start by completing the square in
x? — 6x + 25, so as to reduce the integrand to one of the above types.
We have

2? — 62 + 25 = (x — 3)? +16 = (x — 3)? + 4’,

so that

lee xz? —6z + 25 ITS

(x — 3)? + 42
(putting u = z — 3, so that du = dz)
u
“i|Vut +42
To deal with this integral, we put u = 4tan0, so that du = 4sec? 0 d6.
We also see that

u? + 47 = 16tan? 6+ 16
= 16 (tan? 6 + 1)
= 16sec? 6.
From this, we see that
(i sees sec? 6 dé
Vu? + 4? V 16 sec? 6
= / 4 sec? 6 dO
= 4sec6

= [sec9a8

= In|sec@
+ tan 6| + c*
Vu2+42 u
=l1n
ees +c*

(using a suitable right-angled triangle to find sec @ in terms of u)

V(r —-3)?+47 «2-3 i
n ri ae mi +c

(replacing u by z — 3)
=| (c—-3?+4+2-3 c*

207 1¢%
aided
"hood sivna -
 130 Ch: 16

= In|V2? — 62 +25 +2—3/—Ind te"

(using one of the laws of logarithms)

= In| a — 62+ 25+2-3| +0,

where at the end we are combining the arbitrary constant c* and the
number — In 4 to give a new arbitrary constant c.
Notice again in this example, we
use a right-angled triangle to enable
us to express our answer in terms of
the original variable x. In this trian-
gle, one of the angles is 0, and we may
assume that since u = 4tan@ the
sides of the triangle are u and 4. The w=
7 —3

hypotenuse of the triangle will have

length Vu? + 42, by Pythagoras’ the-
orem. Since we took u = x — 3, we
may therefore find sec@ and tan@ in
terms of x, as we did.

Example 3. Find /V 12-42 — x? dz.

Solution. As in the last example, we start the solution by

completing the square and obtain

12 — 42 — 2? = 47 — (2 + 2)?.
We will therefore make the linear change of variable z + 2 = 4sin6, so
that dz = 4cos@d6. Also

4? _ (x +2)? = 4? — 4? sin?6
= 4? (1 _ sin’ 6)
= 4? cos? 0.

It follows therefore that

[Vive as =) — («c+ 2)? dr

= [ve cos? @ x 4cos

6 d@

=e /2 cos? 6d

=8 [(1 + cos 20) dé

(on using the double angle formula 2 cos? z = cos2z + 1)
 Ch: 16 131

=8(04 5sin 26)+c

= 8(0+sin9cos@)
+c
(on using sin 26 = 2 sin
@cos @)
2
= Bein (42) 432%? 1- (24?) +c

(on substituting for 6 in terms of z)

genie A
=8sin (2
24?) 4 2(¢ + 2) a te

2
= 8sin™ (ee
aa )+5
~(4#
+ 2)V12 — 42 — 2? +c.

This example is again a rather

complicated one, involving as it does
the use of trigonometric formulae to
convert the integrand to a convenient
form, and again to turn the answer
back in terms of z. There is also 5 are
some manipulation of fractions inside
a square-root sign, and finally the use
of
ieee £ +2 /6\ a

sin 7) Wak Jin—4e—2?

This of course arises from the fact that we have put +2 = 4sin 6,
2
from which we have sin@ = ud , so that the above formula gives 0.
(We are assuming implicitly here that the value of 6 is chosen to lie in the
first or second quadrants, to make this conclusion correct. Since there
is sufficient difficulty in examples of the above type, we have chosen
to sacrifice strict accuracy in favour of a slightly simplified approach,
which will not lead to incorrect conclusions.)
The above example does bring out one important fact about Math-
ematics, and that is that it is better to tackle problems by doing a large
number of small steps accurately, rather than by trying to do a small
number of large steps, which almost certainly lead to mistakes.
 132 Ch: 16

EXAMPLES 16
1. By making an appropriate trigonometric substitution, find each
of the following:
2
@ {—“..
(9 — 2)! oo) |an
(4-22)
(c) i V 25 — x? dz; (d) fiVv 36 — x? dz;
2
(e) {/— (f) ee eee5 dz;
(9+ 82 — x”)? (5 — 4x — x?)?

® {3 dx
(x? + 9)
a) f=.
(x?
dz
— 22 +5)?
Delt pein |x + V2? + kl, where k 0 is a constant, show that

dy = —. Deduce that
dr z7+k
dx
a alae + Va? +8] +0,
=
where k is any non-zero constant, and c is the constant of integration.
Hence find

@ [= o) fa
(c) |< (d) learnt
(22 40. oO)2 (x? — 6x + 3)?
Also use the above result to obtain a shorter solution for Example
1 in the text, and for 1(h) above.
 133

CHAP TERA 7.
Partial Fractions
Given a rational function—that is, a quotient of two polynomials
having no common factor—such as
224 + 2° — x? — 152
+ 62
r34+2r?-3r-10 ’
where there is no obvious relationship between the denominator and the
numerator, then the usual way to proceed, if asked to integrate such an
expression, is to use the method of Partial Fractions.
We can think of Partial Fractions as providing a technique for re-
versing the results of expressing a sum of fractions over a common de-
nominator, as for example in
5 Ze O(a i2) a oe) gar 16
PES eo G3)+ 2) ee
8 (a 2)
The method of Partial Fractions only applies to proper fractions—
that is, those for which the degree of the numerator is less than the
degree of the denominator. If this is not the case, then we need to use
long division of polynomials in order to express the rational function
as a quotient term together with a term which is a proper fraction.
The technique of long division of polynomials is explained in detail
in the chapter on basic techniques. Here, we shall content ourselves
with an example to practise the technique. Recall that at all stages the
important thing is to keep the highest power of z correct.
Example 1. Find the quotient and remainder when the expres-
sion z+ — 3x3 — 7x + 6 is divided by x? — 52 + 4.
Solution. In setting up the traditional long division format,
notice that we keep terms involving the same power of x under one
another, and notice that we insert a term + Oz? in the dividend, in
order to create a column for the x*-terms. We always do this if there is
a term missing from the dividend.
a? + Qe + 6
2? — 5c + 4| ot — 323 + Oc? — Tr + 6
g**— 52° 4) 4?
22° — 427 — Tr
22?) Se l0z hake Sz
62? — 152 + 6
6z? — 30r + 24
liz — 18
Hence the quotient is 2? + 22 +6, while the remainder is 15z — 18.
It follows that
z* — 32° — 72 +6 = (x? — 52 +4) (x? + 27 + 6) + (152 — 18),
 134 Chil7

and so, on dividing by x? — 5z + 4, we have

x4 — 32? —72+6 2 15z
— 18
Pi ne ee StS tee Bn D ———_—_____—_..
z?—57r4+4 aera Te eae ee
Using the above method, we can express an improper fraction—that
is, one where the degree of the numerator is at least as large as the
degree of the denominator—as a sum of a polynomial together with a
proper fraction.

The next step in the method of Partial Fractions is to factorise the

denominator completely, as a product of linear factors and quadratic fac-
tors, having real numbers as coefficients, such that the quadratic terms
cannot be factorised further using real numbers as coefficients. In doing
this, the Remainder Theorem can help us to determine a linear factor,
and synthetic division (or long division of polynomials) can then help
us to determine the remaining term which still needs to be factorised.
The methods of the following example should all be familiar from
the earlier examples we did in the chapters on Curve Sketching, and are
included here purely for revision purposes.

Example 2. Factorise x* + 423 — 8x? — 48x + 80 as a product of

linear and quadratic factors.
Solution. Let f(z) = x4 + 423 — 82? — 482 + 80. Then

f(0)
=80£0,
f(1) =1+4-8- 48+ 80 = 29 £0,
f(-1) = 1-4-8+:48 + 80 = 117 £0,
f(2) = 16 + 32 — 32 — 96 + 80 = 0.
Hence z = 2 is a root of f(z) = 0 and so z — 2 is a factor of f(z).
We shall use long division to obtain the other factor, although we could
equally well use synthetic division.
ge? +464 4 Ap 3 40
Ti 02 yee dg? (87?) =. 487 7 80
re 4 = 2°
62° — 82?
6a° ge tor”
42? — 48¢
477 — 8e
— 402 + 80
— 40z + 80
0
Hence we have f(x) = (x — 2) (x* + 62? + 42 — 40). We shall let
g(x) = xz? + 62? + 4x — 40.
 Chithi 135

There is no point in trying z = 0, r = 1 or r = —1 as roots of

g(x) = 0, since these values of x were not roots of f(x) = 0. So the first
value of z that we try is z = 2 again. Now
g(2) =8 + 2448-40
=0.
Hence z — 2 is a factor of g(x). We use long division again to find the
other factor.
z? + 8r + 20
z — 2/29 + 6x? + 4x — 40
L 3 oh Dd
822 + 4r
82? — 16z
20x — 40
20x — 40
0

It follows that g(x) = (2 — 2) (x? + 82 + 20). Since x? + 82 + 20

cannot be factorised any further, (its discriminant is —16 < 0), we see
that
xz! + 4% — 82? — 4827 + 80 = (a — 2)’ (x? + 8a + 20).
We now wish to try to break proper fractions of polynomials into a
sum of terms, each having a single term in its denominator. There are
a number of different cases to consider.
(24) The denominator has distinct linear factors.
In this case the fraction can be expressed as a sum of terms, each of
which is a constant divided by one of the terms from the denominator.
The values of the constants can most readily be found by using the
cover up rule, as in the following example.
Example 3. Find partial fractions for
z? +82 —5
(x — 2)(x + 1)(x2 + 3)

Solution. We assume that we can write

seats ty WRC
(cx-—2)(x+1)(x
+3)
2x-2
i 2x43
r+1l
Gre
To find the value of A, we put z = 2 in the left-hand side of the equation,
‘covering up’ the (x — 2)-factor. So we have to find the value at z = 2
of
2? +82 —5
Ms:
+ 1)(2 + 3)
Then
PP 8902 = 5 §8 15
ib
* 2+ 1+ 3) 23x 5
 136 Chilly

Similarly, we ‘cover up’ the term z + 1 on the left-hand side of the

equation, and put z = —1 in the remainder. This will give us the value
of B. We have
=a
= £23 S207
We can find C in a corresponding manner. This time we obtain

C a —2.
(—5) x (—2)
From this we conclude that
rz? +8r—5 ag: 2 2
GU poe Te ears:
You might be interested to express the right-hand side of this equation
over a common denominator, and show that it is indeed equal to the
left-hand side.
(2) The denominator has a repeated linear factor.
In this case we can express the rational function as a sum of terms,
each of the form of a constant over each non-repeating linear factor,
together with a sum of terms of the form of a constant divided by each
power in turn of the repeated factor in the denominator from 1 up to
the highest power occurring in the denominator.
We determine the constants which are divided by the highest powers
of the linear terms as in (1) above. To find the other terms, we have to
cross-multiply by the denominator, and then either equate the various
coefficients of x involved, or substitute different values of z into the
equation we obtain. The method will become clear once the following
example has been studied.
Example 4. Find partial fractions for
zr +62? 411
(x —1)' (2 +2)
Solution. We assume that the above expression has a partial
fraction decomposition
a? + 627 +11 eee B C D
2G.)
Gat. +Gn
Using the cover up rule, taking x = 1 and covering up the (x — 1)°
term on the left hand side, we can find that

A=>=
We can also use the cover up rule to find D, as we did in the previous
example. We find
ont 27
D = =o = =e
(—3)*) (=27
 Chi] 137

We now multiply both sides of the above equation by the denominator

of the left-hand side, as we do so substituting the values for A and D
that we have just found. We have

(x3 + 62? + 11)(x — 1)° (x + 2) _ (2-1) (2 $2)

(2 —1)° (4 +2) (=a)
, Bz = 1) (@ +2) C(z-1)*(r+2) (e-1)* (+2)
(oa? 1)i r—1 r+2 /
and, on cancelling terms from the numerator and denominator,

x? + 62? +11 = 6(z +2)

+ B(x — 1)(x +2)
Cty (oc 2) —(2= 1).
There are two possible ways of finding the remaining constants, B and
C. We could multiply out each of the terms and collect up the terms in
each power of x. Then we could equate the powers of xz on both sides
of the equation. This is the method we shall use in the next example.
Here, we shall substitute values for z and obtain a pair of equations
from which we can find B and C’. The values z = 1 and x = —2 have
already been used in the cover up rule, so we shall try different values.
We shall take x = 0. Then we have
11 =12-—-2B+42C
+1,
and, taking z = —1, we have

16=6-2B+4C+8.
Rearranging these equations, we have
—-B+ C=-1,
= tC =" 1,
from which we see that C = 2 and B = 3.
We therefore conclude that
z* +62? +11 6 3 2 ih
@2iy Gh) @eaiy usCee Pes
Seo
As with the last example, you may like to check that this is correct
by putting the right-hand side of the above equation over its common
denominator. |
You may be wondering why we chose the values z = 0 and z = —1
to substitute. The answer was simply to make the calculations simple.
z = 0 is usually a good value to substitute, and after that we wish to
choose the smallest values of z that we can. However the answers we
obtain would not have altered had we chosen alternative values for z.
 138 Chrel7

(3) The denominator has an irreducible quadratic factor.

In this case we deal with any linear terms as in (1) and (2) above.
All (non-repeating) irreducible quadratic factors in the denominator
contribute terms of the form
linear term
irreducible quadratic
to the sum, where the linear terms will have the form Az + B, with A
and B being constants which we are to find by the methods described
in (2) above.
Example 5. Find partial fractions for
I 107+3
(x? +5) (x — 2)

Solution. We notice that 2? +5 is an irreducible quadratic term,

and so we have to use the method described above. Hence we assume
that
227—-10r+3 _Ar+B Me C
(x? +5) (@—2) -2? +5 — 2 —=2
We can find C by putting z = 2 and using the cover up rule. So we
obtain
dpe ANE SA =
C = ———__ = — = -1.
4+5 9
Proceeding as in the previous example, we multiply both sides by the
expression (z? + 5) (z — 2), and substitute C = —1. So we have
22? — 102 +3 = (Ar + B) x (x — 2) + (—1) x (2? +5)
or

Qn? — 102 +3 = Az? + Br —2Azr — 2B — 2x? —5,

which reduces to

2x? — 102 +3=(A—1)x?+(B-2A)z — (2B +5).

If now we compare the coefficients of the various powers of z in the
above identity, we can see that we must have 2= A—1, -10= B—2A
and 3 = —(2B +5). From this we can obtain, on solving the equations,
A=3and B= —4.
Notice that we only needed to compare the coefficients of two of
the powers of z in order to find the two constants, A and B.
We conclude from the above that
22? — 102 +3 _ 32-4 1
(224+5)(2-2) 22+5 2-2
The reason that we are allowed to equate the coefficients in the
above solution is that we are trying to find values of A, B and C for
 Ch: 17 139

which the equation holds identically. In other words we want the

equation to be valid whatever value we care to substitute for x. If this
is to be the case, then it can be shown that all corresponding coefficients
on the two sides of the equation have to be equal.
(4) The denominator has a repeated irreducible quadratic factor.
A term in the denominator such as fe | would contribute
terms of the form
Aria Ge libations Fh
(x2 +2)? (a? 42)? 2? +2
to the partial fraction decomposition. All other terms are dealt with
as described in (1)-(3) above. The constants, A down to F, are then
found using the methods of cross-multiplication and either equating co-
efficients or substituting values of z, as in (1)-(3) above.
Since there is no real extension of our knowledge to be gained by
going through the (usually) complicated details of examples of this type,
we shall not do such an example.

Our main use for the method of Partial Fractions is as a way of

breaking up a rational function into a sum of terms each of which we
can integrate. Most of the terms we have obtained in the above can
be integrated quite easily. However, there is one slightly more difficult
case. This is the case involving a term such as
52 +8
g2+4r4+8
The method for integrating terms such as this one is as follows.

Example 6. Find
/ ox + 8
iE
z7+47+4+8

Solution. Let

fast /ait SS
2*7+4r+8
We start the solution by completing the square in the denominator, and
then making a suitable linear change of variable. We easily see that

a? 4+42+4+8=(rc
+2)? +4,
and so we put u = z+2, from which du = dz and also z = u—2. Hence
our integral becomes

=
5(u —2)+8
| —————-du=
(fee du.
: ii u2+4 a ay
 140 Ch: 17

To deal with this integral, we break it into two parts, by dividing each
of the two terms from the numerator by the denominator. So we have
du —2 5u du

aii hai is of a standard type leading to a tan~!

The final one of these
term. To deal with the first integral, we shall make a second change
of variable. We notice that the numerator is a constant multiple of
the derivative of the denominator, which suggests that we should put
v = u* +4, which means that dv = 2udu, or udu = 5 dv. We have

sofa tfae
= 2Info|—2.5 tan ve

= 5Inju? +4) — tan? S +e

9) 2
= 5 In|z? +42 +8 tan"? == +.

We now combine together most of the results of this section in the

following quite complicated example. As we have seen with earlier ex-
amples, however, it is not difficult to do such examples if we concentrate
on doing one small step at a time.
Example 7. Find
/ et stag Ae age 4
SS Hfbe

x? — 7x? + 162 — 10

Solution. Since the degree 4 of the numerator is greater than the

degree 3 of the denominator, we must start by long division to reduce
the rational function to the format to which we can apply the Partial
Fraction method.
z+ 3
a? — 7x? + 162 — 10 | c* — 42% | '027"+ 1dr pes
che To? +) 1622 9—e10r
3x3 — 16x? + 24¢ + 4
32° — 212? + 48r — 30
52? — 247 + 34
It follows from this that
z4— 47°34 147+4 52? — 242 + 34
is maar oo Gl RUMOR dee +3 4+ ———______..
xz? — 7x? + 162 — 10 xr3 — 7x2 + 162 — 10
Our first priority is therefore to find partial fractions for
52? — 242 + 34
z3 — 7x2 +162 —10
 Ch: tt 141

Before we can do this, we need to factorise the denominator. So we

let f(z) = x? — 7x? + 162 — 10. Now
f(0) = -10 #0,
f(1) = 0.
It follows that x — 1 is a factor of x3 — 7x? + 16z — 10.
We shall use synthetic division to find the other factor. (Note that
there is no deep significance in this choice—we merely chose to do this
to add a bit of variety to the proceedings.)
lL =—7 16 ,=10
La. (80" =.leet6an 10
1 -6 10 0
Then, from the synthetic division, we have

f(z) = (2-1) (2? —6z +10).

Since r* — 62 + 10 has a negative discriminant, it follows that it
cannot be factorised further into real factors.
We are now ready to find the partial fractions. We shall assume
that
52? — 247 + 34 A Br+C
(x — 1) (2? — 62 + 10) + Gra es Seay
By the cover up rule, we see that
Pb 24-+34 5 15 =
“16+ 10 6. 5 ae
Multiplying the above equation by (x — 1) (2? — 6z + 10), and sub-
stituting A = 3, we see that

5a? — 242 + 34 = 3(2? — 62 +10) + (Bx +C)(z — 1).

On multiplying out the right-hand side of this equation, and equat-
ing coefficients, we deduce that
5=3+4+B,
SPS awe
34 = 30-C.
From the first and third of these equations, we see that B = 2 and
C = -4.
We now need to find
34 3 fi 2x —4 ae
f{e+ 71 Wereoe10
2x —4
= 527 +324 3In[2—1) + / :
eee tie
 142 Ch: 17

Let 9 4
5) xe SS Dp
C=

aes
We shall use the method of Example 6 to find J. Notice firstly that
x? —6r2+10 = (x —- 3a +1. We therefore put u = x—3, so that du = dz
and also z = u+ 3. On substituting, we have
we Qc —4 Me
TP? BGS)
ae eean
i u2+1
D) y)
= fae
u2+1
2u du
= d vi
baa Bis loa
x — +2 f du
it Vv u2+1
(on taking v = u? + 1, so that dv = 2u du, in the first integral)
=In|v] + 2tan*u+c
=In|u? +1/+2tan tute
= In |z* — 62 + 10| + 2tan7!(a —3) +.
From this we can finally deduce that
/ ge 4g age 1
——_——. dx = =? + 324 3ln|zr-1
P=.ibe No a oe
+I1n |x? —62 + 10) +2tan7!(z —3) +c
As a far simpler example of the method of Partial Fractions in
action, consider the following, where a is any non-zero constant.
Example 8. Find

Solution. | We can factorise x? — a? as (x — a)(z +a), and so we

have, on using on fractions,

(are “| eiera
II /<= att fore

-/= 1 dr
~ Qa r-a 42a r+a
(using the cover up rule to determine the constants A and B)
 Ch: 17 143

1 1
= 5,lnle—al—5 nit +al+e
n
= 5, lin|e — a] —In|z +l] +¢
cae! zr—a 1
caer
The above result is sometimes very useful, and so should be added to
the list of standard integrals to be learnt.
EXAMPLES 17
1. Find the quotient and remainder when:
(a) 2°4+527+7x-—3 isdivided by zx? +32 —-1;
(b) 2444273 4+4274+2-—5 isdivided by 2?+32 +4;
(c) 2t—23+452 isdivided by 2? +2242.
2. Express each of the following as a product of real linear and
irreducible real quadratic factors:
(a) «rt — 32° — 7x? + 272 — 18;
(b) ct —82? + 8x +15;
(c) 2«t —52° 492? —- 8244.
3. Find partial fractions for each of the following:
3x2 +9 1227? — 18
(a) (x —1)(x +1)(z + 2)’ 2) (x poe +1)(x + 3)’
a? +2044 227° —6z —4 }
ON) esas ES eR
2 le 4r3 — 162? + 1627 + ae

Giy Gnu (x-1)'(e-3) ©

(g) 24 —2 ; (h) Ar? + Tr :
6) (a +2)(227 +9)’ (x —2)(x? +6)’
22° — 2? — 34-1 132 + 16
Soe IR PS a eae eaeracers
(z — 1)’ (2? +2) (z+ 2)° (2? +1)
4. By completing the square in the denominator, and making a
suitable linear change of variable, find each of:
224+3
UT oeoc b panez+1
eS
(a) les (b) /=iSs q
 144 Ch: 17

5. Use partial fractions to find each of the following:

zr—4 ap
OM emaisr are I iNeraices; De
(c) ifzo — a a +2 eo (d) i:
zg aso 22 ig

baer ile arfea ae

8 122? ,

2r +12 15¢ — 2?
(8) dee ea
aa (h) (2 + 3) (2?249) 0%
0 geagpen fins
sSCrSA
obAETEl
(i leaeaes® Of @—3)(e? +62 +10) °
6. By first dividing the numerator by the denominator, and then
using partial fractions, find each of the following:
(a) A cae (b) 3a? — 272 —4
—————— i} ———_——_- dr;
(x + 2)(2 —1) (x — 3)(x +1)
(c) iI x re
8x B aah ee (d) 4 3
z*+2°4+3
(2) (+1) (x —1)* (x? +4) :
 145

CHAPTER 18
Two Methods using Complex Numbers
In Chapter 15 we learnt a method which enabled us to integrate a
function, such as cos? z or sin? x cos? z, which involves only even powers
of sinz and cosz. We used the formulae
1
sin? z = 5 (1 — cos 2z)

and
1
cos? ¢ = 5 (1 + cos 2z)

to express the integrand in terms of sines and cosines of multiples of the

number z, which we could then integrate.
An alternative method is to use complex numbers as in the following
example.
Example 1. Use complex numbers to show that
; 4 i 1
sin?
zcos‘ z = —— cos6z — — cos4z + — cos2zr + —,
32 16 32 16

and hence find [se zcos' x dz.

Solution. Let
Z=e = "Cos 2+ isin7,
so that
—=e€e = cosx —isinz.

Adding these two equations, we see that

1
z+—=2cosz,
Fe,

while, on subtracting, we have

z—--—=2isingz.
Z
Furthermore, we have, for any integer n,
y= e Hd= COSN I ASIN NZ,
so that
1 1 -. a0
—
CAL
= —
etre
=e '"* =cosnz —isinne.

Adding these two equations, we see that

pees |
z" + — gn
=2cosnz;

while, on subtracting, we have

2" — == 2isinnz.
 146 Ch: 18

To avoid having a solution that is littered with fractions, we shall

consider first the expression (27 sin 2)*?(2 cos z)*, from which we can find
sin? z cos‘ z by division. We have

is in 2) ?( 2c os z) * = (:via (<4 1)
(2
( 1

(rd G3)
1 2
2-22? + — 4224-44 —[427-S4+a
z z z
1 2 3
ll 2° +224 — 2? -4- + —
z we

1 1, 1

(ct+ 2) #2 tes eee

= 2cos6z + 4cos4z — 2cos2zr — 4.

It follows that
(22)?2* sin? z cos* r = 2cos6z + 4cos 4x — 2cos2z — 4,
so that, on dividing by —2°, we have
ei ine 1 1 1
SImecicOSm re — 55 cose _ 57 Coase + ee + oy

= ——cos6zr — ei cos 42 + =a cos 2z + a

32 16 32 16:
as we had to show. It follows therefore that
cnet 4 1 1 if i
[x GCOS Paps / (—55c0s6e _ noo” + eee + a) dx

1 1 1 1
= ———sin6z — — sin4z + — sin2z + —r+c.
192 64 64 16
As a second example using complex numbers, consider the problem
of finding [ e°* sin3zdz. One possible way of finding this integral is
to use integration by parts twice over, thereby retrieving the original
integral, and obtaining an equation which can be solved to find the
integral. We shall use this method in Chapter 28.
As an alternative approach, we can use complex numbers. In par-
ticular, we use the fact that
1
[evra =—e™7 +¢
m
 Ch: 18 147

for any non-zero constant m, whether m is real or complex.

Example 2. Find

|e°* sin 32 dz.

Solution. Consider

dss cos 3z dz +2 /e°* sin 32 dz

= [e(cos3x + isindz)de

= feteets dz

saefaclet 30 de

1 ,
= (5+32)z
Parry
rF35° +c

1 Soe sta
5432 i ts
5 — 32
= Gasca 0 3x +isin3zr)+c

(multiplying top and bottom by 5 — 3)

3;
= ssz 507 (cos 32 +isin3zr) +c
52x
= ase — 31)(cos 3x + isin3z) +¢
eot
= 34 (5 cos 32 — 32 cos 3z + 52 sin 3z

+ 3sin 3x) + cy + 2c2

(writing the arbitrary constant in the form c; +72c2)
Sr
7 (5cos32 + 3sin3zr) + c
5zr
+2 (F (5sin3x2 — 3cos3z) + ca),

Hence, on equating real and imaginary parts of the above equation,

we have
ay 7

Nee cos 3x dz = Fy (500832 + 3sin3zr)+¢

and
5zr
Hite sin 32dz = ora sin 3z — 3cos3z) + c2.
 148 Chis

As you see, we obtain two results for the amount of work required
to obtain the one integral, and also the work involved in integrating
using this method is less than that involved in the method involving
integrating by parts twice. The difficulty comes from manipulating the
complex numbers involved.
EXAMPLES 18
1. Use complex numbers to find constants a, b, c such that
sin’ z = acos4z + bcos2z + ¢,

and hence find

[st dz.

2. Use complex numbers to find constants a, b, c such that

cos* = acos4z + bcos2zr +c,

and hence find

/cos‘ x dz.

3. Use complex numbers to find constants a, 6, c, d such that

sin‘ 4 x cos* «= acos6z + bcos 4z + ccos2z + d,

and hence find

Js! z cos’ x dz.

4. Use complex numbers to find each of:

(a) [a cos 42x dz; (b) bgt cos 22 dz;

(c) et sin 32 dz; (d) [etsina aa:

 149

CHAPTER 19
Tables of Integrals

We have so far dealt with a number of ways of tackling different

types of integrand. However we have not been able to give a method
which will enable us to tackle any integral which is set to us. One reason
for this is that there are integrals which cannot be solved.
One idea that we have so far not discussed is the idea of using a
formula which gives us the result of integrals of particular types. Lists
of such formulae are known as Tables of Integrals. The formulae are ob-
tained in various ways, frequently using Integration by Parts, or some-
times more advanced techniques which we have not discussed.
Many Calculus text books give extensive tables of integrals. Here
we shall content ourselves with using some of the better known of these
formulae. Such formulae can be readily checked by showing that the
derivative of the answer to the integral problem is equal to the given
integrand.
Consider firstly the formula

| 1 £
Gye
dz
pede Se
2(n nt 1)a? = a g2)""} +( n 3) |(a2 A aad ’

which holds provided a # 0 and n £ 0, 1. We can use this formula in

the following example.

Example 1. Find

where a is a non-zero constant.

Solution. We apply the formula given above, firstly with n = 3
and then with n = 2. We are then left with an integral of the same
form, but we cannot use the formula with n = 1 since this is a value of
n for which the formula is not valid.

1
eae Pilg, Ong? oar ‘Tam

= ai
ai
Sea
s
meiidallo
a
(a? —a
aa
 150 Ch: 19

1 a 3 a 3 Els x r—a ne
~ 4a? (a? — 2?) + ga4 a? — x? 8a4 2a |r+a

1 Ae 5 i 3 r—a
= — |——__| + — |——__ | - —=]1
4a? =| + 3a Ferd 16a5 2 zrt+a HG

The final integral here was obtained using the method of Partial
Fractions as we did in Example 8 in Chapter 17, for the almost identical
integral.
Now we consider the formula
1 ei ee Li), il . =—?2
[ost ade = ——sin"7! 2 cosx +:——_/ sin"~° zdz,
n n
which holds provided n # 0. (Of course if n = 0 the integral reduces to
f 1dz, for which no complicated formula is necessary.)
We use the above formula in the following example.
Example 2. Find f sin®
zdz.
Solution. We apply the formula given above, in the cases n = 6,
4 ana 2.

[so* xz dz

1
= —g sin? x cosx + >fsint x de

ae 5 Let :
= —g sin? ecosz + = [-Fsin’ vcose + $ sin? ede]

ie Dae 15
= —=sin” rcosz — — sin’
xcosz + — | sin? xdzr
6 24 24

= ii sin? x cos x7 — sin? zcosz

; + —
~ ! sinzcosz + =:
|—-sIi sin?
sin’ z cd dz
6 24 24| 2 D
“2 :
mate | np rn ge ae deg als rahe eae ldz
at GEG 24 48 48
5 3
= ——sin° rcosz — — sin xrcoszr— 1g sncoss + as +c.
6 24
In this example we have obtained an answer in terms of powers of
sinz. We have earlier seen methods we could use by which we could
solve the same problem by expressing the integrand in terms of cosines of
multiples of z, which we could then integrate. This method of solution
would yield an answer that would look different to the answer we have
obtained above. In fact though, by using our trigonometric formulae, we
could show that these two answers would be equivalent to one another.
 Ch: 19 151

The following formula is obtained from the standard trigonometric

formulae. It uses the formulae in a form that we have not encountered
them, although it is readily deduced from the addition formulae we have
met. We have

[sams sinnz dz = _sin(m + n)x sin(m —n)zx “tc,

2(m +n) 2(m — n)

provided that m? F n?.

Example 3. Find

/sin 62 sin 3z dz.

Solution. Using the formula given above with m = 6 and n = 3,

we have
sin9z sin3z
[sino sin 3t dt = — 18 6 +c.

Our final formula is valid provided m # —1 and n # 0. It is

established by integration by parts. The formula is
wauaamial Ale:qe n
fem (In z) n eel
=
eee ar <s
fs m
oe) n—1
Laz.

Example 4. Find

[ita z)*dz.

Solution. Using the above formula with m = 3 and with n = 2,

and then with n = 1, we have

3 3 ae 2 3
jtone)ar = PUONEe ~ ¢ f 8dnz)'de
2 D
2 5 4 3
= ze? (Inz)’ — ; |otineds

Dae
= <2?(Inz)*
— = = em
4 | 22(Inz)! ~ = f ehdn2)? de
5 5 : =
aes fl ED 2 a
os ze? (In2)’ = 5 [eet ne - =f ar|

Ano 4
= eee ae [eetine - rad +c

| Q vin
5 8
no
i 2? Ing + 28, oh +c
—a
SE) bo
20
| 125
a
 152 Ch: 19

Several of the formulae we met above expressed an integral in terms

of a simpler integral of the same type. Such formulae are known as
reduction formulae or recurrence relations.
EXAMPLES 19
1. Use the table of integrals given below to find the following:

(a) [re (b) lo"

(c) Joes — x7)? de; (d) [sx 2x dz;

(e) [cost 3x dz; (f) [se 22 dz;

(g) i,7z cos 3x dz; (h) tleIn z dz;

(OEE
=
(i) [os (j) /e°” sin 2z dz;

(k) faa
sin 22 dr; (1) [eot* 3z dz;

: 3 /

TABLE OF INTEGRALS

In this table, it is assumed that a is a non-zero constant, and that

b is also a constant.

|
; sin axcosar n-l1 sie ee
[sis? az de = -7 SE [ow 2ardr (n>1)
na n

cos""larsinaz n-1
Jcos” az dx = ————_———__ [cos aridz A(t >:1)
na n

ms sec n—2 aaxtanar n—2

sec” az dz = pies Ener
TR Sates
SD chats ale tla
Sao |sec n—-2
azdz4(n >.1)

t72—-1
[eo Ge dgs —— S peor? arariGy 1)
a(n — 1)

n
; zr n &
[ #tsinarde = = cosar + ® fo" ‘cosardr (n #0)
a a

agn
n we Be
[2 cosar dz = = sinaz~ > [2° ‘sinardzr (n #0)
a a
 Ch: 19 153

jfsimaz.cosbe de = -ST OS - SBE tc (a? # b’)

peel eat
fe sin br dr = 2 ae (asin br — bcos br) + ¢

grhtl ght
jfeinac de = —— Inaz — ——~ > +c (n#-1)
n+1 (n+ 1)?

5 ae Seay gmae
[ Ay ae =- ant 2 - te (a > 0)
£ a =

1 1, ja+Va? +2?
—_——— _dr = --ln +
evar +? a c

[A
Py
a
pa
a= Vir-anf OE IFe
\ /aa ae pe

ic _ 2r"(ax +b)?
He Vazr+bdz= a(on +3)

2nb 4
-aaee |? VvVar+bdzr (n>0)

/ Jaren
——
+b
dr =
Qc" JVar+b
—c———
a(2n+1)
— ———
a(2n+1)/
2nb
eeee
Var+b
50)

/ dz = Var +b
a*far+6 o(n—1)z"-!
a(2n — 3) dz
rat pene peng: (On 150 0)

2 azr+b
— tan7!,/———_+c, 65<0
i dz Aye —b
sJar+b |) 1, |vart+b-vbl, 4g
Vb Jar +6+ Vb ;

z .
KG - 22)? ar= -5 (22? — 5a”) Va? — oe sin~* =e (a > 0)
 154

CHAPTER 20
The Definite Integral
So far we have concentrated on developing methods for tackling
different types of integral. Before seeing how to apply our knowledge of
integration to various practical problems, we need to introduce a further
piece of notation—we need to define the definite integral.
If f(x) is any real function of z, having F(x) as an indefinite inte-
gral, we define the definite integral of f(z) from + =a to r = 8,
which is denoted by
b
/ f @idz,

to be
b
/ f(x)dz = F(b) — F(a).

It is convenient to denote F(b)— F(a) by [F(z I, so that

b

/ f(x)dx = [F(2)|’ = F(b) - F(a).

We call a and b the limits of integration, where a is the lower limit
and 6 is the upper limit.
We only make the definition given above when f(z) is a continuous
function for all z between z = a and z = b. For example, we do not
define

il
since the function — 5 becomes infinite at z = 0, which lies in the interval
[—1, 1], which is ah range of integration.
Similarly, we would not define
z
x

/ sec’ dz,
0
since
1
sec? z=
cos? x’
and so sec* x tends to infinity as x tends to 2/2, since cosz is zero at
Tie
F(z) can be any indefinite integral of f(z), and so we may choose
F(z) to have 0 as its arbitrary constant. If we choose not to do this,
the arbitrary constants would just cancel out anyway.
For example, consider
x
2 it
ii cosz dz = [sing +c]? = (sin= + c)_ (sin0+<) = 14+c—0-—c=1,
0 =~
 Ch: 20 155

which does not involve c at all.

When we change the variable in the course of finding a definite
integral, the normal procedure is to change the values of the limits to
the values of the new variable which correspond to the values of the
original variable at the limits. Then we substitute these new values into
the indefinite integral without returning to the original variable.

Example 1. Evaluate

od 3
i 2° (1 + ie) az’.
0

Solution. We notice that the integrand is in two parts, one

of which is a power of 1 + z*, the other being a constant times the
derivative of 1+ 24. This suggests that we should use a substitution in
order to find the integral. Specifically, we will put u = 1+ 24, so that
du = 4r°dz, from which we have r3dz = idu. We also calculate that
when z = 0 then u = 1, and when xz = 2, then u = 1+ 24 = 17. Hence
we have

1
== |289v17 — 1].
The following facts concerning the evaluation of definite integrals
can all be easily checked from the definition. In these formulae, a, },
c and k are all constants, and f(z) is any function having F(z) as an
indefinite integral.
b b
( ia(9) gi k | f(a)dz,

since both of these equal kF(b) —kF (a). This tells us that we may take
a constant multiplying a definite integral outside the definite integral,
just as we could for indefinite integrals. Examination of the solution of
the last example shows that we actually used this fact in the course of
that solution.
b a

i,f(a) ie
de / f(a) ride,
 156 Ch: 20

since both of these are equal to F(b) — F(a). We can think of this
as telling us that interchanging the order of the limits of an integral
corresponds to multiplying the integral by —1.

[ reac [ eye = [sera

since both of these are equal to F(c)—F(a). This tells us that integrating
from a to b and then from b to c is the same as integrating from a to c
in one step.

[ sa)ae =o,
since the integral equals F(a) — F(a) = 0.
The other rule that we need is the rule relating to Integration by
Parts, where we are dealing with a definite integral. This rule is

|tae ital |ae:

where, as before, g denotes any indefinite integral of g', so that we can
choose g to have arbitrary constant zero.

Example 2. Evaluate
2
i)zetdrz.
0

Solution. The integrand is a product for which there is no

obvious connection between the two parts of the integrand. It seems that
Integration by Parts is the appropriate method. We shall let g’ = e7,
so that g = e” also, while f = 2? so that f’ = 2x. Hence we have
2 2
| ze dr = [x?e7]> -{f e* 2x dz.
0 0
To integrate this last term, we use integration by parts again, this time
taking g’ = e* so that g = e*, while f = z so that f' = 1. So we have
2 2
i ret dr = [x?e*] >-f e* 2rdz
0 0
2

= [ze*] —2 ([xerl -[ ede)

0

= Qtet = 026-12 (2c? — 0e° — (es)

= 2267 = 072" 2 (Oe? = 0c" een)
ede” — Ae? 4 De? 29
= Qe? — 2.
 Ch: 20 157

In the following examples, we shall solve a number of different types

of integration problems. One of the most difficult steps when confronted
with an integral is to recognise which is the appropriate method to apply.
Example 3. Evaluate

5 sin” 2 x cos
z dz.
=
x

Solution. Since this integral involves an odd power of cosz

together with a power of sinz, the method will be to put u = sinz, so
that du = cosrdr. When z = 7, u = sin% = 3}, while when z = ,
u=sin= = oe So we have
x —.

Saree
sin’ x cosz
dz = oaks sin”
FF) x cosz dz
x
— zr
6

24
Example 4. Find
1
i ztan~)2dz.
0

Solution. The integrand here is a product of two terms which

have no apparent connection. This suggests that Integration by Parts
might be the appropriate method. Since integrating tan~!z is some-
thing we would rather avoid, we shall let g’ = z, so that g = $2, and we
shall take f = tan7! z, from which f' = i “ 5: Applying the formula
x
for Integration by Parts, we have
1 1 : a a 1
fiztan !2dz = 52”tan? | — | (52 x :)dz.
0 2 0 oO Ne l+zcz
To deal with the integration of x?/(1+ 27), the best method is to write
the numerator as (1 + z”) — 1, and then divide each of these two terms
 158 Ch: 20

by the denominator. This puts the integrand into the form of a sum of
two terms, each of which is easily integrated. So we have

1 Soucek ae |zx” tan |; = il= FR PS 1 2 zr

/ z tan

i : i! pene
= Featan; 7 aw:
0
/ 1+ 2?

= 52?tan “|— xf (1— a dr

1
=
1
Fa tan“ s| — L
2 0 2
1
1
52?tata 4 = .
2 0 Z
1 1 il < =
= 51’ tan“! 1— 50° tan? 0-5 (1-0) +5 (tan 11—tan'0)
Belen, eee
Se Se
_t bigs
78 9 yes
Bes 1
oe
Example 5. Evaluate
1
/gee eM be
0 V 1- x4 ;

Solution. In this integral, our first reaction is that we are not

sure which method is appropriate. However, if we notice that z is related
to the derivative of z?, and that z* can be written as (x?)’, then we
have a possible approach. For we can then make a change of variable,
putting u = 27, so that du = 2x dz, and rdz = idu. When we make
this change of variable, we see that u = }2 corresponds to z = von while
u = 0 corresponds to x = 0. Hence we have
pe x dr a xzdr

Nle
Mle
Nl
 Ch: 20 159

In our integrations so far, when we have made a change of variable, we

have included the variables z and u with the limits to make it clearer
what we are doing. While this is sensible in complicated examples, it
is not normal notation. We shall in future use the normal notation,
changing the limits as we change the variable, but without putting the
names of the variables in the limits.
Example 6. Evaluate
[ dx
_2 27+42413°

Solution. We shall start by completing the square and making

a linear change of variable. Now
a? +42+4+13 =(r+2) +9.
So we shall put u = 2+2 so that du = dr. When xz = —2 then u = 0,
while when xz = 1 we have u = 3. Hence we have
iL dx -[ dz
9? +4r4+13 J_y (x42)?+9
-[ du
as 0 u2 + 3?
ul?

=I,
= tan} z — en u
Ae) OF has) 3
tie
vo4
Re
mi
Example 7. Evaluate
é 32? +2
—$—_—_—_—_— a2}
5 (x —4)(z+1)
Solution. For a rational function such as we have here, the usual
method is to try Partial Fractions. That is what we do here. We will
assume that
327 +2 4 i B i C
(2 — 4) (x +1) g—4 z¢+1 (41)
 160 Ch: 20

By the Cover Up Rule, we can see that

3x47+4+2 50
— Dh
x 25 25
Also by the Cover Up Rule, we can find
_4)2
PAE tal
LL SDE LET
(—5) —5
To determine B we need to multiply the earlier equation by the ex-
pression (zr — 4) (z + 1)’. Then we obtain, on substituting A = 2 and
C=—1,
37? +2 =2(2 +1)? + B(x +1)(z —4) -1(x—4).
If we compare the coefficients of z* on both sides of this equation, we
see that 3 = 2+ B,so that B = 1. Hence we have to find

[ 2 1 1
soe — ——, | dz
Bae eee eel)

The first two of these integrals are standard In-type integrals. The
third is also very straight-forward, but involves a linear change of vari-
able. So, for the third integral only, we shall let u = x +1, so that
du = dz, and when z = 5, u = 6, while when z = 8, u = 9. Hence we
have
‘i 2 1 1 )
ss comma eee ae
a \t=4 al. (eee)
8

= sgt emetel 12h aay5 era EG

=[21n|2 — 4|]5 + [In |z +1], — ireras

6
= [2In|x —4{)§ + [In|z +15 + [uo],
wh
= (21n|4| — 2In{1|) + (In|9| — In|6|) + €ie SS3)
1
= nl6+1n
1n16+1n9 —1n6
n6 — —
i8

= ]n24 -— er,
18
 Ch: 20 161

EXAMPLES 20
1. Evaluate each of the following:
4 3
(a) i}(x* — 32? + 42 — 1) dz; (b) ‘i(62? + 2x — 3) dz;
1
1 oe
(c) fi eda; (d) i = dz:
a 0

(e) iP sin 2z dz; (f) ‘fs cos 3z dz;

: a7
(g) i sec 2z tan 2z dz; (h) ie cosec’ z dz;
0 &
? 4 dz
(i) iH sec xdz;
dr (j) / asi

(k) fom dz 4) ips dz

0 /4 — 72 ‘ 0 16 -+- x?

2. Evaluate each of the following definite integrals, by making a

suitable change of variable:
9 dz eat
cliwyivare a rer
: £ 4 cose
———.,, dr; d St
©) i Grae ¢) | lfsin’s ee

t 3 3
(e) i cos z sin? x dz; (f) / z* (442°)? dz;
0 0

Ope Piriots nest

ka Setibetcaees .
(0) fee
nig sah ee == i .

0i [zs Soe
9 t*—4r+8
Mey
:
= = — 2?
SSS
2 Vd+4a

(k) 1 sin z cos® z dz; (1) ilesin‘ cos x dz;

° &
(m) he sec’ 2x dz; (n) iG sec® x tan z dz;
0 )
$k z
fo) dz; Pp | ———;dz.
(0) / 1+ <4 (p) F (4422)?
 162 Ch: 20

3. Use integration by parts to evaluate each of the following:

x

(a) i z cos z dz; (b) | re°* dz;

0 0
3 e
(c) i)ae 2? dr: (d) / z* In z dz;
1 1
(e) ie r cos2zr dz; (f) ip x? sin 2z dz;
oe

(g) i In(z + 4) dz; (h) / tan! adz

—2 0
4. By using the double angle formulae,
1 1
sin?A = 5 (1 —cos2A) and cos*A= 5 + cos2A),

or otherwise, find each of the following:

=

(a) je sin? 4a dz; (b) - cos” 2zr dz;

0 0
(c) i sin? — cos? ~ de; (d) 1p cos‘ xdz .
0 2 2 0
5. Use partial fractions to evaluate each of the following integrals:
* 3a43 2 Saks
© bearer———_— dr;
0 ©
b
sheagarn®——_.——_—~ dz;

blr.—6 * 2? +4+22-2
t°) f re=He9
a
FA) ——_—_——~ dr;
ins
d
8
Engecanycetea) OF
———__——_——. dr;

» [ tr
ae SATIN
z+ 5. (f) / ee 6 i
(e-1) (e+) 3 (x +2)" (¢—-1)
* 20? + 6 —6 ; 6x — 4
g —————_.—~ dz; Se ene
te) o (t+1)(2?
+4) ) I (x + 1)? (2? +1)

6. If f(z) is a real function which is continuous on the interval

[a,b], then the average value, or mean value, of f(z) on [a,] is
defined to be
— [ f(x) de.
Find the average value of
(a), ere on [1,4]; (b) cost on (0,4
(c) «3+1 on [1,1]; (d) e?? on = [0, In 2].
 163

CHAPTER 21
The Area under a Curve
Consider the problem of trying to find the area of the region be-
tween the curve y = f(z), the z-axis and the lines zr= a and z = 8,
where f(z) is a continuous real function such that f(z) > 0 for all
xz € [a,b]. We are therefore assuming
that the curve y = f(z) lies above or !
on the z-axis throughout the range of
z-values in which we are interested.
We could obtain an approxima-
tion to the above area by means of n
rectangles in the following manner.
Let n be a positive integer, and
divide the interval [a, b] into n pieces
(not necessarily of the same width),
where the width of the ith piece is
Az;.
In the 7th subinterval, choose an
z-value x}. Form a rectangle with the ith interval as base and of height
f(z) [the height of the function f above the z-axis at r = zt]. The
area of this rectangle will be
fae) sAz;,
and the sum of these areas will be
n

oy (fix Aa,
i=]

You may recognise the first steps of the description we gave above.
At school, pupils are often asked to cover a semi-circle, or similar shape,
approzimately with a number of rectangles of coloured gummed paper,
usually using rectangles all of the same width. By a simple measurement
it 1s then possible to find an approzimate value for the area of the semt-
circle. .
If we let n, the number of rectangles, tend to infinity in such a way
that the width of the widest rectangle tends to zero, then we obtain
n

Jim )) f(z?) x Ag;.

i=1

We would expect that the method described above would give us

a way of evaluating the area we require, although the work required is
considerable, except for the case of functions such as z, x? and 2°. |
Fortunately there is a much better way of evaluating the area under
the curve y = f(z).
 164 Ch? 21

The Fundamental Theorem of Integral Calculus tells us that

lim )~ f(2f) x Ag; = [F(z)|,

i=1

where F(z) is any function such that F’(z) = f(z) for all z € [a, 6].
Put another way, for a non-negative continuous function, the area
of the region between the curve y = f(x), the z-axis, and the lines x = a
and zr = 6 is given by the definite integral

[ f(x) dz.

If we introduce the idea of signed areas, which are positive for

regions lying above the z-axis and negative for those regions which lie
below the z-axis, then, for any real function f(x) which is continuous
on the interval [a, 5],
b
[ tea
will give the sum of the signed areas between the curve y = f(z), the
x-axis, and the lines zr= a and r = b.

Example 1. Find
=
2

ii sin xzdz.
are
2

Solution. We can easily see

that
3 x
i sin zdr = [— cos x]? 5

= (of +em(-$))
2

=0-0=0.
The region in question is formed
of two equal pieces which are sym-
metrically placed with respect to the
origin. One of the pieces is above
the z-axis, and has a positive signed
area, while the other piece is equal in size, but lies below the z-axis.
Hence it has a negative signed area. It follows that the sum of the two
signed areas, which is what we found by integration, is zero, as we found,
because the two signed areas cancel each other out.

If we required to find the physical size of two pieces of cardboard

cut out in the shapes shown above, then we would need to find
 Ch: 21 165

be ¢ £ 0
/ sin cdr -f sin zdz = [—cosz]? —[—cos r]_s
0 a.
1 Ls
= (—cos 5 + cos0) — (—cos 0 + cos (-5))

=1—(-1) =2.
Notice that in the above calculation we add the area of the region
lying above the z-axis to the negative of the signed area of the region
lying below the z-axis. We do this because the negative of a negative
signed area will give us the positive value of the physical area we require.
Example 2. Find the finite area contained between the x-axis
and the curve y = z* — 62 + 5.
anihen We can write x mete = (zr — 1)(x —5). Notice that
y = z* — 6r + 5 is a quadratic func-
tion which comes to a minimum value
5
somewhere between z = 1 and xz = 5.
It follows that the curve will lie below
or on the z-axis for z € [1,5].
Hence the finite area we require
will be the negative of the signed area
of the region bounded by the curve O
y = x? — 6z +5 and the z-axis, be-
tween the z-values z = 1 and z = 5.
So the required area is —4
5
-f (2? — 6x +5)dz =~ |52°— 32? +5]
eeu
125 5 7p ehigomy Sant 5
= 2a?Saat ian
124 156-124 32
= —-— + §2 = ————_ = —..
3 3

Example 3. Find the area of the region contained between the

curve y = /9— 2? and the z-axis. Hence find the area of a circle of
radius 3.
Solution. y= /9—2? gives the points on the circle z? + y? = 9
for which y > 0. In other words, we have those points on the circle
which lie above (or on) the z-axis. The required area is therefore given
by :

/ V9—-— 2? dz.
=3
 166 Ch: 21

The integrand here involves /9 — x?. This is of the type we met

in the Chapter on Trigonometric Substitutions, where it was suggested
that the appropriate substitution in such a case would be z = 3sin@.
Then dz = 3cos@d0. When x = —3 we need sin§ = —1, and so
6 = —£. When z = 3 we need sin@ = 1, and so 6 = 5. Furthermore
V9—2? = /9-9sin?6 = ,/9(1—sin? 6) = V9cos?6 = 3cos8. It
is valid to take the square-root of 9cos? 6 as being 3cos@ since we are
considering the interval —$ < 6 < $, and on this range of values, cos8
will be non-negative.
The required area is
3 $
jk vo- sas = | 3 cos 9. 3cos 6 dé
-—3 x
ei
3
x

= 9 | cos”
6 dé

i (1 + cos 26)dé

Hence the area of the semi-circle

or
above the z-axis is =, and so the
area of the whole circle is 9x. This —
is of course the result that we would dee vee
have expected from the formula mr?
for the area of a circle of radius r.
(The justification for that formula,
which we were taught at school, is
in fact the proof we have just given —3
in the particular case r = 3.)
Example 4. Find the finite area in the first quadrant contained
between the curve y = z° and the line y = 4z.
Solution. It is a good idea in examples such as this one to sketch
the curves involved, so as to get a clear idea of the area you are trying
to calculate.
The curve y = z* and the line y = 42 meet where z*° = 4z, or
z (x? — 4) = 0, or r(x — 2)(x
+2) = 0 which gives z = 0, t = 2 and
r= —2.
 Ch: 21 167

Let A; be the area of the region in the first quadrant between the
line y = 4z, the line z = 2 and the z-axis. Similarly, let Az denote the
area of the region in the first quadrant between the curve y = z°, the
line x = 2 and the z-axis. As can be seen from the diagram, the area
we require is A; — Ay. Now
2
Aj =| 4rdzr = [227]; = 8,
0
while
2 1.7?
A; = / z3dz = Fa = 4,
0 4-15
It follows that the area of the re-
gion we need is 8 —4 = 4.
The following example is quite
similar to the previous example, but
we shall adopt a different approach
to it. This is purely done for variety, not because there is anything
wrong with using the methods we have introduced above.
Example 5. _ Find the area of the region contained between the
curve y = x” — 2z — 8 and the line y = 2z — 3.
Solution. | We can factorise x? —2z —8 as (x — 4) (x + 2). Hence
y = zr? — 2r — 8 is a quadratic curve
which crosses the z-axis at —2 and 4
as shown. y = 2z — 3 is a straight
line meeting the axes at the points
(0, —3) and (3/2, 0).
The line and the curve meet at
the points where
x? — 27 —8 = 2r — 3,
which is the same as
zr? —4¢ —-5=0,
or

(cx —5)(x+ 1) =0.

Hence the intersections happen at the points with z-values z = 5
and z = -1.
We shall approximate the area of the required region by means of a
series of rectangles, one of which is shown in the diagram. Its height is
(22 — 3)— (2? — 2 — 8), and its width is Az, so that an approximation
to the required area would be
»s ((2x — 3) — (x? — 2x — 8)) Az,
rectangles
 168 — Ch: 21

and, on taking the limit as we increase the number of rectangles, en-

suring that the width of the widest rectangle tends to zero, the area
is
lim Dia ((2 — 3) — (x? — 2z —8)) Az,
aos rectangles

which, by the Fundamental Theorem of Integral Calculus, can be eval-

uated as

de = fi(40 +5 —2?) dz
[ (@2-3)- (et -22-8))
5
= 2? +52 — ial
3 =i)

EXAMPLES 21
1. For each of the following, sketch the given curve, and find the
area of the region between the given curve, the z-axis and the given
lines:
(a) y=2? +3, f= 1 cand eg
(b) y=6e—2% Gal Juanda css:
(c) y=e2eos2z, #210) S-andtixer

2. For each of the following, sketch the given curve, and hence
determine the total area of the region bounded by the given curve, the
z-axis and the given lines:
(a) y=a?-22-8 2=2 and 2z=6;
(b). Meta? @=-—1_.ands .2=3.

3. Find the area of the finite region in the first quadrant which is
bounded by the curve y = z° and the line y = }z.
4. Sketch on one diagram the arcs of the curves y = cosz and
y = sin 2z for which 0 < z < §, and hence find the total area of the two
finite regions bounded by these two arcs and the line z = 0.
 169

CHAPTER 22
Volumes of Revolution
Let f be a continuous real function of z and suppose that A denotes
the region contained between the curve y = f(z), the z-axis, and the
lines z = a and z = b. If we rotate A through one complete revolution
about the z-axis, then we will obtain a solid object, having a volume
which we shall denote by V.
We can approximate V as follows. Divide the interval from zr = a
to z = b into n pieces, the ith of which has width Az;. Choose an
zi in the 7th subinterval, and consider a circular disc, of thickness Az;
and of radius f(z‘). So we are approximating the volume by a series
of circular discs, of differing thicknesses, stacked side by side with their
centres lying along the z-axis. Now
the volume of the zth disc will be
2
m(f(zz)) As; ,
area of circle thickness

so that V can be approximated by

> *(F(2}))? Azi.

i=1

As in the chapter on areas, we

let n tend to infinity in such a way
that the thickness of the thickest disc
tends to zero. Then we see that

V = lim )ox(f(29))? Aci,

i=1

which, by the Fundamental Theorem of Integral Culculus, gives

b b
v= i m(f(z))? dz = / my*dz.

Example 1. Find the volume generated by rotating the region

bounded by y = 3z*, the z-axis, r = 0 and x = 2 about the z-axis
through one revolution.
Solution. The required vol-
ume is
2 2
il™ (32*)’ dz = a 92° dz
0 0
Saif
= 2°x.
 170 Cn:°22

Example 2. By finding the volume obtained by rotating the

semi-circle x? + y? = r? for which y > 0 about the z-axis through one
revolution, obtain the formula for the volume of a sphere of radius r.
Solution. The required vol-
ume is y

which is the well-known formula for

the volume of a sphere of radius r.
Example 3. By finding the volume obtained by rotating the
region bounded by the line y = ;2, the z-axis, r = 0 and x = h about
the z-axis through one complete revolution, show that the volume of a
right circular cone of height h and base radius r is amr7h,
Solution. If we rotate the region described through one revolu-
tion about the z-axis, we will form a right circular cone of base radius
r and perpendicular height h, lying on its side.
The volume of the cone will be

[ntae= [ox(52) ae :

as required. O ha
This formula, and the formula for the volume of a sphere which
we gave earlier, were formulae which we stated in an earlier chapter,
promising to prove them later. That promise has now been kept.
Suppose now that we have two continuous real functions f(z) and
g(x) such that f(z) > g(x) > 0 for all x € [a,b]. Consider the volume V
obtained by rotating about the z-axis the region A contained between
 Ch: 22 171

the two curves and between the lines

z= aandz = 6b. Wecanthink 4
of this volume as being the volume
left when we have formed the volume
from rotating f(z) about the z-axis
and then removed from the centre of
this volume a volume obtained by ro-
tating g(x) about the z-axis. So we SSX
bi IRIEL OS ©
have a volume with a circular cross- ML vena es
2% (
section having through the middle of Q
res Ly q
it a hole also having a circular cross- XL ‘ ALR Lysay,x2
section. Oo; 4 be
Let A, be the region contained
between the curve y = f(x), the z-axis, and the lines r = a and z = b.
Let V; denote the volume formed by rotating this volume through one
complete revolution about the z-axis.
Let Az and V2 denote the corresponding quantities for y = g(z).
Then the required volume is

V=V,-V2
6 6

= f rlfleyP
ae-fx {o(a)P ae
=7 f {UG - lola} ae
Example 4. Find the volume of the body obtained by rotat-
ing about the z-axis through one revolution the finite region enclosed
between the curves y = —3x? + 242 — 10 and y = x? — 82 + 18.
Solution. In this type of problem it can be very useful to sketch
the curves to see what is happening.
The two curves meet where
x? — 82 +18 = —32?
+ 242 — 10
which we can rearrange as
42? — 322 + 28 =0
ly = 24x — 327 — 10
which is the same as

2? —8r+7=0

or
(x —1)(x—7)
=0.
So the curves meet at the points
with z-values r = 1 and z = 7.
 172 Ch: 22

The required volume is therefore

7
7 | {[-32?+242—10]” — [2? - 8¢ + 18]"}de
1
7
=a | {[92* — 1442 + 636x? — 480z + 100]
1
— [c* — 162° + 100x? — 2882 + 324] }de
7
yr / {8x4 — 1282* + 536x? — 1922 — 224} dz
1
7
= 15 Ee — 3224 + ae — 962? — 224]
1

r{(E75 — 92x 74 + S879— 96 x 7? — 224 x7)

8 536
ie efi
— |— —32+ — — 96 — 224
)}

EXAMPLES 22
1. For each of the following, determine the volume obtained by
one revolution about the z-axis of the region bounded by the given
curve, the z-axis and the given lines:
(A) eas Za ORE se as &
(b) pape ee’ fo lee and. Ae Soe2
1
(c) 52? fate aid’ 2 =:
(d) y=tanz, t=0 and «=F

2. Find the volume of the body obtained by rotating, through

one revolution about the z-axis, the finite region enclosed by the curve
y = =andthe line 22 + y =7.

3. Describe the body obtained by rotating, through one revolution

about the z-axis, the smaller of the finite regions enclosed by the circle
z*+y? = 13 and the line y = 2. Find the volume of the body so formed.
4. A torus is the quoit-shaped body obtained by the rotation,
through one revolution about the z-axis, of a circle which does not
intersect the z-axis. Find the volume of the torus which would be ob-
tained by rotating, through one revolution about the z-axis, the circle
x? +y? —4y4+3=0.
 173

CHAPTER 23
Arc Length and Surface Area of Revolution
Suppose we are given a function y = f(z) which is differentiable
for all values of x between the points
A(a, f(a)) and B(b, f(b)) lying onthe /
graph of the function. Suppose that
we require to find the length along
the curve from the point A to B.
As we have done before, we shall
divide the length of curve between A
and B into n pieces. Between each
consecutive pair X; and X;4, of di-
vision points on the curve, choose a
point z} such that the tangent to the
curve at the point (x7, f(z})) is par-
allel to the straight line joining X;
to X,;41. We shail approximate the
length of curve between X; and X;4
by the straight line from X; to X;41, lines of
which has gradient f'(z*).
As can be seen from the dia-
gram, the length of the curve be-
tween X; and X;4) is approximately

V1+ [f' (22)? Az.

Taking the sum over all of the n segments into which the arc length
is divided, and then taking the limit as n tends to infinity, with the
length of the largest segment tending to zero, we see that the length of
curve between A and B is
B

dimSO Vt LF (en? Aas,

A

which, by the Fundamental Theorem of Integral Calculus, we can eval-

uate as ‘

/ Aj A + (y')*dz.

Example 1. Find the length of the curve y = zi between the

points z = 0 and r = 2.

Solution. lf y= z? then
dy = set, and hence
dz
2 2

1+ (2) =1+ ($24) eneee toe

dr 4 4
 174 Ch: 23

Hence

AA = Vito.
It follows that we need to find
2
>f V4+4+ 92 dz.
0
In order to integrate this, we shall let u = 4+ 9z, so that du = 9dz, or
di 3 du. We also have to change the limits from z = 0 and z = 2 to
u = 4 and u = 22. So we have
2 1
a Vive ds = 5 | uz? x —du
0 24 9
speeder
=—.—.— lu
293 4
ok

Example 2. Find the length of the curve y = Vr? — x? between

the points x = 0 and x = r. Hence show that the circumference of a
circle of radius r is 2rr.
Solution. Notice first of all that y = Vr? — x? gives points with
a non-negative value of y, and so the
length of curve we are asked to find
is the length of the part of the circle
z* + y? = r* which lies in the first
quadrant. So the total circumference
of the circle will be four times the
length in the first quadrant.
Now y = (r? — 2?)?, so that

oh= 5 (0? 2?) (22)

1

a —£
oe
Hence
 Ch: 23 175

eee
Therefore

ie @)
ae
- =
Hence the basa) of the quarter of the circle is
= e ee Re
dz =r |sin 12)
/ Vr2 — x? rJo
Pes ye a=
=rsin-! - —rsin7! —
e i
T
lie ene

= -—TT.
2
As explained above, it follows that the total circumference of a
circle of radius r is 2zr.
As we saw in the last chapter, if we take a part of a curve and rotate
it through one revolution about the z-axis then we obtain a solid body.
In the last chapter we found a for-
mula which enables us to find the vol-
ume of such a body. Now we would
like to have a formula which enables
us to calculate the curved surface
area of such a body. It should be
carefully noted that the formula gives
the area of the curved surface of the
body, and does not include the areas
of the circles at the two ends of the
body.
Given a differentiable function
y = f(z), the curved surface area
generated by rotating about the z-axis through one revolution the part
of y = f(r) between the points x = a and z = bis
6

: onl f(2)|y/1 + (y)* de.

Example 3. _ Find the area of the curved surface generated by
revolving the arc of y = 2x2 between x = 1 and x = 3 about the z-axis
through one revolution.
d
Solution. Given that y = 2x2, then = 3 272, so that
2

1+(2) RRS ite 2M

dz x
 176 Ch: 23

Hence

It follows that the surface area generated is

3 3
J an |204 yt ar = 4 | at [22> de
1 z 1 z
3
=4n [ Vz+1dz
1
4
= ar| uidu
2
(on putting u=2+1)
2
3
= 5 (43 — 28)
(4y/4 — 2V2)
Example 4. By rotating the portion of y = Vr? — x? for which
—r <2 <r about the z-axis through one complete revolution, find the
surface area of a sphere of radius r.
Solution. Note that if we take the semi-circle y = Vr? — 2?
(for which y > 0) and rotate it about the z-axis through one complete
revolution, then we shall form a sphere of radius r.
2
In Example 2, we calculated 4/1 + (#) when y = Vr? — 22,
and we use this in the present Example. The required surface area is
r

[ 2=|vA=a aos = anf rde

= 2m [pale
= 2n (r? —r(-r))
= 4rr?,

as we expected.
The formula for the curved surface area of a right circular cone
of base radius r and height A can be obtained by applying the above
formula to the line y = ;z between z = 0 and z = h.
 Ch: 23 Tha

EXAMPLES 23
1. For each of the following, find the length of the arc of the given
curve between the points with the given z-coordinates:

(a) whe ce and “7 = 4:

(b) y=2(2?-1)?, 2=2 and 2=4;

z4++12
Pere eh Andis f=.3"

y
pie
Se Pe tame
(c)

Qi eas), ape!
She, pa
(ce) y=52?—Ing ei Se
2. Find the area of the surface generated by revolving about the
z-axis, through one revolution, the portion of the curve
(8), nin) .20/Z for which B= <8:
(b) \ecguer® for which Okra:
(e)Griiy =e (et es”) for which Oa.

3. Find the z-coordinate of the point of intersection of the two

semi-circles given by the equations y = V16— <2? and y = V8z — z?.
Find the area which lies inside both of these semi-circles. Find also
the volume and the surface area of the solid body which is obtained by
rotating about the z-axis, through one revolution, the region which lies
inside both of the semi-circles.
 178

CHAPTER 24
Numerical Integration
There are four main reasons why we should want to carry out Nu-
merical Integration, which involves determining the approximate value
for a definite integral of a function from knowledge of the values of the
function at certain z-values.
1) Sometimes it is impossible to find the integral of a particu-
lar function. For example, e? and \/4-—sin’ z are both fairly simple
functions fur which no indefinite integral exists in terms of the standard
functions.
2) We can find the values of expressions such as In2 or 7 by
finding numerical approximations to definite integrals which equal these
values.
3) In an experimental situation we could need to find the in-
tegral of a function, because the integral has some significance, while
our only knowledge of the function is from taking measurements of the
function at various moments in time. It is therefore more sensible to
calculate the definite integral directly, if that is what is wanted, rather
than going through the detour of determining a function which has the
values obtained in the experiment, and then trying to integrate this
function.
4) A computer can be very easily programmed to carry out nu-
merical integration, while it is considerably harder to teach a computer
(or a person, for that matter) to integrate more than the most straight-
forward of functions.
An important feature of numerical work is that we can obtain a
measure of the possible error involved in the work. This means that if
we are finding, say In 2, then we can show that the value of In 2 must lie
between 2 values. For example, if in using the numerical methods given,
we could show that In2 must lie between the two values 0.693144 and
0.693149, then we could deduce that the value of In2 must be 0.6931
correct to 4 decimal places.
However in this book, we shall not be interested in finding formulae
for the errors involved in our numerical approximations.
We shall assume that f(z) is a non-negative continuous function
on the interval [a,b], and we shall give two ways of approximating the
definite integral
b
/ Se) de:

Each of the methods is a refinement of the original idea of approx-

imating to the definite integral by means of a number of rectangular
strips.
 Ch: 24 179

The Trapezium Rule.

We divide the interval [a,b] into n equal pieces, each of which will
have width ons , by using the division points
Qe" Ot ay Sit. Ke eT eee =U,

We then approximate the area

between the curve, the z-axis, and approxunating
straight line “Pe
the lines ¢.=12;— and. to=: 2; by.
means of a trapezium of width aoe
and height f(z;-)) at one side and
height f(z;) at the other. [Note that
we are using the word trapezium to
describe a quadrilateral figure with
two parallel sides. Many American
text-books use the word trapezoid to
describe this figure, with trapezium
being used for something different.
Such books refer to the Trapezoidal
Rule, instead of the Trapezium Rule.]
The area of the 7th trapezium will be
[feu se) pee)
n

average height width of base

Taking the sum of the areas of the n trapezia, we have, as an

approximation to the area under the curve between z = a and x = ,
b—a\ [f(to)+f(x1) , f(ai) + f(z2) , f(x2)
+ f(zs)
(A=*) |Bee tee , Ree Re, Rep
WSenti
+ i(te)
2

= (7) (rte) + 24(er) + 2flen) + 2fles) +

b-—a

+ 2f(2a-1) + f(zn)|,
and, provided we make n sufficiently large, this should provide a rea-
sonable approximation to
b
iif(z) dz.
a

We can remember the Trapezium Rule as

[ f(z)dz = (*=*) [f(Initial point) + f(Final point)

+2 x {sum of f(Intermediate points)}].

 180 Ch: 24

Here we are using the symbol ~ to denote ‘is approximately equal

to’, in contrast to its earlier use in curve sketching to denote ‘behaves
asymptotically like’.
Example 1. Use the Trapezium Rule with n = 3 to approximate
5 2
——_ d.
/ eth aid
and hence find an approximate value for |n 2.
Solution. | Weneed to divide the interval [2, 5] into 3 equal pieces,
each of which will have length 1. Hence 2, 3, 4 and 5 are the points of
division. Taking
f(z) = a
we have f(2) = 2 f(s) = 2 f(4) = 2, and f(5)= =. Hence, using
the Trapezium Rule, we have

ik oda pg ean Benen iat

= Taree 2S MIS 8 15° 24
3 [12°).4 4 2
-ls+statal
1/40 30 16 5
-s|Statatal
ere ol
=5 [55
91
120
However, using Example 8 from the chapter on Partial Fractions,
we know that
5
z—1 |
[ wae |b z+
= (alse)-
ee
ai
=Inz-In;

== 1n.2:
From this it follows that In2 ~ 21 ~ 0.7583. [A calculator gives
the value of In2 as 0.6931 (correct to 4 decimal places), suggesting that
the above approximation is not a particularly good one.] However if
we increase the value of n, then we would increase the amount of work
involved in finding the approximation, but we should obtain steadily
improving approximations to the value of In 2.
 Ch: 24 181

Simpson’s Rule.
In the Trapezium Rule we approximated the curve y = f(z) by
means of a series of line segments. For Simpson’s Rule we shall approxi-
mate the curve y = f(x) by means of a series of arcs of quadratic curves
y=azr*?+brt+c.
To approximate f{if(x) dz, we divide the interval [a, 6] into n in-
tervals of equal length b—a where n is an even number. Suppose that
the points of division are
Q= 70 < te <— 77 <2. Sia Sry =O:
For each even integer i between
2 and n, we approximate y = f(z)
between z = 2;-2 and z = 2; by a
quadratic curve which we choose to
pass through points (z;~2, f(zi-2)),
(zi-1, f(zi-1)) and (2;, f(z;)) on the
original curve.
The area beneath the quadratic
curve can be expressed in terms of
the three points we have made it go
bee) Vi-1 Vix
through, and we can show that it is

(AS) [F(zi-2) + 4f(2i-1) + F(z:)]

= (75°)
b-—a
loa) + 4f(einn) + Sle).
If we add the area under each of the quadratic curves, we obtain
[soa a (7=*) {[f(x0) + 4f(21) + f(2-)]
3n

+ [f(z2) +4f(23) + f(za)] +...

+ [f(tn-2) +4f(tn-1) + f(zn)]}
= (*) [1le0) + 4f(en) +2420) + 4f 0) + 2420)
+20. +2f(2n-2) +4f(@n-1) + f(n))-
This can be remembered as

iFf(z)dz = (?= *)[f(Initial point) + f(Final point)

3n
44x {sum of f(Odd points)} + 2 x {sum of f(Even points)}].
It is important to note that there are several different versions of the
formula for Simpson’s Rule. This should cause no confusion provided
we stick to one version of the formula, and remember what the symbols
stand for in the version we are using.
 182 Ch: 24

Example 2. Use Simpson’s Rule with n = 4 to find an approxi-

mate value for
1
I V4 — 2?’
and hence find an approximate value for 7.
Solution. |Weneed a Une the interval (0, 1] into 4 equal pieces,
and so we need to have 0 ’ re 12? 34 and 1 as the points of division.
We shall let 1
f(z) = Tiras
so that f(0)= 3, f(4) = ee = 45) f(z) = tug S
1G)S aan |0) eeeoa
at Simpson’s Rule with n = 4, we have
1-0\/1 2 4 1
2x —4+4x —+4+—
«(a [5tp wee V15 ee ae
iSs
= (4) [3+ 16 *So iraegt AO piel
v6s V15 55 V3
=(3) (0.5 + 2.0158105 + 1.0327956 + 2.1574396
+ 0.5773503]
= 0.5236163 (correct to 7 decimal places),
on using a calculator.
Furthermore, we know that

petieues
1 8

0 V4—-—<22 250
i

and so it follows that

= 0.5236163,
O|Aa

and hence
m = 3.1417, correct to 4 decimal places.
Since the actual value of 7 is given by the calculator as 3.1416
(correct to 4 decimal places), we have quickly obtained a good approx-
imation to the value of 7.
Increasing the value of n would lead to improved accuracy in the
approximation we obtained, again at the expense of increasing the work
involved.
 Ch: 24 183

As a general rule, Simpson’s Rule will usually give a better approx-

imation for an integral than the Trapezium Rule will, although, as with
any generality, it is possible to find exceptions to that statement.
Sometimes we are asked to apply Simpson’s Rule, or the Trapezium
Rule, with a specified number of ordinates. The number of ordinates
is the same as the number of evaluations of the function, which, as in
the examples above, is one larger than the value of n. So, for example,
if we are told to use 9 ordinates, then we should take n = 9 —1 = 8 in
the formulae.
We shall conclude this Chapter with a ‘practical’ example in which
we calculate the integral of a function (approximately) from given values
of the function, without trying to determine what the function actually
is.
Example 3. The velocity v m/sec of a car is noted at time
intervals of 20 sec and is as tabulated.
[tse [0] 20[40 |60|80|100|120|140|160]
OE
By using Simpson’s Rule with 9 ordinates to approximate i yy dt,
find the distance travelled by the car in the first 160 seconds.
Solution. Since we have 9 ordinates, we shall take n = 8. Hence

ts 160 — 0
vdt & nah [1 x v(0) +4 x v(20) + 2 x v(40) + 4 x v(60)
)
+2 x v(80) +4 x v(100) + 2 x v(120)
+4 x v(140) +1 x v(160)]
= (F) [1x 0+4x 1242 x 2844 x 16+2 x 35+4 x 20

+2x35+4x22+1x7|
= (3) [0 + 48 +56 + 64+ 70 +80 +70 +88 +7]
= = x 483

ee AI

Since the integral of the velocity with respect to time will give the
distance travelled, we conclude that the car has travelled 3220 meters
after 160 seconds.
It may help us to remember the Trapezium Rule and Simpson’s
Rule if we notice that in the case of the Trapezium Rule we are multi-
plying the values of the functions by the coefficients
1-0) 210000, 3-0-1.
 184 Ch: 24

while in the case of Simpson’s Rule the coefficients are

Ty4, 2;Ay cs ae ys os hs
and then we are dividing the sum of these terms by the sum of the
coefficients we have used.
EXAMPLES 24
1.For each of the following, use
(a) the Trapezium Rule, and (8) Simpson’s Rule,
with the given value of n, to approximate the given integral:
6 T
(a) i! zx? dz, n= 12; (b) [ sin® z dz, n=4;
—6 0

(c) _—
I 14 24’ n=4;
St (d) [2008
6 de ’ n=6 :

2. For each of the following, use

(a) the Trapezium Rule, and (f) Simpson’s Rule,
with the given value of n, to approximate the given integral. Further-
more, find the exact value of the integral, and hence find an approxima-
tion to the given number:
1 dr ee
@) ofpp neh teh
(Bpduar 7
hate2 : 1 =)9).0 In:

3. For each of the functions tabulated below, use

(a) the Trapezium Rule, and (f) Simpson’s Rule,
to find the best approximations you can to

(a) [ seas (b) [ stz)ae.

T= [00] 05 |10 [is |20,7 25 [30 |
[= [000 [025 [050 [075 | 1.00
(2)
4. In order to find the length of the arc of the curve y = sinz for
0<z< §, it is necessary to find

z
i V¥1+ cos? zdz.
0
 Ch: 24 185

Use
(a) the Trapezium Rule, and | (8) Simpson’s Rule,
in each case with n = 2, to find an approximate value for this arc length.
5. Use
(a) the Trapezium Rule, and () Simpson’s Rule,
in each case with n = 6, to find an approximate value for the volume ob-
tained by rotating around the z-axis, through one complete revolution,
the arc of the curve y = x(x — 2)e? for which 0 < x <2.
 186

CHAPTER 25
Newton’s Method
Many equations that we are asked to solve in mathematics have
been contrived to have a nice solution. For example, z? — 3z — 4 = 0
can be rewritten in factorised form (x — 4)(z + 1) = 0, from which we
deduce that either zc — 4 = 0 or x + 1 = 0, and so the solutions of the
equation are x = 4 and ¢ = —1.
In practical situations, the equations that arise are not likely to have
nice solutions, and so we need to have a method that will enable us to
find at least an approximation to one solution of any given equation.
There are several such methods, of which the best known is Newton’s
Method.
Let f(z) be a differentiable function of z, and let a be a fixed value
of z for which f'(a) # 0. Newton’s Method tells us that if z = a is
one approximation to a solution of the equation f(z) = 0, then another
approximation will be
r=a
f(a)
f(a)”
Although it is possible to give conditions under which the ap-
proximations provided by Newton’s Method improve successively, we
shall not concern ourselves with that ,
here, since the conditions are fairly y
complicated both to remember and
to apply. Instead we shall adopt the
more pragmatic approach that if the
approximations are settling down to
approximate more and more closely
to a value, then that value is likely
to be a solution of the equation.
The idea of Newton’s Method is
seen from the diagram. We approxi- 4
mate the curve y = f(z) by the tan-
gent to it at the point (a, f(a)). This
tangent will have gradient f'(a) [since the derivative evaluated at z = a
gives the gradient of the tangent to the curve at z = a]. Hence the
equation of the tangent to the curve at (a, f(a)) will be
y — f(a) = fi(a)(x —a).
This tangent meets the z-axis where y = 0, and so

0— f(a) = f'(a)(z—a),
which, as we require, can be re-arranged as
 Ch: 25 187

Example 1. By applying Newton’s Method to f(r) = z* — 7,

with initial value x = 2, obtain an approximation to ¥/7.
Solution. Notice that the only real root of z3—7 = 0 will be ¥7.
Starting from z = 2, and applying Newton’s Method, we would hope to
obtain an improved approximation to the root. Now if f(z) = 2* —7
then f'(r) = 3z?, and so f(2) = 1 and f'(2) = 12. Applying Newton’s
Method, a new approximation is

If we applied Newton’s Method again, starting from a = 23/12 then

we would obtain the new approximation z = 36430/19044, which the
calculator gives as approximately z = 1.9129385. Starting from this
value, a further use of Newton’s Method yields xz = 1.9129312, which is
the value that the calculator gives for W/7, correct to 7 decimal places.

This is typical of what frequently happens, which is that on using

Newton’s Method repeatedly to obtain approximations to a certain num-
ber then the approximations we obtain will approach (as accurately as
we wish or can calculate) the true value of the solution of the equation.

Example 2. Find an approximation (correct to 3 decimal places)

for the smallest positive value of x for which the graph of y = tanz
meets the line y = 2z.
Solution. We can sketch the y
graphs of y = tanz and y = 2z as in
the diagram, and from the diagram
we can see that there will be an inter-
section of the two graphs for a pos-
itive value of z somewhere between
xz = 1 and z = }. Remember that
we are measuring z in radians. At
a point of intersection we will have y=tang
tanr = 2z, or tanz — 2x = 0. So
ee
2wld
we shall apply Newton’s Method to
find an approximation to the small-
est positive root of f(z) = 0, starting
from z = 1, where f(r) = 2z — tanz, so that f'(r) = 2 — sec’ 2 z.
Starting from zr = a, the next approximation will be given by
2a—tana 2a—asec 2a—2a+tana
(0C0ereCUrRvXKX—X—_—_—_ =

2 — sec’?a 2 — sec?a
sinacosa—a
~ 2Qcos?a—1
(on multiplying numerator and denominator by cos? a)
 188 Ch: 25

‘ sin2a—a
~ cos 2a
(on using the double angle formulae)
_ sin2a — 2a
~ 2cos2a
(on multiplying the numerator and denominator by 2).
Putting a = 1 into the above formula as a first approximation, a
new approximation will be
sin
2 —2
~ 1.310478,
2cos 2 hoe
from the calculator.
Taking a = 1.310478 as a new approximation, the calculator then
gives 1.223929 as the next approximation. Substituting this value for
a in the above formula, we then obtain 1.176051, followed by 1.165927,
then by 1.165562, and finally by 1.165561, which value repeats from
then on. In all cases, we are giving the value from the calculator correct
to 6 decimal places. It follows that the required root is 1.166 correct to
3 decimal places.
Suppose now that we are given an equation f(z) = 0 and asked to
find an approximate solution of the equation while given no indication
as to where the solution is to be found. What should we do then?
One possible approach would be to take a value at random and
apply Newton’s Method repeatedly, hoping that the approximations ob-
tained will settle down to one particular value, which will be the required
solution.
Another possibility is to use the Intermediate Value Theorem.
This applies when f(z) is a real func-
tion which is continuous on the in-
terval [a,b]. Assume that f(a) > 0
and f(b) < 0 (or that f(a) < 0 and
f(b) > 0.) Then there is at least one
value c between a and 6 such that
f(c) =0.
In less formal language, the the-
orem says that if y = f(r) has a
graph with no breaks between z = a
and z = 5, and if f(a) and f(b) are
on opposite sides of the z-axis, then
the curve must meet the z-axis some- f(b) < 0
where between z = a and x = b.
 Ch: 25 189

Example 3. Find the nearest integer to the positive solution of

x? + 22? — 52 — 5 = 0 and use Newton’s Method to find the solution
correct to 4 decimal places.
Solution. Let f(x) = r* + 22? — 52 — 5. Now f(0) = —5,
while f(10) = 1145. By the Intermediate Value Theorem, the equation
f(z) = 0 must have a root somewhere between x = 0 and x = 10. We
shall evaluate f(x) for integer values of zr between 0 and 10 in order
to get a better approximation to the root. Then f(1) = —7, f(2) = 1.
It follows that the root must lie between 1 and 2, by the Intermediate
Value Theorem.
To determine which is the nearer integer to the solution, we shall
find f(3/2). Now f(3/2) = —37/8, which is negative. Hence y = f(z)
changes sign between x = 3/2 and x = 2, and so zg = 2 is the nearest
integer to the solution.
Now we apply Newton’s Method to f(z) = 2° + 2x? — 524 — 5, so
that f'(x) = 32° + 42 — 5. Starting from z = a, a new approximation
to the solution will be
a®
+ 2a? —5a—5 i 3a?
+ 4a? — 5a
— a® — 2a? + 5a+5
Cah HA gee 3a2
+ 4a —5
_ 2a? + 2a? + 5
Rae
teAG a
Substituting a = 2 in this formula, a new approximation is 29/15.
If we then substitute this value, we obtain 1.930805, and after a further
substitution we obtain 1.930802, which value then repeats itself. Here
we have chosen to work correct to 6 decimal places as given by the
calculator. It follows that, correct to 4 decimal places, the positive
solution of the equation is given by 1.9308.
As a matter of interest, if we had started from the obviously wrong
approximation z = 10, then it would have taken 8 uses of Newton’s
Method before we obtained the above value for the solution. The moral
of this is that you pay for a poor initial approximation by having extra
work to do to obtain an accurate approximation, but in many cases even
a poor guess will eventually yield a solution after a sufficient number of
applications of Newton’s Method.
EXAMPLES 25
1. By applying Newton’s method once to f(z) = z? — 10, with
initial value a = 3, obtain an approximation to 10.
2. By applying Newton’s method once to f(z) = x* — 9, with
initial value a = 2, obtain an approximation to 33.
3. Starting from the initial value a = 1, find an approximation
(correct to 3 decimal places) for the positive value of x for which the
graphs of y = z” and y = cosz intersect.
 190 Ch: 25

4. Starting from the initial value a = 1, find an approximation

(correct to 3 decimal places) for the value of « for which the graphs of
y = z° and y = e~® intersect.
5. Find the nearest integer to the solution of 23 —3r*+5zr—12 = 0
and use Newton’s method to find the solution correct to 3 decimal places.
6. Find the nearest integer to the positive value of z which is a
solution of 4 — 4x3 + 227 — 52 — 20 = 0 and use Newton’s method to
find the solution correct to 3 decimal places.
7. One of the classical problems of Greek mathematics was that of
the duplication of the cube. This problem was concerned with finding
the length of the sides of a cube whose volume was twice that of a given
cube. It was eventually proved that the problem of the duplication of
the cube did not have an exact solution. Find an approximation (correct
to 3 decimal places) for the lengths of the sides of a cube whose volume
is to be twice the volume of a cube with sides of length 3 units.
8. Another of the classical problems of Greek mathematics was
that of the trisection of an angle. This problem involved constructing
two lines which would exactly trisect a given angle. This problem was
also eventually shown to be insoluble. In order to trisect an angle of
60°, it would be necessary to find an exact solution of the equation
823 —6z—1 = 0 having a certain form. Find an approximate solution of
this equation, expressing your answer as a decimal (correct to 3 decimal
places).
 191

CHAPTER 26
Maclaurin Series
As an alternative to the numerical methods which we have used in
the last two Chapters, it is sometimes useful to be able to represent a
function by means of a power series.
A power series in z is an expression of the form
» 3
Qo +a,2
+ a0r +a3r +...,

_ where the coefficients ag, a,, a2, ... are real numbers. So a power series
can be thought of as being the same as a polynomial, but extending to
an infinite number of terms, rather than stopping at a highest power as
a polynomial does.
There is a major difference though between power series and poly-
nomials. If we consider the polynomial

f(@@)Hltzete’?
+z,
for example, then we can substitute any value of z that we like, and
work out the corresponding value for f(z). However, if we consider the
power series
g(a) =1lt+aete?
+r? t+at+...,

then we find that there are some values of z for which g(z) has a finite
value, and other values of x for which it does not. For the series we have
chosen is a geometric progression, with first term 1 and common ratio
xz. The sum of this geometric progression is
1
G\2) = eae
provided —1 < z < 1 and does not exist if x > 1 or if zc < —1. We say
that the series
gx) =1+242?4+22 +244...
is
1 :
cose to i ifaa |<1
divergent 7 Boee Mea
A similar result holds for all power series, and so besides knowing
a power series it is also important to know for which values of z the
power series converges. In this book, we shall state the values for which
power series converge, but without trying to justify the results we give.
Suppose now that we are given a function f(z) and we wish to
choose constants dg, @;, @2, a3, ... such that

f(x) = a9 + a,x + apr? + a3z* + agz* +...,

 192 Ch: 26

at least for values of x close to zr= 0. The approach we adopt is to make

the function and the power series have equal values, equal derivatives,
equal second derivatives, equa! third derivatives, ... at r = 0.
Taking z = 0 gives
f (0) = dag.

Differentiating with respect to z and then putting z = 0 gives

f(z) = a; + 2agr + 3a3r? + 4agz°>4+...,
from which
i (0) = a}.
Differentiating again gives
f(x) = 2az + 6agz + 12ag27 +...,
from which we have, on putting z = 0,
(2)
f (0) = 2a2 or a2 = f {0

A further differentiation, followed by putting r = 0, gives

IOY
(3)
6
Continuing like this, differentiating with respect to z and then
putting z = 0, we obtain

iw SSO)
abeaotag!
’

for all n > 1, where f‘")(0) denotes the nth derivative of f(z) evaluated
atc = 0:
With the constants chosen as described above, the series
ag + a,zr + ax? +a32° + agz* =f

is called the Maclaurin series for f(z).

Example 1. Find the Maclaurin series for f(x) = e*
Solution. If f(#)i = e2,—then 1 i = je7, and SUEUR
fO(z) = fee fO(x) =f(z) = | Hence f(0) =e° = 1,
and f(")(0)= e® = 1 for all n > 1. So we require

= f(0)=1
and
AO
pats ple sen
for n > 1. Hence the Maclaurin series for e” i

ltae+5S24
TT a2 ig
317 aie GO ba e
 Ch: 26 193

which is

Example 2. Find the Maclaurin series for f(x) = (125+ x)8, as

far as the term in x*, and use this series to find an approximation for
23 correct to 5 decimal places.
Solution. We need to differentiate 3 times with respect to z in
order to calculate the Maclaurin series as far as the term in 2. We have

f(z) = (125 +.2)5,

f(z) = =(125 + x) 8 '

FO%(2) = 5(=F) (125-42) 8,

Olen ;(-5) (-3) (125 42)73,
Taking z = 0 in each of these, and remembering that 1253 = 5, we
have

f(0) = (125)3 =5,

Coy -g_iadidi
fmaK), = ae) ABNOR 7b
ins att ae -§_ 221

yon, - 2 f_2 -) -§_ 101

FOV(0) = 5 ( =)( 3 ees) DT. 58
Hence the Maclaurin series for (125 + 2)8 is

sdf 75
ieOe2enieat pelle?
hee) tee 27 Ore Oe
as far as the term in 2°.
Assuming that the Maclaurin series gives a good approximation to
f(z) for small values of z, we shall substitute z = 3 and obtain
a 1 Pea nga ei iUsa
OE a 3 eer—3--aS
e) ty omg tS SS
aT EE a
from which
i il 1
(128)? 5+ 25 «5S 3x57
Notice that 128 = 64 x 2, so that 1283 = 4 x 23, from which we
have
4x 23 +5 +.0.04 — 0.00032 + 0.0000042,
 194 Ch: 26

or, on dividing by 4,
23 1.25992,
which agrees with the value that the calculator gives for 23, to 5 decimal
places.
In the above example, we substituted the value « = 3, and found
that the terms in the Maclaurin series rapidly became very small. This
meant that we only required to use a small number of terms in order
to obtain a good approximation to 23. On the other hand, if we had
substituted a value such as z = 118 in order to calculate 33, we would
have needed to take considerably more terms in order to get a reasonable
approximation to 33. Finally, if we took a value such as z = 131 in order
to find 23 , then the terms in the Maclaurin series would actually increase
in size, so that the Maclaurin series would not have a finite sum in this
case.
We sum this up by saying that the Maclaurin series for (125 + z)3 P
which is

8 1 x4
~ 3) \5) at
(-§ AnD (tafe?
3 3 Cc dus a ea

convergent when |z| < 125

divergent when |z| > 125.
For practical use as an aid to calculation, we need to have series
which converge rapidly, which happens when z is small.

The following are the Maclaurin series of some of the common math-
ematical functions, together with a note of the values of x for which the
series converge.
The series
. CA tegia
e Tel toy eee ey ete

converges for all real values of z, as do the series

eee eee
sinz =z— —+ +
and
ee eee tea e
COS OS Cetin kay meyer

However the series

2 3 4 5
ie oe ee ee
2 3 4 5
converges only when —1 < zx < 1, and the series
—1 neat itn(n —1 t(n —2 ia-323
(a+z2)" =a"+na™ 's+ nm~

a n(n —1) Se 2) (n — 3) nea

baa

converges for —a < xr <a.

The first of the series given above was found in Example 1 above,
while the last one was found in Example 2 [in the particular case a = 125
and n= i, as far as the term involving 2°].
While we cannot prove it here, you may assume that the above
Maclaurin series do add up to the functions of which they are power
series expansions. So, for example, if we wish to find sin z for a value of
x, it is perfectly in order to substitute that value of z into the Maclaurin
series for sinz, evaluate the sum of a number of terms in the series to
give the required degree of accuracy, and to assume that in evaluating
the Maclaurin series for sin z you have also found the value of sin z itself.
The series for (a+ 2x)” should look familiar as it has exactly the
same form as the binomial expansion for n a positive integer. When
n is a positive integer, the binomial expansion holds for all values of
z, since it is a finite series in this case. Even when n is not a positive
integer, the binomial series given above will be valid, for the values
of x specified above, although in this case the series will be an infinite
series.
Notice that if we differentiate the series for e” term-by-term then
we obtain the series for e”. Differentiating the series for sinz, term-
by-term, yields the series for cos z, while the derivative of the series for
cos x is the series for — sin z.
It is als» true, but rather more surprising, that if we take the series
for e* and substitute « = 16, and equate real and imaginary parts,
remembering that e’? = cos@ +isin0, then we obtain the two series for
cos 6 and sin @.
Besides differentiating series term-by-term, it can also be shown
that we can integrate series term-by-term, and that we can add series,
subtract them, and multiply series term-by-term, provided that after-
wards we only substitute values of z which lie inside the set of values
of x for which all of the original series were convergent. We can extend
this to division as well if we add the extra proviso that we must not
 196 Ch: 26

substitute a value of z which makes the series we are dividing by equal

to zero.

There is another useful result which tells us that the power series
that represents a function is uniquely determined. This is an important
result because it says that if we can find a power series expansion for a
function f(z) by some method or other, then this power series must be
the Maclaurin series for f(z). We use this fact in the following examples.
Example 3. From the power series for e* and sinz, obtain the
Maclaurin series for e*'"* as far as the term in x*.4
Solution. We know that
erie Zabevdin wat
e = Fie, Tecgue egis bates

and
} pee ee heh
Sn = Fe bist ayiet

Hence, replacing x by sinz in the series for e*, we have

(sinz)’ (sinz)* (sinz)*
ene {a (sinxz) +
2! 3! 4!
: 2

yahaueagh
uh ldo

AG eagla Jo
3 3 gs 4

(2s eee
4

hai) bis ae
1 1 34 pe 1et hee
wings 1 RET 1+ 83 he1 ot
tae 6° lee Pas hey ng a
eee
il 1
1 ee ag" te
ees

In the above expansions, we ignored all terms which involved powers

of z above x 4 , since we were asked for the expansion as far as the term
inz 4
 Ch: 26 197

Example 4. From the Maclaurin series for (1+ 2)~’ find the
5 : i ; A : :
Maclaurin series for The’ and by integrating find the Maclaurin series
for tan—’ =.
Solution. Taking a = 1 and n = —1 in the binomial series for
(a+z)", we have

(pr) S177 + (=-1)1z 4 D2) 342, CDE)449

2! 3!
DE)mea 3) Ces 49
ae

which simplifies to
(ttealoe wi = pet pie thas
Replacing z by ¢? in this formula, we obtain

(1+ ?7) Fee ie tft

We now integrate both sides of this equation. We obtain
Ee il Bg

eee eat ta ee a

which gives

veal tee t
3 —, Ee gee
S 5 ees
a 9 eee ;
F °

Hence we finish up with the Maclaurin series

a ui we BSc ETH, O29
(GRO QR SS I ee SS
3 5 7 9
which can be shown to be valid for all z strictly between —1 and 1.
EXAMPLES 26
1. Prove that the Maclaurin series for
(a) sing, (b) cosa, (c) In(1+z)
are as given earlier.
2. Find the Maclaurin series for f(z) = secz, as far as the term
in z*, by finding f(a), f(x), f(z) and f(x). Hence find an
. . 1
approximation to sec (;).
e€
3. Find the Maclaurin series for f(r) = , as far as the term
l-gz
in x3
(a) by finding f(z), f(z) and f(z),
(b) by multiplying the Maclaurin series for e” and for (1 — zt).
 198 Chii26

4. Find the Maclaurin series for f(z) =(1+ 2) sinz, as far as the
term in x‘,
(a) by finding f(z), f(z), f(z) and f(z),
(b) by multiplying together 1 +2 and the Maclaurin series for
sin z.
5. Find the Maclaurin series for f(r) = (27+ x), as far as the
term in z?, by finding f(z), f(z) and f(z), and hence find an
approximation to 53 by taking z = —2 in this series.
6. From the Maclaurin series for In(1 + z), write down a series
: : : tie se
for In(1 — z) and hence write down a series expansion for In (; ).
— 2
By taking c = 5 and then x = 15 in this final series, calculate In 11
and 1n13, each correct to 3 decimal places. [You may assume that
In 3= 1.09861, correct to 5 decimal places.]
7. From the Maclaurin series for e*, write down a series expansion
0.2
for e' and hence find (i et dt, correct to 3 decimal places.
0
8. From the Maclaurin series for cosz, write down a series ex-
0.3
pansion for t cos 2t and hence find | t cos 2t dt, correct to 3 decimal
0
places.
9. Use the binomial series to write down a series expansion for
me Se dt
(11—t?)
— 2?) ?, as far as the term in t®. By eae
findin ——— find a
z, as far as the term in 2’.9
1 .
series expansion for sin7

10. Use the binomial series to write down a series expansion for
= ; es :
(1 — t?) a By finding / <a find a series expansion for In € aS );
0 | =
 199

CHAPTER 27
Differential Equations
We shall start this chapter with an example of a very similar type
to the example we solved at the start of the chapter introducing Inte-
gration.
: Example 1. The acceleration of a particle at time t is given by
d°r
= tsint, where x denotes the displacement of the particle from the
dt? | ;
origin at time t. Find an expression for x in terms of t, given that at
time t = 0 the particle is at the origin and is moving at 5 units per
second in the direction of positive values of z.
Solution. We start by integrating both sides of the equation
dz
= tsint with respect to t. We will need to use Integration by Parts
dt?
;
to integrate tsint.
: We must also remember that at2
dase. the derivative
190 F
: d d*x dz
with respect to t of = so that / dt = —. Hence we have
dt? dt
= = t(—cost) + [costat

= —tcost+sint+c.
Now we integ-ate again with respect to t, and obtain
z= —tsint —-2cost+ct+d.
From the initial conditions we are given, we know that when t = 0
dz ; : ‘
we have x = 0 and — = 5, since the distance of the particle from the
origin is zero when t = 0, and the velocity is 5 units per second at that
time. Substituting these values into the above equations we find c = 5
and d = 2. Hence we conclude that
z = —tsint — 2cost
+ 5t +2.

In the above example, we solved a differential equation, which

is an equation involving two variables together with at least one of the
derivatives. First of all we found the general solution, [which in this
case involved two arbitrary constants since the equation involved a
second derivative], and then we found the particular solution that
satisfied certain boundary conditions, or initial conditions.
Differential equations arise in many different practical situations,
and the ability to solve such differential equations is crucial to the suc-
cessful application of mathematics to many practical problems.
In this chapter we shall develop methods for solving some of the
simplest types of differential equations. In order to classify different
types of differential equations, we shall introduce some notation.
 200 Ch:°27

The order of a differential equation is one if the differential equa-

tion involves the derivative etd alone, and the order is two if the dif-
dz
‘ d
ferential equation involves the derivative —F) and possibly - but no
others.
The degree of a differential equation is the power to which the
highest order derivative in the differential equation is raised.
For example, the differential equation
dy ‘
—=y-sinz
dauteats
has order 1 and degree 1, while the differential equation

d’y dy é
qe? tgp + (sine )y2 =e cos r

has order 2 and degree 1. Again the equation

d? dy \*
=f em () —y*? =3tanz

has order 2 and degree 1, while the differential equation

dy\* /dy\* re
2) es ae
has order 2 and degree 3.
We shall be concerned here with differential equations of degree 1,
and of order 1 or 2.

Type 1. First-Order Separable Differential Equations.

A differential equation of the form

a = f(z)g(y),
where f(x) involves z alone, and g(y) involves y alone, is called sep-
arable. Such equations are solved by separating the variables onto
opposite sides of the equality sign, and then integrating both sides of
the equation.

Example 2. Solve the differential equation

dy 32?
dr 4y3"

Solution. We can rewrite the equation in the form

d = 32”,
ays
 Ch: 27 201

and if we integrate both sides with respect to z, we obtain

[ova os [s2?ae,

from which we have

yi=r'te.

Example 3. Find the general solution of the differential equation

dy (3 + 1) = zy’,
dr
and the particular solution for which y = 2 when x = 0.
Solution. If we divide both sides of the equation by z? + 1 and
by y? then we will separate the variables. We have

and so, integrating both sides with respect to z, we have

1 ae
—dy= | ——~dz,
i oe / 14230"
which is equivalent to
1 1
Bh &.

The left hand side of the equation was integrated by regarding it

as [ y~?dy, which is a standard integral. The right hand side was dealt
with by a change of variable. We took u = 1+ 2°, so that du = 327dz.
The above was the general solution of the differential equation. We
need to find the particular solution which is such that y = 2 when
xz = 0. Substituting these values into the general solution, we have
—5 ee i|n |1aa 03 |+c. Since In1 = 0, we must have c = —i. Hence the
required solution of the differential equation is

L he ee

Example 4. Solve the differential equation

dy = 79-7.
dx

Solution. In order to separate the variables in this differential

equation, we need to use our knowledge of the index laws, to write
e2¥-= = e?¥ x e~*. Then multiplying both sides of the differential
equation by e~? gives
emtdy PZ
dx ,
 202 Chzi

which on integrating with respect to xz gives

perry =e fetes,

from which we have

1
—-e *¥.=—e* +6,
2
which, on multiplying both sides by —2, can be rewritten as
e~?¥ = 2e7* +d,
writing d = —2c, where both c and d are arbitrary constants.
One type of differential equation which can be made separable is
the type which can be expressed in the form
dy
ae i res y),

where f(z,y) is a homogeneous function of z and y. A homogeneous

function of xz and y is one which can be expressed in the form 9(4),
a
for some function g.
One way of checking whether the function f(z, y) is homogeneous is
to replace y by vx and then see whether or not f(z, y) can be expressed
purely in terms of v. f(z,y) is homogeneous if, when we write y = vz,
then f(z,y) can be expressed purely in terms of the single variable v,
with all the z-terms having cancelled out.
In particular, f(x,y) is a homogeneous function when f(z,y) is a
quotient of two polynomials where the total degree of each term in each
polynomial is the same.
For example,
zt — 7x3y + 327y? — 8y4
z2y? — 4y4
is homogeneous since each term in the numerator and denominator has
total degree 4. [The total degree is the sum of the degrees in x and y
in each term.] Notice that if we put y = vz in the above expression we
would have
t* — Ta’ vr + 327v?2? — 8vt24* =.1 — Ty + 8v? — 84
ry? — 4y4 24 J v? — 4y4 i
which is a function involving v alone.
On the other hand, the function
x? + 527y — Try” + 8y
2x7y — Try + 4y3
is not homogeneous, since one of the terms in this function has total
degree 2, while all the others have total degree 3. Notice that if we put
y = vz in this expression then we would obtain
zr? + 522ur — Trv?2? + 8°23 Ltt 5ur — Trv? + 8v°2
222 vx — Trvz + 4323 7 Qua —Tv+4v3r”
 Ch: 27 203

from which the z-terms cannot be cancelled.

The method for solving a differential equation with a homogeneous

right hand side is to put y = vz, where v is taken to be a real function
of z which we will have to determine. The differential equation can then
be solved to find v, using separation of the variables, and then we can
find y in terms of z.

Example 5. Solve the homogeneous differential equation

dy 3y?—4zy
dz 2z(y—-z)’

Solution. Since we are told that the differential equation is ho-

mogeneous, we know that the method described above is the appropriate
one to use.
We put y = vz, so that, by the Product Rule for differentiation, we
have

Hence the differential equation becomes

i dy. 3u*z? —4rur = 3? — 40
vt 2— = ——_ = —
dz 2zr(vzr — zx) 2(v —1)’
on dividing the numerator and denominator of the right hand side by
z*. Subtracting v from both sides of the differential equation we have
dv _ 3v?
— 4v
OST E no
_ 8v? — 4v — v2(v— 1)
2(v — 1)
3v? — 4v — Qu? + 2v
ay 2(v
—1)
a v? — Qu
~ Wv=1y
We rewrite this equation with the z on the right hand side, and all
the terms involving v on the left hand side. To achieve this, we divide
both sides by v? — 2v and by z, and multiply both sides by 2(v—1). We
have
Av—1)dv 1
v?—2u dr x’
and, on integrating, this gives

[FSe- [re
 204 Ch: 27

The left hand side of this equation is a rational function for which
Partial Fractions is the appropriate method. Using the Cover Up Rule,
we have ;
[G+ : )do= f Fae
v wv—-2 x

which gives
In |v| + In |v — 2| = In|z| +.

We can improve the appearance of this solution if we write c = InC,

where C is an arbitrary constant. So we have

In |v| + In |v — 2| = In|z|+InC,
which we can rewrite, using the rules about logarithms, as

In |u(v — 2)| = InC|z\,

or as

v(v — 2) = Cz,

and on replacing v by Y this gives

£
y(y — 22) =Cz’*.
You will notice that in the above we dropped the | |symbols. From
now on we shall normally omit the | | symbols when integrals lead to
In-terms. This is not strictly accurate, but we shall find one permis-
sible solution of a given differential equation. We do not wish, in this
simplified treatment, to get involved in questions concerning whether
expressions involved in the solutions are positive or negative. Notice
however that if there is a condition from which we can find a particular
solution, we will then let the choice of the arbitrary constant decide on
the form of the particular solution.

Type 2. First-Order Linear Differential Equations.

Equations of degree 1 of this type will all be expressible in the form

H+ Play = Qe),
where P(x) and Q(z) are functions which involve z and not y. Examples
of this form are:
dy 3
a ;

d
= + ysinz = e??,

dy 2
at = ya int:
 Ch: 27 205

Sometimes it may be necessary to rearrange the terms in the equa-

tion, in order to put the equation in the form

Y + Play = Q(2),
where P(x) and Q(z) are functions which involve x and not y.
The method of solution is then as follows.
1) Find f P(x) dz.
2) Find the integrating factor Al alae
3) Multiply both sides of the differential equation by the in-
tegrating factor.
4) Write the left hand side of the differential equation in the
d
form om (y x Integrating Factor).
5) Integrate both sides with respect to z.
We shall see this method in action in the following examples.
Example 6. Solve the differential equation
d
ai! + 327y = 274 2°.
dz

Solution. In the notation given above, we have P(x) = 32”, and

so
[Pe aa = [xe dz =.

We are just trying to find any one of a number of possible integrat-

ing factors. Introducing an arbitrary constant in the above integration
would lead to multiplying the integrating factor we are going to use by
a constant. Since this is not going to affect the answer we obtain, we
shall, in this and all future examples, ignore the arbitrary constant when
finding f P(x) dz.
The integrating factor is e”. If we multiply both sides of the dif-
ferential equation by this integrating factor, we obtain

Recs + 322ye” = (x? + 2°) e”,

dr
which we can rewrite as

~ (e*") = ler aa z2) e*”,

The reason for choosing the integrating factor we did is so that the
left hand side of the differential equation can be written in the form
given. To check this, try differentiating ye™ with respect to z, using
the Product Rule.
If we integrate both sides of the above equation, we see that

ye” = i (ai - Zip) e? dr = / (1 + a) e® «dr.

 206 Ch: 27

In order to integrate the right hand side of this equation, we will

put u = 2°, so that du = 327dz. Then we need to find
sh
[30 +u)e"du = gue +c,

where we use Integration by Parts in order to integrate ue”. So we have

3 1 3
ye” =—-2%e" +¢,
3
or, on multiplying by e7 a as we have
1
y= sa? + ce-™
3

Example 7. Find the general solution of the differential equation

d
nae 262.
dz
and the particular solution for which y = 2 when rz = 1.
Solution. In order to put this differential equation into the
standard form, we need to divide both sides of the differential equation
by z. Then we have
d
mele caer
7 ars

Comparing this with the standard form, we have P(r) = st Hence

2
2
[Pde = f 2dr = 2102] = In|2*| = Ina’.

(As before, we do not need to include an arbitrary constant here, since

it would have no effect on the final answer we obtain.)
The integrating factor is

on using the connection between the logarithmic and exponential func-

tions.
We must now multiply the equation in standard form by the inte-
grating factor. This gives
dy
z PhFE
ak ae
Belt Kecyiat
ey ae
Garis
or
d
g?¥ + 2ry = 62°,
dr
which we express as
d
er (yx?) = 62°,
 Ch: 27 207

oe ae FG!
on noticing that the derivative of yz? is 22 SY + 2ry.
If we integrate both sides with respect to z, then we obtain

yr? = [sz°ae =r°+¢,

which, on dividing both sides by z?, yields

y= z+ +cxr~*

This is the general solution of the given differential equation.

For the required particular integral, we substitute the given values.
So we want y = 2 when z = 1. This gives 2 = 1+, from which c = 1.
Hence the required particular solution is
va es

Before seeing how to apply the techniques for solving differential

equations to various practical problems, we shall examine methods for
solving one final type of differential equation.
Type 3. Second-Order Linear Differential Equations with
Constant Coefficients.
An example of a differential equation of this form is the differential
equation
5 f’y dy
Wee 45 as — y== 5a°.

In general, a typical differential equation of the type we are going

to consider has the form
dy
ned7 tbs + cy=f(a),
where a, b and c are constants (with a # 0), and f(z) is a function
involving z alone.
The method for solving such differential equations can be broken
into three convenient stages.
1) Find the complementary function.
2) Find the particular integral.
3) Find the general solution, which is
y = complementary function + particular integral.

We shall now examine the methods involved in each of these steps,

applied to the general differential equation of the type under consider-
ation,
at
ay dy
ipo
“de? | dz + cy = f(z).
1) a) Write down the auxiliary equation,
az*+bz+c=0,
 208 Ch: 27

obtained by replacing a by 2? 2 by z, y by 1 and f(z) by 0.

1) b) Find the solutions of the auxiliary equation.
1) c) Write down the complementary function in a form which
will depend upon the solutions of the auxiliary equation found in b).
There are three possible cases to consider.
Case 1. The auxiliary equation has two distinct real solutions z;
and z2. In this case, the complementary function is y = Ae*!? + Be*?’,
where A and B are arbitrary constants.
Case 2. The auxiliary equation has two distinct complex conju-
gate solutions a +728. In this case, the complementary function is
y = e** (Acos Gz + Bsin Br), where A and B are arbitrary constants.
Case 3. The auxiliary equation has one (repeated) real solution 2.
In this case, the complementary function is y = (A + Br) e*!7, where A
and B are arbitrary constants.
Before proceeding to the next stage of our method of solution, we
shall work through stage 1 for several typical differential equations of
the type we are considering.
Example 8. Find the complementary functions for each of the
differential equations
d*y dy
a) aa
qa? et ev
70 6yino
= 1l5e Ee
d?y dy
By) a
qa? ae
oe + 10y== 250r? + 52sin 2z;

c) d’y
an? dy +4y = 4e°* on + 50sinz.
<\ta5

Solution. a) The auxiliary equation is 2z? + z —6 = 0, which

can be written as (2z — 3)(z +2) = 0, which has the two distinct real
solutions z = —2 and z = 3. Hence the complementary function is

y = Ae 2? + Be??,
where A and B are arbitrary constants.
b) The auxiliary equation is z? — 2z + 10 = 0, which has two dis-
tinct complex conjugate solutions z = 1437. Hence the complementary
function is
y =e” (Acos3z + Bsin3z),
where A and B are arbitrary constants.
c) The auxiliary equation is z? — 4z + 4 = 0, which can be writ-
ten as (z — 2)” = 0, having one repeated real root z = 2. Hence the
complementary function is
y = (A+ Br)e??
 Ch: 27 209

where A and B are arbitrary constants.

Before turning to the next stage in our solution—the finding of the

particular integral—we should note that the complementary function
contains two arbitrary constants. Since we would expect that solving
a second-order differential equation would involve two integrations and
therefore two arbitrary constants, it follows that we would not expect
our particular integral to contain any unknown constants.
Now we describe the routine for finding the particular integral for
the differential equation. The operation of this routine depends upon
both sides of the original equation. It is as follows.
2) a) Guess the correct form of the particular integral, depend-
ing upon the form of the function f(z) on the right hand side of the
differential equation.
2 b) Differentiate the particular integral you have guessed and
substitute it into the differential equation, so as to determine any un-
known constants from the requirement that the particular integral sat-
isfies the original differential equation.
In choosing the form of the particular integral, there are many
different cases that can arise. Amongst these are the following.
Case 1) If f(x) =0, then take y = 0 as the particular integral.
Case 2) If the function f(z) is a sum of terms then take a partic-
ular integral which is a sum of the particular integrals which arise from
each term.
Case 3) If f(x) is a polynomial in z, then try a polynomial of the
same degree. For example, try y = az? + br + c if f(z) is a polynomial
of degree 2, where a, b and c are constants to be found.
Case 4) If f(z) involves a sine or cosine term, then try a particular
integral involving both the corresponding sine and cosine terms. For
example, if f(x) involves sin5z, then try y = csin5z + dcos5z, where
c and d are constants to be found.
Case 5) If f(z) involves an exponential term, then try a similar
exponential term. For example, if f(z) includes e47, then try y = ce‘’,
where c is a constant to be found.
Note however that any term which occurs in the complementary
function cannot also occur in the particular integral. If our suggested
particular integral does contain a term we already have in the comple-
mentary function then we multiply the suggested term in the particular
integral by either z or z”, so as to get a different term. For example, if a
complementary function is y = Acos2z + Bsin2z and f(r) = 4cos2z,
then for a particular integral we would try y = crsin2z + drcos2z,
so as to avoid repeating terms already included in the complementary
function.
Probably the best way to understand the method is to work through
some examples.
 210 Ch: 27

Example 9. Find the particular integrals for each of the differ-

ential equations in Example 8.
Solution. a) The differential equation

2—~
+ — — 6y= 15e**
L

y = Ae ?* + Be2”,
where A and B are arbitrary constants. Since e?” does not appear in the
complementary function, it may appear in the particular integral, and
so as particular integral we shall try y = ce*”, where c is a constant we
wish to choose in order that the econo equation may be satisfied.
If 4 = ce**, then dy = 3ce** and oy = 9ce**, and so, on substi-
dr dz?
tuting, we need
2 [9ce**] + [3ce>*| —6 [ce**] =
From this we see that we need
(18c + 3c — 6c) e®* = 15e°*,
from sabich we need 15c = 15, or c= 1. Hence the particular integral is

i ) The differential equation

d’y d
Teak ae :
+ 10y = 2502? + 52sin2r

had complementary function

y =e” (Acos3z + Bsin3z),
where A and B are arbitrary constants. As sin2z and cos2z do not
occur in the complementary function, and neither does any power of z,
we shall try as our particular integral,
y=cr*+dzr+e4+fs+g
in2z
cos2z.
Differentiating, we have
d
= = 2cr + d+2f cos2zr — 2gsin2z,

and on differentiating again we have

d*y :
a? = 2c — 4f sin2z — 4g cos2z.

We now substitute these into the differential equation, and compare

corresponding terms, in order to determine the unknown constants. We
have
[2c — 4f sin 2x — 4g cos 2x] — 2[2czr + d + 2f cos2z — 2g sin 2z]
+ 10 [ex +dzr+e+ fsin2zr + gcos 2z| = 2502? + 52sin 2z.
 Ch: 27 211

Rearranging the terms we have

(10c) z? + (—4c + 10d) x + (10e — 2d + 2c) + (—4f + 4g + 10f) sin 2z
+ (—4g — 4f + 10g) cos 2x = 250x” + 52sin 2z.
From this, we need 10c = 250, —4c + 10d = 0, 10e — 2d + 2c = 0,
6f +4g = 52 and 6g —4f = 0. We can solve these equations and find
c= 25,d=10,e = —3, f = 6 and g = 4. Hence the required particular
integral is
y = 2527 4+ 10z — 3 + 6sin 2x + 4cos2z.

c) The differential equation

d’y dy 2z :
ee ee + 4y = 4e°* + 50sinz

had complementary function

y= (A+ Br) es

where A and B are arbitrary constants. Since e?? and re”? both occur
in the complementary function, we will need to try ze?” as part of the
particular integral. To correspond to the term 50sinz we will need to
try both asinz and acosz term. Hence we have as a particular integral

y = cr*e?* + dsinz + ecosz.

Differentiating with respect to z, we have
d :
©Y = Qcxr¢e?? + 2cx7e?* + dcosz — esinz,
dx
and, on differentiating again, we have

d?y
a2 = 2ce** 22 + 4cxre** 22 + 4cre** 22 + 4cr“e*”
73 Pde 2
— dsinz — ecosz.
x
We now substitute these into the differential equation, and we obtain
[2ce?* + 4cre?* + 4cxe?* + 4cx7e?* — dsin x— ecos x]
— 4 [2cre?* + 2cx*e** + deosz — esinz|
+4 ler e-® + dsinz + ecos z| ze [4e?* + 50 sin Z| ‘

which we can rewrite in the form

(2c) e?* + (4c + 4c — 8c) ze?® + (4c — 8c + 4c) x?e?*
+ (—d + 4e + 4d) sinz + (—e — 4d + 4e) cosz = 4e”* + 50sinz.
It follows from this that we need 2c = 4, 4c+ 4c — 8c = 0, 4c —8c+
4c = 0, 3d + 4e = 50 and 3e — 4d = 0. From these equations we find
that c = 2, d=6 and e = 8. Hence the particular integral is
y = 2x7e?* + 6sinz + 8cosz.
 232 Chi27

Notice that it was not an accident that the equations from com-
paring the ze?” and z?e?* terms both collapsed to 0 = 0. This must
happen in the situation we have here, where we have needed to include
extra powers of z in order to avoid a term which had already appeared
in the complementary function.
Stage 3 in the solution of these differential equations is the simplest
of all. We write down the general solution of the differential equation,
simply by adding together the complementary function and the partic-
ular integral that we have already found. We then find the values of
the arbitrary constants required to give a particular solution if we are
required to find a solution satisfying given boundary conditions.
Example 10. Find the general solutions of the differential equa-
tions occuring in Example 8, and find the particular solution of a) for
which y = 6 and oy = 0 when z = 0, and the particular solution of b)
for which y = 0 whente = 0 aud ae) 7iwhen ot oe
Solution. a) The differential equation
dy +
2—— | —dy — by = 15e**
EP er Belt ak yah
had complementary function
y = Ae~?* + Be??,
and particular integral
y= et

so that the general solution of the differential equation will be

y = Ae~?* + Be2™ + e°*.
To determine the required particular solution, we need to differen-
tiate y, and substitute the given values for y, x and dy: We have
dz
dy_ —2Ae~*7 4 “Bei + 3e37,
dz
and so we need 6 = A+ B+1 and 0 = —2A+$B+3. Solving these
simultaneous equations, we see that A = 3 and B = 2. Hence the
required particular solution is
y = 3e7?* 4 2¢27 + e3?,
b) The differential equation
dy dy
peed 2 + 10y= 2502? + 52 sin 2x

has complementary function

y =e" (Acos3z + Bsin3z)
 Ch: 27 213

and particular integral

y = 252” + 10z — 3 + 6sin 2zr + 4cos2z.
Hence the general solution of the differential equation is
y =e” (Acos3z + Bsin3z) + 252? + 10zr —3 + 6sin 2z + 4 cos 2z.
To find the particular solution which satisfies the boundary condi-
tions we are given, we substitute these values into the formula for y. We
have 0 = A —3+4, from which A = —1, and also

5a —T=e3 a (—B) +25 (=)

1 \2
+1057 =3-4,
from which
Bet Bu 257
2

4
or
to. 2572
~ det
Hence the particular solution which satisfies the boundary condi-
tions we were given is
251°
yisve” (-cos 3z + (3us )sin) +252? +102—3+6
sin 2r +4 cos 2z.
e2

c) The differential equation

d?y dy or .
Wed Fa + 4y = 4e°* + 50sinz

has complementary function

y = (A+ Br) e**,

and particular integral
y = 2r7e** + 6sinz + 8cosz.
Hence the general solution of the differential equation is
y = (A+ Br) e?* + 227e?* + 6sinz + 8cosz.

EXAMPLES 27
1. The acceleration of a particle at time t is given by the equa-
@? d
tion — = 2cost + 6t — 2e~*', where x denotes the displacement of the
particle from the origin. Find an expression for z in terms of t, given
that, at time t = 0, the particle is at the origin and is at rest.
 214 Ch: 27

2. Find the general solution of each of the following first-order

separable differential equations:

(a) Ya ay? (b) aay ty)

@ 2-24 w B-ea+y
() Bast y Bae,
(g) ae (h) Srrrert
3. For each of the following first-order separable differential equa-
tions, find the general solution and also the particular solution satisfying
the given boundary condition:

(a) a sin
xsec y; y= 5 when r= 5;

(b) a = 27(1+y); y=0 when ci

L
(c) 2 = ry’; y=-2 when eS;
dy cos2z
d — = : =— 2 h —f,
le de aay EX GCa\ aes ous
4. By putting y = vz, where v is a function of z, find the general
solution of each of the following first-order differential equations of the
homogeneous type:
dy 4. dy _ F
(a) a(z+y)o ="; (b) (@-y)S = ety
dy _ Quest dy _ y
(c) dx (x? + 2y?)’ "de aes (¢) ;
5. For each of the following first-order linear differential equations,
find the general solution and also the particular solution satisfying the
given boundary condition:
d
(a) = hey = ; y=3 when 2c=0;
dy 5
(b) i a : y=4 when z= 2;
d
(c) = + 4y = ze~**; y=3 when .2=0;

(d)
dy
de + ¥ = Sin2e; y=4 when cx=0.
 Ch: 27 215

6. For each of the following second-order linear differential equa-

tions having constant coefficients, find
(a) the complementary function,
(G) the particular integral,
(y) the general solution,
and, where appropriate,
(6) the particular solution satisfying the given initial conditions:

(a) dy
a dy ae 182? — 182
— 28;

(b) CY 4 Y _ Gy = 22 cos 2x — 6sin22;

3 d d
(c) ob 4 —5y = 90"; y =4 and 5" =5 when x= 0;

(d) dy 9y = —9r? + 9x? + 152 — 20;

Sota
r
dy 4!dy 4 By S192,
SF
‘2 ax oa :
aed aie dy
(f) aor t ae v= 90°; y= 2 and — =5 when x = 0;
2
(g) CY 4 4ay = 40? — 80 + 14
2 d
(h) oY 46H 4 oy = 90? — 62 + 17;
wie
(i)
¥ ing cy
aE ipmmace pa aed2 y=eand dy = 6e when ¢ at
— = 1,
2
(3) ou 4 oy = 136-%;
d’y
(k) an? + 16y = 36sin 2x — 12 cos 2z;

(1) dy ~ tgp
qe? dy t Sy = 5a" — 3a — 12;

(m) dy 27,
7 dy + ly — Ye”;ae
———._=)
dy
y —= 3 and —Fe = 0 whenx = 0);

2
(n) EY _ 4% 4 by = 82? — 82 — 22;
z L
(0) FY 4X 4 5y = 62-114 20%;

(p) Aed ee=

cal
—L£r

dx? dz
 216

CHAPTER 28
Applications of Differential Equations
In this Chapter, we will use our ability to solve different types of
differential equations in order to solve some ‘practical’ problems.
Example 1. Find the curve in the z, y-plane which passes through
the point (3,1) and whose tangent at the point (x, y) has gradient z*y?.
Solution. The slope of the tangent to the curve at the point
d
(z,y) is given by = Hence we are given that
£
dy De
gets es
This differential equation is of the separable type, and so we solve
it by separating the variables, and then finding the particular solution
which satisfies the given condition.
Separating the variables and integrating gives

[S- [ee
y
from which

=Ye = =a +c.

We require y = 1 when z = 3, from which c = —10. Hence the required

curve has equation

One of the most common types of growth and decay is that covered
by the laws of exponential growth and exponential decay. This
happens in those situations where the rate of change of a quantity at
any given time is directly proportional to the value of the quantity
at that time. This happens for example where the rate of change of
a population is directly proportional to the size of the population, or
where the rate of growth through interest of money deposited in the
bank is directly proportional to the amount of money deposited.
To see that such situations lead to the laws of exponential growth
and decay, let us consider the following example.
Example 2. Let y denote the value of a quantity present at time
t. Assume that the rate of change of y is directly proportional to the
value ofy at time t. Find an expression for y in terms oft.
Solution. We know that the rate of change of y with respect to
Eee yer: é
t, which is given by is directly proportional to the value of y. This
implies that
dy
HTRY
 Ch: 28 217

where k is the constant of proportionality. This differential equation is

separable, and on rearranging the terms we have

[te [ee
y
from which
Iny
= kt+c,

where c is an arbitrary constant.

If we apply the exponential function to both sides of this equation,
then we have
ely = ektte

or, using the index laws,

from which

y= Aer’,
writing A = e°, where A will be an arbitrary constant. This is the
well-known rule giving exponential growth and decay.
A gives the value of y at time t = 0, while the constant k determines
whether y is increasing (when k > 0), or decreasing (when k < 0).
Also the size of k determines how rapidly or slowly y is increasing or
decreasing.

Newton’s law of cooling states that the rate at which a body

cools is directly proportional to the difference in temperatures between
the body and the surrounding medium. The following example shows
that this leads to a modified form of exponential decay.

Example 3. A cake is mixed at room temperature of 75°F. It

is then placed in an oven which has been heated to 375°F. Assuming
that the oven maintains its temperature throughout the 20 minutes that
the cake is baking, find the temperature of the cake when it is removed
from the oven. The temperature of the cake is known to be 275°F two
minutes after it is placed in the oven.
Solution. Let T°F be the temperature of the cake ¢ minutes
after it is placed in the oven. Then we are given that T = 75 at t = 0,
and that T = 275 when t = 2. We need to determine the value of T
when t = 20. La ;
We may assume that —,, the rate of change of the temperature of
the cake, is directly proportional to 375 — T, the difference in tempera-
ture between the cake and the air in the oven which is surrounding it.
This is a modified form of Newton’s law of cooling which will apply to
things which are getting hotter.
 218 Ch: 28

Hence we know that

dT CS
a
pairs =k (375 _T -T)
from which, on separating the variables, we have
dT
—— = | kdt
/ 375 —T /
so that, on finding the integrals, we have
—1n(375 —T) =kt+c,
which we can rewrite as
315 Ae,
if we let A=e°.
We now substitute the two pieces of information that we were given
into this equation in order to find k and A.
When t = 0, we have T = 75, and so, on substituting we find
375 — 75 = Ae~*°, from which we have A = 300.
When t = 2, we have T = 275, and so, on substituting we find
375 — 275 = Ae~**, from which we have 100 = 300e~?*. Using the
index laws, we can rewrite this as

(eae = ; from which e~* = (

At time t = 20, the temperature T is given by the equation

375 — T = 300e~?°*,
which we can rewrite as
T = 375 —300(e*)”’,
Or as
; 20
T = 375
— 300 (V3) ;

which is equivalent to
1 10

T = 375 — 300 (5)

or, on using a calulator,
T & 374.99.
It follows that, in the circumstances described, the cake has reached a
temperature of 374.99°F after 20 minutes in an oven maintained at a
constant temperature of 375°F.
It is worth noting that in the above example we avoided finding k,
managing instead to use the index laws so that we could work with e~*.
The advantage of this is that it cuts out one step from the solution,
and also removes a potential source of inaccuracy, arising from finding
 Ch: 28 219

a (necessarily approximate) value for k. However, there would be no

objection to a solution that did find the value of k as a step towards
finding T.
Before the example, it was described as a modification of exponen-
tial growth. The way in which it can be regarded as being an example
of exponential growth is that if we measure temperatures from a base of
375°F rather than from the base of 0°F then we would have exponential
growth of the temperature from the relative value —300° towards the
value 0°.

Example 4. A body is falling under the action of gravity through

a resisting medium which applies a retarding force to the movement
which is related to the square of the velocity of the body. Assuming
that the body starts from rest at time t = 0, and that the velocity v at
timet satisfies the differential equation
dv 1 5
ae = 6 (81 <i ))

find the velocity of the body at time t = 0.5.

Solution. We can separate the variables in this differential
equation and rewrite it as
dv 1
——, = — / dt
la —v? =f
so that, using Partial Fractions on the left hand side, we have
i 1 1 1
— —— + —— }dv=-— | dt
alles ay i ul
and on integrating both sides we have
1 1
jg (in le + 9] — In|9 — vf) = git ¢
We know that v = 0 when ¢ = 0 and so on substituting we have

= (inlo+9 Sag He 0+,

from which we must have c = 0.
Assuming that the velocity v will always stay below the value 9,
we have |9 — v| = 9 — v, while from the given differential equation, if v
always stays below 9 then the acceleration of the body will stay positive,
and so in particular the velocity of the body will stay positive. Hence
also we may write |9 + v| = 9+. We can therefore rewrite the above
equation as

In(9 + v) — In(9 — v) = 3t,

which we can put in the alternative form
 220 Ch: 28

or
9+vu 3t
Oates
Solving this equation for v in terms of t, we have
e®4
—
»=9(S55)-
Substituting t = 0.5 into this equation, to find the velocity of the
body at t = 0.5, and using a calculator, we have

correct to 3 decimal places.

Example 5. The current «x in a certain electrical circuit at time
t is given by the differential equation
dz
pt OF slg cons:

Find z in terms of t, given that x = 0 at time t = 0.

Solution. The given differential equation is a first order lin-
ear differential equation, which we shall solve by the integrating factor
method. Now the integrating factor will be

ed eae
Multiplying the differential equation by the integrating factor, we
have
dz
e*'— + 5re** = e** — e®* cost,
dt
which we can rewrite in the form

di (xe°*) =e?! — 6 cost.

Integrating both sides of this equation with respect to t gives

1
ret = Boe a Je cost dt.

To complete the solution, we need to find I = f e*' cost dt. Earlier

we found an integral like this one by using complex numbers. This time,
for comparison, we shall use Integration by Parts twice. We have

T= eMsint—5 fe sinedt

=e sint —5 |e cost +5 fe!costa

= e*' sint + 5e*' cost — 25 if

e°' cost dt

= e*' sint + 5e*' cost — 251 — 26c.

 Ch: 28 221

Solving this equation to find I gives

26] = e*' sint + 5e*' cost — 26c,
from which
1
I
~ 26 (e** sint + 5e°! cos t) — Cc.

Substituting back into the equation, we have

1 1 ?
Ceo Se oe (ext sint + 5e°! cos t) +c,
5 26
We now need to substitute the initial values r = 0 and t = 0 in
order to find the value of c. Then we have 0 = ; - 4 +c, from which
we have c = — 50° Hence we conclude that

1 1
Ag ee (sint + 5cost) — ao
or :
ci 130 (26 — 5sint — 25cost — e-**) :

Example 6. A weight is attached to the lower end of a spring, the

upper end of which is attached to the ceiling. Let ycm denote the ver-
tical distance of the weight from the ceiling at time t sec. Suppose that
the relation between y and t is determined by the differential equation

d? d
atl
dt2
Oa
dt
+ 41ly = 410.
Find y in terms of t, given that at time t = 0 the weight is at rest and
is a distance of 20cm below the ceiling.
Solution. For this second order linear differential equation with
constant coefficients, the method of
solution involves finding a comple-
mentary function and a particular in-
tegral. Now the auxiliary equation is
z?+10z+41 = 0, which will have the
two complex solutions z = —5 + 41.
Hence the complementary function
will be
y =e °'(Acos4t + Bsin4t).
As a particular integral we shall
take y = c, where c is a constant.
d ok
Then we know that te Ons ee
and so we need 41c = 410, from which we have c = 10. Hence the
particular integral is y = 10.
 222 €h:228

The general solution of the differential equation is therefore

y=10+e °*(Acos4t + Bsin4t).
We are looking for the particular solution which satisfies the initial
conditions y = 20 and = = 0 at t = 0, since we know that the distance
from the ceiling is 20cm and that the weight has velocity 0 at time
Lo:
Substituting y = 20 at t = 0 we have
20=10+A, sothat A=10.
d
We differentiate in order to find cs We have

4 = —5e~°' (Acos4t + Bsin4t) + e~°' (—4A sin 4t + 4B cos4t).

d
Substituting = 0 at t = 0 we have

25
0=-5A+4B, sothat B= oe

We finally conclude that the required solution of the differential

equation is
25
y=10+e (10cos 4t + in sin4t) F

This solution gives us the result that experience tells us would hap-
pen. We start with the spring and weight stretched beyond their equi-
librium position. When the weight is released at time t = 0, it will
oscillate about its equilibrium position, with the magnitude of the os-
cillations gradually reducing, until finally the weight will appear to be
motionless at its equilibrium position. This is precisely the behaviour
of the function we have found in the above solution.
EXAMPLES 28
1. Find the equation of the curve in the z,y-plane which passes
through the point (2,1) and whose tangent at the point (z,y) has gra-
dient x*y.

2. Find the equation of the curve in the z,y-plane which passes

through the point (1,0) and whose tangent at the point (z,y) has gra-
; mete
dient A
zy
3. <A joint of meat has been cooked in an oven at 400°F for a
long time, so that the whole joint has reached this temperature. It
is then removed from the oven and carved in the kitchen where the
temperature is a constant 70°F. Given that it takes the housewife two
 Ch: 28 223

minutes to carve the meat, and that when she has finished carving it the
temperature of the meat has dropped to 300°F, how long can the family
delay further before coming to the table, if the housewife likes to present
their meals ‘piping hot’, which in this case means at a temperature of
at least 150°F?
4. In a room maintained at a constant temperature of 20°C, a
heated body cools from 100°C to 70°C in 15 minutes. How long will it
take the body to cool to 40°C?
5. A solution of chemical A reacts with a solution of chemical B.
If x is the concentration of A which has reacted in time t, and K is a
constant determining the rate of the reaction, then

— = K(0.01 — z)(0.02 — 2).

Given that z = 0 when t = 0, find the particular solution of this differ-
ential equation.

6. The radioactive decay of Uranium 238 is a complicated process,

part of which is governed by the differential equation
dy
—+Ky=
dt + Ky Ale",
e€

where K, A and L are constants with A # L. Given that y = 0 when

t = 0, solve this differential equation to find an expression for y in terms
of t.
7. A reservoir contains a volume V of water, which flows into
it and out of it at a constant rate v. Suppose that c measures the
concentration of pollution in the water flowing into the reservoir, and
that y measures the concentration of pollution in the water flowing out
of the reservoir.
Given that the reservoir is initially not polluted, so that y = 0 when
t = 0, and that c is a constant, find an expression for y in terms of ft,
given that they are related by the differential equation
dy vy ve
d—V—_
8. If glucose is being fed intravenously at a constant rate, the con-
centration y of glucose in the blood at time t is given by the differential
equation
dy
—=C-—ky,
dt g
where C and k are positive constants. Solve this differential equation
to find y in terms of t, given that at time t = 0, y = A, where A is a
constant.
 224 Ch: 28

9. The current 7 at time ¢ in a certain electrical circuit is given by

the equation

where R, C are constants. Find 7 in terms of t if 1 = zo at time t = 0

and E = Eo(1—coswt), where Eo and w are constants.
10. The equation of motion of a certain body is
d
ee +4y =e”,

If v = 0 at time t = 0, find an expression for v in terms of t.

11. In a certain galvanometer, the deflection 6 satisfies the differ-
ential equation
d?@ dé
de pricey
Sat —— 154 6 = 2e 2e7**,

Given that 6 = 0 and bad= 0 when ¢ = 0, find an expression for @ in

dt
terms of t.

12. A particle is performing as a damped harmonic oscillator, sub-

ject to the differential equation
dr dx
ids tae

Find an expression for z in terms of t, given that r = 0 and = =8

when t = 0.
 225

CHAPTER 29
Basic Techniques
In some of the earlier chapters, we promised to deal with certain
further details which we put off then so as not to disrupt the continuity
of the course. This chapter is divided into various unconnected sections,
each of which deals with one of the topics mentioned earlier in the book.
Section 1. Factorisation.
Given a quadratic expression such as x? — 7z + 6, we frequently
wish to write it as a product of two factors. The reason for wishing to
do this is that if we can write
x? —72+6=
(zr —6)(z—1),
as we can, then we can find the roots of the equation z? — 7z + 6 = 0,
since (z — 6)(x — 1) = 0 precisely when either zr -6 = 0 or r—-1=0.
Hence the solutions of the equation z? — 7z + 6 = 0 are the two real
numbers':a = 6 and ¢ = 1.
It follows that factorisation can be a help on the way to finding the
solutions of an equation.
The reverse is also true. If we know that a is a root of an equation
of the form f(z) = 0, where f(z) is a polynomial in z, then it can be
shown that z — a will be a factor of the polynomial f(z).
There is a result which says that if we can factorise a polynomial
which has integers as its coefficients, then all of the numbers involved
must either be integers or real numbers which are not rational. Sup-
pose that f(z) is a polynomial having integer coefficients which can be
factored into linear factors not involving irrational numbers. Then the
implication of the above result is that all the coefficients in the factori-
sation can be assumed to be integers. Furthermore the numbers at the
beginning and end of each bracket must divide the coefficients of the
highest and lowest powers of z involved in the original polynomial.
Suppose then that we are given a quadratic polynomial such as
z* + x — 6, and we are asked to factorise this quadratic as a product of
two linear factors. By the result quoted above, we must try to write
z?+2—6=(azr+b)(cx+d),
where a and c divide 1, and where b and d divide —6, with a, 6, c and
d being integers. We might as well take it that a = 1 and c= 1. What
are the possibilities for b and d?
We can choose 6 = +6, b= +3, b= +2 and b= +1. These are the
only integers which divide —6. Once 6 is determined, the value of d will
be fixed because we need bd = —6. So for example, if b = —2, we would
have to have d = 3.
Now on expanding (z + 6) (x + d), we have
(zc + b)(x+d)
= 27 +(b+d)2+bd,
 226 Ch: 29

and so we need to choose b and d with bd = —6 and b+d = 1. We want

b and d to be roughly the same in size in order that b+ d= 1. This
suggests that b = +2 or b = +3 might be more suitable than 6 = +1
(which means d = +6), or b = +6 (which means d = F1).
We now use trial-and-error to try to find the appropriate values for
b and d. If we take b = 2 then d = —3, so that b+ d = —1, which is not
what is wanted. If we take b = —2 then d = 3 so that 6+ d=1, which
is what is wanted. Hence we have the factorisation
z?4+2—6=(z—2)(x
+3).
Let us now try a slightly more difficult example. Suppose we wish
to factorise 327 — 2x — 8. We shall assume that
32* — 2r —8 =(ax+b)(czr
+d),
where we shall assume that a, b, c and d are integers. (By the theorem
quoted earlier, if these numbers are rational numbers, then they must
in fact be integers.)
Now if the identity above holds, then we must have ac = 3, with
a and c being integers. Hence one of the numbers must be +3 and the
other must be +1. We shall assume that a = 3 and c= 1.
Next we need to choose 6 and d such that bd = —8 and also such
that ad + bc = —2. b and d must be factors of —8, if they are integers,
and so we must have either b = +1, b = +2, b= +4 or b = +8, with
d = —8/b.
Remembering that we have already chosen a = 3 and c = 1, we
use trial-and-error to find that b = 4 and d = —2. Hence we have the
factorisation
32? —2r —8 = (32 + 4)(x— 2).
It is important to check at the end of examples !ike the above that
the factorisation you have written down is correct. It is all too common
to see students guess at what they think is a correct factorisation, and
proceed to work on carrying a mistake, because their guess was wrong
and they omitted to check their guess properly.
Section 2. Right Angled Triangles.
In the first diagram, O is the centre of a circle of radius r, and
A is the point where the circle cuts the z-axis. P is the point on the
circumference of the circle chosen so that the angle AOP is =. X is the
foot of the perpendicular from P onto the z-axis. Since the pane of a
triangle add up to 7 (radians), and since the angle PXO is$, it follows
that the angle OPX will be $. P’ is the reflection of P in the Z-axis.
By congruent triangles, we ea see that the angle OP’X will also be
Furthermore the angle P’OP will also be =. Hence the triangle OPP"
will be an equilateral triangle (one with allsides having the same size
and with all angles being equal).
 Now the length of OP is the ra-
dius of the circle, which we took to be
r. Hence PP’ will have length r also,
and so, by symmetry, PX will have
length 5 (since X is the mid-point of
PP’). Using Pythagoras’ theorem on
the right-angled triangle OPX, we
see that

|OX|? + |XP/’ = |OP/,

so that

loxP 2 + (2) =r,

Le 2 B23

from which we conclude that

|OX| = VB

In the second diagram, O, A and r are as defined above. This

time the point P is chosen so that
the angle AOP is 7. As before, X
will be the foot of the perpendicular
from P onto OA. Hence the angle
PXO will be } and so we see that
the angle OPX will also be {.
The triangle OPX is an isosce-
les triangle, which is one having two
sides of equal length, since the sides
opposite to the equal angles OPX
and XOP must be equal. We now
apply Pythagoras’ Theorem to the triangle OPX, remembering that
|OP| =r, and that |PX| = |OX|.
We easily see that

The third diagram is similar to

the first, except that A, the point
where the circle meets the z-axis, is
also the reflection of O in the line
PX, and the angle XOP is =. Then
the triangle OPA is equilateral, and
 228 Ch: 29

by a similar argument to that used earlier we conclude that |OP| = r,

|OX| = £ and |PX| = #r.
Now we use the definitions of sine, cosine, and tangent given in
Chapter 2 to obtain the values given in the table there.
Section 3. The Addition Formulae.
We shall prove the addition formulae only in the case where z and y
are positive angles whose sum is less
than a right-angle. We refer to the
diagram shown.
In this diagram, the line OCE
is chosen such that the angle AOC
is x radians, while OP is chosen such
that the angle COP is y radians. Let
P be chosen so that |OP| = r, and
let C' be the foot of the perpendicular
\
s
:Z

from P onto the line OEC, while A

and B are the feet of the perpendic-
ulars from P and C respectively onto
O A B
the line OAB. D is the foot of the
perpendicular from C onto AP, while E is the point of intersection of
the lines OC and AP.
Since the lines OB and C’D are parallel, both being perpendicular
to AP, it follows that the angles AOC and ECD are equal. Hence
ECD=n,.
From the right-angled triangle CDE we can deduce that the angle
CED is } —z. Now, by considering the right-angled triangle ECP, we
see that the angle CPD is z.
Now we consider the right-angled triangle OCP, in which |OP| =r
and the angle COP is y. Applying the definitions of sine and cosine to
this triangle, we can deduce that
ICP|
|OP|
cosy = Mee and siny
aloe
= ——

from which we can see that |OC| = rcosy and that |CP| =rsiny.
Next we consider the right-angled triangle OBC in which the angle
BOC is z and |OC| = rcosy. From this triangle, we can deduce that

from which we have |OB| = r cosy cosz and |BC| = rcosysinz.

The next right-angled triangle we consider is the triangle DPC in
which the angle DPC is z and |PC| = rsiny. From this triangle, we
can deduce that
 Ch: 29 229

from which we deduce that |PD| = rsinycosz and |DC| =rsinysinz.

Notice that ABCD is a rectangle, and so in particular we know that
|AD| = |BC| = rcosysinz, and also that |AB| = |DC|=rsinysinz.
Now we are ready to calculate sin(z + y) and cos(z + y). We have

cos(z + y) = |OA|
|OP|
from which

cos(x + y) = oA

rcosycosz —rsinysinz
r
Il cosy cos Z — sinysinz,

as required. Also
: AP
sin(z + y) = on

_ |AD| + |DP|
|OP|
_ reosysinz +rsinycosz
r
= cosysinz + sinycosz.
The extension of these results to numbers z and y which do not
satisfy the conditions given above follows by consideration of a number
of special cases, none of which is particularly difficult, using formulae
such as
sin(a + 2) = —sinz.

Although it was suggested earlier in the book that formulae such

as this could be deduced from the addition formulae, they can also
be readily deduced directly from the definitions of the sine and cosine
function that were given. This avoids the need to use the addition
formulae in trying to prove the addition formulae, which would of course
lead to a circular argument.
Section 4. The Sine Rule.
Suppose we are given a triangle ABC in which the lengths of the
sides opposite to A, B and C are a, b A
and c respectively. Let the angles in
the triangle ABC at the vertices be
A, B and C respectively. Then the
Sine Rule states that
a b c B C
snA snB_ sinC’ x
 230 Ch: 29

To prove the Sine Rule, we consider the diagram shown, where AX

is the perpendicular from A onto the line BC, extended if necessary.
Then
|AX| _ |AXx|
sinB = and sinC = [AC]
|AB|
From these two equations we have

|AX|=csinB and |AX|= bsinC.

Equating these expressions for |AX|, and dividing by sin B sinC’ we

conclude that
Ga
sinB sinC’
By dropping the perpendicular from B onto AC’, we can conclude
similarly that
Gn ft SG
snA sinC’
The Sine Rule follows immediately from these two equations.

Section 5. The Product and Quotient Rules.

Let f(x) and g(x) be differentiable functions, and let us consider
the product fg, which is defined by (fg)(z) = f(x) x g(x). Suppose
that h # 0. Consider

(f(z +h) — (faz) _ fle +h)g(z +h) — f(z)g9(z)

h h
_ fle t+h)g(z +h) — f(z + h)g(z) + f(z + h)g(2) — f(z)g(z)
h
ails tih)ole ot b) al ee ated alot miesdt alate)
h h
tien (“ a . a a) fortes (Es aoe=H) ,

Now we assume that h — 0. Since f(z) is a differentiable function

of z, f(x) will be continuous. Hence it follows that f(z +h) — f(z) as
h + 0. Furthermore, since f and g are differentiable functions of z, we
know that, as h — 0,

Sa h)— fle) Weds giz+h)—g(z) dg

h Cdeuens h widdes
From these results we conclude that, as h — 0,

Gas zr+h)—-
+ N= (Ne) “77.) an anal
|

as required.
 Ch: 29 231

Now we consider the differentiable functions f(z) and g(x), and we

(0-8
consider the quotient f which is defined to be the function such that
g

where we are assuming that g(x) #40. Suppose that h # 0, and that h
is sufficiently small so that g is not zero between z and x + h. Consider
f f fe+h) —f(x)
ie)ie) (5)(#) _ oath) gz)
h h
_ f(x t+h)g(z) — f(z)g(e + h)
hg(z + h)g(z)
_ f(z +h)g(z) — f(x)g(2) + f(z)9(z) — F(z)g(z +h)
hg(x + h)g(z)
SO eee eae)
hg(z + h)g(z) hg(z + h)g(z)
a) (LEAM AION jpn (SEED)
g(x + h)g(x)
As in the last proof, as h — 0, we know that g(x +h) — g(x), and
also that
f(zct+h)-f(z)
h
_ ae
_
of and
g(z+h)-g(z)
Fy —) Fe
dg
Hence we conclude that, as h — 0,

(£) (ah) = (£) (x) vfg(x) x ¢ — f(z) x 2

h [9(x)]”
’

as required.
Section 6. Long Division of Polynomials.
Earlier on, we saw that the technique of long division of polynomials
can be very helpful. The method is very similar to the method of long
division of numbers, which sadly is no longer as familiar as it was, in
these days when calculators are so common.
Suppose we are required to divide 2° + 524 + 52° + 32 + 2 by
x? + 322 —22 +4. We start by setting up the familiar long division
format. In order to keep corresponding terms in the correct positions,
we put the term Oz? into the dividend. As we calculate, we concentrate
on the highest powered terms in the divisor and dividend.

goeteere
— Or + 4) a> + -5e* 4 h2* + 02? £<S2l+ 2
 232 Ch: 29

Since z® divided by z° is x”, we put z? in the quotient and we multi-

ply each term of the divisor by z?, writing the answer in the appropriate
columns under the dividend. So we have
x?

g = 27 + 4a
£37? Sart te Or? i Or" 479242
po iol Spt. Det te Age

Now we subtract the result of the multiplication from the dividend,

and then bring down the term involving the next power of r. So we
have
x?

a? + 32? — 22 4+ 4| 2 405244 Sr?) 027)4+ 32 + 2

2” 4+ 324 —\ 22?) 42?
2244+ 7x2 —) 42? + 32
Now we repeat the above process. We start by dividing 22* by
z*, so that we have 2x which we put into the quotient and then we
multiply the divisor by this 27, subtracting the result as before, and
finally bringing down the next term.
zc? 4 Qe
ge + 327 — 27 44 | 220+ 5x* + 52° + Or? 4°32 4.2
z> + 324 — 27° 4 42?
224 + Tx* — 427 + 32
2z* + 623 — 427 + 8
a* + Orc? — 52 + 2
Finally we repeat the process once more, when we will reach the
point where the degree of the remainder is less than the degree of the
divisor. At this point we could only continue by introducing negative
powers of z. For some purposes this is meaningful, but here we shall just
leave the solution with a quotient and remainder which are polynomials
in z. We have
oe ore tl
z+ 3a? — Qe 4 4 |e? +) 524, 4 52304022,
+. 3a 492
¢° chy da ae2s ie. 427
2c4 + 7x? — 42? + 32
2c* + 623 — 42? + 8
z* + Oz? — 52 + 2
TL 32°) S00; tn 4
— 327 — 3r¢ — 2
From this we conclude that, when 2° + 524 +52? +3242 is divided
by x? + 32? — 22 + 4, the quotient is x? + 27 + 1 and the remainder is
—32z? — 32 — 2.
 Ch: 29 233

EXAMPLES 29
1. Factorise each of the following as a product of two real linear
terms, where possible:
(a) 2?+42r—5; (b) 2?+4+52r+4;
(c) 27+452+6; (d) 2%?4+2r—-12;
(e) 227+ 52 +3; (f) 3274+13244;
(g) 227+ 72 +13; (h) 32? —-32 +412;
(i) 42? —-8r+3; (j) 327+52—-2;
(k) 327442 +2; (1) 32?+4+4r+1.
2. In each of the triangles ABC described below, use the Sine
Rule to determine those of the lengths a, b and c of the sides and the
angles A, B and C which are not given:
(a) = A= 50°3C. =o anda = 3;
(b) A= 40°, a=3 and c= 4 with 0° < C° < 90°;
(¢)"s A= 30°, b= 2 and a=7 with 90°<-C° < 180°.
 234

Numerical Answers to the Examples

This section provides numerical answers to the examples in the

book, together with hints for the most difficult examples.
Examples 1.
3. (a) (x—3)? —2; (b) (x +2)? —8; (c) 3(x—1)* +4; (d) —2(z—-1)? 411.
4. (a) y =(z — 2)? —4; (b) y= (rz — 1)? —4. 5. (a), (d) No real roots;
(b) 2 distinct real roots —4, 1/3; (c) 2 distinct real roots —1/2, 3.
6. (a) y = (x — 2)? —1; (b) y= 4—-(2 +1)3 (Cc) y= 12 - 3(@ — 2);
(d) y = 16—4(x —1)?. 8. (—oo, -2) U (1,3). 9. [—4, -1] U [3, 00).
10. (—2,3) U (3, 00). 11. (—oo, —1] U [5, 00).
Examples 2.

16 (9) Bs Be? 4,, V3; (b) Je, Je, 1, V2, V2, 1; (c) 8,3,

V3, =e, 2, Paid) AsO ND adoND Os (Ot ines-2,

$2, -V3;
(f) — —3,-
i, V3”
3 —2,- Fi (g) avs
= —i, V3, SR
— E=2) Fi (h) 0,

—1,0,ND, —1, ND; (i)0,-1, 0, ND, -1, ND; (j) -¥, -4, v3, -4,
—2, Yx; (k) -1,-4 1
a ae age (1) — Fy i=—/2, - 2,

1; (m) —1, 0, ND, =I, ND, 0; (c= von ae ak; -V2, V2, ="hs (0) =,
A -4,,-4,
0, ND, -1, ND, 0. 2. -¥8, 1 2 2,v3. 3. -3,-4, -§,-§,4.
AW A) 2%,0%/3, 25D) 27/3, 0.0; (¢) +/2, 0;\(d) 2x /3ie%/4, 3225218) m.
6. 8.9 km. 7.6.8 cm. 8. 79.9 ft. 9. 409.5 m. 10. a verve
— J3. 11, Vetv2, Va=V6 _9 _ V3. 12. 1/2- 1/a4 V5
/2—1. 19. 2Asin(t/T) cos(x/X).
Examples 3.
eye V2sin (4a + 2), ads Fae: i= 17sin(2z
— 6),
17, 0.54, 7. 3. 3. 4. 1/V1+22. 5. Jl—2?. 6. 1V¥5—42
=2,
eo) Neer: ie SO vaaerererres‘s/o
8. (a) ¥; (b) —F; (c) $ (d) . 9. (a) 2; (b) -. 10. (a) 2 =
2katF (k € Z); (b)z = kr+ 4 ore = kr+5" (k € Z); (c)a = *4_ 18
=
or z = *4™ 4 It (k € Z); (je=¥2-Zorrza ty 7(k € Z).
11. (a) ¢ = 4® +2 org =kr—forz = kr+ 22 (k € Z); omens
t= kn+§(k € Z); (c) c = 2kw or
z = 2kn+2(k € Z); (d)z= kn+%
or z =kn+%(k€ Z). 12. 91.2°, 1.2°. 13. 18°.
Examples 4.
1. (a) 3524 — 162° + 242” — 10x + 2; (b) 2024 — 152? + 1427 — 8;
eA = 4 12 ; 94.2 4 4 20
(c) 182 6z + 5 am aa 7? (d) 362 247° + 3 — oe + = ae
 Answers 235

2. (a) 7x® — 302° + 242° + 9x? — 542 + 14; (b) 902° + 22024 +
13223 + 219%? + 1247 + 57; (c) (x3 +42) cosz + (32? + 4) sing;
iA 2 2
(a) (1024 — 272?)Inz+2c4-—92?. 3. (a) Se Ele
(x? + 9)
(b) a(2 bs 2) (2 n 2) /(2? es +4)?; (c) 2cos£ +t 3 COsz i520 =:
; 32 UEe
(2+ 2%)
(d) (32? — 5 - 62? Inz)/z (3x? — 5)’. 4. (a)7 (2x? — 32? — 52 + 2) x
(x4 — 273 — 52? + 42 — 7)?; (b) (5/4)
x (82° ~ 42° — Tx? + 32 +2)* x
(1524 — 12x? — 14x + 3); (c) 3sec? x tan? z; (d) 5sec® rtanz; (e) (4 —
10z) sin (5x? — 42 + 7); (f) sec? (2? — 42 + 2) cos (tan (2? — 42 + 2)) x
(22 — 4). 5. (a) 2(7x — 3)(x — 5)"(z + 3)°; (b) 42(32 + 4)°(7r — 2)? x
(52 +1); (c) 3cos? x cos 42; (d) 4sin® x cos 5z; (e) 6sin® z cos? 2z cos 32;
(f) 2sec2zrcosec? x(tan2z —cotz); (g) (3sec? /ztan? /z) /(2V/z);
(h) 32 ¥ sine’; (i) —423 cosecr cot x3; (j) —3tanzsec? z x
(1—tan?z)?; (k) —2?/(zsinz+cosz)?; (1) 3 sin” x cos 2z/ cos? 32;
(m) 5 cos 42z/cos® z; (n) —6 cos 3z cos 2/ sin* 22; (0) 3/V1 — 92?;
(p) -5/V1—252?; = (q) 3/./—(3
+ 127 + 922); (r) 4/(1
+ 162”);
(s) 3(sin7? x)” /V1i—<a?; (t) 2/(13+6x+27). 6. (a) 3/(32 — 5);
(b) 62? /(1 + 2x*); (c) 2/(6+ 27); (ad) 82/(32?=7); (e) —tanz;
(f) +22 In 22; (g) + sec?(In x); (h) 4(cos 4z + sin4z) / (sin 4x — cos 4z);
4 =

(i) «(3x2 — 10)In (x + 3) 5

|
2 Qa (2% — 527).
ot Nk ee Le (2 +7) (62? —5)
SS

pa ; Q) (2x? — 5)
A + 7)ln (22es ie
(22 52);a0 dye
(k) eiialie See Sap
clear (1) SPS TETT
(32? + 4)In 10’
Lefts
(m) —;4;tan4z; (n) 45 cosec4z; (0) ind (3) + log,(x + 4);

(p) 7e’*; (q) —3e*~3?; (r)4r3e*; (s) e* cosa; (t) 8°? seczr tanz;
(u) 22(1 + 22)e4*; (v) x? (323 — 27 +4) et —2t+5. (w) 4 x 247 1n2;
(x) sec?a x In5 x 5*; (y) ——3!™*; (z) 2?.757(3 + 521n7).
sinz
7. (a) (AEF + sec? zinz)2*™*; —— (b) ( : teoszinz) « sinz, :

(a — 2)(2x + 2)
(c) (Insin z+z cot x)(sin zr)”; (d) | +In (x? + 22 + 5)x
(x? + 22 + 5)
(x? +2245)"
Examples 5.
-—% or 474+ 4(keE Z). 3. (a) y =3-2,
1. (—1,—-2). 2.2 = *4*
y = +1, (2,1); (b) y = 18 —1lz, lly= z — 46, (1,7), (—2,40).
4. (a) 4, (0, —2); (b) —32, (3,—125), (—2,-90). 5. (a) 2sinzrcos3z,
xc = kn or z = 42 + 2(k € Z), 0, 1/6, 1/2, 50/6, 7, 7/6,
3/2, 11/6, 27; (b) —4cos* rsindz, c = 2kn + F or kr (k € Z),
 236 Answers

0, 1/5, 2m/5, 1/2, 30/5, 40/5, 7, 60/5, 7x /5, 3x/2, 8x/5, 9/5, 2;
(c) 3sin? xcos 2z/ cos? 3z, x = kn or kn+ $(k € Z), 0, mids 37/4, 7,
5/4, 7/4, 20; (d) —4cos 62/(sin® 22 cos? ren c= *+2(k eZ),
m/12, 7/4, 51/12, 7/12, 32/4, 117/12, 132/12, 50/4, 17m /12, 197/12,
7x /4, 237/12. 6. 48, 142. 7. 10 —2cos3, 12+2sin3. 8. 14504. 9. —26.
10. 50+ 6e®. 11. A= $27, 5. 12. 36z.
Examples 6.
1. (a) —2sin2z, —4cos2z, 8sin2z, 16 cos 2z; (b) sin3z + 3z cos 32,
6 cos 3x—9z sin 3x, —27 sin3z—27z cos 3z, 81z sin 3x—108 cos 32; (c) r+
22 Ins, 3+2Inz, 2/2, —2/2?; (d) —3e—3*, 9e~3*, —27e 37, 8le~9*.

Examples 7
1. (2y — 22 — 5)/(6y — 2x — 4). 2. (ycosz —siny)/(z
cosy —sinz).
3. 4, —25. 4. 3, 16. 5. 4y+2=6,y=42 +10. 6. y+ 32 = 77/6.
Examples 8.
No numerical answers.

Examples 9.
1. (—00, 4). 2. (—o0, —1) U (2,00). 3. (1, -3) and (—3, -19). Both
are points of inflection. 4. (—1,6) is a point of inflection; (1,—2) is
not. 5. 0, —48; concave down, and concave down becoming concave up.
6. (2,26): —10, 18, 22; concave up in each case. 7. (—2,—172), (3, 43);
142, —48, 168 and —82; concave down, up, up becoming down and down
becoming up.
Examples 10.
1. (—o0, —2) U (1,00). 2. (—oo, -3) U (—2,2). 3. (—00,1) and
(4,00). 4. (7/6,7/2) and (57/6,7). Hint: Use a table of sign to
find when 2sinzcos3z < 0. 5. (a) (—1,0), (1,0), (3,0) and (0,3);
wa, — a) maximum, z = 1+ a0 minimum; (1,0); (b) (1,0), (4,0)
and (0,—4); z = 1, maximum, z = 3, minimum; (2,—2). 6. (a) (4,4),
maximum, (10,16), minimum; (—2,0), (6,0) and (0,12/7); x = 7;
y =z. 7. (a) RIES maximum, (5,32), minimum; (—7,0), (1,0) and
(Ol) te 2. y oe Dae ee :) (5,1/8), maximum; (7,1/9); (3,0) and
(0,-3); ¢ = i y & 1/x. 9. (a) (—4,-1/4), minimum; (—5, —2/9);
(—3,0) and (0,3/4); = -2; y = 1/z. 10. (a) (0,4), minimum;
(—4,0), (4,0) and (0,4); c = +2; y = 1. 11. (a) (3,16), minimum;
(—5,0), (11,0) and (0, aig Pit. ly es By cel (1 ala),
4 (2) (c), (g); (8) (b), (d), (f), (h). 13. (a) (4,e7*), maximum;
2e~*); (3,0), (0,-3); y + 0* as t + co; y 4 —co as t + —00.
( (- ), minimum; (—4,—2e7*); (—2,0), (0,2); y — oo as
— oo; y + 0° as x + —oo. (c) (0,1), maximum, (1,0), minimum;
ie5 (-g-1) 4), (3.(4- 1).7); (40), (0,1); y + 00
 Answers 23%.

as t+ 00; y + 0* as zr + —oo. (d) (0,4) maximum,aL 0), minimum;

(-v3, (2 v9)2 eV), (v2,
(2+V2)’ e-¥?); (-2,0), (0,4); y + OF
as I —> 00; y > 00 as Z + —00.
Examples 11.
1. 13, —3. 2. 10.75, 8. 3. 625cm?. 4. 100V/3cm?. Hint: If the
two equal sides of the triangle have length y cm, and the other side has
length z cm, the area of the triangle is +2/900 — 30rcm?. 5. 8, 16.
6. 150cm?. 7. Base radius 3cm, height 6cm. 8. u?/(u — 75), 300cm.
9. Va? + 27 + \/b? + (c—2x)?. 10. (a + 6)/2.
Examples 12.
1. y = 24-227! 43, 18. 2.0 =t? + cost +4, e = it? +sint
+ 4t,
nm? + 3,20, 3m? +4m. 3. (a) $2442? —22 —2-3 +c; (b) 24 +3271 +
2r~* +c; (c) $22 +x? — 82+; (d) $2° — 223 +246; (e) tanz—z+c;
(f) 2tanz+2secr—xr+c; (g)—cotr+cosecr+c; (h) 57/In5+c.
4.y=2°+4+sinz+1. 5. y = —2cosz —3sinz +5.

Examples 13.
3 Z
1. (a)¢(2? +5)? +e; (b)4 (32° ae c; (c) —} (x? +4) * ke;
(d)+In|z4 +91 Kr (e) &(22° 430? 15 Ae f) 35 (a* — 22? bEoR) + ¢;
(g) —cos(tanz) + c; (h) $tan®x — ae Tipe ey (i) ceo +e;
(j) ag3*"* +c. 2. (a) —} cos5x+} sin3zr+<¢; (b) eende) cos 22 +c;
(c) sin~* 2 +c; (d) sin™’ 2 +; (e) }tan-1 24+; (f) Ht no} = + ¢;
(g) 3ANE (h) —e°*
+ ¢; (i) Ulett (j) In eae
Brie) gas (32 + 5)°+ p(82+5)" +0; (b) 55
2 (22— 3)°° + tog (20— 3)°4 +c;
(c) 3(r+7)8 +; (d) Y(4e — 5)? +c (ec) 3(42 + 3)2 +; (f)
2(52— 4)2 +c; (g) + sin(3x—2)+c; (h) —} cos(5z+2)+c; (i) > tan 7+¢;
() -—icot4r+c; (k) }sec(3z + 4) +c; (1) —}cosec(2r + 5) + ¢;
(m ) $n |3z+4|+c; (n) —} In |5—2z|+c; (0) —fe*-** +c; (p) Fe" **+c;
(4) -h2* 7 +6 4, (a) ¢tan] 24 +c; (b)
tan—} 2=1 + ¢; (c) sin7* 2+ + ¢; (d) sin7? %* + ¢; (e) sin’ 4 +¢;
f 3sin! zat? + ¢.
f)
Examples 14.
1. (a) 4(3a —1)e** +c; (b) §(sin3x — 3xcos3x) +c; (c) 72% x
(4Inz — 1)+c; (d) 4 (52 sin 5r+cos5zr)+<; (e) —}2? cos4z+pz sin4c+
33cos42 + ¢; (f)+ze — 152 + 5)e** +; (g)$(11 — 82 — 22?) cos 2z +
(+2) sin 2x+c; (h)135
33 (252? + 48) sindr+22 cos 5a+¢; (i) —+Inz—
i
1 +; (j) esin™*22 + V1 — 42? +c.
 238 Answers

Examples 15.
1. (a) ait C; (P) bias (c) 2sin? 2 +c; (d) dsin®
2—
7 sin’t+c; (e) —} cos? t+icos* x+c; Dens xr+c; (g) 3 cos ® r—
33 cost r+c; (h) —} cos® ee Seeel 22 enh+ sin 2r+¢;:
x—+ cos’ t+c. 2.(a)4
(b) $2+4sin 2z +c; (c) ee aac 3sin 42 +c; ae g2+44sin2z¢+
yy sin4dr +c; (e) gx —- 4 sin4dc — #sin 322+; (f) $2 — fsin2c+
sigan Anais’ Qr+c. 3.(a) +tan® c+c; (b) 2 tan? z+; (c) }tan® c+
+ tan® z+c; (d) }tan*z —tanr+2+¢; (e) i sec’ x — i sec? x + ¢;
(f) 2sec® x— 2 sec z + secx + c; (g) ant —- 2 sec? x + tsec?zr+c;
h) 39 sec? r—37 sec’ 2+ 25 sec® r—}3 sec? r+<; 12 secx tanz—} In| sec z+
tan z|+c; (j) —} cot? x+<; (k) ere (1) —} cot? r+cotzr+zr+c;
1
(m) —4cosec? 1
x + cosec x + c; (n) }cosec* 3 x — i cosec® x + ¢.

Examples 16.
1. (a) z/9V9 — x? +¢; (b) c/ V4 — 2?-sin™ E46; (c) Bsin7* 2+
dV/25—2? + c; (d) 18sin7’2 + 42V36—2? +c; (e) F(x -
4)//9 + 8a —2? +c; (f) (132 — 10)/9V5 — 4a — 2? — sin7’ #4? + ¢;
(g) Ftan7? = + 2/18(9+27) +; (h) InjVz? -22+54+2-1] +6.
2. (a) Is oie tines (b) In|z + Vz? —7| + ¢; (c) Injz —2+
Vx? — 42 + 5| +c; (d) In|c —3+ V2? —6z +3] +c.

Examples 17.
1. (a) +2, 22 —1; (b) 2? 4 2 — 3, 62 +7; (c) z? —32 +4, 32 —8.
2 (a) (£1) E—2)(2—3)(243)
Y
(b) (2921)
1
(242— 42 +5);
1
(0) (2
De et ie oe a east
(fy ee es
diet ) nea Pag gone Che ees omens:
4 3 3
— OO -ti—t— ; (d) —— ed — ———————

es (x + 1) Sir. ea eet Genie al

MBI WRey pcoLll)..0 1G Tae enes AM) ogame ek
r+ tee diet r—1 eee r—3 (zr - 3)?’
(Gr 29 *biigtelot 1 gt 3,
i ar) Weazie, st “he Deen eo:
Sa?
be 3 Med
Geno
car a (ec
OD ao o geal. 4. (a) In|x? +42+5| tan7!(2+2)+c;

(b) 3 In|z? —6r+13]4+2tan7! +e. 5. (a) 2ln|x +2|—In|z—1|+c;

(b) 3ln |z +3] —21n|z+2|+c; (c) 2In|z+1]+In|z—1|—31n|z—2| +c;
(d) 41n |z|+1n |z —2|—51n |z+1|+c; (e) 21n |z|—21n|z-2|—4/(x—-2) +c;
(f) 3ln | —1]+9ln|z +1] +6/(x +1) +6; (g) 2In|x+1]+2tan-! 2—
In |x? + 4| + ¢; (h) —31n|z + 3] + 3tan-} Z + In |x? + 9] + ¢; (i) In|x —
2|—2tan~! 242 — 3 In|2? +42 +8] +c; (j) (j) 21n |z — 3) +2tan71(2+3)+4
$ In|z? +62+10|+c. 6. (a) 22—21n |z4+2|+41n|z—1|+<; (b) $2? +62 —
 Answers 239

In |z—3|—51n |z+1|+c; (c) $2? +22—31n |x —2|+1n |z+-1|4+3/(2—2)

+c;
(d) c+ In|x — 1] + In|z? +4] — 3tan-! 2-1/(2- 1) +<.
Examples 18.
oe i —i, 3, ogi 20ogre 2. i, i, 3, $2+tsin2zr +
5 sin4z +c. 3. roe -4,- rot ce ray sin6z — A sin4dz — dsin2r +
nae 4.(a)gi
+ asin Art oe a) to (b) te?*(sin 2x +cos 2r) +c;
(c) —e~7(sin
3x + 3cos 3z) + c; (d)4e**(4sinz —cosz) +c.
saatin oe 19.
3+ /9+4+ 42?
1. (a) V9+ eovtehitiae F +c; (b) a,Spee: re

22 —
aE s nz Sec (derei sin® o}cos
(c) ap (22? — 125) V25—- 2? +38
#sin2rcos2z + 22 +c (e) *cos 5 Br sin32 + 3cos’ 3rsin3z +
=48 cos3zsin3z + oa + C Hp x sec® 2rtan2r + 2= sec 2x tan2z +
G ppleree ane o (g) -a conl0e at eeu (h) $2° Ine —
3pt° +c; (i) —sin”* $- V4 — 2? /x+c; (j) He" (5sin 2x — 2 cos 2x) +c;
(k) (—$24 + $2? — 2) cos2z + (23 — 22) sin2z + ¢; (1) —} cot? 32 —
Ere V4x +3 1 |M4e +3 = v3
= In|sin 32|+c; (m) ————— +¢;
3 3a “3a! V4e 434+ V3
—=} ——, (4
22 — 20g? + B02 — $00) (3245)? +.
Examples 20.
1. (a) +44; (b) 40; (c) Ao 7%
(h) ai Oh + VA(j)In|3]; (k
(d) tan7 na (2) ae (f) 4 ae

(i) 3; (k) $5 0) 9845 (m) V3; (n)

Seledhupels (Phe lag pa aioe 4
(eo “i x4ig) 22
1n2— 6; 3 ABS NG EY EA iy eet
(c) & 3 @) F _ 5. (a) In 22; (b) In 22; (c) In (2%
2 #*); (d) 141n5;
(e) ng — 4; (f) in #2
2 + 2; (g) ee “(h) 3m3 _ 3 —In2. 6. (a) 7;
(b) 22 } (c)1; (d) ae

Examples 21.
1. (a) 30; (b) 24; (c) ¥. 2. (a) 24; (b)42. 3. 2. 4. 2.
Examples 22.
1. (a) 48"; (b) *47*; (c) 28F; (d) w(1— 4). 2. 8". 3. 367.
4. 4. Hint: think of the circle as formed by the two curves y =
2+ Jl — <2? and y = 2— V1—2?. Then find the volume obtained by
rotating the area between these two curves about the z-axis.
 240 Answers

Examples 23.
1. (a) % (373 —- 108); (b) 728; Hint: ./1+(y’)? simplifies to
2x? — 1 for 2 < x < 4; (c) %; Hint: /1+(y’)? simplifies to
4
(d) In (42); (e) 4+1n3; Hint: ./1+ (y’)? simpli-
©
aS
fies to — + a 2. (a) eee Hint: Show that the surface area is
£
4n es Vy leeds; (b) 7 (103 _ 1); Hint: Show that the surface area
is 27 hs a? V1 + 9x4 dz; (c) = (e? +4-—e7”); Hint: Show that the sur-
face area is xf (e7 + e-*)? dz. 3.2, ¢(4x—- 3/3), 89", 327; Hint:
In each of the integrations, you need to deal separately with the cases
0<2<2and2 <2 <4. For the curved surface areas, you finish up
needing to find 2 {4dz in each case, with appropriate limits.
Examples 24.
1. (a) (a) 146, (@) 144; (b) (a) $(2+ v2), (6) } (14+v2);
(c) (a) 1.073 (to 3 dec. pl.), (8) 1.057 (to 3 dec. pl.); (d) (a) =, (8) =.
2. (a) (a) 0.2314065..., 0.4628130..., (2) 0.2318263..., 0.4636526...,
1 tan-11; (b) (a) 2.2734126..., 1.1367063..., (8) 2.210052...,
2
1 1050264..., 2In3. 3. (a) (a) 6.087, (G) 6.081; (b) (a) 1.8134,
(8) 18100. 4. (a) $(Vv2+V6+1), (8) 3(vV2+2vV6 +1).
5. (a) 32.474435..., (B) 33.267892....
Examples 25.
1. 12. 2, 28. 3. 0.824. 4. 0.773. 5. 3, 2.762. 6. 4, 4.100. 7. 3.780.
8. 0.940.
Examples 26.
2.14 52? + Sa*t..., 1.00786.... 3.14274
$2? + S23 +...
4.2+27 — zx? —iat+.... 5. 34+ i2- art’
+ ty2°, 2.9240216....
6. — (< + pu? + ia> + lot +...), 2(r+ pare + a ee 2.398,
2.565. 7.1+¢7+5t4+
2194 Ato +..., 0.203. 8. t-2t + 245-At7+...,
0.041. 9.14 1 + 33444 St9 4 3548 4 tod 4 Roh 4 Sog7 4 38 99
347
10.1427 +¢4+ #94 t8+...,2(2+ 3034+ 2054 4274 129 +...).
Examples 27.
1. —2cost + 8 — fe# —¢ 43. 2. (a) -y? = 1 +e;
(b) In|y/(1+y)| = 2? +c; (c) —coty = sin 124+ 6¢; (d) tan-}y =
5e** +c; (e) fy? +cosy = sinr—42° +c; (f) e4¥ = 2e?* +c; (g) tan“! y =
tan’ x +c; (h) secy = cz. 3. (a) siny = c—cosz, siny = 1 — cosz;
(b) Injl + y| = i («° +c), Injl+y| = z (2° — 1); (c) t+ 2y7? =e,
z* + 2y-? = 483: (d) 3sin2z + 2cos3y = c, 3sin2r
+ 2cos3y = 3.
 Answers 241

4. (a) In|y] = c— _ ly: (b) tan“? 2 = In|c fz? + y? i

(c) y? +27 In |y|= cxz?; (d) In|y/z|= cz; Hint: Putsu=lInv to solve
the harder integral. 5. is y = 2+4+ce~**, y =2+e77*; (b) y = 2° +cz?,
y= 2° =—Te?*(c) y= i(22—l)e"* teen #,y = }(2r-1)e**
+ Be-*?;
(d) y = i(sin2r — 2cos2z) + ce~*, y = z(sin2x — 2cos2z) + 22ae A
6. (a) (a) y = Ae** + Be?*, (8B) y = 32” + 22 — 4, (7) y = Ae? +
Be? + 3x? +22 —4; (b) (a) y = Ae?* + Be~**, (8) y = sin 2x —2cos2z,
(y) y = Ae?* + Be-3* + sin2zr — 2cos2z; (c) (a) y = Ae** + Be-?,
(B) y= ert) (7) y= Ae®* A Bet San (6) y= Jeot a Se 7? c err:

(ORG Uae see eva te, omy =

Ae** + Be~3* +23 — x? —2 +2; (e) (a) y = Ae?* + Be”, (8) y = —6ze",
(7) y = Ae** + Be? — 6ze"; (f) (a) y = Ae? + Be”, (8) y = —3ze-**,
(vy) y = Ae~*7+Be*—3zre~**, (5) y = —2e-?*
+4e7 —32e-?*; (g) (a) y =
(A+ Br)e’*, (8B) y= 2? +3, (y) y=(A+ Br)e™* +27 +3; (h) (a) y=
(A+ Br)e~**, (8) y = 2? —22 +3, (7) y= (A+ Br)e—3* + 2? — 22 +3;
(i) (a) y = (A+ Bz)e*, (8) y = 2x%e*, (7) y = (A + Br + 22? )e”,
(6) y = (2x? + x — 2)e"; (j) (a) y = Acos3zr + Bsin3z, (8) y = e~?7,
(y) y = Acos3zc + Bsin3z + e~?7; (k) (a) y = Acos4z + Bsin4z,
(B) y = 3sin 2z — cos 22, (y) y= rare + Bsin4z + 3sin2z — cos 2z;
(l) (a) y = e?*(Acosz + Bsine)- (By yeaa? + zr ="2? (7) y =
e?7(Acosz + Bsinz)+ 274 2 —2; (m) (a) y = e?(Acos3z + Bsin3z),
(Byres (ary = e* (Arcos 32 +P sin sa 1), toy y = €* (2083a —
sin3z + 1); (n) (a) y = oe Nco-e + Sree (3) 29 =Sa 3,
(y) y = e?7(Acos2z + Bsin2z) + x? — 3; (0) (a) y = Ae” + Be**,
(B) y = 22 —1— 2e?”, (y) y = Ae? + Be®* + 2x — 1 — 2e?*; (p) (a) ey
Ae~**+ Be", (B) y = e* +2e 7", (4) y= Ae 724+ Be ** +e7 +2e72*.
Examples 28.
1. Iny = $24 -— 4. 2. tan“! 2 = InV/2? +y?. 3. 5.8504910....
4, 44.243095.... 5.2 = 3 (a= ets)/( — 2685). Glu. =
a lets = ert): Ty=c(l- env Bees Aew*t4 3 (1- eu),
Orr = oH (sinwt — wRC coswt + wRCe~Xe )+ ige Re. Hint:
Use integration by parts twice, or the complex number method, to find
f exe sinwt dt. 10. v = 3 (e-# - esi), 11.0 = (t—})e"** + Fe-*.
12.7 =2 (ems! = et)
Examples 29.
1. (a) («© — 1 +5); (b) (2 + Iz +4); (c) (2 + 3)(z + 2)
(d) (2 + 4)(a — 3); (e) (22 +3)(z +1); (f) (32 + 1)(e +4); (g), (h) not
possible; (i) (22 — 3)(2z — 1); (j) (32 — 1)(z + 2); (k) not possible;
(itor (etl) 2ata eB — 715, b= 3:7827798.. ..,1ci=.3.2079811...-. ;
(b) C = 58.98696...., B= 8101503 sae, b = 4.6098767...;
(c) B = 8.2132107..., C = 141.78678..., c = 8.660254...
242

Some Standard Integrals

In the following formulae, a is a constant.
 243

Index
The numbers refer to pages in the text

Absolute maximum, 91 cooling, Newton’s law of, 217

absolute minimum, 91 cost, marginal, 55
absolute value, 9 Cover Up Rule, 135
acceleration, 54 critical point, 52,77
addition formulae, 19 cube, duplication of the, 190
amplitude, 22 cubic functions, 7
angle, trisection of the, 190 curve, length along the, 173
arbitrary constant, 100,199 curved surface area, 175
area, curved surface, 175 cylinder, surface area of, 91
areas, signed, 164 cylinder, volume of, 91
asymptote, horizontal, 65
asymptote, vertical, 64 Decay, exponential, 216
asymptotically, behaves, 65 decreasing function, 7,77
auxiliary equation, 207 definite integral, 154
average value, 162 degree, 200
axis of symmetry, 2 degree, total, 202
degrees, 13
Behaves asymptotically, 65 derivative, 39
binomial series, 195 derivatives, higher, 57
boundary conditions, 199 derivatives, standard, 40,45,47
differentiable functions, 40
Calculus, Fundamental Theorem differential equation, 58,199
of Integral, 164 differential equation, linear, 205
CAST, 16 differential equation, separable,
Chain Rule, 43 200
change of variable, integration by, discriminant, 6
106 displacement, 22
change of variable, linear, 108 divergent, 191
change, rate of, 53 division of polynomials, long,
closed interval, 9 133,231
complementary function, 207 division, synthetic, 141
completing the square, 2 dominates, 85
composition, 42 double angle formulae, 20
concave down, 72 down, concave, 72
concave up, 72 duplication of the cube, 190
conditions, boundary, 199
conditions, initial, 199 Ellipse, 60
cone, surface area of, 91 equation, auxiliary, 207
cone, volume of, 91,170 equation, differential, 58,199
constant, arbitrary, 100,199 equilateral triangle, 227
constant of integration, 100 explicitly defined function, 60
convergent, 191 exponential decay, 216
 244

exponential growth, 216 Integral Calculus, Fundamental

Theorem of, 164
Factor, integrating, 205 integral, definite, 154
factorisation, 225 integral, indefinite, 99
formula, quadratic, 5 integral, particular, 207
formulae, addition, 19 integral sign, 100
formulae, double angle, 20 integrals, standard, 101,108,122,
formulae, reduction, 152 125,142
fraction, improper, 134 integrand, 100
fraction, proper, 133 integrating factor, 205
function, 1 integration by change of variable,
function, complementary, 207 106
function, explicitly defined, 60 integration by substitution, 106
function, homogeneous, 202 integration, constant of, 100
function, implicitly defined, 60 integration, range of, 154
function, periodic, 18 integration, variable of, 100
function, polynomial, 58 Intermediate Value Theorem, 188
function, positive, 9 interval, closed, 9
function, rational, 64,133 interval notation, 9
function, strictly decreasing, 7,77 interval, open, 9
function, strictly increasing, 7,77 inverse trigonometric functions,
function, strictly positive, 9 30-31
functions, cubic, 7 isosceles triangle, 227
functions, differentiable, 40
functions, inverse trigonometric, Law of cooling, Newton’s, 217
30-31 length along the curve, 173
functions, trigonometric, 15 limit, lower, 154
Fundamental Theorem of Integral limit, upper, 154
Calculus, 164 linear change of variable, 108
linear differential equation, 205
General solution, 199 local maximum, 78
geometric progression, 191 local minimum, 78
growth, exponential, 216 long division of polynomials, 133,
231
Higher derivatives, 57 lower limit, 154
homogeneous function, 202
horizontal asymptote, 65 Marginal cost, 55
marginal revenue, 56
Identically, 139 maximum, absolute, 91
implicitly defined function, 60 maximum, local, 78
improper fraction, 134 mean value, 162
increasing function, 7,77 minimum, absolute, 91
indefinite integral, 99 minimum, local, 78
inflection, point of, 73 modulus, 9
initial conditions, 199
 245

IN, 58 Rule, Sine, 229

Newton’s law of cooling, 217 Rule, Trapezium, 179
normal, 51 Rule, Trapezoidal, 179

Open interval, 9 Separable differential equation, 200

order, 200 series, binomial, 195
ordinates, 183 series, power, 191
origin, symmetric about the, 20 sign, integral, 100
orthogonally, 63 sign, table of, 8,67,74,78
signed areas, 164
Particular integral, 207 similar triangles, 13
particular solution, 199 Simpson’s Rule, 181
period, 18 Sine Rule, 229
periodic function, 18 solution, general, 199
phase shift, 22 solution, particular, 199
point, critical, 52,77 sphere, surface area of, 91,176
point of inflection, 73 sphere, volume of, 91,170
point, turning, 52,77 square, completing the, 2
polynomial function, 58 standard derivatives, 40,45,47
polynomials, long division of, 133, standard integrals, 101,108,122,
231 125,142
positive function, 9 strictly decreasing function, 7,77
power series, 191 strictly increasing function, 7,77
Product Rule, 40 strictly positive function, 9
progression, geometric, 191 substitution, integration by, 106
proper fractions, 133 surface area, curved, 175
surface area of cone, 91
Quadratic formula, 5 surface area of cylinder, 91
Quotient Rule, 40 surface area of sphere, 91,176
' symmetric about the origin, 20
R, 30 symmetric about the y-axis, 20
radians, 13 symmetry, axis of, 2
range of integration, 154 synthetic division, 141
rate of change, 53
rational function, 64,133 Table of sign, 8,67,74,78
recurrence relations, 152 tangent, 39
reduction formulae, 152 Theorem of Integral Calculus,
relations, recurrence, 152 Fundamental, 164
Remainder Theorem, 52 Theorem, Intermediate Value, 188
revenue, marginal, 56 Theorem, Remainder, 52
Rule, Chain, 43 torus, 172 .
Rule, Cover Up, 135 total degree, 202
Rule, Product, 40 Trapezium Rule, 179
Rule, Quotient, 40 Trapezoidal Rule, 179
Rule, Simpson’s, 181 triangle, equilateral, 227
 246

triangle, isosceles, 227

triangles, similar, 13
trigonometric functions, 15
trisection of the angle, 190
turning point, 52,77

Union, 9
up, concave, 72
upper limit, 154

Value, absolute, 9
value, average, 162
value, mean, 162
vanishes, 9
variable, integration by change
of, 106
variable, linear change of, 108
variable of integration, 100
velocity, 54
vertical asymptote, 64
volume of cone, 91,170
volume of cylinder, 91
volume of sphere, 91,170

y-axis, symmetric about the, 20

ZZ, 30
 247

ALSO AVAILABLE IN THIS SERIES ARE THE FOLLOWING:

ADVANCED CALCULUS FOR ENGINEERING

AND SCIENCE STUDENTS
Ian S. Murphy. Fourth Edition, 1993. 248 pages.
Arklay Publishers, 64 Murray Place, Stirling.

Contents: Double and Triple Integration, Beta and Gamma Func-

tions, Differential Equations, Laplace Transforms, Partial Differentia-
tion, Errors and Exact Differentials, Vector Calculus, Line and Surface
Integrals, Fourier Series, Maxima and Minima of Functions of Several
Variables, Eigenvalues and Eigenvectors of Matrices, Systems of Linear
Differential Equations, Lagrange Multipliers, Hints and answers to the
examples, Index.

BASIC MATHEMATICAL ANALYSIS: THE FACTS

Ian S. Murphy. Third Edition, 1991. 246 pages.
Arklay Publishers, 64 Murray Place, Stirling.

Contents: Basic ideas, Limits of sequences, Limits of real func-

tions, Continuity, Metric spaces and compactness, Derivatives and ap-
plications, Series, Taylor and Maclaurin expansions, More on series,
Power series, Exponential, logarithmic, trigonometric and hyperbolic
functions, Integration, More on the convergence of sequences and series,
Miscellaneous examples, Hints and answers to the examples, Index.

PROBABILITY: A FIRST COURSE

Ian S. Murphy. First Edition, 1991. 120 pages.
Arklay Publishers, 64 Murray Place, Stirling.

Contents: Basic ideas, Counting cases, Equally likely outcomes,

Hypergeometric distribution, Conditional probability, Bayes’ theorem,
Independence, Discrete and continuous random variables, Binomial,
Poisson and Normal random variables, Expected value, Variance and
standard deviation, Normal and Poisson approximations to the Bino-
mial, Two dimensional random variables, Chebyshev Inequality, The
Central Limit Theorem, Appendix on the Gamma function, Table for
the Normal Distribution, Hints and answers to the examples, Index.
 ue :
ONO
bp. —_. .3
TAT BHA ERATE ZIT MtBIAIAVW OBI
: — > >>

—_ vier-tore <}

o>" DMPA ISUT AOT 2U09IA9 CADMAVOR

os aTuAGY TS? BIASIQe OKA -
Samy BLE 01 cotttbS diwot «dart .2 nal.
= | gaibve gaat? wine & sodsildod yabiA
= 7 ;
eet SNR! bes at-G notempetal hdinT bas sideoll .
MAEM tiie] ccorctuntl soalted sandauys leita -
tid |wt}
Mite her sil? wivels) roigoV alsanetshid id bas eerndl en
freA Reed iftood Ww amici baa amixall ons? wie elatesini
WANS carole tureld to molosvesni ba raclevesgie coldectsV
1G anew bes aol amilgine 4 synerged eolaupS WwilgrrsRid
~ebrl gainers.
“2 - ‘< A? ' ¥

— a ; 3 ae

OSTOMY SGT -CIRYJAVA JAOTTAMSRTAM SIBAS

de 7
cia os @hBig OX. LOL wovikt budT = agen el
Oe 8) Rititae
xe Pa 1salemT Et,
M 3d .evkdaihdi| aublen
’

Mane tase Ys esha eorsupss 1 eri@il exobi cee gia) ~

ORE PAPA etinadinon brie ‘Saye oitisM xtientlin) acid:
Aa) wo sol. Aamoindogne gual brs solysT jesirst seri tadily
“silted bo neMscioluegicsd ,sinidtisyo! Jevisnogss siget xwod
eeriee ters exoicsup Yo sategrernes sdf no sto soliep
elel agoiisndt ~
ashal colqmsxs ad) ofcwene bos etal slgmaxs qraasilsnlh

#eAYOO TeRIt A «=YYMIARORG ea an

weasq OCI (GQi ,aowibs pai: solu 2 ual |
amie soeld werk 8 subti j Ie
This book provides a clear introduction to Calculus. The
main topics covered are Differentiation, Integration, Curve
Sketching, Trigonometry and Differential Equations.
The book also features applications of Calculus to many
practical situations. Worked examples are provided
throughout the text to illustrate specific points and there are
many problems for students to try for themselves. Answers
and some hints are given at the end of the book.

ARKLAY PUBLISHERS
64 Murray Place,
Stirling,
Scotland.

vgn
tem
1/47 47:55:STSOB:KA
mat
Pp
su

ISBN 0 9507126 9 8 LLY

AU
UA
ALAN
QU
228-9