st

Revised 51 Editil(i;f.
<=::;;.-
Re" ' ;ended Text Book for ~he students of under-graduate
M,\ I .IF.MATICS ' by the Indian Universities, like Universities of
CaICI. (/, Burdwan, Jadavpur, Kalyani, North Bengal, Vidyasagar,
D...CCt Patna, Bihar, Bhagalpur, Nalane'!, Sambalpur, Berhampur,
UI;;o/, Guwahali, Dibrugarh, Tribhuban, elc.

, /"

· D . IFERN~Al
' ,-
CALCU·LUS, \

B. C. DAS, M.Sc.
PROFESSOR OF MATHEMATICS (RETO.);'
PRES1DENCY COLLEGE, CALCUTTA;
EX·LECTURER IN "PPLIED MATHEMATICs. '.
CALCUTTA UNIVERSITY

B. N. MUKHERJEE, M.A.
; .'. Pret:nChand Roychand Scholar
PROFESSOR OF MATHEMATICS (RETO.).
SCOTTISH CHURCH COLLEGE. CALCUTTA

C J: N. DHUR &SONS PRIVATE LIMITED

KOLKATA- 700 073 .
 © reserved by the authors

1st Edition 1949

10th Edition 1962
20th Edition 1977
25th Edition 198J
30th Edition 1987
35th Edition 1991
40th Edition 1994
42th Edition 1996
44th Edition 1998
45th Edition 1999
[ 6th Edition 2001
47th Edition 2002
48th . Edition 200J
49th Edition 2004
50th Edition zOO6
Reprint 2006
Reprint 2007
51 SI Edition 2009
Reprint 2010
Rc-print 20 12

ISBN-8 t -85624-85-0

Price: Rupees One Hundred Eighty (J

'published by : Dr. Purnendu Dhar, M.Sc., Ph:D.

far U.N.DHUR & SONS PRIVATE LIMITED
2A, Bhawani Dutta Lane, Kalkata - 700 073
Phone No : 2241 9573 132923854
Mobile: 9830169816/94328 89588
e-mail : undhur@vsoLnet

Printed by : GRAAFl KA
36C, Aurobindo Sarani
Kolkata - 700 006
 PREFACE To THE 51st EDITION
This edition of Differential Calculus, is almost the reprint of
the previous one with afew rectifications of the printing mistakes.
• Some rearrangements and additions-alterations also been
done in afew chapters to upgrade and comply with the recent
trends of paper setting in different examinations.
W e are thankful to Dr. Dinabandhu Chatterjee, M.Sc., D.Phg
of Dinabaudhu College, Howrah (Retd.) for revision of this
edition.
Our thanks are due to Dr. Purnendu Dhar, M.Sc., Ph. D.,for
bringing out this edition in time.
A ny suggestions for the improvement of this book will be
received, with thanks.
Kolkata,
January, 2009 Copyright Holders.

PREFACE To THE 50th EurrIoN

This edition of Differential Calculus, is thoroughly revised
and updated to comply with the latest syllabus of the Indian
Universities.
A few chapters like Number System, Function, Maxima and
Minima, have been modified.
A n attempt has been taken to creole interest about the basic
need to study Calculus by demonstrating and explaining A Few
Real-Life Events or features by help of Calculus.
A lot of interestingMiscellaneous W orked out Examples have
been incorporated to most of the chapters for the benefit of the
students.
W e are thankful to Dr. Dinabandhu Chatterfee, M.Sc., D. Phil,
of Dinabandhu College, Howrah (Retd.) or revision of this
edition.
W e convey thanks to Sri. Santana Kr. Ganguli, F.C.A . for
introducing some new ideas.
W e convey thanks to Dr. Purnendu Dhar, MSc., Ph. D., for
publishing this revised edition in time.
A ny suggestions for the improvement of this book will be
received, with thanks.
Kolkata,
June, 2006 copyright Holders.
PREFACE TO THE FIRST EDITION

This book is prepared with a view to be used as a textbook for

the B. A. and B. Sc. students of the Indian Universities. We have
tried to make the expositions of the fundamental principles clear as
well as concise without going into unnecessary* details; and at the
same time an attempt has been made to make the treatment as much
rigorous and up-to-date as possible within the scope of the
elementary work.
Proofs of certain important fundamental re s ults and theorems
which are assumed in the earlier chapters of the hook for the
convenience of the beginners and average students have been given
in the Appendix: A brief account of the theory of infinite series,
especially the power series, is given in the Appendix in order to
emphasise their peculiarity in the application of the principles of
Calculus. Important formulae of this book are given in the beginning
for ready reference. A good number ot typical examples have been
worked out by way of illustrations.
Many varied types of examples have been given for exercise in
order that the students might acquire a good grasp of the
applications of the principles of Calculus. University questions of
recent years have been added at the end to give the students and idea
of the standard of the examination. Our thanks are due to the
authorities and the staff of the K. P. Basu Printing Works, Calcutta,
who, in spite of their various preoccupations, had the kindness to
complete the printing so efficiently in a short period of time. Any
criticism, correction and suggestion towards the imprnvement of the
book from teachers and students will be thankfully received.

CALCUTTA B. C. D
June, 1949 B. N. M
 LATF.ST SYLLABUS FOR
DWFERENTIAL CALCULUS (G~)
. Rational number~ . Geomttrical representations .. Irrational numbers. Real
number represented as point on a line - Linear Continuum. Acq,!aintance
with basic propeflies of real number (No deducti"n or proof included). '
Sequence: Definition of bounds of a sequence and monotone sequence.
Limit of a sequence. St~emns of limit· theorems. Concept. of convergence
and divergence of monotone sequences - applications of the theorems, in
particular, definition nf e.
Statements of Cauchy's general. prindple of convergence .and its
opplication.
Infinite series of con s.tant terms : Convergence and Oi vergence
defij,ton~)( . Cauchy ' ~ principle as applied to inft~ ~eri (application
only). Serie,s of positive terms: Statements of Comparison te~, O.Alembert's
Ratio test, Cauchy's /IIh root test and Raabe's test - Applications. Ahernating
series: Statements of Leibnitz's test and its application .
Real-valued functions defined on an interval : Limit of a function (Cauchy's
definition). Algebra of limits. Continui"ty of a function at a point and in an
interval.
Acquiantance (no proof) with the important 'properties of continuous
functio~ on c1o~ed intervals. Statement of existence of inverse function of
a striclly monotone ·function lind i.ts continuity.
Derivative - its geometrieai ancrphysic;.I interpretation. Sign of derivative-
Monotonic increasing and decreasing functions. Relation between continuity
and derivability. Defferential - applicaiion in finding approximation.
Successiv.e derivative- l:.eibnilZ's Theorem and its application .
.Statement of Rolle's Theorem and its geometrical interpretation. Mean
value Theorems of Lagrange's and Cauchy. Statement of Taylor's and
Maclaurin's Theorems .with Lagrange's and Cauchy's form of remainders.
Taylor's and Maclaurin's Infinite series for functions like
~'. sinx, cosx. (I+x)" . log(l+x).[ with resttictions wherever necessary I

lRdeterminate Forms : L'Hospiial's Rule. : Statement and pn;blems only.

Application of the principle of Maxima and Minima for a functiori of
single variable in geometrical. physical and other pf!)blems.
 Functions of two and-three variables Their geometrical representations.
Limit and Continuity (definitions only) for functions of two variables. Partial
derivatives : Knowledge and use of Chain Rule. Exact differentials (emphasis
on solving problems only). Functions of two variables - successive partial
derivatives : Statement of Schwarzs Theorem on Commutative property of
mixed derivatives. Eute?s Theorem on homogeneous function of two and
three variables. Maxima and minima of functions of not more than three
variables - Lagrange's Method of undetermined multiplier - Problems only.
Implicit function in case of function of two variables (existence assumed)
and derivative.

Applications of Differential Calculus Tangents and Normals, Pedal

equation and Pedal of a Curve. Rectilinear Asymptotes (Cartesian only).
Curvature of plane curves. Envelope of family of straight lines and of curves
(Problems only).

Definitions and examples of singular pints ( viz. Node, Cusp. Isolated

point).

University of Calcutta
Senate House
Kolkata 700 073
 CONTENTS

CHAPTERS
I. A few real-life events explained by Calculus ix

H. List of important formulcr xix

HI. The Historical Development of Calculus xxix

1. Numbers System
2. Functions 15
3. Limit 47
4, Continuity 87
5. Sequence 111
6. Infinite Series 139
7. Differentiation 170
8. Successive Differentiation 246
9. Expansion of functions 273
10. Maxima and Minima 315
11. Indeterminate Forms 349
12. Partial Differentiation 375
13. Extremaof Functions of two or more Variables 420
14. Tangent and Normal 433
15. Curvature. 470
16. Asymptotes 505
17. Envelope of Straight Lines and Curves 530
18. Associated Loci 546
19. Concavity and Convexity Point of Inflexion 558
20. On Some Well-known Curves 566
21. Singular Points 580
Index 589
GREEK ALPHABETS USED IN THE BOOK

a (alpha) p (rho)

P (beta) cr(sigrna)

y (gamma) r (tau)

45  (delta (xi)

0 (phi) fl(eta)

f(psi) (zeta)

K (kappa) E (epsilon)

). (lambda) O (theta)

u(mu) L% (cap. delta)

V (nu) E(cap. sigma)

IV (pi) F (cap. gamma)

 A FEW REAL-LIFE EVENTS EXPLAINED B y CALCULUS
Calculus, invented from the visions of master minds, took little time to
break through the barrier of theoretical enquiry to practical fields of human
activities. The apparently unrelated areas of application of this pattern of the
novel methods of calculus enabled it to determine the dimensions of a con-
tainer in the shelves of the grocer to the angle of elevation of the rainbow in
the sky, the rate of drying of a moist piece of cloth kept hung in the air, or to
predict the laws governing the motion of planets in the heavens far from our
planet. -
Let us make a voyage to a few.of the areas where calculus reigns.
I. To A ssess the Dimeftsjons of a can to minimise the cost of
material.

A cylindrical can is to be made to hold I litre of oil. To find the dimensions

that will minimise of the metal to manufacture the can.
Let us consider a cylindrical can of radius rand height h. The volume of
the can is

7tr 2 h= 1000,'for 1 litre.iyjc.c

(i)

'E7
Now, total metal sheet required to
make the can
= total surface area of the can (A)
= 2it r 2 + 2nrh
21t r 2 +2 7tr x[.P) -

using (i)

= 27E r 2 + 2- Area 2(irr2) Area (27rr)h

Now, we are to minimise A

Calculus teaches that for a maximum or minimum of A, we have d =0
dr
2000
i.e., -r0 5
or, r3=—.
r
2
d A 4000
Now,4n + — i-->0
;;;2 ,.
X DIFFERENT/AL CALCULUS

500
indicating that A i s minimum when r = i.e.,
i, =
7t N 7t
and the value of n corresponding to this value of r is given by

h =2J=2r.

Now you have got the logic behind the fact that cylindricaal containers
are made so that their heights are approximately double the radius.
II. The calculus of Rainbow
Rainbow in the sky have fascinated mankind from ancient times. Formation
of rainbow is a beautiful phenomenon of nature caused by scattering of
sunlight by waterdrops floating in.air.
The formation of a rainbow can be ex1ained in the following way.

4^4
to
observer
Formation of the primary rainbow
A ray of light meets the spherical surface of the raindrop at A, after
refraction it takes up the path along All
inside the drop, at B the ray suffers total royv rro,n So,,

internal refraction to move along BC. At

138
Cit is refracted again to emerge out of
the raindrop. --
If a be the angle of incidence of the so,
ray of light at A, then Mot), the amount ) .
of clockwise deviation that the ray has
observer
undergone during this three-stage
process.
Thus, Iu=(u-3)+(m-213)+(a—f3)=7T+2(X-4I3.
A little calculus will show that D is minimum when a 59 . 4° and
= 138°.
The above figure shows that the angle oielcvation from the observer up
to the highest point of the rainbow is 080' - 138°) = 42°.
 A FEW REAL-LIFE EVENTS EXPLAINED BY CALCULUS

Ill. Problem of moist piece of cloth.

A moist piece of cloth in the open air loses its moisture at a rate
proportional to the moisture content. If a sheet hung in the wind loses half
its moisture during the first hour, when will it lose 99% of moisture, weather
conditions remaining unchanged ?
To solve this problem found in our everyday life, let m0 be the initial
moisture content and m be the value of ,n0 after t hours.
These facts can be put in the language of Calculus as
dm
= km where k (<0) is constant of proportionality.
dr

This, being a differential equation, has its solution m = rn0ekl.

Given that m = - 3 ,when 1= I.

then mo=noeeK=,k=_log.

Now, if the cloth loses 99% of moisture after Thours

—m0=ni1eKT
too
I loge 100
eKT = T
100 log, 2 hrs. = 6 hrs. 39mjns.

The techinque of this problem can be used to the population growth of a

country, or, the decay of a radioactive substance, the growth of bacteria in a
medium and in many other investigations.
IV. Calculus in planetary motion.
Kepler formulated three laws of planetary
motion, for they fitted the astronomical data.
Now you can see how calculus can help to
deduce the laws.
Since a planet during its motion round
the sun, is acted on by an attraction towards
the sun, the force or acceleration along the
tangent to the path is zero, which can be represented by an equation of the

form - - - -- r 2 = 0 where r is the distance of the planet from the Sun

r dtt, di)
and 0 is the angle traced out by the planet in time I. Integrating the above
relation, we have

, dO
r- - constant.
di
xii D!FFERENIJAL CALCULUS

dO -
But - I r 2 -- is the rate at which the radius vector r sweeps out area
around the sun.
Thus we find that the rate of description of sectional area by the radius
joining sun to the planet is constant, which proves Kepler 's Second. Law of
planetary motion which States that the radius-vector joining any planet to
the sun sweeps out equal areas in equal intervals of time.
The other two laws can be similarly deduced.

V. W here should a pilot of an aircraft start descent

An approach path for an aircraft, as
shown in the figure, satisfies the following Y
conditions
(i) The cruising altitude is h when the Y = r(x)
descent Starts at a horizontal distance I
from the touchdown at the origin.
(ii) The pilot mahtains an uniform
horizontal velocity v through out. () x
(iii) The absolute value of the vertical
acceleration must not exceed a constant Ic which is less than the acceleration
due to gravity.
If the altitude y and the horizontal distance from the origin hex, then we
may assume
y'&+bx2 +cx+d.
6hv2
The conditions (i), (ii), (iii) give - j— k

For a given k, v and Ii, the path and the horizontal distance I can be
calculated.
VI. W here to sit in a Cinema Hall.
We suppose that the cinema Hall has
a screen positioned 10 ft. off the floor and
is 25 ft. high. The first row of seats is
251 '
placed 9 ft from the screen and the rows
are 3 ft. apart. The floor of the seating area —
is inclined at angle (x = 20. above the -———
horizontal and the distance up the incline loft
where an observer Sits is x. The cinema r -
hall has 21 rows of seats, so 0:5 x:560. 91

An observer decides that the best place to sit is in the row where the
angleO, subtended by the screen at the observer's eye is a maximum.
 A FEW REAL-LIFE EVENTS EXPLAINED BY CALCULUS xiii

Now, let us translate the problem in mathematical language. Obviously,

this is a maximizing problem, and we are to find the value ofx which makes 9
a maximum.
Now, 2ah COS O=a 2 +h2 _(25)2
a2+b2-625
i.e., cos9= —
2ab
Where a 2 = (9+x COS cz)2 +(31–xsincz)2
2
and b 2 =(9+xcos(X) +(xsina-6)2
dO
Here numerical computations are complicated, but 0 will give the
required value oft
VII. Flow of Blood through Veins and arteries.

The laminar flow of a fluid is govermed by the law:

v(r)=_! _(R2_r2)
4111
which gives the velocity v of a fluid that flows along a tube of radius R
and length 1, at a distance r from the central
axis; and P is the pressure difference between ..
the ends of the tube 'i being the viscosity of
the fluid.

Now we use this law to calculate the rate of

flow orflux (i.e. volume per unit time) of blood
within blood vessels (i.e. veins or arteries). Let
us consider smaller, equally spaced radii
r1 , r2 ,•• We see that the approximate area of
the ring with inner radius , and outer radius , is

21çAr, where Ar = , — p_

When Ar is small enough, v can be assumed to be constant having

value v(r,).

So the volume of blood that flows per unit time can be approximated by
the expression

(2mir)v(r,) = 2n,v()Ar.
Thus the total volume of blood flowing across a Cross-section per unit
timejs

2iti;.v(i )Ar.
xlv DIFFERENTIAL CALCULUS

Using the concept of Definite Integral as the limit of a sum, the volume of
blood that passes a cross-section of a blood vessel per unit time is given by
F=lim21trjv(r)A r

f27trv(r)dr
0
= J2nr.._._(R2 _r2)dr

ItPR4
i.e., F=--------.
8ijI
VIII. Cardiac Out Put
The figure below shows the human cardiovascular system. Blood
returns from the body aorta
through the veins, vein I
enters the right atrium pulmonary -
arteries
of the heart, and is pulmonary
pumped to the lungs arteries
for oxygenation. It
veins
then flows back into the
left atrium through
pulmonary veins and pulmonary left
veins atrium
then Out to the rest of right
the body through the atrium
aorta. The Cardiac
vein
output of the heart is
measured by the volume of blood pumped by the heart per unit time. i.e.,
the rate of flow of blood into the aorta.
The dye dilution method is used to measure the cardiac out put. Dye
is injected into the right atrium and flows through the heart into the
aorta. A probe inserted into the aorta measures the concentration of dye
leaving the heart at equally spaced times over a time interval 10, TI until
the dye has cleared. Let c(s) he the concentration of the dye at time t. If
the interval [0, fl is divided into a large number of subintervals of equal
length A t ,then the amount of dye that flows past the measuring point
during the subinterval t = t to t = i i is approximately concentration
xvOlumc = c(t1)x(FA t),
where F is the rate of flow that we are trying to determine.
A FEW REA L-UFE EVENTS EX PLA INED BY CA LCULUS

Thus the total amount of dye = c(11)FA t.

Making n -300, the amount of dye A is given by
T -
A = F > c( i 1) & = Ffc(t)dt
lot 0
A
so that F =
fc(z )dt
0

Thus the problem reduces to the evaluation of a definite integral.

As concrete example, if A = 5, A t = I, T = 10.

F=
-j -, by numerical integration using Simpson's Rule.
fc(t)dz
0
= 0.12 litre/Second.
IX. A pplication to the Technology of DA M.
A dam has the shape of a trapezoid as shown in the figure below. The
height of the dam is 20 metre and width is 50 in the top and 30 m at the
bottom. We are to determine the force on the dam due to hydrostatic pressure
when the water level is 4-rn from the top of the darn.

I0

i. 16--i

(b)
(V
We choose a vertical x-axis with origin at the surface of water as shown
in the figure
The depth of water is 16m, so we divide the interval [0, 161 into
subintervals of equal lengths with end points x- and e [xa.x j ]. The ith
horizontal strip of the dam is approximately a rectangle with height & and
width ivi . Now from similar triangles [See fig. (b) above]
a 10
xvi DIFFERENT/A L CA LCULUS

and so w =2(l5+a)=46-j

If A i be the area of the ith strip, then

A , =wx=(46-i1 )Ax .

If Ax is small, then the pressure Pi on the ith strip is almost constant,

and

= pgd = I000gi,.
The force F due to hydrostatic pressure is given by

= Force xarea = P1 A 1 =l000gi,(46-11)Ax.

Adding the forces and taking the limit as n —*

F= urn 1000gI1(46-i1)ix

16
= Jl000gx(46- x)dx

= 1000 x9.8[23x2
]
=4.43xIO7 Newton.

X. Differential Equation to solve problems of Electric circuit..

Let us consider an electric
R
circuit having a resistance R, an
inductor L and a capacitor c is
series.
A battery or generator supplies
I..
an electro-motive force E.
If the charge on the capacitor

at time i be Q(t), then current

C

=I= and the drop of voltage

across the resistance, inductor and capacitor are RI. L-'- and
di C

respectively.
 A FEW REA L-LIFE EVENTS EX PLA INED BY CA LCULUS xiii

By Kirchoff's law
L-+RI+= E(t)
Since != LQ , the above equation becomes
dt
(1)

which is a second order linear differential equation with constant

coefficients.
Let us move to particular case, where R = 40g.
L = IH, C= 16x 10-4 and E(i)= lOO cos (lOt) and initial charge and
current are both zero.
Then equation (1) becomes

= lOOcos(lOt)
cit 2 di
The auxiliarly equation is n2 +40m + 625 = 0
which gives m = —20±15i
And the general solution of (I) turns up to be
Q(t) = e-20' (Acosl5t + Rsin 15t) + 4_ {21cos(10t) + l6sin(l0t)}.

The itial conditions Q 0, =0 at r = 0 gives

dt
84 464
697 2091
So,
4 f_.- {—ó3 cos (IS,)—! 16sin(!5t)+21 cos(lOt) + l6 sin (lOt)}1
Q( t) =

and
= [-{_I920cos(15r)+ 13060sin(15t)}

+120116cos(10f) - 21sin(l0t)}]

For large values oft, e 20' - 0.

Q(i) {21cos(10t)+ 16sin(lOt)}.
B-
 DIFFERENTIA L CA LCULUS

X L W hen a dog chases a rabbit.

A dog sees a rabbit running in a straight line across an open field and
gives chase. In a rectangular co-ordinate system (as shown in the figure
below), we assume:

.0
.0
C,

'I (L,o)

(i) The rabbit is at the origin and the dog is at the point
(L. 0) at the instant
the dog first sees the rabbit.
(ii) The rabbit runs up the y-axis and the dog always runs straight for the
rabbit.

(iii) The dog has the same speed as the rabbit.

If the dog moves along the path whose equation is y = 1(x) ,,then it can
be shown that y satisfies the differential equation
d2y ( dy'2
x— =l 1+— i
(I)

With the initial conditions " = 0, = 0 when x = 0

CLX

Llog -
4L 2 .L
It will be interesting to find that the dog will never be able to catch the
rabbit.
 LIST OF IMPORTANT EoRMuL

I. Important Limits.
sin x
(i) L i = 1, where x is in radian measure.
-4O x

(ii) L i (i+ !] =e ar, Li (i + =e.

n

(iii) Li -! log (l+ 9=1. (iv) L i?_ =1

-) x _.* o x
(1+x) -1
(v) Li — =n (vi) Li
-o x '-.- n!

(vi) Lf x=O(-<z1).

II. Standard Derivatives.

(1(1 L
dx -, dx X)

d d/-\ 1
dx TX 2.J
di\
e; —ta )=a log, a.
dx dx
d d I
— (log 4 = —; ( X) = — log o e.
dx x
d.
— sin x) = cos x; d- (cos
I x) - sin x.
dx dx
d 2 d ( 2
(tan
—x)= sec x; _x)r-COSCC
Cot , X.
dx dx
d
-- (sec x) = sec x tan x; (cosec x) = - cosec x cot x.
dx A
I d( I
(sin x)=- --os X3
dx dx
 xx DIFFERENTIAL CALCULUS

d 1d 1 -1 I
-tan xj - cot xj = - ______
dx ' 1+x dx 1+x2
-,I
x)= d (cosec
d(_ -\ I
dx
xj=
xJx -i dx
d
—(si nli
dx
9= cosh x; (1
—(cosh
dx
4= srnh x.
d
- (tanh 4 = sech - x d
—.(coth x) = —cosech2 x.
dx dx

9 = -seth x tanh x;!_(cosech 9 = - cosech .r coth x.

-f-- (sech
dx (IX
d. I I
- tsinh
-1
xi
di
icosh XJ = x> I
dx ' 2 iL

(x <1);

di _-
—cosech x)=-_
I
dx

III. Fundamental theorems on Differentiation.

(1) (ii)

.(iii) 4(x)±v(9}= '(x)±w'(x).

(iv)-^- (O (4x,,ti (9}= (x ),i'(x)± w(4ø'(x).

Derivative of the product of two functions
=first function x derivatvc of the second
+second function x derivative of the first.
(v)

= (x)(4 (4. ...J+ 0, (x){01 (x)ç (x) .... }

 xx
USTOF IMPORTANT PORMUL

= '(x,(x)_vi'(4(x) ,(x)* 0.
)}
(vi) .±{±_
dx
Derivative of the quotient of two funCtionS
= (Den y , of Num.)xDeflOm.- (Den y , of Denom)xNum.
(Denom.)2
(vii)lf y = f(v), v = q (4,
dydy dv
dx dv dx
Ady dx
1^1(dy dy
-and-^ 0
(viii) —Xi.e,—.= l dy dx
dx dy dx

IV. Meaning of the derivatives and differential..

the tangent at any point
(1) ". = tan its, where 'i is the angle which
(Lt
of the x-axis.
to the curve y = f (x) makes with the positive direction

dy
= rate of change of y with respect to x.
dx

dy = f'(x)dx.i1y =j(x).

V. The nth derivatives of some special 1unctIoiS.

D" n!.

x (m 1iflg aposWive integer > ")•

(rn-n)!
D(e')=a"e"; D(?)=e'.

Dn( —L) (_!!;{

x+a (x+a)"
log (x+)}(1T.
(x+ar

D" {sin (ax + b)) = a" si(2!!. + ax+ b).

D'1 {cos (ax +b)}=a" cos. (j0I7)'.

D" (sin ax) = a sin(!!. + ax); D (cos.ax)

= a" cos(- . + ax).
xxvi DIFFERENTIAL CALCULUS

D' (e sinbx) = ( a2 + b27/'2e sin (bx + n an' b/a).

D' (e ' cOS bx)= (a2 + li2)/2e cos
(0-11
+n can' b/a).

D(2 +a2) (-1)n!., o sin (n+ i)o,

where 0 = tan (aix).

D (tan'
4= - 1)" n —. I) ! sin" 0 sin no where 0 = x.
Lelbnftz's Theorem
(uv), = UV+CU,_1 v1 +'c2U, 2 l 2 +

VI. (I) Mean Value Theorem.

f(x+h)=f(x)+,f'(x+oh) 0<0<1.

(II) Taylor's Series ( finite form),

f(x+h) f(x)+hf'(x)+!!f'(x)++ (fl-1)!)

RemalnderR. = + oh), 0<0 <1 (Lagrange 'sforrn)

h (i —e) f(x+eh) (Cauchy'sform)

(n—I)!

(x- a)2
f(x)= f(a)+(x_q)f '(a)+ f(a)+
2!

(x—' f._I(a)+(x_aY' fa+o(x-a)),

+
(n-I)!

0<0<1.
RemalnderR, (x—a'

(Ill) MaclauEin's Series (finite for,,,)

(x)= f (o)+ xf"(o)+ j'(o)+

+ x '-I
f"'(0)+±_f"(Ox), 0<0<1.
(n-I)! n!
 xsiii
LIST OF IMPORTANT FORMULAE

(iv) Taylor's Series ( extended to infinity)

f(x+h)f(X)+hf'(X)+f(X)+ .... . to °.

(v) Maclaurin's Series (extended to infinity)

...................
(vi) Expansions of some well-known functions In series.

X x2
We

(b) Sill X = X - - ......+ i) x2

2n+
( 1)! for

(c) Cos x=x t +- ,-- ..... . +(-i).-+...

2.! values

x.3 x of X.
(d)sinhx=x+-+--+ .....

(e) cosh x=x+-+-+....—+)+

4 21

(1) log (i+x)=x--+j- .......

for -1<x51.
for -1< x <1.
(g) (1+ x) = T X+X 2 - X 3 +...
(I-x)' =I+x+x2 +x 3 +...
for -I <x <1.
for -1< .x <1
(I +x) -2 =1-2x+3x 2 -4x3 +...
1 1.3 2 1.3.5 for -1< x <I.
I-x) =I+—x+ — X +-X +...
2 2.4 2.4.6

(h.tan'x=x_j-+j-_...+(-1Y 1+.
for' -15 x !^ I.
1 x3 1.3 x5 for -1 !^ x !^ 1.
(i) sin
xxiv DIFFEREWT(AL CALCULUS

VII. Maxima and Minima.

(i) If f'(c)= 0 and f'(c) is negative, then
f (4 is a ?flaxi?flum for x = C
(ii) If f'(c)= 0 and f'(c) is positive, then
f x ) is a minipnu,n for x = c.
(iii) If f 'I:c) = f ' c) = ... = f ' c)= 0 and j ' (c);6 0, then
(a) if r is even, then f (x) is a maximum or a minimum for
x = c, according as f' (e) is negative or positive;
(b) if r is odd, there is neither a maximum nor a minimum for
f (x) at x =
(iv)A lternative criterion for maxima and minima:
(a) f (x) is a maximum if f 'x-i- i) changes sign from
+to - and
(b) f (x) is a minimum if f'(x + h) changes sign from
- 10+,
as h, being numerically infinitely small, changes from - to +.

VIII. Indererminate forms.

(i) Form

Li 0 W = Li
.
( V Hospital's Theorem
.r-.Ly (x) w ' (- ) '
(ii) Form .., Li Li
-*'. V -' w (4
W. Partial Differentiation.
(i) If u = f(x, y) and if j ,, f,., exist and J,,, ( or f ) is
O ii
continuous, then -a2ii =
ax i)Y ay ax
(ii) If f(x, y) be a homogeneous function of degree n in x, y, then

+ Y af = nf(x, y). [£u!ers Theorem]

ax ay
aj aj —aj
Similarly, x— + y—+ z = nf(x.y,z).
ax ay az.
 .111.
LIST OF IMPORTANT FORMULAE

(iii) if J(x,%') =0, ±- = _L_ (i *o)

(iv)If u=f(x,y),whrex=Ø(t),y=(s),
du au dx au dy
—=—.—.--.. +—
di ax di ay di
(v) if u = =
j(x, y), du J d. + fdv.

Similarly if u = f(x, y, z = du jdx + f ),dy + )dz. -

(vi) If u= f(x 1 ,x 2 ), where x 1 =Ix,y),x 2 =2x,.y),

aU aUaXaU aX2
aax, ax aX2 ax' ay a 1 a ax 2 ay

Tangent and Normal.

(i) Equation of the tangent

Y- y = - x), or, (x - x)f +(y - )f = 0.

(ii) Equation of the normal

y-y=-(X-x)
dy f,
(iii) Cailesian suhiangens = y/v 1 subnormal = yv1

Length of tangent = y1 + y2 /y1

Length u/ normal =y,J I +
(iv) A ir-differential (Cartesian)
dv . dv
— =COSIj/, --=SIflIJI, —=Lanl.
ds ds dx
2
dv .d\2 dy2 ds2

( )+(-J =(.)
=1;
+(). di

±=
x
d
as I+ dx ) dy !
F
Fl^
ds2=d':2+dy2
(v) A ngle between tangent and radius vector (0 )
rdO
in = -; co9=—,
di
tan4i=—;
rdO
sØ
dr ds ds
XXVI
DIFFERENTIA L CA LCULUS

' =0+0.
(vi) A rc differential ( Polar)
2 f 2
Ir (d
dO ' d0) dr dr )

ds= Ir2+Ido; ds=F

d0dr I+( )
ds 2 = dr2 + r2dO 2
2d0 dO I
(vii)) Polar subtangent = r - = --. where r = -.
dr du u
dr
Polar subtangeni =
dO
Perpendicular from pole on tangent (p )

I I I Idr 2 (du\2
p=rstn0 — =— +— 2 rI-----1 =u +1—
4 dO) d0
(viii) Pedal equations of some well-known curves
(1) Circle x 2 + y 2 = a 2 (centre) ...r = p.
(2) Circle x 2 + y 2 = a 2 (point on the

circumference) r 2 2ap..
(3) Parabola y 2 = 4ax-(focus) p 2 = or.

(4) Ellipse x + !.. = I (focus) . -- = -

a b P2 r

V
(5) Ellipse (centre) =a2+b2.
a 2 b2

(6) Hyperbola a2b2 = I (focus) •..._ = +1.

(7) Hyperbola. j __Z=i (centre) ... 4__r2 =a 2 P.

a b p
(8) Rect. Hyperbola x 2 - = a 2 (centre) p 2 = ar.
2a
(9) Parabola r = (focus) p 2 = ar.
I ± cos I)
(10) Cardioide r = aØ ±cosO) (pole) ... r 3 = 2ap2.
USTOF ,MPORTAWTfORMULA3

'2r.os2O 1
(1 1) Lemniscate •••r 3 =a 2p.
r2 : :2 sin2O
N -
r a" cos nO
(12) r" (pole)
=a n sin nO

X1. Curvature.
ds1
p= Lf(W)].
dW

[y=f(x)];y2 O.

= Y 12
[x y = w(' )J.
.ty -yx

p= 2 [f(x,y)=O].
-2fff, +ff

= (r2+ri2)
[r=f(0)I.
r2 +2r 12 - rr2

(2+2)
[u=f(0)].
.. (u+u2)

p rdr dp
.(r)].
=_ [p= j— p=p+
.-i [p=i(w)L
4 d 'ji

p= 11!– [attheorigin,x-axis(y = o)beingtangentj.

2y
 DIFFERENTIA L CA LCULUS

p = Li
f- I
at thc origin, v - axis(x = O)being tangent].

2 2
2
= +b .Li X (at theorigin,
ax + by

ax+by=O being tangent ].

Chord of cun'alure through the pole 2p sin çb.

Chord

Centre
of curvalure

of curvature
I parallelto x - axis = 2p sin v'.
paralletto r - axis = 2p cos V.

- 2
- .' ( I +Yi - Yi
x=x — y=y+1 Ly=j(x)]
Y2

- i+x12 - x1(I+x12) (i(,)1

x=x— , y =y —
Y2

x=x—p sin vl , y=y+p coi.

d2s
Radius of curvalure of the ewilute =
dW
 THE HISTORICAL DEVELOPMENT OF CALCULUS
[] ,

• The English and German mathematicians. respectively, Isaac

Newton and Gottfried Wilhelm Leibniz invented calculus in the 17th
century, but isolated results about its fundamental
problems had been known for thousands of years.
For example, the Egyptians discovered the rule
for the volume of a pyramid as. well as an

I -..'
approximation of the area of a circle. In ancient
- Greece Archimedes proved that if c is the
circumference and d the diameter of a circle, then

Isaac Newton - 3!d <c<3 1 . His proof extended the method

7 71
of inscribed and circumscribed figures developed by the Greek
astronomer and mathematician Eudoxus.
Archimedes used the same technique for his other
results on areas and volumes.Archimedes discovered
his results by means of heuristic
arguments involving parallel
slices of the figures and the law
of the lever. Unfortunately,z-,the f ^k
Method was only rediscovered in G. W Leibnjz
the 19th century, so later
mathematicians believed that the Greeks deliberately
concealed their secret methods. During the late middle
ages in Europe, mathematicians studied translations
Archimedes of Archimedes' treatises from
Arabic. At the same time, philosophers were
studying problems of change and the infinite, such
as the addition of infinitely small quantities. Greek
thinkers had seen only contradictions there, but
medieval thinkers aided mathematics by making the
infinite, philosophically respectable.
By the early l7t'h century, mathematicians had
developed methods for finding
f areas and
i volumes F. B. Cavalien
of a great variety of figures. In his Geometry by Indivisibles, the Italian
xrx DIFFERENTIAL CALCULUS

mathematician F. B. Cavalieri (1598-1647), a student of the Italian

physicist and astronomer Galileo, expanded on the
work of the German astronomer Johannes Kepler
on measuring volumes. He used what he called a
"indivisible magnitudes" to
investigate areas under the
curves. Also, his theorem on the
volumes of figures contained
between parallel planes (now
called Cavalieri's theorem) was J. Kepler
known all over Europe. At about
the same time, the French mathematician Rene
• Descartes' La Geometric
appeared. In this important work,
John Wallis Descartes showed how to um,
algebra to describe curves and obtain an algebraic
analysis of geometric problems. A codiscoverer of
this analytic geometry was the French/
mathematician Pierre de Fermat, who also
discovered a method of-finding
the greatest or least value of Isaac Barrow
some algebraic expressions-a
method close to those now used in diffefential
calculus. About 20 years later, the English
mathematician John Wallis (1616-1703) published
The Arithmetic of Infinites, in
which he extrapolated from
Bernoulli patterns that held for finite
processes to get formulce for
infinite processes. His colleague at Cambridge
University was Newton's teacher, the English
mathematician Isaac Barrow (1630-77), who
published a book that stated geometrically the
inverse relationship between problems of finding
tangents and areas, a relationship known today as
5") Eluer

the fundamental theorem of calculus.

Although many other mathematicians of the time came close to
discovering calculus, the real founders were Newton and Leibniz.
THE HISTORICAL DEVELOPMENT OF CALCULUS wI

Newton's discovery (1665-66) combined infinite sums (infinite series),

the binomial (q. v.) theorem for fractional
exponents, and the algebraic expression of the
inverse relation between tangents and areas into
-. methods we know today as calculus. Newton,
however, was reluctant to publish, so Leibniz
la - became recognized as a codiscoverer because he
published his discovery of differential calculus
in 1684 and of integral
ç
calculus in 1686. It was
A. L. Cauchy
Leibniz, also, who replaced
Newton's symbols with those familiar today. In
the following years, one problem that led to new
results and concepts was that of describing
mathematically the motion of a vibrating string.
Leibniz's students, the Bernoulli family of Swiss
mathematicians (Sec Bernoulli, Daniel). used N. i. Lobachcvsky
calculus to solve this and other problems, such as
finding the curve of quickest descent connecting two given points in a
vertical plane. In the 18th century, the great
Swiss-Russian mathematician Leonhard Eider,
- who had studied with Johann Bernoulli (1667-
- 1748), wrote his Introduction to the Analysis of
Infinites, which summarized known results and
also contained much new material, such as a
strictly analytic treatment of trigonometric and
exponential functions.
Despite these
P. G. L. Diilcblet advances in technique,
calculus remained
without logical foundations. Only in 1821 did
the French mathematician A. L. Cauchy
succeeded in giving a secure foundation to the
subject by his theory of limits, a purely arithmetic
theory that did not depend on geometric intuition
or infinitesimals. Cauchy then showed how this J. W. R. Dedekind
could be used to give a logical account of the
ideas of continuity, derivatives, integrals and infinite series. In the next
xUii DIFFERENTIA L CA LCULUS

decade, the Russian mathematician N. 1. Lobachevsky and German

mathematician P. G. L. Dirichlet (1805-59) both
gave the definition of a function as a correspondence
between two sets of real numbers, and the logical

SL
foundations of calculus were completed by the
German mathematician J. W. R. Dedekind (1831-
19 16) in his theory of real numbers in 1872.

Gauss
 DIFFERENTIAL CALCULUS

AN IDEA OF NUMBER SYSTEM

ILIII!
1.1 Introduction Calculus - origin and extension.
Calculus, fundamentally different from Arithmetic, Algebra or
Geometry, is essentially concerned with change and motion; calculus
deals with quantities that approach other quantities. When there is
continuous and gradual change, however small the change be, Calculus,
with its novel concept of 'limit' and limiting operations, is the right
mathematics to apply..
Though invented initially to mccl the mechanical or geometrical
needs, today Calculus and its extensions in mathematical analysis are
far reaching. Besides being used in theoretical fields ofenquiry, Calculus
is now used in determining the orbits of artificial satellites and space-
craft, in predicting population size, in estimating how fast the price of an
agricultural cohimodity rises, in forecasting weather, iii measuring the
cardiac output of the heart and in a large variety of other areas.
However, diverse be the area of application of subject, the common
theme is the way or manner in which one quantity changes with another
when the change in the later is very small or, more properly, with the rate
of change of one quantity with the other.
In these investigations one has to deal with the relations bctweer
pure numbers which represent the magnitude of the quantities. That is
why we begin our study of Calculus with a short discussion on number
system.
1.2. Numbers.
The earliest concept of numbers originated from counting, and the
first set of numbers which was known to men, was the set of positive
integers. The arithmetical process of subtraction needed an extension
to negative integers, and zero was included as a number. The process of
division required a further extension to rational numbers, which are
defined to be numbers of the form al where in and n are integers,
ultimately prime to each other, n being positive and not equal to zero. It
may he noted that terminating decimals, as also recurring decimals, which
are expressible in the form fall under this category.
2 DIFFERENTIAL CALCULUS

1.3. Geometrical representation of rational numbers; rational points.

I 1111111 I X
X' p' 0 A
Fig 1.2.1
7
Take a line X OX extending indefinitely in either direction, for
reference, and a suitable point 0 on it as origin. A suitable length OA on
it being chosen as unit, if we divide OA into n equal parts, and take a
length OP (or OP') equal to m such parts (towards the right of 0 if m
be positive, and towards the left if in be negative), the length OP
(Or OP'), or the point P (or P', as the case may be) represents the
rational number . The point P, representing a rational number, is called
a rational point.
1.4 Properties of rational numbers.
(I) Rational numbers are well-ordered. This means that of two
unequal rational numbers a and b, either a > b or a < b; also if
a> h and b > c, then a> c ,ctc. In other words, rational numbers are
well arranged in respect of their magnitudes, points representing higher
numbers always falling to the right of those representing smaller ones,
and vice versa, in their geometrical representation.
(ii) Rational numbers are everywhere dense; in other words, between
any two rational numbers, however close, or within any interval on the
axis representing rational numbers, however small, there is an infinite
number of rational numbers or points.
This may be easily seen from the fact that, however close the two
rational numbers a and may be, -(a+b) is a rational number lying
between them. Similarly, between a and -(a+b), as also between
(a + b) and b, we can insert rational numbers, and so on. Thus there
is an infinite number of rational numbers between a and b.
1.5. Irrational numbers.
Whereas all rational numbers are represented by points on the axis,
and though in any interval, however small, there is an infinite number of
rational points, still the converse, that every point on the axis must
represent some rational number, is not true;
e.g., OP is not rational.
AN IDEA OF NUMBER SYSTEM 3

OA being unity, if AB be taken at right angles to OA and equal to

it, OR is joined and on OX, OP be cut off, equal to OB. then OP
represents a number equal to 'I , which is not rational.

, I

() A p y
Fig 1.4.1

Proof: For, if r2 = rn/n where m, n are integers prime to each

other, then in = 2n 2 .showing that m 2 and so is an even integer (for,
the square of an odd integer is evidently odd). Let m = 2rn' where m'
is an integer. Then we get n 2 =2m 2 and son is also an even integer.
Thus m and n, which have a common factor 2, cannot be prime to one
another, thus leading to a contradiction.

Similarly, equations like x 3 = 7, 4x 4 = 13, etc. cannot be solved in

terms of rational numbers alone. Besides radicals, there are other types
of number like e, n, . . . (called transcendental numbers) which arc not
rational.
There are, therefore, numbers other than rational numbers, which
are called irrational numbers, thus leading to a further extension of
numbers.
1.6. Relations of irrational numbers to rational numbers;
representation of numbers (rational as well as irrational) as sections
of rational numbers.
Consider the number J2 There is no rational number whose square
is 2. The system of rational numbers, therefore, can be divided into two
classes, say L and R, such that all numbers of the L-class have their
squares less than 2, and those of the R-class have squares greater thati
2. Hence, every number of the R-class > every number of the L-class.
Thus, I, 1 . 4, 1 41, 1 . 414, 1-4142 ..... belong toL-class
and 2,15. 1-42, 1-415,14143.....belong to R-class.
The differences of the corresponding numbers of the two classes
are, respectively.
1,0 . 1.0-01,0-001,0-0001 .......
4 fl/lIE/tENT/A L CA LCl/l.US

Proceeding in this manner (by expressing ./5 in a decimal form,

which will lead to an endless decimal not recurring, and choosing the
rational numbers of the two classes by stopping at any stage) we can
find a member of the L-class and a member of the R-class which differ
from one another by as little as we please. Our common sense notion,
therefore, demands the existence of a number x, and a corresponding
point P on the axis, such that P divides the class L from the class R.

But this number x is not rational and belongs to neither of the two
classes. Further, x 2 is neither > 2 nor < 2.

For, if x 2 > 2 , let x = 2 + C . Then however small c may be, we can

get rational numbers of the R-class whose squares being >2 will differ
from 2 by less than c. Such rational numbers of the R-class will lie to the
left of .v, and so the assumption that r is the point dividing the two
classes is untenable. Similarly, x 2 , 2

X = 2, or, .v = and being not rational as proved before, it

belongs neither to class L nor to class R. The point P is thus only a
point of Section of the two classes of- rational numbers L and R defined
before, not belonging to either class, and representing the irrational
number ' j -

This leads to a new idea of defining numbers as sections of rational

numbers, as follows

"If by some means or other we divide all rational numbers into two
classes L and R, such that each class contains at least one rational
number, every rational number belongs to either L oi' R, and each
number belonging to R-class > every number of the L-class, then we
obtain a section of rational numbers which defines a number, rational or
irrational; the particular mode.of division defines a particular number by
its section."

Three cases may arise : (i) That L-class has it greatest number, but
the R-class has no least; e.g., let all rational numbers > 5 belong to R-
class, and the number 5, as also all rational numbers <5 belong to L-
class. The section in this case represents the rational number 5, which
belongs here to one of the two classes, namely, the L-class. (ii) The L-
class has no greatest number, but the R-class ha' a least one; e.g., all
rational numbers <-3-5 belong to L-class and 3 . 5 with all rational
numbers greater than this belongs to R-class. Here the section represents
the rational number —3 . 5, and the numbcr itself belongs to R-class. (iii)
AN IDEA OFNUMIJER SYSTEM 5

The L-class has no greatest number and the R-class has no least number;
e.g., all rational numbers whose cubes are < 7 belong to L-class, and
those whose cubes are >7 belong to R-class; there is no rational number,
as can be shown, whose cube is equal to 7. The section in this case
represents the irrational number V7 and belongs to none of the classes
Land R which consist of rational numbers only.
It may be noted that the case in which the L-class As a greatest
number and the R-class has a least number simultaneously is not possible,
for otherwise, 'between these two rational numbers there would be an
infinite number of rational numbers as proved before, and they would
belong to none of the two classes.
This extension of our conception of numbers as sections of rational
numbers gives us a more satisfactory basis of defining all numbers in a
uniform way. We no longer think of numbers as isolated members, but as
an aggregate of rational numbers divided into sections.

1.7. Real numbers.

All kinds of numbers, rational as well as irrational, positive and

negative, including zero, constitute what are called real numbers.
The contents and classification of real number s y stem will be
understood at a glance from the scheme given below.

Real Numbers

Rational Numbers Irrational Numbers

I 'I.
Integers Fractions

It may be thought that just as from rational numbers, by dividing them

into two classes by sections, we get, in addition to rational numbers, a
new type of numbers, namely irrational numbers; similarly by sections
of real numbers again, we may expect a further extension of numbers.
But this is not true. In this Connection we state the follou ing theorem
(given in the next page)
6 DIFFERENTIA L CA LCULUS

Dedekind's theorem (On sections of real numbers)

If real numbers be divided into two classes L and in such a way that

(i) every real number belong.s to one class or the other,

(ii) each class contains at least one number, and
(iii)any number of the L-class is less than every number of the R-
class,

then there exists a real number 'a' which effects this section, i.e.. which
has the property that all numbers less than 'a' belong to L-class, and
all numbers greater than 'a' belong to R-class, the number 'a' itself

I I
may belong to either class.
For proof, see lifard.vs Pure Mathematics.

Thus as sections of real numbers we get real numbers alone (unlike

that in case of rational numbers), and not any other new type of numbers.
Thus no further extension of numbers is possible; and the aggregate
of real numbers is complete. The correspondence (one to one) between
all the points on the line X'OX without exception (called the linear
continuum) and the system of all real numbers, rational and irrational
(constituting what is called the arithmetical continuum), is now perect.

1.8 Fundamental Properties of real numoers

Properties involving 'addition' and 'multiplication' of real numbers

(i) If a, bare any two real numbers, then (a + b) and ab are also real
numbers, i.e.,
(a-i-b), ab€R, foralla,heR.

(ii) a+bb+a and ab=ba, for all a,bER..

(in) a+(b+c)=(a+b)+c and a(bc)=(ab)c,for all a,b,c€ R.
(iv) The real number 0 (which is an integer and a rational number) has
the following properties:
a+0=O+a=a, aM=Oa=0 for all aR.
Division by zero is meaningless in the set of real numbers.
(v) For every real number a, there exists a real number -a, such that
a + (—a) = —a + a = 0.

(vi) For every real number I (which is an integer and a rational number)
has the following properties:
a.l=1.a=a for all a€IR.
AN IDEA OF NUMBER SYSTEM 7

(vjj) For every a (.0)e R, there existS!E R, where a.— = —.a = I.

a a a

(viii) a — be R forall a,b R, wherea — b is defined as

a - b = a + (— b).
(ix) a..(b+c)=a.b+a.c,foralla,b,cER.

(x) If a is any real number ( 0) and b is any real number, then

- is defined as
a a a
(xi) a.(— b)=(— a).b— (a.!') and (—a).(—b) =ab,forall a,b€R.

(xii) l0.
(xiii) For a,b e R, a .b = 0 at least one of a and I, must be zero.
1.9 Properties regarding order relation in R.
(i) For any two real numbers a, b one and only one of the statement:
"a>b","q<b","a=b" must .bc true.
(ii) a>b and b>c = ' a>c for all a,b,c€IR.

(iv) a) band c>0 =ac>bc, for all a, b, c€RR.

(iv) a>b=a+c>b+c, for all a,b,c'ERR.

(v) a>0,if and only if—a<0.

(vi)a>b, ifandonlyifa—b>0 and a<b, ifandonly if a—b<0,
where. a,bE R.
(vii) a>b and c.<0=tac<bc,where a,b,c€R.

(viii) a2 2:0 for all a E R (X 2! y means x> y, or, x = y)

(m) Between any two distinct numbers, there exist an infinite number of
rational numbers as well as an infinite numbers of irrational numbers.
(x) If a (>0) and bare two real numbers, then there exists at least one
positive integer it such that it a > b.
Note: 1.. For any real number a, one an only one Of'a>O', 'a<0','a=O'
must be true.
2. A real number a is said to be positive or negtive according as a>0,
or, a <0. -
 DIFFElENi1AL cALCULIJS

3. It should be remembered that the real number 0 (which is also a

rational number) is neither positive nor negative.

1.10. Integers
In Art 1. 2, it has been discussed that in the process of counting number
of elements ofa finite set (e.g., the number of rooms in a house, the number
of trees in a garden, the number of students in a class, etc.) the natural
numbers denoted by the symbols, 1, 2, 3, . . . we obtained, and the set of
natural numbers is denoted by N.
Now for every ntbral numbern, the numbers given by the symbol —n
is introduced together with the introduction of the number 'zero' expressed
bythesymbol 0, where x+O=O+x=x forevety XE N. U{—nn c NT u(OI
and , i•(— n) = — fl +n = 0 f&every xc N.
Elements of the set No (—a n c N) u(0) are called integers- The set
of all integers is denoted by Z.
Remarks : The integer 0 is neither positive nor negative.
Factors of an Integer

An integer a (* 0) is called a factor ora divisor ofan integer b. ui, can

be expressed as b = a c (or, ca), for some intcgerc. In this case c will also be
a factor of b if c* 0

For example, as 2.3 = (-2) (-3) = 6, 2, —2, 3, —3 are factors o16.

We observe that if a is any non-zero integer, then 1. - I, a, - a are
factors of a.

Prime Numbers
An integerp (> 1) is called a prime number if p has no factor besides
—1, p. - p. For example, 2,3,5,7.. . . are prime numbers. The inteeer 2 is
the least prime number.

Remark: From the definition of prime numbers, it follows that the integers 0,
I are not prime numbers.

Relatively Prime Numbers

The integer a and b are said to be relatively prime or, co-primes or,
prime to each other if the integer I is the only positive integer which is a
common factor of a and b. We note that 6 and 29 are relatively prime integers,
while, 9 and 24 are not relatively prime to each other.
AN IDEA OF NUMBER S}7EM 9

An Important Property of Prime Numbers

If p be a prime number and a, b are integers where p is a factor of the
product ab, then p is a factor of at least a and b.

Even and Odd integers

It can be shown that any integer is either of the form 2m, or, of the form
of 2m-l- 1, where in is an integer.
An integer of the form 2,n (where in is an integer) is called an even
integer, while an integer of the form (2m + I), (where n, is an integer) is
called an odd integer.
Thus 0, ±2, ±4, ±6,.. . are even integers and ±1. ±3, ±5,. .. are
odd integers.

Note : The integer 0 is an even number.

1.11 intervalsinlR..

Let a, be Rand a <b.

(i) The set (x:x E Rand a<x<b) is denoted by (a, b)and is cal led as
pen interval.

(ii) The set (x: x ER and a !-< x !^ b) is denoted by Ia, b and is called
an closed interval.
(iii) The sets (x:xE Randax<b) denoted by [a, h) and
x lR and a < x !^ h) denoted by (a, b] are called Semiopen or
Semiclo.sed intervals.

The set R is also regarded as an interval and is denoted by (-. oo)

where we write R= {x-oo< x<oo).

It should be noted that the symbols do not represent real

numbers.

For any real number a, the set (x x IR andx ^: a) is denoted by

[a, oo) and this is a semiopen interval.

Similarly the sets (x:x€ R and x> a) . (x:xE Rand .s !^ a)

x X Rand x <a) are respectively denoted by (a, oo) , (_oo, a). (---n a)
and these sets are also called intervals.
 to DIFFERENT/AL CALCULUS

1.12. Complex numbers.

In order to fill up the gaps and bring about a uniformity in the theory of
equations, as also in all other theories of higher mathematics, it has
been found necessary to introduce a class of numbers, called complex
numbers. A complex number has been defined by modern
mathematicians as an ordered couple of real numbers, i.e., a pair of real
numbers united symbolically in a particular order for the purpose of
technical convenience. Thus a complex number is, strictly speaking, not
a single number at all, but a pair of real numbers with a proper order. If
the order is reversed, we get a different complex number. A complex
number may be expressed in the Form I a, h i, where a and b are two real
numbers. It is also represented. For convenience, in the form a + ib,
where the symboli has no meaning by itself; it merely indicates the
order in which the real numbers a and h are considered. In defining all
ordinary algebraical operations with regard to complex numbers it has
been found convenient to associm m e the symbol i with the property,
=_I in which case all operations consistent with the algebra of real
numbers may be applied to the case of Complex numbers.
For geometrical representation and further introduction into the
algebra of complex numbers see Chapter VI, Das and Mukherjees' Higher
Trigonometry.

1.12 Miscellaneous Worked out Examples

Ex. 1. Prove that JF3 is not a rational number. B. P 1997]

Solution Since, 1< 3 < 4, 1< < 2, which shows that vF3 cannot
be an integer.
Now, if possible, let vf3 be a rational number.
We assume,
q (I)
where q >1 and p and q are positive integers prime to each other.

From (l)=3,i.e., R_=3q

... (2)

Since p and q are positive integers prime to etch other, p 2 and q are
also positive integers prime to each other. Again, since >1, - p-- represents
a rational number, which is not an integer, but 3q represents a positive integer.
So, from (2), we get, a positive rational number which k not an integer is
AN J1)1A OFNUMIJERSYS7EM II

equal to a positive integer. But this is not possible. Hence, our initial
assumption cannot be true, i.e., 63 cannot be a rational number.

Ex. 2. Prove that 1092 6 is an irrational number.

Solution 1092 6= 1092(2x3) =1092 2+1092 3= 1+1092 3.

10 92 6 will be a rational number, if 1092 3 is rational If possible. let us

assume that 10 92 3 is rational and 1092 3= -11 when q * 0 and p, q are
q
positive integers prime to each other.

Now, .• log 2 = =3 or, 2"=

Obviously, 2 and 3 are prime to each other, and p and q are also assumed
3'I cannot hold.
to be integers prime to each other. So, the equation. 2" =

Therefore, log 3 cannot be a rational number.

Since, the sum of a rational number and an irrational number is irrational,

109 6 cannot be a rational number.

Ex. 3. Prove that JF3 + F2 is an irrational number.

Solution: Let us assume the contrary, i.e., J3 + vr2 is rational.
(J+J)(J-J)=i

is also a rational number, since it is the quotient of two rational numbers 1

and (J+Ji).
Thus, vr2 = {( J + J) - (J _62)) , being the difference of two rational
numbers, is rational. Thus we arrive at an absurd conclusion.
Hence, our initial assumption that (.J+ J) is rational is wrong.
So, + is an irrational number. -

Ex.4. Show that no positive integer ni other than a square number has a
square root within the aggregate of rational numbers.
Solution: Let m be a positive integer which is not a square number. Then
we are to prove that Jrn cannot be rational.
 12 I)1F1.EREN71AI, CALCULUS

Since in is not a square numhcr, we have a positive integer ?l such that

fl 2 < in < (n - 1)2

or,
whence it follows that Fin cannot bean integer.
If possible, let us assume that Ftn is rational.

Then ./ = - ,where q> I andp and q arc integers prime to each other.

or, —=m

2
or, = nq.
q ... ()

p and q being integers prime to each other, ,2 and q are also positive

integers prime to each other. Also, since q >1, J?_ is rational number which
is not an integer. 1

But, in and q being both positive integers, mq is also a positive integer.

So. from (I) we get a rational number which is not an integer = a positive
integer, which is impossible.
Therefore, our assumption cannot be true, i.e., J,—), be a rational
number.

Ex. S. (1) If rand s are any two rational numbers, prove that (r+) and
(rxs) are also rational numbers.
[ C. P 2003, B. P. 1992, 93, 95 1
(ii) Give examples to show that the sum and product of two
irrational numbers may be rational or irrational. B. P. 1994

Solution (i) If possible, let r+s=p, where r and s are rational

numbers, while p is an irrational number.
r=p—.v
since the difference of an irrational number and a rational number cannot
be a rational number, so r cannot be it number, which contradicts
our initial condition that r is rational. Therefore, the sum of two rational
numbers cannot be irrational; i.e., (P +s) is it rational number, when r
AN IDEA OF NUMBER SYSTEM 13

and s are rational numbers. Next, if possible, let the product of two
rational numbers be an irrational number.
i.e., let rxs=q, where r and s arc rational and q an irrational number,
r= (s*o
since division of an irrational number by a rational number is not a
rational number, - is not rational, but r is rational Thus q cannot be

an irrational number, it must be a rational number.

• ><.t is a rational number.

(ii) We have F
2 and 63 are two irrational numbers. Their sum

s1i-4 is also an irrational number. [ Sec Ex. 3. above

Again, product of two irrational numbers and

If pi)ssi ble, let JF

6 be a rational number.

then - where p and q are positive integers,

prime to each other and q >1

p 2 =6q.
ó=--,i.e., —
q
since p and q are prime to each other, p 2 and q have no common fior
and so .-. cannot be art integer, but óq is obviously an integer.
'1
Thus, a fraction is equal to an integer, which is not possible.

So, product of two irrational numbers F

2 and 63 is also an irrational
number.

Next, let us consider two irrational numbers 5+,C2 and 5-2

'here sum (5 + + ( s - I3) = 10. a rational number;

and their product (s+J)(s_J)=25_223.

a rational number.
Hence, it is shown that the sum and pioduct of irrational numbers may
be irrational or rational.
 14 DIFFERF.N Tint CALCULUS

Ex. 6. Examine whether log, is a rational number.[ B. P 1999, 2001]

Solution : If possible, let log 10 5 be rational and log 10 5 -, where p

q
and q integers, prime to each other, q > p
P
(lO)q=5

or, 10p=5q

or, 2p•5p=5q
i.e., 2 = 5q-p ... (I)

since p and (q - p) are both positive integers and 2 and 5 are prime
to each other, equation (I) cannot hold.
Hence log 10 5 cannot be rational, i.e., log 10 5 is an irrational number.

EXAMPLE-1
I. Define a rational number. Show that is not a rational number.
[B.P. 1981, '86, '95, '97, C.P. '98]

2. Show that -.i is not rational. [B.P 20021

3. Prove that log 10 7 is not rational.

4. If, ..h + = 0, where x and y are both rational, prove that

x=0=y.
5. Define an irrational number. Give examples to show that the sum
and product of two irrational numbers may be rational or irrational.
-
[B.P.1994j
6. Given that r and s are two rational numbers, prove that
r + s, r' s, rs and r/s (. # 0) are rational numbers.
[B.P. '82, '93, '951
7. Prove that the sum of or the difference between, a rational number
and an irrational number cannot be a rational number.
 FUNCTIONS
I L1I
2.1. Introduction.
In higher mathematics and various branches of science very often we
have to deal with changeable quantifies which are interrelated to one
another, and in many such cases we have occasions to investigate how one
of these quantities changes with a gradual change in the other. For example,
in a given amount of gas enclosed in a cylinder with a movable piston, and
kept at a constant temperature, the volume and pressure are interdependent,
and a change in one produces a corresponding change in the other ; or
again, for a falling particle, the height from the ground depends on the time,
and changes with it; the area of a circle changes with its radius, etc.
In Differential Calculus we deal with the way in which one quantity
varies with another when the change in the latter is ultimately very small, or
more properly, with the rate ofchaA ge of one quantity with another, as also
other allied problems.'
In these investigations we shall be dealing with the relations between
pure numbers which represent the magnitudes (with proper signs) of the
quantities, and not with the concrete quantities themselves, so that the
results will be general in nature, applicable to any pair of interdependent
quantities under similar mathematical conditions.
In the following discussions we shall be concerned with thesystem of
real numbers only, meaning by real numbers, zero, integers, rational and
irrational numbers, positive or negative.

2.2. Preliminary Definitions and Notations.

Aggregate or Set: A system of real numbers defined in any way whatever
is called an 'aggregate' or . 'set' of numbers.
Illustration The aggregate of positive integers; the aggregate of all
negative rational numbers; the aggregate of all real numbers positive or
negative; the aggregate of all rational numbers from —3 to +7; the aggregate
I I 1 1 I 1
of numbers - . - . - . - . - . - ' etc.
I —2 3 —4 5 —6

W hile investigating problems of this type, Newton (in England) and

Leibnitz (in Germany) were independentLy led to the investigation of the
principles of calculus, towards the close of the seventeenth cenflo-11. The
principles of Calculus in some form, were also known to the Hindus in
India much earlier.
16 t)/FFEREN7IALC'ALCIJLUS

Variable: Let be a Symbol used during any ia III e mat ical investigation,
o which may he assigned any numerical value out of a given set of real
numbers. Then xis called a 'variable' or a 'real variable', and the totality of
the values of x constitutes what is called the domain of x.
Illustration: In the expression x !. x may be considered a real variable
whosc domain is the aggregate of positive integers.
Note. Variables are usually denoted by latter letters of the alphabet, such as
X, y, z, U, V, W, g. It, etc.
Continuous Variable: If x assumes successively every numerical value
of an aggregate of all real numbers from a given number 'a' to another
given number 'b', then x is called a 'continuous real variable'.
The donmain orinterval (as it will be sometimes called) of x in this case
is denoted by a, b] or a 15 x < I;.
If t be omitted front domain, it is indicated as a < x < b
III last case the domain is said to be open at the left end, whereas
the domain a !^ .v < b is said to be closed. The interval a <x < b is open
at both ends, a and h being both excluded horn the domain of possible
values of x.

Illustration : III expression /T)i- -i- 3), x is a continuous real

variable whose domain is —3!^ x !^ 5; again, in ,the real
variable x has the interval —2 f^ x <'1 . lii sin -i x, the interval of x is
—l:5x<l.
The domain of the variable x in any expression containing x, as in the
above cases, consists of those values of .r for which the expression has a
definite real value.
The interval [a, bj is very often graphically represented on the x-axis
by means of the length bounded by the two points A (x = a) and B (x = b).

() A Ft (
Fig 2.2.!
The length of the interval [a, hj is obviously AB = OB - OA = b - a.
Constant: A symbol which retains the same numerical value throughout
a set of mathematical operations is called a constant.
Note. Constants (other than numerical constants like 2, -3, e, it , etc.) are
usually deno ted by the earlier Jciicrs of the alphabet, such as a, b, a, 13 . etc.
FUNCTIONS . 17

Absolute Value : By absolute value of a quantity x, as distinguished from

its algebraical value, we mean its magnitude or numerical value, taken with
a positive sign. It is represented by the notation I x I which is =x. 0 or -x
according as x>, or <0.
From the very definition the following results are apparent, viz.,

(i) Ia±b a I + b I or more generally, -

Ia±b±c± ...... l:5laI+lbI+lcI+..........

(ii) Ia±bI 2^ ja - bI, ie., I jal - IbII.

Illustration: I_2l2,l6l6.I_2+6'l< 2+6,
1 - 2 - 61=2+6,1 - 6 - 21 >6 - 2,12-61 = 2- 6, etc.

Note., Meaning of the symbol Ix - al<o.

Sincelx-al<5, ifx>a,x-a<S, i.e..x<a+6;andifx<a,
a-x <5, i.e., a -ö <x. Hence, combining the two, we see that
I x-al <5 means a-S <x< a+5. Similarly,I x-al :55 means
a-5!5x:5 a+S. Symbol 0<lx-a!58 means a-3:5x5 a+S but

x*a.

Thus, j xj <8 means -S <x <5.

Functions: By afunc:ion ofx definedfor a given domain, is understood

a quantity which has a single and definite value for every value of x in
its domain. [See note 1.1

In other words, "If x and y be two real variables so related that,

corresponding' to every value of x within a defined domain, we get a
definite value of y, the,i y is said to be a function of x defined in its
domain."

In this case, the variable x, to which we may arbitrarily assign different

values in the given domain, is referred to as the independent 'variable (or,
argument), and is called the dependent variable (or.function).

[See note 2.1

We shall generally denote functions of x by such symbols as
f (x), V (x), F (x), 0(x), etc., where the mathematical forms of these
funcitons may or may not be obtainable.
18 DIFFERENTIAL CALCULUS

Note 1. When an expression or equation which defines a function givs

two or more values of the function for each value of x, we call the function
multiple-valued. The definition given above refeçs to a single-valued
function with which we are mainly Concerned in all mathematical
investigations. A multiple-valued function, with proper limitations imposed
on its value to be used in any particular investigation, can in general be
treated as defining two or more different single-valued functions of x; e.g.,
y=sin x(—l!5x!5l) can be broken up into y = sinx,
where(i) - . -it ^ y :5
(ii) I n:5y!57t,
'fit 5 )' ^ 4t, etc.:
again y 2 = x can be broken up into y = +,Fx and y = -, and so on.

More generally (without restricting to single-valued functions only).

a funciton of x may be defined as follows:
If two quantities x and y are so related that, corresponding to
values of x, there are values of v, then y is said to be a function of x.

Note 2. If y be a function of the variable x, it will generally be open to us

also to regard x as a function of y by virtue of the functional relation
between x• and y, the proper domain of y being taken into account in this
case, because it may so happen that the domain in which y is defined is not
the domain in which x is defined. For example, y =1FX can be written as
x = y 2 , the domain of xin the former relation being x^, 0, and that of y in
the latter is the aggregate ofall real numbers, positive or negative.
In the latter case, y will be the independent variable, x the dependent
one.
Note 3. A function may be undefined (i.e., may not have a definite value) for
some particular value or values of x in a given interval. In this connection
we may make the following remark:
a
Division by zero (symbols , ) is undefined.

The quotient of two finite numbers a and b (viz., ) is defined as the

definite finite number such that a = bx. Now, obviously, in the division,
zero value of b is excluded: for, if b = 0, then a ( = bx) = 0, and x can be
any number. Hence, the above definition rules Out division by zero.

Therefore, forms , are undefined.

 FUNCTIONS 19

The following simple illustration show' how division by zero leads to

fallacious results.
Suppose, x=y(x*0. y*O), .. x2=sy.
x2—y2=xy--y2
or, (x+y)(x—y)rry(x—y).
Hence, dividing out by x - y, x+ y = y, i.e., 2y = y, or 2 = 1.
The fallacy is due to the fact that we have divided by x - y which is
equal to zero.

Similarly, the assumption =1, on the basis that anything divided by

itself is 1, leads to fallacious results, as shown below.
0 03x0 0
3x—=3x1=3; again, 3x—=----= —=1, .. 3=1.
0 00 0
From the above remarks, it will be apparent that
x2-25
the function f(x) = is not defined for x = 5;
x-5

the function f(s) =sin

sin!I is not defined for x = 0; etc.

Note 4. lff(x) denotes a certain function of x, then in casef(x) is given by

a mathematical expression involving. thenf(a), i.e., the value of the function
for x= a, may, in general [but not always, as explained in note 3 above, and
also in (iv), Art. 2.4], be obtained by putting a forx in the expression for fix).
Thus, If f(s) = sinx, f(0) = sin 0 = 0:
If f(x)sx 2 -5x+1, f(l)=-3, f(—l)=7;

If f(x)=x 2 , f(x+h)=(x+h)2 2xh + h 2 ; etc.,

whereas, If f(x)=x cos !, .f(0) isundefined.

2.3. Graphical representation or functions.

Let =f(x) be a real valued function with domain A (c 9 1 ). Then the
graph of the function y =f(x) is defined to be the set of all points in the
plane with cartesian coordinates (x, f(x)). where y = f(x), XE A.

Taking the straight line X OX ,with origin 0 on it as usual, to represent

the real variable x, the value of the function, y of(x), may be represented
parallel to the line rOY drawn at right angles to X'OX , as in ordinary
graphs. Corresponding to every value of x (in the assigned domain) the
 DlFFEREN1Al.CALClJLUS
20

point is plotted whose ordinate gives the corresponding value of the

function. The assemblage of the points, which may or may not form a
continuous line, represents the graph of the function.
In drawing the graph it is not necessary to know the exact mathematical
relation between x and (which may or may not be obtainable). It will be
sufficient if we know the definite value of y corresponding to every value
(at least a large number of values) of x in the defined domain.
The graph at once presents to the eye the way in which the function is
related to, and changes with the argument.

2.4. Some remarks on functions.

From the very definition the following points should be clear:

(i) It is not essential for afunction lobe expressible by mathematical

form always. F example, suppose x hour after noon on a certain day, the
temperature of a patient is Tdegrec. Now, to each value of x (up to a certain
number, depending on our contemplated period of observation), there
corresponds a definite value of T. Hence, Tis a function of x by definition.
But T cannot be expressed analytically by a mathematical expression in
terms of x. Nevertheless, we can draw a graph which is the temperature
chart of the patient, giving an idea how Tchanges with the time x. For other
examples, see (vii) of the next article.

(ii) In some cases a function may have different mathematical forms

for 4ifferenf ranges of its domain of existence; for illustration, see (v) of the
next article.

(iii) A function may be undefined fpr some value or values of the

argument, as has already been remarked and illustrated in note 3, Art. 2.2.
Also, every function Cannot be defined in every interval; thus, sin-' x cannot
be defined in the interval (2, 3), for sin-' 2 has no meaning, there being no
angle whose sine is 2. -

(iv)A function may be defined arbitrarily. For instance, we may define

a function as
f(x)=x 2 when x<O,

f(o)=3.
f(x)=!x when x>O.

The function is thus definitely defined for all real values of x.

FUNC71ONS 21

(v) The functions x -25 and x +5 are different functions. The

x-5
former is undefined at x= 5 , and so its domain of existence is the aggregate
of all real numbrs excepting 5 for the arguement x. The latter exists for all
real values of x. Hence, though for other values of x the two functions are
equal, there is a point of distinction at x = 5.
x2—
A third function might be defined as 1(x) 25 whenx # 5,
=
and f(5)= 20. Then the function is again different from either of the first
two. It exists for all real values of x including x = 5, but at its value is
different from that of the second function x + 5.
= X 2_25
If we define a fourth function by saying that 1(x) when
-
x * 5, and 1(5 ) = 10, then this funétion is identical with the function x + 5.

2.5. Examples of functions.

Below is given a number of examples of functions of a variety of types,
with their graphs in certain cases which will help to form a clear notion
about functions and will further elucidate the remarks of the previous article.

2 2 x +
(I) Analytical functions like x 7 , etc. • or more generally,
+9
X2
polynomials in x of the type f(x)=— a 0 x" + a 1 x ` +.. . . + a,, x + a,,
(where n is a positive integer), or rational algebraic functions of the type
P(x)
7) , where P (x) and Q (x) are polynomials.

The domain of these are generally the set of all real numbers; in the last
case the zeroes of the denominator are excluded, for, the function is not
defined at these points.
('ml) f(x)=x when x>0
r
=0 when x=0, B

=—x when x < 0.

The graph, as shown in Fig 2.5.1, consists
of two lines QA and 37
B which bisect the x' o x
angles LXOY and ZYOX' respectively.
This is also the graph of the function
Fig 2.5.l
f(x)=xI.
22 DIFFERENTIA L CA LCULUS

(iii) f (x)=,f i

x' x
0.

Fig 2.5.2

f(x) is defined for x = 0, and all positive values of x; the graph is a

continuous curve (parabola ) in the flu-si quadrant.
(iv)f (x)=x!

or, f(x) = sum of the first x terms of _- + + + .....

The functions are defined only for positive integral values of x.
The graph in each case consists of a series of isolated points.
(v) The height y from the ground, at a time x. of a perfectly elastic ball
originally dropped from a height h.
Y

x •or.3r 5J

Fig 2.5.3

Here y is defined for all positive values of x, but expressed by different

mathematical terms for different ranges of the values of x.
Thus, denoting the time of fall (from start to first impact).
i.e.. J2hfg by x 1 , -
y = h— gx 2 , when 0 15 x 5 Xu (i.e., before first impact),

y =(x — x j)f --g (x — x 1 ) 2 , when x 1 < x!53x1

(i.e.. between first and second impacts),
FUNCTIONS 23

y = (- 3x 1 )Iij - - g (x-3x 1 ) 2 , when 3x 1 < x :55x1

(i.e., between second and third impacts), etc.
The graph, as shown, consists of a series of parabolic arcs, on the
positive side of the x-axis.

x2
(vi) y
X
For x * 0, y = x ; for x = 0, y is not known (undefined).

B
Fig 2.5.4

The graph is that of the straight line y = x, with the origin left out.
(vii) y = (x 1, where [x] denotes the greatest integer not exceeding x.
Foi05x<1.y=0; 1!5x <2,y =1;

2 < x < 3, y = 2
—1!5x<0, y=—l;
—25x<-1, y=-2;etc.
I

—2-11
IH. x
2 3

Fig 2.5.5
Thus, the graph consists of parallel segments of lines in which the
right-hand end-points are left Out.
24 DlIFERENTlA L CALCULIIS

(viii) y = x sin

ML

Here v is not defined for x = 0. Thus, the domain of x is the aggregate

of all real numbers except 0. Whetherx is positive or negative as the numercal
value of is very large, the value of yapproaches I, while always remaining
less than I.
The graph shown in Fig 2.5.6, which is Continuous everywhere
excepting at x = 0, where a point is missing on the graph. Near 0, on either
side, the graph has an inifinite number of oscillations with gradually
diminishing amplitude. The graph is comprised between the lines v = x and
Y = -x.
(ix) Functions like e', log x, sin x, cosx, sin x,cos x, etc., which
are not algebraic functions, are called Transcendentalfunct ions. For graphs
of first two, see Art. 19.9, and for some others see next page, i.e. Fig 2.5.7.

2.6. Bounded functions and their Bounds.

Let f(x) be a function defined in the interval (a, b). If a finite number
Kcan be found such that f(x) :5 K for every value of x in the interval,
then f(x) is said to be bounded above in the interval.
Similarly, if finite number k exists such that f(x) 2: k forevery xin
the interval, then f(x) is said to be bounded below
If 1(x) is bounded above and below in the interval, then its said to be
simply bounded.
If f(x) is bounded above, then it easily follows from Dedekind's Theorem
that there exists a definite finite number such that M 2! 1(x) for every
value of x in the interval, but e being any pre-assigned positive quantity,
however small, there is at least one value of x in the interval for which
f ( x) > M - E. This numborM is called the upper bound of the function
in the interval.
 25
FUNCTIONS

In a similar way, if f(x) be bounded below, then there exists a definite

finite number in such that in 15 f(x) for every x in the interval, but given
any pre-assigned positive number E, however small, there is at least one
value of x for which f(x) < m + e. This number in is called the
lower

bound of f(x) in the interval.

We know, sin x, cos x are bounded functions in the interval [-p, p1, the
upper bounds of both being 1 and their lower bounds being —1.

Ph

Fig 2.5.7
 26 L' 'FFERENTIAL CALCULUS

- For sin x, M = 1; now, taking £

, we can find at least one value,
2t if I
say ,of x, such that sin > i - = Other values of xcan be
obtained from tables, for which sin x > -. The function defined in
Ex. (v), § 2.5 is a bounded function in the interval [0, co), the upper bound
being h and the lower bound being zero.

For the interval 0 :5 x < 2., tan x has the lower bound zero, but no
upper bound. For the interval - -< x <- 0, tan has no lower bound,
but its upper bound is zero. ln the j nterval -- :5x < !, tan x has neither
lower bound nor upper bound. Thus we see that a function may have
different upper and lower bounds in different intervals.
The function X2 + 3x -t- 5, in the interval I 5 x :5 2 lies between 9
and 15; so its upper bound is IS and the lower bound is 9.

2 +x
The function f(x) = /_ is not bounded above in the
y(5—x)(x-3)
interval 3 < x < 5, in which it is defined.
Let x = 3 + e. where £ .is a small positive number.

.F 5+e F5
= > I > f— , which can be
y(2—e)e 2e V2e
made greater than any positive quantity by taking C smaller and smaller.
Hence f(x) has no upper bound.

2.7. Monotone Function.

Let x1 ,- x2 be any two points such that x 1 < x 2 in the interval of
definition of a function f(x). Then f(x) is said to be monotonically
increasing if f(x):5 1(x 2 ) and monotonically decreasing if
f(x1 )2: f(x 2 ). Thus in O,!J , sinx is a monotonically increasing

function and cos x is a monotonically decreasing function.

Sometimes the following definition is used. If forrx 1 <
f (x 1 ) < f (x 2 ) then f(x) is said to be strictly monotonically increasing,
and if 1(x 1 )> 1(x 2 ) then f(x) is said to be strictly monotonically
decreasing.
In the interval' 0 5 x <oo the function e' is. a strictly increasing
function, since e 1, < e". when x 1 <x2.
FUNCTIONS 27

In the interval 0 :5x < -, the function j 3x+5

is a
ftx) =
2x+1
strictly decreasing function, for. f(x 1 ) > 1(x 2 ) when X 1 <
The example (vii) of § 2.5 is an example of a function defined in the
interval (0, 3), which is monotonically increasing but not strictly increasing.
2.8. Classification of Functions
(I) Even and Odd Functions
Let f(x) be a function defined in a domain D(o R) where D is such
that X D - x D. The function f(x) is said to be an even function
if f(—x) = f(x) ,for all x e and 1(x) is called to be an odd function if
f(— x)=— f(x) for all XED.
The graph of an even function is symmetrical about the axis of y while
the graph of an odd function is symmetrical in opposite quadrants.
Every function can be expressed as the sum of an even an odd funtion.
H should be noted that inverse of an even function is not defined.
Examples: f(x) = x 2 , f(x) = cosx when x€ Rare even functions, for
f(—x) = ( — x) 2 = f(x). F(—x) = cos(—x) cosx = P(x).
Again •(x) = X 3 , W(x) = sink, where XE R are odd functions, for
= (—x) 3 = — x 3 = —$(x). V( —x) = —sinx = —W(x).

I) Periodic functions
A function 1(x) defined in a domain D is said to be aperwdicfunction
of peiod s if jibe the least positive real number such that f(x + jt) = f(x)
forall XED [Here, x+t€D,foral! xED].
f(x) = cos x, x€ R periodic function of period 271 , since 2n is the
least positive number such that f(x + 2n) = sin(x + 2it) = sin x = 1(x) ,for
all XE R.

(llfl Explicit and Implicit Functions

If D (ç R) be the domain of a functionf, we can express the function
as y =f(x),xED. ... (I)
If a function can be expressed in the form (1), the function is said to he
expressed explicitly and we say that the function is explicit.
2$ DIFFERENTIAL CALCULUS

f(x) = +2.X+ lOx, X IR, is an explicit function.

Now, let x, y be two variables where the relation between x and v is
expressed by an equation, say, 4(x, y) = 0, then it is called an implicit
function.
If x 2 +y 2 =a 2 , then (x, y)=x 2 +y 2 —a 2 =0 is an implicit
function.

Here, y=±Ja2_x2, —a^x^a.

So, we havç two explicit functions, viz.,

y1 =a2_x2, —a!5x5a
and y2=_.1a2_x2,_a:5x!^a

(IV) Parametric Function

Let x = f(r) and y = be two functions of the variable tin the
interval a <— 15 b.
By eleminatingi forn the relations .v = f(i), v 4(i) we shall have it
relation connecting x and v, i.e., y can be regarded as a function of x
Such functions are called parametric functions. If x = a! 2 , v = 2a, , we
can easily see that y 2 = 4ax i.e.,.y2 — 4ax = 0, which is an implicit function
of x and y.

Here, x = at 2 , y = 2at together constitute a parametric function, a'

being called parameter.
2.9. Composite function Function of a function
Let y = f(u) be a real valuesd function defined in a domain D1 (ç R)
and u = g(x) be another function with domain D. where it = g(x) € D1 for
all x eD.
Here, f:D1 —9 R, g:D - Rare two mapping where the range of g i.e.,
g(D) is a subset of D1 Then the composite mapping f rag: D --> R can be
defined where (f o g) (x) = f lg(x) }, x D.
This composite mapping is called the com/,osirefunci:on of two real
value functins y = f(u), u C D1 and u = g(x). u € 0 where g(D) ç D1
and the composite functions is given by y= I k(x)}. x€ 0 and so the
composite funciton can be called a function off i,nction.
 29
FUNCTIONS

fog

Fig 2.9

2.10. Inverse of Function

Let f : A —+ B be a function defined by , = f(x). such that! is
bijective, i.e., both one-one and onto. Then, there exists as unique function
g:B-4A , such that f(x)=y4rg(y)=x. for all X EA an for all yEB.
In such case a situation g is said to be the inverse off and we write
gf : B -)A
1ff and g are inverse to each other. (fog) (x) = ( go f ' (x) = x.
i.e., f(g(x)) = gf(x)).

Domain off Range of I

f

a
Range of f'
Fig2.IO
3.
Domain off

2.11 Miscellaneous Worked out Examples.

Ex. 1. Show that 1(x) = log(x + is an odd function of x.

Solution: f(x)+f(_x)=log(x+1/i7)+IO(_X +/iT7)

= iog{ (x+J_x+i7)}

=Iog{l+x -x}= log t=O

i.e.. f(x)=-f(-x)
Thus f is an odd function of x.
30 . DIFFERENTiAL CALCULUS

Ex. 2. Prove that any function of x, defined for all real values of x, can
be expressed as the sum of an even and an odd function oft.
Solution Let, 1(x) be any function of x defined for all real values of x.

We can write,

=t(x)+W(x), say

Now, as

x) = !{ f(_x) + 1(x)) =

So, 4(x) is an even function of x.

A lso, w(x) = !{ f(x)_f(_x)}

= !{ f_x) - f(X )1 = -! { f(x) - f(— x)} =

so that 119( x) is an odd function of x.

Thus any function 1(x) of the real variable x can be expressed as the
sum of an even function and an odd function of x.
Ex. 3. (i) Define a Periodic Function.
(ii) Find the period, if any, of the following functions
(a) sin (ax);
(1') !cosxl;

(c) 2cos-(x-7t);

(d) sin 4x+cos4x.

Solution
(i) If f (x) be such that, f (x+k) = 1(x), for all values of .v within the
domain of definition off (.v), then f x) is called a Periodic Function, and
k is called its period. Here k is generally taken, if its exists, to he the
least number except 0.
(ii) (a) f (x) = sin ax
1(
sins a . -
2n')
- = ,in(o -t 21t) = slnav,
\ 'I
2it
Sill is periodic Itiiictioii oF period
FUNCTIONS 31

(b)Here, f(x)=I
cos xI=/=.!.(1+ cos 2x)
Since the function cos2x is periodic with period
2it 1i,
— =it, l-(I+cos2xl "
/ =1
cosxl
2 12'
is also a periodic function with period it.

(c) f(x)2 cos -(x_it)

(x it
=2cosi
3
x
= cos-x-
+ 3 sin—.
.3 3:
Obviously, f is a periodic function with period 67t.

() f(x) = sin x+cos4 x = (sin2 x+cos 2-2sin 2 xcos2 x

= l_!.Sfl2 2x -+-cos4x
2 44
it
since cos4x is a periodic function with period 2n
= , f(x) is a

periodic function with period

Ex. 4. Find the domain of definition of the following functions

(i) f (x )=,i+J.T B. P. 1993 1

(ii) f(x)=.J(3x-1)(7— x). C. P 1994]

(iii) f(x)=f8+2x_3x' I C. P 19961

(iv) f(x) =log (x2_5x+6) C. P 1993, 2000, 2006]

I 4x-x
(v) f(x)=.log [B.P 1995; c.P2007]

I 5x_x2
(vi) f(x)=log C. P 1997

I
(vii) f W C. P 2005]
32 DIFFERENTIAL CALCULUS

Solution: (i) f(1)=i+iT

Since f (x) is real, the values of x must be such that both
and are real quantities, which requires that (x - l) -eO and
(5-x)-aO
i.e., x2:1 and x:55.
So, domain of definition of fix) is 1!5x:55 or, 1 1,5 1
(ii) f(x) = (3x- l)(7-x)
In order that J(x) may be defined.

(3x - 1)(7 - x) must be non-negative.

i.e.. (3x- 1)(7-x)^:0

i.e.. either, (3x- 1) ^! 0 and (7- x) ^ 0 ... (1)

or, (3x- i)!50 and (7-x):50 ... (2)

From (I),

Again from (2) and 75x

But these two relations cannot hold simultaneously.

So, the domain of definition of fix) is x!9 7

(iii) f(x) = Jx_3x2

For f(x) to be defined,
8+2x-3x2 ^:0,
i.e.; (3x+4)(2-x)^!0
i.e., either, (3x+4)>0 and (2-x)2!0 ... (1)
or, (3x+4)!50 and (2-x)!50 ... (2)

From (1) -5x:52.

Again. from (2) and x>2

But these relations cannot hold simultaneously.

So, domain of definition of fix) is - <- x 5 2.

FUNCTIONS 33

(iv) f(x) = log W _5x+6)

fix) is defined for all real values of x that make x2 - 5x +6 > 0
or, (x-2)(x-3) >0
This inequality holds for all real values of x, except those that lie
between 2 and 3, including x = 2 and x = 3.
So, domain of definition of f(x) is all real values of x, except
2!5x<3.

(v) f(x) = log

X- X
2
f(x) is defined, if log >0, i.e., ^! lug l

or, 4x-x 2 >-3 or, v 2 -4x+3!CO or, (x-1)(x-3)<3

This inequality is satisfied if 0 < x < 4.

(vi) f(x) = log -

f (x) is defined for those values of x which make

5x-x 2 = x(5-x)>0
Domain of/tx) is 0 <x < 5.
=
(vii)f (x)
I VI x
f(x) is defined, when I x >0
i.e., xI>x
and this inequality is satisfied for all values of x < 0.
So, domain of definition of fix) is (- oc, 0) or, - <x <0.
Ex. 5. Show that the domain of definition of the function
f(x)=log— is -1< x< 1.
1+x
f(xi)+f(x2)f[2)
Also, show that for x1 , 2 E(-1,l);

Solution log x is defined for positive values of x only,

l-x
fix) is defined only when >0
i.e., when both (l - x) > 0 and (I+ X) >0
or, when (l-x)< 0, and (l+x)< 0
3-
34 Dil-FERENTIAL CALCUL(1.s

These two sets of inequalities are satisfied for l< x < I.

Thus the domain of definition of fix) is (-I, I).
1_X1+x2
l+x1x2
Now,
12
)= log
'+X )
f(IX2 i+xi+x2
1+x1x2
•= log (1-x1XI-x2)

l- X1l
log - v,
l-x1 l-x2
=f(xj)+f(xj.

Ex. 6. Findd the domain and the range of the function f(x)
- =lxi
x
-.
I C. P. 1995, 2008 1

Solution: Here, f(x) = lxi

X
Obviously, fix) is defined for allreal values of x, except x = 0.
Hence the domain of fix) is —oc =
< x <=, except x 0
Again, : i x i= x , whenx>0
=-x, when x < 0,
lxi
—=1 whenx>Oand —=-1 whenx<0.
lxi
X x
So that range of fix) is [-i, iJ.

Ex. 7. If f(x) = max. (x, .1W) for x> 0, where max. (a, b) denotes the
greater of the two real numbers a and b, find the value of f(c) i(J
for c>0.[C.P.1988]

Solution: .. f(x)= Max. (x. ±)

when c ^! I, f(c) = max. (C' 1) = c

FUNCTIONS 35

and when O<c<l, f(c)

i
•
A gain, f
(1'
, yr Imx. (1
, — , c)= c'_)(c)
-- (
f ( c)f (!) {f(c)} 2 2
when c^ 1,

and whenO<c<l, f(c)f(!)=(!)4.

EL 8. If f(x)= and c(*O) be any real number, show that

If(c)—f(—c)l=. fC.P. 19941

Solution: We have x = x, when x> 0
=-.x,,whenx<0
So, when c > 0, f(c) = I and f(—c) = —1
and I f(c)—f(—c)II1+1I'2 ... (I)

Again, when c < 0, f(c) = —1 and f(—c) = 1

So, lf(c)—f(--011-1-112 ... (2)

Combining (I) and (2), if c(* o) be any real number,

Ex. 9. If f(x ) x +J x I. find f(3) and f(-3) B. P , 1995 1

Solution: f(3)=3+1313+36
f(-3)=3+I-31=3+3=6.

Ex. 10. If 2f(-)_f(x)r5x. find the value of

•Solution : We have, 2f()_f(X)5X ... (I)

Replacing x by -,

2f(x)_f(-!)= (2)
36 DIFFERENTIAL CALCULUS

From (1)and (2), f(x) =—+

5x 10
3x

1 1 51 l'\ 10
f jx+– x+– 1+--
x) 3 x) (
3i x+ –X
I

- 51(x+—l)'+2

3(x+!)

Ex. 11. (i) The function f satisfies the equation f(x -i- 'P') f(x) + f(y).
Show that
(a)f(0)=0,

(b) f(x) is an odd function.

(c) if x is an integer and (1) =a. then f(x) =ax

(ii) Find the natural number a for which

f(a+k)=16(2"_l),

where the function f satisfies the relation f(x + y ) = f(x) f(y) for all
natural x, y and J(l)=2.

Solution: Given that f (x + y) = f (x) + f (y) ... (I)

(a) Putting x = y = 0 on both sides of (I),

f(0)f(0)+f(0), i.e., f(0)=0

(b) Putting y = –x on both the sides of (1)

f(0) = f(x) + f(–x)

i.e., f(0) = f(x) + f(–x), i.e., f(x) = –f(–x) . f(0) = 0

f(x) is an odd function.

(c) Again, putting x = y = I on both the sides of (1),

f(2)=f(l)-i-f(l)=2f(l)=2a [. f(1)=a

f(3) f(2+ I) = f()+f(I) = 2a+a = 3a

FUNCTIONS

If x be a positive integer, then

f(x) =f(I+x— I)=f(j)+f(x-1)=a+f(x— I)
f (x-1)'f (1+x-2)=f (i)+f (x-2)a+f (x-2)
Similarly, f(x-2)=a+f(x-3)=2a+f(x-4)
Thus, f (x)=a+f (x-1)2a+f (x-2)'3a+f (x-3)
=4a+f(x-4)

=(x— l)a+f(1)

If
= (x— l)a+a = ax
he a negative integer, putting x = —y, where y is a positive
integer, we get
1( )
f(x) f(— y) = — y [.* f (x) is an odd functiin
= -. ay = a ( —y ) = a x
Thus, when x is any integer, f(x) = ax.
(ii) Here, f(x+y)=f(x)'f(y) (1
Putting x=y1,
1(2) = f(l) . f(l) = 2 . 2 =22 f(1) = 2]
Putting x 2, y
f(3) = f(2) .1 (1) = 2 2 .2= 2
Similarly. f(4)=f(3).f(l)=23.2=24
and in general, f(n) = 2'

Now, ': f(a+k)=i6(2'_1)

f(a) • f(k) = l6(2_1) fio.n (1)]

or, f (a)[f(1)+f(2)+f(3)+...+ f(n)] = 16(2 -

or, f(o)[2+22+23+...+...+2]16(2_1)

M f (a)X 2116(2'_l)

or, f (a)=8=2=f (3)

a=3.
38 DIFFERENT/AL cALCIJUIS

EL 12. Solve: 4{x}=x+[x],

where {x} and [x] dcrnne the fractional and integral parts of a real
number x respectively.
Solution: lfx is any real number, x=xJ+(x)
4(x)=x+[xi
4(x)=(x]+ (xi +(xJ

or, 3(x) = 2[x] i.e., {x) = (1)

Numerical value of (x) is less than I. So, the only integral values of
[x] which will satisfy (I) are I and (–I).

So. {x)=-, when, IiiI=l

and (xJ – , when, Ix I= –1

Ex.13. If f(x) = cos (lgx), then show that

AX) .f(v)_-{ f(!)+f(.*Y )}

Solution: . f(x)= Cos (log x)

f(x)f(y)=cos (log x) cos (log) ... (1)

and 4.E) + f(AY) = Cos{log(1)} + cos{Iog(Ay)}

log I i+log(xy) 14:-l –
log(xy)
_ty)
=2cos
2 2
(x '
log– . xv) log–'-
=2cos-2—'cos y
2 2
(kgx 2(i I
= 2cos - - i . cosI –log—
. 2 ) t,, 2 y2

= 2cos (log x)cos(log y) [ cos(–O) = cosO J

FUNCTIONS

f(x).f(Y)_f{f()+ f(xv)}

= cos (log x) cos(log y) - .2cos(logx) cos(log y) 0.

Ex. 14.Iff(x+3)=2x 2 -3x+1.find f(x+1)

Solution: f(x+flf(x-2+3)

=2(x-2) 2 —3(x--2)+1

=2x2-1lx+15.

Ex. 15. Find the range of the following finction.c

(i) f(x)=
2 - cos 3x

• (ii) cos(2x—)+sin(2x—j)

(iii) log, sinx—cosx+3'I

Solution: (i) Y f(x) 2— cos 3x

or,!=2_cos3x -
y
or. cos3x=2---
y

—1!5cos3x:51,

or, 5y:51 [y>O, . —1cos3x:511

Required range is [!, i].

(ii) Let y=cos(2x—)+ sin (2x—D)

40 DIFF-ERENTIAL CALCULUS

=.Jcos2x+_)

_1:5 Cos (2x+_)<1

,i.e., —,h.y:5

range is [_f -. .J•]

1092 sin x—cosx+3,/
(iii) Let , y =

sin x — cosx+3J ( n\
2y =sin i x-- i+3
\ 4)
It
or, 2"-3=sinlx--
4

—l:^'sin(x—^1
4)

—t:52"-3<1
or 2!52'<4
i.e.. 2'!52'< 22

i.e., 1:^y!52 range is [i, 2].

EL 16. Find the domain of 1(x) = + (B. P 20041

Solution: f(x) has real value if4+x^&and 9—x2!0.

i.e., x 2: —4 and x < 9
i.e., if XE (-4,o) and xE(—,9].

Now [-4, oo) fl(—o',9) = 1-4. 91.

Hence, the domain of f(x) is [-4, 91.

"f. 17. Find the domain of function 1(x) = 11

FUNCTIONS 41

Solution: f(x) will have real value if

(1)
2—Ixl
and 2— lx I* 0 ... (2)

The relation (1) holds if

(1) 1^:lxI and 2>fxl,or,if(J1) l -I x I!50,and 2-1xI<O
(I) iM!x and 2>IxI --
then -1-.g x!^1 and -2<x<2.
.e., XE[- I . 1) and x€(-2,2)
i.e., x eL-I, t]fl(-2, 2)
i.e.. x 1-I. I] ... (3)

(II) If 1 -I x I:5O and 2- jx

then I:g lxl and 2<lxI
i.e., lx 1>2
• i.e., x >2 Or, x < -2
i.e., x (-o,- 2) U (4 o) .. (4)

Now (2)holds if lx I* 2, i.e., if x ±2 (5)

From (3), (4) and (5) the required domain is
[-1,IJU(--°', -2)U(2,°').
_3X-10).
Ex. 1& Find the domain of the function 1( x) = log(21_5)(x2
Solution: f(x) is defined if.x2 -3x-10>O :.. (1)
and 2x-5>0, 2x-51 ... (2)
-Relation (1)holds if (x-5)(x+2)>0
i.e.,ifeither, (x-5)>O and (x+2)>O
or, if (x-5)<0,and (x+2)<O
i.e.,ifx>5 and x>-2,or,ifx<5 and X<2
i.e., i f x < 5 or x < -2
So relation (l)is valid if x€(5, o)U(—oo, -2) ...
42 D/FFERFNTIA L CA LCULUS

Again, relation (2)holds if x> i and x vE 3

i.e.,iI xE(, )-(31 (4)

Hence the required domain is the common portion of (3) and (4),
i.e.,(5,o).

= log2 sin .x - cosx+ 3'I

Ex. 19. Find the range of the function f(x)

Solution: Let yIog2 sinx-cosx+3,J

or, 2" =(.j_sinx-*cosx)+3

or, 2'-3=sin(x-)
Since for all real vaues of x
-l:5sin(x-):5l, -l52'-3:5I

or, 2:52 Y 54* 2 ' :52 ' :522 .'.l:5y!52. since the base is2>l.
Hence the raigc is [1.21.
Ex.20. Find the period of each of the following functions:
(i) cot f , ( ii) 3sinf+4cosf.
Solution: (i) cot(lt + x) cot x, , cotxis a periodic function of period ,t.
So, cot A is also a periodic function, the period being = 21t

(ii) sin(2,r + x) =.sin x, so sin x is a periodic function of the period

2x and hence sinf isa periodic function of the period

Again cos* is a periodic function of period 8n

Since 871 is a rational multiple of 41t, 3sinf+4cos f is a periodic

function, the period being the l.c.m. of 41t and 8n,i.e., 87t.

Ex. 21. Show that sin 3 x+cos3 x is a periodic function. Find the period.

Solution: sin-' X+COS3 x=-L(3sinx-sin3x)+-L(cos3x+3cosxi

= 4. 13(sinx 4-cosx)+(cos3x-sin3x)}
FUNCTIONS 43

Both sin x and cosx are periodic functions, period of each of the them
being 2it.
Again cos 3x and sin 3x are als&)eriodic functions, period of each of
them being 2f.
Now I.c.m. of 2n and is 2ir.
Hence 1(x) is a periodic function of period 2n.
EL22. Find the inverse of the function 1(x) log(x+.'/7i).

Solution Let y= log, (x+J7Ti).

then e' =x+'17i

, x2+1(e_x)2=e_2xc3+x2

or, e2-2xe'=l

or x =
2e"
• 2I
Interchanging and y, we have (x) =-- = + V -e).

EXAMPLES-il
I. If y = 6 for every value of x, can y be regarded as a function of x?
2. If y = the number of windows in the house numbered x on a particular
road, is y a function of x?
3. Given 1(x) = x 2 — lOx + 3, find j(o) andf(-2).
4. If 1(x) sec x + cos x, then show that 1(x) = f(—x).
—a+a
5. If 1(x) =b b , then show that
b-a a - b
f(a)+f(b) = f(a+b).

6. If 1(x) = x 2 — 3x + 7, then show that

{f(+h) -f(x)}/h=2x--3+h.

7. Show that
I- tan x
() cos - sin x is not defined for x = 4, r.
(ii) Jx2 — 5x +6 is not defined for 2< x <3.
44 D!FFhRhNT!A I. CALCULUS

x2-5x+6
(iii) is not defined for x = 2.
x2-8x+12
Also find f(-5) and 1(6) in this case.
8. Draw the graphs of the following functions:
(i) y = 1 when x > 0.
= 0 when x=O,
=-1 when r<0.
(ii) y =x for x^1,
= 2 for x=1.
(iii) 1(x) = 1 when x is an integer,
= 0 when x isiiot an integer.
(iv)

(v)

(vi) y=x-Ix],
where [x] denotes the greatest integer not greater than x.

(vii) f(x)=Ji_(x_l)2.

(viii)f(x)=i-l.
sinsrx
@) f(x)=i-
stnirx
(x)
where the positive sign of the square root is to Ye taken.
(xi) 1(x) 0 when xI>1,
= 1+x when -l:5x5 0.
1 - x when 0< x :5 I
9. (i) Show that .f(x)=sec x, in the interval 0 55 x <fir, has the
lower bound 1, and no upper bound.
(ii) Show that f(x)=2x 2 +4x+6, inthe interval 0:5x:5l. has
the lower bound 6 and the upper bound 12.
 FUNCTIONS

(iii) Show that f(x)=( when 0<0<1 , —1< x <1

1 +Gx
and n a positive integer, is bounded. -

10. (i) Show that 1(x) =----

x+l
is monotone ascending for x > 0.

(ii) Show that

I 1 1
+—, x>0,
x+1 x+2 x+n
is monotone descending.
) X,
(iii) Show that f(X I+ x > 0, is monotone ascending.

11. Given the relation y 2 - 6y —.x + 7 = 0, which of the following

statements is true?
(I) The equation defines x as a function of y for all values of y.
(ii) The equation defines y as a function of x for all values ofx.
12. A taxi company charges one rupee for one kilometre or less from start,
and at a rate of (i) 50 paise per kilometre (ii) 50 paise per kilometre or
any fraction thereof, for additional distance. Express analytically the
fare F (in rupee) as a function of the distanced (in kilometre), and draw
the graph of the function.
13. Find the domain of definition of the following functions:
(i) f(x)= 1, whenx is rational.
= 0, when xis irrational.

—
(u) f(x)= log—.
1x
1+x
(iii) f(x)= 1--t . [C.P.1995]

(iv)
f (x )
=
log 5x2 .
[C.P.1997]

(v) f(x)= .

(vi) f(x)= ,jjTj + . B. P 19931

(vii) . f(x) = log — 5x + 6). c.P 1993, 20001

46 Dh1FERENTIAL CALCULUS

(viii) 1(x) =
Fil__
(Jx) f(x)= log

14. If the function! satisfies the relation f(r + y) = 1(x) + f(y), forall
real values of x and prove that:
(i)f(0)=O;
(ii) A -x) -A x);
(iii) f(x) = ax, wherexis any integer and 1(i) = a.

15. If 1(x) = 2' (- ') for I :5x show that

fx)= !(i +
2
+ 4 log, 4.
ANSWERS
1. Yes. 2. Yes. 3. 3: 27. 7. (iii) ft; does not exist.
11. (I) True. (ii) Not true: true only for values of x - 2.

12.(i) V = 1, for 0<d!51;

(ii) F=1. torO<d!gI;

F=l+m. for m<dSui +1, where m isapositive integer.
13. (i) The set of all rational numbers;
(ii)(—Ll); ()_o.<x<oo,except x=0
(iV)O<x<5 (v)-'..<x<O;
(vi) 1< x <5; (vii)altx,except 2 15 x :5 3
(viii) _1:Sx1,2<x<. -<x<-2;
(ix) i<x<.
 ILi. LM
3.1. Intniduction.
The idea of 'LiMIT forms the most outstanding concept in Calculus
and plays an important role in the development of the subject. It is this
process of limit, or limiting operation, which marks the line of difference of
Calculus with Algebra, the latter being based upon the four fundamental
operations, viz, addition, subtraction, multiplication and division. The real
essence and strength of this subject, an important part of Mathematical
Analysis, lies in the concept of limit upon which is built the new and broad
structure of Calculus.

3.2. Liflsjtofan Independent Variable.

Suppose, x is a real variable which takes up different values
.v= 1 . 9, 1 . 99 1 . 999 i . 9999 1 1 . 99999 .,. it is obvious that as the
variables passes through successive values, the difference of x from a real
number 2, gradually diminishes and finally becomes and remains less than
any pre-assigned positive quantity, however small. We say that x
approaches or tends to the value 2, remaining always less than 2.

Again, we consider the values x=2 . 1, 2-01, 2 . 001, 2.0001,

200001,... etc. Here also the difference of the successive values of.r.
from the real number 2, gradually diminishes and ultimately becomes
and remains less than any pre-assigned positive quantity, however small.
In this case, we say x approaches or, tends :02, remaining always greater
than 2.
In either case, J x - 21 <C where e is a pre-assigned positive
quantity, however small, we may imagine and we write:
limit x—t2 or, liin x—,2 or, Lt x—,2.

33. Geometrical idea of the LImito1a Variable.

Suppose, the point A on the real number axis Xx represents the
real number x 2, while a point P represents a real variable x. Further, let
us suppose that the points A 1 , A 2' A ....... . etc. represent the real
numbers 19, 1 . 99, 1 . 999, ... etc respectively and the poiiits
B, B.,, B. ....... etc. represents the real numbers 2.1,2.01,2.001......
etc.
4$ DIFFERENTIAL CALCULUS

2-0 2+0

x'_l Ii A1A2A3AB3fi2R

Fig ill

The variable x gradually approaches towards the real number 2 by

assuming the successive values 1-9. 1-99, 1-999. ........ etc. and the
point P gradually approaches towards A from its efz side after passing
successively through the points A , A A - etc. .but it never meets
the point A: In this case, we say that the point 1' appioachesAfroni the left,

and denote it by the symbol x - a - U or, simply by x -9 a . Agair, let

us consider the sequence of numbers 2-1,2-01, 2 001 ...... etc. When
the variable x gradually approaches towards 2 by assuming the successive
values 2- 1, 2 . 01, 2-001 ...... etc., then the point P gradually
approaches towards A from the right side after passing successively thwugh
the points B ' B 2 B ....... . etc., bul it never meets the point A. In this
case, we say that the point P approaches A front right and denote by
the symbol x - a + 0, or, simply by x -9 a +.
Note 1. If x -4 a -0 (or, x - a .- ); the assumed values of x are always
less than a and the numerical difference between the assumed value
of x
afld a, i.e., I x - al is less than any pre- assigned positive quantity, however
small, but xis not equal to a (x ;^ a).

Note 2. If x -4 a + 0 (or, x -->a + ), the assumed values of x are always

greater than a and the numerical difference between the assumed value of
.x and a, i.e., I x -a I is less than any pre-assigned positive quantity, however
small, but is not equal to a (x * a). --

Note 3. The symbols: x -+ a is read as "x tends to a",

x -, a -0- (or, x -* a - )is read as 'x tends to from the left"
and x -+ a + 0 (or, x -+ a +) is read as "x tends to a from the right".
3.4. Idea of Limit of a Function.
Let y = f () be a function of a real variablex. A question may arise,
what happens to the function f (x) as x - a 7
JMIT 49

We examine the case by an example. Consider the function defined as

•ollows:

-4
1. f (x )= x-2 ,when x*2.

=3, when x=2. . •.. (1)

Obviously, when x * 2. 1(x) = x + 2.
We prepare the following table showing values of x and 1(x), where
he variable x approaches 2 either from the left or from the right.

2.0I)1

H,
1.9 1 . 99 1 . 999 • 2.1 2 . 01
X
j$ 3 . 9 3 . 99 3 . 999 .4-I 4•01 4-00I

It is clear from the above table that as x gradually approaches 2,

Issuming values either less than or greater than 2, and sufficiently close to
2, the values of f(x) gradually approach the number 4, or, in other words,

I 1(x) - 4 1, i.e.. the numerical difference between the value of 1(x) and
can be made less than any pre-assigned positive number, however small.

We write 1(x) - 4, when x - 2 or, symbolically

f(x)=4
.L12

x2 4
or, LI =4.
-*2 x-2
2
x
It is interesting to note that Li 4= 4 does not necessarily
-.2 x-2
imply that f (2) 4.

2 ' =
In the example cited above, Lj x 4, but 1(2) 3, i.e.,
.-.2 x-2
Li .f(x)*f(2)

2. If, on the other hand, we define 1(x) as

f (x )=X
4 when x*2
x-2
=4. when x=2 ... (2)

4.
 50 DIFFERENTIAL CALCULUS

then. Li fx) = 4 and 1(2) 4, i:e., the limiting value is numerically

equal to the value of the function at the point in question.

Here, Li f(9=f(2).
-'• -22

= -4
Further, if f x) X ( 3)
x-2
f(x) becomes undefined at x = 2, but as discussed earlier, Li f(x)= 4.

i.e., Li I (4 exists and has a finite value 4.

Thus, we see that the limiting value of a function at any specific point
is in no way dependent on the vMuc of the function at that point. The
distinction between Li f (4 and f (a) has been discussed in derails
and explained with illustrations in art. 3.6.

- The graph of the function .1(x) as defined in (1) is shown in fig. 3.4.1
and the graphical representation of thef(x) as defined in (2) is shown
in fig. 3.4.2 below.

. /

4 (2,4)
3

1234

Fig 3.4.1 Fie 3.4.2

By the expression. 'the variable x approaches the constant number a'

or simply x tends to the value a', we mean that x assumes successive
valueswhose numerical differences from a, i.e., the successive values of
x - a 1, become gradually less and less, and I x - a I can ultimately be
taken to be less than any small quantity we can name or imagine (i.e., less
than any pre-assigned positive quantity, however small), and we denote
this by the symbol x - a.
Here the successive values of x may be greater than as well as less
!.ban a.
LIMIT 51

If the variable x remaining alway greater than a approaches a such

that ultimately x - a i s less than any Ore assigned positive quantity,
however small (but x 96 a actually), then we say that x approaches or
'ends to 'a'from the right, and denote it by the symbol x -, m + 0 or
;imply by x -+ a+.
Similarly, when x is less than a always, and a - x is ultimately less
han any pre-assigned positive quantity, however small, we say thatx tends
o a'frorn the left, and denote it by x -9 a -0 or simply by x -9 a -

Illustration : When the successive values of x are I 9, I -99. 1 999 .

we say . 2-0, and when the successive values of x are 2 I,201,
.001 ......we say x -9 2 + 0 . If the successive values of x are
2+!, 2+, 2-ti, 2+, 2+....,we say x-*2+.

3.5. Limit ofa function.

1(x) When .v approaches a constant quantity a from either side
LI
(but 9 a ) if there exists a definite finite number! towards which f(x)
approaches', such that the numerical difference of f(x) and! can be made
as small as we please (i.e., less than any pre-assigned positive quantity,
however small) by taking x sufficiently close to a, then I is defined as the
limit off(x) as tends to a. This is symbolically written as Li f(x) 1.

Mathematically speaking, Li 1(x) = I, provided, given any pre-

assigned positive quantity e, however small, we can determine another
positive quantity 8 (depending on C ) such that I f( x) - I I < C for all
< a+ ö,
values ofxsatisfying 0<1 x - a I i5 S ,Le. whenever a- 6 f^- x
but x * a.
x9
E (i).Lt 6. For, if x = 3 + 81, Whether S be positive or
'-.3 x-3
x2-9 (x-3)(x+3) o(6+8)
negative. 6+8 and, by taking
• x-3 x-3
•
.
numerically small enough. the d i fference of -------- and fi can bemade
x - 3

As a panicu!arcasef(x) may remain always equal to 1 when xis sufficiently

close to (1.
 52 DIFFERENTIA L CA LCULUS

as small as we like. It may be noted here that however small 8 1 may be, since
0, we can cancel the factor x-3, i.e.. 8 between the numerator and

X2
denomenator in this case. Hence, Li =6. But when x = 3, the
-.3 x-3
X
29
function is non-existent or undefined, for, we cannot cancel the
x-3
factor x - 3, which is equal to zero in that case. Thus, writing
x2
f(x)=6, whereasf(3)doe not exist orisundefined
x-3 ,-.j
,

Ex (ii). LixsinJ. = 0. For sin i , whatever small value x may have

provided it is not exactly equal to zero, is a finite quantity lying between + i

and - l, and sobytaking x numerically small enough (ji . sufficiently near
to zero), we can make xsin I numerically as small as we like, i.e.. .* sin J.

is less than any assignable quantity. Hence the limit is zero.

Here also the value of x sin . , when ,r is exactly equal to zero, is non-
exMtent.

Ex. (M). Li = —3. For, writing x= —l±ö. wecaà show that the
_.- j x+3

numerical difference of x an d —3 can be made as small as we like by

x+3
taking ö small enough..

x2 7
In this case the value of . when x is exactly = -, is also
x+3
available, and that is also equal to —3.

U1(x): The limit ofa function f(x),asx approaches the value a

from the right ( i.e. from bigger values), is that quantity 1,, (if one such
exists), towards which 1(x) approaches, and from which the numerical
difference of f(x) can be made as small as we please by making x approach
a sufficiently closely, all the time keeping it greater than ti. It is called the
Right-hand limit of f(r) as x tends to a, and is written as
Li
.1 -3 a + 4)
LIMIT

Mathematically, L i f( x) = I I , proiided, given any pre-assigned

- I-tatO
positive quantity e, however small, we can determine a positive quantity
6, such that l f( x ) - Il
< e whenever 0 < x - a :5 6, i.e..

a< x S a+ö.-
Li f(x) is sometimes denoted by the symbol f( a+ 0).

1
E'. (Iv). Li J = 0, as can be shown by writing x=2+6
,-.2-0 5+eJ

where 6 is positive, and then making 6 arbitrarily small when the denominator
becomes arbitrarily large.
Lt 1(x): Ina similar way, we may define the Left-hand limit of a

function f(x) as x tends to a as follows:

Li can be obtained such that,

f(X)=121 provided, a quantity 12
given any pre-assigned positive quantity C, however small, we can
determine a positive quantity 6, so that A X )- 1 2 1 < E whenever
0<a-x:5-6,i.e..a—ö!5x<a.

Li f(x) is sometimes denoted by f( a - 0).

-0

Illustration: Li
I • i 1 =
i
201 -.1- -
5+e'-1

As x -, 2-0, x —2 is negative, and becomes numerically smaller and

smaller, so that e approaches zero in this case.

It may be noted that when Li f(x) = Li f(x) , each of these is

x-.n,+O x-,a-0

Li f(x)
equal to Li f(x) . Conversely, for Li f(x) to exist, each ofs-eD,
s-ta0
N_eu

and Li f(c) must exist, and musu be equalto one another, and this

commonvalueis Li f(x)

If Li f(x) Li f(x) ,or even one of them does not exist, then
i-i40 s-ta-fl
Li f4 x) does not exist.
 54 DIFFERENTIA L CA LCULW

E (v). Ln the above example (iv),

since LJ{'
} 5+e }

does not exist.

Ex. (vi). Again, consider Li x2

2 2
Here, Li x =2 2 = Li [See Art. 3.13, Ex.lJ
x--,2+0 .2-0

Li .x , = 2=4

Ex. (vii). Consider f( x) =

Here, Li f(x) does not exist, since, for values of x< 2 (however

near to 2 ), f(x) does nof exist.

Li does not exist.

3.6. Distinction between Lt f(x) and

The statement Li f(v) is a statement about the value of f(x)

when x has any value arbitrarily near to a, except a. In this case, we do not
care to know what happens when x is put equal to a. But f(a) stands for
the value of f(x) when x is exactly equal to a, obtained either by the
definition of the function at a, or else by substitution of a for x in the
expression f(x) ,when it exists.

Note. Five distinct cues may arise.

(i) f( a) does not exist, but Li f(x) exists. -

This is illustrated by Ex. (I) ofArt. 3.5.

(ii) f( a) exists, but Li f(x) dues not exist.

Suppose f(x) = I when x > 0,

=0 when x=0,
= -I when x < 0.
LIMIT

Li f(x)=
I; because when x, remaining greater than 0,
Here,
becomes arbitrarily near to 0, f(x) always remains equal to I and hence
E1 , for any
f(s) - 11, being = 0, is < any pre-assigned positive number
positive value of x less than 8, however small.
Li f (x) = - 1 ;
because when x, remaining less than zero,
Similarly,
becomes arbitrarily near to 0, f(x) always remains.equal to

Since Li f(x) LI f
f(0) = 0, here.
Li f(x) does not exist; but, by definition,

(iii) f(a) and Li f(s) both

exist but are unequal.

Let f(x)O for .v ;6 0,

= 1 for x = 0.
As in (ii). it can be easily shown here that -
Li f(s) =0 Li f
x-,0+O
Li .r(x)=
0. But, by definition, j(o) = I.

(iv) f(a) and Li j (x) both exisf and are equal.

This is illustrated by Ex. (iii) of Art. 3.5.

Neither 1(a) nor LI f (x) exists.

()
Let j(x )sin-.
Art. 3.131 and j(o) does
Here, Li sin-1- does not exist [See E. 4,
not exist, as it would involve division by zero, and is otherwise undefined.

3.7. Symbols + and -

If a variables, assuming positive values only, increases without limit
(i.e., ultimately becomes and remains greater than any pre-assigned positive
and write it
number, however large), we say that x tends to infinity,
as x-9°=.
Similarly, if a variable x, assuming negative values only, increases
numerically without limit (i.e., —x ultimately becomes and remains greater
than any pre-assigned positive number, however large), we say that steeds
to minus infinit y and write it as x -4 -
 56
DIFFERENT/AL CALCULUS

Note. It should be borne in mind that there is no number such as 00

or -
towards which x approaches. The symbols are used only to indicate that
the numerical value of x increases without limit.

3.8. Function tending to infinity, : Li f () = ± o.

As x approaches a either from the right or left, iff(x) tends to infinity

with the same sign in both cases, then we say
that, as x tends to a, tends
to infinity', (Or
loosely, the limit off(x) is infinite), positive or negative as
the case may be; and write it as Li 1(x) = 0o
or

If, however, as x approaches a from both sidcs,f(v) tends to infinity

with different signs, we say 'does not possess any limit as tends to
a'.
The formal definitions are as follows:
If corresponding
to any /)'e-assgned p051/ut' quantity N, however
large, we can determine a /os,!i ye quantity d, such that
f(s) > N whene,'e,
0<
Ial !^ 5, Wesay

Llf(x)=00.

I! in the abo ve cjrc,,,,,c,0,,,.5 - f(r)

> N whenever O<J s—a
wcs" Li f(X)=—oo.

Similarly, we may define the cases

.12 f(x)=oo, Li

LI f (x) = —oo Li f()

Illustration : Li = ii
-o.o X2.-o-o -o
2
= oo .'. Lj ± =

Li ! = , Li = ,
-O+O X — 0 -0 x
..
-i does not exist. In
X
either case, however, f(0) does not exist,

3.9. Limit of a function as the variable tends to infinity: Lt

r W.
x -,
As x, remaining positive, becomes larger and larger, if there exists a

According to some modern writers, this is described as '

f(xi becoming
infinitely large', and infinite limit is not recognised as a limit.
definite finite number I towards which f(x)
continually approaches, such
that the numerical difference of f(x) and I can be made as small as we please
by taking x large enough, we say Li f(x) = I.

Mathematically, LI 1(x) 1 , provided, given any pre-assigned

positive quantity c, however small, we can determine a positive quantity
M, such that I 1(x)—I I < C for all values of .t> M.

Similarly, Li j( x ) = I', provided, given any pre-assigned

positive quantityC however small, we can determine a positive quantity M,
such that I f(x) -I' < C for all values of —x> M.

In a similar way, we may define the cases

Li f(x) = , Li f(x) = — oo, Li f(x) oo, etc.

Illustration: Li = 0, Li —0,
X 2

Li e =1, fix 2 =00 etc.

3.10. Fundamental Theorems on Limit.

We give below some fundamental theorems on limit which are of frequent
use.

If Li f(x) = 1, and Li 4'(x) = I', where I and 1' are finite

quantities, then

(i) Lt {f(x)±Ø(x)} = I ± 1'.

(d) Li {f(x)x(x)} =
W.

(iii) Li
X-1 =!_, provided I' * 0.
.'-. lP(x)J I

(iv) Li F{ f(x)_ F{ Ii f(x) }, i.e., = F(l),

,
where F(u) is a function of u which is continuous' for u = I.

See next chapter.

58 DIFFERENTIAL CALCULUS

(v) If 0(x) < f(x) < v, (x) in a certiain neighbourhood of the point
V and Li 0(x) =1 and Ii çtt () = 1 .then Ii f(x) exists and is
equal to 1.
In particular, if I f(x) I < g(x) i.e., f( x) lies between —g (x) and

g(x),andif Li g(x) =0, then Li 1(x) =0.

(vi) If Li 0(x) = and Lt '(x) = 1 2 andif0(x)<(x)

in a certain neighbourhood of a except a, then 1 1 12

Proof:
(i) Since Lt f(x) =1, Lt 0(x) = I', we cart, when any
positive number C is given, choose positive numbers 6 8 2 such that

f(x) — iI < -fe whcno<lx — aI ^ (5, (I)

10(x )- 1' ! <e whcno<j I 15 6 x—a (2)

Let be any positive number which is smaller than both 6 and &, then
the inequalities (I) and (2) both hold good when 0 < x -
a j !^ 6

Now {f(x)-1}+{0(x)-1'}I ^ f )- 11+10(x)-I'!,

I{f(x) + 0(x)}—{i+1'}I <

1+ i.e.,< C,

when 0<1x—aI i5 5

by definition, 1+1' is the limit of{f(x )+ Ø(x )} as x -, a.

Similarly, itcan be shown that I - I' is the limit of { f( x ) - 0 (x ) }
as x —9 a.

Hence, u {f(x)±ç(x)} = 1±1'.

(ii) Wehave f(x)çt(x)_lI'={f(x)—I}{O(x)—I'}

+ I'{f(x )— i} + i{ Ø(x )— i'} .

I i( x )o ( x ) -11' f( x ) — J ib (x ) -'

+i I'IIf( x )— I ! + i/Ii O(x)-1'j- ...

 59
L.IIrIiA

Now, e being any pre-assigned positive quantity, and choosing any

other positive quantity k,
e e
k,
' 311'1' 3k
3111

are known positive quantities, and

since Li f(x) =1, L i 0(x) =1',

we can choose positive numbers 62 , 83, 6 4 , such that

< k when o< I x - a I ^ö

when 0<Ix—aI:562,

If(x)'I< 31
E when 0<lx—aI!5 83,

84
and Io(x)-l' <3 1 1 1
vhen O<Ix_al

Hence, if 8 be the least of the positive numbers 62, 8 3 , 8 4 all

the above four inequalities hold when
0 < Ix — a ^ 6.
and so from (I),

jf (x )Ø (x < k.
+
+

i.e., i.e.,<E when O<jX_al!5Si

by definition, Li{f(x)P(x)}= Ii,.

1(x) 1 !'{ f(x)— I} — l{ c(x)— l'}

(in) We have -- =
q x 1'

< I 1 'IIf( x )_ h IH h II p(x)—I'I ... (I)

-
i'j(x)

Now, since Li 0 (x ) = 1' ,there exists a positive number 8,such

that l O(x)_l'I<II'I when 0<1x-al 5 ,for 1'* 0.

I 'I
i — j*(:51
x ).l ø(x)-/'I <
5 .5,. ... (2)
or, IØ(x)I > #j! ' when
 60 DllFtJRENTIAL CALCULUS

Also, there exist positive numbers 8,, 5, such that

If( x )-1l<' for 0< l x — a I 5 S

10( x )—l'l< e' for O< I x—a l 5.c5 3 (3)

where e' is any chosen positive number.

IfS be the smallest of 5 , 5 3,, then it follows from (1), (2), (3)
that, when 0 <Ix - a :5

f(x ) 1 I< EL6

Now c being any pre-assingcd positive number, if we
E=c{I}2I{I1j+I1I } .weget

If(x) - lj < when 0 <t x — al 15 5.

Hence, Ij 1(x) =
(x) 1'

(iv) Let a = 1(x); since F(u)jscontinuous for u =1.

when lu-1I:55,

i.e., when jf(x)—iJ<51 .. (It

Again, since f(x)-+I as x -40,
Jf(x)_lj<Sj when O< Ix—al :55 ... (2)
Combining (1) and (2),

F{f (x)}—.p(1 ) I .ce when 0< Ix - al :5 8,

i.e., Li F{ f (x)} =F(l)

K -ta

(v) Assume that the inequalities (,(x)< 1(x) < ( x) are satisfied
when 0< l x_a <Si.
I
Since Li
Ø(x)=!,lØ(x)_1I< C when 0<Ix—aI!552.
i.e.. l—e<Ø(x)<1+e when 0<Ix_at5 ES,.
LIMIT 61

Similarly,I—e<(x)<l+e when O<Ix- al :5 £53.

1(8 be the smallest of the numbers o5j, 32, 5 4 ,then all the above
inequalities are satisfied when 0 < I x - a 8. Under these conditions,
i— c <0(x)< f (x ).
Also
i.e., lf( x ) - 1 1 <c when O< Ix—al
f(x)— +1 as x —,a.

(vi) Let us suppose that the inequality 0 ( x ) < iy ( x) holds good

when 0<Ix— aI<51, i.e., in the neighbourhood
a-8 1 <x<a+6 1 , x *a, •.. (1)

If possible, suppose I I > 12.

Let us choose e = 4 (i - 1 2 ), a positive number.

Since ,Lt ()= It

,

I0(x)-1iI<e when 0<Ix—al :5 ö,

i.e.,1 1 - 6 <çb(x)<1 1 +C when a-4525xca+52.

—1 7 )<Ø(x ) when a < a+82.

i.e., 4(I1+12)<0(x). .. (2)

Again, since LI

1 2 —E<4((x)<1 2 +E when a-83!5v5a+63.

yt(x)< 1 2 • + 4(.1 I —ta),

i.e.,u,(x)< J1 ( 1 1 +1 2 )whena-8 3 !5xi5a+8 3 (3)
Letö be thesmallestof the numbers 6 1 1 62, 45 3 ,then all the above
inequalities (1),(2),(3) hold good in the interval a—S 5 x S a +6.
from (2) and (3), 0(x)> (i, + 1 4 )> (x I.
 62 DIFFERENTIA L CA LCULUS

(x)>v(v)ina-55x5a+5,

which contradicts our hypothesis that i (x ) < i, (x ) in that interval.

Hence our assumption I I > 1 2 is incorrect.

lj ' 1 1i.e., / :512.

Note. The first two theorems may be extended ffi any finite number of
functions. In languages, the first three theorems may be stated as follows
(i) The limit of the sum or difference of any finite number of/un ctions
is equal to the sum or difference of the limits of the functions taken
separately.
(ii) The limit of the/ - roduct of afinite number offunctioiis is equal to
the product of their limiiv taken separately.
(iii) The limit of 11w quotient of two functions is equal 10 the quotient
of their limits, provided the limit of the denominator is not zero.
As examples of(iv). we get

Lt log J(.v)= log Li tog 1, provided l;.O,

/i f (s)
L i ej(=e540
5-3 a

L j { j(x) = { 0 f(X )} = 1", etc.

3.11.Some Important Lirnils.

sinx
0 IA
-.o
- = I. wherex si expressed in radian measure.
x
From elementary Trigonomctiy', we know that ifx he the radian measure
of any positive acute angle. i.e., 0 < x < tr ,then

sin
Sin A < X < t an
x ,Or, cos; X < -- < I ... (I)
S
sin 1
0<}-- —Cos X, i.e., <2 sin2—x
X 2
Sec Das and Mukherjcc' Intermediate Trigonometry.
LIMIT 63

But 2 sin 1 x <2(x ) 2 . <

2
Hence, 0<l--<-x.
sin I

x
Now, since x2 —0 as x—*O+O , we get

sin x1
-0.0 ( x ) -. O,O X

A lternatively, noting that cos x -4 1 a x -, 0, we can conclude directly

Li sin '-
from (l)that = 1.
x
When - -fit < x <0, putting x = —z, we get 0< z <

sin x = in (—z) = sin

Also.

,.inx sin Z
Hence. Li = Li
-.o-o x ..O+O Z.

Hence the result.

(ii) a) Li
(Il
+- 1-
e (ii .- Co through positive integral

values).
J=
I
(b) Li I =e. -
l x)

Proof: We have already seen that

L i (i+! = e, ii is a positive integer). [See A rt. 5.121

]
Now, let x be any large positive number. Then we can get two
consecutive positive integers n, n + 1, such that

0 X ii+l

and each being > 1, and as a + 1 > x n.

1+—
H)I >1 1+—
x)
I >1
t, f l+l)
or
t,
 64 DIFFERENTIAL CALCULUS

(+±(1+fl'>(1+±> +-
n ,)1 ii) x) ni-I) / n+I

Now when X —' °°, fl -4 also, and n being a positive integer,

/
I I+— I and I I +11+1)
I
- I
i_'•'
both -9 e, as proved before. Also,
fl)

—+ l and I + ---- —)
1. Hence the two extremes jn the above
n n+l
inequality tend to a common limit e, and so ( + -!- — e.
-
Lastly, suppose x = -p, where1, is a large posiIve number; then as
P —) 0O, X—'

Then 1l+) )P

PI •,I p1
I-
I+ = I (wlwre q = !' — I ).
' q)
Now, if p—) . q —, and hence
I
1+-=(I,
^q4l ,
I - 1±-- -e.
q q q

Thus, [i+!J -'easx-3-.

Hence, we see that i

—+- (
+!
X
) = e.

x being not confined to be integral here.

Cox Lt (1+x)' =e.

Proof:

In the above result, replacing v by as x — t co, y —+ 0, and

we get
Ll(1+v)t'. e, or. tf(I+x)''e
 6
LiMfi

(in) Lt .- iog(l +x)= I.

Proof: We have
L i ! log(l+x) = L i log(I+x )"

= log{ ti(l + x ) by3.8(iv)1

= loge 1.

-I
(iv) U e' =1.
• i-.. x
Proof: Put e' = I + z.
Then x =log(l+z ), and as x-+O,z---)O.
'-I Z _
Lt e = Li
Thus -.O X -*O log( l+z)

i/ Lt-{ log(l+z
I :•-•O Z

= ..= I. [lu' (iii)]

(v) Lt na
=1111110-
-. s-a
for all rational values of it provided a is positive.

CAsE I. W hen n is a positive integer.

By actual division, we have

3" ,,-I ,-2 ,,-_3a2+ .......

=x +x a+x
x-a
reqd. limit = L i (_i + x" 2 a + ......+ a_I) na"',

since the limit ofeaChOfthe n terms as x-a as a", and the limit of the
sum of a finite number of terms is equal to the sum oftheir limits (A rt. 3.8).
5-
 66
DIFFERENT/AL CALCULUS

CASE II. When n is a negative integer.

Suppose n = -m, where in is a positive integer and
a * 0.

- a" x " - a fl" x"' -a'

Then
X - (I X - U 5" a' x -a

Now, as x -+ a. the limiting value ofI =1

'-4- and

as x -* a, the limiting value ol = in a , by Case I.

5-a
f-a" I
Li = --ma = -ma
x-a = na

CASE III. When n is a rational fraction. -

Suppose n = p/q, where q is a positive integer and p any integer, positive
or negative.
Let us put x"q y and o b

y" - b" -
x- a - s-a y_b (ve_b(yby

Now, as x-+ a, x' ' -a 1

,:), -h. Again, as)' b, the
limiting value of the numerator of the right side = ph P - I (by Cases / and
II ) and that of the denominator = qb "(by Case I).

= pb"
Li = E,'- qi a" = na"'.
s- a qb" q q

• When n 0. the limit is Li = 0.

5 -a

(l+x)"-1
(vi) Lt =n.
x
Proof: We have
+x
Li = Li
Li it)" -llog(l+x)
x •v-.O log (l + •r ) x

= Li (l+x)"-1 log( l+x)

Li
-+o log(Ii-x) .v-.o x
 67
LIMIT

Now, put (1 +x)'= I +z. Then n log (l+x)10g(l+Z).

Hence,asx-40,log( I +z)-,OandSOZ-40.
L i (l+x)-1 = L i nz
Thus,
._o Iog(l+x) Z-0 log (1+z)

log )}]
z-,O

=n/ Li {!log(1+z)}
z -,O Z

=n (by (iii)l.

Li log (I + x)= Lby (iii).

Also,
A
+ .v )" -
Hence, Li (I =nxl = n.
A
This result also follows by replacing x by x + I and a by I in (v)
above.

3.12. Caucity's necessary and sufficient condition for the existence of a

limit.
The necessar y and sufficient condition that the limit ii ((x)

exists and is finite is that, corresponding to any pre-assigned positive

number c, however sinai! (but not equal •o zero), we can find a positive
number 5 such that x 1 and x, being an y two quantities satisfying

0< l x— a I :55, If(x1)— f(x2)I<

To prove that the condition is necessary, let Li 1(x) exist, and be
finite, and = 1 (say).
Then, given any pre-assigned positive number e, we can find a positive
number 6, such that

lf(x)- lk
when 0 < . 1 x - a < 5. II now x and s, be any two quantities
satisfying 0<I.—al 8
 68 DIFFERENTIAL CALCULUS

then

Ii (x ' )- I I + I ( 2)—i
<E+f ., i.e., <c.
Hence, the condition is necessary.
The proof that the condition is sufficient is beyond the scope of the
present treatise.

Note. In some cases even if we may not know the value of a limit beforehand,
we can determine by the above test whether a limit exists or not. Illustrations
of this are given below.

Ex. i. Show that Li cos .!. does not exist.

-4o X
In order that the limit may exist, it must be possible to find a positive
'\number 8 such that, x and x, satisfying 0 < x 15 8,

cos--cos--- <E
xI x2

where c is any pie-assigned positive quantity.

Now, whatever 8 we may choose, if we take x 1 =l?(2n,r) and
= 1/{ (2n+l )r } , by taking n a sufficiently large positive integer,
both x1 and x2 will satisfy 0 <x:5 8.
But in this case,

Icos(lIx i )_cos(11x 2 )j r Icos2n,r_cos(2n+1),rI = 2,

a finite quantity, and is not less than any chosen e.

Thus, the necessary condition is not satisfied, and so the required
limit does not exist.
Here, the right-hand limit as also the left-hand limit are both non-'
existent.

Ex. 2. Show that L i- does not exist.

-*O 2+e'''

Here, taking x 1 = — I/n, • 2 = 1/n, whatever 8 we may choose, by

taking ii a sufficiently large positive integer, we can make x 1 and x1 both
UMIT 69

satisfy 0 <I x I :5 6. But in this case,

1- I 1=1 1
2+-es
12+e
I I I
= •1. - >
2+e 2+e 2+e 2+e
which is a finite quantity and so cannot be less than any chosen € however

Thus, the necessary condition being not satisfied, the limit in question
does not exisk
Here, the right-hand limit exists and = 0, and the left-hand limit exists
and= 4.

3.13. Illustrative Examples.

Ex. 1. Find the value of L . x2.
By taking successive values of x, which always remaining less than
2 tend to 2, viz.. x = 1-9, 1-99, 1999.....we see that x 2 has the values
3-61,3-9601,3. 996001.... which tend to 4, and we can make the difference
between 4 and x 2 smaller than any positive number however small by
taking x sufficiently near to 2. Hence, the left-hand limit is 4.
Similarly, by taking values of .x, which always remaining greater
than 2 approach 2, viz., x 2-1,2-01, 2.001.... . we see that x 2 has the
values 4.41,4-0401, 4-004001, which continually approach 4. Hence, as
before, the right-hand limit is 4.
Hence, the value of the requied limit is 4.

Note. Exactly in the same way, we can show that L t XA = a. where n

- i-3d
is an integer or a rational fraction (except when a = 0 and n is negative).

Ex. 2. Show that (1) Li sine =0;

e-.o
(ii) Li cosO=l;-
-
0-30
(iii) Li- sin O= sin a;
O-.a
(iv) Li- -coso=cosa.
 70 DIFFERENTIAL CALCULUS

(i) Since, from the definition of sine of a real angle 0 in trigonometry

with the help ofa figure. it maybe easily seen that sin 0-01 . sin
01
can be made less than any positive number C, however small, by making
101 arbitrarily small, it follows that Li sinO = 0
0-*o
(ii) Li (l-cosO)= Li 2sin 2 -9=2x Li (n 0xsinO)
0-40 0-.0 0-0 -
= 2x0 1 b (i) I = 0.
Li cosO=l.
0-)

(iii) sin O-sina = 2sin+(O-a )cos-(0+a).

As 0-a, 0-ct)--,0, .. L i in 1(()

Also. ko(0x I. .. Li (sin0-sincx '=0,

i.e., Li sin0=sina.

(iv) Since cosO - cos u. = 2 sin - . (a-- o) sin -1 a ), it follows, as

in(iii),that Li (cosO - cos(t )=o, i.e., Li cos 0= cos,'(
0-.O 0-.a
EL 3. A pply ( 5, e ) definition of limit to illustrate i/wi
L t ()

Let us choose c=0.0J.

•I(2x2)_61<0.01if12x_81<0.0I,ie.,iflx_41.<O.005,
ie.,8 =0005. Similarly, if e =0•00l,5 =00005; and soon.

Thus, : depends upon E,i.e., the nearer (2x - 2) is to 6, the nearer

x is to 4. We have

.I( 2x-2 )-6<0.01 if 0<Ix'-41<0.005,

I( 2 x-2)-61<0 . 001 if 0<Ix.-41<0.0005,
and generally, (2x_2)_6 .ce if
0<lx-41<+e.
Hence, 6is the limit of 2x-2 as x-4.

Ex. 4. Draw the graph of sin (l/x) and show that neither the right-hand
limit nor the left-hand limit exists as x tends tO mm
LIMIT 71

When .x = 0,sin .! is meaningless'and hence its value is not known.

For all other values of .x, sin ( l/x) exists and may take any value from
-1 10 1. Thus, the graph is continuous curve with a break at x = 0 and is
comprised between the lines y = I and y = -I.

X X

Fig. 3. 13.1

As x -) 0 + 0, by passing successively through values 2 I rnt, where

n is a positive integer which can be made as large as we like, sin(l/x)
passes through values 0, -1, 0, I, etc. taking intermediate values at
intermediate points. Now, it is evident that these values are taken more
frequently as x comes nearer-toO and so sin ( l/x) does not approach any
fixed value as x --* 0 + 0, but oscillates through all values between -1 and
+1, i.e., the function has no right-hand limit. Since, when x is negative,
sin (l/x)= - sin (l/), where z = (- x) is positive, the function behaves
exactly in the same way when x -4 0-0. Hence, the left-hand limit also
does not exist for the function.

Note. Hence, it follows that ii sin! also does not exist.

-.o x
Ex. 5. Give an example to illustrate the following limit-inequality:

If Li •(x)=A and Li q,(x)=B and if tl(x)<iI(x) in a

certain neighbourhood of a except a. then A5 B)

Suppose, •(x)=5-t-x2; (x)=5i-3x2.

Li •(x)=5= Li W(x).

'For proof of this important theorem see Appendix.

72 DIFFERENTIA L CA LCULUS

But, 4(x)<w(x) if x*0.

Thus, the limits of the two functions'are equal, even though
•(x)<W(x) forailvaluesof x on which the limits depend.
If, however, •(x)=5+x2, W(x)=7+3.2,
then of course, Li •(x)< Li w(x).
- x-,0 i-tO

Ex. 6. Evaluate Li
As it stands, theorem (iii) of Art. 3.8 is not applicable, since the
denominator x is zero as x -4 0. But it can he easily transformed into a
form in which the theorem is applicable.
Multiplying the numerator and the denominator by Ji + .Ji
the required limit
=12
o(fi+fi)

=i0 +1.

since L i ,Ji = Li 1 (pulling I + x =

—0v-ti
and similarly L i
i-.0
fi =1.

Ex.7. if -l<x<1, then 12 x" 0. (n isaposi:iveiiueger).

Let us first consider the case when 0< x < I.
Put x=1-p, so that o<p<l. Since (l_p.)(1+p)1_p2,
which is less than 1, we have I - p . I / (I + p).
1 1 1
x =(l-p) c <
(i+p) l+np np

x can be made less than any Oven positive number e by taking

n large nough Ii.e.takingn>_!_ ) ; but x is positive.

-El,)
bx'=O when n-).
Since (-x)" =(-1 rx', the resultalso holds for -1<x<O.When
x=0,? =Oforeveiypositivevalueof n.Hence Lix' =0 when n-.
LIMIT 73

Note. When x >. 1, putting x for l+p in the inequality

(i + p)' >1 + zq > np, it can be shown that .'> k, where k is any
positive number, however large, for all values of n > k/p.

Hence, it follows that, for x>I, Li x =°.

Ex. 8. Prove that (it being a positive integer)

(I) =0 when xI.zl.

(i) when IxI5l.
when x>l.
xH

Li —=0 for. allvaluesof x.

(iv) Li m(m—1 )(m_-2) .... (m_n+1),, =0 when IxI<l.

3.14 MLSCeUaUeOUS Worked . out Examples

Ex. 1. Evaluate the following:

(i) lim(X_ , 1.
2 + x)
I ç. P,1983 1

(n) lim{x—,J(x_a)(x_b)} [C. P 1992, B. p

Solution : () lifl (x_.Ix2 +x) ___

(x_x)(x+x2 +)
= lim
x+Jx2+x
—x
= urn

=Iim
X-4- I
x
.1
hm—=0.
2
74 DIFFERENT/AL CALCULUS

(ii) Jim {x_.J(x._a)(x_b)}

= jim {x-,J(x_a)(x_b) }(x+,j(x-a)(x-b) }

x+-..J(x-a)(x-b)

• x2-(x2-(a-b)x-1-ab
= urn
x-,={x+I(x_a)(x_b)1

ab

I-
X—+(a+h)

x{i+(iJi)}
(a+b)——
=Iim
x- ( aYb
1+j 1--Ill--
I x,) x

, , _ aJ, . 0, =
=-(a+b), . J A —=0, Jm a Jim -=0.
b
2 x x-= x x-= x
Ex. 2. Show that

x2sin[-)
0) hrn =0. A. P 1989. 91 1
x-+O sin
(ii) Iim -
ir -x
= - l.
sin x
B. P 1995]

(iii) Jim (Sec 2x - tan 2x) = 0. C. P 1980 1

X-)-
4
(iv) tim xsin=0 I B. P 1990 1
—o x)
(!
2
(v) Jim (1-x) tan— = -. B. P. 1991, '93
x-i 2 it

sinx—tanx
(vi) Jim A. P 1992 1
A3 2
LIMIT 75.

x2sinh1)
'
Solution: (i) tim = tim x -sin( I
x-+0 sin -o x) sin
x

= urn (x) tim=0.

X-0 x-+O
J 1im-
x-*0 X
.
since, urn x = 0, urn Sin x 1, srn - :51.
x-,0 X-0X
sin = . — O)
(n) lini - urn sin(
x-fl2t — X 0-40 6
where. n-x=O .. x=it-O, and 0-0 as X-47t.

• sinG
= Imi - = 1,
0-40 0

(iii) lint (sec2x- tan2x)

-4--
1_ sin 2x) J(cosx _ Sin x)2 l
= tim
C0s2 = [cos2x- sin 2xj
4 4

cosx - sinx
=lim =0.
cos x + sin x
4
(1
tim xsin I -
(iv) x-40
1X )

= 0, since tim x = 0 and sin- 5 I.

x-*0 I X
71X
(v) urn (1-x) tan —
x-*O

= tim O(an(1_0)}, where l-x=0, i.e., 1-9 and

8-40 12
0-90 as x-31
• (,r -0
= lim0tanl----
8-4(1 2
 76 DIFFERENTIAL CALCULUS

•
8-cot —,
no
oo 2

,to
nO. 1 22
tim cos—• tim = -
0-0
sin
= tim O•
710
— 0-10 2 o-.o 710,
2 2
,to
2

sin x- tan xi ( • sinx

(v,) tim = tim — smx-----
-4O X-)O3 i COSX )
I .sinj 11
J'
• 1 (cosx-i'
sin.lurn 2s rn
X-40X COSX X-40 1COSX x x 22
2
f \ f. \ Sill- I
I 1 i ( SiflXl 2 1 I
= tim I - i tim I - tim -2.- -
X-+OCOSX)X-4O( x x 2

= lxi x(l) 2 (_.)= --.

Ex. 3, Evaluate

(i) urn 2+3'

2- +3
tim x2-xlog+logx-I
(10
x-1

(iii) it...
o tan2x
2 2
tim (a + h) sin (a + h) - a sin a
(iv)
h-O h
S
'(v) tim ____
'

+3+I
Solution : urn
(0 ,.•_ 2l+3
IMIT
-. 77

3i+If(
3
-urn ) n+}
-3
R-4- 3{1+()}

•
for, urn (-) =0 and urn () =0. •

Jim X2-xlogx+Iogx-1
(ii)
x-1
1jx2_I(x_1)Iogx

• x-1 x-1
= Iim[(x+1)-4xj . x-I*0

= 2-limlogx=2,-.-Jim Iogx=0.
X-41
1-cos3x
(in) urn
,
(l Cos x)(1+ Cos x+COSX)
lim -cos2 x
sin2x
(1cosx)(1+c0s2x+cosx) 2
=Jim -cosx
1-cos2x
=Jim • '.. 1-cosx;eO
x-.O 1+cosx
3
Jim Cos .=j.
2 X-40

(a+h) l sjn(a+/z)_a 2 sin a

(iv) urn
x-O • h

a2 ( sin ( a. + h)_ sj na)+2hasj fl(a+h)+h 2 sjfl(a+ h)

= urn
I h

( h. h
a2 2COSIa4----Isin-
•
Jim 2J2
h-,o h
+Ijm 2asin(a+h)+ Jim hsin(a+h)
h-,O
 J)IFFERENT!AL CALCULUS
78

.1.
SIn-
2 urn COS i a+
kin —2=2asina+O
2) i,-( h
2
=a2 .cosaX (l)+2asifla
=a2 cosa+2asina

(v) Inn ____

=lim- -
•— L r-
+4

lim

T'+
F+F
I

= urn _______ where v =

' -- ox
and as x-, y-O

EL 4. Evaluate:
(COS x+sIflX)5
• 4..h -
(i) urn
1— sin 2,
4
xtn2x-2xtanx
(ii) tim
X-30 (l- COS 2x)
I 1ir 1x
urn jtafl — + xJ!
x-+O 4

( IC
(iv) urn I xtan x - —seCx
1-4--' 2


LIMIT
79

(v) Jim -
-O si2x
Solution

J i•  4.Ji–(cosx^sinx) =
(1) m _________ 4
(t–sin2x)

I
–! 1–sjn2x
.4 4
51

4
•
Jim . ----
'fi
- _—cosx+-__sinxj

1
l –sin 2
4

7t'1
1Cos (
=4. urn
' ] -sin 2x
4
=4 . Jim 1–cos5ft it
Whe re , 0 v- - nd 0 (1 it
O-.Oj-c0s20 asx
4
=4 '
Jim — (lcosO) (1+coso + cos 2 o+cos 3 e+co 4 o)
2sin2O
26-
2sin 
=2 . Jim 2 X 
o-.o sin 2 U Jim (1+cos6+c0s20+COS3O+COS4O)

202 sin2
2.JIirn____ix
0-0 02 .
 sin 
—x4 llm_x02 X 02
4 e.oO2

Jim (l+cosO+cos20+c30+COS40)
8-90

(.ei2
1j IiI
2ñm 21i 1

e_o2{
9)
2
'
Jim sin
o

6
(1+cosO 1-cos 2 O +cos 3 0 + Cos 40)

1+1 5,F2.
 DIFFERENTIAL CALCULUS
80

2x tan x
------2xtanx
urn x tan 2x - 2x tan x = lim
(ii)
x-0 (1- cos 2x)2 x4O (2sin2 x)2

2x tan x1 --' l
1— tan jJIjm xtan3_
=lim 2.r-.Ox(1- tan 2x)
4sin4x
1 x tan 3x
=-hm
2x-,O xsinx(1--tan2X)
'3 I 1
1 itanxl
= -lim - x (sinx\4. x—,O1.—tafl2X
X
x
1 31 11
=-(1) x—X-=-
2 (1) 12
(tanx\ sinx I
since, hmt -j= tim —x urn —=lxl=1.
x—O X x-4O COS X.
x )

11+tanxl,,
r (
(iii) urn tanI _+\l
x-9O(
xJf =Iim
•_.i) 11-4-tanxJ

(tonx\ (ianx
1mm!-
-
urn (l+ tanx) taxurn
-4O
= - - = -.------------
(.anx\ (tinx
hml--
X ) -- - 4O X
I-- tanx
taflX
tim (t-tanx) tim (I-tanx)
x—*0

I ''
urn
2 ,where, tan x8,tanx -.
0-401
=__--- -
! I e1

tim (1p)fl
LIMiT.. 81

(iv) lim(xtaflx_Secx)

2
( g

sin '
(2122
J ( J
= Jim 2 =iim
co
Cos z

where, .. z-O as X4.

=Iim-1- Cos z-zcosz--

•z-4OSm z(2 2
Jim -z Cos z 1-cost
= fl. tim
:-.O slnz 2 z-,O sin
Jim Cos z 2sjn
_z-'O -!.jjm 2
• sinz •Z Z
2z-,o2sin-cos-
lim i
Z-+0( 2 2
)
!.-.1im.tan, .. sin !^O as zO.
12z-.O 2 2

2
(v) Let us substitute, cosx $6 - -.

Obviously, , -* I as x -, 0 and = 3, =
and sin2x=l-cos2x=L-112
• _________

• tim
,-.o sin2x
1 t2(tI
3 _12
r Iim -- lim_'
,+i i-rfl (r12 -1)
Jim (t2)

• 12_j
l i nt
1-4111.
- . n
-a-n
1 . tim
z
na
n-I
2i •
12(1)I -z-3a Z-a

6- 12
82 DIFFERENTIAL CALCULUS

EXAMI1FS-ffl

1. Evaluate the following limits:

2x -2. x 2 -3x+2
Li . (ii) Li
+
(i)
.-t 2x+2

a_a2 x2 11 ;2x -JI-3x

() Li (iv) Li
-.o x2 -.O X
2. Find the value of
aox" + + ... + a
Li (b,, * o)
.vO b,x"+ b 1 x +...+b.

3. Do the following limits exist? If so, find their values:

1 sin
(I) Lx—. (ii)Ii —-
x -s37 21 - I .t —.x Jr - I
4. Find the values of:

(i) tan (ii) :a sin(Lt.)

I - cosx l - cosx
Li . (iv) Lx 2
O X 3O x

2 sin (1/x) COSeC I - cot X

(v) L, (vi) Li
,-.o sin O I

sin x*sin-
. (viii Li
X
(vii) Li
-.o x .-.O X

tan 'x . (x) Li —.

Sin
(ix) Li
7-.0 I -3.. I

sin .. x+l
(xi)Li . (xu)Li
'-. X+ Cos X i- +1

(xiI)LI ( , - , )
Sin tan
LIMIT

5. A function f(x) is defined as follows:

f(x) =x when x>O.
=0 when x=O,
=—x when x<0.
Find the value of Li f(.i).
-4O

6. A . function 0 (x ) is defined as follows:

Q(x)=x 2when x<l,
=25 when x = I,
— x 2 +2 when x>J.
Does L1 #(x )exist?
.,-I
7. Do the following limits exist ?
(i) Li [x ], where [x I denotes the integral part of x.

(ii) Li {2+,JTTTi}.

(iii) Li
2
2 bx + c, show that
8. Given f (x)= ax

Lt
h

9. Given f(x) = lxi, show that

Li {f(h)_f(o)}/hdoesnotexist.
h—to

10. If(x) {(x+2Y 4}/x, show that

Li 0 (x )= 4. although 0 ( 0 ) does not exist.

K—tO

2-
11. Show that Li 2 x 8 = 8.

Appi' (s, e ) dc inition. 1 0-I.

84 DIFFERENTIAL CALCULUS

2 ' 2 '
X - a -
12. (i) Is Li - Li -s---- = Li—
x-a x-a x-a

(ii) Is Li (2 -a )xI __!.__ = IJ{(x2 -a 2 )x!}?

13. Evaluate
(1 2 3
(0 L i I —+---+—+
2 ... +--
-.=t

12+22+32 +...+n2
(ij)Li 3
fl -,_ n

14. Does Li f (x ) exist, when

. -.0
(I) f(x)=(2"+2+I12) ?

- (ii) f(x)=[ sin !+xsin!+x2 sin !) ?

15. Evaluate

(i) Li X'

X,,
(ii) Li . I C. 11.1957]
"-= x' +1
x" f(x)+g(x)
(iii) Li . I CH. 19561
X, +I

(iv) Li
' -. x +1

16. Find the value of Li 2 - arc tan nx.

,,-..- n
17. Evaluate
(1) Li sin n,rx.

(ü) Li [CH. 19571

2
-.= 1+nsin.nx
 85
UMIT

(ii)x L i CH. 19571

-3 - 1+nfl 2 nx -
18. (i) Prove that
Li tan --= — 7t, 0 or according as a is negative,
x2
zero or positive.
(d) Draw the graph of the function 1(x) where
(2x x
f(x)=-
Lt tan

19. If f (x )= L i 1 2 , show that f (x )= I. , 010

l+x
according asX<, = or ?l. (C.JL 19501

Draw the graph of f (x ) in this case.

20. The function y = f ( .x ) is defined as follows:

f(x)=0 when .r2>1
f(x)=1 when x2<1
f (x )= when x2l.
Using the idea of a limit, show that the above function can be
represented by
1 [C.P 19491
f (x ) = Li for all values of x.
1_+X 2,

ANSWERS
I . 5
1. (i) , (u) , (iii)j (iv)

3.(i) Does not exist. (ii) I.

b.
4. (i) 1, (ii) 0, (iii) 0, (iv)-- , (v) 0,

(vi) -j (vii) ISO (viii) I, (ix) I, (x) 0.

(xi) 0, (xii) 01(xiii) 0.

5.0. 6. Does not exist.
7. (i) Does not exist. (ii) Does not exist (iii) Does not exist
86 DIFFERENTIA L CA LCULUS

11. 0-05.
12. (i) No, (ii) No.

13.(i) I , (ii)-.
14. (1) No, (ii) No.
15.(i) +co when x>1; 0 when —1<x<1;I when x=1; nolimit
exists when <
(ii) 0 when - I <x <1; -- when x = 1; 1 when x!< —1 or >1; not
defined when x =
(iii) f(x) whenIxI>1; g(x) when IxkI;
{f(x)+g(x)} when x=1; undefinedwhen
(iv) —J when —1<x<I; 0 when x=1; I when IxI:J.
16. 1 when A>0; 0 when x = 0; —1 when x<0.
17. (I) 0 when x is an integer; no limit exists if x is not an integer
(ii) 0.
 çon

4.1. We have acommonSeflSe idea of what a Continuous curve is. For

instance, in Art. 2.5, the curves of example (ii), (iii), (v) are continuous, while
those of (vi) and (vii) are discontinuous, the curve in (vi) having a point of
0. A function f(x) is
commonly said to be
discontinuity at the origin
continuous provided its graph is a continuous curve, and, If there is any
discontinuity or break at any point on the curve, the function is said to be
discontinuous for the corresponding value of x. The general notions of
continuity of a function f(x) for any value of the variable require that the
function should be finite at the point, and for a very small change in x, the
change in the value of f(x) should also be small, or in other words, as we
approach the particular value of x from either side the function should also
approach the corresponding value of f(x) ,and ultimately coincide with it
at the point. If f(x) be non-existent at a point, so that the corresponding
point on the graph is missing, or else, if the value of f(x) suddenly jumps
as x passes from one side to the other of the particular value, or f(x)
becomes infinitely large at a point, then the function is discontinuous there.
We proceed below to give a formal mathematical definition of continuity.

4.2. Continuity.
A function f(x) is said to he continuous for x = a, provided
Li f ( x) exists, is finite, and is equal tof (a).
x.-4a
In other words, for f(x) to be continuous at x = a,
U f(x)= U f(x)=f(a)
0

orbriefly, f(a+O)= f(a—O) =f(a)

This may also be written in the form Li f(a+h)= f(a).

la, b] , it is
If f(x) be continuous for every value of x in the interval
said to be continuous throughout the interval.
A function which is not continuous at a point is said to have a
discontinuity at that point.

Ek (I). f(x)=x t is continuous for any value a of x,.

for, Li 52_a2
- 1— Ia
 88 DIFFERENTIA L CA LCULUS

I
(ii) f(x)=cos is discontinuous at x = 0, since Li cos does

not exist [SeeE.t,'I,l3.l0]

(iii) f(x) =-L

is discontinuous at = o,since Li is not

finite.

(iv) f(x) x sin ! when x * 0, and jo)= 0, then 1(x) is

Continuous atx=0, for Li xin 0.

x-+O X
[See Ex. (ii),* 3.61

(v) If f(x)= when x * 2, and 1(2 )= I, then

J(x ) is discontinuous at x = 2, since Ii f (

x ) L i f(x)bcre,

so that Li 1(x) does not exist.

'-.2

(vi) f(s) = e (' ) is discontinuous at x = a, since though

L i f(x)= Li f(x)=o, i.e., L i f(x)existsand =0,
X-. a*O f(a)
is undefined.
Corresponding to the analytical definition of limit, we have the following
analytical definition of continuity of a function at a point
The function f(x) is continuous at x = a provided f( a) exists and
given any pre-assigned positive quantity e, however small, we can
determine a positive quantity . ö such that I I (x ) - I (a ) I < e for all
values of x satisfying a - 8 :5 x :5 a + 6.
4.3. Different classes of Discontinuity.
(A) If (a + 0) * f (a -0 )thenf(x)issaidtohaveanordinwy
discontinuity at x = a. In this case. I (a ) may or may not exist, or if it
exists, it may be equal to one of f ( a + 0) and f ( a - o) or may beequal
to neither.
To these is to be added the case where only one of I (a + 0) and
f (a - 0 ) exists and f (a ) exists, but is not equal to that.
CON71NW TY 89

Illustration : I (x ) = 2 + e has an ordinary discontinuity at x 0,

for Li f(x)=O,and L i f (x )=-'.

2

Note. Continuity on one side.

In case wheref(x) is undefined on one side of a (say, foi x > a), if
f(a + 0) exists and is equal 101(a) (which also exists and is finite), we say,
as a special case, thatf(x) is continuous at x = a.

(B) lf f (a+0)=f (a-0)* f(a),orf(a)isnotdefined,thenf(x)

is said to have a removable discontinuity at x = a.
Illustration: f(x) = (x 2 - a )l (x - a) has a removable discontinuity at

.For, f(o) is undefined here, though Li

x-4G
1(x) exists, and = 2a.
Again, if f(x)=1 when xa, and f(t)=e when
x a, f(s) has a removable discontinuity at a, for, Ii 1(x) = 0.
x -UI

whereas f(x)=l as defined.

It maybe noted that a function which has removable discontinuity at a
point can be made continuous there by suitably defining the function at the
particularpoint only.
The two classes of discontinuities (A) and (B) are termed simple
discontinuities.

f (a +6) and f (a-i)) tend to + or

(C) if one or both of ••
then .f(i)is said to have aninfiniteiscoUinUiiyata. Hem, f(a)
mayor may not exist.

Illustration: f(x)= e -'° has an infinite discontinuity at x = a, since

f(a-0)_-,co, f(x)= 3x' hasaninfinite discontinuity at x=2.

2
(x-.2)

(D) Any point of discontinuity which is not a point of simple

discontinuity, nor an infinite discontinuity, is called a point of oscillatory
discontinuity. At such a point the funtion may oscillate finitely or oscillate
infinitely, and does not tend to a limit, or tends to to + or -
9() DIFFERENTIA L CA LCULUS

Illustration: j (x ) = sin! oscillate finitely at x = 0.

1 sin—' oscillates infinitely atXa. Li x (x = -i)

• f (s) —

oscillates finitely, and Li x" (x <-I) oscillates infinitely as it

4.4. Some properties of continuous functions.

(1) The sum or dtfference of two continuous funcitons is a continuous
function
i.e., if f (x ) and 0 ( x ) are both continuous at v = a, then

f (x ) ± 0 (x ) is continuous at x a.
For in this case, by definition of continuity, L i f (x ) exists, and
S -40

= L t f(a),asalso Li 0( x ) = 0(a).
x-t0
Hence, L :{ f ( x) ± Ø(x )}= L if (x )± LtØ(x )

lSee 3.8(1)1
=f (a)+ 0(a),

whence, by definition, f ( x ) ± 0 (x ) is continuous at x = a.

Note!. The result may be extended to the case of any finite number of
functions.
Note2. If f (x) is continuous at x = a, and 0(x) is not, then
(x).
f (x) ± $ (x) is discontinuous at x = a, and behaves like O
(ii) The product of two continuous functions is Continuous function;
i.e.. f(x) and 0(x) being continuous at x = a, f(s) x 0 (x) is
continuous there.
Proof is exactly similar to that in the above case, depending on the
corresponding limit theorem (See § 3.8 (ii)].
Note. This result may also be extended to any finite number of functions.

(ii,) The quotient of two continuous functions is a COnhiflW '145 fUflCitOfl,

provided the denominator is not zero anywhere for the range of values -
considered;
i.e.,if 7(x) and 0(x) be both continuous atxa,and , (a 0
CONTINUITY 91

then j (x ) / 0 (x ) is continuous there.

Proof depends on the corresponding limit theorem [See § 3.8(iii)].

(iv) If fix) be continuous at x = a, and f(a)*O, then in the

neighbourhood of x = a,f(x), has the same sign as that off(a), i.e., we can
get a positive quantity 8 such thatf(x) preserves the same sign as that of
f (a) for every value of x in the interval a-8 <x< a+ 8.

Let f(x)= sins, a = I n; then f(a)= land hence * o and positive.

Lei us take S = it. Then in the interval -ir — 4ir <x< -!r + + 7T i.e.,
-L it <x< . I(, f(s) is always positive.
Sincefx) is continuous at x = a, from definition, if e be any chosen
positive number, we can get a positive quantity & such that

If(x)—f(a)kc. i.e., 1(a)—c <f(x)< f(a)+e

(1)
foi all values of x satisfying a-8<x< a-t-8-
As f (a 0 here, if f (a ) be positive, choose e = - I (a )
then from (1), 1(x)> f(a )— E, Le., > f(a),
and isaccordingly

positive when a - S <x <a +5.

If 1(a) be negative, choose F = -- 1(a), and then we have
fi'oin(l), f(x)< f(a)+E i.e., < f(u) - f(a),
and is accordingly negative when a — S < x < a + 5.
Thus, whatever be the sign of f(a) , we can find 8 such that f(s) has
the same sign as that off(a) in the range a-ö < x< a+8.

(v) if f x) be continuous throughout the interval [a, b], and if

f (a ) and f (b) be of opposite signs, then there is at least one value,
say x, of x within the interval for which f ( ) = 0.

Let f(x)= cos x, a = 0, b = it. Then f(a)= 1. f(b) - I.

Now, cos x = 0 if x = 1 7r, which obviously lies in the interval (0,n). and
so here 4 = -ir . Similarly, ifwetake a=0. b=3n, we get another value

of t, viz.,.-,r,besides .ir.
92 DIFFERENTIAL CALCULUS

Let OA =a, OR=b. Bisect the interval lB at C,.lff(x) be not zero

at C, , it must be opposite in sign to one off (a) and f(b) which are given
to be of opposite signs. Suppose f (.) has opposite signs at C 1 and B.
Bisect C, B at C2 . If f(x) be not zero at C 2 , it must have opposite signs
at the extremities of one of the intervals C1 C 3 or C2 B. Bisect that particular
interval at C3 . Proceeding in this manner n times, unless f(x) is zero at one
of these points of bisection, we can get an interval C. - , C. (say) within
AB, at the opposite extremities of whichf(x) will have opposite signs. This
interval C,,_ 1 , C. is clearly is clearly 1/2" of the interval AR. i.e..
= (b— a) / 2" ,and taking n large enough, can he made as small as we like.
But f (x ) being a continuous function for every value of x within
the interval AB, corresponding to any point C . in it, given by x = c say, if
f(c) be not zero, it must be possible [hs (iv) above Ito get a positive quantity
6 such that f(x) will retain the same sign, namely that of f(c), in the interval
[c-6,c+ö ]. Now whatever 6we may choose, =(b—a)/2" can be made
less than 6 by taking n large enough, and it has been shown that the
extremities of the interval C,,-,, C,, which falls within [c —6, c +8 1, f (x )
has got opposite signs. We are thus led to a contradiction if f (x ) is not
zero anywhere within the interval AB. Hence there must be some point in the
interval, given by x = 4 (say), where f ( 4 ) = 0 under the circumstances.

(vi) If f (x ) is continuous throughout the interval I a, b ] and if

f (a ) * f (b.), then f (x ) assumes every value between I (a ) and
f (b ) at least once in the interval.
Let. [(x)=x2,a=O,br4; then (a)= 0, .f(b)=I.
• Let cbe any number between O and l.Then f (x ) X 2 c, which
evidently lies in [0, 11.
5
Let f(x)= sin x.a =0,b= ; then f(a)=0 f(b)= 1, so
2
f (a I (b ).Let c be any number lying in [0, 11. Then sin x c, if
x = n,r + (— I)" sinc, n = 0. +1, ± 2,... Now, for n=0, 1,2 only,

xfles in the interval O, 1. That is, when x = sin c • or 1r—sin c, or

L 2J
2r + sin c, we have f ( x )= . Thus, f ) assumes the value c at
least once ( here 3 times and in the previous example once only).
CONTINUITY 93

Let k be any quantity intermediate between 1(a) and f(b) which are
given to be unequal. Let 4, (x ) = f (x ) - Then sincef(x) is continuous
in the interval [a,b], 4' (x ) is also continuous. Also 4, (a ) = f ( a )- k
and 4, (b ) = f (b ) - k are of opposite signs. since * lies between f(a)
and f(b). Hence by (v) above, there is a value x = 4 in the interval, for
which 4,() = O,ie., f() = k .In other words, f(x) assumes the value
k at some point in the interval.

(vii) A function which is continuous throughout a closed interval is

bounded therein.
The function! (x) = sinx is continuous in -the closed interval
0 x :5 2r, and has the upper bound at x = - - 7r and lower bound at
x=0 or p, hence it is bounded.
Let the functionf(x) be continuous throughout the closed interval
[a, b]. Let us divide all the real numbers in the interval into two classes L.
R, putting a number x in L if f(s) is bounded in (a, x), and in R otherwise
Members of L-class exist in this case, sincef(x) being continuous at a (u'
the right), corresponding to any pre-assigned positive number a we can
get a positive numberdsuch that I (a )- e < I ()< I (a )+ E (and
accordingly f(x) is bounded) in the interval (a, a + 81 ,so that a+ö belongs
to L-class. If now numbers of R-class also exist in the interval, then by
Dedekind's theorem, ther exists a definite number c (say) in the interval,
which represents the section. f To include all real numbers in the classification,
we put all numbers less than a in L, and all numbers greater than b in R
here. I
Now sincef(x) is continuous at c, for any given positive quantity
A 'te can determine a positive number d such that
1(c)—c < f(x)< f(c)+ . e withintheintervai(c --S, cs-S ),
i.e., f (x) is bounded therein. Also c - S belonging to L-class. 1(x) is
bounded in [a • c —6 1. Hence, f (x) is bounded throughout the interval
(a • c +5 ). But c + 5 belonging to R-class, f (x) is not bounded in
[a, c - &]. This contradiction shows that no number of the R-class can
exist in the interval [a, b]: in other words, f(x) is bounded throughout the
interval [a, b J.

(viii) A continuous function in an interval actually attains its upper

and lower bounds, at least once each, in the interval.

The iunctionf(x) = sinx is continuous in the interval 0 :5 x :5 it. Its

94 DIFFERENTIAL CALCULUS

upper bound I is attained at the point .x = it and the lower bound 0 is

attained at the point x = 0 and x = p. Thus, f(x) attains its upper and lower
bounds, at least once each (here the upper bound is attained once, whereas
the lower bound is attained twice).
(tx) A function f(s), continuous in a closed interval I a, b J ,atfains
every intermidiate value between its upper and lower bounds in the interval,
at least once
Let f(x)=x 2 . a=-1. b=2. then the upper bound of f(x) is 4 and its
lower bound is 0. Let c be any number in 10, 41. Now, if 0 ^5 c !^ 1, then
f(x)= x 2 = c, if .r = + rc which liesin (—I. 2); and if 1< c <4,
then f(s) = x 2 = c if .x = ±'.J, of which only + vrc lies in the interval
(-1, 2). Thus,f(x) attains the value c at least once.
Letf(x) be continuous in the closed interval L a, b 1, and let M and nz
be its upper and lower bounds in the interval. If possible, let there be no
point in the interval wheref(x) = M. Then M - f (x )> (I for all points in
the interval. Now, since f (x ) is continuous, M - f (x ) is also
continous, and so I / { M - f (x ) } is continuous in the interval. Thus
1/ {M - f )} is bounded in the interval, i.e.,
1 11 M - f (x ) } 5 k, where k is a fixed positive number.

M_f(x)'a, or, f(x)!5M-.

This contradicts the assumption That M is the upper bound of f(s) in

the interval.
Hence,f(x) must assume the value Mat some point in the interval.
Similarly, it may be proved thatf(x) assumes the value m alin in the
interval.
It now follows from (vi) thatf(x) assumes every intermediate value
between M and m.
4.5. Continuity of some Elementary Functions.
(i) Function x. where is any rational number.
We know that Ii x' = a ", for all values of n, except when a = 0
and n is negative[ See Note, Ex. 1, § 3.111.
Hence, x' is continuous for all values of .x when is positive, and
continuous for all values of x except 0 when n is negative.
 CONTINUITY 95

When n is negative and = —m, say, where m is positive

=x=l/x",
which either does not tend to a limit or .—, as x -, 0.
(ii) Polynomials.
Since the polynomial an x + a 1 x' +... + a" isthesumofa
finite number of positive integral power of x (each multiplied by a Constant)
each of which is continuous for all values of x, the polynomial itself
by §4.4(i)] is continuous for all values of x.

(iii) Rational Algebraic Funclion.i.

Rational algebraic functions like

a 0 x"+a 1 x" t + ... +a"

b 0 x " + b 1 x "-+
' ...+ b
being the quotient of two polynomials which are continuous for all values of
x, are continuous for all values of x except those which make the denominator
zero I b y § 4.4(iii) 1.
(ii') Trignometric Functions.
Since the limiting values of sin and cosx when x -4 a ,where a has
any value, are sin a and cos a [See £z. 2, § 3.11 ] ,it follows that sin .s and
cos x are continuous for all values of .x.
Since tan x = sin x / cos x, tan x is continuous for all values of x
except those which make cos x zero, i.e., except for x = (2n + I )4,r

Similarly, secx is continuous for all values of x except for x (2n + 1 )'r
and cot x and cosec x are Continuous for all values of x, except when
x = 0 or any multiple of it when sin x = 0.
(v) Inverse Circular Functions.
Inverse circular functions being many valued, we make a convention of
defining their domains in such a way as to make them single-valued
Throughout the book we shall suppose (unless Otherwise stated ) that
Sin x, tan x, cot -1 x, cosec -l x lie between ---fl' and j- it (both
values inclusive) and cos x, sec' x lie between 0 and it (both values
inclusive ), which are the principal values of these inverse functions. It should
be noted, however, that and have no existence outside the closed interval of
x, and and have no existence outside the closed interval [-1,1] of x, and
cosec - ' v n"-1 sec t x have no existence inside the open interval (- II ).
96 - DIFFERENTIAL CALCULUS

All the inverse circular functions are continuous for all values for which they
exist this follows immediately from the continuity of the corresponding
circular functions.
(vi) Function e.
Corresponding to the positive number c, however small, we can choose
n sufficiently large such that (i + e )" >e ,since (l+eY' >1 +ne, and
e

is finite.
Thus, e-l<e.
Hence, if 0<x<1/n,e-l<e -l<c, •o*

and therefore Li (e - I )= 0, or . L i e' = 1.

If x be negative. putting x = -y, Li e' = L i 1 1 eV =1.

_,-.o--

Hence, Lt e = 1.

Li =1. i.e., L I eV =e.

e' is continuous at any point x = c.

(vii) Function log x, x >0.
It should be noted that log x is defined only for values of x >0.
Let logx=y and log(x+h)y+k.

Then e"=x and e"'=x+h;

h= -e'.

As e is a continuous function of y. -' e

i.e.,lI-40ask-*
Thus. { log (x + It ) - log x} -* 0 as k -4 0,
i.e., as h -4 0.
Hence, log x is Continuous.

4.6. Ulustratve Ewnp1s.

Ex. 1. A function 1(x) is definedasfollows:
f (x)rx when x>0,f (0)=0,f (x)x whcnx<0.
Prove that the function is continuous at x = 0.

Here. L i f (x )= L i x=Oand L i f (x)- L i (-x)=O.

x-,O+O
 - 97
CONTINUITY

Thus L i f (x )= L i f(x)f(0)'0 here.

,-.0-0
Hence, f ( x ) is continuous at x = 0,
For its graph, see figure of § 2.5 (ii).
Ex. 2. A function 1(x) is defined as follows:
f(x)=xsin- for x*0
=0 f or x 0,
Show that 1(x) is eon: inuous at x = 0. [V P 19991

Since Isin(1/x)I 15 I ,bymaking IxI<e.

wecanmake lxsin(1/x)ke,
where € is any pre-assigned positive quantity, however small.
Hence, Ii xsin--=0. Also, f(0)=0.asdefined.

Thus, L i f (x )f (0)
For its graph, see figure of §2.5 (viii).
Note. It should be noted that the function .x sin (1/x) is continuous for all
values of x, except for x = 0; because when x = 0, .x sin (1/x) is undefined.
In the above example, the discontinuity of x sin (Ifs ) at x = 0 has been
removed by definition of f(0).
Ex. 3.: A functionf(x) is defined as follows:
f (x )-x when

when x=-
2 2'
when 4<x<1

Show that f(x) is discontinuous at x

He, 12 f (x )=li(x =_=0.

_x)= 3_ 1
ii f (x ) 12

2 2
7-
98 DIFFERENTIAL CALCULUS

Since Li 1(x) does not exist,

2
1
hence f(x) is discontinuous at x =
Ex. 4. A function f(x) is defined in (0, 3) in the following way
f(x)=x 2when 0<x<1,
=x when 1:5x<2,

= l x3 when 2:5x<3..
4
Show that f(x) is discontinuous at x 1 and x = 2. [C. P. 1941 1
when x=I, f(x)=x. :. f(i)=i.

Li f(x)= Lt
1-41-0
Li f(x)= Li X=I.
also,
x-.I+O

Hence, Li f(x)= Li f(x)=f(1)here

.—.lco
f(x) IS Continuous at x= 1.
Similarly, it can be shown (from the definition of the function in the
relevant ranges) that Li f (.v )= Li f(x ) = f(2) = 2

Hence, f(x) is continuous at x = 2.

Ex. 5. Show that the function f ( x ) = I x t-t-I x - 1 + x.- 2 1 is continuous

I I
at the points x = 0, t, 2.

1XXXX+-(x-1)-(x-2)=-3x+3, for.x<0
Here,f(x)=- x-1 - x-2 =-x+3, for 0^x'<l
x-I - x-2 =x-t-1, for 1^x<2
+ x-1 + x-2 =3x-3, for x^t2

Now, Li f(x)= Li (-3x+3)=3

Li f () = Li (—x+3)=3 and f(o)=3.

CONTINUITY . .. . 99

Li f(x)= Li f(x)f(0). .1( x ) iscontinousat x = o

Again, Li f(x)= ii f(x)=f(1)=2,

and Li f (x)=L s f(x)=f(2)3

1(x) is continuous at x = land x 2.
Ex-6- Show that the function f defined by f (x )= x - [x j , where [x]
denotes the integral part of x is discontinuous for all integral values of x
and continuous for all others. -
for
x—(a—l), for a—I<x-<cc
Wehave, f (4= O, x=a
I x -mfora<x < a +I
where a is an integer.
Ltf(x)=Li(x—a+l)=l

Ltf(x)= LI (x—a)=O

Li f(x)* Li f (x)=f (a)

x-a-
1(x) is not Continuous at = a.
Since a is any integer, f(x) is discontinuous for all integral values of x.
1(x) is obviously continuous for other values of

Ex. i. Show that the function f defined by

f(x)= X_l. for X^1

=0, for x=1

is continuous at x = 1.
el _-ooas x-1—=' Li f(x)=0
e — o as x—l+= Li f(x)=0

So LI f(x)f(l) Li
f(x)rrO

Hence; 1(x) is continuous at x = I.

too !)IFFERENTIAL ('ALCULUS

4.7. Some Miscellaneous Worked out Examples

Ex. 1. Discuss the continuity of the following functions at the points

indicated:

(1) f(x)=x when O<x<L [C.? 1989, 97, 20051

=2.-x whenl!5x!^2. atx=2
=x---x 2 when x>2

(ii) f (x)=--- ' , x#0 [B.? 1994]

3x al x=0
=, x=0

(iii) f(x)=x 2 +x, 0:5x [C. P. 19921

=2, x=1 atx=1 -.
=2x 3 —x+1, 1<x—<2

= . +2
+4
(iv) f(x) x#0 [C.? 19941
sInx - 0
at5—
=0, x=0

Solution (i) Here urn f(x)= lim (2-4=0

Urn f(x)= urn (x_! x2)=0

• ,—.2* r-2+ 2
and f(2)=2-2=0
Jim f(x)= urn f(x)=f(2),
f(x) is continuous at = 2.
(ii)We have,
tan 2 x . Isinx 1 x
Jim f(X ) = Jim =
-'O 3x x Cos 2X 3

Jim
tsInxI ,
Jim I
I I .
•

ii
,,

x) .r—O',,cc2) .-..ot.,3
=(I)2 x(1) 2 x(0)=0...
 101
CONTINUITY

But f(0)=
tim f(x) * f(0), f(x) is not COntinUOUS at x = 0.

(iii) Here, tim 1(x) = tim (x2 +X)= 2

urn f(x) urn (2x3_x+1)2

and f(l)=2
So 4im f(x) tim f(x)f(l)
1-41+
Hence, f(x) is continuous at x = I.

f(x) =
+4x3+2
, when x 0,
(iv)
nnx
+2 lim (x 3 +4x2 +2) 2 -
x3 +4x2
urn
1-40 1-0 sm x si'\
i. (sin I
t-m

But f(0)=O
Thus tim f(x)*f(0)
Hence, 1(x) is not continuous at x=O.

Ex. 2. (i) Find 1(0), so that .f(x) = xSin().

C. P 1982, '86]
for x * 0 may be continuous at x0.
sin(a2x2)
f(x)— '
x^O and f(0) =k. Find the value of
(ii)
B. P 1992 1
for which f(x) is Continuous at x=0.
defined by
(iii) What should be the value of f(0) so that f
2
x—x C. P 1997 1
f(x) = -, for x *O.•bc continuoUS at x = 0?

Solution : ( Here, f(x)=xsifl(!}X^0.

(!) ^ i , iini .x sin()= 0.

for all real values of x
 102
DIFFERENT/AL CALCULUS

If 1(x) is co ntinuous at x = 0, urn

f(x) = f(0)
f (0)=O, . Jim f(x)rO

(ii) Here, f (x )= ('L 2, x*o

x

Jim f(x)=Iim sin(a2x2

-*o .,-.o

si n(a2x2)
= urn -------a2x

sin(a2v2)
a2•lim
x-0limx----•a2xlx(0)—O
a 2x -..O
Since 1(x) is COntinuous at X0,f(0) Jim f(x) i.e., k =0.
..0
X2 _ X
(iii) f(x) = -, - t (j
x
xtO
Jim 1(x)

So, in order that f(x) may be Continuous at x = 0,

1(0) = Jim f(x) i.e.. 1(0) = — I.
..0
lOg(l+ax)_Iog(1_bx)
Ex. 3. The function 1(x ) -
x is not defined at
x = 0. Find the value of 1(0) so that f(x) is continuous at x = 0.

Solution : Here, Jim f(x)= urn Iog(l +ax)_Iog(l_bx)

x

=a-Iim I0(1f_(_)1j
ax -o (—bx)
= ci X I + b x (i) = a + b.
For 1(x) to be COflt j flUou
at x = 0, f(0) should be defined and
the value of 1(0) must be equal to the limiting value of f(x)as
x-0. Hence, PO) =a+/,
 - 103
CONTINUITY

of a and b such that the function

EL 4. Find the values

f(x)=x+.Jasnx 0:5x'-5

=2xcotx+b, —<—x<--
4 2
=acos2x—b sin x,
4
is Continuous for all values of x in the interval 0:5 x 5 7t.
Solution : The function f(x) will be continuous for all values of x in
TE
and x So, we must have,
0:5x!5lt, if it is Continuous at

Urn (1)
lini f(x) f(x)=f()

It (2)
and urn 1(x) = urn f(x) f( 2)

Now, urn tim f(x+,ñasiflX)=_+a

f(x) 4

limf(x)=lflfl (2xcotx+b)=+b

and

Also, tim f(x)= lint (2Xc0tX+b)

urn f(x) urn

2 2

and
From (1) and (2) we have,
(3)
or, a—b—
(4)
b=—a—b, a-2b
It It
Solving (3) and (4), a= - b=--1-.
 104
DIFFERENTIAL CALCULUS

Ex. 5. The function is defined as follows

f(x)'-2 sin x, -7tx<-1

asinx+b, ---<x— itit

2 2'

2
If 1(x) is continuous in the interval —n < x < it
find the values of a
and b.

Solution : For continuity of fix) at x = -

71
urn f(x) = urn f(x)=f(_

i.e., rn(_2snx)
l Jim (asflx+b)=_2fl(...11)

i.e., —a+b=2
(1)
And, for Continuity off (x) at x =

Jim f(x) 1im

i.e., Jim (asinx+b)= Jim (cosx)=cos(J

2 2
i.e., a+b=O
(2)
Solving (I) and (2), we get, a=-1,b=1.

Ex. 6. Find the points where the function

f(x)= - is
discontinuous. log I X

Solution: 1(x) = _____

logx
Obviously, the function is not defined for x = 0, and hence x = 0
s a point of discontinuity of the function.
Again, when log Jx I = 0, i.e., x ±1, the function is not defined.
Hence, the function has three points of discontinuity, viz.,
=-1, x=0 and x1,
 CONTINUITY 105

Ex. 7. Show that the function f(x) [x]+[—x], where'her [xJ denotes
the greatest integer not exceeding x, has removable discontinuity for
integral values of x.

Solution: Let, x = k be any integer.

— k and
[k]_

f(k)=k— k=o
Now, urn 1(x) = urn f(k + h)

urn [k+h]+Jirn [— ic— h] =k— (k+l)=--I

h-.O

and urn f(.r) = lim f(k — h)

h-.O

lim[k— h]+lirn[_k+h] =(k— l)— (k)=-1

h-.O h-.O

limf(x)=-1, but f(k)=O

So, f has a discontinuity at x = k, where Ic is any integer. If

however, we define f(k) = — 1, then the function becomes continuous
at x=k.
Hence the function has a removable discontinuity for integral values
of X. -

Ex. S. Let 1(x) be a Continuous function and Q(x) be a

discontinuous function. Prove that f( x)++( x) is a discontinuous
function.
Solution : Let, tji(x) = 1(x) +,(x), where f(x) is a continuous and
$(x) is a discontinuous function.
If possible, let W(x) be a continuous function.
Since, 1(x) is a Continuous function, W(x) — f(x) is also a
continuous function, i.e.. •(x) is a continuous function. But this
contradicts the given condition.
So, W(X) i.e.. 1(x) + $<x) must be a discontinuous function.

Ex. 9. Let, f be a function, such that for all real values of x, y.

f(x+y)=f(x)+f(y). Iff is continuous at x=a, then prove that
fis
Continuous for all real values of x.
 DIFIERbJ'ITIA L CA LCUluS
106

Solution Since f (x) is continuous at x=a

Jim f (x )f (a)

or, f(a) = tim f(a+h) = tim [f(a)+f(h)].

[.f(x+y)f(X)+f(Y)]

=Jim f(a)+lim f(h)=f(a)+li m f(h)

h-.O h-0 h-.O

tim f (h)0 . .,. (1)

h-.O

Similarly. f(a) = urn f(a - h) = urn f(a) + urn f(—h)

= f(a) + urn f(—h)

l,-.0
limf(—h)0 ... (2)

Now, let k be any real number.

tim f(k +h) urn [1(k ) + f(h)J
h-*O h-*O

=f(k)+lim f(h)f(k) r from (1)]

and, tim f(k — h) = Jim [f(k)+f(— h)]

h-90

= urn f(k)+ Jim f(h)

h-.O

=f(k) L from (2)]

Also, putting x = 0, y = 0, we have f(0) = f(0) + f(0),

so that f(0)=O
f(k)__f(k+0)__f(k)+f(0)f(k)

Thus tim f(k+h)= urn f(k+h)— f(k)

1(x) is continuous at x = k.
R.
Since k is arbitrary, 1(x) is continuous for all XE
sinx is continuous for all real values of x.
Ex. 10. (i) Show that f(x) =
(B.? 1999
 CONTINUITY
107

(ii) Apply e –8 definition to show that the following functions

are Continuous at the indicated points

(a) f(x) = x X,0 L .0 P 1981, '95 1

X=O
f(0)=O

(b) 1(x) = x2 cos(±) X * 0 at [B. P 1996 ]

f( 0)=O

Solution (1.) f(x) = sinx will be Continuous at x a, if for any

E)O, we can find 8, such that Inx-pina <e, whenever,

x–aI<6
x+c(
We have, Isinx-sinaI=22sin
sinx–a
—cos.---

x–a x+aI
=

Since, cos x 15 1, for all real values of x

and sin— x-a

x-a <1for0<
- x–a it
2 2 2 2

I Siflx–s! flQI1sin___
I x-a x+a x-aI
cos-- <2 -----

1xakE for lx_al<8

The relations are satisfied by taking S =E
So, I slnx — sinal<E, whenever,
So, f(x) = sinx is continuous at x=u. Since a is any real number,
sinx.is continuous for real values of x.

(II) (a) f(x) = rsin( !-), x 0

x) r at x=0
f(0)=0
i.

Now,
 DIFFERENTIAL CALCULUS
168

.1

A
=IxI

15x.since sin— ^l
A

for X-01< E.
The relations are satisfied, if S =e. So, f (x) is continuous at x 0.

(b) Here, f(x) = x 2 when x *0

cos (-),
f(0)=0.
For continuity of f W at x = 0, we are to find a 5 depending upon
, such that
I f(x) — f(o) 1 — ' for Ix_01<&

i.e., x2 Cos ._ 0<€ for, lxI<5.

Since cos-1- !^ 1, relations are satisfied, if we take l- <€ for

lxl<5.
So, the relations are satisfied if S =
Hence, j (x) is continuous at x =

EXAMILFS.JV
1. A function I (x ) is defined as follows:
f ( X ) = X2 when x * 1, f(x) = 2 when X = 1.
Is continuous at A = 1?
2. Are the following functions Continuous at the origin?
;e 0, f(0) = 0-
c') f(x )= sin(l/ x) when x
(fi)

(iii) f(x) xcos (l/x) when x

* 0, f(0) = 1.

sin -
(iv) I ( ) = when x 0.

x
=1 when x=O.
 CONTINUITY 109

= sin x
(iv) I (x ) when x 0.
=0 when x=O.
3. A function 0 (x.) is defined as follows:
Ø(x)=x 2 when x. < 1,
= 2.5 when x = 1,
=x 2 +2 when x>1.
Is 0 () conIiiious at x = 1 ?
4. A function 1(x) is defined in the following way;
f(x)=-x when x!50,
=x when 0<x<l,
=2—x when x^l.
Show that it js continuous at x=0 and =I. IC.P 19421
S. A function f (x) is defined as follows:
f(x) = 1,0 or — i according as x >,= or <0,
Show that it is discontinuous at x = 0.
= x 2 716
6. The function f (x) 6 is undefined at x = 4.
What value must be assigned to f (4), if f (x) is to be continuous at
x=4?
7. Determine whether the following functions are Continuous at x = 0.
i(o)=
(i) f(x) =(xI + X3 +2x 2 )/sinx, 0.
(ii) f(x)=(x+4x+2x)/sinx, •f(o)=o.
8. Find the points of discontinuity of the following functions:
x 3 +2x+5 .. ) x3+2x+5
0) 2 '
xr-8x+12 x 2 —8x+16
9. A function f(x) is defined as follows:
f(x)=3+2x for —^
•..x<o
=3-2x for
=-3-2x for x>'

Show that f(x) is coitinuous at x = 0 and discontinuous at x =

 DIFFERENTIAL CALCULUS
110

10. The function y =f(x) is defined as follows: f(x)

= 0 when. 1(x) =
X 2 <1, f(x) = when x 2 = I. Draw a diagram of the
when
function and discuss from diagram that, except at points x = I and
x = —1, the function is continuous. Discuss also why the function is
discontinuous at these two points although it has a value for every
value of x.
Examine the continuity of thc functions at x = 0 (Ex. 11-14)
11. f(x)=(I+.x)'1'. when 0

=1' when x=0.

12. j' (x )= (i + 2x ) 1/1 , when , 0

e2 when

13. f ( x ) = e- '/-' when x()

=1. when x=0.

14. f(.r ) = when x 0

I +e
=
15. The function I when x =0.
is defined by

f ( x )= 2x - Ixl+ sin '-,for x^0

=0, for x=0,

where fxI denotes the greatest integer not greater than x.
Examine the continuity of I (x ) a x = 0 and x = 2.

ANSWERS
1. No
(iii) No. (iv) No. (v) Yes
2. (i) No. (ii) ' Y es.

3. No. 6. 8. 7. (i) Continuous. (ii) Discontinuous

8. (i) 6.2. (ii) 4. 11 Discontinuous. 12. Continuous.
13. Discontinuous. 14. Discontinuous. 15. Discontinuous.
 SEQuENa

5.1. Set ofRealNumbers.

A set of real numbers is well-defined collection of objects which
are called members or elements of the set. The term 'well-defined' means
that given any real number, it can be determined without ambiguity
whether the real number belongs or does not belong to the set.
Examples:
1. The set of natural numbers: N = { 1,2,3,...... n, ... }
2. The set of all integers: Z = f0,,± 1, ± 2,...., ± n,... }
3. The set of all integers between 3.1 and 8.7. X = (4,5,6,7,8)
• A set is finite, if it is empty or contains a finite number of elements,
otherwisea set is infinite. The set defined in example 3 above is finite,
while the sets inexamples land 2 are infinite.
5.2. Greatest and Least Members of a Set.
A number is the greatest member of a set Sofreal numbers, if
I. is itself a member of S. and
L
2 L2: x, where xis any element of the set
Similarly, I is the least member of a set of real number, if
1. lisa member of S, and
2 l:5x, where xis any member of the set
Examples:
1. In the set of natural numbers (1,2,3...........}, I is the least member,
but it has no greatest member.
2. For the set of numbers I is the greatest member,
but it has no least member.
3. For the set of numbers f7,8,9, 10, 1 L 12 } .7 is the least member and
12 is the greatest member.
5.3. Bounds of a Set.
Given aset S of real numbers, if there exists a number G such that
x < G. for every number x of S. then we say that the set is bounded
above and G is an upper bound, or a rough upper bound.
 DIFFERENTIAL CALCULUS
112

If G is an upper bound of a set S, then any number greater than G is

also an upper bound of S. So. if a set is bounded above, the number of
upper bounds is infinite. The least M of all the bounds is called is the
exact upper hound or, the least upper bound or Supremum.
Similarly, if there exists a numberg such that x ^! g ,for every number
in S. we say that the set is bounded below and g is called a lower
bound or, a r.ugh lower bound.
The least m of all the lower bounds is called the exact lower bound
r the greatest lower bound or, infimum or the lower bound of S.
5.4. Neighbourhood of a Point: Points of Accumulation.

(i) Let 4 be a real number and E be an arbitrary positive number.

rhen the Set of real numbers in the open interval ( - e, 4 + C ) is called
to C -neighbourhood
of . For each separate choice of e , we may form a
;eparate neighbourhood of .
(ii) D.leted Neighbourhood The set of real numbers in the .pen
nterval ( - c, + e )
excluding the point itself is called the deleted

E -neighbourhood of , where is a real number and e is an arbitrary

positive number, however small.
(iii) Point of Accumulation: A number 4,which mayor may not belong
to a set of real numbers, is called a point of accumulation or cluster point
of S. if every neighbourhood of , however smalr, contains an infinite
number of members of S.
Evidently, a finite set of real numbers cannot have any point of
accumulation. Cluster point is also called Limiting Point.

Examples:

1. L=l, m=O, M =1, limiting

point is 0, but l does not exist.
2 For the set..... L=. m=, M rl.

limiting point is 1. but L does not exist.

5.5. Sequence of Numbers.

A set of real numbers X1.X2,X3 ...........
x,, .....such that
corresponding to every positive integer n, there exists a real number x,,
of the set, is called a sequence. The individual numbers are called
flh element is x,, is briefly
elements of the sequence. The sequence whose
denoted by {x}. If the sequence terminates after a finite number of
SEQUENCE 113

terms itis called afinute sequence, otherwise,it is an infinite sequence.

In what follows, we shall be concerned with infinite sequences only and
the word infinite may not be used always.
Examples.

1. { 2,4,6,8, 10 } is a finite sequence

2. { n } is the infinite sequence { 1,2,3.....n,... }

3. 1 is the infinite sequence i,!,!!,...

23
IIn 1  n

4. {x,. = (- iy' } is the infinite sequence {i, -

5. {x,, i+ (-
i)" } gives the infinite sequence 1 0, 2, 0, 2,... }

6. If x it), then {x,,

we (.- r F I does not give a sequence, for x,
becomes undefined when n is the square of an odd positive integer.

5.6. Bounded Sequence.

Given a sequence I x. 1, if there exists a number K, such that K is

greater than or equal to any member x,, of the sequence, i.e., K 2! x,,,

where x,, is an y element of {x0 }, then the sequence is said to be

bounded above, K being called the rough upper bound. Of all the rough
upper bounds, the least one is called the exact upper bound of the
sequence. If K is the exact upper bound of the sequence {,,}, then
there exists at least one member of {,}, x,, > K - e where e is a
preassigned positive number, however small.
Similarly, given a sequence { x }, if there exists a number k. which
is less than or equal to any member of the sequence, i.e., if k 2^ {x },

for all n, then the sequence {, } is said to be bounded below, k being

called the rough lower bound. Of all the rough lower bounds, the greatest
one is called the exaci lower bound of th& sequence. If k is the exact
lower bound of the sequence {x,,}, then there exists at least one
member of {x,,} ,such that x,, < k +C

If a sequence is bounded both above and below, it is called a

bounded sequence.
8-
I i4 DIFFERENTIAL CALCULUS

Examples:

1. The sequence {i + - } is bounded above, the upper bound being 2.

2. The sequence {i- } is bounded below, the lower bound being 0.

3. The sequence 12 + (- i) } is bounded, for it is bounded both

above and below, the upper and lower bounds being 25 and I respectively.
4. 'Me sequence {0,3,-2,5,-4,7,...},i.e (1+(—l)'n) is unbounded,
as it has neither upper nor lower bound.

5.7. Monotonic Sequence. -

A sequence {,, } is said to be monotonic

(a) increasing -(or, more correctly non-decreasing), if x,, 15 x,

for every n
(b) decreasing (or, non-increasing), if x,, ^! x,, 1 for all n;

(c) strictly increasing, if x,, <x *1 for all it;

(d) strictly decreasing, if x > x,, for all n.
Monotonic sequences are also called monotone sequences.
A sequence is said to be monotonic sequence, if it is monotonic
increasing or monotonic decreasing.
Examples: -

1. The sequence {x,,}, where x = -.--, is strictly increasing.

n+1 -

2 The sequence }. where x = fl_t_i, is strictly decreasing.

3. The sequence 11, 0, 1, 0, 1.0.... ) is neither increasing nor decreasing.

5.8. Limit of a Sequence.

The idea of limit forms the most outstanding concept in
Mathematical Analysis and it plays an important role in the discussion
of convergence of an infinite sequence.

Let us consider the sequence Ix,, }' where x, =

If we put n = I, 10, 100, 1000,.. . successively, the respective values
ofx, are 1.01.0010001.
SEQUENCE 115

Obviously, ac it increases, - steadily decreases, but always remains

positive. For large value of n, the difference of x.from 0 is very small and
we can make this difference less than any preassigned positive quantity,
however, small, by making n sufficiently large. For example, if we like to
make this difference less than 0000001, n should be greater than 106.
Thus, the value of x can be made as near to as we please by taking n
sufficiently large. This is expressed as
. —+0 as n -+ oo or, limx = 0.

Therefore, we hve the formal definition of limit of an infinite

sequence. -
Definition: A sequence {x,, } is said to have finite limit!, if for any
pie-assigned positive quantity E, however small, there corresponds a
positive integer N, such that I x,, - I I < C, for n > N.
This state of affairs is expressed as urn x, =

5.9. Convergent Sequence.

An infinite sequence {x,,} is said to be convergent and has the
limit!, if corresponding to any arbitrary small positive number e, we can
find a positive integer N (depending upon e ) such that
x,, -11< e, for n > N.

i.e.,I—e<x,, <l+e. when n^:N.

This is expressed by saying that 'x,, tends to the limit 1, as n tends
to infinity' and expressed as lim x,, = 1.
,7 -4

By the symbol n -, oo , it is meant that n takes up successively an

endless series of integral values which ultimately become and remain
greater than any arbitrarily assigned positive integer.

Find the limit of the sequence {.L } as n -3 00.

when -

by taking N = -- or, the integral part of -- (when it is a fraction),

C C
- 0 < C, if n > N
1
116 DIFFERENTIA L CA LCULI/S

Therefore, urn = 0. Hence, the sequence { } converges to 0.

Note: The limit of a sequence may or may not be a term of the
sequence.
For example, the convergent sequence (, - ,.......... } has the
limit 0, but no member of the sequence is equal to (I.

Again, let us consider the sequence { x }, where x,, = , when it

is odd and x, = 0, when n is even, i.e., the sequence {i, 0,.0, , ..

Obviously, the sequence is convergent and converges to he limit 0
which is equal to an infinite number of terms of the sequence.

5.40. Non-Convergent Sequences.

A sequence {} is said to diverge to +00 , if for any number -K
whatever is assigned, there corresponds a positive integer N, such that
x. > K for all n > N . This Situation is expressed by urn .t =

In this case the sequence {x} is called divergent.

Here —is no real number, it is a symbol to denote a large positivc
number greater than the greatest number one can imagine.
A sequence { X . } is said to diverge to -00 if, when any number. K
whatever is assigned, there always exists a positive integer N such that
x 4 < K for all n>N, and wewrite tim .v,,
I -4
Here, K is generally chosen a negative number, large in absoluit
value. - -
A sequence which is neither convergent nor divergent is caltecl,an
Oscillatory Sequence.
In an oscillatory sequence f x. 1, if a constant c exists such that
< c, for all n, then the sequence is said to oscillate finite/y.
otherwise, it is said to oscillate infinitely.
Note. A monotonic sequence cannot oscillate.
Examples.
1. - The sequence 1 2 , 22.32i.e., {112} diverges to +
2. The sequence —I,— 2,.— 3.......i.e.. n } diverges to -00.
SEQUENCE Ill

3. The sequence - 1, 1, - j.......i.e., {(_ i)" } oscillates finitely between

—1 and I.
4. The sequence - 1. F2, - 'J, 2— F5 ....... i.e., {(_ iy' .1 }osciflates
infinitely.

5.11. A Few Important Theorems.

Theorem 1. A convergent sequence determines its limit uniquely.

Proof: If possible, let I and 12 be two distinct limits of a convergent
sequence

Since 1 * 12,wecantake I'

—12 1 = ö ,where isanon-zeropositive
number. Now, let us choose a positive number E , such that e < 8.
Since {., ).,possesses two distinct limits 1 and 12

Ix,,-1it<e for n>N1 and .x,,—l 2 I<4e and n>N,,

where N 1 and N, depends on the given C.

Thus, for n > N = maxN 1 , N2)

l i —1 2 1 = (x,,
I —'2)— x1, - ')I
!^lx —i 2 +x,, 1 1 1

e e
< — + — = E for n > N.

Thus, 6 < c and we arrive at a contradiction. Thus, the assumption

that the sequence (x, J has two distinct un' s is not true. Hence, the
theorem.

Theorem II. Every convergent sequence is bounded.

Proof Let the sequence {,, } be bounded having a unique limit 1.

II
Then I x,, - <C for all n ^: N , N being a positive integer
depending upon r, however small.
i.e. 1-6 <x. <1+ewhen n 2! N.
Let L and M be the least and the greatest of the numbers
xM_l,l.C,I+•E.
118 DIFFERENTIAL CALCULUS

Then we have L :5 x,, < Al for all values of n.

Thus the sequence I x . } is bounded.
Note. The converse of this theorem is not always true. For example,
the sequence It + (- i)" }, i.e.. (0,2,0.2.... } is bounded but not

II, - .1,—3 , 1....

convergent.

I L.
Also, the sequence is bounded but not
12
convergent.

Theorem HI. A In000ronic increasin g sequence which is bounded above is

convergent and converges to its exact upper bound or supremum.

Proof: A monotonic increasing sequence {} is always bounded

below, for x ^t x 1 for all n. Again, since .r } is bounded above, if
exact upper bound or the supremuni of {,, } is M,

(i) x 15 M or for all n, and

(ii) for any given c (>0 ) there exists at least one member of the sequence
{ x,, }, say X N , such that X N > M -
Since the sequence {,,} is monotonic increasing, x > M -
for n ^! N. Again from x,, < M , for all n, whereby x,, < M + e for
each n.
Thus, M - c<x <M + c ,for all n -f^ N and so lim x,, = M.

Hence, the theorem is established.

Theorem W A monotonic decreasing sequence bounded below is
convergent and converges to its exact lower bound.
The proof is exactly similar to that of Theorem HI proved above.
Theorem V. A monotonic increasing sequence diverges to + o, if ills
not bounded above. -

Proof : Since the sequence {x,,} is monotonically increasing,

x,,1 ^! x,, for all n, and {x, } being not bounded above, there exists at
least one member, say, xi,,, of the sequence such that x,, > M where M
is a large positive number. The sequence being monotonic increasing,
,,,41, X.. 2,.. . are all greater than M.
 SEQUENCE 119

Therefore x, > M for all n ^ rn.

i.e., lim x =
Dl-.

Thus, the sequence { X . } diverges to +

Theorem VI. A monotonic decreasing sequence diverges to - o, if not

bounded below.
The proof is similar to the proof of Theorem V.

5.12. An Important Sequence.

The sequence {x, }, where x,, = (i + is convergent.

C. P. 2004
It will be shown that the given sequence is monotonic increasing
and bounded above.
Using Binomial expansion.

( IY' 1 n(n-1) 1 n(n—lXn--2)

X Dl, =1—I
1+ =1+u.—+
n) n 2! n2 31

nIl

i( i( IV 2
= 1 + I + —Il - -—Il - - UI - - 1+...
2! n' ) + 3! n)t n)

i( l'( 2'l ( n-1'

...+—l1--lIl--l...I1--I ... (I)
n! It fl) ) A
Replacing n by (n + 1). -
1( 1 '\ 1( i V 2 "i
X =l+1+—Il---------l+---ll--lll--I+...
2fl n+1) 3U n+1)k, 17+1)

I ( I_ I 2 1 In"l
—lI1----------l...I1--1 (2)
(n+l)U n+1) n+I) n+I)

From (1) and (2), we observe that

i) the first two terms of x and x,,, 1 are equal, each being I;
12() DIFFERENTIAL CALCULI/S

(ii) _L < I, i- > 1— and so on thus excepting

n+1 n n+1 n
the first two terms, every term of x, is greater than the corresponding
terms in x,,.
(iii) x contains (n + I) terms, while contains (n + 2) terms and
all the terms are positive.
Hence .r 41 x, for all n, i.e., the sequence {x, } is monotonic
increasing.

Next we note that x, 2^ 2 for all n, i.e., {x, } is hounded below and
2 is a lower bound.
Also, x 1+l+11l_ IIl_iil
2 1. 3! n) n

ñ (I
n)

.... i( ) 2 ( n—I
n!( n a) fl

1 1 1
<1+1+ — +—.+...+----
2! 3! a!

<1+1+_+ —+...+---
2 22 2

=1+ = 3-2' <3

2
i.e., x,, < 3, for all n.
Thus, 2< x, <3 for all n.

Hence, the sequence fr,, } is bounded.

Since the sequence {} is increasing and bounded , it is
convergent.
Note. Since the sequence is convergent, its limit exists and this limiting
value is denoted by 'e',

i.e.iin(i+ ! J = e, where 2<e<3.

This number e is a transcendental number.

SEQUENCE 121

5.13. Bernoulli's Inequality.

For every positive integer a 2, and I + p >0, (1+ p)" >1 + up
Proof The method of mathematical induction will be used to
establish the result.
We note that, when a = Z (i+ p >l+2p.
Thus the inequality holds for n = 2. Let us assume that the relation
holds for any particular value k (^ 2) of n.
(i+pYc>i+lqj

or, (i+pXi+pY >(i+p)(i+kp), since l+p>0.

or, (i+pr' >1+(l+k)p+Ai'2 >1+(k+l)p,

Thus, we see that the relation is true for n = k + 1. if it is assumed
to be true for n = k.
But, it has been proved to be true for = 2, so it is true for n = 3, and
as it is true for ,j= 3,itis true for n = 4 and soon.
Thus, the inequality holds good for any integral value of n > 2.
Note The above inequality is true for n >1, even if a is not a positive integer.
5.14. Null Sequence.
A sequence is said to be a null sequence, if Jim jr. = 0, i.e., for
any positive number e, however small, there exists a positive integer N,
such that I; 1< E for all n>N.
Exane:
Then sequence {" } is a null sequence if x < I.

If x = 0, each number of the sequence is Wand jr! - 0 as n -

When x * 0, I x us a positive proper fraction, we write

I I
Now, Ix"=
(i + h) 0l+nh
(1+ h)"> 1+ nit [Bernoulli's inequality]
0 1 part I
\ Il < - < C ifn>N, where Nis the integral Of.

Hence, the sequence {x"} is null sequence if I x < 1.

 122 DIFFERENTIAL CALCULUS

5.15. Cauchy Sequences.

A sequence is called a Cauchy Sequence, if corresponding to

any pre-assigned positive number however small, there exists a
positive integer N, such that for n ^! N,

Ix, , - x,, I <.E for all positive integral values ofp.

Example: is a Cauchy Sequence.

mi

Here.Ix,p..-x,I= _! zr! i_—!_. <!< for >N

1
fl+P n n l+ n
11 1

and taking N > ! or the integral part of

Hence, {! } is a Cauchy Sequence.

5.16. Cauchy's General Principle of Convergence.

A necessary and sufficient condition for the convergence of the
sequence { x,, } is that corresponding to any pre-assigned positive
number e, however small, there cxists a positive integer N, such that

for n 2! N. x,,,, - x, < r, for all positive integral values of p.

Condition Necessary : Let the sequence {,, } be convergent, i.e.,
the sequence has a finite limit, say 1. Then, forgiven e, however small,
there exists a positive integer N, such that

- e forall n > N.

Then it follows that

—i<c forall n ^! N and p>0

Hence, - x,, I = I x,,,, -1 + 1 - X.

<— 'X
I — l+; — I l

<tE + - e

Le.,IX,,+p_x,,I<E,forn>Nandp>0
Thus the condition is proved to be necessary.
,tQUtNCE 123

Condition Sufficient: Next let us suppose that —; I <C.

for n ^ N, p being any positive integer. Then

x
x,, - I < < x,, + C • for all positive integral values ófp.

t bounded as p-400.
.+p I is thus 

Let, L and M be the lower and upper bounds respectively. Then

L 15 a,, - E and M :5 a,, + E.

Thus, M— L:5(a,,+c)— (a,,— c)=21

Since eis arbitrarily small, this implies that M - L = 0 in the limit.
Therefore. M - e :5 x,,^,, 5 M + c.

X. }
It follows that x,, + , - M for all integral values ofp.
Thus, the sequence { is convergent and the condition is proved
to be sufficient.

5.17. Theorems on Limits of Sequences.

Let {,,} and {y} be two sequences, such that
urn x = A and urn y,, =B

Then, (i) Iirn(x,, + y,)= A + B,

(ii) uirn(x— y,)=A — B.

(iii) Iim(x,, .y )= A . B,
I, -3

(iv) lim = ,Provided B 0.

Y.
B

Proof: (i) Since urn x = A , given any c > 0, there exists a positive
integer , N1 , such that I; - A I< 1 c for all n > N

Also, since urn y,, = B, for any c > 0, there exists a positive integer
N2 , such that I y,, - B < 1 c for all n > N2.
I

If N = max{N1, N21 thenfor n > N

and
124 DIFFERENTIAL CALCULUS

Hence, I(x+y.)r_(A+B)I!^Ix,,—AI+Iy,, —BI <L forall

n> N.
Therefore, by definition urn (x,, + y,,) =A + B.

(ii) Proof of this part is similar to that of

(iii) We have.

- A B I =x ( y - B)+ B(x,, - A )l

y,, - BI+IBhIx,, - A

Since { X. } is convergent, it is bounded and there exists a positive

number M, such that {x } < M ,for all values of n.

Then, Ixy,1_ A B I<'MIy_B I+{ IB I+A } IX n — A l... (1)

where I is any positive integer.

Now, let ebe any pre-assigned positive number, however small.
Then we can find two positive integers N1 and N2, such that

I._AI<2{IB+ All for n>N l... (2)

and IY. — B I < ----, for n>N 2... (3)

If N be any positive integer greater than both N1 , N2. we get from

1), (2) and (3)

xy - AB I <- E + = C ,forall n> N.

Hence, lint x,, y,, = AB

Note. ln(l)we have taken 11 B I + A } instead of I RI in (1); otherwise,

the inequality (2) fails if B = 0.
(iv) Here,

x. - A = B.x,, - A.v = B (x,, - A) + A (B - y)

B B.y0 B.y

BIIx - A+ I A II I (1)
I II y.
SEQUENCE 125

Since, tim y = B * 0, there exists a positive integer N 1 , such

that

I B-yI<IBI,forn?Ni
ot, I—IyI<IB—yI<HBI

or, lBI<yj ... (2)

From (1) and (2)

L_ IBJIx_AI+lAIIB_yI
Y. B

2 2{IA+AJIB_yI
<—jx, -AI+
(3)
BI ' fBI2

where A is a positive number.

Let e be a positive number, however small ; then (here exists positive
numbes N,, N such that

...

1B1 2 1
and B - < { IAi. for every ii 2 N3... (5)

If N be a.positive integer, greater than each of N1 , N2. N3 then using

3), (4) and (5), we have

x ,,A 1 1
- -- <— E + = E,forevery n2N

Hence, urn x,,

- = A-
"-" y,, B

.18. Illustrative Examples.

- I
.x. 1. if xII= - + 1- +3.4 1
- + ...
1
then show that1 x
1.2 2.3 n(n+l)
.s a bounded monotonic increasing sequence.
[C.P. 1963, B.P. 1984, 1994]
126 DIFFERENTIAL CALCULUS

( i H I) (I I) (I
Here, x,, +
2) 2 3 3 4, n i)

n+l n+l
n+l
Similarly, r,+,=

n+I fl _________
>0. for all
So, ;,1-x,, =
ii (n+l)(n+2)
positive integral value of n.
Hence, {,,} is a monotonic increa.si!lg sequence.
n
Again, since - >0. x,, >0.
n+1

Also, < 1, for all positive values of n.

n+i
Thus, O<Xn<l.
Hence, the sequence (,} is bounded also.
I
Ex. 2. If =
; In--, prove that the sequence (,.} is monotone

increasing and bounded. I V .P. 20001

= 30 - I and x,, 3n + 2
= -i-
Here, .x,

7
- X, > 0, for all positive integral values
= (n + 2)(n + 3)
of n.
Hence, the sequence is monotone increasing.
3n—1 3(n+2)-7 = 37
Again, . = ,z+2
n+2 n+2

Since --- is positive for any positive integral value of n, x, <3.

Also, x,, Z!
i.e., 1 ^ .,, <3.
Hence the sequence is bounded.

Ex. 3. Prove that Urn n b" = 1, where n is a positive integer.

V I' 1998]
 SEQUENCE
127

For n>i, >i,

Let, Gn = I + h, , where h > 0,

then fl=(l+)l+.,+n(h1),,,,2+^,,fl

>
4 n (,z — since all the terms are positive.

0<!, <J2/(n-1)

urn !=0,hence jim

E.. 4. Show that the sequences given by

G)) x,, = 2 + (- i)". 2 " and (ii) X,, 1 . rnr

= — Sin -- are Convergent.
n 2
Find the limits. [8.1'. 1965]
(I) We have 2 - = 2" and 2 — x < E where e is any
positive number, however small,

if 2"<E,i.e,2" >!, i.e.,if n>.2_iPi)

e log
Thus, if we choose an integer log
N >
log J 2 - x,, < C • for all
n >N.
E,. 5. Use Cauchy criterion to show that
(i) the sequence Ix. I defined by x,,
+ (._ I ) " ' . ,j is convergent;

(ii) the sequence {x,,} defined ; I + . 1 + 1 + I + 1 is

divergent;
(i) Choosing m> n,

1 1 I I
xIn —x
n = ----__. +--...---
n+l n+2 n+3 m
— I (1 lHl I
n+l ln+2 n+3n+t n T1

<e, if n+ I>if n > ! i

C
 DIFFERENTIA L CA LCULUS
128

- I]. then
If now N be so chosen that Nis equal to the integral part of

X,,, -X <e.whenever n > N.

Hence, the sequence converges by Cauchy's Criterion.

(ii) If we choose m.=2n,

IX. - = I X2. -
1 •1 1
=-+----+".-
n+l ,i+2 2n
I I
> fl.. =
2n 2

Thus x,,, - x, I is not less than any pre-assigned positive number.

Hence the sequence does not cnverge by Cauchy's Criterion. Since
the sequence is monotonic increasing and does not converge, the
sequence diverges to +
Ex. 6. (i) Show that the sequence

,
tends to a definite finite limit and find the limit. [ C.P. 1960]

(ii) Show that the Sequence .J, converges

to 2.
(i) If the flh term be ; ,then x.
(I)

Again. ;

... (2)
or, x,—x120
From (1) and (2)

x,,2+l - X .2 x., - -'

or, (x,.+1 + x,j(.s 1 - x,,) x,, - x,,_1

if
This implies that x,,, 1 > x,, - if x, > x 1 and A 14 1 < .,, ,

< Xo_I
SEQUENCE 129

So, the sequence is monotonic increasing or decreasing according

as a2 > a or a2 <a'.
Here, obviously x2 > x1.
Thus, the sequence is monotonic increasing.
Again from (1) x 2 - x, -2<0, x <x1

or, (;- 2) (; + i) < 0

Therefore, lies between -I and 2, i.e.. -1 < • x < 2.
x,,
Thus the sequence is bounded.
Since the sequence is monotonic increasing and bounded, it must
tend to a definite limit, say!.
Then, ; = I, x, 11 =1 as n -

So from (1) 12_I -2=0,

1=2-1.
Since, all the terms of the sequence are positive, I = 2.

(ii) Here. x1 = . J, x2 = T272 =

x.= x. +.=

2; and x.2 = 2;_i

.2_I
x.2+, - .A,, - -

or, + -;)= 2(x, -,,_1)

Therefore, if ; > x. -I , then ; >

Since, x2 = 12T21 X1 = h, 52> SI

Therefore, x3 > x2 , x4 >

So. { ; } is monotonic increasing.

Again, 4+1 =2;.and >x,

4 <2x,, or, x,,(x,, -2)<O

Thus,. {,} is bounded, and so it is convergent.

Let, lim x, =I
II -4
.9
130 0/Fl- ERENTIAL cALculi/S

Jim 4. = Iim2x,
,I-
or, 12 = 21 or, I(/ —2) = 0
since x, >0, 1 *O, . -. 1 =2.
Hence, {x,,} converges to 2.
5.19 Miscellaneous Worked Out Examples
Ex. 1. Prove that the sequence {x}, where x,,=(-1)" is not
convergent. I C. P /I J

Solution : Let us assume that x,,} is convergent.

Then x, -4 1 (a finite quantity ).

So, for a finite number E = -1-, (say) there exists a number N, such

th.t I(1Y_hI<>forn>N.
When n is even and n > N, Il-1I<

ie <t<—

and when n isoddandn>N,

Thus our assumption leads to a contradiction. Hence, {x,,} is not

convergent.
f4n + 3 )
Ex. 2. (i) Prove that the sequence is bounded and monotone
increasing. F C. P.

(ii) If x,, = find the least integer m, such that -!I<_

for n>m. [C.P20001
Solution: (i)
n+2
if 4n+3:^4n+8, i.e., if 3 < 8, which is true irrepecIive of the values
of
SEQUENCE 131

Hence the given sequence is bounded.

4n+3
Let,
x- = 2
The sequence {x,, } will be monotone increasings,
if .x,,:5x+,
4n+3 4(n+l)+3 4n-7
i.e., if — ^
n+2 (n+I)+2 n+3

i.e., if (4n+3Xn+3)5(n+2X4n+7)

i.e.,if4n2+15n+9:54n2+15n+l4
i.e., if 9:5 14, which is always twe
Hence {x,, } is montone increasing.

ii 1 . I2n+5 II 1
(ii) x–j<j gives for n>m

26 1 I 3(6n-1I)
<iotm
3(6n-1l)1000
> ' 26
26000
i.e., 6n< --+I1

Le., n> 1446

-j, for n> 1. and
Hence the least integer is 1446.
m

3n+1
Ex. 3. Show that if x,, = -, then the sequence {x,,} is strictly
increasing. Is the sequence &tergent ? Justify your answer. Also find
its limit. [C. P 1993, 94 1
Solution The sequence {x,, } is monotone increasing if x <x,,+1
3n+1 3(n+L)+1
i.e., if - <
n+2 (n+1)+2
i.e., if (3n+1)(n+3)<(n+2)(3n+4)
i.e., if 3n2+lOn+3<3n2+I0n+8.
which is evidently true for all n>O.
Hence the sequence is strictly increasing.

Again. .. :t:^ 3. if 3n+13n+6,

 132 DIFFERENTIAL CALCULUS

which is true for all n > 0.

Hence the sequence {x,j is bounded.
Thus {x.1 is monotone increasing and at the same time bounded
above, hence it is  convergent.
1
3n+1 3
Also, lun - = Jim = - 3.
n+2 a-'" I

Hence, the limit of the sequence is 3.

Ex. 4. (i) Examine the convergence of the sequences
(a) 22,42,62,...
[C.P 19891
.
(/i') 1 .2
t  c. i' 1991 i
(ii) Use Cauchy's criterion to prove that {x}converges, when,
11 1
+ IC.P. 19801

(üz) Evaluate: Jim {4 +

+ .......+!!-} [C. P 1.985]

Solution : (i) (a) Here, .*,, =(2n) 2 and x,, as  n-0c
Hence the sequence is not convergent.
(b) Here, x, = 2 —oc as n —oc
Hence, the sequence is divergent.
1 I 1 1
We have, -=
n! 2.3.45....n 22.22.....2 2''
IXI flI + (fl ^ 2)j + *jJ . if m >.i

)
=-i-x
2
2
•1 I
<—.2=---.
2-2"
 SEQU&V CE . 133

Thus I x,,,—x,--O as n—)°.

So, for any € > 0, there exists a positive integer N such that

for all n>N.

i.e.,x,,1 —x, <E for aim > Nand m > n.
I

Hence, by Cauchy's Principle, {x} is convergent.

2 33 fl31 113
i) urn (n
--+----+•••+--=lim
n +i n
I— n J

,.-
1)2
1 ( 1
=—•lim 11+—I
4—k. n 4

EL 5. (z) Show that the sequence {x, }, where x,, is bounded.

C. P 1997, 2006, 2008 1

(ii) Show that the sequence Ix.), where x,, = is decreasing

and bounded.
Correct or justify : Ix, } is convergent. [ C. P 1995

(a:) If x,, =-1) ' and Y. !(xJ ±x 2 +x 3 + i-x,,) (n= 1 2 3 )

prove that the sequence [Y . 1 converges although the sequence IX .)
does not do so.. IC.P. 19851
Solution: (i) Here, x = <3
n+1
if, 3n+1:53n+3, i.e., if 1<3, which is true for all n.
Hence the given sequence is bounded above.

(ii) Here, X. = ---, so X,,. . 1 =

5+6n 6n + 11 -
-1 1 •-6
<0, for n>0
x,,41 — x,, < 0, i.e., x,, 1. 1 <x, so that x,, 1 <x, is a decreasing
sequence.
134 DIFFERENTIA L CA LCULUS

Again, Jim x urn __!_ =

"- 5+6n
Hence, {x} is a bounded sequence.
Since the sequence is decreasing and bounded, it must be
convergent.
(iii) See Ex.1., to prove that {x,}, where ; =(—i)" is divergent.
• x1+x2+x3+...+x,,00
y., — - -
n n
• . xi +x2 +x 4 +...+x I
if n is even and y = = --, if n is odd.
n n

As ,; urn y,, = iini

,I-.
1_! = o
jI-+ I?)
Hence the sequence {y,} is convergent and it converges to 0.

EXAMPLES-V
1. Define the terms 'bounds', 'supi-emum', 'infimum' and 'point of
accumulation' in connectioh with a set.
2. (a) Show by suitable examples that the supremum ofa set, if it exists.
may or may not be a member of the set.
(b) Give illustrations, where the infimum of a set is and is not a member
of the set.
3. (i) Define the terms 'limits', 'bounds' and 'cluster point' as applied to
a sequence.

(ii) Show that the sequence is bounded. V P. 19971

4. Prove that a sequence can have at most one limit.

5. What do you mean by a monotonic sequence?
State whether a monotonic sequence tends or does not tend to a limit
under any circumstances. [C. P. 1980, 94, 20001
6. When does a sequence converge? Prove that a convergent sequence
is always bounded. Comment with reasons on the validity of its
converse proposition. [C. P 1981, 98. B. P 19971
7. i) Define a monotone Sequence.
 35
SEQUENCE

(ü) Prove that the sequence I x . }, where x,, = (i + .. )" is

(a) monotone increasing, (b) bounded and (c) converges to a
limit e,where 2< e <3. [C. P. 1985, 88, 98]

8. (i) Show that the sequence { x, }, defined by,

I 1
X. =—+---+....+—
" n+1 n+2 2n
is monotonic increasing and also show that it is bounded.
(ii) Show that the sequence { x,, }, is monotone decreasing, where

(i .+ C. H. 19551

9. (i) - Discuss the behaviour of the sequence 1x" }, where x is any real
number.
(ii) Prove that the sequence r' } is convergent, if I r I < 1.
[B. P. 19981

(iii) Show that the sequence {i } converges to 0.

10. (i) When a sequence is said to be a Cauchy Sequence?
(ii) State Cauchy's general principle of convergence of a sequence
and apply it to show that the sequence {x}, where
11 1
X,, =1 + + +.... + is convergent.
V P. 2002; C. P. 2006]
11. (1) Give an example of a sequence which is neither monotone
increasing nor monotone decreasing. [ V P 20011

12. (I) Show that urn = 1, (x> 1). [B. P1996.1997]

(ii) Prove that lim;.x" = 0, for I x I < I.

(in) Show that urn -Ir" =0.
a-.- n!

(iv) Prove that ijm L_ =0, for Ix <I

(v) Prove that urn ii" = 1. B. p 19961

13. What is a Cauchy Sequence? Show that a constant sequence is a

Cauchy Sequence.
 136 DIFFERENTIAL CALCULUS

14. Define Cauchy Sequence and show that is one such

I. n + l J
sequence. (VP. 19971
15. Determine the bounds of the following sequences, if there be any:

(u)
(iii) _2_22,_23,_24,...
3 5 5 7 7 9 9 Ii
(iv) - , - , - , - , -, - , - , -,...
22334455
16. Prove that the sequence L. } is monotonic increasing and bounded;
where
3n+1 -
0) X, = I B. P 1985, V. P 1999]
1 1 1 1
(n) X.
' 1.3 3.5 5.7 (—i)(+i)
j I
(iii) x =-+---+-+...+
i 1
-
" n+1 n+2 n+3 2n
4n+3
(iv) X. = [C.P.10]
n+2
ni-i
17. (i) If x show that the sequence is strictly
2n + ,
monotonic decreasing and hence prove that it is convergent.
CF. 1996

(ii) Show that the sequence is monotonic increasing,

In1J
bounded above and converges to 3. . [C.R 1991, 20001
{4i +
(iii) Show that the sequence 3.} is bounded and monotonic
n+2
increasing. [C.P. 19891
18. (i) Show that the sequence { x }, where

X. =(_1)' _T-'and x,, =!+(_1)".2

oscillate finitely.

(ii) Show that the sequence { x,, }, where x,, = (- I )" (n + 1) oscillates
infinitely.
SQUFJV CE 137

19. (1) Show that the sequnce J('!LJ.') J is convergent. I VP 19951

1"' ) I
2_5
(ii) Show that the sequence {x,}, where X,, is not
6n
convergent. [VP.1996]
3n+1
(w) Show that the sequence (x j, where x = — i- is strictly
increasing. Is the sequence convergent ? Justify your answer.
[C. P. 1993]
20. Find an integer N, such that

(i) _oj<o.000l for every n>N.

Ct')
I j_ 0 k 0.1fo t 1.
[C.P.1988]

2n+5 I I <_!_
for n>N.
6n-11 3 1O
21. Show that the sequence (xe) is convergent where

2+ (- i ) ) ;
[ C.? 19921

2n-1
(ii) ,,=
n
1
(In) x, =7 sin

(iv) x, —sin ---;

.n 2
2+1
(v) x, =
2
4n+1
tC.P,1980]

(vi) .x,, =i+(..!)". [,C.P.1968]

22. Show that the sequence { x, } is divergent, where

(ii) X = (ni) x = log !

n+I ln
 DIFFERENTIAL CALCULUS

23. Consider the behaviour of the following sequences with respect to

convergence or divergence
(i)

I +I

1n2
(iv) [ C.P 1981]
2n 2 +i

24. Show that the sequence . J _L. , where p > 0, is a null sequence.
I" i _______________

25. (a) Show that the sequence , 1+ , 7+ +

converges to the' positive root of the equation
- 1-7 0.
(b) Show that the sequence F5, /i 5, + ,...

converges to

(ii) If x,, = .JK + x . where x1 and Kare positive, show that the

sequence { x,, } is increasing or decreasing according as x 1 is

less or greater than the positive root of t 2 - I - K = 0 and has,
in either case, this root as limit.

ANSWERS
15. (i.) lower bound —i . upper bound 1;
(ii) no upper bound, lower bound 0;
(iii) upper bound —2, no lower bound
(iv) upper bound 5 , lower bound --
19. (iii) yes, convergent.
20. (I) N = 1000; (ii) N = 99; (iii) 1446.
 Immirm SERIES
Mi
6.1. Infinite Series.
Let us consider an infinite sequence of numbers. U 1 , u 21 u3 , ...
or {u, J . The series derived from the terms of this sequence, viz.,
ui+u2+u3+...+u,,+...

u,, or, simply j u . is called an Infinite Series.

=1
Now, let us form the sequence of the successive partial sums {s,, }
of the above infinite series, where s 1 = u 1 , s 2 = +
S3 = U 1 + U 2 + U 3 ......5,, = U 1 + U 2 + U 3 + ... + U,,

6.2. Convergence of Infinite Series.

If the sequence Is.} of the partial sums is convergent, the series
Eu,, is said to be convergent.
If the limit of the sequence {,, } be a finite number S. then we say that
the series Eu,, converges to Sand that S is the sum of the series. It may
be noted that S is not a sum in the normal sense of the term, actually it is the
limit of a sum. -

E
Definition: An infinite series u,, is said to converge to S. if corresponding
to an arbitrary positive number r , however small it may be, there exists a
positive integer N. (depending upon e ), such that
S. - S I < c , whenever n > N.

If we denote (,,, 1 + u,,, 2 +... ) by R,,, obviously, for the

convergenceofu,,, IR,,I<E, for n>N.
If s,, does not tend to a definite finite limit, but , s,, —), ± or
- o , the series u,, is properly divergent and diverges to +
or to - respectively.
If. however, s,, does not tend to a definite finite limit, or to + or to
— , but oscillates finitely or infinitely, then Yu,, is said to oscillatory, or
improperly divergent.
Divergent or oscillatory se,is are generally called non-convergent.
140 DIFFERENTIAL CALCULUS

Examp:
(i) Show that I + + + + .-.. is convergent, and find its sum.

3
Here,s = =-I3--I-- -,asn--3oo,
l- 2l 3"1) 2
So, the series is convergent and its sum is •j-.

(d) Show that the series I ++ +... is convergent; find its

SUM.
I " l 1)
r(r+l) (— r r+1

( i 1 i\ It i 1 i iut I
2) (1 3) L3 4) 1n-1 n) n n+I

= 1 — ---- — I , as a —, .
n+1
So, the series is c onvergeni and its sum is 1.
(iii) Examine the convergence oftheseries: t + 2 + 3 + - + n +

Here, 5,, = a (a + —* -,as n-4--

The series is divergeizi, as S diverges to + .

(iv) Show that the series — I)l = 11 + I - 1 + . oscillates

finitely. I
Since, S =1-1+1 -1+... -
S. = 1, when n is an odd integer.
= 0, when a is an even integer.
Hence S. does not tend to a definite limit, as n — . The series
oscillates finitely.

(v) Theseriest-2+3-4+5-6+-•• = (-l)'.r oscillates

infinitely, for I

S,,=(n+l), when nisodd,

=--n, whenniseven
and it does not approach a finite limit as n —
 INF!NfTESFJ?1ES 141

63. Convergence of Two Important Series.

L Geometric Series. The infinite geometric series
a+ar+ar 2 +ar 3 +... (a>O)is
(,) Convergent when I r I <I,
(ii) Divergent when r2: 1 and
(iii) Oscillates finitely, if r = —1, oscillates infinitely if r < —I.
Proof. Obviously, the a th partial sum S, is given by

2
S,,=a+ar+ar +..+ar
= a .•., (ri)

(I) Now, if r < I, r" -4 0 as n -4

S. a
urnS
l—r

and the series converges to -f---.

I—r
(ii) If, on the other hand, r> I, r" -9 as n -9 .
urnS.

and the series diverges to

(iii) If r = I,
S,,=a+a+a+...+ana

and no, i.e., S -) as n -

So, in this case also, the series diverges properly.
(iv) When r <-1, (r") oscillates infinitely and the series is
oscillatory or improperly divergent..
(v) lf,again,r = —1.
S, =a—a+a—a+...= a, when nisodd
= Q. when a is even
and the series oscillates finitely between 0 and a.
From the above discussions we arrive at the conclusion that the
geometric series converges if I r I < I and does not converge if I r 1.
 142 DIFFERENTIAL CALCULUS

IL The p-series.
I I. I I
The infiniteseries + + + ... + +
(I) is convergent if p > I, (ii) is divergent if p :5 I

Proof. (i) We suppose that p> I.

Let us consider the partial sums of order. 2" -1, i.e., S,, S, S,. S13,

(1 I (I 1
S 2'-1 1+I_+—I+I—+—+...+
(2" 3") 4" 5" 7"

(I 1 . I
+1 -+-+ . . . +- 1+...
8" 9" 15")
(i
+1
(2" -I)"

(1 l'\ (I I I
I,2" 2") I,4" 4" 4"
(I i 1'
+I-+--.-+...+-
•,8" 8P 8P

I iI
2 -1 r + (2"-'
11 1 ,_i I
2" 4" 8" (2'''

1 I I 1
=l+—+ +
2" (2p_l)2 (2_i)"-t
(2P-'

1-rn I
= , wherer= 2T.
1-
r

S 2, 1 < constant.

2P 1

All the terms of the series being positive, the partial sums are monotonic
increasing.
 INFINITE SERIES 143

And for any positive integer m, there exists another positive integer ft,
such that 2k_I >m,
1
S. <S2 ,_ 1 < = Constant.
1

Thus, the partial sums are bounded.

The sequence of partial sums, being monotonic increasing and
bounded, must converge.
Hence, the series is convergent if p > I.
(ii) Next, we suppose that p :5 1.
Since p I, n 1' n

11 I•l'i 1'I
Thcn.S =1+—+ —+—I+l—+..,+_i^...
2 2" (3P 4') t5" 8')
1 .. 1

i (i r (1 1
2 'L3 4) (5 6 8).

(2 -2" 2"
1 1 1 I I
=l+-+-+--+...+--=1+-n
2 2 2 2 2
Therefore, given any number G >0, however large S > G whenever
n> 2(G-1).
Further, the partial sums are strictly monotonic increasing and not
bounded.
Hence, the given series is divergent, when p ;-> I.

6.4. Conditions of Convergence.

L Cauchy's General Principle
The necessary and sufficient condition for the convergence of an
infinite series u, is that corresponding to any arbitrarily chosen
positive number e, however small, a positive integer N can be found that
forall n 2t N,
144 DIFFERENTIAL CALCULUS

Is,,+_s,,I<Iu,,.i+u,,.2+...+u,+I<c.
for every integral value of p.

Let, S. = it2 + U1 + . ..+ it,

Then Eu, is convergent, if and only if, { S. } is convergent. Now by
Cauchy's general principle of convergence of a sequence. IS. } is
convergent, if and only if,
- s, 1< e for n2! N and for every positive integral value ofp.
Thus, we have it,, is convergent, if and only if

I -U,,,1 + U,, 2 + + U,,.,. <C,

when is , N for every positive integral value of p.

II. Pringsheim's Theorem.
If the terms of the series u,, of positive terms steadily decrease,
then it is necessary for its convergence that urn n . u,, = 0.
Let the series I u, be convergent. Then for a given e, however snlall,
there can be found a positive integer N, such that, for all values of n _> N
we have

u.+1 +U,,•1 +U, 3 +"+UN <E

Since the terms of the series steadily decrease, each of the terms U,,,
.... . it,, is greater than or equal to it,,
U,,.,. 2

Hence, (n - N)u,, < E, when n N .

Since, tim u, = 0, we can choose i > N ,such that

Nit, <e, when n^p..
Thus, nu,, <c, when n2^U.
This gives limflu,, =0.
Note This condition is necessary but not sufficient. If we consider the
= —i----, tim nu,, = 0, but u ,, is divergent.
series u,, .where U.
it kg n
INFINITE SERIES 145

6.5. Tests of Convergence of Series of Non-negative terms

In order to ascertain the convergence of infinite series it is not always
convenient to find the limiting value of S. as ii - 00. So, a number of
n'rethods and rules have been developed for testing the convergence of
infinite series. In this section important rules and methods will be discussed.
I. A necessary condition for convergence of 7, u,, is that km u,, = 0.
Proof: Since the terms of u,, are all positive, it follows from Cauchy's
general principle of Convergence (art. 6.4).
S. -s_ < e,i.e., 1 u 01.< E
Hence urn u,, = 0 is a necessary condition for convergence of U,,
Note : This condition is not sufficient, for example the series -! is
divergent, although urn u,, = 0.

IL If a series I u,, of positive and decreasing terms be convergent, then

lim(nu,,)= 0.
This theorem as Pringsheim 's theorem, has already been discussed in
art. 6.4.
UL Comparislon Test
Statement:
Let u,, and v,, be two series with non-negative terms, and
suppose that there exists an integer N such that u, :5 v,, for n > N
Then (1) Z u,, converges if Y, v; converges and (ii) Y, v,, diverges
if Y, a,, divçrges.

Proof: (1) We denotç the n" partial sums of ru,, and Y, v, is. S. S'
respectively.
Since V. converges, given e (>0) we can find N, (depending
on ), such that for all p 2: 1, and a> N1.

Let N 2 = max(N,iv 1 ). Then for a> N 2 we have

10-
i's,,, I, -s,, I < js P - s H
146 DIFFERENTIAL CALCULUS

and since this inequality holds for all positive integral values of p. it follows
then ) u, is convergent
(A) If n> N,wehave
S - S S. - SN

or, S^:S,-SN+S

If now, n -, oo, S -9 - ,since S. -3 00.

Hence the theorem.
Another form of comparison test
If 'u,, and iv,, be two series of positive terms and if

0 < Jim < 00 , then either both of them converge or both diverge.
n-. V0

IV D'Alembert's Ratio Test

Statement:

If > 0, and if = p, then

(i)u 0 converges if p <I,

(ii) u diverges if P> I If however,, p = ithis testis inconclusive.

Proof: (i) If p < 1, for any given e (< I - p), we can find a positive
integer N (depending upon C ) such that for n ^: N

p+c. or. u,, 1 <(p+e )u,,.

In particular.
U N+I <(p+C)uN
c )2
UN+2 <(p + C )u, j < (p + UN

Since 0 < (p + e )< 1, the series ( p + e )" U N converges,

being a geometric series with common ratio <l. and hence i  i 

N*I
u,, or u0

is convergent by the comparison test.

INFINITE SERIES 147

(ii) If p> l,foranygiven e p - 1);we can find a positive integer

N, such that for n 2! N

or, u. 1 >(p—E)u.
U.
Thus, as in (i), we have for m 2^ 1,

UN+,,.>(PE)MN
By comparing with the geometric series with common ratio
(p—a)>!,
we can conclude that the series u,, is divergent.
V. Cauchy's Root Test.
Statement:

If u,, >0 and if iim(u, 1" = p, then

(i) Ufl converges if p <1,

(ii) diverges if p>!,

When p = 1, the test is inconclusive.
Proof: (i) If p < l. given any e (<i -vp),

we can find a positive integer N, depending upon a, such thati N,

z
Comparing with the geometric series with common ratio (p + a ),
where 0 < (p + a ) < 1, we conclude that the I u, is convergent.
(ii) Next, let us suppose that p> 1.
Given any a (< p —1),we can find an infinity ofn,
say N1 , N 2 , N3, such that for these values of n, (a,. )" > p -
or, u,, >(p—c).

Since (p - )> 0, u, cannot tend to zero, so that u,, is divergent.

 148 DIFFERENTIA L CA LCULUS

VL Raabe's Test
Statement:

Let u, be a series of positive terms satisfying urn - =

then
(i) Eu,, converges if p>l,
(ii) E u,, diverges if p< 1,
This test gives no information regarding convergence, if p = I.
'IlL Logarithmic Test
Statement:
The series Eu,, of positive terms is convergent or divergent according as

Jim nlog---1>lor<1
"- 1 U-1 J
This test fails, if the limit is 1.
VIIL Gauss's Test
Statement:

If for a series u,, of positive terms _f !L_ be expressed in powers of

- U,,41
1/11, so that
U /1 (1
U,,41 fl2)'

then u,, is convergent if p > I ,and divergent if u!5 I.

Note, The notation 0 (i/n 2 ) denotes such a function j (n ) that for

every n 2t no (a definite positive integer), If (n )J < k -4, where k is a

finite quantity independent of n -

6.6. Mixed Series: Absolute and Conditional Convergence.

The series I u,, which contains both positive and negative terms is
said to be absolutely convergent, if the series E I u,,I be convergent.

I II
For example, the series I - 2 + - -_ +... is absolutely
22 2
convergent.
INFINITE SERIES 149

If the series I u,, be convergent, but the series be divergent,

then the series Y u, is said to be non-absolutely convergent.

The series I - + ! -! +... is non-absolutely convergent.

234
A convergent series which remains unaffected by rearrangement of its
terms is said to be unconditionally convergent, while the series which is
affected by rearrangement Of its terms is called conditionally convergent.

- -!- + -4 - _- + ... is an example of unconditionally convergent

series, and
- -!- + - -I 1-... is a conditionally convergent series.
234
6.7. Alternating Series.
A series in which the signs of consecutive terms are alternatively
positive and negative is called an alternating series.. -

THEOREM I. Leihnitz c tes:for Alternating Series.

An alternating series u - u 2 + u 3 - u 4 +... is convergent if {u,, }
be a sequence of positive terms decreasing monotonically to zero.

THEOREM H. An infinite series in which the terms are alternately positive

and negative is convergent if each term be numerically less than the
preceeding term and Jim u,, = 0.

• Note: I. When we say 'u,, is absolutely convergent we are to test the

convergence of Y I u.1 and not that of

2. To determine the absolute convergence of series we are to use
the test deve)oped for positive series.
3. If the terms of an absolutely convergent series be rearranged,
the series remains absolutely convergent and its sum also remains unaltered.

68. Power Series.

A series of the form

..a,, (x-x0 ) =a0 +a 1 (x.— x0 )+a 2 (x — xo)2 +

 150 DIFFERENTIAL CALCULUS

or, ia,, x = a,+ a 1 x + a 2X2 +. . + a,,x" +

where the coefficient cz, a 1 , a 2 ,..., a,,,... are independent of x. is called

a power series in x.
The simplest and the most important power series is the geometric
series: l+x+ x 2 +
From the discussions of art. 6.3, it is obvious that the above geometric
series is convergent (and also absolutely convergent) for I < 1, i.e.,
Jx
—1 < x < l. diverges to +00 for x > 1, and oscillatory when x <_i.
The above example and other similar situations reveal that as the
variable x, in a power series, changes, the terms also change and the series
may change from a convergent to a non-convergent one. This leads to the
Conclusion : a power series is convergent either for all values oh, or for a
certain range of values of x, or for no value of x except zero. It is, therefore,
important to ascertain the value or values of the variable x for which a
power series is convergent. That is why comes the idea of interval of
Convergence'.

6.9. Interval of Convergence.

Definition: The interval of Convergence of a power series in x is the
collection of values of x in an interval such that the series converges for
every value of x in this interval, but does not converge for values of x
outside the interval.
If a power series a,, .r 0 cnverges absolutely for all values of x,
inside the interval ofconvergence —r < x < r, and diverges for j x I > r,
then r is called the radius of convergence of the power series.

6.10. Determination of Interval of Convergence.

D'Alembert's Ratio test and Cauchy's Root test will be useful for the
determination of interval of convergence of Power Series.
For the power series >a,, x".

(i) If lim .-* A , a finite quantity other than zero,

"-" a,
(6) if limlIa,I }l/ = A where A # 1,

then the interval of convergence of the series is (- r. r). where r =

INFINITE SERIES

Note : The interval of convergence as determined by the above tests is

x = ± r.
open. The series may be or may not be convergent at the end points
one should first find
To determine the complete interval of convergence
absolutely convergent and then test
the values of x for which the series is
the end-points.
6.11.Properties of Power Series.
Here we state (without proof) certain properties of infinite Power Series,
which are often used in obtaining new series.
1. Within its interval convergence, a power series represents a conhinOUS
sum function and has not more than one power series representation in a
given interval.
H. Two power series converging in the same interval of convergence can
be added and subtracted tern, by term; thus if
f(x)=>ax" and

f (x) ± ra,,x' ± bx' (a, ± b,,)x"

III Two power series converging in the same interval of convergence can
be multiplied term by term.
Thus, if 1(x) = a,,x" and 0 (x) bx"

f(x).O(x)

=(a0b+a1b,,_1.+ . . . +abo)x". -
This product series is absolutely convergent in the same common
interval of convergence.
IV. The quotient of two power series ax" and Ib,,x° (' * 0) is

another power series , provided x remains within a sufficiently small

interval in which the denominator does not vanish and both numerator and
denominator are convergent series.
a0+a1x+ax2+
2 ...
Thus, = C0 + c1 x + c2 x 2 +... + c x"
b0+b1x+b2x2+...+bX

Since. Y,
ax" = !b,,x.Zc,,x' , a 0 = b0c0 , a 1 = c0 b 1 +
etc. whence co, c 1 , c 2 ,... can be calculated.
c1b0,

V Limits, term by term, are permissible ip case of power series within its
interval of convergence.
 152 DIFFERENTIAL CALCULUS

Thus, if 1(x) = u,,x" lint f(x)= urn

VL A power series can be differentiated or integrated, term by term, over

any closed interval lying entirely within its interval of convergence, as
many times as one wishes.
If f(x)=Ia. x",

f'(x1 )=nax' . j(x 1 )=n(n1)ajf2,etc.

and f'f(x) d fat"di-= _1(rz1
provided x and x are both interior points of the interval ofconvergence.
VII If two power series I a0 x" and I b,,x"
both converge in the same
interval and both represent the same function f(x), then they are identical,
i.e., a,, = b, for all values of n.
VIII If y = f(x)= sax" and t() Eb,f

then F{f(x)}=b0 ±bi{a,,xJ+b2{a,,x"}+...

= b0+a1x+a2x2 + (a +a x+a

= Co + ci x + c 2 x 2 +

for every value of x for which a,,x" converges and has a sum less than
the radius of convergence of bay"

6.12. Illustrative Examples. I

EL 1. Prove that the series
1 1
+ - + - +... is convergent.
1.3 3.5 5.7
[V P.19971
Here, U,,
(2n — I)(2n + i)
I I1 1 1
5 =u
,,2l2n—I2n+l

2t .31 23 51 2 I. 2n-1 2,z+l

Ii 1] 1
as n . .
= 2n+1
Hence, the series is convergent and its sum is 1.
 NFINITESERIES 15:

Ex. 2. Examine the convergence of the series:

1 1 1 1
1! 2! 3! (n—I)!

1. 1
Heir, s =i+i.r!+-L.i-
2 2.3 2.3.4 2.3.4...(n —.i)

(n>)
2 2.3 3.4 (n-2).(n-1)

1 (i ñ (I i ( 1 1
2 t2 3) t3 4) t,n-2 n—i

1(1 I
=1+1+—+i---------
2 (2 n — i

= 3— <3.
n—i
{ S. } is monotonic increasing sequence and bounded above.
Hence, the series is convergent.

Ex. 3. Show that the series + -- + 2. - ... + _!_ is convergent.

2 22 2 2
[V P. 20011
n n+ I
Here, u' = - and z, •7* =
2
u 11+ n+l ii n+1 I( ñ I
U, 2-1 2 11 2n n) 2
21L
Hence, the series is convergent.
EL 4. Test the convergence of the following series:

ci) (_L + -J_) .ii) _L_ [C. P 19981

%F
n .jj log n-

_____ + .ç ji-:- +
(i) Hem, u
=T

Let us introduce a comparison series v, = AT where

V11 =
154 DIFFERENTIAL CALCULUS

+ .
• urn = urn = urn += 2
vn fl) -

But, Y, v,, is divergent, for in the p-series, here p = <1.

,=I fl
Hence, the series U,, is also divergent.
(ii) Iogn<ri for all n>1.
1 1
- > - for all it > 2.
logn n
Let us compare the given series with the divergent series
1 1 1 ... (1)
- + + - + :.. + + .,.
234 n
Each term of the given series exceeds the corresponding term of the
divergent series. Hence, the given series is also divergent.
Ex. S. Examine the convergence of:
12 22 32 42

2 22 2 2 2
[C.P 1993, 2003, 2008 B. P. 1'Q81

n 2(n+1)2
Here, U,, = and U,,1
= 2I

u, (n+1)2 n 2i( 1t 21
-= 1
+--=--I 1+—I as n -
U,, 2- 2 2. nJ 2
Hence, the given series is convergent.
Ex. 6. Examine the convergence of:
(3\ 3 14\4
I '2"
-+1--i +!-t +1--I +... [C.P.1986,'92,20071
3 5) 7) t9)
-
Here, (- fl
U
2n + 1

(,,)V" =
2n+1

lirn(u = urn __!T = <1.

o-.2+-- 2
Hence, by Cauchy's Root test, the given series is convergent.
 INFINITE SERIES 155

Ex. 7. Examine the convergence of

3 4 5 (ni-i)
2+-+-.-+— +...+ 3 +... [C.P.1997,20031
8 27 64 n
- n-+ 1 - 1 +
Here, U,, - —i--- -
n n

Let us introduce another series , where v,,

Evidently, v, is convergent, for here p = 2 > I.
U (i+ 1 I i
Now, lim--=1imI-----'-xa2I=IjmIl+_J1
"-.-v,, "-. n ) "--1 nJ

Since, v,, is convergent, the given series u,, is also convergent.

Ex. & Examine whether the series
3x 24x 3 (ni-1)x"
8 27
is convergent or divergent (x > o).
Since x > 0, each term of the given series u,, is positive and

!!!±i..= (,i+2) n3 ._L

U,, (n+i)3 (ni- I) x"

n 3 (n+2) 1+1
(j)4 •x -,x asn -+oo
(i+
Therefore, by D'Alembert's Ratio test, the series is convergent, if x < 1,
and divergent if x>1.
For x = 1 ,this test fails.
n+i
When x=1,u,,= -.--1--.

Let us take another series v,, , where v = -i-. -

v,,being a 'p-series' with. p = 2 > 1, is known to be convergent.
tim - = urn - = tim (I 1+-l\ l,isafinitequantity.
n+l
n-_ ,- fl 1t'I fl)

Therefore, Zu. is also convergent.,.

Hence, the given series is convergent if x !^ 1, and divergent iix > I.
156 DIFFERENTIAL CALCULUS

EL9. Examine whether the series is convergent or divergent:

1 x 3 1.3 x 1.3.5 x7
[C.P.1969]
2 3 2.4 5 2.4.6 7
Denoting the given series by U 0 + u, + U2 +..., we have

1.3.5...(2n-1) x2"
U.=
2.4.6...2n 2n+1
1.3.5...(2n-1)(2n+1) x2'"3
u *I =
2.4.6 ... 2n(2n+2) 2n+3

u,, - (2n + 2)(2n + 3) I

(2,i + I)(2n + i) X 2
I
-- ,as n — 3 --

Thus the given series is convergent, if -- > I , i.e., if x

- x2
2 < 1, i.e., if

—1<x<1 and it is divergent if -- < 1,i.e.,ifx>1 or x< --I

x2
u,, (2n + 2)(2n. + 3)
When x = ± 1, -- =
(2n + 1)2

im - 1]} = -= (+ 02n(6n+5) 3 >1

•-1 Un urn =—
2

So, by Raabe's test, the series is convergent for x = ± 1. Hence the

given series is convergent, if —1 :5 x 15 I and is divergent if x > I or
X < — 1.

EL 10. Examine the convergence of

=1 fl
[C.P 1996, VP 2000, B.P 2001 1

Here, u. =- and u,,1

= (,*1)!

U. - n' (n + 1)! - n'

—
- (n+If, (i+i

Iim_±<1, 2<e<3..
e
fl,
is divergent.
INFINITE SERIFS 157

Ex. 11. Examine the convergence of

7
(i) 1+—+—+---+... [B.P.1998]
2 3! 4!
2 2 . 4 222.42.62
2
(ii) 2x + x6 + +... > o)
3.4 3.4.5.6 3.4.5.6.7.8
[C. P. 1965,N. B. P. 1981,VP.19881

(2n-1) (2n+1)
(i) Here,u U
n! - (n+1)

(n+1)(2n-1)
,,
(2iz + i)

11m--> 1, is convergent.

(ii) Denoting the first term by uo,

22.42.62... (
2.)2
U. 22
3.45 ... (2n+1)(2n+2)

22.42.62...(2n)2(2n+2)2
and U,,1
= 3.4.5...(2n+3)(2n+4)

U,, (2n+3)(2n-i.4) I
- (2n+2)2

lim----= tim
"- .- (i + y x2 x2

Therefore, the series is convergent

.4 > 1, i.e., if x 2 < 1, i.e., if
-1< x <1 and it is divergent 1,i.e.,if x>lor x
if-4<
The ratio test fails when x = ± 1 and we apply Raabe's test.
( n 1(2n +3)(2n + 4) 1 6n2+ 8n
I-1I
We have, n U. 1
(2n+2)2 (20+2)2
J
• 6 3
Iirp n( u,, = Jim =->l.
ff
(2 + 2) 2 2

Hence, the given set ies is convergent, if x = ± 1.

158 DIFFERENTIA L CA LCULUS

Thus, the given series is convergent, if —1 15 x < I and divergent, if

x> 1 or x < —1.
Ex. 11 Determine the region of convergence of the series

(i) 1+4 [C.P.1998j

x-3 1 (x-3)2I (x-3)3

(ii) -+-' 32 +– 33
3 2 3
(i) We have, by the Ratio test.

U. = n+1
x

urn -- = 0< l, for all values ofx.

z__ U4
Hence, the series is convergent for every value of .x and the region of

convergence is – < x <

(ii) Here, hrn

I UU.ttL=hrn( 1)3' n. 3#

=
3 -..'n+1 3

Thus, the series is tonvergent ifI X3 < i, i.e., - I < 1,

3 . 3
i.e,O<x<6.
The series is also convergent for x – 3 = —3, i.e., x = () , but not for
x-3=3,i.e.,x=6.
Hence, the interval of convergence 0 :5 x < 6.

6.12 Miscellaneous Worked Out Examples

EL 1. Vxamine the series for convergence.
12 2 2 32 42
I C. P. 1993
2 22 2 2
– (n!)2
-i.
Solution Here, u,, = —
2 and
an u4.1—
24+1

1 1
So.2
U4 2 fl22('
n-)
 INFINITE SERIES 159

U.
tim = 1 <1
U. 2

Hence the series Yu, where u. = !_ is convergent.

EL 2. Test the convergence of the series:

[C. P 1994, 2003]

Solution: Here, u and U,, 2'

(n+i)

hrn—=hm
'-•" U,, "' x" .(n +

= urn 2 lxi
11+-
" n
=jxl.
Hence u,, converges when I x < I and diverges when I x> I-
When jr = 1, the ratio test fails -
1 1 1
At x=1, )u,, —+—+--+ .......
12 22
is a 'p' series with p 2>1.
which is
(i I 1 1
At x=-1,
\.1 2 3 .4

since the series i+_! + 4+ ..... being a p-series with p = 2>1, is

convergent, by Leibnitz's Test, Y u, is also convergent.
Hence, the given series is convergent when N:5 1 and divergent when
lxi>!.
Ex. 3. Apply Raab'es Test to examine the convergence of

1 1.3 1.35
(1) 1+ + C. P1990, 93, 94, 2004
23 243 2.4.6.7"" I
160 DIFFERENTIAL CALCULUS

(ii) !., where M. = I C. P 1989 1

7.10.13 ... (3 n + 4)

Solution (1) Denoting the given series by

+ U1 +M2 +
1.3.5 ..... (2n—l) i
we have U.
2.4.6......2n 2n+1

1.3.5.....(2 n+1)
and "246.
2n(2n+2)2n+3

(2n+2)(2n+3)
so that
(2n+l)2

Iu 1 . n(6n+5) 3
lim ni--I= urn
2(2,1+1) 2
Hence the series is convergent.

•0 3.6.9.....3n
(ii) Here, u. = 7.10.13.....(3n ;4)

- 3.6.9 ...... .3n(3n+3)

- 7.1013........

U,, 3n+7

,,+i 3n+3

!!.__ 1 1 = urn .—fL-

By Raabe's Test, uJ
rn
,,..3n+33>l
",,•I J'
Hence the series is convergent.
EL 4. Use Root-Test to examine the convergence of the following series

1, 22 3
(0 - 1 (- I C. P 1986, '92. 2007]
5) +() +...

(iii 'u,,, where u,, {(nI ) F C. P 1990, 20()0]

n )j
INFINITE SERiES 161

Sn
n
Solution: (I) Here, U, = ( • so that

n
(u )n
2n+1 2+_

and lim(un)=!<1.
- 2
Hence the series is convergent by Cauchy's Root-test.
(i) Obviously, here

(u,,)
= (i)+1
_(n +1- 1 \" I
(i+!}["
- U U •f l
J

lim (u
p){( i) } Ji<1

Hence the given series is convergent.

Ex. 5. Examine the convergence of the following scries
2 3 4
X x
(0 x+ +—+----+•.. (x>G) B. P 1997]
23-4

1 x31-3 x-' 1•3•5 x7

(ii) x+—•—+---+. .—+ ... . (C. P 19941
2 3 2 . 4 .5 2•4-6 7

a + a(a + 1) + a(a +
(iii) —
+ 2) id
b b(b+I) b(b+1Xb+2)

where a, b are non-negative integers and b * 0. [ C. P. 1997 1

1+cz (l+aX2+a)(1+a)(2+aX3+a)
('
l+3 (1+3)(2+) (1+3)(2+3)(3+) +

- C. P.
where a.
11•
 162 DItHi'ENTIAL CALCULUS

-
Solution : (i) Here, ii,, - and U, 1 =
n n+l

u. n+1 I . ( I '\ .1 1
=lim---.—=.lim I
,,- -- ,,-, n X •-( X X

Hence the series is convergent if x<I, and divergent if x>1

If x = 1, the Ratio-Test is inconclusive.

when x=1, the series is

234
and it is divergent, by the p-test ', p = I here.
Hence the series is convergent for Ox< I and divergent for x ^: I
Denoting the given series by n0 + 14 1 +11 2 + 11 3 -i
(ii)

1'35''''(2n—I) x2"
we have, u,
2.4.6'.. (2n+1)

1 -3.5 .... (2n -I)(2n + I) x21'3

and, u1
= 2,i(2iz 1-2) (2n+ 3)

urn (2n+2)(2n + 3) I
urn --- =
(2n 1- 1)(2n + t) x 2 - A2

Thus the given series is convergent if

i.e, if x2 <1, i.e. if < x<l and it is divergelil

if -<1, i.e., if
if.>1 or, x<-1
x
when x = ±1, the above test fails and we proceed to Raabe's Test

u,, 1 n(6n+5) - 3
urn nL --i= tim
"-- u,,.1 J -(2n+I)2 - 2>1

So, the series is convergent for x=±1.

Hence, the given series is convergent if —1 :5 x 5 I and is divergent if
1, or, x<-1.

4
 163
INFINITE SERIES

u b+n
(lit) Here, = -
a+n

and urn U— =1
n—_

So ii Alembert's Ratio Test fails. We proceed to Raabe's Test.

1 .
Jim n--ll' Jim ni—
Ib+n— I
I;
u,+i J
• n(b— a)
=hm =b— a
n— a+n
Hence the series is convergent if (b—a)>!, divergent if (b— a)<1, and
when b =a , the series becomes
1 + 1 +1 + ..... . which is obviously divergent.
(1+a)(2+a).'(n+a)
(iv) Here, U,
=

(1+ctX 2+a) ... (n+cx)('i+l+(L)

and u,,1
= ( 1+1)(2+)-••(n+X n+ t+)

1+n+.
that --=
l+n+cx

U
lini .-----=I -
Us,.,
We proceed to Raabe's Test
I
lim n i --1=hm
u ni
1t3-a 1
t+t J - 11+n+a

n (13— a) • ____
tim =hm =3—a.
.s—.' 11 +1+U

Hence the given series if convergent if (3 - (x) >1 and

divergent if (—a)<1.
 164 DIFFERENTIAL CALM UJS

EXAM11VI
1. (I) When does an infinite series is said to converge?
(u) Give an example each of (a) a convergent series, (b) a divergent.
series and (c) an oscillatory series.
(iii) Show that the addition or removal of a finite number of terms at the
beginning of an infinite series will not affect the convergence or
divergence of the series.
(iv) Prove that multiplication of each term of an infinite series by a
constant term, different from zero, does not affect the convergence
or divergence of the series.

2. (t) Show that the p-series ._L converges only when p > i.

Also-show that the harmonic series X is divergent.

lB. 1' 1296.1999]

(ii) Show that the series __L

3 + -__ + _L._ + ... converges.

23(C.? 1987]
3. (i) State and prove 'comparison test' for convergence or divergence
of a series of positive terms.

II
(ii) Use comparison test to prove the convergence of

2 32 42 [C.P.1988]

4. (I) State and prove D'Alembert's ratio test for convergence or

divergence of a series of positive terms.
[C.P. 1982, '87;B.R 19991

2 22 2
(d) Hence, show that 1 + - + - + - + ... converges.
1! 2! 3!
[C.. 1992]
5. (1) State and establish Cauchy's Root test for convergence or
divergence of an infinite series. [Ci'. 19801
(ii) Use Cauchy's Root test to examine the convergence of
3 (4\4
I (2 2
+ L:j () '• ICP.1986. '92]
INFINITE SERIES . . 165.

6. State Raabe's test for convergence or divergence of a series of positive

terms. . . IC.P. 19941
7. For the following series, compute the partial sum S ,and then obtain
their sum:

(i)

1
(••) [C.R 1999, 20061
+ 3)(n + 4)

(2n-1)(2n+1)

(iv)
n (n + I)(n + 2)
8. Use comparison test to examine the convergence or divergence of the
series:
1 . I
(i) 1 + 3 4 + + (2,, -
[C.P 19851

3 4 n+l
(ii) 2+—+--t-...+-----+... •[C.P.19971

1 1 1
[C.P.1998]
log log log
1 1 1
(iv) + +...+ +....
21og2 31og3 n log n

1CR 19941
(v) ,(2n-1)(2n+l) .

+ ,1 i_) . CP.19981

9. Apply D'Alembert's ratio test to examine the convergence of:

(i) IC.P.20061

1 222 32
(ii) —+—+—+...+— +...
2 22 2 2'
4 42 42 42
—+—+---+...+-- -'-••.
1! 2! 3!
166 DIFFERENTIAL CALCULUS

10. Use Cauchy's root test to investigate the convergence of the series
f \2 3
1^2 r2+21 ( 3 +2 (n+2
(i) —+I---I +— +...+l—I +...
2.1 t, 2.2) 2.3 2n )
I 1 1 ____
(ii)
2 3 2 43
++—+•••+
(n+I)"
+...

2 222 2 2'
(in) 33 44
22

(iv)
[n
2] ^ ("')
11. Examine the following series for convergence or divergence

(i) C.Pii]
1! 2! 3! n!
1.2 2.3 3.4 4.5
(ii) —+—+—+---+...
3 5 7 9
I' 22 32 42
(iii) —---+ + ——...
+ +
2 2- 2 2
1+2 1+2+3 1+2+3+4
(iv) - + + C.P. 1991, 2(")4 I
22 33 41
+...

(V)!++++... IC.!'.J9)31
2345
(vi)!+4+._+ ... +-_+ C. P. I qw)

(vu) I [C.P 19861

+ 2.3.4 + 345+.;.

(viii)jj+j— +22 l^2

••' [C.P.I9)IJ
1 1.3 1.3.5
(Ix) —+—+
2 2.4 2.4.6
1.2 3.4 5.6
(x) ,
3_42
+
2 212 42
(xi) I+_+L_+_+...+_ EC.PJ9'/l
2! 3! 4!
 167
INFINITE SERIES

12. Test the convergence or divergence of the of the series where

(I) u,,

(ii) u =siJ;

(in) U,, =
n +1
n
(iv) u,, = (x > 0); [C.R 19871

n!
(v) U,,
it

(vi) u,, =---x

It
2
(vii) U,, =
(n +1)
=
(viii)u
" (i + i)"'
2.4.6...(2u +2)
(IX) u =
" 1.3.5...(2n-1)

=
(x) U.

13. (i) Show that the series

1 1 .1
+ + converges.
i+1 2 1 + 2 2 1 + 32
(ii) Show that the series
1 1 I
1+ + + + convergent.
2.12 + 1 2.22 + 1 2.32 + 1
[B.P.1984. 94j

14. Apply Cauchy's general principle of convergence to test the following

series

(i)
23 n
 168 DIFFERENTIAL CALCULUS

(ii) i—+_!+...+(_iy'-

15. Test the convergence of - by Cauchy's criterion.

,,=on.
16. Show that the following series are divergent:
l.3.5 ... (2n-1)
(I)

1.3.5...(2n–l) 4n-i-3
2.4.6...2n 2n-f2'

(iii) —;
,,,

(iv) JTi}.
17. Prove that the hypergeometric series
a.13 a(a+1)f3(fl+l)
+ ly + 1.2.y +i)
(where a, /3, y are real and, none of them is zero or negative integer)
converges if y > (a + 0) and diverges if y :5 (a + p) . IC.? 1989]
18. Discuss the convergence or divergence of the hypergeometric series

1 a/3 a(a+1)13(j3+1) 2
X+ X
1.2.y(y+I)
(where are real and none of them is zero or negative integer) for
x I < I. I x 1>1 and for x1.
–1 .
19. Show that the series converges for J x < 1.
+1
[C.? 1999]
20. Find the interval of convergence of the following series:
x x2 x3
(1)
[C.? 19871

+LL_+.
2 3

22
32 [C.P.19941
INFINiTE SERIES 169

1 x 3 1.3 x 1. 3.5 x7
(iii) x+-"—+--+ •—+... (x>O)
2 3 2.4 5 2.4.6 7

2 x4
(vi)I x
2! 4!
x' x5
(vii)x+—+--+...

ANSWERS

(a) ; (b) (c) (- i)".

5. (d) Convergent.
7. (i)2; (ii); (iii); ( iv)f.
8. (I) Convergent; (ii) Convergent; (iii) Divergent;
(iv) Divergent; (v) Convergent; (vi) Divergent.
9. (i) Convergent; (ii)Convergent; (iii) Convergent.
10. (i) Convergent; (ii) Convergent;
(iii) Convergent; (iv) Convergent;
11. (i) Conver2ent; (ii) Divergent; (iii) Convergent;
(iv) Divergent; (v) Diverged; (vi) Convergent;
(vii) Convergent; (Viii) Convergent; (ix) Divergent;
(x) Convergent; (xi) Convergent.
12. (i) Convergent; (ii) Divergent; (iii) Convergent;
(iv) Convergent; (v) Convergent;
(vi) Convergent when x :5 1, Divergent when x> I;
(vii) Convergent; (viii) Divergent; (ix) Divergent; (x) Convergent.
14. (i) Divergent; (ii) Convergent.
20.(i)-<<+..; (ii)xI !5I ; (iii)-I5x:5I;
(iv)o<x:52; (v)-I<x:5I; (vi)_<x<+oo;
(vii) - < x < +.
 DiFnRENTiAinciN
IL1I
7.1. Increment.
The increment of a variable in changing from one value to another is
the difference obtained by subtracting the first value from the second. An
increment of .v is denoted by A x (read as delta x) or h. Evidently, increment
may be positive or negative according as the variable in changing increases
or decreases.
If, in v = f (x ), the independent variable x takes an increment A x
(or /i), then Ay (or k) denotes the corresponding increment of y, i.e., of
f (x), and we have
y+Ay=f(x+Ax), i.e., Ay=f(x+Ax)—f(x)

or, y+k =f(x+h ), i.e., k =f(x+Ji )—f(x)

Illustration: Let y =
Suppose. x increases from 2 to 21, i.e., Ax =01; -
then y increases from 4 to 4-41, i.e., tsy =0-41.
Suppose. x decreases from 2 to 1 .9, i.e., Ax =-0l
then y decreases from 4 to 3-61, i.e., Ay —039.
Increments are always reckoned from the arbitrarily fixed initial value
of the independent variable x.
If decreases as x increases, or the reverse, then Ax and Ay will have
opposite signs.
)rom a fixed initial value 2 of x, if x increases successively to
2. 1,2.01, 2 . 001, etc. then although the corresponding increments
A x (=0.l.0.0l,0.001,..) and Ay(=O .41,0.401,0. 004001,. ..)aregetting
A y'
smaller, their ratio, i.e., being 4. 1,4 .01,4 .001.....is approaching a
definite number 4, thus, illustrating the fact that the ratio can be brought as
near to 4 as we please by making Ax approach zero. Thus, the ratio of the
A
increments has a definite Finite ljmit4as Ax - 0, and, consequently,
Ay - 0.
 IA TION LI'
L)If' I'IK iIV I

7.2. Differential Coefficient (or Derivative).

Let y = f (x ) be a finite and single-valued function defined in any
interval of x and assume x to have any particular value in the interval. Let
Ax (or h)be the increment OfX, and let Ay = f(x+Ax )- f(x) be

the corresponding increment of y. If the ratio A of these increments tends

to a definite finite limit as Ax tends to zero, then this limit is called the
differential coefficient (or derivative ) of f (x) (or y) for the particular

value ofx, and is denoted bY f'(x), --{f(x )},, D{f(x )}.

dx dx

Thus, symbolically, the differential coefficient of y = f(x) with respect

to x (for any particular value of x) is

f'(x) or, Li!-=

- dx -.o Ax A-*O Ax

, dy f(x+h)-f(x)
I, k'J or, - Lt Ii
provided this hm,texists.
dx h-O
If, as Ax 0, A y/Ax - + oo or-
-4 oo. then also we say that the
derivative exists, and -8 + or - 0O
Note 1. The process of finding the differential coefficient is called
differentiation, and we are said to differentiate f (x)
and sometimes to
differentiate f(x) with respect lox, to emphasise that x is the independent
variable.

Note2 !ty-stands here for the symbol .!_(vi. a limiting process, and
dr dx
hence must not be regarded as a fraction dv divided by dx. although, for

convenience of printing, It may sometimes be written as

dx
of
Note 3. The differential coefficient of f(x), for any particular value a

x, is often denoted by f'(a) . Thus, from definition,

f(a+h)-j ,
provided this limit exists .
f,(.)= It

Note4. If f'(a) is finite, f(x) must be Continuous at x = a

f(a+l;)-f(a)
)'(a)zo Ii
Ii
 172
DIFFERENTIAL CALCULUS

We can write f(a+h)_f(a)

ii {f(a+h)—f(a)} = I f(a+h)_f(a)h}
h ^0 h-0 h

Ij f(a+h)_f(a)Uh
h-.O l.-.0

= f'(a)xO

= 0, since f'(a) is finite.

Li f(a+hf(a)
h-,.O
from the definition of continuity, it follows that f (x) is continuous atx = a.
Hence, for the differential coefficient off W
to exist finiieh' for any
value of x, the function f(x) must be continuous at the point.

The converse, however, is not always true, i.e., if a function be

continuous at any point, it is not necessarily true that a finite derivative of
the function for that value of x should exist. For illustration see Ex. 4, § 7.5.
Again, a function f(), though discontinuous at a point, may have an
infinite derivative at a point. [See Ex. 7(u), Examples Vl1(A).]
Note5 The right-hand limit jj f(x+h )—f(x)
for any particular
It
value of x, when it exists, is called the right-hand derivative of f(x) at that
point and is denoted by Rf'(x). Similarly, the left-hand limit
L j f(x+h)—f(x)
or, Lt f(X_h)_f(x)whenjtexjstsis
h ,,-.o+o —h
called the left-hand derivative of 1(x) at x, denoted by Lf'(x). when
these two derivatives both exist and are equal, it is then only that the derivative
Of f(x) exists at x. When, however, the left-hand and right-hand derivatives
of f(x) at x are unequal, or one or both are non-existent then f(x) is said to
have no proper derivative at x.
Thus, though f'( x) may not exist at a point, one or both of the right-
hand and left-hand derivatives may exist (the two being unequal in the
latter case).
For illustration, see § 7.5. Lx. 4.
 DIFFERENTIATION 171

7.3. Differential coefficients in some standard cases.

(i) Differential coefficient of x'.
Let f(x)=x'.

hi)" -.x"
Then from definition, f'(x)=L.,
Now, writing X for .r + h, so that h = X -x, and noting that when h
X-,x,we get

f'(x)= Li =nx" , for all rational values of a.

X^ X-x
[See * 3.9(v)

Thus, .!_(" )= rix fl ,for all rational values of n.

dx
Otherwise:

f ' (x)= Lx (x+h)'-x" = Li x"'. (I+lz/x)"-I

i,-.o h h-.0 h/x
[supposing x 0J
,,_i (1+z)-I
=. fputtingz=/i/x

[See §3.9(vi)J
The result can also be derived for any rational value of n [ s (I J from
the well-known inequality.'
nX"'(X-x) X .— X . nx "'(X -x )

fupper sign if n>li,r >01

(and lower if O<n<l
-x
Whence nX"(X-x) X"
X .-x
Now putting X =x+h, and letting h -0, we get
(x+h )" — x " ,,_
Li =nx
h [See §3.8(v)I,
h-30

since both extremes tend to the same limit ax" -

When n is a positive integer, the result can also be proved as/kllows:

See any text book on Higher Algebra ( e.g., See § 10, Chapter XIV. Barnard
& Child).
174 DIFFERENTIAL cALCULUS

(x+/i)"–x'
f '(x)=IJ
h-.O h

= {"
Lj

(By Binomial Theorem)

=Lt
h-O 17

= ILK'
W hen n is not a posilire intege?; for an alternative proof, see Ex. 1,
§ 7.13, See also Ex. 2, § 7.13 for the case when it ny real value, not
necessarily rational.

d dfr\ 1 d(1'\ n
Cor. —(x)=1,—yx) ---,
dx dx 2'& dx'.x) x

Note. It is to be noted that in the above formula we tacitly assume those

values of x as do not niakex or.t' 'meaningless; e.g., if a Le afraction of
even denominator, zero and negative rabies of x are exc!ude' and if n - 1
be negative, zero value fr v is excluded.
Following the definition it may be seen in particular, that it 1(x) =
then f' (0)=Q, when a> I. f ' (0)= I. when n= 1, and f'(0) is non-
existent if n<l.

(ii) Differential coefficient of e'

Let f (x) = e Then from definition,
–e =
f '(x)L J
i,-.o h I, *0 II

= , Lt0 (€"' h = I See § 3.9(iv),)

since

Thus,
dx
(iii) Differential coefficient of a'

Let f(x)=a'.
DIFFERENTIATION - - 175

x-i-h x h
a -a a -l
Then f'IX)= Li -
h-+O h h
I,
a
Now, 1 =Ls .loga
h -,O h h-+O h log a
h
al
= .Joga [where h'=hloga]
11-94) h'
a -1
= toga, since Li
l?-40 h
I See § 3.9(i')
f '(x )= aX log a

Thus,
dx
(iv) Differential coefficient of log x.
Let f'(x)=logx
log (x + h) - log x
Then f'(x) Li
I?

x+h
= Li - . log
I-.() I? x

I x ( h
= L i - . -log 1+-
o x h x

LI ! log (i+ z [where := }

= . I See § 3.9 (iii) j

Thus,
dx x

cor. Proceeding exactly as above it can be easily shown that

—logx) loge.
dx x
176 DIFFERENTIAL CALCULUS

(v) Differential coefficient of sin x.

Let f(x)=sinx.
)=L4 sin(xi-h ) .-sin x
Then f'(x

• 2sinlhcos(x+Ih)
2 2
=Li-
h

2 ,
= .cos x+-'-h)'=cosx
h-.O2

because as h - 0 • cos x being a continuous function of x,

- cos(xi-h)--cosx;
also, by § 3.9(i),
sin
Ll -j'

d—
Thus, (sin x)= cosx
dr
(vi) Differential coefficient of cos X.

Let f(x) =cos x

f'(x) Cos (x+h)- cos x

h
-2sin 12 h sin (x+1h)
2
_j2
h
. sin1h
2
= Ii -sin x+J-h).
2

=-sinx [as in (v)]

Thus, —(cos x)=- sin x.

dx
Note. It should be noted that in finding the above differential coefficients
of sin x and cos x, we tacitly assume that x is in radian measure, because
we make use of the limit sin h/(--h ) = 1 as Ii -4 0, which is true when
It is in radian. Hence, the above results require modification when x is
given in any other measure,
 DIFFERENTIATION 177

(vii) Differential coefficient of tan x.

Let f(x)= tan x
tan (x + h)- tan x
Then Li
ft
I sin (x+h) sin x
= Li
h -0lcos(x+h) cosxj
= Li J
sin (x+!I–x) j
h-,O L11cs(x+h)cos
1.sinh I
= Li
h-.O t1coS(x+h)cOSX }

= cs2
[x ^ --(2n +i)lr ]
sin h
Li —=1, and LI cos(x+h)=cosx
k-.O ft

Thus, -_(tanx)=sec2x. [x*4(2n+I)nJ

(viii) Exactly in a similar way, we can get

-!.(cot x)=–cos ec 2x.. [x * nit]

dx
(x) Differential coefficient of sec x.
d cc
_
—secx)
( Li
d ft

=Ll!I_'
h-.O h cos (x + h) cos x
__L

*-.o 1
I Cos x– Cos (x+h)1
hcos(x+h)co ix
Li 2 sin -hsin(r+4h)
=
h-.o hcos(xi-h)cosx

sin
= Li sin (x+!h)
2 cos(xi-/,)cosr

12-
178 DIFFERENTIAL CALCULUS

Then (secx) = 1-sin x = tan x see x,

dx cos2x

Li sin(x^h)=
sin x
l,-.0 h-O

and Li cos(x+ h)= cosx,

1- 0

Thus, --(secx)=Secxtanx. [x;kJ-(2n+1)n1

(x) Proceeding exactly in a similar way, we get

- (cosec x )= -cosec x cot x. # nit I

dx
Note. For an alternative method of differentiating tan x • cot x, sec x and
coscc x front knowledge of the derivatives of sin x and cos x, see § 74,
Theorem V.

7.4. Fundamental Theorems on Differentiation.

n the following theorems we assume that 0 (x ) and m,, (x )arc
continuous, and 0 ' ( x ) and V ' ( x ) exist.
Theorem I. The differential coefficient of a constant is zero.

i.e.. —(c)=O, where c is a constant

dx
Let f( x ) = c for every value of x.

Then f'(x) = Ii ( + h) -f (x)

It
=12 C c
=1A =O.
h-.O It ,.-o It

Theorem II. The dif ferential coefficient of the product of a constant and
a function is the product of the constant and the d if ferential coefficient of
the function,

i.e., ---{cq,(x )}=cp(x), where eisa constant. -

dx dx
cØ(x+h)cØ(x)
For,---{C(X)}= Li
dx ,,-o It

=c.Lt
h-0 Ii
 DIFFERENTIATION 179

Theorem III. The differential coefficient of the sum or difference of two

functions is the sum or difference of their derivatives,

i.e.
dx
Let f(x)=Ø(x)+(x)

Then f(x+h)=0(x+h)+i(x+/:)

Now f'(x) Li f(x+h)-f(x)

h

= Li
h

=Ll h h

w(-, +h
= Li Li
h + h-.O h
=
Similarly, 1ff (x) = 0(9- ,i (4 thn f ' x) = 0 ' (4- i,ti ' (x)

Note. The above result can be easily generalized to the case of thestn,
or difference of any finite number offunciIons.

Illustration:
If f(x)=e -4sinx+x 2 +5,then f'(x)e-' -4 cos x+2x.
Theorem IV. The differential coefficient of (lie product of two functions
= first function x derivative of the second
+ second function x derivative ofthefirsi.

i.e., --{,(x )x(x )}=,(x ).'(x )+ W (x ).,'(x).

dx
Let f(x) = O(x) x(x)

Then f(x+h)=Ø(x+h) x

Li f(x+h)-f(x)
f'(9
It

• •.•
= ,
h
180 DIFFERENTIAL CALCULUS

Lt 4(x + h)qi(x + h) - _(__+ h)1(x) +$(x + h)W(x) -

h h

= Li
h-0 I h Iz

= (x ).,'(x )+ i (x ).'(x),
by the limit theorems, and the definitions of 4,( x ) and 0'(x), noting also
that Li •(x + h) = Ox), since Ø( x) is continuous for 0' ( x) to exist.
h-.O -

Note. This result, by repeated application, may also be easily generalized

for the product of a finite number of functions in the form

di
,1(x).p2()ip3(x)., (x)
dx
=p'1 (x).{ 2(x), 3(x)...}+ç2(x).{tp1(x)4P3(x) ... }
+ .....

Illustration :lf f(x)=esinx, then f(x)= ex.cosx+sinx.e

If f(x)=x 3 tanxlogx, then
f'(x.)=3x 2 .tanx logx+sec 2 x.x 3 iogx+-Lx tanx
= x2(3tanxiogx+xsec2xlogx+tanx)

Theorem V. The differential coefficient of the quotient of two functions

(Djff. Coeff. ofnum)xdenom — (DtO Coeff ofdenom)xnum
Square qf denoin

{,(x)}2

provided (x) #O

Let f Or) = 4,(X)

(x), Li f(x+h)—f(x.)
Then f '
h
0(x) 1
= Li !1) _
h_.0h 141 ( x+h ) 4,(x)
 DIFFERENTIATION 181

i(x+h)W(x)-VI(x+h)Ø(x)
Li
h-.oh ji(x+h)W(x)

= Lj W(x){#(x+h)-4(x)}-(x){JiI(x+h)-1iF(x)}
hyi(x+h)141(x)

= Li
h-.OW(x+h)I)/(x) h h J
).'(x )-(x ).i'(x )},
{w(x)} {w( x
=
2

by the limit theorems, and the definitions of 0 ' ( x ) and tv ' (x),and noting
also that Li ii(x+h )= w(x) since (x) is continuous for i#"(x)
l,-40
to exist.
(1)1fSIflX •cosx-2xsinx
Illustration: then f'(r)=

Cos x
(2)If f ( x ) = cot r = -,
sin
(-sinx).sinx-cosx.cosx I
then f(x)=
sInx
2 sInx
=- cosec 2x.

(3)If f ( )= cosec x = ilflX

O.sinx-cosx.1 cosx
then cosec XX
sin2 x sin2 x
7.5. fi istrative Examples.
Ex. 1. Vind,from the first principle, the denvativeof ,[ (x>O)

Let f (x )=

f (x ) = Li by definition I
h

(x+h)-x -
- Li - Li
 182 DIFFERENTIAL CALCULUS

Ex. 2. Find, from the first principles, the differential coefficient of tan' X.

Let tanx=yand tan(x+h)=y+k.

Then, as h— O, k— O Also, x=tany, x+h=tan(y+k).

h=(x+h)— x= tan (y+k)—tany.

d tan'(x+h)—tanx
— tan -1 x = Li
cfr i-.o
k
Li
k tan (), +k)— tan y

= L i --. Cos (y+k )Cos y

k sink
2 I I
'
Sec2 5= 1 + tan2 y I + x -'
Note. In a similar way, we can work out the derivatives of other inverse
circular functions from first principle. These have, however, been worked
Out by a different method in § 78.
Ex. 3. Find, front the differential coefficient of log cos x.
Let us put cosx=u. cos (x+/i)=u+k
k = cos (x + cos x and so, when h —*0, k *O

d , ., log Cos (x+h)— log Cos x

• — log Cos x)= Li
dx h-.O

Li log (u+k )— logu k

-
k h

- Li log(l+k/u) ! k.
k-.O k/u u
As k -* 0, k/u — *0 .. limit of I"factor= I. [See § 3.9 (iii) I

k Cos (x+h)— cosx sin (Jr +h).sinh

• gain, h h-.O h

as k *O, j. e., Ii .—, 0. k/u -4-Sm x Also. u = cos x.

d —sinx
—(logcosx)= =—tanx.
Lx cosx
DIFFERENTIATION -I

Note. Differentiation from 'first principle' or 'definition' means that we are

to find out the derivative without assuming any of the rules of differentiation,
or the derivative of any standard function, but we are permitted to use
fundamental rules of limiting operations ( § 3.8 ) and the standard limit
results ( § 3.9).
Ex. 4. A function is defined in the following way:
f( x )=I x j, i.e.. f(x)=x, 0, or, -x, according as x>, =,

or, <0 show that f'( 0 ) does not exist. I V P. 40001

f(o+/z)-f(0)
Li Li
It i.-o h

f(h)
Now. = Li Li =1.
,,-o+O Ii I,-.o+O h

Li Li
•h hO-O- h
inc&the right-hand derivative is not equal to the left-hand derivative.
the derivative at x = 0 does not exist.
Ex. . A function is defined in the following way:

f(x)xsin .! for x*, f(0)=0.

Show that 1(0) does not exist.

f (0+h)-f (0)= L i hsin(l/h)

Li
It /-.o It

U sin — , which does not exist, [See §3.11, Ex. 4]

I_.o /2

f '(o) does not exist.

Note. In both the Examples 4 and 5, f(x) is continuous at x = 0 (See § 4.6,

Ex. land Ex. 2) but f(x) does not possess derivative at x = 0.

Ex. 6. if f(x)= x
2
sin!) when x* 0, and f(0)=0, find f'(Q).
 184 DIFFERENTIAL CALCULUS

f (O+h)f (0)
f••() Li -!-(h2Sifl!_o
h *4ohI¼ h

=LJ (,sif l!_=.

h )

when At is not exactly zero, sin is finite, not exceeding I numerically. I

Ex. 7. Find, from firs: principles, the derivative of x' (x> 0).

Let f(x)=x =e°.

e)b0. ) xlqgx
f'(x)= Li
h-,O h
e '' )I.g(+h ) - - Iog. -
= Li e'1°•
h-,O h.

=e' L i
h-.O z
where z=(x+h)log(x+h)-xlogx
and hence z -+ O,as h-90,

f'(x)=x . Li Li != . Li
z-.O Z h-.oh h-,oh :-.O Z

Now b-=Li x{ log(x+h)-logx} +h log (x+h)

A -0 A l h-.O h

= Li -xlog(1+-
h'
Li log (x+h)
h-.oh X 1 h-.O

= Li ! log (1+k)+ log x,rl+logx.

h-.O k

where kbeing h/x-)0, as h -*O.

f'(x)=xr(l+ log x).
DIFFERENTIATION I8

EXAMPLES -VII (A)

Find, fromflrsl principles, the derivatives of (Ex. I -5):

1. (j)X 3 +2x. (ii)x4 +6. (iv)l/x (x * 0)

(iv) i/ .I; (x > 0). (v) J. (vi) 2+a2
(vii) x+.1x2
e " . (iii) 2
2. (i) e . (ii) e '/x.

3. (i) log 10 x. [CP.1941] (ii) x log x.

(iii) log sin (x/a ). C. P.19301 (iv) log sec x.

4. (i) a sin (x/a ). [C. P.1937) (ii) sin 2 x.

(iii) sin x 2 . (iv) sin -1 X. (v)

(vi) (sin 4/x. (vii) x 2 tan X.

5. (i) e at x = 0. (ii) log cos .r at x = 0.

6. (i) f(x)=x2 Cos (1/x)for x*O;f(0)=0.

Find '(o).
(ii)f(x)=X for 0:5 .x:5 f(x)=1_x for-<x:5I.

Does i'(+ ) exist ?

7. (i)f(x)=3+2X for —<x<-0,
=3-2x for 0<x<.
Show that f ( x ) is continuous at x = o but does not exist.
C. P. 1943 1
(ii) f(x)0 when O^x<4 f )=i, f(x)=2

when I <x :5 1
- Prove that although f(x) is discontinuous at x = -, f'(-) exists
and its value is infinite.
& f(x)=1 for x-c0,
=1+ sin x for 0!5x<-it,
=2+(x--ir for
show that f'(x) exists at .x = - jr but does not exist at x = 0.
 186
DIFFERENTIA L CALCULUS

9. f(x)=5x-4 for 0<x:5I,

=4x 2 -3x for I<x<2,
=3x+4 for x^!2.
Discuss the continuity of f ( x ) for x I and 2, and the existence
of f' ( x ) for these values. -
10. (I) f(x)=x for 0<x<l,
for l<xi^2,
for x>2.
Is 1(x) Continuous at x = I and 2? Does
f'( x) exist for these
values ?

(ii) (x)4ç (b 2 _2),

for 0!5x!^a

=b 2 —x 2 (a/x). for a<x<—b,

it 3
=-b —a 3I
yx, for x>b

Show that 0' ( , ,,, S Continuous for every positive value of x.

-- [C.P.1944J
Find the differential coeff j c j ens ofth.folIowing With respect to x
(Ex. 11-13).:

(x 3

... (ivi.{x+2)(x+l)2.

(14(3x+4x223 (v) (l+x)/x.

(vii)6 2 —3x 4. viii) 4x - 6x + 2.

: I
-' 2 3
x. —4—_-+---..+-._-.

+ 4 f[T) +
(x)
+ 2 3/(-;.-i—) 
+

(xi)xJ+x
TX TX
DIFFERENTIA TION 187

5x 33x lx
(xii) ++ 12
-

()iii) 2sinx---logx— L e' —6tanx-7cosecx.

(xiv) log o x + log x + e ig + log e' + e

12. (I) x " e'. (ii) x 2

log X. ( iii) x 2 log x 2

(iv)e' sin x. (v)2 sinx. (vi) 1O xL0.

(vii) c0s2 X. (viii) sec x tan x. (ix) (x2 + i

(x) (3x —7 )(3 - 7x). (xi) (x2 + 7 )(x 3 + io).

(xii)( sin x + sec x+ tan x )( cosec x + cos x + cot x ).

(xiii) cosec
X. (xiv) x tan x log x.

(xv) &.ev sec x .(xvi) (I+x)(I+2x)(I+3x[

(xvii) x( I —x )(i - x2 ). (xviii) x sec x log (xe' ).

(xix) x. cot x. log (x ).e'.

lOrXf.

in x (ii) (iii)
13 (i) !___

cosx cosx x

x4 cot
(iv ) -----. [CR 19401 (v ) --.
sinx C

x I+x
(vi)x -. (vii) (viii) -.
log '—i 1—x

+X2 Slfl X
(x) + (xi) +
i —x2 I —slnx

I —cosx sin x+... cos x

(xii) (xlii)
I+ Cos x lT sin 2x

cosx— Cos 2x e' +e'

(xiv) (xv) x -
1— Cos x e +e
 188 DIFFERENTIAL CALCULUS

tan x sin x+ cos x

X
(xvi) - log -- (xvii)
) sIflx — COsX.

cot x+corecx 1+x+x2

(xviii)
COtx — CoeCx I— x+ x 2
x 3 -2+x 3tan
(,) (,j) - e log X.
x-2+x X

() sin
COSX 2
sin x + cos x
xe.
14. If y=.i_Ji+L±_, find for x=2.

X 3 —8x 2
15. If f(x)= X 2 —llx+1O
+ 13x-6 find the values of x for which

Is there any value of x for which f' (x ) is non-existent?

16. From the relation
. x n*I
l—x
deduce the sum of the series 1 + 2x + 3x 2 + ...+
and hence, show that
1+2x+3x 2 + ... to o =(I_x) 2 ; O<jxf<l.
17. If 1(x) =i+x for x<O
=1 for 0:5x:51
=2x 2 +4x+5 for x>l,
find f'
( x ) for all values of x for which it exists.
Does Li f' (x) exist ?

18. (1) If f(x)=.x2 for x_<O

and f(x)= x"sin(I/x) for x >0,
find whether f' (0) exists for n = I and 2.
DIFFEREI'TI7A flON 189

(ii) If f (x ) = [x ] where [] denotes the greatest integer not

exceeding x,find f(x) and draw itsgraph.

ANSWERS

1. (0 3x2+2 (ii) 4x 3 . (iii) _i/x 2 . (iv) -4x

X

(vi)
(v) +x
7( + a 2 ). (vii) J(x2 + i)

(ii) e '.cos x. (iii) 2-2 . log . 2. 2x

2. (I) e 4—. / 2 JT.
)IX 2
(iv) (xe' - e'

3. (i) x.log me. (ii) I +Iogx. (iii) a' cot(x/a). (iv) tanx

4. (j) cos (x / a ) (ii) sin 2x. (iii) 2xcos

(•) xcos-sinx
(iv) i/J(i_ 2 ) (v) 2J)

2
(vii) 2x tan x + x

5. (i) 0. (ii) 0.
6. (i) 0. (ii) No.
9. Continuous for x = I and 2, but f'( x) exists for .r = I and does
hot exist for x = 2.
10. (i) Continuous at x = 1 and 2; f'( x) does not exist for x = 1, but
exists atx=2.

11. (i) 15x 4 +28x 3 -4x-1. (ii) 6x 5 -36x 3 +54x. (iii) I+x+—+-.

(iv) 3x 2 +8x+5 (v) 9x2._4x2_6x

(vi)-x 2 +3+2x. (vii) -12x3+3x2.

(viii)3x' +3x. (ix) 8+IO+x 2 -4x 3 -9r4.

I 9! 25!
(x) +2+-x +8x+—x.
2 2
(-xx ++4 +4x — x
 190
DIFFERENTIAL CALCULUS

I 1 2 I;
NO 13x + x •+ 5x

(xiii) 2cosx- . x -e -6sec 2 x+7cosec xcotx.

(xiv) xloge+ &x +2+e.

(xv) esecx(1+2x+2xtanx)/2.f
(xvi) 18x 2 +22x+6 (xvii) 4x 3 -3x 2 -2xi-I.
(xviii) sec x{j + x+ (I + x tan )(x + log ) }.
e'{x cot x(i + 2 log x+ x log x)_v 2cosec2 x log x}.

(xx)iO cot x 1( - 2cosec2x + IFtjoglo+-LX-12)lox+x

].
13. (I) sec 2 X (ii) sec xtan v (iii) (x Cos x - sin )/x 2
(iv) x 2 (4sin- xcosx)/sinx. (v) _e_(cosec2x+ cotx)

5 (J X)-1
(Vi) x"1 ( it log x - I )/ ( log )2 (vii) (

2 4x
(viii) — (i.\' -______-__ (x) ----------
(l-x)2 (x2)2 (- )
2 Cos x 2 sin x
(xi) (xii) (x) 0.
-sinx )2 +COS X)2-

(xiv) —2 sin x. (xv) e 2 (i * 2x ).

-2
(Xvi) - x -i tan x + (i - log .v ) ec 2. (Xvii)
)2

2 cosec x
(xviii) 2(l_X2)
(cot x—cosecx)2 (xix) _________
)2
( - + 2
(xx) 2(x + t - x -2 _3).

€ -v 2X
() —f{x + ( - 1 )tan x }log x + tan xJ.

(xx) C
2x ) co s
1 (sin x+cosx)2

14. 21. 15.4. 16 non-existent at I. 10.

DIFFERENTIATION 191

- (n + t)x + nx
16. (I-x)2
17. lifx<O,OifO<x<l,4(x-i-l)ifx>l;No.
18. (i) No, Yes.
(ii) f'(x ) = o for all values of x except zero and integral values, for
which it does not exist.
7.6. Differentiation of a Function of a Function.
Let y = f (v ), where v = 0(x), and f ( v ) and 0(.v) aie
continuous. Thus y is also a continuous function of x.
Let f'(v) and Ø ' (x) exist, and be finite.
A ssume v +v =O(x +A x ) and y+A v=f(v+Lw).
It is evident that when &-0. AV - O,andas AV - 4 A-40.

L.ytw
iy-=-.-.
Now, [AV ^1O
Ax AV AX

Li = Lt . Li . [^ ,o]
---oAx A—oAV
dydy dv
i.e.,
dx dv dx

If AV 0, then Ay = 0. [Otherwise , i.e., f'(i') would not be

finite.] .. Ay/Ar=O [v A#0 I

d . . dv
Hence, - Li - 0. Similarly. - = 0.
dx -.aAx dx
Hence, the above relation is true in this case also.

Illustration : Suppose y = sin x

then we can write y=sinv, where v=

Iv dy di,
—=--=cosv.2x=2xcosx
dx dv dx
The above rule can easily be generalized.
Proof depends on the corresponding limit theorem, see § 3.8 (iv).
 192 DIFFERENTIAL CALCULUS

Thus,if y= f(v),where v=4(w),andw..ru,,(x).

dy dydvdw
then = -j- and so on.

dy dx dy /dx
7,7. —x—=i, /—,pmvdedIdva.
dx dy dx jdy
Suppose y = f (x ) , where f (x ) is continuous. From this, inmost
cases, we can treat x as a function of y.
Let
Itis evident that when Ay - 0, Ax– 0.
V X 4X
Now, II
Ax Iiy x /Lty
/ Ay
( /1%A
Li ---= b I'i—i,
-.ax A y)

dv /dx d y dx
i.e., - = 1/—, Of, — x — I.
dv /dy dx dy
7.8. Differential Coefficients of Inverse Circular Functions.
(I) Let y = sin x. [I x !^ 1 1' .. x sin y.
dv '-2 y rT
=cosy='I–stn v1–x
(I'
• I dx
for x^ J,or, -1, ±Y Ji
dx /dy
=
Thus, --- (sfl ')
dx [–i <x <iJ
(ii)Let y= cos ' [X:gi]' .'. X=tOSy.

dr
=_siny=_,J1_cos2y._,/jT.

= I/ =
-______
for x #1. or, - I,
dx •/ Ji.x2
'The domain for wh y exists.
 DIFFERENTIATION
193

d 1
Thus, __(Cos ')=_
dx ,
Ii_x2
Note. This also follows immediately from the relation

coslx=,r_sin_Ix.

(iii)Let y=tanx. x = tan y.

2
— =sec y=l+tan2 y=1+x 2
dy
• dy_ 1 /dx I
dx /dy 1+x2

- ' I
Thus, - tan 'x)=
dx 1+x2
(iv) Let y = cot-' x. x= cot y.
dx 2 / '
—=- cos
dy y=-l+cot 2 '
• y.) = -J+x 2

• dy1/dc I
dx /dv l+x2

d I
____
Thus, - cot x1=----
dx • 1+x2
Note. This also follows immediately from the relation

cot' x 1t-tafl X.

(V) Let Y=sec'x. [jx jai} . •. x= sec y.

dX
sec y tan y = sec y y - 1 =x
dy =
forx or,
dx /.dy
=i/4= jTj
1 X.

d(. I 1
*ec x )=
, 771 [lxi>!]

(vi) Let y = cos ecx. X = c

For which exists.
13-
194 DIFFERENTIAL CALCULUS

dx
- = —cosecycoty = –cosecyJcOSec y–1 = —x'ix –I

dy Idx
for x#I,or. - I, —=1/—=–
' I°Y xJx2_l.

d ' I
Thus,
dx
cosec-'x)=- EIx>l1
Note This also follows immediately from the relation
cosecx= -,r – secX.

7.9. Derivatives of Hyperbolic Function?.

d e + e
d(e–e'')
—(sinhx)=—; 1= =coshx
dx dx'., 2 ) 2

d d(ex+e_'l e –e
—(coshx)= — 1 1= =slnhx.
dx cfr!i 2) 2

d d (sinh x ' cosh 2 x - inh 2 x

—(tanhx) =— 1 1=
dx dx cosh xJ cosh 2 x

=sech2x.
- cosh 2 x

Similarly, -- (cothx)= –cosech 2 x

dx
d I "Oxcoshx–sinhx
( ______
_(sechx)'_
dx dxcoshx) cosh2x
- sinhx
-- –sech x tanh x
cosh x =

Similarly,(cosech x ) = –cosech x coth x

dx
Let y=sinh x. x = sinh.

dx r
—=coshv=,iI+sinh x='jl+x
2 r 2

dy

For the definitions and properties of Hyperbolic Functions, see Authors'

Higher Trigonometry, Chapter XII.
 DIFFEREW TIA TJON
'95

Since 4L,/i_ I
A /a'y i.

Thus, (SIfl -I
I
dr
71 --x

Similarly, !..(COSh_lX)_ (x> l)

-(tanl'x)= ---, (x<l)

dc l-x2
(x>l)
x2-i
di 1
—cosechj=-
-'x
-

-dl sech x)= - ____.- (x<l)

xl_x2
The derivatives of inverse hyperbolic functions can also be obtained
by differentiating their values. vL'.,

Sinhx=Iog(x+); cosh ' x= log ( x+ 177

-I I l±x I
tanh x = - log . coth-Ix = - log x+l
• 2 l-x' - 2 xl'
l+2. l+l_ X
cosech x = log -_--_ sech x = log
X

7.10. Logarithmic Differentiation.

If we have a function raised to a power which is also a funtion, or if we
have the product of a number of functions, to differentiate such expressions
it would be convenient first to take logarithm of the expression and then
differentiate. Such a process is cajied the logarithmic differentiation.

(i) Let y = { f ( x )}ø() . ofjfl --

dr
Uei, logy = (x).logf (x)
196 DIFFERENTIAL CALCULUS

Differentiating both sides with respect to x,

= i(x ).__-_)f'(x)+sP'(x ).logf(x)

(il)Lety=f1(X)Xf,(X).....f(x); to find
dx

Here, log y = log f (x ) + log f2 (x ) + .... . + log f (x).

Differentiate each side with respect to x.
ldyf( X ) f'2(x) ^L

Now, multiplying left-side by y and right-side by

f1(x).f2(x) ..... j(x)
dy
f(x).f2(x).f(x)...f,,(X)
dx

Hence, the differential coefficient of the product of a finite number of

functions is found by multiplying the differential coefficient ofeachfwtction
taken separately by the product of all the remaining functions and adding
up the results thus formed, as already obtained otherwise.
See § 7.4, Theorem IV, Note.]

7.11. Implicit Functions. ..

In many cases it may be inconvienient or even impossible'io solve a
given equation of the form f (x, y ) = 0 fory in téms of x. However, the
equation may define y as a function of x. In such cases, y is said to be an
dy
implicit function of L If y be a differentiable function of x, then dx
may be

obtained as follows:
Differentiate each term of the equation with respect to x, regarding y
dy ,
as an unknown function of x having a derivative and then solve the
dir

dy
resulting equation for
 DIFFERENTIATION

Illustratlon:Fjnd, if
Differentiating each term with respect to x.

3x2

(6y_2)y 2 _3x 2 ,. '_ y2-3x2

- dx dx 6y-2.xy
7.12. Parame'trjc Equations.
Sometimes in the equation of a curve, x and y are expressed in
terms of a third variable known as a parameter.

In such cases, to find it is not essential to eliminate the

dx
parameter and express y in terms of x. We may proceed as follows:

Le t x=O(t), y = V(t).
Then x may be regarded as a function of t and also y is a funcion
of.:.

dy_dy th_dy /dx

Now_.._._/__ (dx — ^O

(By §7.6 and 7.71

For illustration see § 7.13, Fm 6.

7.13. Illusfrative Ezamples.

Ex. 1. Show that -4--x"=nx" when n isapositive integer by the

dx
product rule.
[Note, Theorem. IV, Art. 7.4)
Let f(x)=x"=x.x....x (n factors)
f' (x) =x.x ... to (n–i) factors+x,x ... to (n –i) factors

+x.x... to ( n –1) factors + ...... to n terms

=nX n-I
 198 DIFFERENTIAL CALCULUS

(b) Assuming that x" = nx' when n is a positive integer,

dx
show that the same result is true when n is a negative integer or a
rational fraction, positive or negative.
When n is a negative integer, suppose n = - m, where m is a positive
integer.

dx dx
d 1 OXx" -mx xl
- = —mx = n.E

Next, let us suppose n is a rational fraction, positive or negative and

let n = p / q, where q is a positive integer and p any integer positive or
negative.
11q and yz".
Then y =x " =x ". Let ; then

Thus, dY=/!=iz/q =7-I

ILX
dx dz dz q

{e*(') }= e Ø'(x) for all real values of x,

Ex. 2. Assuming that
dx
deduce that - (x- ) at' for all real values of x.
dx
n nlogs,
Let y=X =e

dy r.-(e7b0 ) e b0 ._(nlogx) x".n.- = nx

Then
dxdx .dx x

Ex. 3. Find the differential coefficient of sin (log Sec x).

Let y={sin(logsecx)}2
=u 2 , where u=sin(Iogsecx)sinV , where v=Iogsecx
log w, where w= sec x.
dv dydudvdW = 2u. 1
sec
cos v— . X tan X
dx dudvdwdX w

=2sin(logseCi )cos(Iogsecx).tanx
sin (2 log sec . tan x.
 DIIFEREmiArIoN

Ez.4. Differentiate (secx ) 's".

Let y =(scx )'. .. logy =tanx.logsecx.
Differentiating both sides with respect to x,
ldy 1
= tan x.—.secxtanx+ 2 xlogsecx
ydx secx
= tan 2 x+ sec 2x log secx
dv
1-=(secx) tan2 x+Sec 2xlogsecx

Note. Writing the given function as e Ufl x.Iog Sec X, we may proceed
to diftèrencjate it.

Ex. 5. Find ,
Y(x-3)(x-4)
Taking logarithm Of both sides,
logy ={1og(x_l)+Iox_2)_log(x_3)log(x4)}.
Differentiating both sides with respect lox.
I +I -I
y d - 21jx-l) (x,2) (x-3) (x-4)

- (x - i )(x-2 )(x-3 )(x-4 )•

dy 2x2-IOx+Il
-

EL6.Find , if x=a(O-sin8), y=a(1+coso).

dx

• dydy/dx asinO 2siniOcosiO

-COtO.
dx del dO a(I-cos9) 2 sin 2-O

dx •• 1+sinx+JI-sjnx

2sjn2x
On rationalising the denominator, y = tan - i-cosx = tan 2sin- r cos- .r

tan Laan4x=+x dy
... dx
200 DIFFERINT1A L CA LCULUS

Note. Sometimes an algebraical or trigonometrical transformation as

'shown in this example considerably shortens the work. The next example
also illustrates the same method.

Ex.& If y—tan

J, l seco-1 1—cosO
Putting
Putting x= tan8, = =
X tanO sinO
2sin210
= 2 =tan-LO
2sin-OcosO 2

Hence, y=tan tan-6=8 =-tanx.

•dy I 1
dx
2 1+x2
EL 9. Differentiate sin x with respect to x2
Let y=sinx, z'X2.
• dy dy dx dy /dz cosf
dz dx dz dx l dx 2x
Note. This is an example of the differential coefficient of a function of x with
respect to another function of x.
Ex. 10. If sin y = xsin (a + y), prove that
tsin2(a+y)
dx sin
From the given relation, we have
smy . (1)
•x=
sin (a+y)
Hence x is a funciton of y.
differentiating both sides of (1) with respect to y,

sin (a-+ y)cosy— sin ycos(a + y)

dy sin 2(o+y)

sin{(a+y)— y} = sina
sin 2 (a+y) sin2(a+y)
 201
DIFFERENTIA TION

Idx
Since, by Art. 7.7,dy- 1/ -, the required result follows.
dx /dy

Es. 11. Find the derivative of ( x). where

f (x ) #(x ) V (x )
A (x)= 1 2 (x) 02( x ) W2(x)
f 3 (x) # 3 (x) V3 (X)
and f 1 (x), f 2 (x ), f (x ), 0 1 (x), e:c.aredjfferentfunctionsofx.

Fint Method:

f1.(x+h) # i (x+h) v1(x+h)

ti(x+h)-'A (X )= f 2 (x+h) # 2 (x +h) v2(x+h)
if 3(x+h) 0 3 (x+h) p'3(x+h)

f 1 (x) #(x) w (x)

- 12( X ) 020) I'2(X)
f 3 (x) 030) tV3(X)

f1(x+h)-f,(x) #i(x+h)-#i(x) y' j (x+h)-' (x)

f 2 (x+h) #2(x+h) tv2(x+c)

f3 (x+h) 03(x+k) p(x+h)

f(x) #i(x) ivi(x)

+ f 2 (x+h)-f 2 (x) 02(x+h)-02(x) 2(x+h)-Iv2(x)
f3(x) •3(X) v3(X)

f(x) v(x)

f2(x)
f3(x+h)-f3(x.) #3(x-4-h)-#3(x)
01(X ) y,3(x±h)-W3(x)

(I)
The right-side on simplification can be easily shown to be equal to
the left-side.
 202
DIFFERENTIAL CALCULUS

Dividing (1) throughout by h and letting h -+0. we get

f (x) O'i (x) (1(,(x)
1(x ) •x ) ' j(x )
A(x)= f2(x) y1z(x) + f '2(x) 0 2 (x) (11'2(x)
13 (x) o( (x) 13(X) 03(X) p3(x)

f, (X o(x) ui(x)
+ 1 2 (x) 0 2 (x) 412(X)

f 3 (x ) 0 '3(X) (113(X)
Second Method:

Clearly A(x)f1(x){02(x)413(x)03(x)412()}

+ f (x ) {0'2 ( )41 3 (x ) - 03 (x ) 41 2 (x ) }

f (x ) # j (x) (x) f1(x) Ø'i(x) V 1 (x)

'(x) f (x ) (x) t'(x) + fl
( x ) 0'2(x) 412(X)
f'j (x) 0 3 (x) (1(3(X)
I
fa(x) 0'3(x) (x)
Oi(x) ip'1(x)
X)
+ f2 (x )) 02(x) 0"2(
f. (X
13(X) 03( x ) V3(x)

Thus, the derivative of a third order determinant A ( x) is equal to the

sum of three determinants, each obtained by differentiating one column of
A(x) leaving the other columns unaltered. Similarly, A( x) is the sum of
three determinants, each obtained by defferentiating one row of A( x)
leaving the other rows unaltered.
The similar result is true whatever be the order of the determinant.

EXAMPLES -VU (B)

Find the differential coefficient of [Ex. I —7]
)7.
1. (i) (OW}'. (ii) (x2 +5 (iii) j(x 2
+ a 2)
(iv) I/(ax +b) (v) (eX)3. (vi) 1j.
DIFFERENTIATION - 203

(vii) sin" X. (viii)tan sI X. (ix) sec 3x.

I. .; . I
(x) srn x) . (xi) tan xj (xn) J(x).

2. (1) e e"'. (iii) e"1"'.

(iv) e . (v) e " .. (vi) e

(vii) e - e

3. (I) a( (ii) 7 (1)

4. (i) log 0 (x). (ii) log sin x. (iii) log cos x.

(iv) log (x + a ). (v) log (ax + b). (vi) log .J.

(vii) log (ax 2 + bx + c ). (viii) log (log x).

(x) 10 (x) log tan x.

(xi) log (sec x+ tan x) (xii) log. a.

(xiii) log . sin x •(,iv) log (a + x ).

(xv) log ^ i n x X (xvi) log s,,, (sec x ).

2
(xvii)log tan (+ n + 4 x). (.Viii) log +

io(2x+.J2+1).
(xix) log (+ .i) . (Xx) log

• log l+x
(x);)

5. (1) sin 0(x). (ii) cosØ(x). (iii) tan 0(x).

(Iv) cosecØ(x). (v) secØ(x). (vi) cot *(x).

(vii) sin ax. (viii) cos (ax+b ).(ix) COS 2 x.

(x) tan mx. (xi) cosec 3X. (xii) sin 2x cos; x,

(xiii) cos 2x cos 3x. (xiv) sinx° (degree). (xv) e' sin bx

• (Xvi) e' Cos (bx+c). (xvii) tan 3x + cot 4x.

• (Xviii) sin xsin 2x sin 3x. (xix) a tan 2 x +.b cot 2 X.

/
204
DIFFERENTIA L CA LCULUS

(xx) sin x cos "x (=) sin' x / cos x.

(xxii) cot x coth x (xxiii) tanh x - -} tanh3x.

(xxiv) log tanb x.

6. (') sin Ø(x). (ii) tan-'O(x)

(iii) sec
Ø(x) (iv) sinx2.
(v) tan'(.J). v I
( i) tan (x /a).
v
( ü) sin (x/a ). j
(vii )sec' xJ.
(ix) cosJ(ax+b) (x) cot -1 (e' ). .•
(xi) Sec A tan x). NO tan '(sec x). -
(Xin)tan
-if
1+x+.r (xiv) cos -' (8x 4 2
- 8x .+ i).
(xv) sin -j (3x - 4? )
(xvi) sec (tan
[ C. P. 1940
(xvi) tan (sin x (xviii) tan' (tanh .. x).
NO cot (cosec x + cot x).

(xx) tan ( sec x+ tan x) (Xxi) cot 1 (I1+x 2 ..)

(xxii) cot -' (xxiii) cos 1- 2
l—.x I +•x2
(xxiv) tan a + bx (xxv) sin 2x
b - ax 1 + x2•
(xxvi) Sec ' x2 +1 2x
X2 (xxvii)tan
-1• 1- x2•
(xxviii) tan'
(C. P 19431

(xxix) tan —__

[C. P.1938)
 205
DIFFERENTIATION

1
3x-x
() tan'.
t-3x2

(xxxii) sech x - cosech x. (xxxiii)tanh (tan - ).

(xxxiv) tanh_ t {(x 2 x2 +i)}

7. (1) cos (ii)

(H) e. [C.PI943,1948]

)2 rF-;;1.
(iii) e . (iv) e

(Vic log tan -1 x. (vii) ,J (log sin x). (viii) (log sin x )2

(ix) cos{2sin'(cOs x)}. (x) s in 2 (16g 2)

(xi) log sec (ax+b)3. [C.R19411

xi')log {2X+4+J(4x2+l6x_12)}

xii,J(l+ log x log sin x). L. P. 19441

(xiv) tan log sin (e" ).

(xv)A(x+2_1)+Bx_JX2_1).

8. Find the differential coefficients of:

(i) (ii) (i +W. (iii) x. (iv)

(v) a. (vi) e ' . (vii) (viii) x '.

(Ix) •(sinx). (x) x C. P. 19441

(xi) (sinx)b05x. . IC.P.19431

(xii)x. . IC.P1937]

(xiii) (sin x) +(cos

, (xiv)(tanx)+( cot x)'.

9. Find the differential coefficients of:
(i) (1_x)(l_2-i)(1--3x)(l-4x).
 206
DIFFERENTIA L CA LCULUS

1( 1+x
(ii) jx(x+I)(x+2). (iii)

.21
(iv)( 2 x2 
a2+x2 
J
Iog{ex(i:_)2}.
(v)

3FX
x+1

+4
(vi)x
. (C.P./941J
+3

x3Ix2_12
(vii) for x=4.
Y20 -
-3x

(
x
(VW) 1 C.R1935]
1+1_2 )

(4x+I)
()

(x) ( x
x
X x J r x(
dy
10. Find  tn the following cases:

(I) 3x4-2y+2yo. [(.P.1941J
(ii)x 4 +. x2 V2 +  y = 0. 1 C. P 1939]
(iii)
x3+y3+4x2y_250.
(iv)x3 + y 3 = 3a.
(v). x a• 3+y3=
(vi) ax 2 + 2/uy + by 2 + 2gx  + 2j5 + c = 0
(vii)x =  y log (xy ). (viii) xFy 4 = ( x + , )p+q
(ix)y=x'. IC.Pl940]
(x) x = Y'. F C. P. 1945, V P 2002
(xi) x y ' ='i. [C.? 1943, V P. 20001
NO (cos x)' = (sin y). [C. P2007
1
DIFFERENTIATION 207

(xiii)e'' -&ty ,= 2.

(xiv) log (ty)=x 2 +y 2 . IC.R1943]

dy
11. Find when
dx
(i) x=acosØ, y=bsinø.
(ii) x=a Cos 3 O,y=b sin 38.
iii) x = at 2 • y = 2a1.
(iv) x=sin 2 O, y=tanO.' IC. P 19431
(v) x= asec 2 0, y-= a tan- 3 9. IC. P.19421

(vi) x a(cos: +Jogtan - : y asin:.

(vii) x= a (cos t+:sin:), y a (sin t-: cos :j.
(viii) x= a(2 cos r+ cos 2r), y = a(2 sin i- sin 2t )
(ix) x= 2a sin 2 : cos 2t, y =2a sin 2 t sin 2,
(x) x = 3at /(l+ t 3 y=3a1 2 /(l+1 3 ). FC.P19411
21 x= 2:
(xi) tan y = sin C. P. 19441
1 .-: 2 1+1 2
T is independent ofx.
dv
12. If y=e" and z = e° then show that

13. Differentiate the left-side functions with respect to the right-side ones:
(i) x 5 w. r. t. x 2 (ii) sec x w. r. t. tan x.
log 10 x w.r.t. x 3 . ( iv) tan-x w.r.t.
(iii) x2.

_ 1 1-x 2 -t 2x
(v) cos j— 7. r. t.• tan
+x 1-x2

(vi) tan - ____________ w. r. t. tan 1 X.

x
(vii) x' w.r.t. sin x. (C.P 19381
14. Find the differential coefficients of

(') (II)
208

FI
DIFFERENTIAL CALCULUS

L, + xx2
(iu) (iv) log x2+x+l
x+2 x2 —x+l

If l+Sinx}
(v) log (vi) log IJ1_cosx}
VI l — sinx V tl+cosx

(vii) tan
V i+x)
(viii) tan [CR 1942, 1944, V P 19981
I + sin x

I 1—COSX). COSX — SjflX

(ix) tan (x) tan
VI i+cosx) cos x + sin x

(xi) sin 1 x+ sin _1J(i2)

(xii)sin {2a.r((1_a2x2 )}.

I+;)_J(l_x2 )
(xiii)tan'
(i + x 2 ) +X2)
.(-

(xv)sin{ 2tan
V+x)f
i

(xv)
(1 + xj + . 1( 1 -
(xvi)hj(0 )+,J(a2 x2)
(a +) (a 2 x2
-t
(xvii) log - tan' x.
l._x)

F(a osx - bsin

(xviii) log
OSX+frsjflx
 DIFFERENTIATION

a+btanx
NO 109 [.c.P.194j
a - b tan x

0o) ta{i) tan }.

a+bcosx .1+5cosx
(xx) sin (xxii) cos
b+acosx 5+3 cos x

I1 + —
X 2 +x
(xxiii) log I
J1+2_
V

J(i+ x)+(1_x.)J
(x-i")log (Xxv)x+
v L7(1+77-(1_7j
• x+—
x

Find in the following cases (Ex. 15-23):

15.

16.yfx,(a 2_ x 2 )+ka 2 sin _.

17. y= log(xfrjx2_a2)+sec1&

18. YX(x2+a2)+a2lo8(x^Jx2+a2).

19. x=(a2_y2)+!Ia2.

1-+x• 1 1+x+x2
20. y=log—+-log ,
3
l - x 2 l - x+x 1-x

j iixf+x2 I
21. '
I-x2
14-
 210 DIFFERENTIA L CA LCULUS

::
i+x " "+x ' 1+x '" +x

+
1+x ,"_ p + X'-P
- b+ 2x

24. If jr (x ) = [) show that

b+x)

h22
-a I(' \.a
f'(0)=[2lo, a+ C.II946J
ab

______(a + bx ) - - bx)
25. if

find for what values of x, =0.

f(x)

IT (2,r
26. If sin x sin - + x sin I +x
fl
)
(n — i '1 .Slfl flX
.sinI —,r+x
-
1=—
I
)
L
then show that

71" I2t
cotx+cotl —+x I+cotl
In ) fl
) .

n
+ cot ( 7r + x n cot nx C. P. 1945

27. (1) From the relation (when n is odd)

- 2ir\ I 41r
2 cos B cos (0+ —Cos i
It) \. Il

) Ir
0+ _COSnO
Ii
)
 DIFFERENTIA TION 211

deduce that

( '\ I 4,r '\

tan o+ tan l 8+ 27r j+ tan I O+ 1+...
n) t n)

I 2(n— 1),r \
+ taf + = n tan n9
n J
(ii) From the identity
0 0 0 .0 sin
cos — cos —cos— ....... . COS— =
2 22 2 2 2" sin (6/2")
show that
1 9 1 0 0 1 0
— tafl --+ ----tan --+ 1
—(an--. ........ tan
2 2 2 222 2 .2 2" 2."

1 0
= - cot - - cot 6
2 2"
28. Find f'(x ) in the following cases and determine if it is continuous for
x = 0.
i) J'(x )= O or x cos(l/x.) according ax isorisnotzero.
(ii) f ( ) = 0 or x 3 cos(l / x ) according ax i's or i's not zero.

29. Tf (i+x)"=c 0 +c 1 x+e,v 2 -- ..... +c,,.", then prove that

(1 ) c 1 +2c 2 +3c 3 + ..... + nc',,n.2"* -

(ii) c 0 +2c 1 +3c 2 + .....

at a2x
.30. If y = 1 + +T
- x—a 1x _a1)(x_a2)

a3x
-ai)(x_a2)(xa)
then show that

dy = v a
+ + a3
dx x a 1 — x a, — x a3-x
212 DIFFERENTIAL CALCULUS

(x -a) 4 (x-a) 3 I
31. If A(x)= (x—b) 4 (x—b) 3 I show that
(x -c) 4 (x -c) 3 1

I
(x—a) 4 (x— a) 2 1
A' (X ) = ( x—b) 4 (x—b) 2
4 (x— c)2 1
(X—C )

sin cosx sin

32. If A ( x) cos x - sin x cos x , then show that A' ( x ) = I.
X 1 1
1.1 1 1
1 X
32. If f(x)= 1 1 , prove that f'(x)=3(x—I)2.
lixi
IlIx
34. If the determinant of the 4' order
x a a a
ax a a'

a ax a

a a ax

be denoted by A 4 , show that A'4 = 4A3.

ANSWERS

1-0- ( ii ) 14x(x2 5)6 (iii)x/J(x2 +2).

(iv) -a / ( ax +b) 2 . (v) 3(e )3. (vi)i

(vli)nsin'xcosx (viii)S tan 4 xsec 2 x. Ox) 3scc4 xtanx.

3 (sin x') 22tan'1x .

. 2f()

2. (I) e.Ø'(x) (ii) ae'. (iii) (2+b).e2''.

DIFFERElTA TION - 213

(iv) 4x.e 4. (v) scc2x.e.

'71 77
e e
2 J( x + I ) - 2J(x—I)

Ioga.a().9'(x). (ii) 2(x+ I )772+2

3. 4)

(iii) IogIo.lo'°" .(l+Iogx).

4. 4) Q'(x)/(x). (i) cox. (iii)—tan.x. (jV)I/(x+a).

(v) a/(ax+b). (vi)I/(2x). (vii) (2ax+b)/(ax2+bx+c).

(viii) I/(x log x). (ix) lO'°.log,II.coIx.

(x) sec (xi)x

x cot x log x— log sin x
(xii) —log a/{x(Icx)2}. (xiii)
x(Iogx)2
xlogx—(m+x)Iog (a+x) I.gsinx— xcot-xlogx
(xv)
(xiv) x(a+)(Iogx)2 x(Iogsinx)2
tan x log siila+ cot x log cos x (xv .
(xvi) ii) sec X.
( log sin x)2

('iII) Jx
2 ±a 224x-a)(x-b)
2 log 1e
J4x2 +1 -

S. (i) cosQ(x).*'(x). (ii) —sinQ(x).'(x).

(iii) 20(x),(x) (iv) -.cosec(x)cot*(x).Ø'(x).
(v) sec(x)lanØ(x).'(x). (vi) cccc2(x).G(x).
(vii)acosax, (viii) —a sin (ax+b).
(ix) —sin2x. (x) m
see
2mx•

(cj) COSOC (xi) 4(3cos3x+cosx).

(xiii)—(5 sin 5x+ sin x). (xiv) cos x0(degrce)

ISO
(xv) a sin bx + bcos ).
214 DiFFERENT/AL CALCULUS

(xvi) e''{acos(bx+ c)-b sin (bx fC )}.

COS 2X + 2 COS 4x - 3 COS 6x).

(xvii) 3 sec 2 3x - 4 cosec 2 4x. (xviii)(

(xix) 2 I a tan xsec 2 x - bcot xcosec x

(ix) sinmxcos_lx(m COS 2x_n Sin 2x).

SiflX( 2
(xxi) yncos2.x+nsln x
cos x
2
(xxii) —cosecxcoth x - cotxcosech x -

(xdii) sech x. (xxiv) 2 cosech 2x.

_ Øx)
6. (i) (ii) 14{0(1-)12(iii)
)}2 ()}2 -1
1{Ø(x)}2
2x I a
(iv) (vi) a2
(v) 2(1+x)J

1 3
(vii) (viii)
F.
cx
(ix)-- (x) 2•
1 —(ax+ b) 2( + b)

sin x
(xi) (xii)
i- 2X
x
2
sin x4sin x - cos

l+2x 4
(xiii) 2 (xiv) (xv)

x 3
(xvi)f . (xvii)(i_x 7) (xviii) 1 sech x.

I I
(xix) - (xx) (xx —.
2 1+x

2
(xxil) - _L (xxm) -. (xxiv)
1+x 1+x
2 2 2
(xxv) (-Vi) (xxvii)

(xxviii) -_____ ().(xxx) Ja)(b-x

J(2)
DIFFERENTIATION 2

3 .. 1 1 I
(xxxi) — i- (xxxu)

(xxxiii)- sec x. (xxxiv)1/x.

_xsin((i+x2)) _____x _____

J,)
(i+x2)
(ii 2sin2f

JT ( -
(iii) - e (iv) 2sin
,151n x gl—x

(l+2x)log33J(i+x+x)
(v) (Vi) cosecx.

(vii) -(logsinx cot x. (viii) 2cotxlog sin x.

(ix) 2sin 2x (x) 2x'sin(4logx).

(xi) 3a(ax4-b) 2 tan (ax +b)3 (xii)

2 + 4x 3
log sin x + xcot dog x
(xiii)2
(1 + log x log sin x)

(Xiv) sec 2( log sin e2 )cote2J e 2 .2x.

(xv) fl {A(x+x 2 _1) +B(x_x2_l)}.

8. (i) x(logx+1). (ii) (1+x){log(1+x)+_--}.

t+x
(iv) l+2{(2X+1)logx+x_I+1+x}.
(iii)

(v) a"'.a(loga )2 (vi) e el.ex. (Vii) e'.x (logxi-1 ).

)W .(sec 2 x log sin x + I).
(viii) x' + log x ). (ix) (sinx

log + Cos x
(x) x
J(i_2) x
t
(xi) (sin x)b0 {x' log sin x-i- log x cot x}.

(xii) x.x{logx(lox+l)+l/v}.
216 DIFFERENTIAL CALCULUS

(Xiii)(sinx)CO '{cosxcotx—sinxlogsinx}

+(cos x){cosxIogcosx—sinxtanx}.

(XiV)(tanx)

+(cot x){sec2x(IogcoIx_1)}.

3x2+6x+2
. 0)
96x 3 -150x 2 +70x-10• (II)
3{x(x+IXx+2)}

(i-x (iv) (a2 +x 2 )Ja 4 -x 4

x2(3X4+2ox2+36)
(v) X22 (vi)

(vii) 120.

fly
(viii) y being the given function.

_(s+2X+18X2)
(ix) 6. (x) 0.
(4x+I)4(2x+3)12(5x-1 )5

2x(6x 2 _y) X(2X2+y2 3x2+86y

10. (I) ___________
x , -6y2 Y(X2+2Y2j3Y+4x

x 2 -ay (vi) ax+hy+g

(iv) (v) - - . -- __________
ax-y2 hx+by+f
YJ,
y(x-y) y _________
(vii)() . (viii). (ix)
x x(I.-ykgx)

(xi)- y(IIOgx) y tan x+ log sin y

(x) (x ii)
x(ylogx — x) x (I —Iogy ) Iogcosx — xcoly

(xin) I
—. (xiv) i
X xI_2y2)
11. (1) —(b/a )Cot ø. (ii) —
tanO. (iii) If'.
3 cosec 0.
(iv) .} sec (v) . tan 0 . (Vi) tan:.

NO tan:. (viii) - tan -1. (m) tan 3

DIFFEREJvTIA DION 217

• ('0 :(2_t3)(1_2I3).

110, ! . (ii) sinx. (iii) x3 log 0e.

(iv) 1/ {2x(1 + x2 )} (v) 1. (vi) .

I— x
(vii)x Iogx+sin j x•
x

14. (I) 4{(i_xy —

(l+xr J
1 1
0 2(a-b).J(x+a)J(x+b)

l-x2 2(I _x2)

(iii) I (iv)
$I+x+x )1-x+x2 )'
2 1+x.2 +x 4

(y) sec X. (vi) cosec X. (vii)

(viii) (ix) . (x) -I.
2a
(xi)a (xii)
J(.i_a2x2)

NO . (
xiv),_____-. (xv) - 2Ji
a 2"I
(xvi) - -j-I l +- I. (xvii)
X 4ax) I—x
-ab 2ab
(xviii) 2 2 sinx
2
a cos2 x-b2 sin2 x
(di) 2 2
a cosx-b
4a 2 _b 2b2_a2 .. 4
(X* . (xxi) .(xxu)
2(a+bcosx) b+acox 5+3cosx

'I . -1 x4+x2+2
(xxiii) . (xxiv) . (xxv)' - 2
J i+x 2 2j-•;- (x3+2x)

15. • 16. JaT7,2). .17. !IIT.

Y 6
18. 2J(a+x). • 19. J(a2_y2) 20.
218 DIFFERENTIA L CA LCULUS

21. 22. 2x2+i 23. 0. 250.±(u/b)

l+x4
28. (I) f'(x ) 0 or sin (I/x )+ 2.Y cos I Ix) according as x isorisnot
zero; f'( x ) is discontinuous for x = 0.
) = 0 or 3x 2 cos( i ix) + xsin (i / x ) according as x is or is
not zero; f'( x) IS Continuous for x = 0.

7.14. Significance of derivative and its sign.

A very important aspect of a derivative, following from its definition,
is as a rate-measurer. This will be clear from the following examples.
Let s denote the length of the path covered by a moving particle in
any time. Clearly, as the particle moves continuously, s has a definite
value for every value of t, and, accordingly, by definition, s is function
oft. If s + As be the value of $ corresponding to the value I + At of t,
then Li S denotes the distance moved over by the particle in time At
As
the ratio - in the limit, when A £ becomes infinitely small, represents

the rate at which the particle is describing its path per unit of time at the
As
moment. But on the other hand, Li - is, by definition (Since
- .r-o At
As and denote corresponding changes in s and I), the derivative of s
ds
with respect to g . Thus, the derivative - represents the rate of change
di
of s with respect to 1, i.e., the speed of the moving particle.
More generally, if y be a function of the variable x, changing
continuously with x, then A y being the change in y corresponding to

a change Ax of x, the derivative - = Li represents the rate of

dt Nv -OL\x
change of y with respect to x.

Thus, v being the velocity of a moving particle at time I,

di
represents the time-rate of change of velocity, i.e., its accleration; again,
if V be the volume of a quantity of gas enclosed in a flexible vessel at a
constant temperature when its pressure is p . which we can change at
dV
pleasure, represents the rate of change of volume with pressure and
dp
so on.
DIFFERENTIA nON 219

Next we may note, that y changing with x, if y increases when x is

increased and diminishes when xis diminished, the corresponding changes

and A y and A x are of the same sign, and accordingly, the ratio -'- is
Lx

positive. Hence, A (when it exists and is 0) is positive.

dx
Similarly, ify decreases when x is increased, or increases when x is

diminished - is negelive.
dx

Conversely, a positive sign of -- at a point c indicates that in the

di
neighbourhood of the point y increases or decreases with x, i.e., both y
and x increase or decrease together. On the other hand, a negative sign of

means that y decreases when x increases and vice versa near the point.
dx
A formal proof of the above result is given below.

A theorem on the sign Of f'(x).

if f' ( a)> 0, prove that f ( x) < f ( a ) for all values of x < a

but sufficiently near to a, and f(x)> f(a) for all values of x > a
but sufficiently flea,' to a,

Since

Li f(a+h)_f(a)>Oand
h

Lt
- h

for all sufficiently small values of h, we have

f(a—h)< f(a)< f(a+h).

In other words, there exists some neighbourhood (a - , a + 5) of

a in which

f(x)> - f(a) for x> a, f (x)< f (a) for x <a,

i.e., f ( x) > f ( a ) for x > a but sufficiently near to a,

and , f(x) < f ( a ) for x <a but sufficiently near to a.

220 DIFFERENTIAL CALCULUS

When the function y = I (x) is represented graphically, a

geometrical interpretation of the derivative . corresponding to any

value of x may be given as follows

Fig 7.14.1

Let P be a point (x, y) on the curve, and Q a neighbouring point

(x+Ax, y + A y ) which may be taken on either side of P,sothat
A x may have any sign. The equation to the line PQ is (X, I denoting
current co-ordinates)

Y — y= (x_x)=i (X — x).
x+Ax—x Ax

If 0 be the inclination of this line PQ to the x-axis, the slope of the

line, i.e., tan = A Y ... (1)
Ax
Now, Jet Q approach P along the curve indefinitely closely, so that
A x –, 0. If the straight line PQ tends to a definite limiting position
TPR as Q approaches P from either side, then TPR is called the tangent
to the curve at P. In this. if t, be the inclination of 7PR to the x-axis, then
- — A y dy
as PQ tends to TPR, 8 - ji . Also, as A x -3 0, LI
s,-.oAx dx
from definition. Thus, (1) leads to

tamp =
dx
DIFFERENTIATION 221

Hence, the derivative for any value of . when it exists, is the

dx
trigonometrical tangent of the inclination (otherwise known as slope or
gradient) of the tangent line at the corresponding point P on the curve
y f(x).

Also, if tan .y ) be positive, y is acute (as at P in the figure

dx

below), and at that point y increases with x. If Le., tan u,i be negative
dx
ji is negative (as at Q), or is obtuse (as at R), and y diminishes when x
increases, or vice versa.

X.

.At a point where = 0, the tangent line is parallel to the x-axis (as at
dx

F), and at a point where A .- o, Le., ,the tangent lineis parallel to

dx dy
the y-axis (as at G).
7.15. DIfferentials
If f'(x) is the derivative of 1(x), and Ax is an increment of x,
then the differential of f(x), denoted by the symbol df(x), is defined by
the relation.
df(x)=f'(x)Ax. ... (1)
If f(x)=x, then f'(x)=l.,and (I) reduces to dr = A x.
Thus, when x is the independent variable t,diffcrenLiaP of x ( = dx) is
222 DIFFERENTIAL CALCULUS

identical with Ax. Hence, if y = f (x ), then the relation (1) becomes

dy=f'(x)th ... (2)
i.e., the differential of a function is equal to its derivative multiplied by the
differential of the independent variable.
Thus, if y = tan x, dy = sec x dx.
From the definition of the differential of a function, the following formula:
for finding differentials are obvious:
d(c)=O, where c isaconstant;
d(u+v—w)= du+dv--dw;

(u' vdu - udv

d(uv)=uth'+t'du; 2

Differentials are especially useful n applications of integral calculus.

Note L The students should note carefully that although for the independent
variable x, increment A x and differential dx are equal, this is generally
not the case with the dependent variable y, i.e., A y ;e dy generally.
Note 2. The relation (2) can be written as dy/dx = f'(x); thus, the
quotient of the diffcrentals of y and x is equal to the derivalive of y with
respect to X.
Probably on account of the position that f'(x) occupies in equation
(2) above, f'(x) is called the differential coefficient..

7.16. Approximate Calculations and Small Errors.

If v = f (x ), since Li A)
-- = f' (x ), A y is approximately
^ 0Ax
=f'(x ) for small values of Ax. Thus dy and A)' may be taken as
approximately equal, when A x (= dx) is small. Hence, when only an
approximate value of the change of a function is desired, it is usually
convenient to calulate the value of the corresponding differential and use
this value.
Small errors arising in the value of a function due toad assumed small
error in the independent variable may also be calculated on the same principle.
As an illustration, let us consider the following case
The radius of a sphere is found by measurement to be 7cm; If an error
of .01 cm is made in measuring the radius, find the error made in calculating
the surface-area of the sphere
 DIFFERENT/A T/ON 223

If S cm' be the surface-area of the sphere of radius r cm,

S = 4nr2.

dS=8itrdr.
Here, r=7 and dr=0O1.
approximate error in the calculation of the surface-area

=dS=8xx7x'01=l•76cm2

Note 1. The actual error is 4t{(7.01 )2 _72 }cm 2 whichis very ileat-ly

equal to I'76 cm 1.

Note 2. If dx is the error in x, then the ratio (i) and (ii) 100 are

X per Unit .r) and

called respectively the relative error (i.e., error x the
percentage error They may be easily obtained by logarithmic defferentiation.

7.17. Illustrative Examples.

Ex. 1. lIthe area of a circle increases at a uniform rate, show that the rate
of increase of the perimeter varies inversel y as the radius. ( C. P 1930]
At any time 1, let be the area, P the perimeter and r the radius of
the circle.

Then A=71r 2P=2itr.:. P 2 =4mA.

differentiating both sides with respect to i, we have
dPdA dP 27 d4 I dA
2P4it-- ie
di di' 'dt P di rdi

dA dPi
Since - s constant,
i .. - -.
di di r

Ex. 2. A ladder All, 2.5 rn long, leans against a vertical wall If the
lower end A, which is at a distance of '7 m .from the bottom of the wall,
is being moved away on the ground from the wall at the rate of 2 rn
per second, find how fast is the top B descending on the wall.
Let the distances of A and B from 0, the bottom of the wall, at time!

be x and y. Then the velocities of A and B are and - hence, as

di di
dx
given here -- =
di
224 DIFFERENTIA L CA LCULUS

Now, x 2 +y 2 =252.
Differentiating with respect tot, 2x + 2y =O.
• dy x dx
di ydt
When x=7, y'=2 . 5 2 2 5.76=2.4 2 .24

Hence, when x=•7, y=24, — =•2;

di
dy .7 7 5
from (1), -i-- -j-X.2=- 1 m per second= -5 cm/sec.

the end B is moving at the rate of 51 cm per second

towards 0 (. dy/dzis negative), i.e.. 13 is descending at that rate.

Ex. 3. The adiabatic law for the expansion of air is PV L 4 = k, where k is

a constant. if a! a given time the volume is observed to be 20 cm 3 and the
presure is 5Q dyne per square centimetre, at what raze is the presure
changing, if the volume is decreasing at the rate 012 cm 3 per second?
PV '4=k
Taking logarithm of both sides and differentiating with respect to t.

1dP 1 d
P di V at

Since V =20, P=50 and

di
I dP 1 dP
—x—+l.4x—.(-2)=0 . —=l.4x50=7
"50 dt20 •dt
Thus the pressure is changing at 7 dyne per second.

Ex.4. if y=2x-tan' x- log (x+ 'Ii 77), show that y continually

increases as x changes flvai zero to positive infinity. [C. P 1942

Here, dy = 2__L__1
1 2x
dx
l+x x+2J

-2
I+x 2 Jj•
 DIFFERENTIA TION 225

Since (i + x 2 ) and .JjTT are each greater than I

and are each less than I.

1+x 2 fj T -
dy
dx is positive; also y=O, when x=O.
for positive value of x, y must be positive and continually increases
as x increases from 0 to
Ex. 5. Find approximately the values of tan 460 ,given 10 = 01745 radian.
Let y= tan x: •.dy=sec2xdr

Thus, taking x= 450 dx =l 0 x•01745 01745, we have

4
dy = 2x . 0I745 = 03490

Hence, for an increment of 1 in the angle, the increament in the value

of tan 450 is 03490
tan 46 0 = tan 45 ° + 03490 1-03490 (approximately)

Ex. 6. If in a triangle the side c and the angle C remain constant while
the remaining elements are changed slightly, show that
da db
— +— =o
cosA cosB
a b c
In any triangle,
— ---- = - -
sinA sinB sinC
Since c and C are constants. •. c/sin C = constant = K suppose.
• a
K
sin sin

••a=KsjnA ,and ;.da=KcosA dA

Also,bKsinB,and •. dbKcosB dB

— — +----=K.(dA +dB)Kd(A +B)

cosA cosB
= K.d(g— C)
= KxO = 0 (. C is a constant)
15-
 DIFFERENTIA L CA LCULUS
226

7.18 Miscellaneous Worked Out Examples

1. (i) A function 1(x) is defined in 0:5x:92 by
f (x)=x2+x+1. 0:5x!51,

=2x+1, 1:5x:52•
Examine the continuity and derivability of 1(x) at x 1.
(ii) 1(x) is defined in [0,2] by
f(x)=x 2 +x , for 0!5x<1,
=2, for x=1,

=2x 3 —x+1 for 1<x:52.

Examine the continuity and differentiability of 1(x) at x1.
C. P 1992, B. P '95 1
Solution (i) For continuity at x= 1, we have

Jim f(x) = tim (x 2 +x+ 1)= 3

X-4i-

Iif(x) Urn (2x+1)3

and f (I)=21+13
Thus, Jim f(x) = Jim f ( X )=f (l). So, f(x) is Continuous at

x=1.
For derivability at x 1,

f(1+11)— f(l)-
Lf (1)= )
h-49 •h
h2+3h - .(:..h*01
lim lim
It h-0-

f(1+h)— f(1) - { 2(1+h)+1} -3

Rf'(l) = tim
It h

2h
= Urn — =2 Lh*O
h-.O+ h

Thus, Lf'(1) # Rf'(l)

Hence f'(l) does not exist, i.e., f(x) is not derivable at x = 1.
 DIFFERENT/AT/ON 227

(ii) For continuity at x = I,

urn f(x) = Jim (x2+x)=2

x-tt-

Jim 1(x) =' Jim (2x—x+1)=2

and f(l)=2

Hence, 1(x) is continuous at x = I

For differentiability at x = 1,
2
+x)_2
Jim f(x)1(1)
= urn
x-.l- x-1 -.l- i—I

= tim (x-1)(+2)
.'-4I- (x-1)

= urn (x+2), -. x-1*0

I-
3
i.e., L/'(1) = 3

Jim f(x)—f(1) Jim .(2xxt1)_2

,_.I+ X-1 .-i+ x— 1

= Jim (2x2+2.+IX._1)
(x — i)

= Jim (2x2 + 2 + i). ... x —I * 0

=5,
i.e.. Rf'(1)=5

since Lf'(l) * Rf'(l),

1(x) is not differentiable at x = 1.

EL 2. (1) Show that the function f defined by

f(x)=3+2x for

=3-2x for 0<x<

is continuous but not differentiable at . =0. [B. P 1999, 2006
 DIFFERENTIAL CALCULUS
228

(ii) If f(x) = x, 0< x <I

=2—x. 1:5x!52
1
=x--x 2,

show that f(x) is continuous at x = i and at x = 2, and that

f'(2) exists, but f'(l) does not. [ C. P. 1989. .93. '97 1

Solution : (i) For continuity of f(x) at x = 0,

Jim f(x)= Jim (3+2x)=3

W-40-

urn f(x) = jim (3-24 3

and f(0)'3+2XO3

Thus, jim f(x) = jim f(X )=f(0)

Hence, f(x) is continuous at. x=0.

For differentiability of f(x) at x = 0,

4f'(0)= urn
h

{3+2(0+h)}-3
= lim
/2

= jim =2 (vht0)
h-.O- I?

f (0+h)— f(0)
Rf'(0)= tim
h

= urn
h

—2h
= u rn ----=-2 (vh;60)
h-.O+ /2

Since if'(0) * Rf'(0)

f(x) is not differentiable at x=O.
at x = 0.
Hence, f is continuous but not differentiable
DIFFERENTIATION 229

(ii) Continuity at x= I and at x=O.

tim f(9= tim (x)=1
i-Il-

Jim f(x)= Jim (2–x)=1

-r-.i+
and f(1)=2-1=1.
tim f(x)= urn f(x)=f(1), f is continuous, at x=1.

Again, Jim
Jim f(x)= tim (2–x)=0

Jim f(x) = Jim

• 2
and f(2)=2-2=0.
tim f(x) = Jim f(x)=f(2), f is continuous at x=2.
Differentiability at x = 1 and at x = 2.
Lf'(I) = jim –1(1) = hm
i-I- x-1 i-i- x–I
x *I

Rf'(I)= tim f(x)–f(I) = tim (2–-1

.-*i+ x–I -.i. x-1
–(x--1)
= tim -=-1, •.• x*.1
5-41+ x–1

Lf'(l) * Rf'(J), f'(l) does not exist.

Again, Lf'(2) = tim f–f(2)

s-32- x-2
rum (2–x)–O
:x=2*O
s-.2-- x-2
Rf(2)= iimj_f(2)
s-.2+ x-2

(x±x2)o
= tim
s-2+ x-2
–x(x–)2
= tim —=-1, •. x-2.0
s-.2* 2(x-2)
V L/'(2) = Rf'(2), f'(2) exists and f'(2) = –I.
230 DIFFERENTIAL cALCULUS

EL 3. (I) Show that the function

f(x)
'()'

is both continuous and derivable at x = 0.

C. P 1987, '96, 2000)
(ii) Show that the function

f(x)xCos x*0
-
XxI )I
=0, x=0
is continuous at x = 0, but has no derivative there.
C. P. 1981, 93
Solution : (i) For continuity at x=0,
we arc to find a 5, depending  upon E. such that

lf(x)-f(0)I < E Ix-0I< 8.

for

or, < c for <8. IxI

X2 <e
Since. sin '  1:5 1, the above telations will hold. if we take 

•forl xI <8, i.e. , &=,i.

So. f( . ) is continuous at x=0.
For derivability at x = 0,

x sin (_)
f(x) - f(0) =
We have, P(0) = lint

=lim xsinI-) '.

-° kx)

( x1
= lint (x)x lim sin I -
__*o .-.o '

= 0, '. sin ^ 1, and urn x = 0.

DIFFERENTIA TION 231

Hence f'(0) exists and f'(0) = 0.

Thus f(x) is both continuous and derivable at x=0.

(ii) 1, by making, I x I < e,

We can make Cos (!)<E.

where E is any pre-assigned positive quantity, however small.

So, lim xo-i-)=o. Also, f (0)=ó by definition.

Thus, urn f(x)f(0)

Hence, f is continuous at x = 0.
For derivability, at x =0. we have

f(0+h)—f(o) h
(1)
fj0)=lim =lim
h h-,O h

=iirnco !"1. h*O

k-.O
which does not exist.
I-knce, f'(0) does not exist.
So, f(x) is continuous at x = 0, but has no derivative there.

Ex. 4.Iff(x)21x1+Ix-21 find f'(l). [C.P.1992,2000, '02]

Solution: We have I x i for x >0
=0 for x=0 ... (1)
=—x for x<0
and Ix-21=x-2, when x>2
=0, when x=2 . . . (2)
=2—x, when x<2.
To find f'(1), we are concerned with values of x in the
neighbourhood of x = I.
 232 - DIFFERENTIA L CA LCULUS

From (]land (2), f(x)=2x+(2— x)

=x+2 for O<x.z2
f(x)— f(1) (x+2)-3
Now, f'(i) = urn = lint
-.i x—1 i-I x-1
x-1
=hm—=1, . x—l*O
-i-" x-1

Ex. 5. Let f(x) = x 2 , when x is rational

= 0 when x is irrational
show that f'(0)=O. [ ..P. 1996, 2001

f(0+h)— f(0)
Solution : P(0) = Jim
I.o h
• h2-0
= tim (when h is rational)
h-.O h
=lim h=0

f(O+h)— f(0)
Again, f'(0) =II'im
h-.O h

Jim= (when h is irrational)

h-.O h
=0.
f'(0)rrO

Ex. 6. If f(x) be an even function and f'(0) exists, show that

f'(0)=O. IC. P. 19831
Solution •; f (s) is an even function, f(x) = f(— x).
or, f'(x)=— f'(— x) .. (I)
since f'(0) exists, putting x=O in (1)
f'(0)=— f'(0) i.e., f'(0)=O.
Ex. 7. Show that the derivative of a differentiable odd function is an
even function.
Solution Let f(x) be a differentiable odd fuction.
Then, f(x) = — f (— x)
 DIFFERENTIATION 233

Since, f(x) is differentiable,

f'(x) = -{-f'(-x)} = f'(-x),
so that f'(x) is an even function.
EL 8. Find, from definition, the derivatives of

(i) sin(J), (x >0) I C. P 1985 J

(ii) sin(logx), (x >0) (iii) x2 COS X.

Solution : Let f(x) = sin (v') (x > 0)

sinJ_sin'/
f'(x) = Jim
h

=Jim (1)
!01 h

Now, Jim sin

,,_.o
sin(y+k)-siny
=Jim where = y + Ic
k
" k" . k
2coslY+-J.sin.
as h-+0, k-4()
=lim
k

•
I
i
\ Sifl
=hm COS iy+-ix Jim — 2
!,. 2) -.o k
2
= cos y x I = cos y = COS IFX

4. (/T - +
and Jim
h-.O h
=Jim
h-.O
h(ii +

=u rn
•
h-.o h+J)°
x+h-x
iT+J' . h^O

=.
234 DIFFERENTIA L CA LCULUS

Using these results in (1)

1 1 1
f (x)=cos,jxx =—cos.Jx
2 x 2J
(ii) Let 1(x) = sin(log4, (x>0)
f(x+h)-f(x)
From definition, P(x) = Lim
h-.O h

sin{Iog(x + h)} - sin{iog(x)}

= lim (1)
h-.O Ii
Let, logx=u
If the increment of u be k corresponding to an increment Ii of x,
u+k =log(x+h).

i.e., k=log(x+Ii)-u=log(x+h)-logx
Obviously, k -4 0, as h -, 0, and h -+0, when k -, 0.
From (1) we have
• sin(u+k)-sinu
f (x)=hm
h
lim sin(u + k) - sin u k
= - x HIM -
k-.O k h-.oh

k
2 Co u+— ,Sifl_
2) 2og(x+h)- log x
= lim Xlim
k-.o k h

k (x+ h)
k \
sin
X
= urn cos('u+ — Jxhm —2-xlim
k -0k -0 k i-o h
2

1o4'l
cosuX(l)xlim +X
h-.O h
—XX
x

1iog(1+:) h
cosux - .log where -=1 and t-,0 as h-0
x,_.o I

cos(logx) 1
= xI=-.cos(logx)
x x
 235
DIFH.REi%l'lATION

(iii) Let, f(x) = cosx

From definition,
f(x+h)-f(X)
f'(x)=lim
h

(x+h)2C0S(X+h)X2CO5X
= urn

x 2 { cos(x+ h) - cosx}+2xhCOS(X+ h) +h 2 cos(x + h)

hm
h-.O h

X2 cos(x+Ii)—COS.X
=—

+2xXIim cos(x+h)-1-Iim (hcos(x+h))

( h'
2sin x
=x2xIirn
h-,O h
• (h
sin! -
• • ( h\ • _____
=–r xhrnsinl x+ ixIrm—- +2xcosx
2) h-.O At

= –x2 sinxxi+2xCOSX 2xcosx — x 2 sinx

EL 9. () If I +J =a(x–y), show that

FL2 B. P19911

(i1) If Ji' +.Ji' =a"(x" – y"). show that

dy - ()
dx y) V1_x2

Solution: (i) Let us assume, x =sin 0 and y =sin

dxdy
—=cosO, ---=cos+
dO d$

J+J=a(x-v)
236 DIFFERENTIA L CA LCULUS

cos0+cosqi=a(sino_ sin )

or,

, cot! (o—q)=a, . cos -!-(o+ct,)#o

or, 0—'2cota=constant

or, or, i.e.,

dO

Now,
dx d4 dO dx cosO

(ii) Here, Ij - x2" + = a"(x" - v") (1)

Let us substitute f = sjn0 and y" = sinp
dx dx cosO
Then, f ix - cosO, i.e., (2)
dO (/0 nx"
,, y cos
and ny (/- = C0s4,i.e.,dy
- = - ... (3)
dod ny"
From (I), O+JIS if l24,(.oi,)

or, cosO+cos=a"(sinO_sin)

or, 2 cos -(O+4)) Cos ± (O_)= 2-a"

or,

i.e., 0_0=2cot_1(a)=constant
d4 d$
Of 1---O ,e
dO '' dO
dy fly 4 dO CO54)
Now, = -X IX
;;;;:i I From (2). (3) 1

I \'I \I I
cos$ I I—y
2
cO L) 1_y2
 237
DIFFERENTIATION

2x
Ex. 10. -(1) If y = tan prove that
I-i-15x 2
dy_ 5 - 3
dxl+25x 2 1+9x2
(ii) If the sum of first n terms of a OP. with common ratio r is S,,,

prove that

(r— 1)- = (n— 1)S—nS,_1.

dr

2x 5x-3x
Solution(j) y=tan tan
1+15x 2 1+5x.3x

= taii'(5x)—tan_'(3X)
dy 5 3•_ 5 3
dx l+(5x) 2 1+(3x) 2 - 1+25x 2 I+9x2

(ii) Let the first term of the G. P. be a.

a(r _i)
Then S,, =
r-1

or, (r—l)S,, =a(r" —i)

Differentiating both the sides w. r 1. r,

(r —1) dS = n r"
dr

Now, the nth term of the G. P.=ar' = S. —S,,_1

(r-1)-+S,, =n(S—S_,)
dr

i.e., (r— I) = (n— 1)S —nS,,..1

dr

Ex. 11. (1) If to =, show that

dy cosx
dx 2v-I
 238 S DIFFERENTIAL CALCULLIS

I 2x 4x3 8x7 I
(ii) Show that —+----+----+---^...t=_
—
1+x l+x2 l+x 4 I+x 8 l—x
(O<x<1)
(iii) If y=x+
x+
I
x+
X+ to

show thaL I
dx 
2-

X+tOo

Solution (i) Here, y =

isinx+.Isinx+Jsj-
.-
T.

= Ssin
-X + y
or. Y 2 
= Sinx+y
Differentiating both the sides wr.:. x,
dy dy dy cosx
dx dx dx 2y_i

(ii) We have, (1--x)(1 +x)(l+x2)(j+x4)(1+8)...(1+x'.)

= ( —x)(1 +x2 )(I +x 4 )(l +x 8 )• .. (1 +x")

=(1—x4)(1+x4)(1+xS)...(I+x)

0<x<1, urn x2'=O

n-_
urn {(1—x)(1 +x)(1+x2)(1+x4)(I+x8)....(1+x)}

=Iirn(1—x2)1

log(l —x)+ Iog(I +x)+log(I+x. 2 )+Iog(1 +x 4 )+ log(1 +x8)

+10"=Iogl=O
DIFFERENTIATION 239

Differentiating both the sides w.,1. x, we get

-1 1 2x 4x3
--+O.
1-x 1+x I+x 2 l+x4
I 2x 4x3 1
whence, — +--+---+—
••to— •
1+x j+x2 I+x 4l-x

(iii) Here,

or, y2=xy+I
Differentiating both the sides w.1t. x,
dydy
= x+ v
dr dx
dy y - I -
dr 2y-x 2!

1 --
x+
x+..ioo

Ex. 12. (i) If f(a)=2, f'(a) = 1, j(a)=-I, g'(a) = 2, then find

the value of

lim

(ii) Show that the function 1(x) = x I x , is differentiable at x = 0.

Solution: (0 tim g(x)f(a)- g(à)f(.r)
x-a

f(a){g(x)- g(a)l - g(a){f (x) - f(a)}

=lim
x-a

{g(x)- g(a)} g(a) {f(x) - f(a)}

= urn f(a). - - tim
(x-a) (x-a)

=f(a)g'(a)-g(a).f(a)

= 2x2-(-1)< 1 = 5.
 240 DIFFERENTIA L CA LCULUS

(ii) Here, - f(x)=x 1 , x>O

=0. x=0
=_x 2 , x<0

4f'(0)= Jim j)-j(0) = tim

-.O- x-0 -.o- x

Rf'(0)= urn f(x)-f(0)= urn — =0

-.fl+. x-0 -.o. x
Lf'(0) = Rf'(Q), f(x) is differentiable at x = 0, and f'(Q) = 0.

Ex. 13. (i) If g be the inverse of the function and f'(x) =

f
then prove that g'(x) =I+ l g wl ' -

(ii) If f(x + y) = f(x) .f( y ) for all real values of x, y, f(x) * 0

for any real value of x, and f'(0) is defined and f'(0) =2, prove
that for all values of x, f'(x)=2f(x). Hence find f(x).

Solution (i) . g is the inverse off,

g(x)=f'(x), i.e., f{ g(x)} =x
Differentiating both the sides WI:!. x,
f'{ g(x)} g'(x) = 1.

= = 1
or, g'(x) .•. =
f{g(x)} I l+x

i.e., g'(x) = 1 + {g(x)}.

(ii) f(x+y)=f(x).f(y) ( I)
for all real values of x and y, putting x = y = 0 in (1),
f(0) = PO) -PO)
i.e., f(0)=1, .• f(0)0

f(x+h)- f(x)
Now, f (x) - -
u . rn
Ii

[from (1)1
DIFFERENTIATION 241

=f(x)xlim f(h)-f(0) -f(0)=I

= f(x).f'(0)
= 21(x), f'(0) = 2, given.

f'(x) = 2 1( x ), for all real values of x.

Again, f (x) -
On integration log{f(x)}=2x+logA, A being constant of
integration.
Putting x = 0, log{f(0)} = 0 + log A
or. log A= log l=O, ..A 1
Hence, f(x) =

EXAMPLES-VU(C)

1. A point moves on the parabola 3y =x 2 in such a way that when x= 3

the abscissa is increasing at the rate of 3cm per second. At what rate
is the ordinate increasing at that point?
2. A toy spherical ballon being inflated, the radius is increasing at the rate
of cm per second,, At what rate would the volume be increasing at
the instant when r = 7cm-?
3. A circular plate of metal expands by heat so that its radius increases at
the rate of 025 cm per second. Find the rate at which the surface-.area
is increasing when the radius is 7 cm.
4. The candle-power C of an incandescent lamp and its voltage V are
5V 6
connected by the equation C =
Find the rate at which the candle-power increases with the voltage
when V=200.
5. If Q units be the heat required to raise the temperature of I grain of
water from 0' C to t C, then it is known that
Q=:+10 5.2f 2.3:.
Find the specific heat at 50' C the specific heat being the rate of
16- increase of heat per unit degree rise of temperature.
 242 DIFFERENTIAL CALCULUS

6. A man 15 m tall walks away from a lamp-post 45 m high at the rate

of 4 k per hour.
(i) How fast is the farther end of his shadow moving on the
pavement?
(ii) How fast is his shadow lengthening?
7. If a particle moves according to the law x: 2 ,where x is the distance
(measured from a fixed point) travelled in time I, show that the velocity
will be proportional to time and the rate of change of velocity will be
constant.
8. Water is poured into an inverted conical vessel of which the radius of
the base is 2 m and height 4 m, at the rate of 77 litre per minute. At what
rate is the water-level rising at the instant when the depth is 70 cm?
9. If the side of an equilateral triangle increases at the rate of cm per
second and its area at the rate of 12 cm 2 per second, find the side of the
triangle.
10. If in the rectilinear motion of a particle s = u+ - ft when u and f
are constants, prove that the velocity at time t is u + ft and the
acceleration is f.
11. A man is walking at the rate of 5 km per hour towards the foot of a
building 16 m high. At what rate is he approaching the top when he is
12 m from the foot of the build,ing?
12. A circular ink-blot grows at the rate of 2cm 2 per second. Find the rate
at which the radius is increasing after 2- second.
13. The volume of a right circular cone remains constant. If the radius of
the base is increasing at the rate of 3 cm per second, how fast is the
altitude changing when the altitude is 8 cm and radius 6 cm?
14. Sand is being poured on the ground and forms a pile which has always
the shape of a right circular cone Whose height is equal to the radius of
the base. If sand is falling at the rate of 154 m 3/s, how fast is the height
of the pile increasing when the height is 07 m?
15. The marginal cost of a commodity being the rate of change in the cost
for change in the output, if f(x)= ax. x-t-b
— +d (b > c)bethe total
x+c
cost of an output x, show that the marginal cost falls continuously as
the output increases.
DIFFERENTIA TION 243

I
16. (1) An aeroplane is flying horizontally at a height of km with a velocity
of 15 km an hour. Find the rate at which it is receding from a fixed
point on the ground which it passed over 2 minute ago.
(ii) A kite is 300 in high and there are 500 m of cord out. If the wind
moves the kite horizontally at the rate of 5 km per hour directly
away from the person who is flying 1, how fast is the cord being
paid?
17. If (x) = (x- 1)e + 1, show that •(x) is positive for all positive values
of X.
18. If 1(x) cos x + cos 2 x + x Mn x, show thatf(x)continually diminishes
as x increases from 0 to
19. Showthat,for 0<0<r

sin
(i) diminishes as 0 continually increases.
0
4sin 0
2+cos increases with 0

20. (i) Prove that. If 0< x <

1 2 •1 2 1
(a) I--x <cos x<1--x +—x
2! 2! 4!
1 3. 131 S
(b) 1--x <sin x .<x--x +—x
3! 3! 5!

(ii) Show that, if x >0,

(a)x> log (1+x)>x-.x2 (b) --x 2 +2x+3> (3-x)e
a sin x+ b cos x
21. Given y = ,prove that
csin x+ d COS x
(i) If a=1, b=2, c=3. d=4, then y decreases for all values of x
(ii) If a=2, b=1, c=3, d=4, then increases for all values of x.
22. Find the range of values of x for which each Of the following functions:
(i) x3-3x2-'24x+30 (ii) x 3 -9x 2 +24x-16
(iii) 2x 3 -9x 2 +I2x-3 (iv) x 4 -4x 3 +4x 2 +40
decreases as x increases.
23. Show that the function x 3 - 3x 2 + bx -8 increases with x.
244 DIFFERENTIAL CALCULUS

24. Find the approximate values of the following by the method of

differentials
(i) log10 . 1, given log10=2.303
(n)log 10 101. given log10e=04343
(iii) given J22.5
(iv) sec 2 46°, given I0=00175radian
(v) sin 620, given sin 0_0803
25. What is the approximate change in sin per minute change in 0 when
8 =6ff?(given 1'=0•00029radian).
26. Find the approximately the values of:
(i) x 3 +4x 2 +2x+2 when x=2.000l2
(ii)x 4 +4x 2 +1 when x=l-997
27. Find approximately the difference in areas of two ciivles of radii 7 c
and 7Olcm.
28. What error in the common logarithm of a number will be produced by
an error & 1% in the number ? [log 10 e = 04343]
29. Find the relative error (i.e., error per unit area) in calculating the area of
a triangle two of whose sides are 5 c and 6cm, when the included
angle is taken as 45° instead of 4502g.
30. Show that the relative error in computing the volume of aphere, due to
an error in measuring the radius, is aproximately equal to three times
the relative error in the radius.
31. The angle of efevation of the top of a tower as observed from a distance
of 43 m from the foot of the tower is found to be6O°; if the angle of
elevation was really 600l ,obtain approximately the error in the
calculated height. [1' = 0 . 00029 radian
32. The pressure p and the volume v of a gas are connected by the relation
pv L4 =k.where k is constant. tf there bean increase of 07%inthe
pressure, show that there is a decrease of 05 per cent in the volume.
33. An electric current C as measured by a galvanometer is given by the
relation C .tan0 . Find the percentage error in the current
corresponding to an error 07 per cent in the measurement of 0, when
0 = 45°.
DIFFERENTIATION 245

34. The time T of a complete oscillation of a simple pendulum of length I

is given by the relation T 2itwhere g is a constant. Find

yg
approximately the percentage error in the calculated value of T
correspoinding to an error of 1 per cent in the value of I.
35. (i) In a triangle if the sides and angles receive small variations, but a
and B are constants, show that
tan Adb=bdC.

(ii) In a triangle if the sides a, b be constants and the base angles A

and B vary, show that
dA = dB

a2 -b 2 sin 2A 4b -a2
Sin
2B

36. If a triangle ABC inscribed in a fixed circle be slightly varied in such a

way as to have its vertices always on the circle, show that
da db de
cosA cosil cqsC

ANSWERS
1. 6 cm per second. 2. 56cm 3 1s. 3. llcmpersecond.

4. 96. 5. 1.00425.
6. (I) 6 k per hour. (ii) 2 km per hour.
8. 20cm per second. 9. 8cm.
11.3km per hour. 12. 0•25cm per second.
13. Decreasing 8 c per second. 14. 1 m per second.
16(i) 9 km per hour. (ii) 4km per hour.
22.(i) x>'4 or <-2. (H) x>4 or <2.
(iii) 1 .<x<2. ('iv) x .<0 and 1<x<2.
24.(i) 2.313. (ii) 10043. (iii)2•516. (iv) 297. (v) 08835.
25.01)0015. 26. (i) 30.0036. (ii) 32.856.
27.044cm 2. 28. 00043.
29.000058. 31. 0'05rn i.e.. 5cm.
33.11. . 34.05.
[I1i SucassivE
8.1. Definitions and Notations.
We have seen that the derivative of a funcçion of x, say 1(x), is in
general a function of x. This new function (i.e.. the derivative) may have a
derivative, which is called the second derivative (Or second differential
coefficient) of f(x), the original derivative being called the firs: derivative
(or first differential coefficient). Similarly, the derivative of the second
derivative is called the third derivative; and so on for the n-th derivative.

Thus, if y x3, .:. = 3x

dx
Again, ) =(3x2) = 6x
dxdxdx

d(dy'\ a2 d
Since is denoted by Therefore,
. (i.e., the second
dxyirj dx-
derivative of y with respect to x) in this case is 6x.
d(!Py)=
Again, - d,6x)= 6
dx câ 2 dx

d (d2y
d3y
Since is denoted by — 3-

d3
(i.e., the third derivative ofy with respect to x) is 6 here.

Similarly, the n-th derivative of y with respect to x is generally

denoted by
dx -
If y =1(x), the successive derivatives are also denoted by

yl, y2, - Y31 .........

or, y',
or, y, j, .j ..........
or, f'(x), f(x), f(x) ...........f(x)
or, Df(x) 02f(4, D 3 f(x) ........ D'f(x)

D standing for the symbol d

dx
SUCCESSIVE DIFFERENTIATION 247

8.2. The n-th derivatives* of some special functions.

(i) y = x , where n is a positive integer.
=
Y2 = n(n-1 )x2;

y3 = n (n-I) (n -2 )x and proceeding in a similar manner.

y =n(n- l )(n-2).... {n-(r- l) (r<n)

y =n(n-l)(n-2) . . .
i.e., D(x) =
COr. Since y n which is a constant, y,,, .... etc.'all zeroes
in this case.
(ii) y = (ax + b ) , where m is any number.
yj =na (ax +b);

y =m(m-1)a2(ax+b)'2;
Y3 =,n(m-1)(,n-2)a3(ax+b )'; and proceeding similarly,
y =m (ni-1)(m -2) ........ (m-n+l)a(ax+b).

D(ax+b)'=m(m-1)(m-2)(m-n+1)a(ax +b)
If m be a positive integer greater than n,

Sincem(m- I )(m-2) ...... (m-n+ I )= (rn-n)!

m!
lY(ax+b)'
(in-n)!
m being a positive integer greater than n.

Note. If m be a positive integer less than n, !Y (ax + b =0

When m=n, D"(ax+b)=a.n!.

() y=e
y 1 =ae°'; y 2 =a 2 e"; y 3 =a 3 e .... . y,,

D' (e" ) = a
Cor.(i)D(e)=e'
Cor.(ii) y=a =e"°, ..

* Strictly speaking, in these cases, the n-th derivatives are to be established

generally by the method of Induction.
248 DIFFERENTIAL CALCULUS

1
(iv) y=—
x+a

Y2 (-1)(--2)(x+a)-3 =(-1)2.2'(x+a3
Similarly. y =(-1)3.3!.(x+a) -4 etc.
1) ^( I = (-1)n!
.%
x+a ) (x + a)°1

} (-lYa(m-Fn-l)!
Cor. Proceeding as above, D"
I(+b)' (m-l)!(ax+b)""
(v) y=Iog(x+a)

YI Hence, as in (iv) above

x+a
a)}= (-1)"_1(n-1)!
D"{log(x +
(x + a)"
-1)! a
Co DIog(ax+b)J=
(ax+b)"

(vi) y= sin (ax +b)

Yi =a cos (ax +b)=a sin (4x+ax+b)

Y2
=a2 cos(1t.1ax+b)=a2 sin(2.fn+ax+b)

Y3 =03 COS (2. n+ax+b)=a sin (3. .. n+ax+b) etc

D°{sin(ax+b)} =a sin (x+ax+b)

Similarly, D"{cos (ax + b)}=a n cos (ir+ ax + b)

As particular cases when b = 0,

In
D"{sin ax}=a sin I -n + ax);

(U
D{cos ax}=a" COS l-n+ ax).
SUCCESSIVE DIFFERFRTMTION 249

8.3. The n-th derivatives of rational algebraic functions.

The nth derivative of a fraction whose numerator and denominator
are both rational integral algebraic functions may be conveniently
obtained by resolving the fraction into partial fractions. This is shown
in Ex. 4, Art. 8.4. The rules for decomposing a fraction into partial
fractions are given in the Appendix.
Even when the denominator of a given algebraic fraction cannot be
broken up into real linear factors, the above method of decomposition
can be used by resolving the denominator into imaginary linear factors.
In this case, DeMoivre 's theorem is conveniently applied to put the final
result in the real form. This is illustrated in Ex. 5, Art. 8.4.

8.4. Illustrative Examples.

Ex. 1. If y= sin 3 x,find y,.
sin 3.x = 3 sin x —4 sin 3x
1
.3x=—(3sinx—sm3x)
v=sm

(i2 .
3 sin1— ,z+x - sini— nx+3xJ}
i{ ( L2

Ex.2./fi'=sin3x.cos2x,find y,.

, =!.2 sin 3x cos .x=!( sin 5x+ sin x)

2 2
y =!I5'sin(±n1t+5x)+sin(!n,r+x
2 (2 ) (2
Ex.3. Ify=esinbx.findy,.

Y i = e .a sinbx +e".cosbx . b
=e.(asinbx+b cos bx)
Let.a =rcos, b=r sin $,sothat

r =(a2+b2),=tan*.

Yi = re' sin (bx+Ø).

Similarly, y2 = re" la sin (bx+Q)+b cos(bx+$))
= r2 e' sin ( bx + 20 ), as before.
 250 DIFFERENTIA L CA LCULUS

In a similar way y3 = r 3 e sin ( bx + 3$), etc. and generally

y re'sin(bx + n$ ),

i.e., D(e sin bx) = (a2 +b 2 )esin (bx + n tan'

b)
Note. Similarly,

D (e n' cos bx) = ( a 2 +b 2 ) . " ecOs (bx +n tan k)

Again, if y = e'sin(bi + c),

2
Y.=(a +b 2 ) e sin (bx + c + n tan
J
and if = e' cos(bx+c),

y=(a2+b2)ecos(bx+c+ntan_J

Ex.4. If y= . find y.
X3
+x -6x
x 3 +x 2 -6x=x(x 2 +x-6)=x(x+3)(x-2)
x 2 +x-1 A 9 C
Let
x 3 +x 2 -6x x x+3 x-2
Multiplying both sides by x(x+3)(x-2),we get
x 2 +x- I A (x+3)(x-2)+ fix (x-2)+.Cx (x + 3).
Putting x = 0, -3, 2 successively on both sides, we get

A = !, B =! , .=!.
.6 3 2
II 1 1 1 1
Y =-"-+-.----+-.— :
6 x 3 x+3 2 x-2

I I I I I
Y. (-I) n! — .— +— . +— .
6 xt 3 (x+3)' 2 (x-2)'

Ex. 5. If y = find y,,.

x2+02

I 11 ___
(x+ ,a)(x_:a) 2:a ( x - za x+:a
 SUCCESSIVE DIFFERENTIA TIoN 251

Yn
2ia l(x-ia)' +' (x+ia)'}

= (-1)'n !
- (x +:a) }
Put x=rcosO, a=rsinO

Sothat r=(x2+a2)2, O=tan'a/x.

-(,.*I)
Now
(x—Ia) = r -("+1) (cosO—jsjnO)"

= r -(+ I) {cos (n+I)$+isifl(n+1)0}

=(x + ia)_("+-(.+I)
r (cosO+jsin9)

=r' {cos (n + 1)O—isjn (n + I)O}

Y,, = r" '.2i sin(n+I)8

2ia

Since r =__±_, -*I) a - sin"'G

sinO sin" 0

1 )=LP!,i.,^ O sin(n+1)0,
where O=tan=cot*!
X a

Note. Ify -
I .,41
Osin (n + 1)0,
(x+b) 2 +0 2a"2

where 0 = tan

Cur. If y = tan' x, y,!_ hence

I+x.
D o (tanx)=(—I)"' (n-1)!sjnO shsn0,

where n*L=cot 1i.

252 DIFFERENTIA L CA LCULUS

EXAMPLES - Vifi (A)

1. Find y in the following cases:
(1) y=(a--bx). (ii) y=1I(ax+b)
(iii) y= 11(a—x). (iv) y=Iog (ax+b)'.

(v) y=log{(a—x)I(a+x) (vi) y=4x.

(vii)y=1/'x. (viii) y=(2-3.x).
(ix) y=log(ax+x 2 ) - (x) y 1031.
(xi) y=x/(a+bx). (xii) y=(a—x)I(a+x).
(xiii)y=x/(x-1). (xiv)y =sin 2x.
(xv) y=cos2xcosx. (xvi)y= Cos 3x
(xvii)y=sin 2 .x cos 2 x. (xviii) y=sinx sin 2x sin3x

()ix) y=ecosx. (xx) y=e'sinxsin2x.

(ma) y= e ll sin 4x. (xxii)y=e-' sin 2x

2. Find y3 , if
(I) y=x 2 logx. (ii) y=e.

(iii) y = e (iv) 3 , = sin- 1 x.

3. Find the n-th derivatives of the following functions:

1 1
(ii) x c.P.1993 I (iii) tan X
@ ' x 2 —a2 162 + a

x I
(iv) 2 2 (v) (vi)
x +a _a4 X2 i-x+1

(vii) (x2 (Viii)

+.'T +b2) 4x 2 +4x+5

x 2 +1
(x) [C.P.1993]
(ix) (x-1)(x-2) (x—I)(x-2)(x-3) .
x2 x2
(xi) (X+1)2 (x+2) (xii)
(x—a)(x—b)

I+x 2x
(xiu) tan-' (xiv) Sin

_ —I Ix
(xv) tani (xvi)cot X.
a
SUCCESSIVE DIFFEREWrIAT,0N 253

4. If y x 'i', where n is a positive integer, show that

y,=2f1.3.5...(2n—I))x.

S. If u = sin ax + cos ax, show that

u=a( 1+(—I)sin 2ax). [B. P. 19931

6. If ax 2 +2hxv+bv 2 =1. show that -

y j,2_,
d2
- (hx+by)3

7. Find y 2, if
(i) sinx+cosy=I.
(d) y=tan(x+y).
(iii) x3 + y3 - 3axy = 0.

8. (a) Find .fLl in the following cases:

dr2
(i) If x=acos0, y=bsin8.
(ii) If xa(0+sin0), ya(t—cos0). LC. P. 1988 1
(1,) If x= cos : and y= log r, then prove that at
d2y ( dy)2
0. (J.E.E.1985J
dx

9. If x=f(r), y=4t), then

d2 y - X1 V2 - y1x,
3 [VP.1998j

where suffixes denote differentiations with respect toe.

10. If xsinO +ycosQ =aand xcos0 —ysinO = bthen provethat

oq doq dO" is a constant.

dO" d

d" (
Show that X2I
'1 .
;---iJ = (x2 +1)'

J (n+1)x' _1
#i+1
3 )
+1
n+I

5
)X.- -d
 254 DIFFERENTIAL CALCULUS

12. If y=sin mx,showthat

Y Yi Y2
Yl Y4 Y5 =0.
Y6 Y7 Y8

where suffixs of y denote the order of differentiations of y with respect

to X.

ANSWERS
1.(i) (-ly m (in-1)(rn-2)...Jf ll - n +I)b(a-bx)

(ii) (-I)"m(m - 1) (m - 2).... (in - a +1)a'

U, +,,
(a + bx)

(-1)" pa" (n - I)!

(iv)
(a-i)"" - (ax +b)"
L3.5 ... (2n - 3)
(v) (a - 1)! + - (vi) (-1)"
[(a_x)' (a+i)"J
2"x 2

(vii) (-1)"- 1.3.5...(2n-l)

(viii) (-1) 3"• a!
p1*-.
2"x 2
1 1
() (-I)'(n - I)! + . (x) 10 -2) .(1og 10)
Li' (a+x)'J

(xi) (-I)''ab''.n! a!
(xii)
(a + bx)"" (a +x)"

(xiii) (x (xiv) - 2' cos[n7t + 2x)

- 1)' '

(xv) 21 2 j
} (2
(i
(xvi) cosi -nr+3x

I
(2 )

(xvii) - 22 cOS(! n + 4x

n.
(xviii) !i4 sin nn + 4x+ 2sin l-n,r-i-2x
(I ) -6 SinI!ni + 6x
4 2 (2
)2
 SUCCESSIVE DIFFERENTIATION 255

- 1 "
(xix) 22 ?cos( x+ —n,rl
4)

() !eI2cos(x+!n_ 102 Cos (3x+n tan '3

2 1 t. 4)
i

(xxi) 5"e3 sin 4x

4X+fltafl -
3)

(xxii) eIl _52 cos (2x + n tan 2) .

2

2. (j ) 21x (ii) —eu'' cos x. sin x. (sin x + 3)

(l+6x+6x)e (I+2x)
(iii) (iv) -

(_.1)"n!J 1
3. (•i)
2a

1)'n!sin" 0 sin (n + 1)0 '

4..2 "here
ereO tan --
- x
(iii) (- i)'' (a - I)! sin' 9 sin nO
_________ , a
where = tan -
a' x
cos (iv)
(a (1),,.s,n''
+ 0 1)0, where S = tan -- a
a. x
(-1) n! ( 1 1 2
(v) - — sin "9 sin (n+ l)Ot
4a 7 (x—a)" ( x+ a)'' a''

where x=acotO.

(vi) (—l)'.2"'2n! sin'" U sin(n + 1)0. whereO = tan'{j

I(2x +

(-1)" n! 0 sin(n +1)0 sin" 4) sin (a +1) i1

(vu) ---
—i-
a _b21 b n*2 a„+2

(viii) (-1)"n !! sin”' 0 sin (n+1)0, where cot = x + -'

4 2
16 I
(ix) (- In! { , when u> .
(x— ) (x-l)
256 DIFFERENTIA L CA LCULUS

(x) (_i)"n J 16 5
(x-2r' + (x-3)'J
5

n+I 3 4
(xi) (—iY'ni - , +1
(x +1)2 (5 + ( x+2r+ll

(_Irfl! r
(xu) ________
(a - b) (x-a)
02
(x
b2

+ a) I
_1

(xiii)(-1)(n - 1)!sin 0 sin nO • where cot 0 = x.

(xiv) 2(-1)'(n - I)!sin 0 sin nO, where cotO = x.

(xv)! (n - 1)! sin , 0 sin nO ,where cot 0

(n — I)! sin' 0 sin no Where 0 = cot

(xvi) —" an a

(I) 2(1+?) (iii) -

7
sin' y y

8. (a) (I) —4-coseco (ii) -seC 2.

4a 2

'eibnitz's Theorem*. (n-th derivative of the product of two

functions)
If u and v are two functions ofx, each possessing derivatives upto
nth order, then the nth derivative of their product, i.e.,
(uv) = uv + c 1 u ,_1 v 1 + c 2 u _ 2 v 2 + . . ' cu ,, v, + .
. + uv,

where the suffixes it and v denote the order of differentiations of u

and v with respect to x.
Let y=uv.
By actual differentiation, we have

Yt ='i" + UV1,
Y2 = I2V + 2U I V 1 + UV2 = U2V + 2 c 1 u 1 v 1 +

Y 3 =U 3 V + 3u2v1 +3uv2 +UV 31 -

, U3V + 3 C1U2V 1 + 3C2U1V 2+MV 31

The theorem is thus seen to be true when a = 2 and 3.
*Leibnitz (1646-1716) was a German mathematician, who invented Calculus
in Germany, as Newton did in Englançl.
SUCCESSIVE DIFFERENTIA TION 257

Let us assume, therefore, that

Y. u.,v+" C1 u,,_ i v 1 +' C2U,._2V¼+....+" +....+UV,,,

where a has any particular value.

differentiating,

- =++( +(2
+n
1) _ 12 +....

+ (n cr +" Cr_I +

Since Cr +" c_ 1Cr and "c 1 +1 ='- c1

yn+1 =u n+I v+ 1 C1 U,V j C2 U,_ 1 V 2 +....

+ n+l
Cr U n_r+l V r +....+UV ,
Thus, if the theorem holds for a differentiations, it also holds for n+ 1.
But it is proved to hold for 2 and 3 differentiations; hence it holds for four,
and so on, and thus the therem is true for every positive integral value of a.

8.6. Important results of symbolic operation.

d
If F(D) be any rational integral algebraic function of D or dx
(the
symbolic operator), i.e., if

F(D) = A D -I- A,, 1 D' + .......+ A I D i-A

= I A,D'. where A , is independent of D, then
(i) F(D)e'' = F(a)e.
(ii) F(D) eV = eF(D + a) V , V being afunction of x.
in (ax+b) = F (_a2) Jsin (ax+b)
(iii) F(D2)
cos
Is (ax +b) Icos (ax+b)
Proof:
(i) Sine D'e' =a'e,

F(D)eV A rD'e

=A ,a'e"

A ,. D'

= F(a)e
17-
258 0 DIFFERENTIA L CA LCULUS

(ii) Let y = ?V . Since W e- = a'e,

by Leibnitz's Theorem, we have
y, =e(aV + c1 a''DV + 'c2 a 2 D2 V + ...+ DV )
which by analogy with the Binomial theorem may be written as
D" (eV )=e'(D+aY V ,

F(D)e"V = (A D' )e''V

=A rDrC=V

=e"'>A ,(D+a )v
=e'F(D+a)V .

(iii) We have D sin (ax+b) = a cos (ax +b),andsoon

D 2 sin (ax+b) = (_ a 2 )sin (ax + b);

21 sin (ax +b) =(- 2) sin (ax +b).
D
Hence, as in (I) and (ii), it follows that
2
F (D )in(ax+b) = F (_a2)sin (ax+b)

Similarly, F(D 2 )cos(ax+b) =F(_a2)cos (ax+b)

8.7 Illustrative Examples

Ex. 1. if y=ex 3, find y,,.
Let u=e, v=x 3 . Now, u—a'e".
by Leibnitz's Theorem,

y =a'e''x 3 +n.a e'.3x +

3!
= + 3na2x2 + 3n(n - 1)ax + n(n - 1)(n - 2)}
Ex. 2. if y = a cos(log x) + b sin (log x) ,show that x2y2 +y + y =0.
Differentiating,
1
Yi =—a sin (log x).- +b cos (log x). 1
X x'
SUCCESSIVE DIFFERENTIATION 259

xy 1 = — a sin (log x) + b cos(log x).

Differentiating again,

Y2 +y = _ acosogx).!_bsinogx).!,

x 2 y 2 +xy 1 = — (acos logx+bsinlogx) = — y,

x2y2+xy1+v=0.
Note. This is called the differential equation formed from the above
equation.
Ex. 3. Differentiate ii fillies the equation
(l+x2)y2+(2x)y0.
By Leibni&z's Theorem,
d 2 2 n(n-1)
)1 = y, 2 (I+x .2x+ y,,2
21

— (y2(2x— l)} = y1(2x-1)+n.y,, .2

dx
adding,
(I+x 2 )y,, 2 + { 2(n +1)x-1)y,,.,. +n(n+l)y,, =0.

Ex. 4. Find the value of y for = 0, when y = e' . (C. P. 2004 J

From the value ofy, when x = 0, y = 1

,, Hem Y i = e' .a . (I)

=ay

y12(1_x2)=a2y2.

Differentiating, 2y!y2(I - 2)+ y1 2 (- 2x) = 2a2yy1,

or, (1_x 2 )y 2 _xy i _ a 2 y0 .... (2)

Differentiating this n times by Leibnitz's Theorem as in Ex. 3, we
easily get (i - 2) - (2n + l)iy,,. - (n 2 + a2)y, = 0.

Puttingx. = 0, (Y -2)0= 2 + a 2 )(,, ) .. ( 3)

260 . DIFFERENTIAL CALCULUS

Replacing n by n —2. we get, similarly

o{(n-2)2+a2J(y2)o

=((n_2)2+a2Jn_4)2+a2}(y,4)0

Also from (l) and (2), (Y )o a, (Y2 )o —a

Thus (y. .ia 2 }n_4) 2 +a2}

2 ' )(2'+a)a2
V 'if niseven
4+a
(4

and {(n._2) 2 +a 2 Jn_4) 2 +a2I.....

(32+a2)(12+a2)a2ifnisodd.

Note. The value of v for = 0 is shortly denoted by (y,)0

8.12 Miscellaneous Worked Out Examples

Ex. 1. (1) If F(x) = f (x) and f'(x) 4'(x) = k, (k is a constant),

then show that
fL_+_+.L..
F f
(F(x)*O)

(it) If x = sin i, y = sin k[ where k is a constant, show that

I_ x2 ^LY _ +k 2Y = 0.
th 2 &
Sohitiôn: (i) F(x) = f(x) 4(x),
F'(x) = f(x) 4(x) + f(x) f'(x) .. (1)
Differentiating (I) w.r.t. x,
F"(x) = f"(x) (x) + f '(x) 4'(x) + f'(x) 4'(x) + f '(x) "(x)
= f"(x) (x) + f(x) "(x) + 2k
F(x) = f(x) (x) * 0, dividing the left-hand side by
F (x) and the right-hand side by fix) ii(x), we get
F'(x) f"(x)$"(x) 2k
F(x) f(x) •(x) f('-)(x)
F" f " 4" 2k
F f 0
 SUCCESSIVE DIFFERENTIATION 261

dx
(ii) x . S1fll, —=cost
dt
and y..sinkt, =kcoskl
dt
dydydxkcOskf
dx di di cost
d2 v d (dy) d (kcoskI' di
and

cosi(_k 2 sin kt)–kcoskl(–sinf) • L

• -
- Cos 2 1 cost

• {_0cosisinki+ksinicoskt}

(i_ sin 2 t)cost

1 1 2 (k cos kt
= --t–k sinkt+sinil
1–x2t
t. COSt

1 1 2 dy
=— . t--k y+x—
I.x 2 1dx

X2 d2y d
or, )__j-_x j+k2y=O

Ex. 2. If y" +y" =2x, prove that (x2 —

l)Y 2 +xy 1 –m 2 O•,

dv d
2
where, v 1 =—, y2=—.
dx dx2
I I

Solution: •.• =2x, a 2 -2ax+1=O, where a=y

a= 2X_ =±1/ri

or, y=x±'J
Taking logarithm of both the sides,
•.iogy=iog(x±JTi)
 262 DIFFERENTIA L CA LCULUS

Differentiaing both the sides w. r. t. x,

1i ______
11± x 2< where Yi
dy
m y dx
(± J T ) j

or,iL=± I
nky

or, (2 - =

Differentiating 3n w.r.t.
2y
(2 d
–1)2yt y2 + 2x•y , = nr, 2 Y Yi' where Y2 =

or, (x2_I)y2+.ryi_m2y=0,
2v1#0.

Ex-3. 0) If x=a(0-t. sin0), y=a(I–cos0) verify that

[C. F 1988 J
d 4a 2
d2y
0i) If x=2cos8–cos29. y =2 sin 0+ sin 29 find at
C. P. 1990]
—d.v
Solution : (i) x=a(8+sinO), =a(l-t-cosO)

dy
and y=a(I–coso), —=asin9
dO

.0 0
d dv efr SIfl0
2sjn–cos–
0
- an 2.
dx- I/O A- a(1+cosO) .2cos2
2
d2 dy( 0\ d ( O dO
= — 1 tan– I = — i tan.–
fr 2 dx' 2) dO . 2) dx
I ,O
=-----
sec ----- I
2 2 a(1+cosO)

1 c40
=— . -
4a 2
 263
SUCCESSIVE DIFFERENTIATION

dx
(ii) x=2cosO—c0S20, =2sin2O-2sinO

and . y2sinO+stn20, rr2cs0+2COS20

dO
30 0
2(cos2O + cosO) 2cos--cos
dx dO dx 2(sin2O-sinO) — 2cossin
2 2
dy 0
- = Cot—
Jr 2
d 2 y dl 0'\ dl 0'\dO
= - cot — 1=—I cot-
dx 2 dx ' 2) d0'. 2) dx
2 0
— cosec

- 2.2(sin0-sin0)

-cosec
7t ________j
at 0= 2
4(sin,—sin)

Ex. 4. (i) If sin x=-- coty=!.'_,find the value of

21 dy2
I C. P. 1987 1

(ii) If y = -f-- , show that y, (0) = 5! 1 C. p 1991, 2003 1

= 210.
(iii) If y=2cosx(sinx-COSx)phOW that (yio)o
C. P.'1992, 2000, 2002, 2005, 2007 1

Solution: (1) •.' = __!,

1+t.
. 2i
X=S'fl =2tan(t)
2i
coty=_. -_, tanY-----1

or, y=tan'-j----2taflt(t) (2)

 264
DIFFERENTIA L CA LCULUS

From (1) and (2) xy

dx
— =land d2x
-=
dy
x x+1-1
(I') Y = - =1-(x +1)
x+1 x+1

Y i = ( - 0(-J)(x+ 1)2

Y 2 = (l)(-l)(-2)(x+ 1) -a = (-1)32!(x+ i)-

= (-1)' 3!(x +
Proceeding in this way, we have
y5=(-1)65!(x+1)-6 -.

(Y s)o=(1)65!(1)=5!
(iii) y = 2. cos x(sin x - cos x)

2sinx cos x-2cos' x

= si n 2x - (I + cos 2x)
= sin2x- cos 2x-1

lO .71
Y to _ sin{1O + 2x}_2b0 cos{10 .! +2x}

vide art. 52 (v)

If y = sin ax, y = a" - sin(n . +x and

).

if y = cosax, y a" -cos (n+x)

= 2I0.sin57_2I0cos5i =2'°.O-2'°x(-l)= 210.

EL S. If , y = /T2 I x J <1 . show that

x2)y2_3xy1_y0

00 (i_x2)y+2 -(2n+3)xy,,1 -(n+1) 2 y,, =0 C. P 1993

StJCCESSIVEDIFFERENI'IATION 265

sin-'x
Solution: y=

(I_x2)y2 = (sin' x)2

Differentiating w.r.t. .
(i_x2)2yy1-2xy2= X=2y

or, (l_x2)yi_.y1 (... 2y 0)

Differentiating again w. r. t. t,
I- 2)2_ 2xy 1
• -y- xy 1 = 0

(1_x2)y2_3xy1_y=o ... (1)

Differentiating (1) n times by Leibnitzs theorem,
y,2(l_x2)+n.y+(_2x)+ -I) : (-2)
1.2
-3 •x+n -y - Y.=
or, (1_x2)y,,+2 -(2n+3)y, 1 7 n+1) 2 y =

EL 6. If y cos(lOcos x), show that (i - x2)y12 = 2Ly11

I C.P. 19831
Solution: y = cos(lococ'x) • .
(-10) IOsin(IOcos' x)
.v1 =-sin(IOcos'x).

(l_x2 )y = lOOsin2(iocos_1 x)

Ioo{I —Cos 2(Iocosix)}

r100(1_y2)
from (1).]
I
Differentiating again w.,t. x
(1_x2)2y1y2_2x.y2 =-2.100y.y1

0_x2)y2=.xy_1ooy
(... 2y 1 O) .. (2)
266 DIFFERENTIAL CALCULUS

Differentiating (2) 10 times with the help of Leibnitz's theorem,

10-9.
(1_x2
)Y2 +10.ii(-2t)+-1.2
-----io(-2)

X)' + 10 . Yto .(i) IOOYJO

, (i_x) =21XYii

Ex. 7. (i) If f(x) = x", prove that

-C(l)f "(l) f "(l) f(1)
f (1)+— +---+
1! 2! 3!
(ii) If f(x) =tan x and n is a positive integer, prove with the
help of Leibnitzs theorem that

L CF. 19921

Solution : (i) .• f(x) =

f'(x) = nx"* f"(x) = n(n — 1)x"2

• f (x) = n(n-1)(n-2).(n-r+ 1) x'', ••, f'(x) =

I! 2! 3!
= , + n + n(n — 1) + n(n — 1)(n -2) +...+-
n!
1! 2! 3!
i+c, +c2 +c + ....+c,, = (1 + =
sinx
f(x)=tanx=
cos x
f(x) . cosx = sinx ... . (1)
Applying Leibnitz's theorem to differentiate both the sides n times
w.r.t. x, we get
f" (x)cosx+"C1 f (x )(- sin x)+"C2f'2(x)(- Cos x)

+ C3f'"(x)(sinx)+"C4f" 4 (x)(cosx) .. = sin(n +

Putting x = 0 on both the sides,

f (0)-C2 f 2 (0)+C4 f 4 (0) + ......=

SUCCESS! VS DIFFERENT/A TION 267

EL 8. If y = cos(msin' x), show that

(I) (1_x2)y2_.yi +m2y=O

(ii) (1_x 2 )y,+2 -(2n+ 1).y, +1 +(m2 _n2)y= 0

Also, find the value of y,, when x=O. B. P19991

Solution: Given, y = cos(msin x) ()

Differentiating w.r:. x,
dy
— -1 ____
= YI = _ msin(nisin t):
dx

or, (i - x2)y12 = m2 sin2(nisin x) = rn 2 {i _cos2(,nsin_I )}

from
or, (i_x 2 )y =m 2 (J_y 2 ) (1) j . (2)
Differentiating again wr.t. x
(1_x2)2y, Y2 +y? ( -2x ) = -m2 23y

or, (i_x2)y2 .xy 1 +m2 y=0, 2y0) .. (3)

Differentiating (3) n times by Leibnitz's theorem, we get
(t - x 2 )v,,,2 +"C1 y,, 1 (-2x)+"C,y,, (2) - X -" 1 y,, ( I) + m 2 y ,, 0
or, (1_x2)y,,+2 - 2n xy,,. 1 -n(n- l ) y, -xy 1 - ny + m 2 y,, = 0

or, (i_x 2 ) y ,2 -(2n + +(m2 - n 2 =0 . . . (4)

Last part: From (1), (2), (3), we have y=l, Yi =0, Y2 = -m2 , when
x=0.
putting n = 1, 2, 3 successively in (4), we get

y3 =-m2 y, =-m2xO=0

.}4 =(2.2 _m2 )y2 =_m2(22_m2)

y=(32_m2)y=0

,_ (42 _m 2 )y4 = _,nl(22 _m2)

_m2X42
Thus, y,, =0, when n is odd and.
m2) ... .{(n_.2)2_m2},
Y " = _ n,2(22 when n is even.
_m2X42
268 DIFFERENTIAL CALCULUS

Ex. 9. If y = '° . show that an equation connecting y, Y*i and

2 )y,,+2 —(2n+ l).ty,1 _(2 + i)y., = 0
Y..+2 is given by (1_x
C. P. 1980, B.P1996

Solution: y= (1)

Yi = x , .._.L.. using (I)

= - ,Jj2
or, (i_x 2 )y_y 2 =0 ... C2)

Differentiating (2) again w.r.i. x.

(l_x 2 )2y 1 y2 +y(-.2x)-2y.y1 =0

or. (i_x 2 )y2 _y 1 _y=O. . . (3)

Differentiating (3) n times with the help of Leibnitzs theorem, we have,
(i _x2)y,,,2+C1y,4j(_2x)+c2y,(_2)_.zy,,1
—"Cpy,,(1)—y, =0
or, (i_x2)y,,+2 —(2n+1).*y. _(2
+l)y =0.

Ex. 10. If x = sin 0, y = sin p0, then prove that

(0 (l_x2)y2_xvi+p2yr0

)Y. 0
(ii) ( 1 _x 2 )yn+2 —(2n+l)y,,+1 +(p2
where y, denotes the order derivative of y with respect to X.
nth
F C. P 20021
Solution: Here, x = sinO and y = sin p0
Proceeding exactly as in EL 1. (ii), we have
(1_x 2 ) y2 +p2y=0 ... (1)

Differentiating (1) n times with the help of Leibnitz's theorem, we

get,.
n(n-1)
(1_x 2 )y2+n.y+1(-2x)+
1•2
—n.1-y, +p2y., =0
or, (t_x2)y+2 -(2n+I).ty,,+1 + ( p2 - 2 )y. =0.
 SUCCESSIVE DIFFERENTIATION 269

EXAMPLES - VIII (B)

I. Find y in the following cases:
(1) (ii) y=x 3 sin x..(dli)y=x 3 log x.
(iv)y=x2 tan -'x. (v) ye cos bx. (vi) y =log zx+x2)
(vu)y=r(l —x). (viii)y=xi(l+x).
2. If y = A sin mx+ B cos nix, prove that Y2 + m 2 y = 0.

• 3. if y =A e+B g , Prove that y 2 — m 2 y= 0.

• 4. If y = e sin bx, show that y2 - 2ay 1 +(a2 +b 2 )y = o.

t
.5. If y = lo(x + 4a 2 + x2) ,show tha (,2+ X, ) 2 + xy1
y =

6. If y log(x+.Jl+x2) ,thenprovethat

(t + x2 ) y 2 + X y, - M 2 = 0.
If y = tan- ' x, then prove that
(i) (i+x2)y1=1, and

(ii) (i+2)y, +2iay, +h(n— l)y 1 0.

Find also the value of (y),,
If y = sin' .x, then show that
(i) (i_x2)y2_xy, =0,

(u) (I_x2 )Y,2 —(2n+1 0. [c. P 19971

Find also the value of (y ) ,.
)2 ,
,-9' If Y = sin then show that
(i) (i_x2)y2_xy1_2=0, [C.P.1988,95,961

(fi)
,4. If log y = tan' x, then prove that
(1) (1 +x2)y2+(2X_I)y1 =0.

(ii) (i ± x2)y,2 + (2nx + 2x— 1)y,,, 1 + + 1) y,, = 0.

270 DIFFERENTIA L CA LCULUS

11. If y = a'cos (log x) + b sin (log x), then prove that.

x 2 y,, 2 + (2n + +( 2
+ ii,, = 0.
C.? 1989, 96,2007 BY 1990,97, V P 19991

If y = (x - I then show that

(2 — l)v,, 2 +2xy,, -n(n+1)y,, =0.

13. If v = ell—' ', then prove that

(I - x2)y2 - (2n + 1)xy,,, - (, 2 + a)y,, 0.

[C.? 1985, 94, 98, 20041
44. If v = sin rn sin -' ), then show that

x 2 )v , -xy 1 + In y = 0, [C.? 1990, 2002. 2008]

x2 )Y +? - On + I)xy 1 + (m = o.

15. If y = ( +bx +c)/(l_x), show that (I-x)v 3 = 3y2.

16. If y = ecosx.provethat y4 +4y = 0. B. P. 1998, 20011

17. If y= x" Iogx,show that y,,

[C.P.1985. 2000,8.? 19911

18. Show that

19. If a, p, w be functions of x and if suffixes denote differentiations with
respect to x, prove that

U 1 V1 W1 U V1 W1
d
_dU2 V
2 W2 = U2 V2 W 2.

U3 V3 W3 U4 V 4 W4

20. By forming in two different ways the nth derivatives of x1', show that

i+t+n2(n_I)2 fl 2 (fl _ 1)2 (fl _ 2)2 (2n)!

7 12.22 12.22.32 (n!)2

[Equate the nth derivative of the product x. x" to that of x 2 " .]

SUCCESS!VED/FFEREIVTIAT1ON 271

21. Prove that i_(!)= {Psir{x+ 'J J+ Qcos(x+L!J}/x'

where P=x-n(n-l)x'2+n(n-j)(n-2)(n-3)x"-
and Q=nx' -n(n-1)(n-2)x"3 + ....
22. Prove that

I' nir
Pcosl x+ -- 2)Qsin(x+)}/x
I
)
where Pand Q have the same values as in Ex. 21.
23. Prove that

9=
,r I (p sin x + Q cos 4
(n)x2 (n4
where
Pl2!+Iuj

and Q = x
x-I —+
(") 3)3! 5)5!
24. Show that
d" (logx '\ 1 1
------(-l)----i-1
n! (Logx -----
dx X ) 2 3 ii]
25. Show that
"
d"
— (e- x -1.
Ir " +) (n — r)!' a>
r=O(

26. If f(x) = tan x, prove that

j(o) — c 2 f 2 (o)
+"
c4 f " 4 (o) sin —
2

27. Show that the nh differential coefficient of 1

is

- (- 1)"n!sjn'19 { sin(n+ 1)0-cos(n+ I)O+(sit+coJ"}.

where 0 = cot' x.•

272 DIFFERENTIAL CALCULUS

ANSWERS
e.a_2{a2x2 + 2nax+ n(n_I)}

(ii) x3 sin( i mr+x)+ 3nxsin{..(n_t)+x}+3n(n_1)x

x sin{(n_2)m+x}+n(n_1)(n_2in{(n_3)+x}.

(iii) (- I)°.6(n -4) !L1.

(iv) (- I)" (n - 3)!sin" 2 6 {(n - 1)(n - 2)sin n cos 2 6

- 2)sin (I - i)UcosO +n(n - 1)sin(n - 2)0} where cot = x.

(v) cos bx +° c1 a"1, cos(bx + -!-

+?c,a2b2cosIbx+2.!+...+b cos(bx+n.!fl
- 2
2)

(- 1)ni + I
(vi) -
Lx (x+a)

(vii) n (;— 4° - (" )2(j - x)''. + (C2 )2.i x)"2.x2

(viii)n! /(I +
7. 0, or (- i) " - I) L according as n is even or odd.

8. 0, or {1.3.....(0 _2)}2 according as n is even or odd.

 EXPANSION OF FUNCTIONS
IL]I
9.1. RoIle's Theorem.
If (i) f( x ) is continuous in the closed interval a 5 x <— b -
(ii) f'(x ) exists in the open interval a < x < b
and (iii)f(a)=f(b),
then there exists at least one value of x ( say ) between a and b
[ie., a< <bj, such that f'( )=o
Sincef(a) = f(b) ,if f(x) be constant throughout the interval[a, bJ,
being equal to 1(a) or f(b) ,then evidently f'(x) = 0 at every point in
the interval.
If f(x) be not constant throughout, then it must have values either
greater than or less than f(a) or both, in the interval. Supposef(x) has
values greater than f(a). Now, since f(x) is continuous in the interval,
it must be bounded and M being its upper bound [which is> f(a) in this
case], there must be a value of x in the interval a <x < b for which
f()=M

f (. + ii)— f ( ):^ 0 for positive as well as negative values of h.

( )< 0
- i f h be positive, and ^:O if h be negative.
/

Hence Li Li
f(+h)—f( )>
h h
provided the limits exist.
Now, since f'(
x ) exists for every value ofx in a <x <b
also exists, and so the above two limits must both exist and be equal,
and the only equal value they can have is zero. Hence )=o
f'(
If f(x) has values less than f(a) in the interval, we can similarly
show that f'( , )= 0, where f( )='n, the lower bound of f(x) in
the interval.
f
If (x ) has values both greater than and less than f (a).
then there
f
must be an upper bound M which is greater than (a) and a lower bound
in which is less than (a )and for values ofx toç which either (x ) = M
f f
or f(x )= ni. f'( x ) will be zero.
18
274 DIFFERENTIA L CA LCULUS

9.2 Geometrical interpretation of Rolle's Theorem

rig ,.z.t

Let LM be the points on the number axis OX representing the real

numbers a, b respectively. We draw the graph of the function y = f (x ) and
let A, B be the points on it corresponding to L, M respectively, that is,
L4=f(a) and MB=f(b).
From the condition (i) of Rolle's theorem, we say that the graph is a
Continuous curve between the points A and B: the condition (ii) says
that the curve has tangents at every point between A and B and the third
condition implies that LA =MR.

I
Now, f'( ) is the gradient of the tangent to the curve at x = By

Rolle's theorem (x) vanishes at least once between x = a and x = b.

Geometrictlly we say that we get at least one point C on the graph between

A and B such that the tangent at C is parallel to OX

Note. From the above graph, it is clear that there are more points D and
like C.
EXPANSION OF FUNCTION 275

9.3. Mean Value Theorem. [Lagrange 'S form I

If (i) f(x) is Continuous in the closed interval a !^ x 5- b, and
(i) f'(x) exists in the open interval a< x < b, then there is at
least one value of x (say ) between a and b [i.e., a < 4 < b ], such
that
f(b)-f(a) = (b-a)f'().
Consider the function (x) defined in (a, b) by

(x) = f(b) - f(x) - {f(b) - f (. )}

b-a
Here '(x) is Continuous in a x :S b, since f ( v) and b- x
are so,

f b)-f (a)
O x) = -f'(x)+ exists in a < x < b, since

f (4 exists in a < x < b

Also, y(a)= 0, t,iI(b)= 0, .. 41(a) = '(b),

Hence, by Rolle's Theorem, ti' (9 vanishes for at least one value
of (say ) between a and b, i.e., ui" () = 0,

i.e., 0f (çf (b)f (a)

b-a
Whence f(b)--f(a)=(b-o)f'(4), [a<<b].

Cor. Since 4 lies between a and b, 4 can be written as a+O(b-a)

where 0<0 < I. Putting b= a + h we get another form of the Mean
Value Theorem
f(a+ii)= f(a)+hf'(a -i-oh), where 0<0<1,

Or, f(x-i-h)= r(x)+hr(x+oh), where .c9<i.

Note. The value of 0 usually depends upon both x and h, but there
are cases where it is not so dependent. [See. Ex. 2 and 13, Examples IX
(A) J. Also 0 may have more than one value in a given range in some
cases. [ See Ex. 12, Examples IX (A )].
276 D/EILRENT/AL CALCULUS

9.4. Geometrical interpretation of Mean Value Theorem.

Let ACB be the graph of f (x) in the interval [a, b] and let
a, , b be the abscissae of th' "flints A , C, B on the curve y = 1(x),
such that the relation f(b)— f a) = (b — a) f'(.) is satisfied.
Draw AL, BM perpendiculars on OX and AB, perpendicular on

BM Then A L = f(a), BM = 1(b). Let CT, he the tangent at C.

Then, f(a) = BM - AL BN = tan BAN

b-a LM AN

Since f '() = tan CTX (as explained in § 7.14), it fol!ows from the
Mean Value Theorem that tan BA N= tan CTX, i.e., ,nLBA N = mLCTX i.e.,
AB is parallel to CT.
Hence, we have the following geometrical interpretation of the Mean
Value Theorem
If the graph ACB of .1 x) is a continuous curve having everywhere a
tangent, then there must be at least on poi nt C intermediate between A
and B at which the tangent is perallel to the chord AB

9.5. Taylor's Series in finite form. (Generalized Mean Value Theorem)

If f (x) possesses diffcrential coefficients of the first - 1) orders

for every value of x in the closed interval a x5 b and the nth

derivative of exists in the open interval a< x < b [ i.e., if (x) is

continuous in a S x !^ b and f " (4 exists in a < x < b ], then
EXPANSION OF FUNCTION 277

(b-a)2
1(b) = f(a)+ (b - a)f'(a)+ f(a)+ ......
2!

+(
b - a)
I)!
(n-
I
'(a)+
(b-
f"(,),

where a<<b ... (A)

and If b = a + h, So that b - a = h, then

h2
f(a + h) = f(a)+ hf'(a)+ —f(a)+ .......

h"-1I?"
+ f "'(a)+— f '(a+6h),
(n-i)!
where 0 < 0 < 1 .. (B)
or writing x for a,

li2
f(x + h) = f(x) + lif'(x)+ f(x) +
2!
It
+
where 0<6<1 ... (
Consider the function 1/1(x) defined in f a, b] by
(b-xY'
W(x) =4(x) - a)" 0(a), ...
(b -
where, Ø(x) = 1(b)- 1(x)- (b - x)f'(x)- f(x) -
2!
- (b -
... (2)
(n -I)! f "(x)
Then, evidently V (a) = 0 and i(b) = 0 [since 0(b) .0 is
identically],
Now,
O'(x) = - f'(4 + U'(x) - (b - x)f'(4} +f(b-x)f(x)- f(x)} +...
2!

(b - x r2 (b -

+{ (n-2)! (n-I)! f(x)}

(b - x) "
- f"(x)
-- (n-i)!
278 DIFFERENTIAL CALCULUS

Hence, from (1)

x)
41'(x) = f.(x)+ n 0(a), ... (3)

Since i(a) = iji(b), and t'(x)exists in( a, b), by Rolle's Theorem,

,'()=o, where a< <b.

Substituting for x in (3), and cancelling the common factor

(b -r' we get ultimately

j" (c), and

0 (a) = - (b—a)" since, from (2)

(b - a)
0(a) = f(b) - f(a) - (b - a)f'(a) - f - '(), the
(n—i)!
required result in the form (A) follows by transposition.
Since a < <b,wecan write = a + (b - a)0,

i.e., 4 =a+h0, where . 0<0 <l,and b — a=Is,andhencethe

form (B) follows and writing x for a in the form (B), the form (C) can be
obtained.

Note!. The series (A), (B) or (C) is called Taylor's series with the
rem'der in Lagrange's form, the remainder (after n terms) being

(b—ar t'(), or —fa -fob), or !_-f'(x+Oh), 0<0 <I, which is

generally denoted by R.

Note 2. Putting n = I in Taylor's series, we get

f(a+h)=f(a)+hf'(a+Oh), 0<0<1,
which is the Mean Value Theorem.
So, Taylor's theorem is sometimes called Mean Value Theorem of
the nth order.
Putting n= 2 in Taylor's series, we get

-f"(a+l) 0<6<1,
f(a + h) = f(a) + hf'(a) + h2

which is often called the Mean Value Theorem of the second order and
so on.
Note 3. Yet another form of Taylor's series which is found some times
useful is obtained by putting x for b in (A). Thus,
EXPANSION OFFUNCTION 279

(x-a)
f(x)= f(a)+(x-a)f'(a)+ (x_a)2 f(a)+....+ f(a)
2! (n-I)!

and the function 1(x) is said to be expanded about or in . the

neighbourhoodof x = a.
9.6. Maclaurjn's series in finite form.
Putting x = 0, It = X 141 Taylor's series in finite form (C), we get
X
f"-t(o)+....fn(ox) 0<0<1
2! (n-i)! n!
the corresponding form of the remainder R being

The above is known as Maclaurin's series for f(x), and f(x) is said
to be expanded in the neighbourhood of x = 0.
Note. Putting ii = 1, 2, we get Maclaurin'sseries of the first and second
orders, viz.,

f(x)=f(o)+xj'(ox) andf(x)= f(0)+xf'(o)+f(ox), 0<0<1.

9.7. Cauchy's series in finite form.

In Art. 9.5, if we take V (x) = Øx)- (a), the other conditions

b-a
remaining the same, and carry out the investigation as in that Art.,
we get

(4)

Since v (a) = v (b)and i'(x) exists in (a,b)byRolle's theorem, we

tver'( 4 )= 0, a < < b.
Substituting for in (4), we get
= (b - a)(b )t
r (4 ... (5)
Writing=a+(b-a)0,where 0<0<1,
280 DII- F/RENT/A L CA LCULUS

wehaveb—=b—a--bO +aO =(l —0)(h—a).

(b — )'= (I —O) 1 - 1 (b—a)" '=(l—O)''h'-',since

b—a=h.
from (5), we get

(a+Oh).

Now replacing a by x, the required expression for the remainder R.

would come out as

" / ,n(\X+
- / n— 1 )! ),0<0 <I.

This is known as Cauchycforiii of remainder in Taylor's expansion.

The corresponding form in Maclaurin's expansion is
"
R 0f(Ox), 0<0<1.

This form of remainder is sometimes more useful than that of

Lagrange's form. It should be noted that the value of 0 in the two forms
of the remainder for the same function need not be the same.

9.8 Illustrative Examples.

Ex. 1. (i) If f'(x)= 0 for all values of x in an interval, then f is
constant in that interval.

(ii) If 0 ' (x)= y, ' (x) in an interval, then O (x) and , ( 4 ditfrr
by a constant in that interval.

(i) Suppose, ['(s)= 0 at every point in (a. b).

Let us take any two points x 1 , x2 in [a, b 1 such that x, >x1,
By Mean Value Theorem,
f(x2 )—f(xj )=(x2 —x1 )f'(c wherex <e<x2
= 0, since f'( c ) = 0 , by hypothesis.
f(x)=f(x).

Since x, x, are any two points in a, b 1, it follows that f(x) must

be constant throughout [ a, b 1.

(ii) Let f(x)=Ø(x)-41(x).

f'(x )='(x )-41 ' ( x) = o,everywhere in (a, b).
EXPANSION OF FUNCTION 281

f (x) = constant = k, say, by (I).

Note: The result (ii) is fundamental in the theory of integration.

Ex. 2. 1ff (h)=f(0)+hf'(0)+f"(0 h), 0 < 0 < 1, find 8. when h=1

andf(x) x)2 CR 1944

W e havef(h) = (1— h), sincef(x) = (I — x).

515
f'(h)= _(l_h) ;f"(h)=— (l— h)

from the given relation

(1-12)2 = l--—
5h+ . --( l —0/1)2.
2 15 J7

4
-0/i),1 whence (1-0) =
5 15 0
putting /i = i,o= 1— +

25 25

Ex. 3. Prove that the Lagrange's remainder after n terms in the

expansion of C U in powers of x is

(a 2 +b 2 )" / b'
x'e' 0 Cos b0x + ii tan -), 0<0<1.1 C.P. 19421
n! a
Lagrange's remainder after n terms in the expansion of f(x) is

— f(Qx),0<0<1. (by A rt. 9.6.) (1)

Here, since f(x) = e" cos bx.

f(x)=(a2 +b2)2 e cos(bx+n tan 1... (2)

[See Ex. 38.4I

f1(O x), and substituting this
writing Ox for x in (2), we get
value of f'1 (0 x) in (1), the required remainder is obtained.
 282 DIFFERENTIAL CALCULUS

Ex. 4. Prove that the Cauchy 's remainder after n terms in the expansion'
of (i + x )" (ni being a negative integer or fraction) in powers of x is

o-zo<l:
(n—I)! I+Ox)

Cauchy's remainder after n terms in the expansion of 1(x) is

)I
X" (.1f(0x)fl 0<0<1 (bvArt. 9.7) -(I)

Here f(x)'=(l+x)'",

f" (x)=m(m—l)(ni-2) .-'•• (m—n+i)(i+x)"".

So, the expression (I) is equiva!en to

• (n—i)!
which is the required remainder.
Ex. 5. Show that the Cauchy's remainder after n terms in the
expansion of Jog( 1 + x) in powers of x is

0<0< 1.
I+Ox1+Ox)

Heref(x)=log(i+x), :.
(i+x)

Hence, x" (n1)! (Ox)(_' )"x" (i—o)'

(I+Ox)"
which is the required remainder in Cauchy's form.
Ex. 6. If (i) f'(x) exists in a x 5 b. (ii) f'(a )= a, f'(b )= 13.
and (iii) y lies between a and 13, then there exists a value of
x between a and b such that f'( )= y. [Darboux's Theorem 1
Suppose, a-zy<13 and let

Since •'(x) exists in (a, b), 4) (x) iS continuous in [a, b ] and

therefore attains its lower bound at some point k in the interval. [§ 4.4
(viii)]
EX PA NSION OF FUNCTION 283

• Now. this point cannot bi a orb. since 4)'(a )= f'(a )- y = a-y

which is negative and 4)'(b )= f'(b)-y=f3-y and which is positive.
Hence, the point is between a and b. and 4)' ( )=(). .%f'( )-y=0
f '()= 'y for a<t<b

Ex.7. (a)Jf (i) 4)(x) and w(x) are both continuous in 6:5x:5b
(ii) 0'(x) and '(x) exists in a<x<b.
and (iii) w'(x )* a anywhere in a
theti there is a value 'E of x between a and b for which

= (0) [Cauchy's Mean Value Theorem j

W(b)-W(a) 'v()
(b) If f urther 0(a)=w(a)=O and w'(x)* 0 in the
neighbourhood of a,

then Ii = Lt . if the latter limit exists.

c(a) (x)
• [L'Hospital'smeoreml
(a) Consider the function 1(x) defined by the equation

f(x)=4)(b )-4)(x)- $(b )-(a)

ji(b )-xV
Now, f (a)=.f (b), since each =0 identically.

A lso, f'(x)=-4)'(x)+

Since f (x ) satisfies all the conditions of Rolle's theorem

for some , where a <<b ,whence the required
result follows.
Note 1. The condition is'(x);tO anywhere in (a, b) ensures that
W (a )# Nc(b); for, if i(a ) = c(b), tI(x ) then satisfying all the
conditions of Rolle's theorem, w(x) would vanish at some point x in
(a, b).
Note 2. Putting b - a = h, we get
0(a+h)-4)(a)4)'(a+oh)
W (a+h)-it(a) s'(a+0h)'
284 DIFFERENTIAL CALCULUS

(x+h )-() - '(x+0h)

or. writing x for a, 0<0<1.
W(x+h)-W( x) W'(x+Oh)
Note 3. Mean value Theorem can be deduced from this theorem by
putting w(x)=x
(b) Wehave,for a<x<b.
•() c(x)-$(a)
W(X)_W(a) •(a)=iv(a)=0 here.
()

a<<x , byCauchy'stheorem.
'I' ()
Taking limits, and noting that - a as x -> a

We get Li Li Li
.- * o ;(x ) -r*ü j () . - *o ti (.x )
Again, when b 1 <x< a [assuming b1 sufficiently close to a such
that '(x) and iI'(x) exist at every point in the interval, and
x )# 0 in it], we may similarly write

(a ))-=R ,,
where x< <a 1 byCauchy's
J(a)-W(x) ii(E1)
Theorem, and making x-), a we get

Li Li Li
. - 0 4f (x) .-- ji'(E. ) . --o 141'(x)

Combining the two cases, L'Hospital's Theorem follows.

EXAMPLES - IX(A)
1. Find the value of in the Mean Value Theorem
f (b)-f (a)(b-a)f '()
(i) if f (x )=x 2 , a =1, b=2, -
(ii)if f (x ).i.a4,b9, [C.P.2006]

() if f(x)=x(x-l)(x-2),a0, b=,
[C.P. 1987, BY 19971
(iv )if f (x )=
A x ' +B x +C in[a,b].
EXPANSION OF FUNCTION 285

2. In the Mean Value Theorem

f(x+h)=f(x)+hf'(x+Ok),

if f (x ) = A x 2 + Bx + C, Where * 0, show that 0 =

Give a geometrical interpretation of the result.
3. In the Mean Value Theorem
f(a+h)=f(a)+hf'(a+Oh),
ifa=1,h=3 and f(x)..j, find 0.
4. (I) In the Mean Value Theorem
f(h)=f(o)+/if'(oIi), 0<0<1,

show that the limiting value of 0 as Ii - 0 + is - or

2
according as f (x ) is cosx or sin x. F C. P 1994, 2005]
(ii) In the Mean Value Theorem
f(x+h)=f(x)+hf'(x+Oh), 0<0<1,

show that the limiting value of 0 as Ii —> 0 + is whether .1 (x )

is sin or Cos X. F C. P 1994, 2008]

5. If f(h)= j(o)+ hf'(0)+ f'(Oh), 0<0 <1.find 0,

when /i=7 and f(x)=_±_.

l+x
6. From the relation

f(x)= f(0)+xf'(0)+f(0x), 0<6<1.

show that log (1+ x)> x — x 2 , ifx>0,

and cosx>x--.x2,if0<x<..,r.

7. Show that sin > x—x 3 , if0<x<7t. FV P 19951

8. Iff(x)=tanx, then f(0)=0 and f()=0.

Is Roile's theorem applicable to f (x ) in (0. ,r )?
[c.P. 1982, '86, '96, 2004]
9. Is the Mean Value Theorem applicable to the functions (i) and (ii) in
the intervals [-1,1] and [5,7] respectively?
286 DIFFERENTIA L CA LCULUS

c,) f (x)=xcos for x;t0

=0 for .x=0

(ii) f(x)= 4— (6— x). [C.P.20051

10. If f '(x ) exists and >Oeverywhere in the interval (a, b), then show
that f (x ) is an increasing function in a, b] and 1' ( x ) < 0
everywhere in (a, h), then show that f (x ) is decreasing function in
(a, b).
11. Show that 2X 3 + 2x 2 - lOx + 6 is positive if x>1.
12. (i) In the Mean Value Theorem
f (a+Ii)— f (a)=hf '(a+8h), 0<8<1,

If f(x)=x3— .-x2+2x, and a=O, h=3,

show that 0 has got two values and find them.
(ii) In the Mean Value Theorem
f (b)— f (a)=(b— )f '() a<<b,

find , if f(x)=.-3x—,a=4)-,b- and give a

geometrical interpretation of the result.
13. In the Mean Value Theorem
f (x + h ) = f (x ) + Ii f' ( 0h),

(i) find 0 where, (a) f (.)= i , (b) f (x ) = e

(c) f (x)=logx.

(ii) if f (x)=a+bx+cm
then show that 0 is independent of x.
14. Show that
3 .i 3 1 h3 I
,0<0<l
2 2 2 2L j+6h
Find 0, when x=0.
15. Expand in a finite series in powers of h, and find the remainder in each

7:
Cse:
log (x+ h), (ii) sin ( + ii), (iii) (x + h )".
 EXPANSION OF FUNCTION
287

16. (I) Apply. Taylor's Theorem to obtain the Binomial expansion of

(a + h )'. where n is a positive integer.

(ii) If f ( x) is a polynomial of degree r, then show that

f(a+h )= f( q )+hf '(a)+f (a)

+ .....+f ' (a)

(iii) Expand 5x 2 + 7x + 3 in powers of (x —2).

17. Expand the following functions in a finite series in powers of x, with
the remainder in Lagrange's form in each case:
(i) e', (ii) a', (iii) Sfl X,
(iV) Cos X, (v) log(l+x), (vi) log( 1-x),
(Vii) (I + x )", (viii) tan x, (ix) e' cos x,
(x) e'" sin bx.
W . Find the value of 0 in the Lagrange's form of remainder
R for the
expansion in powers of x.
I—x
19. Expand the following functions in the neighbourhood of x = 0 to three
terms plus remainder (in Lagrange's form):
(i) sin 2 x, (ii) cos 3 x, (iii)
20. Expand the following functions in a finite series in powers of x, with
the remainder in Cauchy's form in each case:
(i) e , (ii) cos x, (iii) (i -. x

21. (i) Prove that Li

2h
provided f ' (x ) is continuous.

(ii) Prove that tj f (a + h )- 2f (a )

= f' (a ),
h2
Provided f' ( x) is continuous.
288 V/Fr/RENT/AL CALCuLuS

22. (I) If f' ( x ) is continuous in the interval [a. a + h I and

f ' (x );t 0, prove that Lu U = .-, where 0 is given by
h -.0

f(a+h)f(a)+hf'(a+Oh). 0<0<1.

(ii) Show that the limit when ii —* a of 0 which occurs in Lagrange's

formofrrnainder f i
+ u/i ) n the expansionof f(x+h)

is ._±__ provided f'(x) is continuous and ;^ 0.

,i+l
23. In Cauchy's Mean Value Theorem,
(i) if O(x) = sin x and , (x ) = cos ..,or
(ii) ifO(x)=e' and VI(.v)=e,or
(iii) if O ( x ) = x ' + x + I and v (x ) = 2v 2 + 3x + 4, then

show that 0 is independent of both x and Ii, and is equal to

24. If f (x )= x', Ø(x )= x, then find a value of E in terms ofa and

b in Cauchy's Mean Value Theorem.

25. If f (x ) and 0 (x ) are continuous in a :^i. < b and differentiable

in a <x< b such that f '( x ) and 0' ( x ) never vanish for the same
value of x, then show that
f()—f(a) -

26. If l/I"(x)# 0 for a < x < b. then prove that

Ø(b)—Ø(a)—(b—a)O(a)
(b)—(a)—(b—a)W'(a)
o)
w(Y
where a < < b

27. If f ( x ) and g (x ) are differentiable in the interval (a, b) then

prove that there is a number , a < 4 < b, such that

f (a) f(b) f' ()

=(b—a)
g(a) g(b) g(a) g ' (4 )
EXPANSION OF FUNCTION 289

28. (1 ) If f (x), (x), ,j, ( x ) are continuous in a <x < b and

differentiable in a < x < b, then show that
1(a) 0(a) v (a)
f (b) 0(b) i1(b) =0

f '( . ) 014 v' (

(ii) If. F( G (x ), H(x) are Continuous fl a x !C b and
diffcntiable in a < x <b, then prove that

1 F(b)-F(a) '(
1 G(b)-G(a) G'( =0.
I H(b)-ll(a) H' (

sinx sincx sin$

29. lff( x ) = cosx cosa cos/3 ,0<a<f3<17r,
tan tan tan
show that f '( ) = 0, where a < < 13. CH. 19551

30. Deduce Taylors Theorem from Cauchy's Mean Value Theorem

[CH. 19611
Assume O(x)= f(b)-f(x)-(b-x)f'(x)-...

and

31. If f(a) = f(c) = f(b) O where a <c < b, and if f'( x ) satisfies
the conditions of Rolle's Theorem in [a, b], prove that there exists at
least one number such that 0, f '(where
)= a < b.

2. Given that

(x-a )(x-b)
+ f (c)f (x)
(c-a)(c-b)
where a < c < b and f ' ( x ) exists at all points in (a, b). Prove, by
considering the function 0(x), that there exists a number .
a<<b, such that
1(a) +f (b) +
(a-b)(a--c) (b-c)(b-a) (c-G)(e--b) 2
19
290 DIFFERENTIAL CALCULUS

33. Given that

f(x) x 2 x 1
f(b) b 2 b I
' '
1(a) a 2 a
f'(a) 2a

and f ( x ) exists at all points in (, h), deduce

f(b) = f(a)+(b—a)f'(a)+b-u)f"(), a<<b.

34. If f (x ) exists at all points in (a, b) and

f (a) f(b)f(c)
c-a h-c
where a <c < b, then show that there is a number such that a < <b
and f'(çr)=O.
35. Given that
)F (a) f(b) f(x
Ø(x)=g(a) g(b) g(x
h(a) h(b) h(x)

and F(X)=O(X)_(xaXxO(C)
(c-aXc-b)
where a < c < b and f ( x ), g ( h' ( x ) exists throughout
the interval (a, b), show that, by considering the function F (x),
(c)= (ca )(c — b )0 ' < <b

36. If f (x ) exists at all points in (a, h) and if .1 (a ) = f (b )= 0

and if f(c )> 0 where a <C < h, prove that there isat least one
value such that f' ( ) < 0, a < b.

ANSWERS

1. (i) 15. (ii) 6.25. (iii) 1 (iv) . (a + b ).

2. The tangent at the middle point of a parabolic arc is parallel to the
chord of the arc.

3. j. 5. --. 8. No. 9. (i) No. (ii) No.

12. (1)0 = (3 ±
 EXPANSION OF FUNCTION 291

(ii) = ±1; the tangents at these two points are parallel to the line
joining the points (a, I (a ) } and {b, 1(b) } which is parallel to the
x-axis in this case.
,jx2 + xh -x
13. (i) (0) (b) x14 . --
(c) log (l+h/x)If 64
h

x+-_
h h2 /1
15. (i) . x+----j-+ ......... . +(-i
n!(x+Olzr
113 ( fl7t
(iJ)sinx+hcosx--sinx-- COS x+...+—sinr x+Oh-f--
2! 3! n! 2
m(rn- ! ),-22
x", + in + + ..........

rn(m - I )( in .. (rn - n + 1 ),
+o,
1?

16. (i) a" + "c a" - 'h+ "c2 a" 2 h 2 +... + "c a_nIi +.-. + h"

(ii) 37 + 27(x -2 )+ 5(x 2)2.

In the following series, in every case, 0 < 0 < I
I
17. (I) l+x+ x
—+ X
—+...+ x"
---e
2! 3! n'
x- x )2
(ii) 1-Fxloga+ --(log a)+ ---(loga +...+—(loga)"a
2! 3!
x3x5 X'. . (nn
3! 5! n!

x 'fl2t
(IV) 1--+------...+--cosl—Gx +
2! 4! n! L2
X2 x3 1t,-1, -i x
(v) x- —+ ---...+
2 4 n (i+oxr

(vi) -x----...+ —1 r•"

2 0 (1-Ox)"

(vii) I•+rn x+ 2 + ........

,
2
rn(t_I)(rn_2)". (m_n^1) "(
+9h r"
292 DIFFERENTIAL CALCULUS

x2 I )"x
(viii) x—
— +—
x3 —.. + ( n (cot x )sin n(cot Ox).
3 5

(ix)i+XJ r'
- cos I.— - ) cos l( 2—4)
1+ ..........
4

(x) — rsjn Ø+ —+r 2 sin 2Ø+ ...... +—re

• x2
9 sin (bOx+nØ).
21 it'

where ,=J02+b2 and 0 = tan -

a

- (i - ) i/ ( +
18. 0 =
X

3 45 315
(ii)1_x 2 - 2- x, -----x(8lsin30x+sin Ox).
2 8 160

(iii) l_.x2
+ X 4 --x e (4O + 150
2
15

:°
2O.()1+—+—+----+ .... + e
1! 2! 3! (n—i)!
x2 x 4x" (i—o )_1 ( ix
(ii) .... ^ cosi —+Ox
2! 4! (n—I)!. l 2
nx
,, / \,,-1
(uI)i+x+x2 +1-
- Ox )"
24. ..(b+a)

9.9. Expansion of functions in infinite power series. -

Taylor's series (extended to infinity).
If f(x), f'(x), f "(x).....f "(x) existJini1e/' however large n may
be in any interval f -8, x + 8 ]enclosing the point xand if in addition
R tends to zero as n tends to infinity, then Taylor's series extended to
infinity is valid, and we have

to [jhl<o}
Denoting the first n terms ofthe expansion ofby S, and the remainder
by R we have, by Art. 9.5.
EXPANSION OFFUNC71ON 293

f (x+h)rS,,+R,, i.e., f (x+h)— R,,S,,,

Now, let n — oo then If R, — ,0,wehave

h2
f (x+h)=IJ

A gain, since f(x + h) — 5,, = ,if f(x + h) = Li 5,,,

- - ,,-_
then Li R,, =0.
l -4=
Thus, L I R,, = 0 is both necessary and sufficient condition that

fx + h) can be expanded in an infinite series.

Cor. A nother form of Taylor's series which is found often useful is

obtained by putting x — a for h in the form (B), A rt. 9.5.

(x —
Thus, f(s) = f(a)+ (x — a)f'(a) +2a) f"(a) +.-.
2!

Maclaurin's series (extended to infinity).

If f(x), f '(x), f"(x).....f" (x) existfinitely however large n may
be in any interval (— 8 ,8 .) and R . tends to zero as n tends to infinity,
then Maclaurin's series extended to infinity is valid, and we have

f(x) = ((0) + xf'(0) + — ("(0) + ,where I x 1<8.

2!
Illustration.
Ex. Expand the following functions in powers of x in infinite series
stating in each case the conditions under which the expansion is valid:
(i) sins, (ii) cos x, (iii) e, (iv) log (I +.r),
(v) 0+ x),-
(I) Let f(x) = sins
f" (x )= sin ( f nit+x), so that f(x) possesses derivatives of
every order for every value of x. A lso, f" (o )= sin -- nit which is or
±1 according as n is even or odd. -

x I, ,,
fl
x .1 nfl
x" f (Ox)=— sinl Ox+—
• R,, =—
n! 2
294 V/FFEREN1IflI, CALCULUS

R
n! 2 )1 n!
X` ^I s (

since sin (Ux+--niv

)I
R *0 as n— eoo, since as 1°° for all values ofx.

IEx. 8,3II
Thus the conditions for Maclaurin's infinite expansion are satisfied.
- X
3 X5 X7
SIflX = X++ ..... . tO°°, for all values ofx.

(ii) Let f(x)=cosx.

F' (x)=cos(n7t+x), ... f" (0)=cos(n.t), which isO

or ±1 according as n is odd or even.

R,, =f" (0x)= cos 10x+ 2

Now proceeding as in the case of sin x. we can show that
R, —*0 as n —* , for all values of x.
COS X2
x=l----+--"- ......t000,
for all vatuesofx.

(iii) Let f(x)=e.

f" ( )= e, :. f" (o )= i , thus f" (o ) exists and is finite,

however large n may be.
n
x
• R5 =—j (Ox)= -e'
n! a!

Now, since e 0 <c 1 (a finite quantity for a given x)

and ---*-0asn—sco,R--*0 asn —*oo.

a
, 3
• e' =i++L_+_+ ...... t000, for all values ofx.
 1 295
EXPANSION OF FUNCTION

• (iv)Let f(x)=log(1+x).

(i +
Which exists for every value of n for x>-1.
f0(0)=(_1)(n-1)!
If R ,, denotes Lagrange's form of remainder, we have
Y-t I ^ x

I
f " (0 x
n 1+Ox)

(i) Let 0 !^ x:5 1, so that -- 0 as n 00, since

I + Ox )

---- is positive and less than 1.

1+Ox

Also, 1. - 0 as a R,, -, 0 as it

(ii) Let —1 <x <0; in this case may not be numerically

less than unity and hence may not tend 10 0 as n 00

J
Thus, we fail to draw any definite conclusion from Lagrange's form of
remainder. Using Cauchy's form of remainder, we have
I
R = f" (o =—' )' .
I+Ox ( I +Ox

• 1-0 ( 1-0
0
j
Now, is positive and less than 1; hence
asn —>oo. 1
since --l<x<O;
Also, x—*0,as n—*, is bounded
1+Ox
R —O as n — *oo,
• x2 .x
is valid for —1<x!^1.
Proceeding similarly we can show that

log(1—x)=—x—----........is valid for —1!^x<1;

 296 DIFFERENT/AL CA LCULUS

(v) Let f(x) = ( 1 + x)", where in any real nzmber.

Binomial expansion]
When mis a positive integer, f0 (x) = 0 , when n > m, for every
(i)
value of x. Hence the expansion stops after the (in l)fl term and the
binomial expansion, being a finite series, is valid for all values of x.
(ii) When in a negative integer or a fraction,
f"(x)= m(m-1). . . (m-n+1)(1+x)"', for x> -1
Hence Cauchy'sTorm of remainder B is

R. -
(n-l)! l+Ox)
Let -1<x-<l, i.e., jx< I; also, 0< O<i.
0<1-0<1+0x.

1-0 ( i-o
0< <1. .. 0<1-! <1.
l+Ox l+0x)
(n. being a positive integer > I

1-0 I -I
I - 0 as n —p oo. Also, (i + Ox) is finite
1+Ox)
whether m - 1 ) is positive or negative.
(in
-I) ... (rn- nfl) -+0.
Again, if IxI<1, m
(Art. 3.11, Ex 8(iv)
x<1, R0 -0 as n-*.
for I xI<1, Maclaurin's infinite expansion for (1+x)' is valid,
in
being a negative integer or a fraction.

9.10. Determination of the coefficients in the expansion of 1(x) and

I (x+h), (Alternative Method).

(i) Assuming thatf(x) admits of expansion in a convergent power

series in x for all values of x within a certiain range, and that the expansion
can be differentiated term by term any number of times within this range,
we can easily get the oefficient of different powers of x as follows:
Let .f(x)oa0 +a 1 x+a 2 x 2 +a 3 x3 +.-- ( I)
where a0 , a 1 , a 2 ..........are constants.
EXPANSION OF FUNCTION . 297

We have by successive differentiations,

f '(x)=a1 +2a2 x+3a3 x 2 +4a4 x 3 + .- (2)
f (x)= 2.1a2 + 3.2a3 x + 4.3a4 x 2 + ( 3)
f (x)=3.2.1a3 +4.3.2a4 x+ • ( 4)
etc. - etc. etc.
Putting x = 0 in (1),(2),(3),(4) .......... we get
f (0)=ao, j'(o)= a1 ,.f (0)= 2!a2 , f (0)=3!a3...

Hence; f (x )=f (0)+x f '(o)+f '(o)+f (o)+ .......

(ii) Letf(x + h) be a function of h ( x being independent of h), and let
us assume that it can be expanded in powers of h, and that the expansion
can be differentiated with respect to h term by term any number of times
within a certain range of values of h. We can easily obtain the coefficients of
various powers of h as follows:
L et f (x +h)=a0 +a1 h+a2 h 2 + 3 h 3 + ......... ... (I)
where a0 , a 1 , a, ......... . are functions of x, and independent of h.

Since f(x+h)-f(z)- [ wherez=x+h]

differentiating (1) successively with respect to Ii, we get

f (x +h)= a1 +2a3 h+3a3 h 2 +4a4 h 3 + (2)

f ' (x + h)= 2.1a2 +3.2a3 h + 4.3a4 h 2 + ( 3)

f" x + h)= 3.2.1 a3 +4.3.2a4 h + (4)
etc. etc. etc.
Putting h = 0 in (l),(2),(3),(4) .......... we get
f (x )=a0 , f '(x )=a1 , f (x )=2!a,, f (x )=3!a3,
h2 h3
f (x + h )f ( x )+ hf '(x )+ f ( x )+— f -(x )+

Note. Althoujh the forms of the series obtained above for f(x 4-h) and
.1(x) are identical with the Taylor's and Maclaurin's infinite series for the
çxpansions of these two functions, the above method of proof if used
for establishing these two series, is considered as defective in as much
as it does not enable us to determine exactly the value of x for which the
 298 DIFFERENTIAL CALCULUS

infinite series obtained from each of the functions converges to the

value of the function. In fact. Taylor's and Maclaurin's expansions in
infinite series do not converge to the functions from which they are
developed unless R,, -* 0 as n -+ , even though the function might
.possessfinite differential coefficients of all order and the infinite series
maybe convergent; e.g., f ( x ) = e" ( , ,o), j(o)= 0.
Here, f (O ) = 0 for every value of r. But Maclaurin's infinite series
for this function, though convergent for all values of x, is not equal to f(x).
9.11. Other methods of Expansion.
The use ofMaclaurin's (as also of Taylor's) theorem in expanding a
given function in infinite power series is limited in applications because
of the unwieldy form of the remainder (i.e., of the nth derivative of the
function) in many cases. So we employ other methods for expansion.
Now, in this connection it should be noted that the operatiohs of algebra
like addition, subtration, multiplication, division and operations of
calculus like term by term limit and term by term differentiation, though
applicable to the sum ofa finite noiiber of functions, are not applicable
without further examination to the case when the number of terms is
infinite, and hence to the infinite power series Ea,x" . Ifa power series
in x converges (i.e., has a finite sum) for values of x lying within a
certain range (called the interval of convergence), then for values of x
within that range, operations of algebra and calculus referred to above
are applicable, as in the case of polynomials, and the series obtained by
such operations would represent the function for which it stands only
for those values of x which lie within the interval of convergence. We
illustrate below some of these methods.
9.12. Illustrative Examples.
A. Algebraical Method.
Ex. 1. Expand tan x in powers of x asfar.as x.

x3 X5
X- + -
3! 5!
Since tanx= sinx
-=
cosx
- x- +
2! 4!

we may, by actual division, show that

See Chapter 6
EXPANSION OF FUNCTION 299

tan x=x+x3 +-jX 5 +0..

(1+x)
Ex. 2. Expand log in powers of x as far as

Multiplying the two series

log (l+x)=x— x 2 +x 3 --x 4 +... (— l<x<l)

and (i +x) =l—x +x2 _ X 3 +X 1 -... (- 1< x<l )

and collecting together the coefficients of like powers of x, we have
log(i+x)
= x_(l+. )x 2 +(i++)x +... for

Ix 1<1 (the common interval of convergence).

B. Method of Undetermined Coefficients.

Ex. 3. Expand l og (i + x) in ascending powers of x. [ V. P 1998]
Let !og(1+x)=a 6 +a 1 x+a 2 x 2 +a 3 x 3 +... •.. (I)
differentiating with respect to x,

= a i +2a 2 x+3a 3 x 2 +... ( 2)

1+x
(l+x)(a j +2a 2 x+3a 3 x 2 + ... ) = l •.. (3)
Equating coefficients of x" on both sides,
na, +(n+1)a,,,1 rO •.. (4)
Putting x = 0 in (1) and (2), a0 = log! 0, a 1 I

Putting n1,2,3,... in(4),weget a 2 =---, a3 31, a 4 =---,etc.

log(l+x)=x—x2+x3—.... ... (5)

A Ite,na1ii'ely
4_ X 2 — x 3 +x4
For IxI<!,(l+xf
Hence, comparing coefficients of like powers of x on both sides of
(2),weobtain a1=1,a2—,a3,etc.

Note. We shall have now to find for which values of x the series is
convergent, and hence represents the function.
It can be shown that the series is convergent for -I C x < I.
 W) DIFFERENTIAL CALCULUS

Ex. 4. Show that

-1 X=X+---+-----.+--
1 x3 1.3 x 5 1.3.5 x7
Sin + ICP 19471
___
Let y = sin x= a,, + a 1 x + a2 X2 + ...... + a,,x" + .........(1)

Since y=sin x, .. differentiating y 1 = ..±_ ... (2)

1.3 1.3.5
Yi =(i-x) 2 =l+_1x2 +x +x + ..... . (3)
2 2.4 2.4.6
for —1.<x<1 by Binomial expansion.
Also, y1 a + 2a 2 x + 3a3 X2+ ...... + na,,x ... + ......... (4)
Hence, comparing the coefficients of (3) and (4), we get

a 1 =1, a 2 =0, a 3 =-, a 4 =0, a =--,CtC.

2.3 2.4.5
Also, putting x=0 in (1), a0 = sin 0=0.
Hence, the result.
C. Method of formation of Differential equation.
Ex. S. Expand (sin x) 2 in a series of ascending poii. ei-.c of x.
Let y=(sin t x) 2 . ... ( I)
= 2sin'x
Differentiating, y . ... (2)

or, y1 2 (l_x 2 )_4(SIfll •) 4y

Differentiating again, and dividing by 2y 1 # 0,
(i_x2 ) y , — xy 1 —2 =0. ... (3)
Differentiating this n times by Leibnitz's theorem, we get
(1_2 )
y ,2 - 2n+, — n(n-1 ) y ,, — xy,, 1 —fly,, = 0,
or, (1_x2 )Y +2 — (2n+I )xy,, 1 — n 2 y,, =0 ... (4)
From (1), (2) and (3), we get y0 = 0, (y),= 0, (y2)0 = 2, and from (4),
putting x = 0, we get

(Y *2 ) =n2 (Y .) ... ( 5)
putting n = 1, 3,5 ...... in (5) we get
(Y3 ) =(v ) =(Y,)o = ........=0
EXPANSION OF FUNCTION 301

and putting n = 2, 4, 6......in (5), we have

(y4)O=22(y2 )=22.2,
(Y6 ) =42(y4 ) =42.22.2.
2
Similarly, (YE ) =6' . 4'.2 .2, etc.
Assuming that a Maclaurin's series exists for this function, the
coefficients are the values of y, y 1 . Y2 y,.. ...... when x = 0.
Hence,
22
2 . 4 2 .4 2
'4
+-2x6 +
1._I 12  " 8
(
s n- '
i =—•2x +-2x •2x +...
2! 4! 6! 8!
Note. It can be shown that this series converges for x2 :^ I.
D. Differentiation of known series.
Ex. 6. A ssuming expansion of sin x, prove that

x 2 x 4 x6
cos x = 1 -
2! 4! 6!
x3 . .
From the series SIbXX-- + + ......

which converges for all values of x, we get the required result by

differentiation.

Ex. 7. Show that sec 2 x=1+x 2 + . x 4 + ......

• Since tanx=x+x3+x5+......
3 15

we get the required relult by differentiation.

9.13 Miscellaneous Worked Out Examples

Ex. 1. (i) is Rolle's theorem applicable to f(x) X2 - 1, 1,1 ?

=
Justify your answer. I C. P 1989 1

(ii) Does f(x) = cos satisfy Rolle's Theorem in the interval

[C.P.1993J

(iii) Is Rolles Theorem applicable to f(s) = I—x 3 in —1 !^x:5 1?

Justify your answer. I C. P 1990 1
302 DIFFERENTIAL CALCULUS

Solution : (I) is continuous in —l:5x:51.

2x 2
f'(x) (-0(2— x2)2(_2x)= which exists 'in 1< x <I.
(2_x2)

And[(1) =
'-j
= L f(—l) = 1

Thus f(—l) =f(1)

So, f (x) satiifies all the three conditions of Rolle's Theorem,
Hence; Rolles Theorem is applicable tof(x) in [-1, IL

(ii) f(x) =(±)

J(0) = COS is undefined

(J .
So, f (x) is not continuous at x = 0 which is a point in the interval
— 1:5 < 1

Again. f'(x) = —'sin(.!_). which does not e.ist at x 0.

So, f (x) does not satisfy the first two conditions of Rolles
Theorem.

Rolle's Theorem is not applicable to f(x) = cos (±) in [—i. 11.

(iii) Here, f(X)=J—X3

if —0,
3x3

f'(0) does not exist.

f (x) is not derivable at all points in the , open interval —1< x < I.
2
Hence, Rolles Theorem is not applicable to f(x)=I —X3 in the

interval —I:5x:51.
Ex. 2. Explain whether Rolle's Theorem is applicable to the function
f(x)
= x I in any interval containing the origin.
[C. P. 1980, '95, B. P '95 1
EXPANSION OF FUNCTION 303

Solution Here, f(x) = x, X>0

=0, x=0
=—x, x<0
Let us consider an interval —a !^ x 5 a, where a > 0. Obviously,
this interval contains the origin.
Here, ((a) = f(—a) = a
And since, Jim f (h) = 0, lint f(h) =0
and f(0) = O, f(x) is Continuous at x =0 and so at all points in
—a !^ x a -
h)—J(0) Jim
Now, Lf'(0)= lim
h—,O+ It

and Rf'(0) urn f(0+h)—f(0) lim

/1 h-,O+ h
is not derivable at x=O.
Lf'(0)?E Rf'(0), J(x)
— a < x <a.
So, it is not derivable at all points in
Second condition of Relies Theorem is not satisfied.
Hence, Rolle's Theorem is not applicable to 1(x) = x in [—a, a],
i.e., in any interval containing the origin.

Ex. 3. (i) Test the applicability of Roile's Theorem for the function
f(x) =(x—a)'".(x—b)"in ax!^b where in, n are positive integers.
in
(ii) If f(x) =(x—a)'"(x—b)° , where n are positive
integers, show that 'c' in Roile's Theorem divides the segment a !^ x 5 b
in the ratio in n. C. P 1998

Solution : (i) f (x) = (x-a)"'-(x- b)"

since in and n are positive integers, f(x) is a polynomial of
degree (m+n) which is continuous at every point, so it is continuous
in a5x:5b.

Also, f'(x) = m(x—a)"'(x—b)" +n (x—a)"(x—b)"

= (x — a)'' .(x—b)'-lfm(x—b)+n (x—a)} . (I)

which exists in a < x < b.
f(a)=0=f(b)
So f (x) satisfies all the conditions of Rolle's Theorem. Rolle's
theorem is applicable tof(x)in asx!5b.
304 DIFFERENTIAL CALCULUS

(ii)Since f(x) satisfies all the conditions of RoDe's Theorem,

there is at least one point c in a <x < b, such that f'(c)-O.
From (I), (c—a)'(c—b)" {m(c—b) + n(c—a)} =0
:a<c<b, c—a*0, c—b;60

mb-I-na
m(c—b)+n(c—a)=0 or, C
M+11
so that the point x = c divides the segment a x!5 b internally in the
ratio m: n.
Ex. 4. Verify Rolles Theorem for the function f(x) = —5x+6 in
1<x<4. [C. P.1991]

Solution : f(x) = -,5.x + 6

f (x) is a polynomial function, it is continuous at every point,

so it is continuous is I !^ x < 4.
f'(x) = 2x-5, exists in I <x <4.
f(1) =2 =j(4)
Thus f(x) satisfies all the three conditions of Rolles theorem.
Then there exists at least one point .r = c in 1< .r < 4, such that
f'(c)=O.
2c-5=0, or. c=— , which lies is l<x <4.

Thus RoDe's theorem is verified.

Ex. S. Discuss the applicability of the Mean value value theorem

a << b.

Find , if the theorem is applicable.

(i) f(x)=x(x—t)(x-3), 05x4 F C. P 1992, B. P 20021

(ii) f(x)=I x I, —1:5x:51 IB.P 1995]

(iii)f(x)=IxI,05x!^1 I
[G.P1988]
(iv) f(x)=x(x1)(x2), a=0, b=— [ C. P. 1987, B. P 1997]
Solution: (i) f(x)=x(x—l)(x-3)

= x 3 —4x 2 +3x
f'(x)= 3x 2 —8x+3.
EXPANSION OFFUNCT1ON 305

1(x) being a polynomial function of x is Continuous in 0:5x:^,4.

f'(x)also being a polynomial function of x, exists in 0<x<4.
So, 1(x) satisfies the conditions of Lagrange's Mean value
theorem in OSx54.
There exists at least one point t in 0 < x < 4, such that
f(b)—f(a)=(b—a)f'(t)

i.e., f(4)—f(0)=(4_0)f'()

or, 12_O=4(32_8+3)

or, 3,2 —8= 0 .. =0,

0< < 4, = (the value : = o is rejected)

(ii) f( x )=I x I, —1:5x<1

Here .1( x) = x I is continuous in —1 :5 !^ I ,but f'(x) does not
exist at x = 0, in the interval [see Ex. 2 1. So, Mean Value theorem is
not applicable for the function f(x)=Jxj in the interval —1:5x:51
(iii) f (x) = J x I , 0<i-<I
f(x) = x I is continuous in the interval -o !^ x:5 land f'(.)
exists at all points in 0< x <I. f '(x) = I. Thus 1(x) = x satisfies the
I

conditions of Mean Value therorem in 0:5 x <'.

So, there exists at least one point in 0<< I,
such that f(l)—f(0) = (1--0) f'() . ... (I)
i.e.,
so the relation (1) is satisfied identically.
Hence, is any number in the open interval 0 <x < I.
(iv) f(x) x(x — 1)(x —2), a = 0, b = -
Here, f(x)=x3-3x2+2x
f'(x)=3x2-6x-t-2

f (x) being a polynomial function is continuous is 0 :^ x ^ -- and

f'(x) also being a polynomial function exists i n <x

 DIFFERENTIAL CALCULUS
306

f (o)=o

f (!)_f (0) (----O)f '() gives

or. l27-24E,+5=0

- 6
I 6—

Ex. 6. (i) For what range of values of x. 2x 3 - 9x 2 + 12x —3

iC. P. 1986, B. P 1989 1
decreases as x increases?
(ii) Show that x 3 + x 2 —5x+3 is monotone increasing when x >1
C. P 1992 1

Show that —2x 3 + 15x 2 -36x+6 is strictly increasing in

C. P 1993 I
2<xe.
+ 24x +,", is increasing on the real line.
(in') Show that 2x 3 - 12x 2
C. P. 1995
V2
(v) Separate the intervals in which 2V3-15 +36k +1 is
I B. P. 1994 1
increasing or decreasing.

Solution i) Let. ' f(x) = 2x 3 —9x 2 'i-12x-3

= 6x 2 - lSx+ 12
= 6(x - 1)(x —2)

f(x) decreases as x i ncreases if f'(x) < 0

Obviously, f'(x) < 0, for 1 <x <2

Hence the required range of values of x is 1< x < 2

(ii) Here. f(x) = x 3 -Ex 2 —5x+ 3

f'(x)3x ij
Here, f'(l) 0 and f'(x) >0, when x >1.
EXPANSION OF FUNCTION 307

Hence,! (x) is monotone increasing when x >1.

(iii) Here, 1(x) = —2x 3 + 15x 2 - 36x + 6

f'(x)-6x2+30x-36=--6(x-2)(x-3)
f'(x) > 0 for all values of x satisfying 2 <.x <3.
Hence, 1(x) is Strictly increasing in 2 <x <3
(iv) Here, f(x)= 2x3 —12x2 +24x+6
f'(x)=6x2_24x+24=6(x_2)2
For all real values of x, f'(x) ^ 0.
Hence, f(x) is increasing on the real number line.
(v) Here, f(x)=2x3-15x2+36x+1
f'(x)=6x 2 —3O.+36=6(x-2)(x-3)
Obviously, f'(x)> 0 if x >3 or x <2
and, f'(x)<Oif 2<x<3
So, f(x) is a decreasing function when 2 < x < 3 and it is an
increasing function in (--,2) and (3,

Ex. 7. (i) Show thatj- — <log(1+x)<x, ifx >0.

[C. P 1987, '89,, B. P 2000 1

Solution Let 4(x)= log (l+x)__L

1+x
Here 4(0) 0
I l+x—x x
= >0, for x>0
J+x (l+x) 2 (li-x)2
•(x)> 0 for x > 0, and consequen;h
x
--< log (1±x) for x>O
l±x (I)
Now let, iv(x)=x— log (1+x)
Here, c (0) = 0

and 'U '(x) = 1— -- = >0 for x >0

1+x l+x
W(x)>0. when xA
i.e., x - log (1 + x)> 0, when x >0
308 DIFFERNT1AL CALCULUS

2)
i.e., log (1+x)<x, for x>O

From (1) and (2). j--< log (l+x)<x,ifx>O.

Ex. 8. Show that decreases steadily in 0 <x

C. P. 1983, B. P 1993

Solution : Let, f(x) = -

sin X
x
XCOSX — sin ...
X
(j)
f (x)=
x2
Let F(x) = xcosx—sinx
then F'(x) =COSX—xsiflx—COSX

=—xsinx< 0, in 0 < x <—

7t
<3.
F(0) = 0 and F (x) is strictly decreasing function in 0 <X

F(x)<0in0<x<.

i.e., xcosx—sinx< 0, in 0 <x<

So, flDrn (1) f(x)<0in0<x<.

sinx it
Hence, — decreases steadily in 0< x
x

2 sinx It
Ex.9. Show that —< <1, for 0< .r <—.
it x 2

Solution Let us define a function 0, such that

sin
4(x) x x0
x
=1 when x=0

Obviously, 4 (x) is continuous in 0 !^ x and derivable in

it , xcosx—snX
0<x< and 4(.t)=
2 x
 EXPANSION OF FUNCTION 309

Let us consider another function W (x), such that

f(x) = xcosx — sin x defined in {o. .]

V'(x)= —xsinx4 for 0< x

1(x) is strictly decreasing in 0:5 x <

W(X)<W(0)=O for all x 0:5x<-

$'(x)<O, for b:5x5.

So4i (x) is strictly decreasing in 0:5x<,

$(0)> $(x)>4(2)for 0<x<-.

sinx -
i.e., 1>—>!
X
2
2sinx
i.e., <1 for 0<x<—.
x 2
15
Ex. 10. Show that—>
tan --,for
.t O<x<—
x sin 2
•
[C.P.1983,B.P. 1993)
Solution Here, we shall have to show that
tanxsinx—x2
>O for o<x< -
xsinx. 2
X sin 0, when 0<x<_
71

it will be enough to show that

tanxsinx—x2 >O, for O<x<

Let F(x)r tan x sin x_x2,

thee F'(x)=sec2xsjnx+tanxcosx_2x
srnx sec 2 x+ sin x —2x
 DIFFERENTIAL cALCULIJ
310

and F'(x) = cosxSec2 x+sin.r2SCCXSCCX tanx+COSr-2

= sec x + cosx -2 + 2 sin x tan x sec 2 x

(JJ_) 2 +2sinxtaflxseC2x
It
>0, for 0<x<

So. F'r) is strictly increasing in the interval 05

.V

i.e.. F(x)>O, for 0< x<

x< and also
Therefore. F(x) is strictly increasing in 0<-
F(0) =0.

So, F(x)>0. for 0< x<

i.e., tanxSinx-x2 >0, for 0<

IT
tanxsiflx — X 2. — . ( . xsinx>O in 0< x<ç)
or, >0, in 0< x<
XsInx
tnx x
> -, when 0< x< -.
X sin

EXAMPLE -IX (B)

h:
1. Expand in infinite series in powers of
cos ( + h ). (iii) e 1' sin (x +
(i)
Ii ).
(ji)

Expand the following functions in powers of x in infinite series, stating

2.
in each case the condition under which the expansion is valid:
(ii) sinh x. (iii) cosh X.
(1) a
(v) cot ' x (vi) e" in hx.
(iv)*tan-'x.
(viii) e sin x. (ix) e' cos X.
(vii) e" cosbx.

(x)II
1+x
3. Show that
)2
sin 20
tan - 'x +(i sinG ).sin0 - .2L
(hsin0
tan(x + h )=

o
)3 sin 3(1
- ..., where 0 = cot-'
+ . (h
EXPANSION OF FUNCTION 311

4. Find approximately the value of sin6OP3423r to 4 places of decimals

from the expansion of sin (x +h) in a series of ascending powers of h
by putting
and of a radian (=34'23nearly).

S. Show that

(I) logx=(x-l)-4(x-1)2+.(x-i)3-...
is true for 0 < x :5 2.

(ii)

is true for O<x<4.

6. Expand e'. in powers of(x.-1).
Verify the following series (Ex. 7- 19):
7. sec x= l+x
2
2 + .24 x4-

& log( l+x)' =x+-_.+...

( \ x2 2 3
2 )=-x+-..-+-x X
9. logi) - x+X
23 4
X3 X5
10. el slhx=x+x
2
+—+--...
3 30
X2
2x3 9x5
e

X 1 .1 2 1
12 1--x+—x --x
2 12 720
13.

2. 3
14. :log(l+sinx)=x__+-_..;

x2 2x4 16x6
Is. log sccx=—+---.-___+...
 DIFFERENTIAL CALCULUS
312

16.

17. =1+x+----'... [C.P.19401

18.

19. (I+x)
20. (j) By differentiation the identity
-I ' 3
(1-x) =l+x+x +x + .... . jxI<1.
show that
- 3.4 , 3.4.5
(1-x) =1+3x+—x+—x +....
1.2 1.2.3
(ii) Differentiating the expansion for log (i + sin x). obtain the
expansion for see - tan
X
21. (i) Show that Fx and 2 cannot be expanded in Maclaurin's
infinite series.
(ii) Given , show that for this function the expansion of
f ( ) =
f ( x + h) fails when x = 0, but that there exists a proper
fraction U such that
f(x+ h)= f(x)+hf'(x)+ h2f(x+Oh)

holds when x = 0. Find q. [C.P. 19491

22. If y=(1-i.x)" =a 0 +ax+a2 x 2 ....

+ show that

(i + x )y, = ny
and hence obtain the expansion of (i +

23. If y = = 00 + a1 x + 0 2 x + a3x3+...,
prove that
(i) (I_x2)y2=xy1+a2y,

and hence obtain the expansion of e"

Deduce, from the expansion of e' "' - " , the expansion of sip - x.
IcR 19451
 313
EXPANSION OF FUNCTION

24. If yeao+aix+a2x2+.. show that

=
(i) (i
(ij) (n+1)a. j +(n-1)a_i -

and hence obtain the expansion of e

25. If y = sin (msin 4 , show that
+ 1)xy + (m 2 n2 =o
—x2 )yn+2 — (2n

and hence obtain the expansion of sin (m sin x C. P 19381

26. If y e cos bx,PrOVe that

y2_oyi+(a2+b2)y0.
and hence obtain the expansion of e" cos bx.
Deduce the expansion of e and cos bx. C. P 19371

ANSWERS
ft2 h3
: (i) 2! 3!
2 )h3
h
—cosx + —s,nx +...
(ii) cosx – hsinx –

sin x+h sin (X+)+2? sin x+

)...

2 2
2.

x3 x5 forall valuesof x
(ii) X+ ++..•
3! 5!
2
(iii) I+—+—+..
2! 4!

(w)x--+-j+...
for-1:5x:51
(v)_(x_+c.–...).
 314
DIFFERENT/AL CALCULUS

(vI)-rsjflO+_r2sjn2+
where r=Jd2+L?
(vil )1+rco$+r2 co and 4=taxi'-.
....

r r7 ( \2
(viii) x2 sin— + - i2 ) sin— +...+ _22 sun— +

(ix)

Ex. (vi)- (ix) are valid total! values olx.

(x) 1-x+ 2 -x 3 i...

1xI<1.

I
(xi) l-x2+4--.+ -l<r<l.
4. 0 . 8710. e11+(x_I)+
(x-i) (xJ
6. +
2! 3!
20. (ii) I - x + . x 2 - -x3 +... 21. (ii) .

22. n(n-I) 2
l+nx+ _—x +...
2!
ax '2 12+1 '
aa 2(2 2
23. I+ax^_+ i + aa +2/ x 4
2! 3!
+ a(a 2 +1 2 ) 2 +32 )
+
5!
sin x = + +

24. I+n+ T .
,x ..— + x +
nz2 (m2 - 8
) x +...
2! 3! 1
4!

_i)n2 _32 )
25.
3!

a2 -b 2 2 a(a2 _3b2) X 3
26 l+IJX+ X + +...
2! 3!
22
ax
e =I+ ax + -
2!
b2x2b4x4
cos bx = I----- + —
2! 4!
FAI MAXIMA AND MINIMA (Functions of a Single V araible)

10.1. By the maximum value 0fhfunction f(x) in Calculus WedOnot

necessarily mean the absolutely greatest value attainable by the function.
A function f(x) is said to be maximum for a value c of x, provided
f (x) in the
f(c) is greater than every other value assumed by
immediate neighbourhood of x = c. Similarly, a minimum value of f(x)
is defined to be the value which is less than other value in the immediate
neighbourhood. A formal definition is as follows:
A function f(x) is said to have a nzaxirnu,fl value for x =
c, provided
quantity ? such that for all values of x in the
we can get a positive
interval
c — ö <x<c+S , (x c) f(c)>f(x);

i.e..if f(c i-h)— f(c)< 0, for I I sufficiently small.

for x = d, provided
Similarly, thefunctionf (x) has a ininitnuni value
within which
we can gel on interval d - <x < d + '

i.e.,if f(d + h) — f (d )> 0,rnr It I I sufficiently small.

Thus, in the Fig. 10.1.1 which represents graphically the function
1(x) (a Continuous function here), the function has maximum value at
etc. and has minimum values at Q1, Q2,
P1 , as also at l' 2 , P 3 . P4 ,
say),
Q3, Q 4 , etc. At P 1 for instance. corresponding t0 x=0C1 (=c1
is not necessarily
the value of the function, namely, the ordinate PC 1
, but we can get a range, say
bigger than the value Q 2D2 at x = 0D2
on either side of it, (i.e., we can find
L1 C1 L2 in the neighbourhood of C1
say) such that for every value of within 4C, L2
a 6=L 1C1 =C1 4.
(cxecept at C 1 ), the value of the function (represented by the
corresponding ordinate) is less than P1 C1 (the value at C1).
 316
DIFFERENTIAL CALCULUS

tlg 10.1.1
Hence, by definition, the function is maximum alx= OC,. Similarly,
we can find Out an interval M 1 D2M2 (MI D 2 = DM, = , say) in the
neighbourhood of D2 within which for every other values of x
the
function is greater than that at D,. Hence, the function at D, (represerted
by Q,D,) is a minimum.
From the figure the following features regarçling maxima and minima
of a Continuous function will be apparent:
(i) that the function may have several maxima and minima in an
interval;
(ii) that a maximum value of the function at some poinit may be Le-
than a minimum value of it at another point (C,P, <D2Q);
(iii) maximum and minimum values of the function occur alternately,
i. e.,betwccn any two consecutive maximum values there is a minimum
value, and vu-c versa.

10.2. A necessary condition for maximum and minimum.

If f (x) be a maximum, or a minimum at .r = c, and if f' (c)
exists, then f'(c )= 0.
By definition, f (x ) is a maximum at x c, provided we can find
a positive number 5 such that

f ( c + h )- f (c)< 0 whenever - S <h <5, (h * 0).

 317
MAXIMA AND MINIMA

• [+ h)— f (c <0 ifhbe positive and suffiCiefltlYsmall, >0

if h be negative and numerically sufficiently small.

Thus, L i L (.i') f('):go,isee &t5,31l]

It
Li f(c+h)—f(c)>0
ànd,similarly.
h

Now, if f'( c ) exists, the above two limits, which represent the
right-hand and left hand derivatives respectively of I () at x = c,
must be equal. Hence, the only common value of the limit is zero. Thus,
f'(c)=o.
Exactly similar is the proof when f(c) is a minimum.
Note. In case f'(c) does not exist, f(c) may be a maximum or a
miflimwn, as is apparent form the figure for points Q2 and P4. At the
former point f(x) is a minimum, and at the latter it is a maxiinum. f'(x)
is, however, not zero at these points, for, f (x) does not exist at all at
these points.
10.3. Determination of Maxima and Minima.
(A ) If c be a point in the interval in which the function f(x) is
defined, and if f'(c ) = 0 and r(c);&
0, then 1(c) is
(I) a maximum if 1 (c) is negative and
(ii) a minimum if is positive.
Proof: Suppose f'(c)=O, and f(c) exists, and *O.
By the Mean Value Theorem',
f(c+h)_f(e)=hf'(c+Oh). where 0<0<1.

=01,2
Oh

Since 0<0<1, Oh -.+ 0 as It .- 0, and writing Oh

f(C + k)— f'(k) = f(c)
the coefficient of 0 h 2 on the right side
2 has the same sign as that
Accordingly, since o h is positive, j (c + h ) - f (c )
of t (c ) when I h I is sufficiently small.

Since f"(x) exists, f(x) also exists in the neighbourhood of c.

 318 DIFFERENTIA l. CA LCULUS

If f (c) is positive, f(c + iz)- 1(e) is positive, whether h

is positive or negative, provided 1h I is small. Hence 1(c) is a minimum,
by definition.
Similarl y , If f(c) is negative, I (c + h )- 1(c) is negative,
whether is positive or negative, when I h I is small.
Hence f (c) is a rnaxi,nu,n.

(B) Let c be an interior point of the interi'aI f deiniiion of the

function f( x ), and let

f'(c ) f(c )= f'1(,)= U, and f" (c ) 0:

the,, (i) If n be even, 1(c) is a maximum or: minimum accoiding as f" (c)
is negative or positive,
and (ii) (n be odd, 1(c) is neither a minimum nor a maximum.
Proof: By the Mean Value Theorem of Higher order, here
I I
f(e+h)-f(c) f"'(c+Oh). 0<0<1,

- Oh" f" (c +0/i )- f"' (c)

(ti-I)! Oh

Since 0<0<1,as h -sO,Oh - O,and

the coefficient of Qh"/ (it ) , on the right side --> f" (c

Now, suppose n is even: then Oh" / (, - j ) is positive.

f(c+h )- f(c ) has the same sign as o f (c ), whether It

is positive or negative, provded h is s ufficiently small. Hence, if
f" (c) be positive, f (c -* Ii )- 1(c) is positive for ithcr sign of h,
whc Ii 1 is small, and so j ( ) is a ,ninin,utn. Sinii lady, if
f' (
is negative, f (c ) is a maximum.

Next, suppose n is odd: then Oh"/ (a - i )! is positive or negative

according ash is positive or negative. Hence. f(e+/i ) -j (c ) changes
in sign with the change cf I, whatever the sign of f' (c ) may be, and
so f(c ) cannot be 'iiliet a rnaimurn or a minimum at x = C.
MAXIMA AN! MINIMA 319

Hence, to determine maxima and minima of f ( x ), we proceed

with the following working rule:

Equate f'( x ) to zero, and let the roots be c, c2 1 c 3 ,.... Now

work Out the value of r (c 1 ). If it is negative, then x makes a
maximum. If be positive, then is a minimum of. Similarly test the sign of
for the other values of x for which is zero, and determine whether is a
maximum or a minimum at these points.
If, ij any case above, f ( c ) = 0 ,usc criterion (B).

Note I. The above criterion for determiniming maxima and minima of

non-existent, even though 1(x)
f( x) falls at a point where f'(x) is
may be continuous.
In such a case we should bear in mind that if 1(x) be a maximum
at a point, immediately to the left of it the value of f(x) is less, and
gradually increases towards the value at the point and so. f'( x) [which
represents the rate of increase of /( .v) ] is positive. Immediately tGthO
right, the value off ( ) is again less, and so . 1(x) decreases with s
increasing and, therefore, f'( x) is negative to the right. Thus changes
sign from positive on the left to negative towards the right of the point.
See point P 4 , in the figure ofA rL 101 1
Similarly, if f ( .v ) be a minimum at any point. 1(x) is larger on the
left, and diminishes to the value at the point, and again becomes larger
on the right, i.e., f(x) increases to the right. Thus f'() changes
sign herc, being negative on the left and positive on the right of the
point.
Thus, we have the following alternative criterion for maxima and
minima At a point where f(x ) is a rnaxi,nurn or a minimum, f'(x )
changes sign, from positive on the left to negative on the right if f (x)
he a maximum, and from negative on the left to positive on the right if
r() beaminimum.
If . r'( • ) exists at such a point, its clt.tngc of sign from one
side to another takes place through the zero value of f 'v ). so
that f'(x) = 0 at the point. If 1'(x)be non existent at tiC p0int,
the left-hand and right-hand derivatives are of opposite signs at the
point.
320 DIFFERENTIA L CA LCULUS

Even in the case where the successive derivates exist, instead of

proceeding to calculate their values at a point to apply the usual criteria
for maxima and minima of f( x ) at the point, we may apply effectively
in many cases this simple criterion of changing of sign of f'(x + h) as
h is changed from negailve to positive values, being numerically small,
[For illustration see Ex. 4, § 10.5]

Note 2. At points where f( x) is maximum or a minimum, f' (x) = 0

when it exists, and accordingly, at these points the tangent line to the
graph of f ( x) will be parallel to the x-axis (as at P, Q 1 1', Q2' P, Q,
etc. in figure of § 10.1). At points where 1(x) is a maximum or a
minimum, but f' ( x) does not exist (e.g., at Q1 and P), the tangent line
to the curve changes its direction abruptly while passing through the
point. A special case is where the tangent is parallel to the y-axis, the
change in the sign of f(x) taking place through an infinite value.

Note 3. A maximum or minimum is often called an 'exiremum' (extremal)

or 'turning value'.

The values of x for which f'(x) or l/f'(r)=o are often

called 'critical values' or critical points of f( x).

Note 4. The use of the following principles greatly simplifies the solution
of problems on maxima and minima.
(I) Since f (x ) and log f (x ) increase and decrease together,
log f ( x) is maximum or minimum for any value of x for which f( x)
is maximum or minimum.
(ii) When f ( x) increases, since I/f ( x) decreases, any value
of x which renders f( x) a maximum or minimum renders its reciprocal

1/1(x) a minimum or a maximum.

(iii) Any value of x which renders f( x) positive and a maximum

or a minimum renders { f( x ) }" a maximum or a rttinimum, n being a
positive integer.

For examples on maxima and minima of functions of two variables

connected by a relation, see Ex. 7 and Ex. 12 of Art. 10.5.
MAXIMA AND MINIMA 321

10.4. Elementary methods (Algebraical and Trigonometrical).

Certain types of problems on maxima and minima can be solved
very simply by elementary algebra or trigonometry. The discussion of
the maxima and minima of the quadratic functions or the quotient oftwo
quadratic functions will be found in any text-book on algebra.'
In solving simpler problems of maxima and minima of functions of
more than one variable, the following elementary results are of great use:
- 1 12 f 12
(I) xy(x+y)j —(x—y)j.
(ii) (x+y )2 = 41-Y + (x — y )2

(iii) x 2 + y2 = x + , )2 + I (x - y)2.
When the swn of two positive quantifies is given, it follows from
(i) that their product is greatest, and from
(iii) that the sum of their squares is least, when they are equal.
When the product of two quantities is given, from
(ii) their sum is least when they are equal.
The above theorems may easily be extended to the cases of more
than two quantities.
Thus, when the sum of any number of positive quantities is given,
their product is greatest when they are all equal, and so on.
For illustrative examples see Art. 30.5, Ex. 9 to 11.
Note. In algebraical or trigonometrical example, by maximum or minimum
value of a function we t'sually mean the greatest or the least value
attainable by the function out of all its possible values. In Calculus,
however, as has already been remarked, a maximum or a minimum value
indicates a local (or relative) maximum or minimum.
10.5 An absolute or Global maximum and an absolute or Global minimum
A real valued functionfx) defined in [a, b) is said to have an absolute
maximum (or, a global maximum) at a point c E [a, b] if f(x) !^ f(c) , for
all xc [a, b] and f (x) is said to have an absolute minu,num (or, global
minimu,n) at x = cc [a, b} , if f(x) ^! f(c) for all xE [a, b], f(c) being
called the absolute maximum or the absolute minimum value of f(x) in [a, b].
To find the Absolute maximum or the Absolute minimum of values of a
continuous function defined in a closed interval [a, b], the points X 1 , A2,
A, in [a, blare determined where f'()L r ) = 0, r= 1,2,... n.
Sec Das & Mukherjee's Higher Trigonometry, Chap. X V Sec. B.
2
See Ganguly & Mukherjee's Intermediate A lgebra Chap. V I, A rt. 6.12.
21-
 322 DIFFERENTIAL CALCULUS

Then the Absolute or Global maximum of f(a) in [a, b] is given by

G = max{ f(a), f(A1), f ( 2 2 ), ... I f(X ^ ), f(b)4 and the Absolute or
global minimum 011(a) in [a, b] is given by L = min(f(a), f(X j ), P h),
,f(A),f(b)}.

Remarks!. Any absolute maximum (or, minimum of f (x) is also a local

maximum (or, minimum) but the co verse is not true.

2. A functionf(x) in l a, b] can have no local maximum (or, minimum)

more
or may have one or points of local maxima (or minima).

Example 1. Find the global nzini,nu,n if the f(x) = x 3 - 6x 2 +9x + I in

[0, 11.

Solution. Here,. f(x) x 3 6x 2 + 9x +1 -

f'(x)= 3x 2 —12x+9
For maxima or minima off(x), f(x)= 0
or, 3(x2-4x+3)=0
or, (x—l)(x-3)=0 :.x =1,3
But 3 [0, 11
Only critical point is x = I
Global maximum value = max )f(0), f(1), f(2)).
=max)1, 5, 3) -
Global minimum = min (f(0). f(l), f(2)1 = min1, 5, 31 I.
Hence the global maximum and global minimum 011(x) in [0, 1] are Sand I
respectively.

Example 2. Discuss the absolute maximum and absolute mnini,numof

f(x)=tanx— log x in l, I.

Solution. f(x) = tan x - log x -.

1 1 I (x—l)2
f(x)=---.—=—
2
1+x 2 2 x 2x(l+x )
f'(x)=Ogives x=1.

Absolute maximum = max f(l).

fitl iijtl
=max—+—loo3. -, -----Iog3 =—+- -log
6 4 4 3 4 J 6 4
liti
MAXIMA AND MINIMA 323

Absolute rtiinimum =min {f(_j_). f(J), f(..J)J

• .Iiti 7t1t-:l
=min-+- log 3,-, --- log3
16 4 4 3 4

= iti
- - - log 3
64
Note : The Absolute max or absolute min does not cor re spond to x = 1.
where f'(x)=O.

Example3. Examine the function f(x) = 1 xl -2:5x:52,

for absolugenzaxi,nu,n and absolute inin)n,um.
Solution. Here, Ixl^:O
for all XE [-2. 2] andf(x)=O,
when x=O.
So, f(. ) ^ f(0) for all
2
xE[-2,2]
Hence f (x) has absolute X X
minimum at x = 0 and the
absolute minimum value off(r)
is 0.
Also, I x j!^ 0 for all XE [-2, 23 andf(2) = 2, f(-2) = 2.
Sof(x) has absolute minima at x = -2, 2 and the absolute maximum
value of f (x) is 2.
Note: Here f'(x) does not exist atx=Oand f '(X)0 at x=±2.
10.6. Illustrative Examples.
Ex. 1. Find for what values of x, the following expression is maximum
and ,nini,nuni respectively
2x3-21x2-t-36x-20
Find also the ,naximunt and nzinimu,n values of the expression
[C.P. 1936]
Let f(x)=x3-2l2+36x-2o.
f'(x )= 6x 2 - 42x+36, which exists for all values of x.
Now, when f (x ) is a maximum ora minimum, f '(x )=o.
we should have 6x 7 -42x +36 = 0, i.e., x 2 -7x+6 = 0,
 324 DIFFERENT/AL CALCULUS

or, (x-l)(x-6)=0; xl or 6.
Again, f'(x)=l2x-42=6(2x-7).
Now, when x = 1, f "(x )= -30, which is negative.
when x=6, f"( x )= 30, which is positive.
Hence, the given expression is maximum for x = 1, and minimum
for
The maximum and minimum values of the given expression are
respectively f (i ), Le., -3, and f (6 ),Le.. -128.
Ex. .2. In restigale Jbr what values of x,

f(x)=5x' -18x' +15x 4 -10

is (1 !llQXjnIUfl? Of tin nirnum.
Herr, f '(x )= 30(x 5 -3x 4 +2x),
Putting f '(x )= 0, we have x3 (x2 — 3x+ 2
i.e.,x3(x_t)(x_2)=O, whence x=0,lor2.
Again, f(x)=30(5x 4 -12x 3 +6x2)
When x = I, f'(x) is negative, and hence f (s) is a maximum
for x= 1.
When x = 2, f( x ) is positive, and hence f (s) is a ,nini,nu,n
forx = 2.
When x = 0, f(x )= 0; so the test fails, and we have to examine
higher order derivatives.
f_(5)=120(5x3._9x2 +3x). jo)=o,

f ' (x )= 360(5x2 '-6x+ .. j o) is positive,

Since even order derivative is positive for x = 0,
for x=0, f(s) isamini,nun:,
Ex.3. Show that f (x ) = x 3 - 652 + 24x +4 has neither a maximum nor
a minimum.
Here, f'(x)=3(x 2 -4x+8)=3{(x'-2)2 +4)
which is always positive and can never be zero.
f(s) has neither a maximum nor a minimum.

Ex. 4. Examine f(x _92 +24x-12 for maximum orminimu,n

)= x3
values.
MAXIMA AND MINIMA 325

Here, f(x)=3(x2 _6x+8)=3(x_2)(x_4).

Putting .f'(x)=O.wefind x=2 or 4.
Now, f'(2-h)=3(-h)(-2--h)-_+ve,
And f'(2+h)=3(h)(h-2)=--ve, since his positive and small.
by § 10.3. Note 1, for x = 2, 1(x) has a maximum value, and
this is f()=8.
Again, f'(4-/i )= 3.(2-h )(-h )= - ve,since h is positive and
small,
f(4+h)=3.(2-fh)(h)=+ve,
by § 10.3, Note 1, for x = 4,f(x) has a minimum value, and this
isf(4)=4.

Note. In this case we could have easily applied rule. of Art. 10.3.
Ex. 5. Find the maxima and minima of
l+2 sin x+3 COS 2x, (0!5x^+1T)
Let f(x)=1+2sinxi--3cos 2 x.
Then
COS x-6cosx sin x.
f '(x )= 2
f'(x)=O when 2 COS x(1-3sinx)=0, i.e., when cos x=O.
and also when Sin .r =
(COS
2 x-sin2 4.
f(x )= -2 sin X_6
When
COS x= 0, x = 2'n, :. sinx = 1, .. f'(x).= -2 + 6 = 4 (i-ye).
for cos x = 0,1(x) is a minimum, and the minimum value is 3.

When sinx r - ,

f (x )= -2sin x_6(1_2sin 2 x)---1-6(1

3
-V 2 ) (negative)

Therefore, for sinx=,f(x) is a maximum and the maximum

value is 1+2.++3.(I_..)=4-.

Ex. 6. Exaftiine whether X . possesses a maximum or a minimum and

determine the same C. P. 141, '45 1
Let y =x , :. logy= .! log x.
326 - DIFIE/?INTIAL (ALCl/U/S

Idy 1 1
--=----2-Iogx=---O_logx), (1)
ydx x x

..when
dv
=O,l_
log xo, ...
logxl=Ioge,...xe.
dx
Again, differentiating (1) with respect to x,

, ( _ I /.V -
I ( dy I d - v x log x)2s + 2 log .r
2j) 2
Yd 2 x4 -

d2 y 1 -3+2 e

I
when x=e,,IX----e. =---- whidi is negative.
2
(1),
for x = e, -c- = ()
dx

for x= e, the function is maximum, and the maximum value is e.

Ex. 7. Find the maximum and nun,,nu,n values of it

4+ 36
u-+— and x+), = 2. C. P.
X y

Eliminating y between the two given relations

4 36 du 3616 % 2 +.v—1
x 2-x ' " dx x2 (2-x) 2 x2(2-x)2

du d2u 8 72
- = 0 gives X=-j or -I. Also--=----+
d - dx . (2-J

d 2u 8 72
Wln x -2 = +_ Which is positive.
fl 7
for x = ,',, it is a minimum.

4 36
minimum value of U
; 2-

d 2 u72
When x = -1,— = - 8+ - which is negelive.
27
for x=-1, J4isa maximum.
4 36
maxiiium value of u=— +----=S
-1 2+1
MAXIMA AND MINIMA 327
-4--I

Ex. 8. Examine the function f(x).'4-3(x-2) for maxima and

minima at x = 2.
2
Here, f'(-)=–
(x-2)
For x = 2, f'(x) does not exist, the left-hand derivative being
+ and the right-hand derivative - ; but i/i'( x) is zero for x = 2.
So, the test of Art. 10.2 fails. Let us apply the criterion of § I 1)3, Note!.
Now, f'(2–h) is positive and f'(2+h) is negative.
Hence, the function has a maximum value forx = 2, and the maximum
value is j . ( 2). i.e., 4.

Ex. 9. A conical tens of given capacity has to be constructed. Find the

ratio of the height to the radius of thebase for the minimum amount of
canvas required for the lent.
It
Let r be the radius of the base, the height, V the volume and S
the surface-area of the conical tent.
Then, V=-tr 2 h ... (1)

and S=nr '1r 2 +h 2... (2)

Here, V is given as constant.

S2=7t2r2(r2+h2)=m2r2(r2+2_.) lfrom(i)J

=n 2 r 4 +9V2.4.

Now, if S is a maximum or a minimum, S2 is so and herce for

of d1 2 \ . d 24
maximum or minimum S. —=O.
S ) I.e. - it r +9V.—)=O,
dr drt r2

I Vi
i.e., 47t2 r3 _18V 2 . _L0. .. r6 9V
._L; r
21E2

S2 )= 12it2r2 +54V 2 . _.-, which is positive for

Now, 7( r4
rT
 328
DIFFERENTIAL CALCULUS

for minimum amount of canvas,

.,n2r4h2
i.e., r6=
27C 2
IV 2
i.e.,r2=h2,.•. r2:2=i or r:h=1;.
Ex. 10. Show that for a given perimeter, the area of
a triangle is
maximum when it is equilateral.

The area of triangle A BC ..js(s— a )(s— b)(s— c):

The area A of triangle A BC
Let s— a=x, s— b=y, .c—c=z,
x+y+z 3s_(a+b+c)3s_2s=s=cosE
Now, A = Since s is a constant, A will be maximum when xyz
will be maximum subject to the condition x + y + z = const, i.e., when
x=y=z, [See §10.41
i.e.,when s--a=s-j=scieab=c
Ex. 11. Show that the maximum triangle which can be inscribed in a
circle is equilateral:
Area A of a triangle A BC inscribed in a circle of radius R
= bcsin A = - . 2Rsin B. 2Rsin C.sin A

=2R 2 sin Asin Bsin C

= R 2 {cOs (A - B)—COS (n + n )}sin c.

Let us suppose C remains constant, while A and B vary. Since R is

Constant, the above expression will be maximum when A = B.
Hence, so long as any two of the angles A , B, C are unequal, the
expression 2R 2 sin A sin B sin C is not a maximum, that is, it is maximum
when A =B=C.
Thus, A will be maximum when A = B = C.
Ex. 12 Find the maximum and minimum values o
asin x+bcos x.
Let arcosO, b=rsinO,
so that r 2 = a 2 + b 2 and tan =h/a.
MAXiMA AND MINIMA 329

Thus asinx+bcosx.=r(sinxcoso+cosxsino)=rsin(x+e)

4 2 +b 2 sin (x+tan'b/a)
Since the greatest and least values of sine of an angle are I and—I,
the requiredmaximum and minimum values of the given expression are
,Ja2+b2 and _Ja2+b2

EL 13. Assuming Fermatc theornz that a ray of light in passing from

a point A in one medium to a point B in another medium takes the path
for which the time of description is a minimum, prove the law of
refraction.

Fig 10.5.1

Let AOB be a possible path of the ray of light, 0 being the point where
it meets the surface of separation MON of the two media and let POQ be
the normal to the common surface MN at 0, and AM, RN the perpendiculars
from A and B on MN.
Let )n L4OP 0, mZBOQ = 0, and let v and v' be the velocities
of light in the two media. If AM = a, BN = b, then AO=a see 0,
BO = b sec 4' . The time taken by the ray of light to travel the path AOB is
asecO bsec4'
T= + ... (I)
V V
and by Fernsat's theorem this is to be a mini mum.
Again, since A and B are fixed points,
atanO+btan$=MO+ON = MN = constant ... (2)
so that 0 and • are not independent, and we can, thus consider$ as a
function of 0 . which is then the only independent variable.
330 DIFFEREtI MI. cALCULUS

For T to be minimum, .i
dO
= 0, giving
a b d4t
- cse0 tan 0 + - sec 4, tan 4 - = 0,
V
Also, from (2),
asec 2 O+bsec 2 4t—=0.
dO
di
From these two, eliminating ,wc :isily get
dO
sinO. si n4, sin v
or, - = = t (say)
V v sin4, v
which is the law of refraction, satisfied for the actual path of the r;.v of
light.
10.7 Miscellaneous 'Worked Out Examples
Ex. I. (i) What do you mean by the maximum or minimum value ofa
function 1(x) at x = c? If f '(x) exists, what willbe the value of
f'(c)? Is it neccessary as well as sufficient conditio&i
[C. P. 1987, '96, B. P '96. 98
(ii) Cite an illustration to show that even if f'(c) does not exist.
f'( - v) may have a maximum or minimum at x = c I C. P 1987
Solution : (i) A function 1(x) is said to have a maximum ( or, a local
maximum) at x = c, iff (c) is greatest of all the values, i.e., 1(x) ^s f(c)
in some suitably small neighbourhood of c.
Analytically, this means
f(c+h)–f(c):^0, for I i
sufficiently small.

Similarly, 1(x) is said to have a minimum (or, a local minimum) at

X = c, if 1(c) is smallest of all the value', i.e., f(x)^ 1(c) in some

suitably small neighbourhood of c, i.e.,

f(e + h) - 1(c) ^: 0, forj h I sufficiently small.

Second part: If 1(x) be a maximum or a minimum at x = c, and if
f'(c) exists, then f(c) =0.
Proof: By definition, f(x) is a maximum at x = c, provided we can
find a positive number & such that
MAXIMA AND MINIMA 33

f(c+h)—f(c)<O, whenever —ö< i<8 (h#O)

f(c+h).-f(c) <0
if h is positive and sufficiently small and
h
f(c+h)—f(c) >o
if h is negative and numerically sufficiently
h
small.

Thus urn 1(c)—j'()

h

f(c+/s)—f(c)
and ,l,

Now, if f'(c) exists, the above two limiting values which represent
the right-hand and left-hand derivati'es respectively of f(c) at x = c,
must be equal.
Hence, the only common value of the limit is zero, i.e.,.f'(c) = 0

Similarly, when f(c) is a minimum, f'(c) = 0.

Third Part : f'(c) =0, if it exists, is a necessary, but not sufficient
condition that f(x) may have a maximum or minimum value at x = c.
As an example, let us consider the function f(x) = x5.
Obviously .f'(0) = 0, but whenever x>0, 1(x)> 1(0). and whenever
<0, 1(x) <f(o). Hence, f(x) has neither a maximum nor a minimum
value at x=0 -
(ii) Let us consider the function 4t(x) = N
Obviously, 4(x) is minimum at x = 0, but $'(0) does not exist.
Ex. 2. (i) Show that the maximum value of ry subject to the condition
•25
3x+4y=5 is j .[ C. P 1991, B. P 1996

(ii) Show that the function 1(x) = x 3 — 3x + 6x + 3 does not

possess any maximum or minimum value. f C. P 1994 1

Solution : 3x+4v=5,

Let U = xv = !x(5._3x) = .!(5x .3x2).

332 DIFFERENT/A L CA LCULUS

du I
—(5-6x)

d2u 6 3
and -=
d,, ---= --.
4 2
For a maximum or minimum of U =

du 5 5 d2u
=•O, which gives x = Also, at x = , •- <0

5
u = ry is a maximum at x = and the maximum value of

I1
21 - 25
u=.y=-5x--3I( )
6 6 48

(ii) f(x)=x3-3x2+6x+3

f(x)=3(x2_2x+2)

f'(x)=O gives x 2 -2x+2=0,or, x=l±i.

Thus f'(x) does not vanish for any real value of x.

Hence f(x) has neither a maximum nor a minimum.
Ex. 3. When does the function sin3x-3sinx attain its maximum or
minimum values in (0, 2it)? [ C. P 1981
Solution : Here, f(x)=sin3x-3sinx
f'(x) = cos 3x-3 cos x
f"(x)= —9 sin 3x + 3 sin x
For, maximum or minimum values of f(x),
f'(x) = 0, which gives 3(cos3x —cosx) = 0
or, 4cosx (COS x_I)=0
cosx=0, I, —I
t 3i
x= —, -- (.• 0<x<2t )
 MAXIMA AND MINIMA 333

it .3n .it
At x= —, f"(x)=-9sin—+3sin— >0

3m .9it .3m
At x= - , f (x)=r-9sin----+3sin_ <0

3m it
f(x) is maximum at X = and minimum at X

Ex. 4. (i) Show that of all rectangles of given area, the square has the
smallest perimeter. [ C. P 1984, 2008 1
(ii) Show that, of all rectangles of given perimeter, square has
the largest area.
Solution : Let x be the length and y be the breadth of a rectangle.
Its area = xy = k say, where k is constant.
k
y=—
x

If S be the perimeter of the rectangle, S = 2 (x + y) = 2 +

dS ( k '
— =2,1----
X2
and d2 4k
dx j*2

dS
For maximum or minimum of S, = 0, which gives
dx
k

• when : = , ( 2k = 2 10
d

• So, S is minimum when x = , and when jr = ' J, y =

(.. y=k),
i.e. x=y.
Hence the perimeter of the rectangle is smallest when the rectangle is
a square.
(ii) Let, x be the length and y be the breadth of a rectangle--
Perimeter of the rectangle =2(x+y)=2k , say, where k is
Constant.
y=k— x
 334 QIFEERMTIAL CALCULI'S

Area A of the rectangle is given by

A=xy=x(k-x)=kr-x2
dA d2A. -
—=k-2x and —=-2
dx
dA
For maximum or minimum of A, - = 0 , which gives k - 2x = 0.
dv
i.e., x=k.

I d2
for, x=- -k, —=-2<O
2 2

A is maximum when .x-k

1
when x.-k, y= -k thus x= V.

Hence, the rectangle of given perimeter has largest area when it is a

square.
Ex. 5. (1) Show that the rectangle inscribed in a circle has maximum area
when it is a square.
(ii) Find the largest rectangle that can be inscribed within the

v2h2
ellipse I C. P 1993 I
Solution Let ABCD be the rectangle
inscithed within the circle of radius a.
0 C

A B C=,

Let Z CA B = 8, then
/0
AC is a diameter of the circle.
- A B

AB = ACcosE3 = 2acosO and

BC=2asin0.
Area S of the rectangle ABCD is given by
S = ABxBC = 42 sinOcos0= 2u 2 sin 20
(IS 2 dS 2
= 4a cos 20, = -8a si n 20
dO dO2

For, .extrcmum of S, - = 0 , i.e., 4u2 cos20 = 0

dO
MAXIMA AND MINIMA 335

7t It
20=—. [ •.• ]
O^O^! i.e., 0=—.
2 2
For, o=! , _...8a2sin!E_8a2<O
4 dO 2 2

Sis maximum when 0-.

Then, AB=2acos!=,ha and BC=2asin!=,Ja

AB = BC, the rectangle inscribed in the circle with largest area is

a square.
(ii) Let ABCD be the
rectangle inscribed in the ellipse
X2
--+--= I. The length and
a
breadth of the rectangle are
parallel to the axes of the x
ellipse which are coordinate
axes also.
Let coordinate of A be
(acosO,. bsinO).

S = area of the rectangle ABCD = 4 X area of the rectangle OPAQ

=4OPPA=4acosO.bsin9=2bsjn20

= 4abcos2O; dS = —Sab sin 20

cis

dO dO2
dS
For extremum of S. = 0, which gives cos29 = 0
dO

20=,j.e, o.=!.
2 4
d2S
For, 0 =-
7E , = —8absin = — 8ab <0
7C

4 dO 22

S is maximum when
4

In that case, length of the rectangle = 20P = 2acos

TE
= ia
 336 DIFFERENTIAL CALCULUS

and breadth of the rectangle = 2PA = 2bsin .!

4
=
and area of the largest rectangle = 2ab.

Ex. 6. (i) Show that the maximum value of log(

x2 is
xj
I 2e
C. P 1989]

(ii) Show that x (x > 0) is a maximum at x = e. Deduce that

e'>7t.
[ C'. P 1992

Solution : (i) Let , f(x) = x2 log( — )= -x2 log X.

XI

f'(x)=-2xlogxx 2 .!= -2xlogx-x

and f"(x)2logx2x.t1-2Iogx_3

For, extrenlum of f(x), j"(x) = 0, which gives -x(2logx+1) =0,

i.e., logx=_ ! , ..x=e 2

At x = e 2,

f(x) is maximum for x = e 2

Maximum value offlx)= fIe2 loge2

J= J=
(ii) Let /(x) =

log f(x)=J-.logx

Differentiating, = iL!- ._!Iogx = -4(i-iogx)

f(x) X . X x x

f'(x) = f( x ){4(1 - logx)}

MAXIMA AND MINIMA 337

f"(x) = f'(x){ -4 (l - logx)} + f(x){ __j.+4logx}

For, extremum of f(x), f'(x)=O, which gives

-4(l— Iogx)=O or, logx=1=loge, .. x = e

f"(e) = f'(e){ 4(l - Ioge)} +f(e){ _4 +..-. ioge}

A1

e
f(x) is maximum at x=e.

and maximum value of f (x) is e.

Since f(x) is maximum for x=e.
f(e)>f(7t)

Of. e r .> 71"

i.e., e >7t. Le, 7r

7. Find the point on the

parabola 2y = x 2 ,which is
nearest to the point (0, 3).
C. P 1990, 1997

Solution Let P(x, y) be

any point on the parabola
0
Y = 1 X2 and A (0, 3) is the fixed point

A P2 =(x_O)2+(y_3)2 =x2+[_3J =f(x) (say)

f '(x )2x +2x (_3) =x3-4x

f"(x)= 3x 2 — 4.
when A P is minimum, A P = f(x) is also minimum.
22
338 DIFFERENTIAL CALCULUS

For maximum or minimum of f(x), f'(x) = 0,

which gives, x(x2 —4) .. 0

i.e., x = 0, 2, —2 at x = 0, f "(x) = — 4<0

at x=±2, f"(x) = 3(4) —4 = 8> 0

f(x) is maximum when x=Oand minimum when x=±2.

AP is minimum when x = ±2. When x ±2, y = 2.

Hence, the points on the given parabola, nearest to the point (0. 3)
are (2, 2) and (-2, 2).

Ex. 8. Prove that the function f(x, y) =x 3 +3x 2 t-4xy+y 2 attains a

minimum at the point I . P /9901

(, — i).
Solution : f(x, y) = x3 +3X2 +4 +

f=4v+6x+3x 2 . f. =6+6x, f=6x+2y , fr = 2, f,. =4.

For, extremum of f(x. y), f 0, fy =0, which give

4v+6x+3x 2 =0 (I)

4v + 2v = 0 i.e., y —2x ... ... (2)

Now, from (I), - 8x + ôx + 3x 2 = 0

or, .v(3x —2) = 0, :. x = 0,

2 4
From (I) when x=0, v=0. and when x—

At ft., =10, =2, f,=4

So, fxJ—(f,) = 10x2—(4) 2 =4>0

(2 4
Thus f(x, y) has an extremum at '

Also, .f=l0>0
(2 4
hence, f(x, y) has a minimum at
j ' -
 MAXIMA AND MINIMA 339

EL 9. Find the extreme value of f(x, y) = 2x2 - xy + 2y 2 - 20x.

I C. P 2000 1
Solution: f(x, y) = 2x 2 —.*y+ 2y 2 - 20x
f =4x—y-20, 4 =—x+4y , f, 4. f_ 1
For extremum of f(x, y), f = 0, f,. = 0
Le., 4x—y-20=0, —x+4y=0
16 4
which gives X= -j-. Y =

Again at 16 4
-
l 3 3

f- - f = 16-1=15>0
and f=4>0
Hence, f(x,y) has a minimum value atand
256 64 32 16 160
f =2x---+--20x—=—_
9 9 9 3 3•
Ex. 19. (i) A wire of length Fis to be cut 'into two pieces, one being
bent to form a square and the other to form a circle. How should the
wire be cut if the sum of the areas enclosed by the two pieces to be a
minimum?
(ii) A wire of length 20 metre is bent so as to form a circular
sector of maximum area. Find the radius of the circular sector.
[C.P1983..961
Solution : (i) Let the wire be cut in two pieces ii the ratio 1: A. Then
lengths of the pieces will _ and respectively.
_-_
be-1^x 1+x
If the length of each side of the square be x and the radius of the
circle r, then
I I
4(1+ A)

IA
and 2,tr=—, I.e., r=
2n(1^A)
 340 DIFFERENTIAL CALCULUS

Let A be the sum of the areas of the square and the circle.
nil k2 12(71+4x2)
12
+=
16(1+A) 2 4 2 (1+X)2 167t(1-t-A)2
(4A — x)
. (1)
dA8it (H-A)3

d 2A 12 (41+37t-8X)
and -=----------------
2
... ... ()
8it (li-A)4

For maximum or minimum value of A , dA = 0,

which gives from (1) A = TE

Also, from (2), when A ,

dAl 2 (4li-3x-21t) 12(41+x)

- 8it (1^A)4 81t (l+A)4
It
Hence, A will be minimum, when A

In that case, the lengths of the pieces of the wire will be----j- and

It
lx—
__.4 respectively,
it
1+ —
4
41
i.e., - and respectively:
4+jt 4+n
(ii) Let, OAB be the circular sector
formed by the wire of length 20 metre, r
metre be the radius of the circle and
Z AOB = 0, where 0 is in circular
measure;
then 2r+s=20, i.e., s=2(10— r)
(I)
MAX/MA AND Al/N/MA 341

Area of the sector A OB = S = - 126

r 1 r2 x± [s=rOJ

=—r .2(1O— r)

or, S=10r—r2
dSd2S
-=1O-2r and -=--2
dr dr-
dS d
2
For an extremum of S, _O,.which gives r=5 and -=-2< 0
dr dr2
Hence, S is maximum when r =5 metre, i.e., radius of the circular,
sector is 5 metre.

EXAMPLES .. X
1. Find for which values of x the following functions are maximum and
minimum:
_ 9x 2 + 15x —3. (ii) 4x 3 —15x 2 + 12x —2.
(1) X3

x 2 — 7x+6 [C.P.1939]
x-10

(iv) x 2 +x+
- 8x 3 + 22x 2 24x + 5.
x2—x+1

2. Find the maximum and minimum values of (iii), (iv) and (v) of Ex. 1.

3. . (i) Show that the maximum value of x± is less than its minimum
value. BY 1990. V P 20001

)()
(2x—1 x-8
(u) Show that the minimum value of is greater
(x—l)(x-4)
than its maximum value.

4. Show that x 3 - 6x 2 + 12x —3 is neither a maximum nor a minimum

when x=2.
5. Show that the following function possess neither a maximum nor a
minimum:
342 DIFFERENTIAL CALCULUS

(i)x 3 -3x 2 +6x+3. (ii) x3-3x2+9x-1.

(iii) sin (x+a)/sin(x+b). (iv) (ax +b)/(cx+d).

6. Show that x 5 —5x 4 +5x 3 -J is ncjflier a maximum nor a minimum

when x = 3; neither when x = 0.
7. Examine for maxima and minima of the following functions:

(i) sin x. (ii) cos X. (iii) X5.

(iv) x 6 . ( v) + x40(vi) e'. sin x

•k x5 -
8. Test the following functions for maxima and minima at x = 0:

X3 2
(I) sinx — x+ —------- (ii) cosx-1+----.
3! 5! 2! 4!
9. Show that

(i) 'Jsin x + 3 Cos x is a maximum for x = fir.

(i) sin x(l + cos x ) isa maximum for x= 31 n. [C.P 1942, 47]

(iii) Sin x cos x is maximum when x 1/3 it

(iv) x sinx + 4 cos .v is maximum for x= 0.
(v) sec x + log cos 2 x is a maximum for x= 0 and a minimum for

X=

(vi) sin 20 (o > o) js maximum when 0 ,r.

10. If y is defined as a function of x by the equations

y = a(l — Cos 0 ), x = a(0 - sin ),
show that y isamaximum when 0 = it.
11. Show that

(i) the maximum value of (i /x ) is e•

[CY. 1990, B.? 1995. VP 2002]
(ii) the minimum value of x/log x is e.

(iii) the minimum value of 4e 2 + 9e 2 is 12. [C.? 1994]

MAXIMA AND MINIMA 343

12. (I) Show that 4 - 8x log e 2 is minimum when x I.

(ii) Show that l2( log x + i)+ x 2 - lOx + 3 is a maximum when

x2 and a minimum when x = 3.
(iii) Show that x2 log (lIX) is a maximum for x = I /vre

13. If f ' (x) = (x —' a) 2'" (x - b)2"', when m and n arepositive

integers, show that x = a gives neither a maximum nor a minimum value
of f(x), but x = b gives aminimum.

14. Find the maxima and minima, if any, of

(x—I)(x— 3)

ax+b
15. If y = has a turning value at (2,—l).find a and b
x—l)x-4)
and show that the turning value is a maximum.

)2
16. Prove that Y, (x - a1 is a minimum when x is the arithmetic

meanof a1,a,,a3.... . a,,.

17. (i) Given x/2 + y13 = I , find the maximum value of xy and minimum
value of x 2 +y 2 . -
(ii) Given xy = 4, find the maximum and minimum values of 4x + 9y.
I VP 2001]
18. (i) If f(x) = - /, whenthesquarerootistobetakenpositive,
show that x = 0 gives a maximum for f(x).

(ii) If f(x)=a+(x_b)t+(x_b)4. show that f(x)isminimum

for xb.
(iii) Show that (x—a)(2x—a) is a maximum for x =-a,a
minimum for x = a and neither for x=a. [a>0J

(iv) If 1(x) = x l, show that f(0) is a minimum although

f'(0) does not exist.

19. Show that

(i) the largest rectangle with a given perimeter a square;
344 DIFFERENTIAL CALCULUS

(ii) the maximum rectangle iiiribable in a circle is a square.

[C..? 19361
20. Find the point on the parabola 2y =X2 which is nearest to the point (0, 3).

21. P is any point on the curve yf(x) and C is a fixed point not on the
curve. If the length PC is either a maximum or a minimum, show that
the line PC is perpendicular to the tangent at P.

22. Find the length ofthe perpendicular from the point (0, 2) upon the line
3x + 4Y + 2. = 0, showing that it is the shortest distance of the point from
the line. Find also the foot of the perpendicular.

23. A cylindrical tin can, closed at both ends and of a given capacity, has
to be constructed. Show that the amount of tin required will be a
minimum when the height is equal to the diameter.
24. By the Post Office regulations, the combined length and girth of a
- parcel must not exceed 3 metre. Find the volume of the biggest
cylindrical (right circular) packet that can be sent by the parcel post.
25. A line drawn through the point P (I, 8) cuts the positive sides of the
axes OX and OY at A and B. Find the intercepts of this line on the
axes so that
(i) the area of the triangle OAB is a minimum;

(ii) the length of the line AS is minimum.

Find also in the above cases the area of the triangle and the length of
the line respectively.
26. P is point on an ellipse whose centre is C, and Nis the foot of the
perpendicular from C upon the tangent to the ellipse at P; find the
maximum value of PA'. [C..P. 19451
27. The height of a particle projected with velocity u at an angle a with
the horizontal is u sin at - 4 g! 2 at any time I. Find the greatest
height attained and the time of reaching it.
28. The total waste per mile in an electric conductor is given by

W = C2R+__K2,where C is the current, R the resistance, and K a

R -
constant. What resistance will make-the waste a minimum ifthe current
C is kept constant.?
MAXIMA AND MINIMA 345

29. The force F exerted by a circular electric current of radius a on a

magnet whose axis coincides with the axis of the coil is given by
5
Foc x(a2 +x2)
where x is the distance of the magnet from the centre of the circle.
Show that F is greatest when x = - a.
30. Assuming that the intensity of light at a point on an illuminated
surface varies directly as the sine of the angle at which the ray of
light strickes the surface, and inversely as the square of the distance
of the source from the point, find how high should a light be placed
in
directly over the centre of a circular field of radius 1 5,F2 in order
to have a maximum illumination on the boundary.

31. (i) Find the altitude of the right cone of maximum volume that can
be inscribed in a sphere of radius a.
(ii) Find the altitude of the right circular cylinder of maximum volume
that can be inscribed in a given right circular cone of height Ii.
32. (I) For a given curved surface of a right circular cone when the
volume is maximum, show that the semi-vertical angle is
-I
sin

(ii) For a given volume of a right cone show that, when curved
surface is minimum, the semi-vertical angle is sin - ' -.-

33. An open tank of a given volume consists of a square base with

vertical sides. Show that the expense of lining the tank with lead
will be least if the height of the tank is half the width.

34. If POP' and QOQ' be any tow conjugate diameters of an ellipse,

and from P and Q are drawn two perpendiculars to the major axis
cutting it at M and N respectively, show that PM + QN is a
maximum when POP' and QOQ' are equi-conjugate diameters.
35. A window is in the form of rectangle surmounted by a semi-circle. If
the total perimeter be 10 m, find the dimensions so that the greatest
possible amount of light may be admitted.
 346
DIFFERENTIA L CA LCULUS

36. A particle is moving in a straight line. Its distance x cm from a fixed

point 0 at any time t second is given by the relation

x=1 4 10: 3 +24t 2 +36i-+12.

When is it moving most slowly?
37. In enclosing a rectangular lawn that has one side along a neigh-
bour's plot, a person has to pay for the fence for the three sides on
his own ground and for half of that along the dividing line. What
dimensions would give him the least cost if the lawn is to Contain
4800m2?
38. A gardener having 120 in of fencing wishes to enclose a rectangular
plo't of land and also to erect a fence across the land parallel to two
of the sides. What is the maximum area he can enclose?
39. A shot is fired with a velocity u at a vertical wall whose distance
from the point of projection is x. find the greatest height above the
.jevel of the point of projection at which the bullet can hit the wall.
40. From the fixed point A on the circumference ofa circle ofraditis the
perpendicular AY is let fall on the tangent at P. Show that the maximum

area of the triangle API is IC..P '9301

41. The intensity of light varies inversely as the square of the distance
from the source. If two lights are 15 in and one light is 8 times
as strong as the other, where should an object be placed between
the lights to have the least illumination ?
42. The boundary wall of a house is 2-7 in and is at a distance 80
cm from the house. Show that a ladder, one end of which rests on
the ground outside the wall and which passes over the wall, must at
least be 1-3 A3 in in order to reach the house.
43. A man in a boat vF
3 km from the bank wishes to reach a village that
is 11 km distant along the bank 'from the point nearest to him. He
can walk 8 km per hour and row 4 km per hour. Where should he
land in order to reach the village in the least time ? Find also the
time.
44. If for a steamer the Consumption of coal varies as the cube
of its
speed, show that the most economical rate of steaming against a
current will be a speed equal to 1 1 times that of the current.
MAXIMA AND MINIMA 347

45. For a train the cost of fuel varies as the square of its speed (in km
per hour), and the cost is Rs. 24 per hour when the speed is 12 km/h.
If other expenses total Rs. 96 per hour, find the most economical
speed and the cost for a journey of 100 kilometre.
46. Assuming Fermat's law, that a ray of light in passing from a point A
to a point S in the same medium after meeting a reflecting surface
takes the path for which the time is minimum, prove the law of
reflection.
47. Assuming the law of refraction, if a ray of light passes through a
prism in a plane erpendicular to its edge, prove that the deviation
in its direction is minimum when the angle of incidence is equal to
the angle of emergence.

ANSWERS

1. (i) x= 1 (max.), x= 5 (mm.), (ii) x = - (max.), x = 2(min.),

(iii) x=4(maL), x= 16(min.), (iv)x= 1(max.), x=-1(min.),
(v) x=1(max.), x=2(min.), x=3(min.),
2. Max. value = I, mm. value = 25 for (iii),
Max. value = 3. mm. value = for (iv),
Max. value = —3, and mm. value = —4 in both cases for (v),
7. (i) x=(2n+-)ir(max.), X=(2n—.+)iL(min.),
(ii) x=2nit(max.). x=(2n+ 1)n (mm.),
(iii) Neither max. nor mm. (iv) x=Ogives minimum
(v) x=0(max.), x=1 (mm.),
(vi) x = 2nit+ it (max.), x = 2mr -- it (mm.),

8. (i) Neither max. nor min. (ii) Max. for x=0.

14. Mm. for x = 0, max. for x = . 15. a=], = 0.

17. (i).
2 X,
13 (ii) Max. value = --24; ruin, value = 24.

20. (± 3.2), 22. 2 units; (- . ). 24. . m,

25. (i) :2. 16, (ii) 5, l0.Area fo the triangle in (i) = 16 sq. units;
length of the line in (ii) = 5,F5 units.
348 JflFFERENIIAL CALCULUS

26. a --b. 27. (a2 sin' a )/(2g ); (u sin a)! g.

28. - units. 30. 15m.

31. (i) A a (ii) -h,

35. Height of the rectangle = radius of the semi-circle.
36. At the end of 4 second. 37. 80m x 60
38. 600m2, 39. (11 g2x2 2U 2g ).

41. 10 m from the stronger light.

43. 1 km from the point nearest to him; I - hour.
45. 24km/h; Rs. 800.
 INDETERMINATE Fouuis (Evaluation of certain limits)
[no
11.1. The limit of o (x )/i, (x ) as •x -, a is, in general, equal to the
quotient of the limiting values of the numerator and denominator [see Rule
(iii) of A rt. 3.7 1, but when these two limits are both zero that rule is no
longer applicable since the quotient takes the form 2 which is meaningless.
We shall consider in the present chapter how to obtain the limiting values
of the quotient in such cases, and also the limiting values of the quotient in
such cases, and also the limiting values in other cases of meaningless forms,
apparently arising out of the indiscriminate use of the rules of Art. 3.7. The
name 'indeterminate forms', as applied to these cases, is rather misleading
and vague. -

11.2. Form (P Hospitals theorem)

if 4' (x ), V (x) as also their derivatives 4,'(x ), W '(x ) are

continuous at x = a, and if 4, (a )= iji (a ) = 0 [ i.e., ii 4 (x)
x —. a
= Li ,(x)=0 ],then
1-. 3 (2

4,(x )_ Ø'(x)
0'(')
(x)V, a)
provided i (a )* 0.
Since Ø(a )= 0 and (a)= 0,we have
Ø(x)=Ø(x)-4,(a) and
Now, by the Mean Value Theorem,
0(x)— 0(a) = (x— a) O'Ia +0(x—a)}, 0<0 1 <I,

W(x) — '(a) = (x—a) 4/{ a + 02(x— a) } , 0<02<1.

0(x) - 4,(x)-.-Ø(a)
(x) ,(x) — ,( a) I{ a+O2(x— a)} '

Li
ji ( a) -*al/i (x)

provided W '(a )# 0.
 350 D//-FLREN7/AL CA LCIJLUS

Generalization:
Incase O'(a ) and 41'(a )are both zero, applying the above theorem
again, we get
O'(x) L O(x)'(a)
-.at ' (x)
-.ai(x) t(a)'
provided Ø'(a ) and i,u(a ) are continuous air = a, and i(
a .O.
If, however, ' (a ) = (a ) = 0, then we again apply the above theorem
and obtain the limiting value as 0(a )/ w(a ), and so on.
[For illustration. see Ex. I, Art. 11.81
Note 1. The above result can also be established by Cauchy's Mean
Value Theorem. [See Ex. 7(a), Art. 9.7 1
Note 2. In the theorem of this article if x send.., to instead of a, then
the substitution 1/1 for x would reduce it to the above form when
tends to zero.

11.3. Form .

If Li O(x)= oo and Li t(x)=o,andzf Li 0'(x)

x -, a ._. a x —3., tv (x
exists then ii 0 (x ) will also exist, and it. value is equal to the
—aji (x )
former limit.

Let Li (-' = Then we can determine a positive number 8,

—a 07
1(7)
such that jn the interval a — 5 < x < a + x A aJ, isasneal-to /
(x)
as we please. Also, since the limit exists, it follows that for x sufficiently
close to a [hut ^, a.] 0' (a ) and tj" (a ) roust both exist, and
a )7^ 0 there.
Now first consider the interval a <x :5 a + 8. and x., be any particular
value therein, and take another value x such that a <x < x.,.
Then, by Cauchy's Mean Value Theorem [see § 9.7, Ex. 7(a)],
o(
—
(
)—Ø(x) -

where x0 and so a a+
INDETERMINATE FORMS 351

1o(x) ii
Okx /l x) J o'()
Hence.
MI
____

0( x )W( x ) ø'()
(1)
vi ( x ) o(x
0(x)
Now keeping x 0 fixed, if we make x-+ a, MI (x ).3oo and
1O(x o ) 0-1
fl
/fw(xo) i
Ø(x)_9oo, and so ' i'
y—' (x)

Also, ,7 is as near to / as we like, by a proper choice of S.

MI ()

Hence, from (1), ( x) is a bitrarily close to 1, as x -, a + 0.

I/i (x)

•
Thus,.
-+o MI ( )
Similarly, considering the interval a- 6 <x < a, and proceeding

exactly as before, we get Li

0 X =1.-
-.--o , (xii)
Ø'(x)
Hence, Li O(x)= Li
MI (x ) MI (x )

We can also prove a modified form of the above theorem as follows:

If Li Ø(x and Li ip(x) are both infinite, then

-11

Lt \X / ( when it exists) = Lt
y x

LI
0(x)
Li
l/MI( X )
x)
f
r.-g(x)
say.

[where f(x)=1/MI(x) and g(x)= 1/0(x)]

which, being of the form -2 ,- F see Art, 11.2, above

 352 DIFFERENTIAL CALCULUS

2
(') w'(x) i(x)
Lt
Ø'(x)

^. 2
0(x) uI'(x)jx w(x
(I)
f(x) -.'(x) 0(x)
(x)
Now, Let ii =1 ... (2)
i.."
Three cases arise:
Case!. I is neither zero, nor infinitely large.
Dividing both sides of (I) by 1', we obtain.
I V, ( X
07( ..1,i.e.. LI .
I V(X )
/ x) 41 ( x )
Case!!. 1=0.
Adding 1 to each side of (2),
X
41(x) 41(x)
by case (l)]
0'(x)
=Lt +1.
X-

i
0(x ) ____

£-41IX) '-''t/1X

Case HI. When I is infinitely large,

141(x)1I/J(x)
1.1
'.-'{O(x)/4I(x)} . —. 0( x ) £-.O'(x)

0(x)
Li
1(X) 41'(x)'
Hence, the theorem is proved in all cases.
For illustration see Lx. 3, Art. 11.08.

Note 1. Theorem is evidently true also when one or both the limits
tend to -

Note 2. By substituting x = lit, it can be shown that the theorem is

also true when x tends to instead of a.
 INDETERMINATE FORMS 353

11.4. Form Oxo.

Such forms arise when we want to find the limiting value of

We can write

Øx ),Ijf x )- 0('),
/ , or, /
JJx
when being of the forms 0/0 and o/o, as .can be evaluated by the methods
of A rts. 11.2 and 11.3.' [See Illustrative Lx. 4, A rt. 11.8]
11.5. Form 00 - 00.
Such forms arise when we want to find the limiting value of
Ø(x)-i (x ) as x-+ a where (.x )_-) and i ( x )-4-as.
We can write

( \ I \
øk X ,V iX ) 1/(x ),(x )}
which being of the form 0/0 can be evaluated by the method of Art. 11.2.
[See Illustrative Example 4, A rt. 11.8 1

11.6. Form 0 0 . 000. i.

These forms occur when we want to evaluate the limits of functions of
the form fj I x;\Iw ( ) as x -*a,
When (i) both (x) and , (x .)-, o as x —4 a;
(ii) 0 (x)_400 and,(x)--.O asx-a
(iii) O(x)-41 and f(x)_±ooas x -a,

If(x)>O, let yr b(x)r log() ,... y=r(x

)log O(x)
Li log y reduces to the form discussed in Art. 11.4, and, hence,
can be evaluated.
Since Li log y = log Li y the required limit Li y can be obtained
See Illustrative Example 6, A rt. 11.8.1
11.7 Use of power series.
In evaluating limits of certain expressions it is sometimes found
convenient to use the expansions of known functions in the expression in
power series in a finite form, and then to take the limit.
1 See Illustrative Example 7, A rt. /1.8. 1
23.
354 DIFFERENTIAL CALCULUS

11.8. illustrative Examples.

E.- 1- If $( '(a )4"(a (a) and (a ),ty'(a

iJ'(a L...,tI"'(a) are all zero, and t j " (a ) * o. then

•(x)
uj" (x )

Put x = a + h so that, when x-,a, h-*0,

Now, by Taylor's theorem

a+h)=(a)+h'(a):(a)+....

+ t (a)+— (a+Ojlz)
(n-i)!

= ---(a+9jlt), where 0<0<1,

Similarly, 1lI(a1h)= !i__ 1l, A (a+O,h)where 0<02<1,

4i(x) 4(a+h) Ii" '(a+e1h)

Lt LI
-..ii(x) i-oti(a+h) -Oh',jj'(a+0h)

=1
•-"() qc"(a)'

provided 4," (x) and " (x) are continuous at x = a.

Ex. 2. Evaluate L-.o x-sinx

e e -2x
[ C.P. 20011

The required limit, as it stands, being of the form 0/0. [see § 11.21

Lt e' + e — 2
= {form ]
-.O l-cosx 0
Ld e -e 0]
= form
-.o sinx 0
e+e
= Li =2,
-.o cosx
since. ii (e" + e ) = 1+ 1=2,..d Li cos .r = I.
-.0
INDETERMINATEFORMS 355

X4
Ex. 3. Evaluate Li -.
.t-500

The given limit, as it stands, being of the form can be written (by
§ 11.3] as
4x3( 12x2(
= LI —I form — 1 = Li —i form-
e' °' J -"' e'
24x( X''l 24
= Li —I form— 1= Li — =0
eT . ) '-"s e'

Ex. 4. Evaluate Li (l—sinx)tanx.

.T 4 ;lt

The given limit, as it stands, being of the form ox o, can be written as

l —nx
si [ fo 0
= Lt I rm —
cOtxL 0

— cos x
= Li =0
•-x—cosec2x
since cosxO and cosecx = I as x--)-,t.

( I
EL 5. Evaluate Li I 2—

The given limit, as it stands, being ofthe form ao — oo, can be written as

--- l [ 0
= Li — I form—
•.4Ix 4 _I L 0
I I
= Li
-Ix 2 +I 2

Ex. 6. Evaluate L1(cosx)c0v. [Patna 1933, VP. 1997J

The given limit, as it stands, is of the form 1.

)cfh.
Let y =(cosx
log y=Cot x log COS x=
• log
x2
COS

tan x
log Cos x 1 0
Now, La' logy = Li I form—
•-..o tan 2 x L 0
 356 DIFFERENT/A L CA LCULUS

—(sinx/cosx) = Lt (_-cos2x)
= Li
2 tan xsec x

=1(•.• Li cos2x=l
-•o
Since Li logy =log Li y, .. log Lt y = — .-.

Li = e'. the required limit e.

^ 0

X Sill X =I
Ex. 7. Show that Li [C.? 1932, 1995 1
.-.o x 6
Writing down the expansion of sin x in a finite power series, we have

I x 3 x5 S (n
x– sin x=x– . x--+— inI —+Ox , O<<L
3! 5! 2 )j
l
S
X x 5(Sit
=--- in
3! 5! 2

3 11 x 2 .(sit
=x < ---sini —+Ox
- [3! 5! 2

x–sinx i x?
sin i —±Bx
x 3! 5! 2

x–sinx I I x 2(5it '\1

Li = L i ---sini —+Ox I
x3 .-.03! 5! 2 3! 6'

(s . (5it
since — sin — + Ox --+0 as x sin x being :5 1.

Note. This being of the form 0/0 can also be obtained by the metkod of Art.
112.

a2 ++x2 2 — ax+x2
Ex. 8. Evaluate Li —

Multiplying both the numerator and denominator by

(a2 +ax+x 2 i,ja2 _ax+x2

 INDETERMINATE FORMS 357

and simplying, the required limit

-L i_______ _______
O2x(f.+x2+Ja2_+X2 )

a(I+JT )
...0(.12++2+42+x2)

Now, the limit of the numerator = a. 2 ,1a and that of the denominator
= 2a. Therefore the required limit = 'J.
Note. An algebraical or trigonometrical transformation often enables us to
obtain the limiting values without using calculus, as shown above, which
case belongs to the form 0/0.
sin 2x +a sin x be finite, find the value of 'a' and the
Ex. 9. If Li
-.o
limit. C. P. 1931. 1994, 2000, 2006 1
The given limit, being of the form 0/0,
2 cos 2x+a cos x
(by § 11.2.)
= -*° 3x2
When x -p 0, the denominator 3x2 = 0 hence, in order that the
limiting value of the expression may be finite, the numerator
(2 cos 2x+a cos x) should be zero, as x-90. :. 2+a=0, i.e., a=-2.
When a = -. 2. the given limit becomes
Ij sin2x-2sinx 10
I form-
-O L 0
2cos2x-2cosx 10
form-
-.O 3x2 L 0
-4sin2x+2sinx - 1
=Ls Ini -
form
6x L 0
Li -8cos 2x+ 2cos x -
6 - 6

ExtO. Evaluate Li (--'1 [C. P. 1947, 1994, 1997. V P. 1999

x
 358 DIFFERENTIAL CALCULUS

'Let u= Li
(lanx y .
X

I (tanx'\ (tanx)/X.
logu=-logIJ=log_
X x ) ,, x
tan x0
Since Li —=1 Li logu isoftheform
.-,o x

Li log u= Li log
0
tan X V X
'-to

x4
( x
x xsec2x_tanx'
tan x x2 (&' §11.2.)

2x-sin2x 2-2cos2x
=Lx =Li
-,0 xsin 2x .-.O sin 2x+ 2x cos 2x
4 sin 2x
L =0.
,-.o 2 cos 2x-4xsin 2x
Since Li log u = log Li u, .. log Li U = 0.
x-.0

Li u =
= I , i.e., the required limit I.
O .'• -t

Otherwise: Writing the finite form of the expansion of tan x by Maclaurin's

theorem,
tanx=x+-x 3 a where U=sccOx(l +2tan 2 Ox), 0<9<1.
I ( tan x I X+X2a I 1
Iogu=-logj---..-- 1=-log =-logI,)+x 2x
X x ) x x x
I \ I
= j—j--logI, l+ 5I x2 a)xcx=-log(l+v).xa,
xc(
where vrxa

When x-30,v---0,atso Lt
"-tO V
I Iog(l+v)=l.
Hence, Li log u= Li .!. log (l + v) Li (L xa
N-tO '-*0 V

Li log u=0. Hence, etc.

-.0
INDETERMINATE FORMS 359

It (e _i) tan 2x
Ex.fl. Evaluate
-,O

= e x -,_(tanx)21
LI
Given limit

e -1
- I tan
= Li -xl Li
-9O X l\ x )

Now, Ii !..IL!(l,eingof the form 'L Li f-. =e 0 l

o x 0) I

tan x
Also, Li
.. the required limit =, 1x1 ? =1.
.-.o x
Note. Such forms are sometimes called Compound Indeterminate forms.
In evaluating limits of such forms, the use of the theorems on limit (Art. 3.8)
is of great help.
11.9 Miscellaneous Worked Out Examples

x
I C. P. 1981
IX-1
i1i-EL1.valute:
im
Solution: l-.i -. i -; I
x logx - x + ii
=lim [Form
-4I (x-1)logx I
I l
Iogx+x I-1
x
= urn [By L'Hospitals Rule I
Iogx+(x_1)!j

xlogx
I Form -1
0
=lim Iogx+x-lJ
.-." x 0
 360 DIFFkRENTIA L CA LCULUS

lim(1+logx)
1+logx 1+0
=lim
-! 2+ log x 1im(2+logx) =j

I 2 1
--
EL 2. Evaluate: Urn j1ix x(er +i)j C. P 19821

Solution lIm i -2
I
._.ox x(i)J

erl
r I
= u •n
•--+O x (e' + [Form
0
• e
=hm
O1+(x+1)e

Jim er

m{l+(x+1)e} i=.

EL 3. Evaluate: Jim [C. P 1995 1

SiflX-X
Solution hm X3 Form -
0
0
=lim cOsx - 1
*O 3x2

-SiflX
Jim
-.o 6x [Form

-cosx
=Iim
-.0 6 6
EL 4. Find the value of

(I) Jim (cosx)?'. C. P 1989, 98 I

INDETERMINATE FORMS 361

(h) lim [C. P 1995 J

(iiOlim(1+x). [C. P. 1996 1

(iv) urn (C. P 1996 1

.-.0

(v) lim(I+sinx)".
.0
C'. P 1993 1.

Solution: (i) Let y = (cosx)1T

1logcos
logy =--•logcos.r=
X, [Base of logarithm is e]
x2
Ilogcosxl 0
urn log y=hm 2 [Foim -1
.-.o1 0
• —tanx
= urn [Form I
.-.O 2x

-sec 2 x -1
= u rn ---- = --
'-*0 2 2

M iog{uim y}=_!

urn y re2
'-$0

(it) Le

1 logx
or, logy — .logx—
I 1-x
• 0
urn {logy}= urn logx
- [Form
-$l I-x
• ( 1'\
=hm
-.1 x)
362 DIFFERENTIA L CA LCULUS

iog{iimy}=_i

urn y=e
-. I

urn =

(iii) lim (i+x)

Let , Z = - then z-40 as x — to'

x

Let y=(I+x)

then IOYZ.uO(I+_)

Thus urn{Iogy}= uimJz.iog(i+!i1

z-.o z-ot . z)J

lo
=çi)
(Form — I

1+
=hm
z-.O
kz
lim z as z—*O, z * 0 I
z-.O I + z
=0

Jog {Iim y}0

urn y.= e0 = I

Thus urn (I +X), =1.

!NDETIRMINA TE FORMS 363

(iv) Let. , = 2in

or, log y=2 sin x• log x I Base of logarithm is e

0
or, urn {logy} = urn 2logx [Form
—O COsecx

2
=lim
.--O -Cosecv cot
2sin2v 0
= urn (Form
-.o -xcosx
-2 Iiin 2sinxcosx
=
_.o C0X —.VSlflX

limsin 2x
-2' -0. =- 2 x=O.
Jim (cOsx — xslnx)

or, log {ujmy}=o

.0

i.e., urn y=e°=I

i.e.. Jim x 2 '"" =1

(v) Let y=(1+ sin x)'

or, logy = cot x log (J + sin X) I Base of logarithm is e]

= log (l+ sin x)

tan x

l 0
urn logy= urnog (1+ sin [Form
x-O.-.o tan x

cos x
= Jim
•—O S€2X

c0s3x
=Iim =1
.—.() I+sinx
364 DIFFERENTIA L CA LCULUS

iog{iim y}=i

or, limy=e'=e

x)cot.
IIm+ sin e

EL 5. Evaluate

lim (,an x )—x

(i) B. P 1995, C. P 1994. '97

ixJi
(ii) lim I C. P. 1990.

Solution: (i) Let, y(tanxJ

tim {log y} = tim {! lo)}

(tanx
0
= tim f Foim x-90, SIflCC as .%-40
-.o x
tanx ) , O
,i.e.. J

sec2.y
=
.-+o

=IimI .-
— .OLSiflxCOSx x1)

sIim 2x- sin 2,x 0

----------- [Form-]
-0 xsin2x 0
 INDETERMINATE FORMS 365

1-cos2x 0
=2km [Form
-.Osjn2x+2xcos2x
2.sin2x 0
2 urn
-.O 2cos2x+2cos2x-4xsin2x L Form
urn sin2x
=4x-=0
tim (4cos2x-4xsin2x) 4
.0

or.uog{Iim y}=0

Or, Jim y=e°=I

.-.0

(tanx
limI—I =1
-.o•x )
I
(ii) Let, y (sinxjF

sinx
1 sinx log—
or, log y=--
log . ------=
x x2
1sinx . (sinx
km —J=l, Jim log—J=0
x) -O 'J X

sin x
log— 0
urn (logy)= km X I FOrm I
-.o

xcOsx-sinx
= urn
2x
k• xcOSx-sinx 0
= m Form
-O 2x2 sinx
cosx-xsinx-cosx
=tim
-O 4xsinx+2x2cosx
366 DIFFERENTIAL CALCULUS

•
= urn
-stnx 0
Form
-o 2xcox+4sinx
= h• m -cosx
-.o 2cosx-2xsinx+4cosx 6
log {iim y} rr_!

or, urn y =e6

km(sinx
—i
•' x )
Ex. 6. Evaluate
Jimm
log (1-x)
(i)
cot (rr.x) I C. P. 1993

(ii) urn (cos ,)X' C. P 2002

km Iog(I --x)
Solution: (1) .-.i cot (ax) I Form

(2-.v)
=urn
-I -ltCoSe(' (ltr)

sin 0
= urn
-i IL( I Form

12 sin (1kv) cos(nx) In

= km = lim (-2 sin 2itx) = 0
nx(-1)

(ii) urn (cosmx)72

Let y=(cosmx)1
INDETERMINATE FORMS 367

then logy = _!- log(cosmx)

log(cesmx)
lint flog (Fo
Form 0
-]
,-40 x2 u
mn . tanmx 0
=-----hm (Form -
2-'o x
m2 n
=_—.Jim sec2mx I
--m 2 n
2-.o 1. 2

log {iimy}=_2m2n

or, lim y = e2

n
urn (cosmx)T1 =e2

x(l+acosx)-bsinx
Ex. 7. WFind a, b such that lim =
X-
C P 1990 1
sin 2x + a sin x
(ii) If lim is finite, find a and the value of the limit.
.-..o
I C. P 1994, 2000]

Here, Jim
• x(l+acosx).-bsinx
Solution: (i) (Forni 0

1(1+acosx)-axsjnx-bcosx
lim
3x 2-

Jim I+(a-b)cosx-axsjnx
= •.. ... (I)
3x
For (1) to be of the form 0
i.e., b=1+a •.. •.. (2)

l-cosx-axsinx 0
So, the given expression = Inn I Form 1
11
3x2
 368 DIFFERENTIAL CALCULUS

siflx — asiflx — LXCOSX 0

hm= Form
[
6x . 0

COx-aCOSX-aCOSX+aXSiflX
= urn
6

(I - 2a) cos x + ax sin 1-2a

=hm =glven
6 6

5
1-2a=6, i.e.,

• 5 3
From (2),b=l

.53
Thus, a=—. b=— —,

sin2x+asinx 0
(ii) [Form
•3

u• 2cos2x+acosx
=rn
-.O 3x
0
For this limit to be finite, the form should be
0
i.e., 2+a=6, Or, a=-2
sin2x+asinx . 2cos2x-2cosx 0
lim = km [ Form -1
x --° 3x 2 0
—4sin2x+2sinx
=lim [ Form 0
-]
—0 6x 0

urn -8cos2x+2cosx
=
6

Hence, a =-2 ai. the value of the limit is —1.

 INDII 'ti!NATE FORMS 369

EXAMPLES-)a

Evaluate the following limits I Ex. 1 -9]

- in x Cos tan x - X
1. (i) Li Li P. 20071
-o x3 .- Ox - Stflx

2.c
e - i x -a
(iii) Li - (iv) Ii
_n log (i + x ) x - a

x3 -2x 2 +2x-4 e +e' - 2 Cos x

[VP 1996]
(v) Li 2
X2 - ( vi) Li XSIfl X
-5x+6

a' - e'+ 'Sin x-1

(vii) 11 (Viii)
_..-+(i -.o log(i+x)

-'. + sin
.5

(ix) Li
I - Sifl' X
(x) Li _e (C. P. 2004
---.o -iO X-SlflX
sjn3 x

tan iix - a tan x

(xi)Li (xii) Li -
. • --O a sin x - sin nx .r-2 3x -

2 sin x - sin 2x Sin

sin Jog (i + x )
(xiii) Li (xiv) Li
•-.o tan x .
'. .-, 0

(sin2x+2 5 in 2 x-2 sin x l Cos x

(xv) Li I .,
cos x - cos x cos x sin 2 x
)

(- ir - x)iog sin x cos x

(xvi) Li - ______ + I
e°' _i+ log +x+

log x2
2. (i) Li
tan Sx
(ii) Li
tan x -*0 log Cot x

(iii) Li - ( n being positive ).

ex
24
370 I)/FF1RENTlA L CA LCULUS

tan3,rx log v – cot

(iv) Li (v) Li
seclZX -.O+ cot 7

(vi) Li 109 1 tan2 2x. (Vii) JJ

1 - 5x2

3. (i) Li x 2 log X 2 . ( ii) Li cosec (r- ) log x.

(iii) Li xlog sin X. (iv) L i sec x(xsin -v -

X.
(v) Li sin log x 2(vi) Li sec 5x cos 7i.

(vii) Li x" (log x ni and n being positive

L t (sec x - tan x). (ii) i

4. (i) (x cot

(iii) Li ( . I CP. 1996, B. P. 19981

- sin x )

(iv)
.'
Li I _L_ - __
ix – I ]ogx)
L (v) Li
--2 x -4
-
x2

I 1
NO L i- - — i log (i + x)
X

(v ii)- –9). (v i) 2x - x).

L,
(i) - (ii) Li x F CR 19961

flu) Li (sin x )21 - ( iv) (cos x

[CR 1989, 19971

(v) i (sin x )' (vi) Li x

C. P. 20051

(vii) Li(C(),'-
Sifl X
(viii) I1_,/x1I 2
1-0
) -' - ii
INDETERM/NA TE FORMS 371

2 )I / log - ) /(I -iog)

( x) Li ( log x
(ix) Li (i - i-It
-.1-

(xi)
(I.' + - ).
x
(xii)
X-O ( X

I C.P 1985, B.P 1996, VP 1995 I

Li (sin x]
(xiii) LJ (xiv)
o

[C.P 1990. 1998, VP 19981

2
a0 x + a 1 x" + + . + a,,
.
box" b1x" - + b2x"' -2 + .. + b,,,

(a 0 ;4 0, b0 + 0 according as n>=or<m (n and ni being

positive integers 1.

7. Li (ax` + a1.v " + + . . . + a,,, , ni being a positive

integer (a 0 + 0).

8. Li 2' sin '?_(a U) [CP1946]

+ Cos x - + - )
9. (i) Li (ii) Li

asin x— sin 2x .
10. If Li - is finite, find the value of a. and Inc limit.
I-*0 tan

IC.P 1997]
11. Adjust the constants a and b in order that

0(1+acos8)—bsinO_1
Li
0-10

12. Determine the values of a, b.c so that

ae' — bcosx+ce
(i) —*2, as x—i0.
x Sin X
372 DiIFEkIN1/A I. ('A L(JI,/.S

(a+b cos x)x- csinx

(u) ----------------------.-. --+1, as.

a sin x - bx + cx 2 + x3
may tend toa finite limit as
2x 2 log( l + x)- 2x 3 +
x -- 0, and determine this limit.

Evaluate the following I Et. 13— /91

xe -Jog(l + x
C. P. 20071
I

(ii) LOi{ !_ log (1+x)}. IC..1 1989, 1991

x - log (1-i-x)
(iii) Li x Cos

tan x tan' X 12
(iv) Li

L. - + 2 sin x - 4x
14 X5

Li Cos x- log (l+x)+ sin x--1

15.
e'-(l+x)

r 2x
16. Li [ 112+c052x - sin x -
[Vt
x sin 2. - +.i COS x J 1,, 2 sin 2x )

17. Li [(a 2 - 2) . cot J ir

.-.K-() 2 V (I + .1
L

18.D
+ 2)-(3x ••2)

19. Li
--*O if 2
-a
2
 INI)ETERMINA TE FORMS 373

Show that [Ex. 20-26]

20. Jja X sin __=O orb accordingaso<a<l,Ora>l.

21. Li { x _ x 2 log ( 1+!)}=.

log (I .* x + . )+ iog(i - x +
22 Li
-.0 sec x - cOsx

e --1 (3 sin x- sin 3xY

23. Li
-O x sin x cos x - cos 3x

Jog., (Cos x)
24 Li -4
.-.o Iog., (cos --x )

25. Lj =-e.

+ a2 1 +...+ a,'
26. Li -
n
)
asinbx-bs-jna
27. Evaluate Ii —.
tanbx - tanax

28. If 0 (x )= x 2 sin (l/x ) and j, (x )r Ian x, show that,

,
although Li 0'(X) does not exist, Li ---- exists and = 0.
-.otj, (x)

ANSWERS

1. (I) f (ii)2. (iii) 2. (iv) na"- (v) -.

(vi) 2. (vii) log (-/b ). (viii) 2. (ix) - - . (x) 1. (xi) 2.

(xii) 56. ( xiii) 1. (xiv) 1. (xv) 16+. (xvi)

2. (i) . (ii) -1. (iii) 0. (iv) -(v) --2 (vi)!. (vii) --

374 DIFFERENTIAL CALCULUS

3. (i) 0. (ii) ( - I / r ). (iii) 0. (iv) -1 (v) 0 (vi) - . (vii) 0

4. (i) 0. (1.1)0. (iii)-. (iv) -. (v) -.(vi). (vii) 0. (viii) I.

5. (i) I. (ii) 1. (iii) I. (iv) e (v) 1

(vi) I / e (vii) 1. (viii) 1. (ix) e.

(x) I/e. (xi) I. (xii) I (xiii) e (xiv)

6. +oor- (corresponding to a(,/b0 being positive or negative)

a0 lb0 according as 11> =,< m

7. 1. 8. a. 9. (i) L (ii)-. 10. a=2; limit =I.

11. a=-, b= — .

12. (i) a1, b2,cI. (ii)a=I20, b=60,cl8O.

(iii) a=6,b=6,c=0; limit =j.

(iii) I1. ( iv) . 14.j. 15.0.
13 (i) .. (ii).

16. - . 17. 4a/Ir 18.-8. 19. i/J.

27. b' h cos br - sin bx) cos 2 bx
 PARTIAL DIFFERENTIATION
Ing (Functions of two or more variables)

12.1. Definition.
If three variables u, x, y are so related that for every pair of values of x
and within the defined domain, say, a x !^ band c < y 15 d, a has i
single definite value, a is said to be a function of the two independe?..
variables x and y, and this is denoted by u = f (x, ),
More generally (i.e., without restricting to single-valued functions
only), if the three variables u, x, V are so related that U is determined when
x and y are known, u is said to be a function of the two independent
variables x and y.

Illustration Since the area of a triangle is determined when its base and
altitude are given, the area of a triangle is a function of its base and altitude.
Similarly, the volume of a gas is a function of its pressure and temperature.
In a similar way, a function of three or more independent variables can be
defined.
Thus, the volume of a parallelojlped is a fuction of three variables, its
length, breadth and height.
Note 1. If to each pair of volues of x and y, u has a single defipite value,
U IS called a single- valuedfunction (to which the definition refers and with
which we are mainly concerned in all mathematical investigations), and if to
each set of values of x and y, a has more than one definite value, u is
called a tnulfiple-valued function. A multiple-valued function with proper
limitations imposed on its value can, in general, be treated as defining two
or more single-valued functions.

Note 2. Geometrical representation of z = f (x. y ).

When a single-valued function z = I (, y ) is given, for each pair of
values of x and y, there corresponds a point Q in the plane OX V and if a
perpendicular is then erected of length equal to the value of z obtained
from the given relation, the points like P describe what is called a surface in
three-dimensional space. Thus to a functional relation between three-variables
x,y, z, therefore, corresponds a surface referred to axes OX , 5i, 57 Z in
space.
Note 3. Continuity -
The function f (x, r ) is said to be continuous at the point (a, b ), if
376 I)/FFERENTI,)L CA LCulI -'S

corresponding to a pie-assigned positive number C, however small, there

exists a positive number 8 s uch that

whenever 05jx—aI:5 5 and 0!^Iv-bI!^8

12.2. Partial Derivatives.
The result of differentiating u = f ( x , y ), with respect tox. treating
y as a Constant, is cal led the partial derivative of it with respect lox, and
all af -
is denoted by one of the symbols - , , A ( x, y ) UI bnefly,f 1. a,,
ax ax
etc.

Analytically, = Lu
av ' --.('
when this limit exists.
The partial derivative of a = f ( x, y ) with respect to ' is similarly
defined and is denoted by , A (x, y) F or briefly,f I, u , , etc.
aY ax
Thus, = Li
)' 8v-O A)'
provided this limit exists.
If it f ( x, v, z ), then the partial derivative of u with respect to x
is the derivative of it respect to x, when both y and z are regarded as
constants.

Thus, = Li
ax —.o Ax

Similarly for L
ax az

Illustrations:

Let a=x2 +xy+y 2 ; then

- 2x+y; -- = 2y+x.

Let 'u=yz+zx+Ay; then

all
au
----=v+z; ---=z+x; =5+)'
ax - ax az
PA/1iAL DIFFERENTIATION 377

Note. The curl a is generally used to denote the symbol of partial derivative,
in order to distinguish it from the symbol d of ordinary derivative.

12.3. Successive Partial Derivatives.

i)IIau
Since each of the partial derivatives -- is, in general, a function
dx ay
of x and y, each may possess partial derivatives with respect to these two
independent variables, and these are called the second order partial
derivatives of a. The usual notailions for these second order partial
derivatives are
a(au d2u
TX YX
i.e., —s- OIL,. etc.

a i(aut
\ du
i.e., 2 or etc.

a
TY i.e., , or f,

a (au'
i.e., or f" . etc.

Although for most ofthefunction.c that occur in applications we have

d 2u -
axdyayax
i.e.. the partial derivative has the same value whether we differentiate
partially first with respect to x and then with respect to y or the reverse, it
must not be supposed that the above relation holds good for ailfunetions;
because the equality implies that the two limiting operations involved therein
should be commutative, which may not be lnie always. Ex. 3, Art. 12.4 will
a2
elucidate the point. We can prove, in particular, that if the funtions —
'th

and both exist for a particular set of values of x, y, and one of them is
dxdy
Continuous there, the equality will hold good.

Proof of the equality ----

axay ayax

If f,, 1,, f,, J., all exist, and f,, (orf, ) is continuous tl,e,i
 78
D1LFERI:NTIAL CALU 1/ tic

Proo: It Ø(x )=f (.r,y +k )_f (x ,,)

(I)
Now applying Mean Value Theorem to 0 (x ), we get
0<0<!,
=h{ f (x +oh,y +k )_f (X +ohY )}
h{ F(\+k)_F(v)} ,say

[where F(y )=f jx +Oh, v)1

=h(k F(y +O'k ) 1,
0<0 ' <l, -
by Mean Value Theorem
= /lk if, (x+oh, y+0'k) 1 (2)
Again from (l), O(x + h )=f (x + h, y + k ) f(x +/z, y)

f (x , y +k )— f (x ,y ) (3)

Now, f . ( 2,V)
+ k)— f (x'
k,O k )

and f. ( x ,
y ) = it
f, (x (x ,y)
+ h,y )_Iv
- h
L f (x +hy +k )_f (x +h,y )_j(X ,+k )+f ()
h-.Ok-.O hk
(x + h )—Ø(x.)
= Li Li [from (3)]
h-.Ok--.0 uk
= Li Li fvv(x+Oh,y+O'k) [from (2)]
! - 0 k -. 0

= f( x, y ) , since f is cofltifluou,.

Illustration: if f(x,y)= xy-— , when x*0,ory#O,

x +y
when x=0, y=O,

a2
show that at the point (0, 0), __f.,._L., [C. P 2004, 2007]
axay aox
i.e., f , (o,o (o,o).
PART/A!. DIFFERENT/AT/ON 379

When x ;e 0, or y ;e 0.

(x2 + 2 )2x - ( x2 - 2 )2x

2-2
f (x ,y )=y
2
-
I x +Y (1
2
+Y
2

2 2 22
•0 (I)
22+2}
I

- 4xy
Similarly, f . (x, y) = x ... (2)
x+ (x2+y2T
f (h,0)— f (0,0)0
h

Similarly, f(0,0)=0.
From (1) and (2), we see that
j1(0,y)=— y(y*0),f y(x,0)x(x*0).

f , (h, 0)— f , (0, 0)h

Again, f'(0. 0) = Li = Li - = I,
h-40 h h—,Oh

f (o,k)— f (0,0)
- k—+O k k—O k

We have similar definitions and notations for partial derivatives of

order higher than two.
If, z = f ( x, y ),the partial derivatives of z are very often denoted by
the following notations:
2 2 2
az az a-z dz

Illustrations:

au = 3x 2 + 2xy 2- = 6x 12 y 2 --- 4xy.

ax a. axa,

These notation was first introduced by Monge.'

 380 D/Ff-Lp/J.'j/nJ, CA LCULUS

2
+2x 1,; - r 6y+2x 2 ; - =4xv.
OY
12.4. Illustrative Examples.
Ex. 1. If = r cos 0, r = r sin O

So that r = J, 2^ y 2 0 lan (v/k ).

show that

3 Ith r) /30
— ^ It---- and
3r 13x 30 ldx
05 3,- r cos U
Here, --= cos i3 _
X 2) r
-+

30 V r sin 0 sine
.30' 3x x - r2
Hence, the required results follow,
Note. If y is a function of a single vai-wb/e x, then we have seen that.

under certain circumstances (see § 7.7), A .cirnilarproperiv

(1.1 . dv
not true, as seen above, when v is a function of fliore Man one variable.

Ex. 2. If u = f s/iowtlza,x- + du = o.
IxJ 3x 3y
U = .1(z), say, where z = y/x
an all az , z y , /
=—.—= .f z
a
—
a ax
Z t — =—
ax -
.1

au .,/\ 3 I'
Similarly, - = f iz j =

du
ax + .v
a)
= VY V

if '(z )+ f'(z )= 0.
Ex.3. Show that

X )' X V
Li Li —* Lu fj
S - I) y *O x + y ,- -O x *O A + y

Left side = Li = Li I = I.
x-+0 A
PARTIAL
TIAL DIFFERENTIA lION is I

Right side = Li
y -+o y y-.o
Hence, the result.

EXAMPLE- XII (A)

I. Find t. 1 , for the following functions f(x. ),):

(i) • ax 2 + 21xy + by (ii) tan ( ) , Ix ).

(iii) + (iv) log( x 2 +y 2 ).

(v) x2/a2 + y2//,2 I.

2. Find f , f , f , f for the following functions f(x. y):

(i) x 3 + 3X 2 + 3x' 2 + y 3. (ii)

(iii) xcos y + )' COS x. (iv) Iog(x2y+xy2).

3. (I) if v=x 2 +v 2 +z 2 , show that xV,. +yV, +zV =2V.

(ii) If u = x 2 v + v 2 z + z 2 x, show that
14+a,+u=(x +v +z )2.
(iii) If u = f (xv:), show that

= ZU:

• V Z X ii u au u
(iv)1fu=-+--+—,provethatx—+z
-t-y ----- --- =O.
a

• X Y au 3,,
4. (0 IT u s in' - + tan-' -, show that x - + y -- = 0.
a a
VP 1995, '971

x(l-y2)+y(I-x2)
(ii) If U = and
1- xv (1+x2)(l+y2)
prove that U,. V = U V

5. (a) Show that ----- +—' = 0, if

3)2
3x2
log ( X
(i) u= 2 +v 2 ). IC.? 19901
(ii) u = tan (y Ix). C.? 1998!

(ii) a = i' ' ( xcosy — ysin )

382 DIV rERENTIAL CALCULI/S

a 2 v a2v a2 V,
(b) If V = z tan -, then + + = 0.
x ax2 V az2
[B.? 1998]
X 2 2 2:
2
)'

6. (I) If f (x, y, z ) = x y

I I

show that ,f -f .1, + = 0.

I 1 I
x 1' 2: W
(ii) If U
= X2
V
2 2 7
X33 3 3 3
' 2:

show that u + u + u + u, 0

7. If V ax 2 + 2hxv + 1n 2 ,then show that

V, ' V — 21/. V V.VV + V. 2 V, = 8 (h 12)V

8. If ulog(x 3 +)''+ --3x),z),(hen show that

(i) + + = [13.11 1989, '91, '971

ax dv di. x ± y - 4-
7
a2 a7 a
z
11
3
(' I ) —i- +--- + )2
ax = ---,;i- +

[ V P. '97, C. P - 46, '85, 2007, BY 20011

9. how that
If V=(i+vFJ). S V VV +V vV +V . =2/V.

10. If V = Ii. 4. y 2 4 that

show z 2 ). V+V.+V,0.

CF. 1976]
H. If u = e l , prove that

d3u
= ( I+3 Z +x 2 y 2 z 2 ) e . C. P.
ax ay a
(X 2 +),2),
12. (I) If V (ax +bv)2 _ where a 2 + b2 = 2,
then show that V + V, 0.
PARTIAL DIFFERENTIATION lxi

(ii) If u =3(ax-4-by+cz)–(x 2 +y2 +z2) and

2 C 22 a2 a2 u a-u
a +b ± I, find the value of - + --•- +
(2 2
Eb'2aZ
(C. p 19341

(iii) If u = ux' +b 2 +c: 2 +2j5'z±2gzx+2hj' and = 0.

ax
show that a ^ b + c = 0.
13. Show• that, if u(, ly,: ) satisfies the equation

a2 ,. a•' u a-u
ax—+----=0, then
—+
2 ' -2
av - C:

• m, aU all
-. - , - satisfy it, and also
ax a az

au au on
(II) x — + -v— + 1 — satisfies it.
' Oz

14. If ii = !og(x 2,2 + z2), prove that

a2l, 0-i, Thu

15. If it = log i- and 1.2 = + + 2, prove that

,21 0 u + a 2 11 + 0 2 11 =1. [C.P1975,2008}

OX 01' i3z2

16. If y=f(x+ct)+(x–ct), show that

a ly
I VP 2001]

v= 0( ax +2/n)'+by) show

17. that :D(v (CH.1934j

18. lf(i) U=x-i-y4:, I'rZY2-4-I2-4-.

IV = x 1 + )' + z - 3xv:,
(ii) U = x + i' + : , = + 1,2+ ,2 W= i'z + :x -4-
384 DIFFERENTIAL CALCULUS

U U y Ur
show that in each case V t'. V = 0
W W W-

19. If a, i, y be the roots of the cubic x 3 + px 2 + qx + r = 0 ,show that

OP &I Or
Oa Oa Oa

Op Oq Or
O5 (?fl 0/3

Op allCr
Oy ay Cy

vanishes when any two of the three roots are equal.

20. Show that

f1(x) f2 (X) j(x)

t' (x) f (x) /" (X )

01 () øU•) Ø(r) = 0() o;() o()

,j() ,,(:) ii(z) ii;(:) l//(:) yi,(:)
where dashes denote differentiations with respect to the variables
concerned.
ANSWERS

1. (I) 2 (ax +hy) : 2(hxi-bv). (ii) -

12
(iv) , --s---;-.
(2 + 2 ) + - + v
(2 + l )

2x 2v
(v)
a b-
2. (i) 6(x+v). 6(x+1). 6(x+3 , ). 6(x+v).

(ii)

2x+)(x+2)+t}. z{(x+2v) 2 +2,where :=e'2

— lcosx. —( Sill .v+ Sill ,). _( sin x+ sill 1). — x Cosy .

I x+ , )
I I
(iv) - +2
-2 -. I *
+ 1 .(. + Y) T(.v + r)
12. (ii) 0.
PARTIAL DIFFERENTIATION 385

12.5. Homogeneous Functions.

A function f (x, y ) is s to ke homogeneous of degree n in the

variables x and y, if it can be = ,ressed in the form x" or in the

form y" 0 (

If V be a homogene'" . function of degree n in x, y, z, then each of

av av av
- . . -, 'a s a homogeneous
i function of degree (n - 1).
ax ayz
, 2.
Since ax +2hxy+by 2 x 2 a +2hX +bI - - 01 '
X I,X) V
+ 2h + b y2 is a homogeneous function of degree 2 in x, y.

Similarly, y/x, x tan( y/x), x 2 log( v/x) are homogeneous function

of degree 0, 1 and 2 respectively.
Note 1 An alternative test for a function f ( x, y) to be homogeneous
of degree n is that f (tx,ty) :" j* (x,y) for all values of t, where i is
independent of x and y.
Note 2. The test that a rational integral algebraic function of x and y
should be homogeneous of degree n is that the sum of the indices of x and
y in every term must be n.
Note 3. Similarly, a function f ( x, y, z) is said to be homogeneous of
degree ,z in the variables x, y, z, if it can be put in the form !J( Y , -
)1-1
if f (:. I-y,Iz) = t"f ( x, y, z ); and soon, for any number of variables.
xx

Thus, f ( x, y, z) = IF, + is a homogeneous function of

degree . , since
tx + rty + Ftz = : f(xy,z).
f(tzly,rz) = ,JF

12.6. Euler's Theorem on Homogeneous Functions.

If f (x, y ) be a homogeneous function of x and y of degree n then
a a
x - + y - = nf (x, y )
T ay
25
 386 DIFFERENTIA L CA LCULUS

Since f (x, y ) is homogeneous function of degree n,

let f (x ,y )=x Ø(y /x )

= x'Ø(v), where v= y/x.

= nx" Ø (V )+ x°
ax ax

=fØ'(v)
ay a)

nfO(v)= nf(x,y).
ax ay

12.7. Differentiation of Implicit Functions.

Let the equation f (x, y) = 0 ... (1)

define y as a differentiable function of x, and let I and be continuous.
f

Then, we can find in terms ofas follows:

dx a' ay
Wehave f (x+&,y+itv)=o.

(2
Now, by the Mean Value Theorem, [see § 9.2]
f(x+ A x, y + A Y )- J(x,y+ A y)

=x — f (x +oj A ,v +) [ 0<0, < 11.

f(x,y+y)—f(x,v)

a
\y— f ( X I y+O2 Lsy ) [0<0,< 11.

Adding these two and using relation (2) and dividing by Ax,weget

)=o... (3)
ax &ay
Since y is a differentiable function of x, when ii - 0, iv - 0, and
since J and f are continuous, we get, by making Av - 0 in (3),
PARTIAL DIFFERENTIATION 387

àrd.ry
a
(r7 * o) •.. (4)
dx !! f,
a,

12.8. Total Differential Coefficient.

Let u = f(x,y), where x=sp(t), y =1(f ).
Then usually u is a function oft in this case.
dx
To obtain the value of
Let us suppose that f , f as also Ø'(i '(z ) are continuous.
When i changes to t + A t ,let x and change to x + A x, y 4 Ay.
Now, u = I 10 (t I i v (i
)1a F (i ), say.
duL i F(f +A f )-F(i)
di -o At

. t (i+A I ))-f { ( : ,()}

= L i f { (t+
Al

f (x i&,y +A y )-f (x ,y )
= Li

I
A:
f(x+x,y+A y)-f(x,y+A y)A x
= Li
t -.o t Ax Al.

f (x ,y +A y )-f (x ,y )A y
AY Al
But, by the Mean Value Theorem. [see § 9.2]
f(x+ Ax, y+Ay.)_f(x,y+Ay)=Av.f, (x+O AX, y +y)
and f (x,y+A y)-f (x,y) A y.f (x,y+O'A y).
where 0 and 0' each lies between 0 and 1.

When Ai-*O,Ax->O and Ay-90; also _90'(I) and

At
Ay
--v i(i).
At
 388 DIFFERENTIA L CA LCULUS

Also f (x +O&,v +1y )_4f (,y ).

f (x,y+e'y)— , f (x,y).

du cL dy
-= f — +1 ' — ... ( 1)
di di di
du au dx an dy
i.e., ... (2)
Note 1. Asa parttcular case. if ii =f( x. y ). where y is a function of x.

dt ax dy dt
du
is called the total differential coejiicienl of a, to distinguish it
dv
from its p artial differntiaj coefficient.
Note 2. The above result can easily be extended to the case when u is a
function of three or more variables.
Thus, If it .f(x, y, z), where x, y, z are all function of 1.
du au dx audy an dz
----f — — +–-.--
di ax di ay di az di

12.9. Differentials.
We have already defined the differential of a function of a single
independent variable; we now give the corresponding definition of the
differential of a function of two independent variables x, Y . Thus, if
u = f (x, v ), we define du by the relation

du = h A x + f,. 4y
Putting u = x and u = y in tum, we obtain
dx = Ax, dv = 4w ... (1)
so that du = f, dx i- f dy ... (2)
Multiplying both sides of the relation (I) or Art. 12.8 by di, and noting
that, since x, y are each function oft,
du dx
du= — dt, dx= .—dt, dv= —
dy dz,
di di dr
we get du = f . dx + f. dv
PARTIAL DIFFERENTIATION 389

which is same in form as (2) above. But x, y here are not independent, but
each is a function oft.
Hence, the formula du = f, dx + 1, dy is true whether the variables
x and y are independent or nor.
This remark is of great importance in applications
Sinrilarly,ifu = f(x,y.z),
du = f dx 'i' f dy 1' f dz,
whether x, y, z are independent or not.

Note. It should be noted that the relations (I) above are true only when x
and y are independent variables, if x and y are not independent but
functions of a third independent variable r, say, x = 0(1) y = 41(1 ), then
dx= Ø'(: )dt and dy = 41(1 )dt , where dt = 1st

12.10. Exact (or Perfect) Differenilal.

The expression

0(x,y)dx+41(x,y)dy ... (I)

is called an exact (or perfect) differential if a function u of x. y exists such
that its differential

du, i.e., }.dx+dY ... (2)

is equal to (1) for all values of dx and dy.

Hence, comparing (I) and (2), we see that if (1) be an exact differential
it is necessary that

auau=
O(x,y) and —=41(x,y).
TX ay
Differentiating these relations with respect toy and x respectively, we have
a 2u -
and au -
ayaxay axayax

a2u d2u
Since, in all ordinary cases, = ,hence, in all ordinary
ayax axay -
cases, in order that (1) may be an exact differential it is necessary that

- a41
aya.x
It can be easily shown that this condition is, in general, also sufficient.
390 DIFFERENTIA L CA LCULUS -

12.11. Partial Derivatives ora Function of two Functions.

Ifu=f(x1,x2)
where x 1 =Ø 1 (x, y ), x2 = 2 (x, y ), and x,y are independent variables,
then

du-
an ax1
. +
au. dx2

du au ax 1 an ax2
ay X1 ay ax2 ay
We have
all au
du= 2 (I)
ax,

a., ax aX2 ax
di1 2
ax a
When values of x, x2 in terms of x, y of are substituted u becomes
a function of x, y; hence
all
— cLt+--dy
dx a

Now, substituting the values of dx 1 , dx 2 in (1), we get

(du du dx 2(du ax, du dx2
ax,
du=i --+ ------ idx +1 +— Jdy
dx dx, dx) dx, dy dv, dy)
Comparing this with (2), since di, (lx are independent, the required
relations follow.
Note. The above result admits of easy generalization to the cases of more
than twovariables.Thus,lf u = f (x 1 , x 2 , x3 ),where x 1 (x.y.z),
X2 , ( x, , z ), x 1 = Ø (x, y, z ) and x, y,z are independent
variables, then
du an dx 1 du dx2 an dx3
---+— — —+—, —
Tx dx 1 Tx- dx 2 Tx- a x' ax
du a n ax1 + an dx2 an dx3
-= -------------. +
ay ax 1 dy dx 2 dV dx3 ay
!!= h+ -!!i+ an ax3
dz ax, ax dx2 ax dx3 dz
PARTIAL DIFFERENTIATION 391

12.12. Euler's Theorem on Homogeneous Functions (generalisation).

iff(x, y, z) be a homogeneous function in x, y, z of degree n, having
continuous partial derivatives,
aj aj a
then x—+y--+z—=nf.
ax ay a

Proof: Since f (x, y, z ) is a homogeneous function of degree n.

j (ix, ty, fz ) = t"f ( X, y, z) ... (I)
for all values of t.
Putting tu = u, ty = u, tz = w,
differentiating both sides of (I) with respect to t, we have

au a, av a at
Y,

a a a.i
• x—+y—+z—=nt f(x,y,z)
au av dw
Putting S =I in (2),
a a a f
- + y - + z = nf
x
ax 3y 3z
Note 1. The above method of proof is applicable to a function of any
number of independent variables.
Note 2. The above result can also be established as in the case of two
independent variables, i.e., by writing

x) x
f df.
and then obtaining .-L a
axay a
12.13. Converse of Euler's Theorem.
1ff (x, y, z) admits of continuous partial derivatives and satisfies the
relation
a af a
x—+y— +z—=nf(x, y, z)
ax
where n is a positive integer, prove that f (x, y, z ) is a homogenous
function of degree n. I C. H. 19601

Proof: Put =-, =

z Z
 392 . DIFFERENT/A L cA LCULUS

then x = , y = 77 z = . Suppose, when expressed in terms of

f(x, y, z)=v(, . C)
Then
ax d dx di dx d dx

=x
(dv I
-- — +
dv
.0 +
dv
.0

jrdV
,

a a
Similarly, v — = —

aj (dv ddv dt7 dv

Zd
dz all dz ac az

a 2 2

dv dv dv
= -- —7) +

Hence the given relation reduces to

dv ldv n
—=nv, •oi; --=-,
vd
whence log v =, n log + a constant
where the constant is independent of , but may depend on and 11 let
this constant by denoted by log 0 (, q).
Then

i.e.,
X, Y)
which, according to the definition of a homogeneous function, shows that
f ( xi y, z - ) is a homogeneous function of degree n.

Note, if a be any rational nuniber, the proof and the result remain
unchanged.
PARTIAL DIFFERENTIATION 393

12.14. uIustrafiveExampI.

Ex. 1. if u =tan - ' show that

au au
+ y--= sin 2u. [C.P 1996, '98, B.P. '95. V P '99,2002]
dx dj,
From the given relation, we get
= +
x3l+(y/x)3}X21y
tanu
x-y x I .-(y/x)} ix
tan a is a homogeneous function of degree 2.
Let i' = tan a; .. by Euler's Theorem,
dv dv
= 2v
dx dy
• a 2 du
xsec2u—+ysec u—=2 tan u
dx dy
du au = . -.
x—+y-------=2sinucosu =sin2u.
dx dy secu
Ex. 7. If A be the area of a triangle ABC, show that
d4 =.R (cos Ada+ cos B db+ cos C dc)
where R is the circum-radius of the triangle.
From trigonometry, we have
b2c2 +2 +2a 2 b 2
+2a2 - b4 -c4 ).

Thus, A is a function of the three independent variables a, b, C,,

Hence, taking differentials of both sides,
2dS = {4Q (b 2 - c 2 -a2 )da+*(c2 +a 2 -b2 )db+4c(a2 +b 2
_C2)dC}

= .1 (4a . 2bc cosA da+4b 2ca cosB db+ 'lc• 2ab cosC dc)
+abc(cosA da+ cos B db+ cos C dc)

...dA=.k( cos Ada+ cos Bdb+ cos CdC)

44
= (cos A da+cosB db+ Cos C dc)
R

Ex.3. If P dx+ Q dy + R ds can be made aperfect differential of some

function of x, y, z on multiplication by a factor, prove that

a P + R (^—
p - ^—Q = 0
dz ay ) +
P (^—Q - Q(' ,
dx dz)
-
dy ax)
(C. H. 1949. 1954 1
 394 DIFFERENTIA L CALCULuS

Suppose u is function of x, z and

p(P dr+Q dyt -R dz)= du, ... (1)
where ji is some function of x, y, Z.

Also, du =±! dr+.!i dy+- dz. (2)

dx dy dz
since u is a function of x, y, z.
Comparing (1) and (2),

dx
AP (3)
du
—JLQ ... (4)
ay

(5)

32 E,
ap ag a a,
6)
dydx ay dv TI:dx
(on differentiating (3) with respect to ')

(on differentiating (4) with respect to x, assuming _f_ = a 2 , ) .

d'dx dxdy

Similarly,

32 u 3Q dp 'dRdu
(7)

32 u 3R di aPdp
dxdz (8)
ax ax dz dz
From (6),(7),(8), we get on re-arranging

JU '^—
Y -^7) = a a (9)
(

ju(^—Q - 'R )= R ap - Q " (10)

(3R P) 3u
(II)
Multiplying (9) by R. (10) by P, (11) by Q and adding together, we ge t
the required result.
PARTIAL DIFFERENTIATION 395

Note. If P a + Q dv + R dz be itselfaperfect differential, then we easily

deduce the conditions that
dQ dR _dR aPaP
dz ayax dzdy a
Ex. 4. If V be afuncf ion of x and y, prove that

av a 2 v_a2 v IdV . I d2v

+2 + r d r + r2 502
TT dr2
where x = r cos 6. y = r sinO. I CH. 1953 1

r = .,J52 +y 2 , 6 = tan' 2..

dx dr x rcosO
Hence, - = cos 0; - _________ = cos 0.
dx Jx2+y2 r

dy ar y r sin O
sin 0 -=----= ,in O.
Tr Ty J+y2
__ 2 y rnO s in O
r
dx -
j±1.
x2 )
+
x2+
y2 x2 - r2 -

Similarly,.! coso .
dv
Since V is a function of (x, y), and x and y are functions of r, 0, so
V is a function of (r 0).

a dV dr dv do
Hence, — =— .----+— .— .
dx dr dx dO d'x
sinO dV •
coso. av ... (I)

av dv drdv 'dO
dy dr dy d0
dV cosO all
=sinO.-1------ (2)
r dO
Thus we have the following equivalence of Cartesian and polar
operators:

d ( d -in Od
—
dx dr .r dO
 396 DIFFERENT/AL CALCULUS

a a
.dy drrd0);

a av av sine av'
cosO--
dr rdo dr rdO)

a2v
c cosO- sm8 a 2 v dv Sin
---
dr2 ,-

sin O a2v a2v a a I

_
coso--SJno__--coso__slno-----
r dr r dO r

=cos20 --
a2 2 sin O cos6 a 2 v sin 2 U d 2 v sin 2 oa
dx2 5 r 2r drdO 2 do 2r dr

+ 2sinOcosO dv
r2 do
Similarly,

a2v 2sinO cosO d 2 v cos 2 O d 2 V cos 2 O dv

- = sin2 O+
ay 2dr2 r drdO ,2 as dr
2r
2sin6cosO dv
r2 dø
d 2 v .J 2 V d 2v I d2v i dv

12.15. Miscellaneous Worked Out Examples

ö 2 u öZu
Ex. 1. (i) If u = find the value of II C. P 1986]

ö2u 82U-
(ii) If u = xY, prove thdt — = -. [ C. P 1986, 2001, 2008 1
öx6y 8v 8X

2, 11 ö 2j
(iii) If f(x,y)= x3y-fe-'Y Show that = [ C. P 1993]

I I
Solution; (i) Here, u=x2 'y2

iu I I 82u 1 ( l'\ — ! I—
2 y 2 and 2 , y 2 __, 2
PARTIAL DIFFERENTIATION 397

8u i ! –! 82u 1 •'- '--

2 and -=–x2y 2

- 82u - if 'i j' i(L2+y2 )

(ii) u .v

su
or, —=xY.Iog
8y

and, —=y.xY'.log+xY._
6x8y x
y . x Y ' . Iog+x Y( I)

8u
Again, — = y.XY'

82u
and — =1 . x Y ' + y . xY' 1og (2)

6 2u 82u
From (1) and (2),

(iii) Here, f(x, y)=xy+e'2

• L_x3+e.ry2(2y)

821 3x2
and
8xöy
=3x2 +2yeYZ (xy2+1) - (I)

Sf = 3x2y+eY2 (y2)
Again,

and _L=3 2 +eY2. (2xv)y 2 +erY2 .2y=3x 2 +2yeY2(xy2 +1)

&y&r
(2)

From (1) and (2)

62f

ix8y y6x
L
Ex. 2. (I) If u r 3 , x2 +y 2 + = r2, then prove that
6 2 , 82U 62u
[C. P ii
&v2 62
398 DIFFERENTIA L CA LCULUS

(ii) If u = x 2 tan 1_2 t an-t !• prove that

82u X 2 _ 2
C. P. 1998
.3x6y X 2 +y2

(iii) If u = .r log y . show that u . = y. C. P 1983, 96, 2003

Solution : (i) r2 = .2 +v 2 +

Sr &x
2r=2x,i.e.,

Sr V Sr z
similarly, - = and -- =
r r

Sit
.3,, , Sr
u=r3

52U St. 3x2

—=3r-l+x.3—=3r+-
r

62u 3y2 82u

similarly, 3r+ and —=3r-t--------
32
T r
.32u .324 S 23
-- =9r+ --(x 2 +y 2 .tz 2 )=9r+_ •r 2 =12r
6x 2
3 2 .3z 2r
.

(ii) u = x 2 tan - . 2 tan-'

y
I I x I
2(:y
x
i+)

X3
= —21ian--+------x--2ytafl'
x2 +y 2- y y

1 2y2 - 'c2—y2
—1-2 X) 2
SAY )x2+y2x2+y2Y
1+1 -

1
(iii) u=x log y .. it, = X 110=—

Again u =lIogv a.id ,,,

=! :. u, =u
 PARTIAL DIFFERENTIATION 399

Ex. 3. (i) tan - ' 1 + sin

Show that f(x, y) is a homogeneous
x y
function of x, y. Determine the degree of homogeneity. Hence, or

otherwise, find the value of [C. P 1991, B.P 1996]

&

(ii) Examine whether f(i, y ) = x3 . y 3 tan - is a homogeneous

function of (x, y). If so, find its degree. [ C_-JP 1993

(iii) Examine whether the function u (x,y)= x 3 + ' is a

x2+y2
homogeneous function of x and Y. B. P 1995]
Solution f(x, y)r tan1 2-+sjn- = tan-' X + cos ec-1

=x0{tan(.)+
cos ec_i()}=

Hence, f(x, y) is a homogeneous function of x, y of degree 0.

By Euler's theorem on homogeneous functions,
ff
x+y-=Oxf=O
ax 8Y here n=O].

(ii) f(x, y)x .x3.Lt

.an h'X)
X -. x)
4

=x .(!J.tan(rX)=x' .t(Y)

J ^ff
Hence, f(x, y) is a homogeneous function of x, y of degree I.

x3(1+2)
(iii) u(x, y)= =xx
X2 I+ l j x

So, u (x, y) is a homogeneous function of x, y of degree 1.

u
EL 4. (i) If uxsin1(2)+ytan.i) find the value of x--i- &4
y- ---
& .5y
at (1.1).
IC. P. 1988]
400 DIFFERENTIAL CALCULUS

6u u
(ii) If u = X.(+X) prove that x—+y----x
x) ox by x)
I C. P. 1988 j

x2 + Ou
(iii) If u -- (x,y)^(O, 0) and x—+y--=ku
+y-8u find the
Ox 0"
value of k C. P 1997, 2002, 2004 I

Solution : (I) u = sin_t()+ytan_J(

=x{sin J+2cot1()}=xi

So, U is a homogeneous function of x and y of degree 1.

-, , Ou Ou
By Euler s theorem, x—+y----=l'u
Ox O)
(y\
=xsinj - I+ y n -
\XJ

At (1, 1),

(ii) Let v = and w =

x ()
The- is a homogeneous function of x, y of degree I, and w is a
homogeneous function of x, y of degree 0.
8v &
So, x ---+y-- =I-V=V (I)
Ox by

Ow Ow (2)
and x—+v----=O'w=O

Ou Ot' Ow
U = v+W,-=-+---
Ox a a
Ou Ov Ow
or, x-=x--+X---
Ox Ox Ox

Ou Ox Ow
similarly. y— =y—+y—
0:1 v y

Ou ö ( o Ov ' ( Ow '
x+y—I X—+y— 11 x o— +Y — 1 = v+0 vX4)l -
Ox 0 Ox Ox) i Ox '5y) ' ' x
PARTIAL DIFFERENTIATION 401

2)

• 3 (
x2+y2 x21+() J
u= _____
__

So, u is a homogeneous function of x and y of degree Hence,

by Euler's Theorem,
öu su 3
(1)
8x 8y 2
& 8U
But given that. x— +y =k u . (2)
X 8?
—)
3
From (1) and (2), k = -.

Ex. 5. (i) If V = 2cos'I ( I, show that

8V 811
[C.P1994,2008J
8x 8y +cot •2

(ii) If u=x Y f (!). prove that x -+y -=2u

x 6x 8y
by Euler's theorem,. ( C. P 1984 1
6u 8u 8u v z x
(ui)Venfy that x —+y ---+z —=O, f u=—+-- +-- .
a
. P 1987, 2005]
8V 8V
(iv) If v=!ogX 33 ,show that x
X.2 +y2
+y 1. [c. P 1978]
5 öy
if = X2
(v) Find that value of, u Y [C p 1990]
8x 8y x--xy
X+)'
Solution: (i) V =2cos'

COS—
r 7= =____
26
 402 DIFFERENTIAL CALCULUS

(V'\
COS) is a homogeneous function of x, y of degree .

Hence by Euler's theorem,

61 v' 81 V'\ 1 V
x—I cos— I+y—I cos— 1—COS—
6x 2) 6yk. 2) 2 2
I I . V"6V I I V'\6V 1 V
or, xi — Slfl I-+yl ---Sin — I— —COS----
t. 2 2)& '. 2 2)öy 2 2
8V öv V
or, X+y - +WtO

()
Hence, it a homogeneous function of x, y of degree 2.
By Euler's theorem
8,4 8u
x - + yby
-=2a
6 x

XYZ

12 2
x3fl'--') +!
- x) x\.x) x

- - Sx' x
31--.
I.x xJ
Hence, it is a homogeneous function of x, y, z of degree 0; by
Euler's theorem.
6,4 6u 6u
öx . 8y. öz

x3 + V 3
(iv) V=Iog ,
x-. +y-

or, e
x3-s-y3 t y
PARTIAL DIFFERENTIATION 403

Thus e" is a homogeneous function of x, y of degree 1.

& ( v'
x— ( e V )+Y —
e
)= 1e
6x 6y
V . 8V , 8V .
or, x•e •— +y . e — =e

8v &v
x—+y—=1
6x sy

x2{J+(2)}
(v) Here, u = x°

I xJ

So, u is a homogeneous function of (x, y)of deg :e 0

8U 8U
By Euler's theorem, =O.u=0.
&x
x+y&y
Ex. 6. (I) if u be a homogeneous functio't of x and y of degree n
having continuous partial derivatives, then r ave that
/
I öu uj
Jx—+y-— I 11=n( –Pu.
t &&y )

2 82 82u
&' 2 
where, —+y---I =x 2 —+2 y _
& 8y) &2 &v ,2
1 • P 1985, '98. 2001; B. P 1993]
2
(ii) if u=---3--_, apply Ei..Ier's theorem to find the value of
x+y
8u 6u o2 5U
x—+y — and hence deduce that x2 —+2xv------f-y 2 --6u.
6J' - &8y 8Y2

[ C. P. 199., 2t) i

(iii) If u=x@(^)+ii(ZJ. ptove that (x+Y-)u0.

[C. P 1990 1
404 DIFFEREi9TJAL CALCULUS

(iv) if f(x, y)=(x2 + y2 ). use Euler's theorem to find the value of

22f 821 821 2
X +Y
f 8f
andhencepmvethatX +2.x y +y 2--+.f =O.
sy
C. P. 1995

(v) if U= tan +, prove that.

x–y
814öu
x—+y—sin2u .B. P. 1995 C. P 1984 89 96. 98]

2 62 l 62u 82a . 2u)sln 2u.

(b) x + + Y2 = (i –4sin
8x8Y 8Y
C. P. 1996, '98 2003, B. P '95 1
Solution (i) Since u is a homogeneous function of degree n,
we have by Euler's theorem,
Su 5u
y (1)
6x
$jfferentiating (1) w.r.t. x partially,
Su 8 2u 82u 8u
(2)
ox 2 (8y Ox
Differentiating (1) w.r.t. y partially,
• 62u Su 82u By
(3)

Multiplying (2) by x and (3) by y and adding

( 82u 82u O2 I 8u Ou
I x 2 — +2.xy-----+y2 -- I+X—^Y 5" J = nx—+y----
I Su —
OxOy 8y 2 Ox &x Oy
t Ox ) Oy

8 2u 62u
where we have taken - -, since the partial derivatives an
OxOy Oyo.x
continuous.
.2 8u2
or, X --+ 2 xv— + - V 2 --+flUflflU
Ox8

1 8 8\2
orIx— +yI un(n – l)u
i, Ox &y)
 PARTIAL DIFFERENTIATION 405

i
X 4l
x22
y = ____ y
(ii) u -
X+y x
x(1+fl
t. x)
So, u is a homogeneous function of x and y of degree 3.
By Euler's theorem, -
öu 8U
(1)
Differentiating (1) wat. x partially,
82u8u 24 8u -

32, 62u. Su 6u
or, x2--+.y----+x--=3x- (2)
ox OxOy Ox Ox
Differentiating (1) w.rj. y partially,
2j 8u 8u 8u

2 02U OU Ou
or, +)' + y- = ,•• (3)

[.. o2 O2'

Adding (2) and (3)

2u 2 02 u ( S u
X 22x y 8 y I
21 x — +-----
Ox2 Sy )

= 2 x 3u = 6u

(iii) U = X0
()+()
=v+w
where, v = , a homogeneous function of x and y of degree I,

and w = a homogeneous function of x and y of degree o.

 406 DIFFERENTIAL CALCULUS

By Euler's theorem,
8V öv
x— y— =1.V =V
& + öy
and
3w ow
=0.W=0
Ox Ty-

ou (Ov 3v '( 3w Ow
x—+y— = I x—+y— 1+1 x—+y—
a o
=v+o= v. (1)
Different)ating (1) w.r.I. x. partially and then multiplying both the sides
by x, we get
32u Ou OV
X2 -- + +x (2)
ax^y 8x = x—
2 6x

• ',32u 3u OV
Similarly, x y — +y --+y—=y— ... (3)
OxOy 3 Oy Oy
52 U 82U
8yOx&&y
Adding (2) and (3)
3 2t 32u 2 6 2 u ( Ou &
X2 -+2YK+Y -++Y

3v ov
=x — +y --
Ox Oy

32u 82 U 82u
or, X 22
__+2y__ __+y2T- = 0 [bv(l)I
Ox2
OX OY

i 2 2
= { t+ ()2 } _(Y)
(iv) A x, y)=(x2 +y 2)

Thus f(x, y) is a homogeneous function of x and y of degree

By Euler's theorem,

x 1 +y 1— f (1
ax Oy 3
PA RTIA L DIFFERENTIA TION 407

Differentiating wrf. x partially

82f 8f 62f 28!
X
8x8x6y36x

or, X
2 82! 2j 8f 2 8f
... ... (2)

. 8 2 f 2 6 2 f 8f 2 6f
Similarly, XY + Y+ =Y ... ... (3)
Adding (2) and (3),
1
or, x2 6 2 82! 2L( X -.-_+
6f 811
v-
8x 2 8x8y 8y2 6x 6v)

.
2( of 6f
=— I x— +y ----
3 6x Oy

2 21 +21821+Y262f + 2 f 2- . 2 f

i.e., X 282
I
1
— +2.xy
82
.L 2 6 2 ! 4
6x8y
+y
2

• 82 1 2.y----+y
x 2 —+ 2 62f 2
• 62j —y + -- f —O
8x 26x 8y oy 9
x 3 +y3
(v) u=tan-
x— y

3Ji+(Zl
or, tanu= - I .x)J =x2j!)
x— y -

So, tan u is a homogeneous function of x and y of degree 2.

By Euler's theorem,
8 6
x . -(tan u)+y- (tanu)= 2 tan u
8x 8y
', 6u
or, xsec2 u+ysec
6u
— u— =2tanu
6y
408 DIFFERENTIAL CALCULUS

SU SflU
x—+y-2----cos2 u=2sinucosu=SIfl2U
514 (1)
Sx Sy cosu
Differentiating (1) w.r.l. x., partially,
82u 81( 52u
x +1•— + y - = 2 cos 2u -
2 Sy' &xv

2u Su
2 8--+-----+x—=2xcos2u--- (2)
or,x
5x2 SxSy Sx
Differentiating (1) wi-f. y., partially and then multiplying by y
5 2 u 2 52 Su
x y -+y — +y -=2v cosu (3)

82. 62,
SySx SxSy
Adding (2) and (3),

52U 82. ( 8U 8U
X2 2 _-=(2cos2u-1)x+y-
8X 5y)
= (2cos2u- l)sin2u [ From (I)

= sin242(1_2sin 2 u)_i}

= sin2u(1-4sin 2 u).

Ex. 7. if f (x, y)= ((Y )) then using Euler's

theorem on homogeneous functions, show that

,2j 52f 2521
x--+2xy +y _--(x2+y
2) tB P 1999]

Solution : Letf(x, y ) = u(x, y) + v(x, y) + w(x, y)

x2n)
(x2 +y2r
whew, u(- v)= 2n(2n-1)
2n(2n-1) 2n(2n-l)
a homogeneous function of x and y of degree 2n,
 409
PARTIAL DIFFERENTIATION

v(x, y) = x4(-), a homogeneous function of x and y of degree 1,

and w(x. y) = a homogeneous function of x and of degree 0.

w (i),
We know, if f (x y) is a homogeneous function of x and y of degree
then
Ô2 2-
+ + y2 —4j- = no- 1)f
&2

Using this general result to u(x, y), v(x, y), w(r, y)

2j 2u ., 8u
x2--+2.xy—+Y — 9-=2n(2n--l)u
x&v öy

... (2)
x2+2x ' +y2 1 . (1-1)v0 •..
6x 28x &v y2

.. (3)
x2__+2y_^Y2_0(0_1)w_0
xy .5y
Adding (1), (2), (3),

x2__ +2Xv+V2 _ u+v+w)2n(2n1)U

&r 2 öxöy y2)

ô2f
+2xy—+
ôxöy
.2L 2n(2n - I) x
(x2+y2)"
2n(2n–l)
y2

then prove
Ex. 8. (i) If z.f(x, y) anox(u+v) 2 , y=(u–v) 2

z+ ( öz 6z'\. C.!'. 1992)

that .

show that
(ii) If V•is a function of(x, y)and x=e u. cost. y=e" . sin t.

+-)
(6v2 (8v'\2 =(x2 2 1(V 2 (ov'2
+v
j—) +--)
[C. P. 1993. 2003. 20081
 410
DIFFERENTIA L CA LCULUS

(iii) If z is a function of x, y and x = e" + e' and y= e

prove
that
T. & Ky [C. R 1997, 2000J

8z- z&
Solution; (i) = - -8zôy
+ -. -
8u x iu y bu

=2(u+v)+2(u— v)
(1)
zz &T 8z 51,
& 8x öv 5y
oz
TX (2)
Adding (1)and (2),
8z 8z
= 2{ u2 +uv+uv +2{u2
& 6v

=2(u+v)2

Idz dz
= 2 x +y -
(, (li dy
Let V = f(x, y). x = e 11 cost,
y e' sin
8V 8V ox ov o

8V 5V
=— .e°cost+--.e' sift
Ox
8V OV
=x— +y— — ...
&':Oy ... (1)
Ov OV Ox OV Oy
WT 8x Oy 8y 8:
Ov 1 '
— -e"s.n1)+
cost
Ox

=— Y M' Ov
TX v (2)
 PARTIAL DIFFERENTIATION 411

(öV 2 v\ 2 ( ov 8v 2 ( ov
I-I +1-I = I x— + y— 1 .+Ix---y—
u) I '. 8Y) 6y ox

2 2l
–(x2+
y2)
{(). +(
(iii) Let. z = f(x, y), where, x = e' + e s', v = eU _V

- Oz Ox Ox - Ox
---.—+----e --e–u-- ... (1)

Ox Oz Oy
8z
—.–+–.–=–e--e– ... ... (2)

From (I) and (2),

Oz Ox Ox .Oz
-- i- =(e" +e)–+(e –e")--

oz oz
=
Ox 0;
2dz It
Ex.9. (I) If z = e ' , x=(COSI, y=tSrnt, obtain at t.
di
C. P. 2001
(ii) If u = f(x, y) . and x rcosO, y = rsin8, prove that

( ö ..2 (o2 (5\2

+-) =;) +-).

1
2 To
)
C. P 1990, 2001, 2008

Solution : (i) z=e and x=tcosi. y=tsint,

8z 8z thOz dy
&Ox di 8v di

= e XY 2 . y2 (cos: - tsint)+ .2.y(sint + (Cost)

= ye 7CY2 {y(cos: - tsini)+ 2 4sint +tcost)}

7C 71 TE IC 71 71

when t-, x= Cos O and y=Sifl.

eXYe°1
 412 DIFFERENTIA L CA LCULUS

Hence, at £ =
2

&
i.1(o_!-"0=_
+ l!L
2t2I 2) J 8
(ii) Here, u = f(x, y) where, x = rcosO, y = rsinO
614 6U 8x 6u 6 y 6u
(I)

oa- Su 8y 8u
._+__..__(_rsjnO)+___.rCoSfi
co 6x68 Ty-
8 6x 8y

18,1 8u
or,= sin0+— cosO ... (2)
8v
From (1) and (2),
8U ) 2 i(8)2 (& s2
öu (64 6u
=j —coso+—sino I +1 —cosO--sinO I
+ r2 6 Q \&C öy ox
) 6y )

+&J
I
212 (co520+sin2O)
=i
)

f \2 = 2 ,-
iOu , iOu , iOu, I iOu
---)——
) + ) +
2xy
Ex. 10. (i) Show that (x, y) Jim (0. 0) 2 + ,2 does not exist.
—*
B. P 2001 1

(ii) If f(x, (x, y) # 0

X22 2'
=0,
Show that j,(0. O)= fyx (0 1 0). II C. P. 1995,2006, B. P. 1999]

y)=xy ,when both x. y O

f (0, 0)=0.

show that (0. 0)*f(0. 0). [ C. P 1988, 1999, 2004, 2007)

 PARTIAL DIFFERENTIATION 413

2xy
Solution: (i) f(x,
x2
Let, y = mx
2x-ma 2m
then f ' " x2
-I-rn 2 x 2 = 1i-rn2

urn f (x, y)
Thus (x, v) -4 (0, 0)) will assume different values for

different modes of aproach of (x, y) towards the origin.

urn f(x, Y) does not exist.
Hence, ( v

2y2
(ii) Here, f(x, y )_ when x 2 + y2 ^0

when x2-i-y2=0

k)_f(/z,0
f.9 (h, o)= urn
k-40 k

h2k2
-0
= lim h 2 i-k2
k-40 k
h2k
= urn =0
k-+0
f(0+k)-f(0. 0)
f (0, o)= urn
k

0-0
= km —=0
k-0 k
f(h, 0)-f (0,0)
f (0, 0)= lim
h-0 h
0-0
= km =0-
h-i-0 h
- f(h, k) — f(0, k)
Again. j(o. k)
= h-0 h

h2k2
---o0
h2 hk2
= urn = urn =0
h-i-0 Ii !i-#0
414 DiFFERENTIAL CALCULUS

f(h, 0)-f(o, 0)
and 0) urn
h-40 h

= urn
h-0 h

k)-f(0. 0)
0)= Jim Jim
k -0 k -0 k
0-0
(2

From (1) and (2), it follows that 4(0, 0 )J, (0. 0)

(iii) Here, f(x, ) 1. 2, when x and y both 0.

2

f(0, 0)=0

f (h, 0) = urn L' k) - f(/, 0)

k

(h2 +k2)
= Jim
k0 k

h(h2-k2)-0
= Jim
t k*O]

=h

jim j(o, k)-f(0, 0)

and fv( 0, o)=
k-I) k

= jim
R-() k

f (h , O)-f(O. 0)
• f (0, 0)= Jim
h

hO - 1
= Jim ... (1)
h ...
PARTIAL DIFFERENTIA T!ON 415

f(h, k)-f(0, k)
Again, f( 0, h
k) = h-*0
(h2_k2
hkl 1- 0
h2 +k2
= lim
h
+2 -k2)
= urn ---- [h*o]
h-0

-k3
A
-A2 -
f(h, 0)-f(o, o)
f(0, O)= liin
- h-30 11
0-0
= urn -- = 0
h-.O h

o, k )-f ( 0 , 0)
f(O. O)= urn •1(
A
-k-0
=urn
k -0 A
From (1) and (2), we conclude that
f (0, 0)*f ,(0. 0)

EXAMPLE-)M(B)
1. Verify Euler's theorem for the following functions
(I) u _2 +7hxv+bv 2 . ( ii) a + y 3 +3x2y+3.y2

x -y - x2+y2
(iii) U = - ( lv) u = sin
x +y xy

(v) a (x3 + v+ x +Y1)

dy
2. Find in the following cases:
dx
(i) .r +y 4a• (ii) x' +y a' LC.H.19441
 416 DIFFERENT/AL CA LCULUS

(iii) (cos x = (sin y )•' . (iv) e+e 2xv.

(v) y+x(x+y).

3. (i) If u = 0 ( H,, ). where H. is a homogeneous function of degree n

in x, y, z, then show that
au all au F(u)
X — + --+Z — =
ax a az F(u)

where F(u) = H,.

(ii) If u = cos' {( + y )/ then show that

+ Fy)),

++ J- cotu = 0. [B.P 1994, VP. '96,2001,2008]

ax a '-

4. If v = sin' {(x2 + y2 ) / (x + y) }, then show that

xV,+vV,, = tan V.

5. If u = x O(y / x ) + i,ei (y/x), then show that

a
(i) • -.- + y - = X0(y/x).
,ix a
(x u + 2xvu, + y u i.,, = 0.
6. (I) If i' f (a), u being a homogeneous function of degree n-in x,
y, show that
a a 3t'
x—+y--=flU— IC P 19481
ax a du'
(ii) If V be a homogeneous function in x, y, z of degree n, prove that
av av and -av
are each a homogeneous function in x, y, z of degree
-
ax ay az
(n—i). [C. P 20061
(iii) If V be a homogeneous function of degree ,, in x, y, z and if

V= j(x. Y, Z), where X YZ are respective'y---, , show

X-
ax ay az
av av av
that X+Y---+Z---V.
—
ax a Z n—i
7. (i) If H be a homogeneous function of degree n in x and y andif
2 2'"
u = I,x + y' show that
PARTIAL DIFFERENTIATION 417

-- hi + --- IH . =0.
Tx- ax ) ay '. ay )
(ii) If H be a homogeneous function in x, y, z of degree a and if
+ _2)_ show that
u = x2 +y 2

a ( au'l a ( au a ( au 'l
H—
— I H— 1+ — I )+ — I H—I = 0.
ax ax) ay ay az t az )
S. If x = r cost), v = r sin 0, prove that
(I) dx 2 +dy 2 -dr2 +r2 d0 2[B .P.2001]

(n) xdy-ydx=r 2 dO. [B.P.20011

9. If q(x,y)= 0, yi(x,z)= 0, prove that

a V, ()o ay-DO
ox ay Ox Ox Oz

10. Express A, the area of A ABC, as a function of a, b, C and hence show that
dA da A
— - + — + cot C dC.
A a b

II. If x 2 + y 2 + z 2 - 2xyz = 1, show that

dx + dy + dz -
- o
- 2 - 2

' 2 =1 and Ix+ my + nz = 0, show that

12. (I) If a 2 +by +cz

dx - d:
My - cmz ci; - w t r amr - bIy
2 .2 2 2 2 2
X + +
(ii) If and
,2 C-1 +A c 2+A -i- A
a2

X(b - c2) + z(a2 - b2) =

prove that +
dx dy dx
13. The radius of a right circular cone is measured as 5 cm with a possible
error of 001 cm, and altitude as c8 with a possible error of 0-024 cm. Find
the possible relative error an, percentage error in the volume as calculated
from these measurements.
27-
 418 DIFFERENTIA L CA LCULUS

14. The side a ofa triangle ABC is calculated from b, c, A. Ifthere be small
errors db, dc, dA in the measured values of b, c, A., show that the error in
the calculated value of a is given by
da = cos B.c/c + cos Cc/b ,ji, sin Ct/A.
IS. If f (p, t, v) 0, show that

(di (c/v
x .— I xI-..--.-I =–1
k di ) r t.. dv) q/i
L
16. If u F (y - :,: x, x - y), prove that

au oil
- +- +. =0
ax a 5 Dz
[C.P 1983, '88, '94, '98, 2002, 2007; V P '95]
17. If u=F(x 2 + 2 +2 )f(x,v+)'z+, -

() U Oil
(, .;--+(z–x)—+(x–y)--=0.

J, 2
18. If U = f (x + 2yz, + 2:x), prove that

/ \13U 'l
Iy –z)---+ ' x 2 –yz)-----+z2 –.iy,l—=O
ax
CE 1981, '95, 2008; C.H. 1947]
V 2 _,2
19. If F( V 2 – x2, 2•_2 )= o,

where v is a function of x, y, z, show that

la ylos' lOs' I
xOx y O y :0: v

20. If u be a homogeneous function of x and y of ii dimensions, prove

that

( a
U = — I )u,

where 1x-- + y_-'1 +2+2

OxOy Oj,2
a Ox2

V P 1998, C. P '85, BR '93, C. H. 46]

PARTIAL DIFFERENTIATION 419

21. Lfu=xØ(x+y)+y(x+y)

a2u alu a2u

2
dx2axay ay2

22. If V be a function of r alone, where r2 = + +

a2vvvv
show that
az2 r2 r a,-

ANSWERS

yxt +y logy
- .xy 1 + x log x

y tan .x + log Sill y - 2)'

(iv)
Jog cosx — xcot y 2x -

Y, logy + yx — (x+ y )' {log (.. + v) -Fl}

(v)
x log x + - + {log ( + )' ) 4 i }

13. 0.007 (relative error); 0.7 (percentage error).

[A ExrRnlA OF FUNCIIONS OF TWO OR MORE VARIABLES

13.1. Extrema with two variables.

A function f(x, y) of two independent variables x, y is said to be
maximum for x = a, y = b provided j( a. b) is greater than every other
values assumed by f(x.y) in the immediate neighbourhood of
x= a, y = b. Similarly, a minimum value of f(x. y) is defined to be the
value which is less than every other values in the immediate neighbourhood.
A formal definition is as follows:
A function f( x, y ). is said to have a maximum value at a point (a, b)
of the domain of f(x, y), f (x, y ), provided we can find a positive
quantity 5 such that for all values of x, y in a— 8< x< a+ 8 and
b-6<y<bi-ö, (x*a,y*b)f(a.b)> f(x,y);
i.e.,iff(a+h, b+k)—f(a,b)<O,

for l hl <8 and lkj<o.

Similarly, the function f( x. y) has a in in zmwfl value at a point (c, d ).
provided we can find a positive quantity 5' such that for all values of X. y in
c-8'<x< c+5', d-8'< y<d+ö',(x# c, yd) f(c.d)< f(zy);
i.e.. eff(c+h. d-i-k)-f(c,d)>O,

forjhl<5' and k1<8'.

13.2. Necessary conditions for Maximum and Minimum of extrema with

two variables.

if a function f (x, y ) be a ,na.xiniunl or minimum as x = a, y = b

and if the first partial derivatives f, (a, b ) and f, (a, b ) exist, then
f,(a,b) = O and f(a,b)O.

ProoL lff(a,b)bea maximum ora minimum value Of f(x,y).

then clearly it is also a maximum or a minimum value of the function f (x, b)
of one variable x for x = a and so its derivative f (a,b)forx=a must
be zero, provided it exists.
Similarly, f5(a,b)=O.
kXI'REMA OF FUNCTIONS OF TWO OR MORE VARIABLES 411

133. Determination of Maxima and Minima of extrema with two variables.

If (a, b ) be a point in the domain in which the function f (x, y) is
defined, and if f, (a, b ) = 0, f ',, ( a, b ) = 0. and
f(a,b).fyv(a,b){fy(a,b)}>0, then f(a,b) is
maximum ora minimum according as f, (a, b < or > 0 (and consequently
f(a,b)<or >0). But if f,(a,b).f,.(a,b) _{fa,b) <0,
f (a, b) is neither a maximum nor a minimum and if
(a,b ) ) 0 further analysis is necessary.
f,, (a, b ). f,, (a, b ) -
Proof of these results is beyond the scope of this elementary treatise.
Note. Points where f = 0 and f 0 are called stationary points. These
poiqts may be a maximum or minimum but in certain cases it may so happen
that the points is a maximum in respect of one variable while a minimum in
respect of the other variable.
13.4. Illustrative Examples.

Lx. 1. Examine for extreme values ofthefunction

x2+y2+(x+y+1)2, C.P. 19951

Let f(x.y)=x2+!?+(X+Y+1)2,
f=2x+2(.x+Y+1)=+2Y+2,

f, =2y+2(x+y+1)=2X+4Y+2,
f=4, f,51=4, f=2.
The equations f, = 0, fy = 0 are equivalent to
2x+y+10 and x+2y+1=0.
• xY I x Y
These give, or,
== -i = - = -
1 1
or, X=—, Y=— 3'

The function may have an extreme value at (- - . -

)z 4.4_22 = 12>0.
Now, at (-•, --k),
Also, f>0.
Therefore, f(x.y) isaminimumat (- .
 422 DIFFERENTIAL CALCULUS

Ex, 2. Find all the maxima and minima of ihe function

4x2 - xy+ 4? + x3y+ xj 4.

We have f, =8x-y+3x2y+y3

f -x+8y+x3+3xy2
ç=8+ a, fyy f=3x2+3y2-1.
The equations f = 0, fy = 0 give

8x-y+3x2y+y30

-x+8y+x3+3K1J=O
Adding () and (2) we get
7(x+y)+(x+y)=0, or, (x+y){(x+y)2+7}0,
or, (x+ y)= 0. .. (x+ y) 2 +7>0
=-x. (3)
From (1) and (3) we get
9x-4.=0, o r, x(-9)=O.

The corrcponding vaucs of y= 0, --. -

The function has three stationary points (o,o). (1,3 .

),

At(0,0),f,=8, f =8 , f,=-1,

f,j_(f2=64_1=63>0and f— 8>
the function is minimum at (o. o).
At(3 ' - fl . f -8 27_ 11
'2 - 2 - 2
' =8--- U
Jyy f =3( . 4^4
2 )_i Z2
2 2'Jt,

f f (f )2.121 625 126< 0

'-i 4 4
so that the function is neither a maximum nor a minimum
EXTREMA OF FUNCTIONS OF TWO OR MORE VARIABLES 423

(3 27 1!
At f = 8-- = - -i-

I f (f ) 2 =-126<0,

so that the function is neither a maximum nor a minimum

The points ( ,-- )ami (- ', - )where the function neither a

maximum nora minimum are called Saddle Points.
Ex. 3. Show that f( x, Y) = + 2-x-' y+ 2x 4 has a minimum at (0,0)
We have f=4xy+&
f,=2y+2x2

f=4y+24x2 , J=2 , f =4x

At (o,o)j=o. j,=O, f=O,

Since f, = 0, f = (o,o) is a stationary point of f(x, y).

2
Again J - = 0. So further analysis is necessary.
f(., Y ) = ( Y+ X ,
We hav . r(o,o)= 0 and ) 2 +x 4 . which is
positive for positive as well as negative values of x.
Hence, Ax, Y) is minimum at (0, 0).
EXAMPLES- XIII (A)

Examine the following for extreme values:

(1) x 3 -i-v 3 -3axy. (ii) 4.xy

(1
+ - - (iv) x3y (i - x - y).
x y

(v) sin xsin y sin (x + y ). (vi) x2y2 - 5x 2 - 8xv - 5y2.

(Vii) X4 +2x 2 y—x 2 +3y 2 . ( viii) 2(x— y) 2 —x 4 —y4.

(1x)y 2 +x 2 y+2x 4 . (x) x2+xv+y2-4x+y.

() xy (6a - x -y). (i) (2 + y2 )e62.

)2
(XiiO(x2 +)' )2 -8 ( X 2 - p2). (XiV)x2 + ,2 + ( a + by + c
424 DIFFERENTIA L CA LCULUS

(xv) x 3 + 3y 2 - 15x 2 - 15y 2 + 72x.

2. Show that the function f(x,y )= x 2 + 2xy + y 2 + X 3 + +
has neither a maximum nor a minimum at the origin.
3. Show that the function (x + y ) 4 + ( x - 3 ) has a minimum at
(3,-3 )
4. Show that the function f (, y ) =y 2 + 3x 2 y + 5x 4 has a minimum
at(0,0).
S . Show that the function f (x, y ) = 4x 2 y - Y 2 - 8x 4 has a minimum

at(0,0).

6. Show that the function f (x, y ) = 3x 3 + 4x 2 y 3;y2 - 4y is

neither a maximum nor a minimum at (0,0 ).

ANSWERS
1. (i) Max.at(a,a ) if,, <0 andmin.at (a,a )jfo >0.
(ii) Mm. at ( .i.- .i) and

(iii) Min. at0.0 (iv) Max.at(4).

J
(v) Max. at (n'+j, n'ir+j-)and

Mm. at (n'Jr - -, n'ir - i- ). n' being any integer or 0.

(vi) Max. atI 0,0 ,)\ no mm.
. ..
(vii) . at I
No max.,-mm
(viii) Max. at (.h,_.J) and (_.j,j).

(ix) Neither a max, nora min. (x)Max.at( _Z,_2

(xi) At(2a,2a )maximum if >0 and min. ifa <0.

(xii)Min.at (0,0)and (-i.o). (xiii) Min.at (2,0) and (-2,0).
( -ac -be
(xiv)Min.at - 2
+ b 2 ' I+a 2 +b2
(xv) Max. at ( 4.0) and nun, at (6.0).
 EXTREMA OF FUNCTIONS OF iWO OR MORE VARIABLES 425

13.5. Lagrange's method of undetermined multipliers.

n variables
Let = f (x1 , x2, x.. ..... .r,,) . . . (1) be a function of
it

which are connected by m equations.

91(x1,x2,x3.....
g2 ( x1 ,x2.x3 ......x,,)=O
.. .(2)

gm(XI,3(2, X3 .....
We have in (< n ) equations in n variables given in (2), so only
n - m variables are independent.
For it to be maximum or minimum_we should have

dud.+----th+....+---d)cO.
ax1 a ax,,
Also theequatlonn(2)giVe,

ax, a. ax,,

dg = cbj + dx + .... + = (1
ax1 a ax,,

Multiplying the equations in (3) by I, A1 , A2. .... .A,,, respectively and

adding we get

Fdx 1 +F, ('r 2 1" dx,=O

+,1"39 2 + . .. + A,,,
where F = --- + A1
aX, . ax

(k = 1,2 .... . n).

Them quantities A I, A.. .... .A ,,, are at our choice. Let us choose
these quantities so as to satisfy m linear equations
F 1 =O, F2=O,...F,,,=O
Then the equation (4) hecomes
F,,,.. 1 dr,,, 1 + F,,, 2 d4,,, 2 +... + F,, dx,, = 0 ... (6)
426 DIFFERENTIA L CA LCULUS

We have already noted that only n - in variables are independent.

To be specific let x,,,, 1 . x,,,2.....Jr. be these independent variables.
Since the quantities dx,,,., dr ,,, 2 .... . dx,, are all independent,
their coeficients must be separately zero.
Thus we get 11 — in extra equations
= 0, F,,,., = 0.....F,, ... (7)
Thus froni (2), (5) and (7) we have n + in equations
g 1 =0, gO,...g,,,=0
F1 =
O. F, 0. ... F, = 0.

From these equations we can find the multipliers A 2 ......A,,,.

and variables.' , x, ,...,x,, for which u is maximum or minimum.
10.06. Illustrative Examples

Ex. 1. Find the ,nini,nwn value of x2 + + z2 , subject ot the condition

2x+3y+5z=30
Let u=x2+y2+z2
and 2x+3y+5z=30 . . (2)
Fora maximum or a minimum value of u, du = 0,
i.e., 2.Lr + 2ydv + 2zdz = 0,

i.e., xdx+ ydy+ zdz= 0. . . . (3)

Taking differentials we get, from equation (2),
2dx+3dv+5dz=0. ... (4)
Multiplying (3) by I, (4) by A and then adding we get
(x+ A )dx+(y+ a)dy+(z+ 5A)dz= 0
Equating the coefficients of dx, dy, dz to zero we get,
x+2A,r O, . . (5)
y+3A=0, ... (6)
z+5X=0, . . . ( 7)
From the equation (5),(6) and (7) we get
x _ y _ z 2x+3y+5z 3015
2 352.2+a3+5.5 38 19
EXTREMA OF FUNCTIONS OF TWO OR MORE VARIABLES 427

i.e., x=-, y = 4. z=.

From the equation (5), A= - -fe.
u has an extreme value when x=4. y=-, z=.
From the equation (2), one of the three variables, say,z can be expressed
as a function of two independent variables x,, that is, z = 6— x - y.
az 2 az 3
ax 5 ' ay 5

Now, = 2x+ 2z = 2x—

ax ax 5
au az - 6z
- = 2y+ 2z— = 2y--
ay ay 5

a 2 1i 4a 8 58
-=2-- - -2+---,
ax2 5a 25 25
a 2 u 6 a 12
ax ai, —5 —
ax 25'

a2u 6az 18= 68

-2
ay2say 25 25

a 2 u a 2 u ( a2u25868 (12\2 152

I-j =•_) =7>o•

a2ti
Also —>0.
ax2
uisminiinumwhen19x, y -, z= and this niinjmum
value of u is
36' +4 + 75f - 855C
361°
EX-2. Find the maximum value of x 2 y3 z4 subject to the londition

x+ y+ z= 18.
428 DIFFERENTIA L CA LCULUS

Let u=x2y3z4.
lag u= 2log x+ 3log y+ 4log z.
Taking differentials we get
du 3 4
—=—dx+—dyi--dz (1)
U X J Z

For a maximum or minimum value of u, iu = 0.

• x +dy tdz =O • (2)

•. x y z

From the relation x+ z 18 • . (3)

we get dx+dyidz=O • . (1)
Multiplying (2) by 1 and (4) by A and adding we get

tX ) jJ ) (Z ')

Equating the coefficients of dx, dy, dz to zero we get

(5)

- + x =0. (6)
Y

-+A=O. (7)
z
From (5), (6) and (7) we get
2_342+3+4_ 9=1
x y z x+y+z 18 2

i.e., x=4, y=6. z=8, and A= — .

From (3) one of the three variables, say, z can be expressed as a
function of two independent variables x, y,
i.e., z=18— x— y.

1
ax '


EXTREMA OF FUNCTIONS OF TWO OR MORE VARIABLES 429

Now 2,c?z4+4N2y3z1_z

= 24z 4 - 4x2y3Z3.

-=3x2y2z4+4x2y3z3
ày

dx ax, dx

= 22--2JI+12Y3Z2

=22-16'qjY+12y32

=2yY(2-thcz+6J).

= 6y 2 z 4 + 8xv3 z 3 -12x2y2z3 - 12x2y3z2

a  a 

= 6 2 z 4 8 3 z 3 - 12x 2 y 2 z 3 + 12x2y3z2

= 2 xy _4yz_6xz+6xy).

6x2 yz -12x2 y2 z 3 -12x 2 y2 z3 +12x2y3z2

a y•

6x2z2
(2 - 4yz + 2y2)

For x=4, y6, z=8.

=-, 92 y3Z2,  .. = -336x2yz2.

=-96x.  2 z2 ,
dx. àxdy ày

32u
a2 (22
= (-192 )y 3 z 2 (-336 )x2yz2 - (96xy2z2 )2
3X 2 aY2 axayj

=192336 2 y4 z 4 -96 96x2 y 4 z 4 >0.

a2
Also —<O.
ai
 430 DIFFERENTIA L CA LCL'LUS

u is maximum when x= 4. y= 6, z= 8 and the maximum

value of u is 426384
EL 2. Prove that of all rectangular parallelopiped of same volume, the
cube has the least surface-area.

Let X . y. 2 be the length of three cotcrminous edges of the

rectangular parallclopiped. The volume is given by -
xyz= cL(given).

IfS he the surface-area then

S=2(yz-t-zx-i-xy)
and we are to minimize S subject to the condition (1).

Fora maximum or a minimum value of 5, dS = 0.

(y+z)th+(z+x),f(X+y)dZO

Also, from (1), yzdx+ zxdy+ xydz= 0. - (4)

Multiplying (3) by I, (4) by A and adding we get
(y+ z+X yz)dx+(z+X +zK)dy+(x+ y+A xy)dz= 0
Equating the coefficients of dx. dy, dz to zero we get
y+z+Ayz=0 -. (5)
z+x+Azx=O
x+y+Axy=0

1 I =1 I 1 1
+ =
—+———+— (=-)
Y Z 2 X y

1 1 :i \1
Le.,x=y=z=(jz)3a
xyz

that is, all the edges are equal.

Front one of the variables, say, z can be expressed as a function of

the independent variables x, y.

(1) gives yz+xq-=O. i.e.,

axax
az
Also __.__ -
Il
EXTREMA OF FUNCTIONS OF TWO OR MORE VARIABLES 431

as dz
Now —=2 y—+z--x-
—+y
dx dx

=2{ (+z)+(x+)(-)}

2 (x-z)y ( yz
=-(xtj+xz-zx-yz) =2 =2j y--
x x x

d 2 S2yz 2ydz
dx2 x2 x TX
_

2yz +
 .! -
I?- xxx2

o.:, 4zx
Similarly =—
dx2 y2

i--=21 x xdy

=21 1-^- =2.

x xy
For x= a, Y = 0, z= a,

a2 s a2s a2s
3T4' dxdy2

2
d 2 Sd 2 S d2S
2 dxdy =16-4=12>0.

Also
d2s
—>0.
Therefore, when x= y= z, that is, when the rectangular parallelopiped
is a cube, its surface-area, is minimum.
432 DIFFERENTIAL cALCUlUS

EXAMPLES - XIII (B)

1. ¶indthe minimum values of x 2 +y + Z2, when

(i) x+y+z=15, (ii) yz + zx + xy = 12., (iii) .xyz=8.
2. Find the extreme values of yz + 3zx + 2.y where x + y + z = 1.
3. Find the extreme values of .xy where x2 + . y +y2 = 1.
4. Determine the maximum and minimum values of 7x 2 + 8xy + y 2 when
x2+y2=l.
5. Find the minimum distance of the point (i, 2, 3 ) from the plane
x +y - 4z = 9.
6. Which point.ofthesphere x 2 + + z 2 is
= 1at a maximum distance
form the point. ( 2, 1, 3 )
7. If x, y, z are the angles of a triangle, then prove that the functions
(I) sin x sin y sin z;
Ir
(ii) cos x cos y cos Z ; are both maximum at X y Z =
8. Find the maximum value of the function x 2 y 2 z 2 . subject to the
condition x 2 + + z2 = c2.

9. Find the rectangular parallelopiped of maximum volume that can be

inscribed in the ellipsoid +--- +

--= I.

10. Divide the number 27 into three parts x, y , z such that

2yz + 3zx + 4.y is maximum.

ANSWERS

1. (i) 75for (5,5,5); (ii) 12foi(2,2.2), (-2,-2,-2);

(ii) 12 for (2,2,2),(-2,-2,2),(-2,2,-2), (2,-2,-2).

2. )- (max.) at (1,o,.);nomin.
3. —1 (min.) at (1, —1) and (-1, 1);

(max.). at( _j.r._jr) and 4. 9._I. 5. 'Iii.

2 1 3 ' 8. C 9. 8abc
6 23' 23' 23
TIT 714 714 27
I 3,F3
FIR TANGENT AND NORMAL

14.1. We shall now consider certain properties of curves represented by

continuous functions If the equation of the curve is given in the explicit
form y =f (x), we shall assume that f(x) has a derivative at every point,
except, in some cases, at isolated points. If the equation of the curve is
given in the implicit formf(x, y) = 0, we shall assume that the functionf(x, y)
possesses continuous partial derivatives f , and f , which are not
simultaneously zero. When the eqution of the curve is given in the
we
parametric form x= (t)' y= tj,(t ), shall assume that $'(, )and ,'(i)
are not simultaneously zero.

14.2. Equation of the tangeut.

Del. The tangent at? to a given curve is defined as the limiting
position of the secant PQ (when such a limit exists) as the point Q
approaches P along the curve (whether Q is taken on one side or the other
of the point P).
(i) Let the equation of the curve be y =f(x) and let the given point P
on the curve be ( y) and any other neighbouring point Q on the curve be
(x+,+y+y)
The equation of the secant PQ is,(X. Y denoting the Current
co-ordinates)

Y -y= Y+A y - y(X -X)= A1(X-X)

the equation of the tangent P is

Y - v= Li
d

provided dy/dx is finite.

Thus, the tangent to the curve y =f () at (x, y) (not parallel to the
y-axis)is

Y-y=-(X-x). ... (1)

dx
(ii) When the equation of the curve is f (x, y ) = 0-

Since = -5-. ( j- ,, ^o),

dX
434 DIFFERENTIAL CALCULUS

the equation of the tangent to the curve at (x, y ) is

(X-x)r+(y-y)r=o (2)
(iii) When the equation of the curve is x = 4(t), y = ji(t )

since t"(t)#O.
dr dif di o'(')
the equation of the tangent at the point 't' is

{
Y-W(t)=-?. x -(i )},
(3)
i.e., '(t ) x —'(: ) y =o(t )'(, )- (t )'(t ),

Note!. When the left-hand and right-hand derivatives at (x, ),)are infinite,
with equal or opposite signs, the tangent at (x,y) can be conveniently
obtained by using the alternatives form of the equation of the tangent
X—x = (Y—y)(d/d),) which can be easily established as before.
[Sce Ex. 32, Examples XIV(A) I

Note 2. In the notation of Co-ordinate Geomctiy, the equation of the tangent

to the curve y = 1(x) at ( 1 . y 1 )can be written as
Y — v,
in the application of Differential Calculus to the theory of plane curves,
for the sake of convenience, the current co-ordinates in the equation of the
tangent and normal are usually denoted by (X, Y ) while those of any particular
pohlt are denoted by (x, y). The current co-ordinates in the equation of the
curve are however, as usual, denoted by ( XI, Y1 ).
-
43. Geometrical meaning of dy
dx

rig 14j.I
 TANGENT AND NORMAL 435

The equation (I) of the tangent can be written as

dy. dy)
Y= X+ Y-X
dx dx
wh i ch being of the form y = t + c, the standard equation of a straight
line, we conclude that
dy
is the ,n' of the tangent at (x, y).
dx
If is be the angle which the positive direction of the tangent at P

makes with the positive direction of the x-axis, then tan i= m =

dx
Hence, the direvative 2LV at (x, y) is equal to the trigonometrical
dx
tangent of i/zn angle which the tangent to the curve at (x, y) makes with
the positive direction of the x-axis. [See Art. 7.14]
I

Note 1. It is customary to denote the angle which the tangent at any point
on a curve makes with the x-axis by iV.
Note 2. The positive direction of the tangent is the direction of the
arc-length s increasing. Henceforth, this direciton will he spoken of as the
directio'n of the tangent or simply as the tangent.

Note 3. tan 4i ,i.e., is also called the gradient of the curve at the point
dr
P(x,y).
Note 4. The tangent at (x, y) is parallel to the x-axis if i, = 0, i.e., if
dy
tan, = 0,i.e.,if = 0.
th

The tangent at (x, y). is perpendicular to the x-axis (i.e., parallel to the
y-axis)ift/i = - 1r ,i.e.,if cot V = 0.

/dy dx
if or,-=0.
Ids dy

14.4 Tangent at the origin.

If a curve passing through the origin h, given b y a rational integral
algebraic equation, the equation oft/ic tOo n (or tangents) at the origin
is obtained by equating to zero. the tc,?n of the looe.s-' degree in the
equation.
436 DIFFERENTIAL CALCULUS

Let the equation of a curve of the n-th degree passing through the
origin be
ax + b 1 y + a2 X2 + b 2 xv + c 2 y 2 + ... + a,,x" +.•• +k,, y' = 0
(I)
Let P (x, y) be a point on the curve near the origin 0. The equation of
the secant OP is Y =2X

the equation of the tangent at 0 is

Y = Li X = mX (say) .(2)
OX
y-0

Thus, the 'm' of the tangent at the origin is Li -

- -

CASE!. Let us suppose that ,n is finite, i. e., the y-axis is not the
tangent at the origin.
(i) Let us suppose b 1 * 0
Dividing (1) by x, we get

a t +b1--+a2x+b2y+ei'..+ ......=0

Now, let x—* 0, y - 0, then Li(!.r)= in.

a1 + b1m = 0, the other terms vanishing.
(3)
From (2) and (3), the equation of the tangent at the origin is
atX +bjY =0,
or, taking x and; as current co-ordinates, a .v + , i y = 0.
 437
TANGENT AND NORMAL

(a) if b1 = 0 .then from (3) it follows that a 1 = 0; now in this case,

let us suppose that b 2 and c2 are not both zero. Then, the equation of the
curve (1) can be written as
X 2 +b2 .xy+c2 y 2 + a3 x 3 + 0 ... (4)
a2

Dividing by x 2 , a2 +b 2 + 2 [) +a 3 x+ ......=0.

When x -4 0, y —p 0, we have
a 2 + b 2 m + c2 m 2 = O,the other terms vanishing....(5)
From (5) it is clear that there are two values of m and hence, there are
two tangents at the Origin and their equation, which is obtained by eliminating
,n between (2) and (5), is
a 2 X 2 + b,XY + c 2 Y 2 =
or, taking x and as current co-ordinates,
a2 x 2 +b 2 xy+c2 y 2 = 0. -

If a1 = b 1 = a2 = 112 = c2 = 0 , it can be shown similarly that the rule

holds good then also; and so on.
CASE II. When the tangent at the origin is they-axis, then Lt (x/y ), as .r
and y both -9 0, being the tangent of the inclination of the tangent at the
origin to the y-axis, is zero. Hence, dividing throughout the equation of the
curve by y, and assuming a1 * 0, and making x and y both approach
zero, we find b 1 = 0 . Hence, the-equation of the curve now being
a1 x + a2 x 2 + b2 xy + C2 V 2 + ..... = 0
we see that the theorem is still true in this case.
Illustration: If the equation of a curve be x 2 - y 2 +x 3 +3x 2 y - y 3 = 0
the tangents at the origin are given by x 2 - y 2 = 0, i.e.,x + y = 0 and
x—y=0.

14.5. Equation of the normal.

Definition. The normal at any point of a curve is the straight line through
that point drawn perpendicular to the tangent at that point.
Let any line (not parallel to the co-ordinate axes) through the point be
(x,y)be
Y - y = m ( X - x).
438 DIFFERENTIAL CALCULUS

This will be perpendicular to the tangent (not parallel to the co-ordinate

axes) to the curve y = f (x ) at (x, y ).
dy dv Idy
i.e.,to Y - y = — (x - x )if in -i-- = —1 , i.e. if m - 1/ -,
Substituting this value of in in the above equation, we see that the normal
to the curve y = f (x ) at (, y ) ( when not parallel to the co-ordinate
axes) is
dy
(Y —y )+ (X—x)=O ... (1)
dx
Similarly, if the equation of the curve is f (, y ) = 0 ,the equation of the
normal at is
X—x - Y
2
()
fy
and ifthe parametric equations ofthe curve are x =Ø(t y = v'(' ),the
equation of the normal at the point 'I' is
)x+'(z )y = (, ) Ø'(z )+ V (' )w '( i ) ... (3)

Note 1. When the tangents are parallel to OX and OY the nor mals are
X = x and Y = y respectively.

Note 2. The positive direction of the normal makes an angle + 1 It with

the tangent, or I it + 41 with the x-axis.

14.6. Angle of intersection of two curves.

The angle of intersection of two curves is the angle between the
tangents to the two curves at their common point of intersection.
Suppose the two curves f (x, y ) 0, 0 (x, y ) = 0 intersect at the
point (x, v).
The tangents to the curves at (x, y) are
X f,.+Y fy— (xf+yf)=O. [by*142(2)]
i- yø — (xO + y )= 0.
The angle a at which these lines cut is given by
f'P -
tan a=
+ fvøv
Hence, if the curves touch at (x, y ), a = 0, i.e.. tan a = 0
i.e.. .t;o. = i.e., f, /0, =
 439
TA NGENTA NIJ NORMA L

and ((they cut orthogonally at at (x, y), a = ,i.e., cot =

i.e., f,b.+fy4v=O.
Note. If the equation of the curves are given in the forms y = 1(x),

y = 4> ( x ) the angle of their intersection is given by tan . ,,

l+fx)4>x
Hence, the curves cut orthogonally if f'(x ) '(x )= —1.

14.7. Cartesian Subtangent and Subnormal.

Let the tangent and normal at any point P ( x, y ) on a curve meet the
x-axis in T and N respectively and let PM be drawn perpendicular to OX

r5. --

Then, TM is called the subtangent, and MN the subnormal at P.

In the right-angled triangles PTM, PNM,
since LNPM — LPTM =/,, and PM = y,
Y/(dy
sublangent = TM = y cot ji =
dx

subnormal = MN = y tan i/i = y d

dx
Note. PT and PN are often called as the length of the tangent and the
length of the normal (or sometimes simply tangent and normal) respectively.
Thus, from g PTM and PNM.
PT=y cosec yJ 2 s = y,Ji+(i/y) (y">2)

2
PN=ysec
 440 D!FFERi2NTIAL CALCULUS

Lt chordPQ
14.8. Proolof =
Q-P arcPQ
Let PP 1 , T, P ..... P,, - , - Q be the sides of an open polygon
inscribed in arc PQ of the curve y = 1(x). If the sum of the n sides
tends to a definite limit when n —* and the length of each side
tends to zero, that limit is defined as the length of the arc PQ.
.P2

2=S Fig. 14.8.1

Let 8 1, 0...... . 0,, be the angles which the sides make with the chord
PQ, and let f '( ,% ) be continuous throughout PQ.

Projecting the sides on PQ, we have

PQ=proj.PP 1 +proj.P1P2+...+proj.PQ

='PP i cos6 1 + P1 P, cos0 2 + ...+ P, Qcos8

it follows that PQ <PP +P P 2 +...+P,,_1Q

and > 1 + P 1 p 2 + .. . + P,, I
Q ) cos o
( p

where 0 isumei- ical ly the greatest of the angles 8 1, 8 "...' 0

Hence, cos 0 < < I.

PP1+P1P2+...+pQ

Since the chords PP. P1P.. .... .P, -, Q as well as PQ are parallel
to the tangents to the arcs at points between their respective extremities (by
the Mean Value Theorem), it follows from the continuity of f' ( x) that the
numerical value of 8 can be made as small as we please by taking Q
sufficiently near to P, and, in the limiting position, cosO -4 land
PP1 — * arc PQ.

chord PQ
Li
(2-.P arc PQ
TANG&VTAND NORMAL 441

14.9. Derivative of arc-length (Cartesian).

Let P(x,y)be the given point, and Q ( x + Ax, y+ ,iy) be any
point near P on the curve.

rig i'.Y.i
Lets denote the length of the arc A? measured from a fixed point A on
the curve, and let .c + js denote the arc . A Q. so that arc PQ = As Here,
s is obviously a function of x, and hence of y. We shall assume the
fundamental limit
Lj Chord PQ
=
Q .-.P are PQ

From the figure, (chord PQ)' = PR' + QR 2 = (Ax )2 + (Ay )2.

hord PQ
•
As ) Ax) Ax)
Now let Q - P as a limiting position; then Ax -, 0 and we have

(dsY (a,:2
I — I =I+I---I •.. (1)
dx) dx)

ds
=I+(_!L)2
or (2)
OX dx

ds dsd . dx
Since - - we get, on multiplying both sides of(2) by -.
dy. ds dy dy

ds [ d.\
(^
(3
dy y)
 442 DIFFERENT/AL CALCULUS

€or. Multiplying both sides of (1), (2) and (3) by d.c 2 . dx, dy, we get the
corresponding differential form
ds 2 = dx 2 + dy2;
2
dx)
dx; ds= dy.
dx) \I (dy
ds1+(
14.10 Values of sin ji, cos w.
From tPQR [See Fig., 14.9], sin QPR =
PQ A s
In the limiting position when Q - P, the secant PQ becomes the
As = -
tangent atP, ZQPR and 0 and 1.
PQ chord PQ
sin ty = Li = . ... (1)
A-0 As ds

Similarly, cost/i = Li = ... (2)

- -.O A s ds
dx
Ay- cot i = -, we get, from (2) and (3) of
Since tan Vand
dy
dxdx
Art. 14.9, - = Sec V. - = coscc
(Lx dy
w,
whence also cos ir, sin are obtained.
2
dx ( dy
+-J 1. ... (3)
()
dx dx ds . dy_dy dx
Cor.!fx=(1)(t), y='qs(t),—=—'—.-----
di dx di di d.c di
dX)2
(dy'\2(dx'2y'2'2
( d tdi) ds) (ds) dt

(dx '\2 (dy "2 (dsY

I—I +1—I = 1—I ... (4)
Idt) (dt) Idt)
Note. Relations (2) and (3) of Art. 14.9 can also be deduced from the values
of sin i, cos ti, tan t41,
14.11. Illustrative Examples.
Ex. 1. Find the equation of/he tangent at (x, y) to the curve
, '2
(x/a)' +(y/b) =1.
 7ANGENTAND NORMAL 443

Here the equation of the curve js f(x,y)(x/a)4 -i-(y/b) — 1=0.

The equation of the tangent is
(x — x)f +(— y)j =0,

i.e.,

i.e., X x /a +Y y /b i =(x/a) +(y/h),

I A, 2
i.e., X x / a' i-Y y 7 b ' 1.
Note. The equation of the tangent should be simplified as much as possible
as in the above example.

Ex. 2. Find the angle of intersection of the curves X 2 - y2 =a 2 and

x 2 +y 2 =a 2 ,h.[ Patna, 1940]
Adding and subtracting the equations of the two curves, we find their
common points of intersection given by 2x 2 =a2(,J + i.e..
and 2y2=a2(j_1),i.e., y=±aJ_i)/.J.
Since the equations of the curves can be written as
f(x,y)=x2 — y — a 2 =0 and ,(x, y)x 2 + 3, 2 _a2=O
hence if a be the angle of intersection of the curves at (x, y), we have, by
Art. 14.6.

tan a = 222 4(_ 2 =

±2xy =1
2x.2x + (-2y)(2y) x2— y2
on substituting the values of x and y found above. Hence, a
the curves intersect at an angle of 45.

.Ex. 3. Find the condition that the conics

ax +by 2 =1 and a1x2+b1y2=1
shall cut orthogonally.
The equations of the conics are
f (x,y)ax 2 +by 2 — 1=0, ... (I)
4(x, y)a1x2 +by 2 — 1=0 ... (2)
Now, the condition that they should cut orthogonally at (x, y) is. by
§14.6.
444 DIFFERENTIA L CA LCULUS

LO, +f'4, =0,

i.e., 2ax.2a1x+2by.2b1y=0,
i.e., aa 1 x 2 + bb, y 2 =0 ... (3)
Since the point (x, y) is common to both (1) and (2), the required
condition is obtained by eliminating x, y from (1), (2) and (3).
Subtracting (2) from (1), (a—a 1 )x2 +(b— b1 ) 2 =0 ... (4)
Comparing (3) and (4), we get
a—a 1 b— b1I Ii 1
or, a a b1 b
aa 1 bb,
which is the required condition.

Ex. 4. If x cos a + ysin a = p touches the curve

'=1
am b"'
show that (acosa)Ti +(bsina) i = p". [C.R 1996]
The equation of the tangent to the given curve at (x. v) by formula (2)
of Art. 14.2 is

(x x)mflx +(r- y) — -0,

i.e., X x'°/a" + Y y"' lb" = x"/a" + y"' /b" = 1 ... (I) -

If X' cos a+ Y sin a = p ... (2)
touches the given curve, equations (1) and (2) must be identical.
- y"''/b' I
Hence, -. sina - p-,
cosa
.x m_h /a_ t ym_h/b_l I
i.e., acos(X - bsina p

acOSa (Y b sin a
a) p 'tb)

(acosa')'+1bsina' irir =1
I\ p t\ 1' ) a) b)
)

i.e., (a cos a)+(b sin a) f ' = p'.

TA NGENT A ND NORMA L 443

Ex. 5. If x,, v, be the parts of the axes of x and y intercepted by the

tangent at any point (x, y) to the curve (x/a ) + (y/b ) = 1. show that
xjh/a2+y12/b2 1
The equation of the tangent at (x, y) to the given curve is, as in Ex. 1.
__1/2 _.t/l
Xx /a'+Yy /b'=l.

Where it meets the x-axis, Y=0. hence X =ax ,i.e., x 1 =ax' and
2 2 I
where it meets the y-axis,X=0, hence Yay ,i.e., Y 1 =by.

±/ 2 2
2/(12
X1 + y 1 2 /b 2 = a 3 x 3 /a 2 +b 3 y 3 /b 2 =(x/a) 3 +(y/b) 3 = 1

EXAMPLES - XIV(A)
1. Find the equation of the tangent at the point (x, y) on each of the
following curves:
x2y2
(ii)

(iii) X+)' ' =a3 (iv) x3 -3a.xy+v 3 0.

(v) (x2+y2)2=a2(x2-y2)_-.

2. (i) Find the equation of the tangent at the point 0 on each of the
following curves:
(a) x=a cos O, y=b sin 6.

(Li) x = acos3 9, y = bsjn3 0.

(c) x=a(6+sin8), y=a(1-cosO).

(ii) Find the equation of the normal at 'r' on the curve

x = a ( 2 cos t + cos 2t ), y=a(2 sin i- sin 2t).

3. (i) Find the tangent at the point (1, -1) to the curve
X3
+xy 2 -3x2 +4x+5y+2 = 0.
(ii) Show that the tangent at (a, b) to the curve
)3 )3
(x/a +(y/b = 2 is x/a + y /b = 2. [C.? 19431
(iii) Show that the normal at the point 0 = - yr on the curve
X = 3 cos 0 - cos 3 6 , y = 3 sin - sin 0 passes through the

origin.
446 DIFFERENTIAL CA LCUI (IS

4. (i) Find the tangent and the normal to the curve

—2 )( x --3 )- x + 7 = 0
at the point where it Cuts the x-axis.
(ii) Show that of the tangents at the pints where the curve
Y = (x — i )(x— 2 )(x-3) is met by the x-axis, two are parallel,
and the third makes an angle of 135 with the x-axis.
(iii) Find the tangent to the curve xy 2 = 4 (4 -. x ) at the point where
it is cut by the line y = x.
S. (i) Find where the tangent is parallel to the x-axis for the curves
(a) v = x 3 -- 3x 2a 9x + 15 (b) r' + 2/zxy + /,y 2 = 1.
(ii) Find where the tangent is perpendicular to the x-axis for the curves
(a) y 2 = x 2 (a - x). (I,) ) + 2/ixy + by 2 = 1.
(c) y= (x-3)2(x-2). [C.R1935]
(di) Show that the tangents to the curve
3x2 + 4xy + 5y 2 — 4 = 0
at the points in which it is intersected b y the lines
3x 4- 2y = 0 and 2x -F 5y = 0
are parallel to the axes of co-ordinates.
(iv) Find at what points on the curve
y = 2x3 - 15x 2 + 34x —20
the tangents are parallel to y + 2x = 0.
(v) Find the points on the curve y = xZ + lv -t- 4 the tangents at which
pass through the origin.
6. Show that the tangent to the curve -t v3 = 3uy at the point other
than the origin, where it meets the parabola y2 = ax ,is parallel to the
v-axis.
7. Prove that all points of the curve
= 4a{x-i-a sin (x/a))
at which the tangent is parallel to the x-axk lie on a parabola.
- [C.P.19981
8. Tangents are drawn from the origin to the curve y = sin x. Prove that
their points of contact lie oo .vy 2 = x 2 - 1 ,2 -
9. (i) Show that the curve (x/a -F ( v/b )' = 2 touches the
straight line x/a + y/b = 2 at he pont (a, b), whatever be the
value of a.
 TANGENTAND NORMAL 447

(ii) Prove that + 1 touches the curve + log l.') = o.

a b a
10. (i) If Li+ my= 1 touches the curve (ax )n +(by )f = I, show
• that (1/a)"-' +(m/b)-' = 1. [8.? 19891
(ii) If lx + my = I is a normal to the parabola y 2 = 4ax, then show
that a1 3 +2a1m 2 =m 2 . [ V P 19991
11. Prove that the condition that x cos a -t v sin a = p should touch
= a'" is p""'m"n" (m + a )"""" sin" a cos'" a.
12. Find the angles of intersection of the following curves
' 2 2
(0 x – y = 2a and x2 + y2 = 4a-'
(ii) x2 = 4y and y(.2 + 4) =8.
Y
= x3 and 6y =7—X2..

13. (1) Prove that the curves x—2+- —y2 x2 v2

= I and ----- + -- = I will cut
a b a b
orthogonally if a -- h =a'— h'.[C:P 1980, '902007 V P. 20001
(ii) Find the condition that the curves a? + b y -3 = 1 and
a? + h'v 3 = 1 should cut orthogonally.
(iii) Show that the curves x 3 -- 3xy 7 = —2 and 3x 2 y—y 3 =2 cut
orthogonally. [C P. 20061
14. (i) Prove that the sum of the intercepts of the tangetit to the curve
+ .jy = la upon the co-ordinate axes is constant.
(B.P. 1993
(ii) Find the abscissa of the point on each of the curves
(a) ay 2 = x3.
(h) FU = a + x, the normal at which cuts off equal intercepts
from the co-ordinate axes.
15. Show that the portion of the tangent at any point on the following
curves intercepted between the axes is of constant length.
0) r + y = a ,3 C.?. 1940]
(ii) x = a cos 3 0 , v a sin 3 0
16. if the tangent at (x,, y,) to the curve .v 3 4 3.3 a 3 meets the curve
again in Or, 32' show that X 2 /-, + =— 1
448 DIFFERENTIAL CALCULUS

17. (i) Show that at any point on the parabola 1 ,2 = 4ax ,the subnormal
is constant and the subtangent varies as the abscissa of the point
of contact.
(ii) Show that at any point on the hyperbola xy = ,the subtangent
varies as the abscissa and the subnormal varies as the cube of the
ordinate of the point of contact.
18. Prove that the subtanent is of Constant length for the curve
logy= .v log a.
3
the square of the subtangent
19. Show that for the curve by 2 = (x + )
varies as the subnormal. C. P 2006 J
- Y2"
20. Show that at any point on the curve x' = k- " y 2" the m' power
of the suhtangènt varies as the n" power of the subnormal.
[C.P. 1995, '97.2002, 2004]
21. For the curve X")" a"', show that the subtangent at any point
varies as the abscissa of the point. -
22. Show that for any curve the rectangle contained by the subtangent
and subnormal is equal to the square on the corresponding ordinate.
[C. P. 2005 1
23. Find the lengths of the subtangent, subnormal, tangent and norn3al of
the curves.
(1) x=a(9+ sin 0), y=a(I— cos 0)at'B'
(ii) x=a( cos t+l sin l). y=a( sin z — t cos t)at':'.
I C. P 20061
24. Find the value of i so that the subnormal at any point on the curve
xv" = a may be constant.
25. Show that in any curve
subnormal = length ofnormal'1
subtangent I\ length of tangent)
26. Show that the length of the tangent at any point on the following
curves is constant
, a a_2_y2
12
(I) x = ja - Y2 - log
2 a+a_y2

(ii) x = a(cost + log tan -r), y = asinl.

(iii) s = a log( a/v ).
TANGENT AN!) NORMAL 449

27. (i) If j, and p, be the perpendiculars from the origin on the tangent
and normal respectively at any point (x, y) on the curve, then show that
p1 x sin I,ff - y cos i, p2 = x cos tjl + y sin !1,
where, as usual, tan ty = dy/dx.
(ü) If, in the above case, the curve be x + yi = show that
4p 1 't-p, =a
28. In the curve - T y" = a'" show that the portion of the tangent
intercepted between the axes is divided at its point of contact into
segments which are in a constant ratio.
29. (i) In the catenary y = c cosh (x/c )show that the length of the
perpendicular from the foot of the ordinate on the tangent is of
constant length. [C.? 19431
(ii) Show that for the' catenary y = c cosh (xfc ) the length of the
normal at any point is y2/c.
30. Prove that the equation ofthe tangent ' to the curve x = af ON (t),
y = aØ (i)/' (i) may be written in the form
X y a

f(r) 00 ,j,(t) =0
f' (t) '(z) v"(:)
31. Find the equation of the tangent at the origin of the curve
y=. 2 sin(1/x) for x'*O
=0 for x=O.
32. Show that for the curve ) = x3 the tangent at the origin is x = 0,
although dy/dx does not exist there.
33. If a and 0 he the intercepts on the axes of x and y cut off by the
• tangent to the curve (x/a)" + (y/b )" =1 then show that28

(a1a)+(b1/3)il
ds
34. Find ' for the following curves:
dx
(1) y 2 =4tir. (ii) x 3 +y' ra'.

(iii) y - a(e" +e

• cos O),ya(O+ sin O)

(iv)x=a(l-
29-
450 1311-1 /RENT/AL Q.4LCULUS

35. If for the ellipse x 2 /a 2 + y2 /b 2 = 1, x = a sin 0 .show that

ds =al_e2sin2O
do
36. Two curves are defined as follows:
(I) x=t
y=t 3 sin (l/t), for i^,0
=0 for t=O.
(ii) x = 2t + 2 sin (I/i ), y = 2 sin (i/i ), for , 0
x=0, y0 for ,=O
show that, for the first curve, although d/dt . dv/dt are continuous
for : = 0, the cun'e has no lange,?t at the point; and for the second
curve, although dx/dt d y/th are not continuous fort = 0, the curve
has a tangent at the point.

ANSWERS

1. (i)
X rY y
=. (ii)
X "", Y v"
=
a 1' a
(ili) X X 1, +} 05.(iv) X (x2_ay)+y(2_ax)axy.

(v) x + y2 )- a2x }+ Y2y ( +

3,2 )
ay }= a2(x2 - 2 )•
2. (t)
x y
(a) - cos 0 + - sin 0 I.
a b
(b) bX sin O +aY Cos 6 absinU Cos O
(c) X sin -0 - Ycos 0 = aO sin -0
(ii) X cosi - Y sin -- t =3a COS -t.
3. (i) 2x + 3y + I = 0.
4. (i) Tangent x-20y-7 = 0; normal 20x+ y- 140 0.
(iii) x+y-4=O
5. i) (a) (3,-12), (-1,20).
(b) Where ax + hy = 0 intersects the curve.
(ii) • (a) (O.0),(a,O).
(b) Where hx + by = 0 intersects the curve.
(c) No such point exists
(iv) (2,4); (3,1). (v) (2,14): (-2,2).
TANGENT AND NORMAL 451

12. (i) ,r. (ii) -1

13. (ii) aa' (b - b' ) + bb'(a - a' )l = 0.

14. (ii) (a) A a (b) ±

23. (i) a sin 0, 2ain 2 1 0 tan . -0, 2a sin O, 2asin +0 tan
(ii) y cot I, y tan I, y cosec t, y sec I
24. –2. 31. y=0.

34 (i)(Ii) (iii) (iv)

14.12. Angle between Radius Vector and Tangent.

If 0 be the angle between the tangent and radius vector a, any point
on the curve r = f (0 ), then

rd0rdO dr
tanZ = -, s it! = - cosø
dr ds' ds
Let P(r,0 ) be the given point on the curve r j (0 ), and
Q ( r + A r, 0 + 46 )be a point on the curve in the neighbourhood of P.
Let QP be the secant through Q, P. Draw PN perpendicular on OQ.
Then, LPQN=A O, PN = ,-sin4O, ON=r cos 40.
Let 0 be the angle which the tangent PT at P makes with the radius
vector OF, i.e., LOFT =
From the right-angled E PQN,
r sin A O
tan PQN
=—PN = PN =
NQ OQ – ON. r+A r– rco.sA O
r sin AO - rsin40
= r( 1 - cos 46) + A r - 2r sin 2 40 + 4,

= r.(sinzi0/A 0) -
--rA 0(sin .40/-A 0)2+(4r/49)

(i.e., dividing both numerator and denominator by 40 1

Now let Q -+ P, then 40 ..-) 0, and secant QP becomes the tangent
PT, and LPQN -9 LOFT, i.e., 0.
452 DIFFERENTIA L C4LCULUS

(,J A
Fig 14.12.1

r.(sin49/48)
tanO = Ld
M?-O ,rziO (sin --A01-A0)2+(4r1A6)

/dr . dO
= ri - ---- (i.e.. = r/r ) = r
idO dr

I since Li (sin AO/AO ) and Ii (sin - A19/1 )are each equal to 1!.
Now, lets denote the length of the arc AP measured from a fixed point
A on the curve and lei s + As denote the are AQ., so that the arc
PQ A s. Here s is obviously a function of 0. and hence of i:
sin PQN= PN r sin AO 46 As
—=
- - I'Q 46 As PQ
Now let Q —s P, then 40 --. 0, As -4 0 and then
As/PQ = (arc PQ)/(chord PQ) -9 1. -
dO
sin O= Li r 46
—r—
-.o As ds

Again, cos PQN = = --- =

PQ PQ PQ
= r(l--cosA9 )+Ar r.2sin 2 I 4O +4r
PQ PQ
(sin i. zie 2 46 As Jr As
= _i i17 0. I -- + - -. -
2
j.-A6) As PQ As PQ

Now let Q - P Then, as before,

.ir
cos Ø Li —=r dr
--
- -o As ds
- Otherwisc:
-. dr rdO dr dO -dr
cosØ=cotØ - sinø --—
rdO ds dO ds ds
 453
TANGENT AND NORMAL

Cor. 1. Form AOPT. ZPTX = ZPOT

+ LOFT
yO+f.

Co2.
r. + r 2( ') 2 = cos2 + sin' , =

14.13. Derivative of arc-length (Polar).

previous article, we have
With the notations and the figure of the
)2
PQ 2 = I'N 2 + QN 2 =( rsifl 40 )2
+(r + 4r - rcos AU

=(rsin49) 2 +(r.28in2+A0+4r)2.
)2 we get
Dividing both sides by (40
21
PQ ..r 2 1+ 2 1.r1-- 40)
As 40) \ 40 )

in the limiting position, when Q —3 P and AO--) O.

2 2
(I)
r+I'—-J
dO) dO)

do (2)
i.e.,ds = r
+
dO
Multiplying both sides of (2) by
dr
-,we get

do ... (3)
11+ r— I
dr ' dr)
2, corresponding
Cor. Multiplying (1), (2) and (3) by dO 4[0, dr, we get the
differential forms
ds2 =dr2+r2d02.

ds=[]d0.

ds
fl( 1+1 r
^dO"2
I dr.
)

Note. Relations t2) and (3) can also be deduced from the values of sillk.
Cos 4t. tan 4.
454 . 1)/I-ti-RENT/A!. C'A!.C(JI.IIS

14.14. Angle of intersection of two curves (Polar)..

Suppose two curves r = f (o ), r = 0 (6 ) intersect at the point P
and let PT, , pi be the tangents at P to the two curves, and let
ZOPT1 =Ø 1 , ZOPT,=Ø.,

Fig 14.14.1
Then, if a be the angle between the two curves, a =
-
tan Ø 1 - tan O,
tana= —
I + tan tan 0 2
Since tan = r/r = .f (0 )If' (a )and tan r/r = 0 (0 )/(e).
0,
we get

tan a = f( 0 )0'( o )- f , (0 ) 0 (0
1(0 )'(o )+o— )0 ( t ).

14.15. Polar Subtangent and Subnormal.

Fil

T
TANGENT AN!) NORMAL 455

Let P be any point on the curve r = f (o ), and let the tangent PT

and normal PN at P meet the line drawn through the pole 0 perpendicular
to the radius vector OP in T and N respectively.
Then OT is called the polar subtangen: and ON is called the polar
subnormal.

Since, LOPT = , OT = OP tan Ø = r.r -

dr
2d9
polar subtangent = r
dr
dr
A gain, ON =OPtanOPN =rcotr--.
rdO
dr
polar subnormal=
dO
1 du ldr
Note. !fu=-,
r dO r2dO
ldO dO
P subtangent = - =-

14.16. Perpendicular from the pole on Tangent

Let p be the length of the perpendicular ON from the pole 0 on the
tangent PT at any point P

T
Fig 14.16.1
Then from A OPN, ON = OP sin 0.
p=r sin . ... (I)
I I l
Again, = - cosec2 0 = I + cor
p r r
456 1)IFFEREN7JAL CALCULI/c

1 1(dr't2
1 +-I-
r 2r2dO

1 1 1(drY
P2
• —=—+—I— 2
r 2 r4dO
The symbol u is generally used to denote hr. the reciprocal of the
radius vector.
du_ I dr
dO rdO
Hence, the relation (2) becomes
(du'2
—=u2+ j... (3)
p2

14.17. The (p, r) or Pedal eqation of a curve.

The relation between the perpendicular (a) on the tangent at any point
P on a curve and the radius vector (r) of the point of contact P, from some
given point 0, is called the (p, r) or pedal equation of the curve with regard
to 0. Such equations are found very useful in the application of the principles
of Statics and Dynamics.
(i) Pedal equation deduced from Cartesian equation.
Let us take the origin as the point with regard to which the pedal
equation is to be obtained, and let .1 (x, y ) = 0 be the equation of the
curve.
The tangent at (x, y) is X f + Y , - (xi, + 0.
If p be the perpendicular from the origin on it,

2 (xi +yf
(1)
L 2 +f '
Also, 2 (2)
y2 = + 3,. . .
and f(x,y)=O ... (3)
If x and y be eliminated from (I). (2) and (3), the required pedal
equation is obtained.
(ii) Pedal equation deduced from polar equation.
Let us take the pole as the point with regard to which the pedal equation
is to be obtained, and let f ( r, 0 ) = 0 be the eqation of the curve.
Let p be the perpendicular from the pole on the tangent at (r,O ):
then
TANGENTAND NORMAL 457

f(r.O )= o; ... (1)

tan O=! i , ... (2)

dr
prr sin . ... (3)
If 0 and 0 be eliminated front (2) and (3), the required pedal
cquatioh is obtained.
Note!. When in any case nothing is mentioned about the given point
with regard to which the pedal equation is to be obtained, the given point is
to be taken as the origin in the Cartesian system and the pole in the Polar
system.
Note2. In some elementary cases, pedal equations can be easily obtained
from geometrical properties. I See Ex. 6of Art. 14.181

U18. illustrative Examples.

Ex. I. Obtain the value of sin 4,, ens 4,, tan ID and arc-differential in
polar co-ordinates by riaiisfor;nazion fi-oln Curie.s ion system.
Since .r = r CoS 0, v rsin 0,
(lx = cos 6 dr r sin OdO
and (Iv = sin 0 dr + CCOS 0 dO
1c 2 + 1v 2 = dr 2 + r 2 dO2,
i.e. d,52 = 1r 2 + r' j2 - -. (I)

Also, xdy— vdx

= ,2
do ... (2)
Again, since x2 + y ' = r, :. xdx+ vdv = rd, ... (3)

Now, s=0+4,, ..
cos4=costcos0+sinti sin 0
dxx dyy xdx + ydy Jr
ibv(3)I
ds r ds r rds d.c -
Again, sln4)=sinslicos0—cos'.lisinO
= 4y. x dx y xdy—yd.c rd0
by (2)]
ds r ds r rds ds
rdE) dr rd0
tan 4, = sin 4, + cos 4)
= ^ =
Ex. 2. Find the angle of intersection of the curves
r=sinO+cosO and r=2sinO.
 458 DIFFERENTIA L CA LCULUS

r cosQ+sjn0 1+ tan O (
Here, tan 1= = tan!7t+0
r cosO— sanO 1 —tan O
r 2sin0
tan t -, = - = - tan 0
- r 2cos0
it+0, 4t2_0.
angle of in = = It.
Ex. 3. Prove that the curves
r" = a" cos nO and r" = b" sin no
cut orthogoitalI
Taking logarithm of the P equation,
ii lug r = fl log a + Jog cos nO
Differentiating with respect to 0.
I dr nsin no
17 - - =-
r dO Cos 110
• cot 1 =—tan n0=coi(--n+nO)
Similarly, from the 2 n ' equation, we get
Cot 2 = cot no
02 =no
angle of intersection = =

Ex. 4. Find the pedal equation of the parabola y 2 = 4ax with regard to
its vertex,
Differentiating the given equation, Y.V J = 2a, :. = 2a/ v.
the tangent at (x,) , )is Y — y= (2a/y )(x —x).
i.e., 2aX —vY+2ax=O (. y 2 =4ax)
2- 4a 2 x 2 4a- 2X 2 -
p
402+y24a2+4axx+a .•. (I)
and r2=x2+y2=x2+4,,. ... (2)
From (1) and (2),
ax2_p2x_ap2 =0 ... (3)
and x2+4ax_r2=O
... (4)
By eliminating x between (3) and (4) the required relation between p
and r will be obtained.
 T,4NGENTAND NORMAL 459

By cross-multiplication
x I
p 2 r 2 +4a 2 p 2 - or 2 — op 2 - 402
( 22 (.-2 )2
P2 +4ap ,4u +p
is the required pedal equation.

Ex. 5. Find the pedal equation of r " = a' cos rnO

Taking logarithm of the given equation,
rn log r = ni log a+ log cos mO
Differentiating with respect to 9,
I dr m sin mO
- _________
I. cM cos mO
cot=—tanmno=cot(4it+mo),
4=--7t+MO.
Again, p=r sin i =r sin (+1r+,nO)= r cos mO

r -f-- from the equation of the curve.

r '" =a p is the required pedal equation.

Ex. 6. Find geometrically the pedal equationof an ellipse with respect to

ajbcus.

Fig 14.18.1
SN, S N are drawn perpendiculars on the tangent at any point P on
the ellipse. SP = r, SP= r', SN p, S?'I'=j, We know from Co-ordinate
Geometry that
r+r'=2a and pp'=b2,
Since LSPN = ZS'F'N', .. ' SPN, S'I-'N' are similar.

rr'r r,r 2a—r)

PP ' PP' b2
460 L)IFFERE VT/AL CALCULUS

r 2r(2a-r) I 2o
__ = ---------- or, = —
/12 r
,n h
which is the required p2dal equation.
i,21, the above
Noting that the semi-latus rectum / of the ellipse =

equation may be written as = - I

V U

Ex. 7. Find the geometrical meaning of -fe-, and hence deduce

d>i:
-, (dp')2
r =p+I---
d v'

We have p=rsin4.
Differentiating with respect to yt.

dj) d4> di
- = r cos 4k— + sin 4>
r
M dr sin 4> r elo
=r Cos 4--+cos4k.r--- fl.--
cAp di d'' COS 4> di
L

= r cos 4k d (0+o
'I"
rcos t ; 0+0

=PN (See fig- /3.16)

projection of the radius vector on the tangent.
From L\OPN. OP7 = ON 7 + PA ' 2 -

2 Idp
r =1'
(It!>

14.19 Miscellaneous Worked Out Examples

Ex. I. (i) At what point is the tangent to the parabola 1 rrx 2 parallel

to the straight line v=4x -5? .[ C. P /982

(ii) Find the points on the curve y = x the tangents at

which pass through the origin. / 1991, 3002 1

Solution y= (I)

= 2x = gradient of the tangent to the curse 0) at

iiv
 461
TANGENTANI) NORMAL

If the tangent is parallel to the straight line y=4x-5. then 2x=4,

i.e., x 2 and so y = x 2 = 4.
Hence the required point is (2, 4).
4x + 9 ... (1)
(ii) Here, equation of the curve is y = . t 2 -

Equation of the tangent at (, v) is

Y— y = 2(. — 2)(X — x) .
(x. Y) being current coordinates.

If this tangent passes through(he origin (0. 0), then

0—y = 2(x-2)(0--x)
2
i.e., y = 2X —4x
or, x2_4x+9=2x24x . EFrom(l)1

or, x=9, i.e., x=±3

when x=3, y=6 and when x=-3, y=30
Hence the required points are (3, 6) and (-3, 30).

Ex. 2. Find the slope of the r ye . + = a 3 at the point (x,, y),

and hence obtain the equations of the tangent and normal at the point.
Also deduce that the portion of the tangent at (x 1 , YO intercepted
between the axes is of constant length.
[C.?. 1983, 91..B.P.1995]

Equation of the curve is x 3 =a ... (1)

Solution: + y3

= _2I = slope of the curve at the point (x 1 , vi).

(!V -

Equation of the tangent at (X, ,Y,) is

Y-yl=—ZL7(x —x1)
xJ

... (2)
or.

is the equation of the tangent.

 462 011-FERENT/,4L CALCULUS

Similarly, equation of the normal at (x1 .'y i) is

Y--y 1 =—'y-(X-x)

or, x(X-x1)=y1(y_y)

Equation (2) can be written in the form

X y
—+---=l.
X , i- , Y, 3a
-

so that the intercepts on the axes are X j 30 and v a respectively.

Hence the length of the tangent intercepted between the axes
X, _ )2 ", ) 21
=

jt7 2
x+'ii)]={[a.aJ} =a, a Constant.

Ex. 3. If p=xcosa+ysina touches the curve (J"' +(2-)"

(hen prove that = (acosa + (bsina)" .

p" [ C. P 2(40/ 1
Solution Equation oT the curve is
•7

(a) b)
(I)

I (x'i I (y dv
or, . II -•l-i -=0
a (a) b tb) dx

i.e.,

Equation of the tangent of the curve (I) at any point (x, v) is

Y-

(x, Y) being current coordinates.

TANGENT AND NORMAL - 463

or, .X+a()L)Y=bx()+ay(1)
bV

(y'
ab +
J( X ) 1a
kb

or. bVJ"1X+a(Z)1Y=ab ... (2)

But Xcosa+Ysinu= p . (3)

touches the given curve. So, equations (2) and (3) should he
identical.
COSC( - sin u. p
—Lab
h(i )"' a(Y')

('t \
acos€x=p•-- bsina=p- ---
ta)

Hence ( acosa) " + (bsina =

' ( Y) }

=p" [From (1)1

Ex. 4. (i) Find the length of the cartesian snb(angent of the curve

F C. P 19831
(ii) For the parabola y 2 = 4aV show that the subtangent is
bisected at the vertex and that the subnormal is constant. [ B. P. 1993
Solution i) Equation of the curve is y =e 2

x -
dy 1----
Y1 —e 2
dx 2

Length of Cartesian subtangent -Y_ =e =1-21=2 units.

464 fl11FERLN7/4L CA LCULUS

(ii) Equation of the parabola is = 4a. (I)

dv dy 2a
2y =4a . i.e.,
y
Equation of the tangent at any point P(x, ) , )on the parabola is
2a
'TY =—(x -x)
y
or, 2a X - y Y = _).2
t- 2ax = -2a A

If this tangent meets the x - a is at T(a, 0), then

2aa-y0=-2a. Y
i.e., a=
-x .
So, coordinates of T(-x..0)
and coordinates of S (x, 0)<
Evidently, the subtangcnt ST is
SN x
bisected at the origin 0,
since j0SI=I0T=L
Also, length of cartesian
Subnormal
dyl 2a1
= y —l=I y — I=2", (constant).
"I I YI
Ex. 5. (i) Find the length of the polar subtangent for the curve
r=a(1+cos0)at 0 —. 7E [C. P. 1990]
2
(ii) Find the length of perpendicular drawn from the pole upon
the tangent to the cardioide r = a (I + cosO) at the point whose vectorial
It
angle is -. [ C. P 1992 1
3
Solution (i) Here, r=a(l+ cos t)
dr
--=-asin0
dO
11 2j0
Length of polar subtangent = i r - I
I dr

a2(1+coso)2 I
-
—a sin
I=a, when
 TA NGENTAND NORMAL 465

(it) Here, r=a(1+ cos O), =-a sin O

do
dO a(1+cosO)0 (it O'\ it 0
-cot-=tanj--4-- I ..
.dr -asinO 2 2 2). —2+-2-
2
Now, length of perpendicular from pole upon the tangent at any
point (r, 0) is,
(it O 0
p=rsIn=rsInI —+— l=rcos-
k.2 2) 2

—it it
At the point where, 0= , r = a(I +cos-)- 3-a
3 32
I it P 3 ,/ 3,Ja
and cos-0=cos-and hence =-ax-
2 6 2 2 2 4.
Ex. 6. (i) Find the pedal equation of the following curves

(a) r = e0. [ C. P 1983, 92, 2000, '02 1

(b) 1=a(j-cos0). [ C. P 19881
It
(c) r=5•e 7.
IC.P. 1989]

(ii) Show that the pedal eqation of the ellipse +-=l with
a 2 b2
b2 2a
respect to a focus is ----1. [C. P. 1985, B. P. 1993]
P r

Solution (i) (a) r = e 0... (1)

or, logr=Ologe=0

ldr it It
or,cot=l=cot-, ...•=—

=rsins=rsin!=_J

2p2 = 2 is the required pedal equation.

(b) r=a(1_CoS8)=2asin2 ... (I)

logr = log2a +2logsin

30-
466 DIFFERENTIA L CA LCULUS

1 9
2OS2
I dr 2
Or, - —c .. cotocot, i, •-.
=
rdO 2 2 2
sin
.2
9
Now. p=rsinrSin - -

2 2
or, 11 =rsi Ir — r [From (l)
2 2a
or, 2ap 2 = r 3 is the required pedal equation.
ft
(c) r=5e 7 ... ( 1)

or, logrlog5+Ocot!-,
7 rdO 7 7
IE
or, Cot 4i = cot—, i.e., • = .

.5jfl!
Now, p=rsin4i=
TE
Hence, p = r-sin- - is the required pedal equation.
7

I- y
(ii) The polar equation of the ellipse , + -- =

with focus as pole is given by I +€COsO (I)

a2-2
where. 1 = semi-talus rectum = - and e =eccentricity = 2
a T '
From (1) on differentiation
I dr . I (dr\ 2 e 22
or ,.4
—'I--I —sin 0
r 2 dO I ) i

e 2 sin 2 8 I (dr'1 21
or, 4) I Vide Art. 14'16
12

or. e 2 n 2 O1 2 ( t rp ) ... (2)

Also, from (I)

2 CO5 (l-r'\
U j ... (3)
r)
TANGENT AND NORMAL 467

Adding (2) and (3)

12(r2_p2)
(l-r)2-
+
p2 r2
j2 ,2 12
21
or, P 2 r2 I

12 21 1 ,2
b
or, ----+—'O 1. 1e
p2 r a a2 a

2a a! b2
or, --1--
r .p p
• b2 2a
Hence, the required pedal equation is = - - I.
p. r

EXAMPLES- XIV (B)

ds
1. Find for the following curves:
(I) r=a(l+cosO). (ii) r=ae0•
(iii) r 2 = a 2 cos 20. (iv) r" = a" cos nO.

2. Find for the curves:

dr
(I) r = a 0; (ii) r = a/9
3. Show that in the equiangular spiral r = a eOcO, the tangent is inclined
at a constant angle to the radius vector. I C. P 20061
4. Showthatfor log r = aO + b, p r.
5. Find 4i in terms of 0 for the following curves:
(i) Cardioide r=a(1— Cos 9).
(ii) Parabola r = 2a/( I - cos 0
(iii) Hyperbola r 2 cos 20 = a2.
(iv) Lemniscate r 2 = a 2 cos 20.
6. Find the angle of intersection of the following curves:
(i) r=asin2O. r=acos20.
(ii) r=6cos0, r=2(1+cos9)
(iii) r = 16 sin 20, r 2 sin 20
2
4.
 468 DIFFgRENTIAL CALCULUS

7. Show that the following curves cut orthogonally:

(i) r=a(1+ cos o),r=b(l— cos o)

(ii)ç= al( 1+ Cos o),r=b/(1— cos o)
8. Show that the curves
r" = a sec (no+a),r'rrh' sec (no+p)
i1tersect at an angle which is independent of a and b.
9. Prove that

tan Xyj
(dy 'I(
/ dy
dX
where $ is the angle which the tangent to a curve makes with the
radius vector drawn from the origin. C. P 1931 1006 I
[Use Ø=qi-6, zanO=y/x]
10. Show that for the curve rO = a. the polar subtangent is constant and
for the curve , = au • the polar subnormal is constant.

11. Show that for the curve r = e 0 ,the polar subtangent is equal to the
polar subnormal. [C. P 2007.1
12. Find the polar subtangeiir of
(i) r =,aea (j) r = a ( I — cosO ).
(iii)r=2a/(1_ cos 9). (iv)r=l/(li-e cos 8).
13. Show that the locus of the extremity of the polar subtangent of the
curve u+J(0)=oisu=f(-,r+o)
14. Prove that the locus of the extremity of the polar subnormal of the
curve r = i . (e ) is r = j'(e — r ).
Hence deduce that the locus of the extremity of the polar subnormal of
the equiangular spiral a e9 "" is another equiangular spiral.
15. Show that the pedal equation of the ellipse

x 2 /a 2 + ),2 /b2 = I
with regard to the centre is a 2b 2 /p 2 = a 2 + b 2 - r2
[CR 1988, '93. 2007, B.P. 5, VP '951
16. (i) Show that the pedal equation of the astroid x + y = a is
r2+3p2=a2. [ C.P.20061
.
(ii) Show that the pedal equation of the parabola > 2 = 4a (x + a) is
p 2 =ar. . . [ C.P.1931. '93, '97, V.P.2000
JANGENTAND NORMAL - 469

17. Show geometrically that the pedal equation of a circle with regard to
a point on the circumference is pd = r2 , where d is the diameter of
the circle.
18. Show that the pedal equation of
(i) the cardioide r = a(l+ Cos 9 ) is r = 2ap2.
[B.? 1997, V P. '991
(i) the parabola r = 2a/(l-. Cos O ) is p 2 = ar.
[8.? 1992. '94, V P 20021
(iii) the hyperbola r 2 COS 20 = a2 i s pr = a2.
(iv) the lemniscate r2 = a 2 cos 20 is r3 = a2p.
F C.?. 1998, 2001, 20081
(v) the equiangular spiral r = a e000 is p = r sin a.
(vi) the class of curves r" = a" sin no is r'* ' = a" p.
(vii)the reciprocal spiral rO = a is p2 (a2 + r2
)= a2r2.
[C. P 19381
ANSWERS

1. (I) 2a cos 8, (ii) rcoseca. (iii) a2 1r. (iv) asec no.

2. (i) 1r2 +a 2 /a. (ii) _Jr2+a2/r.

5. (I) O. (ii) ir- 2 O, (iii) .. ,r-20 (iv) ',t+2O,

6 (i) tan - . (ii) ir (iii) r.

12.(i) r tan a. (ii) (2a Sin 3.8 (COS -O ).

(iii) 2a cosec 0. (iv) I / (e sin O ).'.

 CURVATURE
ItI1l
15.1. Definitions.
Let P be a given point on a curve, and Q be a point on the curve
near P. Let the arc AP measured from some fixed point Aon the curve be s,
and the arc AQ be s + 1ss; then the arc PQ=s. Let TPL, MRQ be the
tangents to the curve at P and Q, and let mLPTM = I' and
mLRMX = V + then mZQRL = AV. Thus, At,ti is the change in
the inclination of the tangent line as the point of contact of the tangent
line describes the arc PQ (= Ar).

Fig 15.1.1

AV is called the average curvature of the arc PQ.

As
The curvature at P (denoted by X) is the limiting value, when it exists,
of the average curvature when Q -, 1' (from either side) along the curve,
i.e.,curvature at P -

x= Li
As ds
Thus, the curvature is the raze of change of direction of the curve with
respect to the arc, or roughly speaking, the curvature is the "rate at which
the curve curves
The reciprocal of the curvature at any point P is called the radius of
curvature at P. and is denoted by p. Thus,
ds
dip
CURVATURE 471

0 T X

Fig 15.1.2
If a length PC equal top is measured from I' along the positive direction
of the normal, the point C is called the centre of curvature at P. and the
circle with centre Cand radius CP (i.e., p) is called the circle of curvature
at?.
Any chord of this circle through the point of contact is called a chord
of curvature.

Note I. The line PG which makes an angle + I it with the positive direction
of the tangent (i.e.. the direction in which s increases) is called the positive
direction of the normal at P. To avoid ambiguities we make the
Convention that p is positive or negative according as C is on the
positive or negative side of the normal.

Note 2. The above formula is convenient only when the equation of the
curve is given in terms of s and y' (i.e., when the intrinsic equation of
the curve is given). So in the next article we shall obtain different
transformations of the above formula for the radius of curvature for
different forms of the above formula for the radius of curvature for
different forms of the equations of the curve, and henceforth whenever
we require the curvature of a curve we shall take th reciprocals of those
radii of curvature.

Note 3. Since the radius of curvature of a circle is equal to its radius

(See Lx. I. § 15.6), it follows that the radius of curvature at any point P
is the radius of a circle which has the same curvature at P as the curve has,
and this explains the nomenclature of the above circle. Since, the curve and
the circle of curvature at any point P (x, y) have the same tangent and the
same curvature, hence x. i i e. y have the same values at Pfor the circle of
curvature and the curve. [See A rt. 15.2
472 DIFFERENTIAL CALCULUS

15.2. Formulae for radius of Curvature.

(A) For the Cartesian equation y = 1(x).

We know tan
= W
differentiating with respect to x,
d2 v , dqi dy' ds
Vi dx
dx dx
dii [ dx
= sec y, a LCOS I'
ds

ds i /d 2r
p= — =sec '/-----
d/' ldx

2
Since I + (d
secw = (i + tan 2 1P ) =
dx
I

{ +() 2}i _____

d 2y Y2
2

where ', 0.
Note 1. Making the convention of attaching positive sign to (i + y 32 )
p is positive or negative according as y, is positive or negative.
Note 2. The above formula fails when at any point y becomes infinite,
i.e., when the tangent at the point is parallel to the y-axis (For illustration,
see Ex. 4.j5.6). In such cases the following formula, for the equation of
the curve as x = ( y), would be found useful.

dy
- = cot v', :. differentiating with respect to y,
dy
d2x di,ii -' dpds
= – cosec 2 = –cosec w I -—
ds o'

=- coscc i' — . I dx
I•.•
P L d.

P = - COSCC 3/'!
/
 u,rA1LRE 413

Since cosec 2 cli - I + cot 2 ip = I +

considering the magnitude only of the radius of curvature

3
dx22
d') }
p=. (1)
2
d x

where x, *0.
(B) For the Parametric equations x = #(t), y =

dv-
Here, (x' o).
dx di di x'
where dashes denote differentiations with respect to t.

• d2y d (Y"1
dx 2 - dxx') di'x) dx
d(y" di x'y' -

x'3

Then substituting the values of i

dvdx
, in the formula (I) above, we get

= (2)

where dashes denote differentiations with respect to t, and where

(C) For the Implicit equation c (x,y) o.

Here.4,,_i, (11. ;'o),..,j+j,S'-L=o

Differentiating düs with respect to x.

j x + , + + + ,. = 0
dx - th dx clx

or. f +2j1 I +i (i) =0

we assume here -f,.

 474 DIFFERENTIAL CALCULUS

whence, replacing for dy and simplifying,

.1, dx

+ jj2
d2y - 2fff,
= -
3
dx2 fy
2y
Substituting the values of 4V d in the formula (1) above, and
dx dr2
considering the magnitude of p only, we get

2)

(3)
- 2f, f f. +

^,
where - + ff2 o

(D) For the polar equation r = f (e).

ds ds dO dc /dt

Now, V/ = 0 + 0 = tJ -t tan -f-. where r dr

Pi dO

diji 1 I 12—rr,
dO r

2r 2 -rr2
2 2
-I + -_" 2 - r +
2' 2
r +i r +l

ds i 2
Again, = Jr + t ... (6)

[See Art. 14.131

from (4), (5) and (6), we get

.r2+2r12—rr2 (7)
CURVATURE 475

Car. For the Polar equation U = f(0), where u= yr.

UL,.
Since u = r1=--i- r2–
tL2'
Substituting these values of r, r 1 , r2 in the formula (7) above, we get

(u2+ui2)
...(7a)
u
(u*u2)'
where u 3 (u + u2 ) # 0.
(E) For the Pedal equation p = I (r).
We have p = rsinØ.
Differentiating with respect to r.

= sin + rcosØ-
dr
dO d41 dr dO d
=r—+r----r--1-r--
ds drds d ds
dij, I
j•.•o+=ig.]

dr
dp
(F) For the Tangential polar equation p = I (ii,).
When the tangential polar equation, i.e., the relation between p and W
of a curve is given,
dp dp ar ds dp d1 dr
_=_..-—.cosØ.p—••cosØ•r—
d' dr ds dVI dr dr dp

=rcosØ. -

p2 + (-?- 2 = r2 sin 0 + r2 cos 2 0 = r2.

Differentiating with respect top,

2p+2-4 -–" = 2r
4w diji 2 d1,(10
d2p
(9)
dip
 410 DIFFERENT/AL cALCULuS

Alternative Method

If p be the length of the perpendicular frm the origin on the tangent

at (x. y), viz., Y - y 1 X + sy1 y = 0,

xvi — y xtant/I-y
then p= =
l+ tan,

p = xsinI - ycosVi.
dp dx. dy.
• - = — Slfli/f — XCOS/I + — COy f. i'Sifl!/f.
diii Ill/i (/111

I.
= xcosiv + ysinlit

dx dx dc dv di, Is
S i n ce ----=--•---=pcosi----=.-.- •----=psiny
L dqi d.c dW dig ds dv'
2
Similarly, - 01 2p(IX (IV
= —cosip -.csln vi+-----slnhJf -cosl//
dip di/i dvi
= 9 COS8 l - S II1 I/f + /) S1fl I/f + V COS I/I
= p-(xsin, y COS ip )= -
Hence, the result follows.
15.3. A Theorem on curvature.
If a circle be drawn lonc/th:ç' a curve at P and cuuin' it at another
Point P 1 , then, as P, -, I', the (trek tends to the eircle o[cun'alure.

0
C P,

P.
Fig 15.3.1

Let C be the centre of curvature at P, and let y = f(s) be the equation

of the curve, where f' (x) and f' (4 exist.
Let ) be the centre of a circle touching the curve at I' and
cutting it again at P 1 .and let r be its radius; also let Q ( x, y ) be any point
on the are PP, of the curve.
CURVATURE 477

. 0Q2 =(—x)2+(-y)2 F(x).

Since OP2 = OP = r 2 , It follows that. F (x) has the same value

both at P and P 1 . Hence by Rolle's Theorem, there exists a point Q1 (x , Y )
between P and P 1 such that F'(-.) = 0,

x 0I — Y, ))=o.
dx

which is evidently the condition that 0Q1 is the normal to the curve at Q1.
Now let I - P then Q, also - P and hence by Art. 15.8, 0, the point of
intersection of the normals at Q and P. tends to C, the centre of curvature
and thus r also tends to CP, i.e., p.
Thus, the circle tends to the circle of curvature.

15.4. Curvature at the origin.

(i) Method of substitution.
Radius of curvature at the origin can be found by substituting x = 0,
y = 0 in the value of p obtained from Art. 15.2, or by directly substituting
the values of (1 )o and ( 2 ) in the formula.

'ii) Method of Expansion.

In some cases the above method fails, or becomes laborious. In such
cases, the values of (y1 ) u and (y2 ) can be easily obtained in the
following way by assuming the equation of the curve to be y I (x) ?d
writing for y in the given equation its expansion by Maclaurin's theorem,

.f' (o) + f (0) + .. . j (o) being zero here, since the curve passes

through the origin], i.e., px + qx2 /2! +..., where p, q stand for ('(a),

f (0), i.e., ( y 1 ) , (y 2 ) 0 , and then equating coefficients of like powers

of x in the identity obtained.
• This is illustrated in Example 9 of Art. 15.6.
cii) Newton's Formula
If the curve passes through the origin, and the axis of x is the tangent
at the origin, we have

x=O, y = 0, (y 1 p =

by Maclaurin's Theorem.
 478 DIFFERENTIAL CALCULUS

y = qx212!+...
Dividing by x212! and taking limits as x —+ 0, we get

ij (Yi'x q.
It should be noted here that as x —, 0, y also -4 0,

(t)
+2 i
But
utfromformulaofAn
15.2. at the p = -,
q q

p= L t-
-.o2y

Similarly, ifa curve parses through the origin, and the axis of p is the
tangent there, we have at the orin

p = Lt
s-.0

Geometrically:
Let the x-axis be the tangent at the origin.

g j5 .4 1
Draw a circle touching the curve atFi u, and passl'tg through a point P
(x, y) near 0 on the curve. Now, when p -, 0, along the curve, the limiting
position of the circle is the circle of the curvature.. (Art. 15.3)
Let OB be the diametclofthe circle, anddraw PN perpendicularto it,
and PM perpendicular to OX Let r be the radius of the circle.
 CURVATURE 479

Then, ON. NB=PN2,i.c..ON(OJ3-ON)=PN.

ON 2 Op-2

ON ON ON

0P .? + ).
2
x2
i.e., 2r = — = = — + Y.
PM y v
In the limit when P 0, x y -'0. r ---I P. and hence we
get as before
P ,
I x-
2 y
Similarly, when the y-axis is the tangent at the origin,

we obtain pr.Lj2._.
2 x
Analytically
The equation of the circle passing through the origin and having the x-
axis as the tangent at the origin is

x2+y2-2fr0.
If :- be the radius of the circle, then r =f
Since (3) passes through the point (x, y) oil curve,'

+ - 2fr = . whence j = (X +y

'2
pfJrr/Jf rL =Lj —
. 2y
General Case: -
If ax + by = 0 be the tangent a, the oratni, then proceeding as above.
we get
2
p2 X2

iy ) 1 J(a +b9

p1Ja2+b2. Li X+Y (4)

2 x-.oax+by
, -.0

Note. It should be noted that as x 0,y -s 0, y/. the rn' of

—> (- ),
the tangent line ax + by = 0. Here, it is supposed that 0. b * 0.
 480
DIFFERENT/AL CALCULUS

15.5. Chord of curvature through the origin (pole).

Let PQ be a chord passing through the origin

0 of the circle of curvature
at P on the given curve, and let C
be the centre of curvature and PT be the
tangent at P.

0 T. x
Fie 15.5.1
Join PC; produce it to D ;join DQ.
Then LPQD =a ii. L, being in a semi-circle.

LOFT = 0 and ZPTX

From tPQI) ,cbord PQ = PD cos DPQ

= 2p cos ( 1 7r -

= 2psin0
dr
=2.r!.E
dp

dr
2p —
dp
Note 1. From above it is clear that the chord
of curvature through the
origin can be easily obtained when the pedal equation
of the curve is given.
Note 2. If the chord PQ, instead of
passing through the origin, makes an
angle a with the tangent PT, i.e., LQPT =
a, then obviously LPDQ = a,
and hence PQ 2p sin a,
Hence, the chord of curvature parallel to the x-axis is 2p
sin
(: here LPDQ=)
CURVATURE 481

and the chord of curvature parallel to y-axis is 2p cos i

(... here LPDQ=7rr-W).

15.6. Illustrative Examples.

Ex. 1. Show that a circle is a curve of uniform curvature and its radius of
curvature at every point is constant, being equal to the radius of the
circle.

X
!I2 1)0.1

Let C be the centre of a circle of radius a. Let P be the Si' [it, Q

a point near it, and let PT. QM be tangents at P, Q and let L =

LQMX=s+ic;join CP,CQ.
LPCQ=ZPRM=Ai'.

• Al.r = anglePCQ = ., since /PCQ is measured ir.

1s As As a
radian.
as in Art. 15. 1,

curvature = IiLt - = - (constant) and hence p = a.

s-.OA s -.oa a

Ex. 2. Find the radius of curvature at the point (s, 'v) of the cun'
s=asectan+alog(sec+ tan V ).

ds i
Here, p==akSeCti.Scc l,+ta1PWsec14l

+ a- -----.--- -- . secjc ( sec + tanW)

secw + trnw
= asectc i + SeC2 sc - i asecw= 2asec
31•-
 482 DIFFERENTIAL CALCULU,,

Ex. 3. Find the radius of curvature at the point (x, y) on the curve
y= a1ogsec(4.

Hem, Yi = .SeC()9'a)tafl() tanc,

a'--
see
Y2 =(Vcsec2(c/4
Also, 1+ y = 1+ tad2 (çia)= se?(ç/c.

(i+12 {sec(x/a)}
= =asec(x/a)
Y2
(i/a)sec2(x/a)

E. 4. Find the radius of curvature of the parabola y 2 = 4x at the vertex

(0,0).
dy
The tangent at the vertex being the y-axis, at the vertex (0, 0) is
dx
infinite. Hence, formula (1) of Art. 15.2 being not applicable, let us apply
formula (la). [See Note 2, A rt. 15.2.
cix 1 d2x 1
Hence,d-V
-----2 y ; 2

at the vertex, x 1 = 0, x2 =

at the vertex, p =
1+x12
() i
= = 2.
2
Ex. 5. Find the radius of curvature at the point '0' on the cycloid
x=a(0+sInO), y=a(1_coso). [CP1944,V P.2000. '961

Here, x'=a(l+cosO), I=asin0,

x'=-asinO, y'=acosO,
by formula (2) of Art. 15.2,

{d2(1+co)2+ci2s1d20} _ a.&00
P d2co(i+co)+&siri2.9 - 2co0

= 4ucos 0.

Note. p can aio be obtained by using formula (I) of Art. 15.2 by first
obtaining the-values of y 1 and v, in terms of 0.
CURVATURE 483

Ex. 6. Find the radius of curvature at the point ( r 0 ) on the cardiolde

r= a( 1—cosO ), and show that it varies as .j. V P. 2002; C. P.007]
Here, r1=1riO. r2=acos0.
by applying formula (7) of Art. 15.2,

- {c?(1_cos0)2+.2s1ri20}
dF (1—cosO+Gsiii0—ccosO(1—cosO)

= a(2-2cos0) 22a(i 2a(

o) 2s1n2
--3 17 cosOT 3 3 l, 2

41
=- sin -0. (I)
3 2.
Since r=a(1—cos0)=a2sin210,

sin-!J(r/2a).
O=

Hence, from (1),p=i.. ... poc,/.

Note. In the cases where it is easier to transform a polar equation into a
pedal one, to find the radius of curvature, it is convenient to transform the
polar equation into the pedal form first, and then use formula (8) of Art. 15.2.

Ex. 7. Find the radius of cari'aare at the point ( p, r ) of the curve

rm =
rm+I dp (.+1)r-
We have p= -•--. .•. =
dn dr d
dr d
dp (rn+1)rm
- dr_ am am

EL 8. Find the radius of curvature at the origin for the curve

x 3 +y 3 -2x +by=O.
Here, y = 0, i.e., the x-xls is the tangent at the origin,

at the origin LL— = 21

484 DIFFERENTIA L CA LCULUS

Dividing the equation of the curve by y, we have

x2x2
x—+y 2 -2—+6=O.
Y y

Now, taking limits as X -4 0, and y - 4 0, we have

-2.+6=0, or,

Ex. 9. Find the radius of curvature at the origin of ihe conic

y-x=x2+2xv+y2 [C.P.19481

First Method: Differentiating the equation successively with respect to x,

Yi -1=2(x-i-xr 1 +v+yy1 ),
and )'2 =2(l+xy, +2r 1 +yy2 +y12 ).
at the origin, i.e.. when x = 0, y = 0, y = I and ), 2 = 8.

(I+%,,2 ) ( i+i) ,
atiheongin, p= - = =—=—=0.35 nearly.
Y2 8 8 4

Second Method:

Putting y px + + .......on both sides of the equation, we have

(, ) _l)x+-+ higher powers of x=(j+2p+p 2 )2

+ higher powers of x.
Equating coefficients of x and x' on both sides,
p-h.-O, i.e.,p=I,
and -q=l+ 2p +- :. q=8.

Since here p and q are the values of y, y2 at the origin, using the
+Y2 )'
formula p ,weget p at the origin.

Third Meth.d(Newtonian Method)

Since y - x = 0 is the tangent at the origin here, by the formula for the
Newtonian method at the origin,
Lj
p=.x 2+ y2
yX
=I.0 x+2xy+y
(from the equation of the curve)
 495
CURVATURE

i+(y/x)2

1+2(y/x)+(y/x)2-
(On dividing the numerator and denominator by 2)

=IJi__±
2 1+2+1

since Li (y/x), being the value of 'rn' of the tangent at the origin,
vi:.

y— x= 0, is equal to!.

Ex. 10. Show that the chord of curvature through the pole of the curve
I
2r
r cos:nO is
in + I
Taking logarithm of the given equation,
in Jog r = m log a + log cos mO.
Differentiating with respect to 0, we have
1 dr –ammO
=–tanm0.
rdO cosm0

i.e.. 0=427t+mO.

p = rsin Q = r cos mO = r.r"/a'" =

dp(m+1)r'
dr a"
dr p dr
chord of curvature = 2p sin 4=2r --2' P
dp r dp

a mrm1' 2r
=2._
(m+l)r m am m+l

Ex. 11. For any curve prove that

' 2
J d2x 1 +IL.
' I2J 2

_ d'4i I
d2x
dx ,inW sin tli... (I)
Wehave — =cosW . ... 'j
ds
 486 DIFFERENTIA L CA LCULUS

dw
and -1= sinqI. CO5141 COSI.— (2)
,is• d 2ds p
Now,squaring (1) and (2) and adding, the required relation follows.
EL 12. For any curve prove that

dO
___where tan 4,= r
Sin dr
dO)

d.
$ dU d
5104) 1+ -- +— .r—
S1fl 4)4- —— S1fl4)=r do
A) dO ds dO dr

(do d4) d9" (do d4)

=rl—+--.—l=rI__.f_
çd do dc) ,, ds dr

d dW
=r–_(O+i1)=r__ (. e+=,).

right side =

ds

EXAMPLES - XV(A)

I. Find the radius of curvature at any point ( s, ii ) on the following curves:

(I) s = a w. (ii) s=4asini.
(iii) S = c tan l/(. (iv) s=8cssin2 w.

(v) s = a (e"' - (vi) s = c log sec cv

(vu)5 s _i). (viii) s = a log tan(Jr +i)
2. Find the radius of curvature at any point (, ) for the curves (I) to
(viii), and at the points indicated for the curves (ix) to (xiv):
(I) y 2 = 4a.. (ii) e 1 ' = see (x/a).

(iii) y = logsin X . (iv) ay 2x 3.

(v) X),= C2. () y = + a .[eo + en].

CURVATURE 487

(vii) x 2 /a 2 + 2 1b 21. (Viii) x I + = a3.

(C.P 1943, B. P. '931

fix) y = —2? + 7x at the origin.

(x) y = 4sin x - sin 2x at x = .4 it

(xi) 9x2 + 4y 2 = 36x at(2,3). (xii) y =e at (0, 1)

x +
(xiii) .JF jy = F,, at the point where v = x cuts it.
[v.P 1988, '96 1
(xiv) y = xe at its maximum point. .
3. Find the radius of curvatute at any point of the curves (i) to (vi), an
the points indicated for the curves (vii) and (viii):

(i) X = acos 9, y = a sin G. (ii) x a, 2 , y = 2al.

(iii) x = a cos , y = b sin 0 .

(iv)x=aseCø, y=b tan ib
(v) x=a(cosr+:sinl), y=a(sini—i'cosf).
g g
(vi) x=a sin 2O(l+ co g 29). v=a co 2O(1— co 29).

(vij)x = acos 3 G, y= a sin 3 G at 0 = .,r. IB.P.19991

(viii)x=a(0+ sin 0). y=a(I— Cos 9)at 9=0.

4. Find the radius of curvature atany point (r, 0 ) for the curves (I) to (xi),
and at the points indicated for the curves (xii) to (xvi):
g _L O .
(i) r=aO (ii) r=a co 0.(iii) r=asec2

(iv) r = a (1—cog). [(P. I999.B.R '89, '91,981

(v) r2=a2 COS 2O. [C.P.19921

(vi) r = ae° a (vii) , = a 3 cog 3O.

g g
(viii) r = a +b co 0. (ix) r" = a co m0.

(x) r 2 cog 29=a2.[V.P.19981 (xi)ra see 2O.

g
(xii) r= 2a co O — a at 0=0.
(xiii) r = a sin nO at the origin.
g [C.P.2005,O6]
.(xiv)r=I/(1+e co 0)at 0=,r, [<i].
 88 DIFFERENTIAL CALCULUS

(Xv)r 2 =a 2 Cos 2O.aj 6=0.

(Xvi) r = a(O + sin O) at 0 = 0. [C..P 19891
S. Find the radius of curvature at any point (p, r) on the following
curves whose pedal equations are
(i) p = r sin a. (ii) r2 = 2ap. (iii) p 2 = at.
IC..P.1982)
(iv) pr = a 2 (v) r1 = 2ap 2 . (vi) r3 = a2p.
-. a 2 b 2
(vfl)--+r 2 a2 +b2 . [ C..P.1998j

6. Find the radius of curvature at any point on the curves:

(I) p = a( I + sin y ). ( ii) p = a cosec j,.
(iii) p 2 +a 2 Cos 2W =0.
7. Find the radius of curvature at the origin of the following curves:
(I) y = x 4 - 4x 3 - 18x 2 . (ii) 2x 2 - xy + y 2 - y = 0.
(iii) 3x 2 +4y 2 =2x. (iv) 3x2+xy+y2-4x=0.
(v) 3x4_2y4+5x2yf2xy_2y2+4x0.
(vi) 4x 4 +3 Y3 - 8x 2 y + 2x 2 - 3 -ty - 6y2 - 0.
(vii) x3 + y 3 = 3 axv. (viii) x 2 + 6 Y 2 + 2x - y = 0.
(ix)x 4 +y 2 =6a(x+y). [C.P.20061
(x) Sx2+y2+6x+8yr0.
(xi) y 2 =x2(a+x)/(a— x).
NO ax + by + a'x 2 + 2h'xy + b 'y 2 = 0
(xiii)y 2 - 2xy - 3x 2 - 4x 3 - x 2 y 2 = 0.
(xiv) y2-3.ry-4x2+5x3+xy—y=0.
8. Show that the chord of curvature through the pole for the curve
p = 1(r) is given by 2f (r)/ f'(r).
9. Find the chord of curvature through the pole of the curves:
(i) r = a(1 + cos0 ). (ii) r2 = a Cos 20
(iii) r 2 Cos 2O = a 2 (iv) r = ae0 a
(v) r' = asin nO.
CURVATURE 489

10. Show that the chord Of curvature parallel to the axis of y for the curve
(i) y = a log sec (x/a) is constant.
(ii) y = c cosh (x/c) is double of the ordinate.
11. Show that in a parabola the chords of curvature
(I) through the focus, and
(ii) parallel to the axis are each equal to four times the focal distance of
the point.

12. Show that for the ellipse x 2 /a 2 + y 2 /b 2 = I the radius of curvature

at an extremity of the major axis is equal to half the latus rectum.
IC.?. 1990. '94)

13. IfCbe the centre ofthe ellipse x2/a2 + y 2 /b 2 = 1 ,show that atany
point P.
CD3 a2b2
- --p,
ab p

where CD is the semi-diameter conjugate to C?, and p is the

perpendicular from the centre on the tangent at P.

A. If P 1 and P, be the radii of curvature at the ends P and D of

conjugate diameters of the ellipse x2/a2 + y2 /b 2 = 1 ,then

P, J + P2 3 = (a2 + b 2 )/(,,b)'3 IC.P 19881

15. Prove that the radius of curvature of the catenary y = a cosh (x/a) at
any point is equal in length to the portion of the normal intercepted
between the curve and the axis of x.

M. Show that for the cycloid x = a0 - sin 0). ,' = a(l - cos 0) the
radius of curvature at any point is twice the portion of the normal
intercepted between the curve and the axis of x. [C. P. 2004 1
17. Show that in a paraboh the radius of curvature is twice the part of the
normal intercepted between the curve and the directrix.

18. If p 1 and p 2 be the radii of curvature at the ends of a focal chord of

the parabola y 2 = 4ax, then show that

p1 T+ p; =(2a) .[ VP. '99,C..P. 1986,20061

490 DIFFERENTIAL CALCULUS

19. Show that in any curve

(2 (dy2
(j = I—I +1—
Idyr) di

1 d 2x /dy d 2y /dx
(ii)

I - d 2 x d 3 y d 2y d3x
(ui)3—j
ds

20. Show that in the curve for which

(i) y = a cos" Vi, p is in times the normal;

(ii) y = ae",, p is in times the tangent.

21. Show that

(I) for the cycloid for which s 2 = Say.

p = 4aJ{l - v/(2a)};

(ii) for the catenary for which - ,2 = c 2 + s 2 , p = y2/c.

22. Prove that in any curve

1 11(dr'2 (dr2
- + 1—I—
____ - d2
p r rlds) iI.-

dO (dO2 j.
' d2r
0i) p=r—r—I-
ds ds) 2

23. Show that the radius of curvature at any point of the equiangular spiral
subtends a right angle at the pole.

24. Show that at the points in which the curves r = aO and rO = a

intersect, their curvatures are in the ratio 3:1.
25. Show that when the angle between the tangent to a curve and the
radius vector of the point of contact has a maximum or minimum value,
P = r2/p.
 CURVATURE 491

26. Prove that in any curve

=3y!y22_y3(1+yI2)
ds
Y22

and show that at every point of a circle

3y1y,2 = ( +. 2)

ANSWERS

I In the following examples, generally the magnitudes of the radii of

curvature are given.)

1. (i) a. (ii) 4a cos '. (iii) c sec2 V. (iv) . a sin I V.

(v) ante". (vi) c tan V. (vii) 3 in see3 V tan V.

(viii) a sec V.

2. (i) 2(x+ a )/..j. (ii) a sec (x/a). (iii) coseex,

(IV) (4a+9x)h h /6a).(v) (2

(vi) y2/a. (vii)(b4X2 + a 4 y 2 )/a4b4

(viii) 3(a.y)1. (ix) (125h). (x) f/i.

(xi) 13 . (xi) . (xiii) u/.J. (Xiv) C.

3. (1) a. (ii) (,2 + ) L () (a sin 2 0 + b2 cos2 ) /-

b-(iv) (a 2 tan 2 0 +b 2 see2 ) /ab (v) al. (vi) 4acos3O -

(vii) I a. (viii) 4a.

4. ii (r2 + a2 )3 /(r 2 +
2.2 a.

(iii) 2a sec 3 (iv) . (v) a 2 / 3r.

(vi) r cosec a. (vii) a314r2

(a2 + 2abcosO + h2 ) a'" t

(viii) (ix) -.
- a2+3abcos0+2b2 n,+l r'"
492 DIFFERENTIAL CALCULUS

(x) r 3 /a 2 . (xi) r(4r2 _2 )/3a 3 . (xii) 4a.

4 na. (xiv) 1 .(xv) a (xvi) a.

5. () tosec a. (ii) a. (iii) 2

(iv) r 3 /a 2 . (v) 1 VT 2ar (vi) a2/3r.

(Vii) 22b2/p3.

6. (i) 0. (ii) 2a cosec 3 f. (iii) a/p

7. (I) Ji' (ii) -4. ( iii) . (iv) 2. (v)

(vi) 2. (vii) a, Ia (viii) .j-L
(ix) 6 . (x) 5. (xi) ±a
(a2 +b2 )
(xii) 2'. (xiii) sJTh;J. (xiv)
''
ab —2h'ah+b'a2
9. (1) -31 r . (ii) - r. (iii) 2r. (iv) 2r. (v) 2r/ (n + I

15.7. Centre of Curvature.

Let (, 5, ) be the co-ordi?atcs of the Centre of curvature C
corresponding to any point P (-, y) on the curve.
Since C ( ., ) lies on the normal at P. viz.,
(x — x)+(Y — y ) y1 = 0,
(—x)+(5,—y)y =0.
Again, since PC = p,i.e., PC2 =

Substituting —(5,— y ) y1 for ( — x ) from (I) in (2) we get

+ ...
)2 2 2 = (i
(5, — (i +yI )= p (3)
y22

I- \2
— 2 l+v12)
i.e., (y — y ) =
 CURVATURE 493

(4)

Again, from (1),

(5)

from (4) and(S), we get

I+v12
______ = ... (6)
Y2 Y2

o, Hence the equation of/lie circle of curvature is

)2 2
(x— )2 +(y-

Note 1. According to our convention, we take the positive sign only in

(4); for, if y 2 is positive, p is positive and hence - y is positive. Similarly
if Y2 is negative, p is negative and hence y - y is negative.

Note 2. Since the normal at P makes an angle ( 1 71+ 14 ) with the x-axis, it
follows from the definition of the centre of curvature that

I
n1+)
Cos (t+)
i.e., i=x-psinw, 5=y+pcosi/ ... (7)

= Yt =
Now, since tan i V = y 1 , sin i, and cos t1
2

Thus substituting the values of p , sin,!!, cOS,!! in terms of y j and

Y2 in (7), values (6) of x, can be obtained.

Note3. By writing the relation (1) as (i-x)x 1 +( .-v)=O, and using

, =( \/
the values p I + x " ) / x 2 , of (from Art. 15.2, Sec. A) we can similarly
obtain
x1(1^x12)
=x+1+xI2 ..
X2 x2

dx d2x
where x1=—,x2=----. -
dy2
This form is useful when v 1 becomes infinite.
 494 DIFFERENTIA L CA LCULUS

Note 4. The centre of curvature can also be obtained geometrical!)' as

follows:
Let C (i, ) be the centre of curvature corresponding to the point
P (x, y) on the curve.

Fig 15.7.1
Then PC=p
Let PT be the tangent at P, so that LPTX = 141
Draw PN, CL perpendiculars on OX , PM perpendicular on CL.
Then ZPCM=lt,
i=OL=ON— MP
=x— PCs1nPCM
=x-p sin W, . . .(1)
5=LC= LM + MC = PN + MC = y + PC cos PCM
=y+p Cos llI
Since tan ic=v 1 , sin i= and cosllI=
VII + 2
Now substituting the values of p. sin 141, cosi in (I) and (2), the
required values of x and y are obtained.
15.8. Property of the Centre ofcurvature.
The centre of curvature C for a point P on a curve is the limiting
position of the intersection of the normal to the curve at P with a
neighbouring normal at Q, as Q P along the curve.
 URVATURE 49,

0! X
Fig 15.8.1

Let P (x, y) be the given point and Q (x + A x, y + A y) be a point near

P on the curve y = f (x); let us suppose y y2 exist at P and y2 ;4 0.
The normal at P is
(Y — y)y1+(X — x)=o, ... (I)
or, (i — y)Øx)+(X — x)= 0, ... (2)
putting y1 = 0 ()
... the normalat Q is
1y )0(x + 44+(X —x-x)= 0 ... (3)
Suppose the normals at P. Q, ii.. (2) and (3) intersect at N (, r
and let (.i, j) be the point C to which N tends as Q -4 P.
Subtracting (2) from (3) and putting TI for Y , we have
- y) (x + A x) - 0 4} - A vØ (x + A x) - 4x = 0 .. 4)
Dividing by Ax, and making A x -+ 0 and noting that in that case
7 -9 y, we have
(—y)O'(x)—y10(x)—l=o,
y)y2_y12_l=O
Again, since (, ) is a point on (I),
(—y)y1 +—x)= 0
The value of (, ) obtained from (5) and (6) are identical with those
of the co-ordinates of the centre of curvature obtained in Art. 15.7.
Hence, (, ), i.e.: C is the centre of curvature.
496 DIFFERENTIAL CALCULUS

15.9. Evolute and Involute.

The locus of the centre of curvature of a given curve is called its
Evolute. -
If the evolute itself be regarded as the original curve, a curve of which
it is the evolute, is called an Involute.
Formul (6) and (8) of Art. 15.7 give the co-ordinates of any point
(, ) on the evolute, expersscd in terms of the co-ordinates of the
corresponding point (x, y) of the 'given curve; since y is a function of x,
these formulx give us the parametric equations of the eolaie in terms of
the parameter x.
Ordinary cartesian equation of the evolute is obtained by eliminating
x and y between the two expressions for i,Y and the equation of the
curve. [See Art. 15. It Ex. 2. )

15.10 Properties of the Evotute

The normal at any point to the given curve is the tangent of the
evolute at the corresponding point of the 'evolute.
Let (it, ) be the centre of curvature corresponding to the point (x, y)
on the curve. Then from Note 2. Art. 15.7.
xp sin Vi, y)'+COSt/.
dji . •dp
.— =l—pcosqi-----sinyi——
dx dx dx
ds dxdVi. dp . dp
= 1-- •-•--sin !JI- = - sln(-
dtjf as dx dx dx

dX dp
Thus, = — slfl/,- . . ( 1)

IFY ... (2)

—
Similarly, —cosljl dp
dx dx
dividing (2) by (I)
dy dx
= - cotij/ = - ,which is the 'in' of the normal at (x, v).
dx dy
'in' of the tangent to the evolute at ( , i) ='m' of the normal to the
given curve at [he corresponding point (x, )), and since both tangent to the
evolute and the normal to the curve pass through the same point (:,
they are identical. Hence the result.
CURVATURE 497

(II) Length olan arc of the Evolute

The tenth of an arc of the evolute of a curve is the difference between
the radii of curvature of the givesi curve, which are tangents to this arc of
the evolute at its extremities.

Fig 15.10.1
Let s be the length of the arc of the evolute measured from some fixed
point on it up to the centre of curvature (, ). Then from (I) and (2) above,
we have
(6^v )2
I di 2 + (d2
dx) dx dx)
2
Also we have F !. )^ 41
dx) . dx) 'dx

& up .

- p = = p + C.
C (a constant), i.e., i

Hence, s,- S2 = p 1 - p , ,where P j , P2 are the values of Pat the.

two points P 1 . P2 on the curve and are te values of of the
corresponding points Cl' C 2 0n the evolute.

Thus, the arc C 1 C2 of the evolute = P 1 C 1 - P, C,.

Hence, if a string is wrapped round the curve C I C 2, it is clear that

when the string is unwrapped, being kept tight all the time, the point on the
thread which was at P, will describe the curve' P, P1
32-
498 - DIFFERENTIAL CALCULUS

(ifi) Radius of curvature of the Evolute

Fig 15;l0.2
Let ' be the angle which the tangent at the point C (, 5') on the
evolute I corresponding to the point P(x, y) on the original curve] makes
with the x-axis, then is the angle which the normal at on the given curve
makes wih the x-axis.

'1" = 4- r + yi. .. = I also from (II) above = I.

dv dp
Let j5 be the radius of Curvature of the evolute at (, 5').

-- di ds dp diji_dp_ d(ds)_d2s
dji'dp dyI

15.11. illustrative Examples.

Ex. 1. Find the centre of curvature at any point ( x, y) on the parabola

Y 2 =4x.

Here, y,= Y2

(x+a)

If (. 5') be the centre of curvature, we have

=x+2(x+a)3x+2a.
 CURVATURE 499

!-t-y 1 2 2L(x+a)
y–y - j-
Y2

r— 2vfx 2
=2ijax– (..y2_4ax)=FX2
,_.
I.

Ex. 2. Find the evoluze oft/ic parabola y2 4ax.

As proved above, its centre of curvature (, ) at any point (x, y) is

given byj3x+2a
2
y=- x' (2)

–2a
From (l),x=----_

- 2 (j-2a i
From (2) , Y-7-------)
squaring and writing r, y for ir, , the required evolute is giv

by 27ay2 =4x-2a)3

Ex.3. Find the equation of the circle of curvature at the point (3, 1)

the curve y =x 2 –6x+1O.

Hem, y j =2x-6;. y 2 =2.

at the point (3,I), y 1 = O, y2=2.
If (i, j) be the centre and p the radius of curvature at (3, 1),

y1(1+12)
- ____ 3
x' =3, y=y-f
Y2 Y2 -

Also, p
Y2

the equation of the required circle of curvature is

(x_3)2+(y7)2 =,

or
 500 DIFFERENTIA L CA LCULUS

15.12 Miscellaneous Worked Out Examples

Ex. 1. Show that the radius of curvature of the curve yc.cosh1

2
is--. EC.P.198095]

Solution y =c-cosh () (1)

dy
-- c—sinh- =sinh -
y j =-
Ac \c}

,12Y
and y 2 =COSh()4
I From (1)

3 1 21X '\)2
h
(i+v?)
Radius of curvature p = ________- =
Y2

3
= c2{cOsh2(i)}2
Icos h 2 x -- sin h 2 x I)

X )1 3
2 {cos ( 2Y3 - 2-

Y c
YC

Ex. 2. (i) For the tractrix s=c log sec J prove that p=ctanp and in

case of the equiangular spiral s = a(e'"'' - i). p = in a

B. P. 1995
(ii) For the curve s=asec1 flog(secw+tan4l), show that
p = secs(1+atanW). IC.?. 1995]

(iii) Find the radius of curvature of s=aseci+log(secU/ tanW)at

any position W. [ C. P. 2001

Solution : (i) s = c logsecw

dssecw tan 14l

=c tan 141
dW see
CURVATURE 501

Again, ' s=a(e"' .i)

ds
p=—=mae "
thji

(ii) s=aecqc+log(secw+tanhls)

ds secwtanw+sectI
p = — = asecw tanqI+
dW see 1+ tan I
sec(secW + tan I)
=a = see (1+atanW).
secw+ tan 1I

(iii) s=a see s+ log (see . tan i)

= a sec iji + log sec W + log tan w

ds see tptan W sec2l4s

p=—=asecwtanw+ +
dW see taint!
= a Sec l tan jI + tan 4I + Cot W.

Ex. 3. Find the radius of curvature for the curve

x=a(0+sin0), ya(1–coso) at 0=0.
C. P 1987, 91, 97, 2000, 2008

Solution: x =a(0+sinO), x'=a(l-i-cosO), x" = –asin0

y=a(1–cos0), y'=asinO, y"=acosO

At 0=0, x=2a; x"=O, y'=O. y"=a

(x2 +2) (2 +o) 8a3

At 0=0,
X3")?'X" 2a-a-0 2

Ex. 4. Find the radius of curvature at any point 'f on the curve

x=a (cost +t sin t), y=a (sin f–tcosz). [C. P 2002

Solution: Here, x=a (cost +t sin:)

x'=aicost, x"=a (cost –tsini)
and y=a(sine'–Icost)
y'=atsint, v"=a(sin:+tcost)
502 DIFFERENT/A L CA LCULUS

Now,

= (x+ {a2t2'(sin2t + cos 2 r)}

xy yx = a2t{stntcosf+tcos t —sIntcost+f sin 2 -

a313
at

Ex. 5. (I) Find the least value of the radius of Curvature of the curve
x=51, y=Slog sect . [C. P. 19831

(ii) Find the radius of curvature of the parabola y 2 = 16x at the

end of its latus rectum. F B. P J988, 1997

Solution : (i) Here x = 5s and y = 5log sect.

x'=5, x"=O, y'=5tan:,y"=5sec2t

- (x12 +y'2) 125(l

- + ta n 2 t)1 2 - l25sec3t
- 25sec 2 r-0 - 25sec2t
oip=5sect
Since numerically the least value of sec t is 1, least value of P is 5.

(ii) Equation of the parabola is y 2 16x.

One end of latus rectum is at (4, 8).

Now, 2y16
dx
• dy 8
i.e., —=y1 =-
dx y
d2 8 dy 8(8)64
Y2

At the point (4, 8), Y 1 Y2 =

CURVATURE 503

EXAMPLES - XV (B)
1. Find the centres of curvature of the following curves at the points
indicated:
(1) xy =12 at(3,4). [C. P 1934, 2008]
(ii) y=x3+2x2+x+lat(O,l).
(iii) xy=x2+4at(2.4).
(iv) y =sin 2 x at(O,O).
(v) x=e' cos 2Z, ye 2' sin 2t att=O.

2. Determine the centres of curvature of the following curves at any point

(x,y):
(i) x 2 = 4ay. (ii) a2 y = x 3 . (iii) x 2 /a2 + y 2 /b2 = I.
(iv) xv=a 2 . (v) x i +y 3 =a•
NO y = i2 a(e4" +e4' ).

(vii) x = a cos , y = b sin 0. (viii) x = a!2 , y = 2a/.

(ix) x=a(O–
sin 9), y=a(I– cos o).
(x) x= a( cos t+: sin l ), y = a(sint – lcosl ).
3. Find the evolutes of the curves (iii), (iv), (v), (ix). (x) ofEx. 2, abOve.

4. If (a, /3 ) be the co-ordinates of the centre of curvature of the parabola

x+
vF vFy = ja, at (x,y). then show that
a+f3=3(x+y).

5. Show that the co-ordinates (, ) of the centre of curvature at any

point (x, y) on a curve are given by
- dy - dx
(i) x=x-- y =v +— .
dtjf di,ií -
(ii) =r.+p2x, y=y+p2y,
where dashes denote differentiations with respect to the arc s.
6. Prove that the distance r1 between the pole and the centre of curvature
corresponding to any point on the curve r = f (o ) is given by
r1 2 =r2+p2-2pp,
where P and p have the usual significance
 504 DIFFERENTIAL CALCULUS

7 For the equiangular spiral r = ae° prove that the centre of

curvature is at the point where the perpendicular to the radius vector
through the pole intersects the normal.
8. Find the circle of curvature of the curves
(1) y = x+ 41x at(2,4). [C. P 2006]
(ii) y=x3+2x2+x+lal(0,l).
(iii) x = at the point where it crosses the x-axis.
(iv) y2 = 4x at at the ends of the latus rectum.
(v) x + y = ax + by2 + cx3 at the origin.

ANSWERS

1. (i) 71)
-g' (ii) (iii) (2,5).
(- 1, •).
' (v) (0,-i).
X (l 4 ),( 3
.? a 29x5
2. (i) 2a±-_1
4aJ (ü)

(a2 -b 2
3 a 2 -b2 y3 3 x3
( iv) 23
a4b 2a2
v3)( -
J2 - 02
( 1 2 2
(v) X3XY' Y + 3x Y J (vi)
( -
a
,2y I
{a2 a- b2
COS ,
b2 — a 23 1
(vu) b
(viii) (a (2+12 ) - 2a13}.

(ix) {a(O+ sin O),—a(l— cos 0)}. (x) {a cos :, a sin t}.

3. (i) (ax) +(by) = ( 2 _b2).

(ii) (x+ y)i — (40L

(x+y)4+(- y) = 2a1.
(iv) x a(O +sinO ), y = -a(1-cos8). (v) x2 +y 2 =a2.
4. (i) x 2 +y 2 -4a-I0y+28=0. (iI)x2+y2+x-3y+20.

x2 +y2 -6x+4y+5=0. ( iV)x 2 +y 2 -IOx±4y-3=0.

(v) (a+b)(x2+y2)=2(x+y).
II ASYMPTOTES

16.1. In some cases a curve may have a branch or branches extending

beyond the finite region. In this case if P be a point on such a branch of the
curve, having co-ordinates (x, y), and if P moves along the curve so that at
least one of x and v tends to + or to - , then P is said to tend to infinity,
and this we denote by P
y
Definition. If P be a point on a branch of a curve extending be ond the
finite region, and a straight line exists at a finite distance from the origin
from which the distance of P gradual!)' diminishes and ultimately fends
to zero as p _ (moving along the curve), then such a straight line is
called on asymptote of the curve.

16.2. Asymptátes not parallel toy-axis.

If y = nix + c be an asymptote corresponding to an infinite branch of
a cun'e, where ni and care -both finite (including zero), then

nz=L1 l and c=Lt(y-mx)

where (x, y) are the co-ordinates of a point P on the branch of the curve.
The distance of the point P from the straight line y - mx - c = 0 is
given by
d y - mx - c ,
and if y = mx + c be an asymptote,
i+
d -* 0 as x -+ and since in is finite here,
LI (y-nu-c)=0,or, Li (y-mx)=c.
if.-.-

Again, denoting y - mx - C, i.e., d fl + m2 by u,

y c+u
- In = -
x x
Now making -v - =' , since u -4 o in this case, and c is finite.

bL-m=0,Or, IJZ=m.
x )
Accordingly, to find asymptotes (which are not parallel to the
y-axis) of a curve y = f(x){or F(x,y )= o}, we first of all find out

Li 2 from the equation to the curve, which may have several finite values
506 DIFFERENTiAL CALCULUS

(inclusive of zero). Corresponding to any such value (m say), we next proceed

to find. Lz( y - mx), using the equation to the curve.

If this limit is.found to be finite, say c, then y = nix + c is an asymptote.

[ See Ex. 7,*16.81
Note. An alternative definition of a rectilinear asymptote is sometimes given
as follows:1! P be a point on a branch of a curve extending to infinity and
if a straight line at a finite distance from the origin exists towards which
the tangent line to the curve at P approaches as a limit when p then
the straight line is an asymptote of the curve.
With this definition also, we can prove the results of the above article;
for, the equation of the tangent at P (x, y) to the curve is

Y -y =(x _.r),or, Y =X +Iy -x

dx ds dx
and as x - ,if this tends to Y = ,nX +c, where m and care finite, clearly
1 dy '\
- and c= Lt I y -x --- 1= Li (y-mx).
m= Li dy
dx . dx )
It should be noted that when P-.-+oo, if the tangent line tends to a
straight line as its limiting position, that line is an asymptote. The converse,
however, is not true, i.e., even if the tangent line has no definite limiting
position when p _ ,there may be an asymptote. (See Ex. 8, § 16.81
163. Asymptotes parallel toy-axis.
The necessary and sufficient condition that the straight lihe x = a is
an asymptote to the curve y = f (x) is that j f (x) I -
when either
x -9 a + 0 or x - a - 0.
For, suppose x - a - 0. Since I yI -* in this case. P being the
point (x, y) on the curve, p in this case [con ve,el?, if p - in this
case, , and hence the necessit y of the condition]. Now the
perpendicular distance of P front line x = a is .v - a (the axes being
rectangular), and I x - a 1-3 0 as x -i a -0. hence, x = a is an asymptote.
Similarly, for the case when x -* a + 0.
Thus to find asymptotes parallel to y-axis, we may write z for 1/v . in
the equation to the curve, and then make z - 0. If then the result leads to
A rigorous proof of this last equality requires the use of integration.
ASYMPTOTES 507

a finite value or values of x of the type x = a, these will give us the

corresponding asymptotes parallel to y-axis. [Sec Ex. 7, § 16.8]
Note. In a similar way we may get asymptotes parallel to x-axis thus if as
y— )b±O, JxI_i oo , (where x , y isa point on the curve) then y=bisan
asymptote.
16.4. Asymptotes of algebraic curves.
The most useful case of determination of asymptotes is for algebraic
curves. The general form of the equation of an algebraic curve of the nth
degree is, arranging in groups of homogeneous terms,
+ a1x -I y + a 2 x - 2 y 2 + ............+ ay )

+ (box" +x 2 y + ............+ h,,_1y )

+ ........=0, . . (1)
which can ao be written as
(y't -2 (y'
- I—I
V) ,,-i ,.-iI—I
X) -2Il ......=0 , ... (2)
X)
where 0, is an algebraic polynomial of degree r.
For asymptotes of this curve, we proceed to prove the following rules:
Rule 1. Asymptotes not parallel toy-axis will al/be given by y = mx + c,
where
in
any of the real finite roots of O,, (m) = 0 and for each such
values of in, c = .- (m)/ (m) provided it gives a definite value of c.

Proof:
The equation (2) of the curve can be put in the form
............................=0

Now if y = mx + c be an asymptote, where in c are finite,

Li (y/x) = in § 16.2). Hence from (3), making x -+ o, since in
finite, and the functions 0,, (m), 0._t (m), etc. which are algebraic
polynomials in in accordingly finite, we get 0,, (m)= o.
- Again, since in this case 11 (y - mx) = c (See § 16.2) we can write
y— nix = c .i-u,where u is a function of x such that u — *O when
508 DIFFERENT/AL CALCULUS

+u
x —*o".Thus, -y =m+ c —
X x
From (3), now we get

c+u i I c+u
0. =.(I"^- I+—Øtjm+--------
X)X

m+- •• =
X )--
Expanding each term by Taylor's theorem, since the functions 0, are
all algebraic polynomials and will each lead to a finite series, and remembering
that 0,, (m) = 0, we get

Ic+u - (c+)
x3! }

c+u (C l-u)2
,
(m)+ O1(m)+ Ø 1 (in)+... }
X 1
x x

C+u
(m)+ --ø,..2(m)+ ... }
+ - {_
+..... . =0. ... (4)
Now multiplying throughout by x and making , x —, . we get
(vu—Onow).
co,', (m) + (m) = U ,or, c = - Ø,, (in).

Each finite root of 0, (m) = 0 will give oic value of c (provided

(m) # 0 for this value), and accordingly x vc get the corresponding
asymptote y = tax + C.

Special cases
If any value of m satisfying (m) = 0 (say in = nil ) makes 0 (m) = 0
also (which requires in, to be a multiple root (m) = 0 of as we know from the
theory of equations), and if 0 ,,_ 1 (in) * 0 for this value, then c - as
in - m - Accordingly there is no asymptote corresponding to this value of rn
ASYMPTOTES 509

If for ni = m1 ,we get si,, (!fl), ,, (rn) ,,-i (. ) ,each 'O, then from (4),
multiplying throughout by x 7 , and making x -* cc, we derive
4c 2 (rn)+ ctt',,_ j (m)+ -2 (in) = 0

giving two values (say c 1 . c, ) of c in general ( provided ,(1' 1 ) * 0] , and

thereby giving two parallel asymptotes of the type 3' =
Y = ,n 2 x +c2.
If ø, (ini) is also zero (i.e., if in, is a triple root of 0,, (in) 0, and if
(n ,, are also identically zero, we shall, proceeding in a similar
manner, get three parallel asymptotes in general corresponding to in =, rn;
and so on. -

Rule U. Asymptotes parallel to ;'-axis exist only when a,, (the coefficient
of the highest power ofy. i.e., of y") is zero, and in this case the coefficient
of the highest available power of y in the equation (provided it involves
x, and is not merely a constant) equaled to zero will give us those
asymptotes.
A similar rule will apply to asymptotes parallel to x-axis.

Proof: After dividing by y", and replacing 1/3' by z, the equation (l) of
the curve can be written in ascending powers of z in tile form

a,,+ z(ax + h+ z 2 (a,,_2x2 +b-2X + c,,)+... = 0 . . .(5)

This will have an asymptote parallel to -axis of the types = a where a

is finite, provided z - 0 when x —*a+0 or a-0.[ See Ex. /6.3]
Hence making Z - 0, since x now tends to a finite value, we must
have the necessary condition a,, = 0.
Assuming this to be satisfied, we get from (5), dividing by: and making
z —*0,
a,,x + b,,_ 1 = 0 ... (6)
giving a finite value of x(provided u,, isnotzero) which makes I
31 -4
and thus represents an asymptote. -
Incase o,, is also zero along with a, in order that we may have an
asymptote parallel to v-axis, since xis to be finite, we must have, from (6),
= 0. Hence, from (5), dividing by :2 and making = - 0 now we get
the asymptotes parallel to v-axis (two in this case) given by

a,,_ 2 x 2 +b . ,x+c,,, =0
510 DIFFERENTIA L CA LCULUS

provided this gives finite values of x. In case a,,, b,,_ 2 , c,,_ 2 are all
Z3
identically zero, we proceed in similar manner with the coefficient of in
(5), ie., the coefficient of j" in the original equation (I), and so on,
proving the rule.
Note. By interchanging y and x in arranging the given equation (1), and
proceeding in a similar manner, (making l/x -* o) we can prove the
corresponding rule for finding the asymptotes parallel to the x-axis.

16.5. Working rule for asymptotes of algebraic curves.

For an algebraic curve of the nth degree with equation given by (I) of
the previous article, first of all see if the term involving v" is absent, in
which case, the coefficient of the highest power of y involved in the
equation (unless it is merely a constant independent of x) equated to zero
will give asymptotes parallel to the y-axis.
Similarly, if the term involving x" is absent, the coefficient of the
highest available power of x equated to zero will in general give asymptotes
parallel to the x-axis.
Next, replacing x by I and ;' by m in the homogeneous nth degree
terms, get 0 ,, (in) [as is apparent from the alternative form ,-"" (y/x) 1
Similarly, get 0_1 from the (n - 1)' degree terms, and if necessary (see
later), ,,-2 (rn) from the (n - 2)1 degree terms, and so on. Now equating
$,, (m) to zero, obtain the real finite roots rn, m1, etc. which will indicate the
directions ofthe corresponding asymptotes (repeated roots giving in general
a set of parallel asymptotes).
For each non-repeated root (m 1 , say), a definite value C 1 of
c = -0,,1 (m/(m)

is obtained, and the corresponding asymptote y = ny ± c, is determined.

For rep eated roots the several values of c may be obtained as explained
under the head Special cases' of Rule 1.

16.6. Alternative method of finding asymptotes 0 algebraic curves.

Let the equation to an algebraic curve be
(I)

where I, [ax +ci1 x''y + ....... ±a,,i' - x"ç, (y/x)j consists of

homogeneous terms of degree n, P , is homogeneous of degree n - I and
 ASYMPTOTES
511

so on. Now m,, m 2 , m........ being the roots of 0,, (m) = 0 , we know fron-
the theory,pf equations that in m1, in in . .....are factors o
0,, (m) and accordingly
P. a,, (y - m1 x)(y - m 2x)... The possibb
asymptotes are parallel to y - m 1 x = 0, y - m 2 x
= 0, etc., as proved in
16.4, and their directions are thus all easily found from the factors of F',,.
CASE L Let y - m 1 X be a non-repented factor of P, 's, Equation (I) cat
then be written as
=0

where Q,,_, [= (y - m 2 x n3
)(y—r I ).....] is a homogeneous expression ol
degree ii - I which does not contain y - m 1 x as a factor, and
F,_ 1 [ P,,_ + +..... I consists of - i)" and lower degree terms.

Now the asymptote parallel toy - m 1 x = 0 is y - x = c1 ,where

= Li (y - rn 1 I 1j (- / Q,,1) [from (2)above], it being

remembered that Li (y/x) = m in this case [See § 16.2 J. In other

words, the particular asymptote in question is

y-- 1 x+ Lt (—F_1/Q.1)=0,
where in determining the limit involved,
x-- we are to put y = in X and then.
make x -, .
CASE II. Let F',, have a repeated factor '_ III I x ). The equation (I) can
then be written as
(y_mx)2
Q,, + P,,. i + F,, 2 =0, . .(3)
where Q,,_ 2 is a homogeneous expression of degree
n - 2 and
F,,_ 2 1 P,,_ 2 + 1P,,_3 +.... . I consists of
(n - 2)th and lower degree terms.
Now the asymptotes parallel to y — m,x = 0 willbe y —m,x = Cr,
where cr = Li (y - in, x ), and this from (3) is given by -
+
Cr2 + LI =
i- Q,,_2
it being remembered that LI (j ) = in here.
512 DIFFERENTIAL CALCULUS

If P,, 1 does not co.ain y — m x as a factor, then it is easily seen

that c, 2 , as given above, does not tend to a finite limit, and accordingly
there are no asymptotes parallcl to Y = n,x
If , on the other hand, ',,_ has a factor y — m ,. x, assuming
= (y - m r x ) R,,. 2 ,wecanwritc (3)intheform
R, F2
(y_m,x) +(y_m,.x )-'_- +-- =

and arguing as before, the required asymptote, ill be given by

(y–mx)2+(y_ in, x)Li —-+ Li

_,-_ Qfl__2 Qn-.2
it being remembered that in proceeding to determine the limits we are to use
Li (y/x ) = m, here.

The two parallel asymptotes corresponding to the two repeated factors

of P,, are thus obtained.
Similarly, we may proceed in cases of factors of P. repeated more than
twice.

Note. If, in P,, the term involving v" be absent, that is, a,, = 0 clearly P,,
will have a factor x , and corresponding to this there will be in general an
asymptote parallel to y-axis, i.e., parallel to x = o. •(In) (which is in
general of degree n) will have its degree lower than ii in this case. If, for
X2
instance, be a factor of I',,, P,, (ni) will be of degree n –2, as x2y"2
will he the term involving the highest power of y in P,, . In this case, there
will be in general two asymptotes parallel to y-axis (i.e., x = o) and n 2
asymptotes corresponding to the roots of 0, (en) = 0, i.e., corresponding to
the other factor of P,
Thus, all the possible directions of the asymptotes of the algebraic
curve (including those parallel to -axis) will be indicated by the factors of
and the asymptotes may be very effectively determined by the method of
the present article. [For illustrations, see Ex. 1-5, § 16.8. 1

A special case (Asyin plates by inspection).

If the (/Ilation to an algebraic curve can be put in the form

F,, + F,,, = 0 , where F,, consists of it and lower degree ter,ns which
ASYMPTOTES 513

can be expressed as a product of n linear fectors none of which is repeated,

and F,,_ 2 consists of terms at most of degree n — 2, then all the asymptote.

are give by F, 0.

For, let F. = (a t x+ b 1 y+ c 1 )a2x+b2y+c2)

(a,,x+b,y+c )

= (a 1 x+b 1 y+c 1 )Q,,_i (say),

where Q_ 1 is of degree n — i.
The equation of the curve can then be written as

a 1 x+b 1 y+c 1 +F,_2/Q,,.1 0,

and the asymptote parallel to a x + b y = 0 is, as shown above,

a 1 x+b 1 yl-c 1 +Ii (F,,_2/Q,,J 0

where, in calculating the limit of the last term, we are to put

y=_(a1/bjx,

and then make x —+ ,and this limit is easily seen to be zero, since F,,2

is at most of degree n —2, and Q,,_ 1 is of degree ,i —1.

Thus, a 1 x + b 1 v + c 1 = 0 is an asymptote. Similarly, each of
a 2 x + b 2 y + c 2 = 0, etc. will be an asymptote. As there are n
asymptotes here, F. = 0 represent all the asymptotes.

Note. If in the above case, F,, consists of real linear factors, some repeated,
and some non-repeated, the non-repeated linear factors equated to zero will
be asymptotes to the curve. The asymptotes corresponding to the repeated
factors, however, will have to be obtained as in the general case.

16.7. Asymptote of Polar curves.

Let r = f o) be the polar equation to a curve. This may be written as

u__=_r=Fo)(saY).

P being any point (r, 6 ) on the 'curve, p — o as r - which

requires F o) -4 0. Let the solutions of F o) = 0 be U = a, /3, y,... etc.
Then these are the only directions along which the branches of the curve
tend to infinity. Consider the branch corresponding to U = a. Let the
33
 514 DIFFERENTIA L CA LCULUS

straight line r cos (o - a) p . (2) be the asymptote to this branch.

0
rig iu.,.i
Then p = ON ,Where ON is the perpendicular from the pole 0 on
the line, and .LNOX = a1 Let OP produced meet this line at Q. if PM be
the perpendicular from P on the line, then
I'M = PQ cos QPM = (OQ .- o) cos QON

= (p see (0 -(X 1 ) .r(o )}cos (o - at)

[Front (1) & (2)
=-f (0)cos (0-a1),
Now since (2) is an asymptote, PM — > 0 as p , i.e., as () - a
for the branch in question.
Li(p-f(0)Cos -a)} =0
0-
or, L i jo) cos _a1)= p and as!) is finite, and f ((9)-

as0-a, ii cos(0-a1)rO;

a-a 1 =-ir or,a1

cos (o - a
Again, p = Li f (ü ) cos (o -a = 0-,a
Li
F(0)

-sin(6.-a1)
Which being of the form = LI
F'(0

sin(a - a) 1
- F' (a) F'(a)
 ASYMPTOTES 515

Hence, (2) reduces to rcos (o - a + 1- 71 ) = -

or, rsin (Q.-a)=1/F'(a)
which is the required asymptote.
Similarly, the other possible asymptotes corresponding to the other
branches are rsin (e - /3) = IJF'(0 ), rsin (o - y ) = 1/F'(y ),elc.

16.8. Illustrative Examples.

Ex. 1. Find the asymptotes of the cubic

x3 —2y 3 +y(2x—y)+y(x—y)+!=o C. P 1949, '971

The curve being an algebraic curve of the third degree, since the terms
involving x 3 and y3 are both present, there are no asymptotes parallel to
either the x-axis or the y-axis in this case.
To find the asymptotes of (he type y=nrr+c, which are oblique,
considering respectively the third and second degree terms (putting I for x
and in for), we gel here
,, (in ) I - 2nz + in (2—ni )= (i—ni )(i +m ) ( i +21z ),
and O._i- (in )=m( I—rn)

Now, 4,, (m)=0 gives n,!, —1, -.,

(in) n( r— 171 )
Also, c= - , and thus for m=l, c 1 =0; for
(in) —6w 2 + 2-2w
=L c 2 =— 1 and for 111 = — -i, C 3 = .

Hence the required asymptotes are

Y = -r , y =—x— j and
i.e., x—y0, x+ v+1=0 and x+2y=l.
Note. It may be noted that the equation to determine in and c might be
obtained in practice by jiulihig V = nix + c in the given equation, and
then equaling to zero the coefficients of the two highest powers of x.
Alternative method
Writing the highest degree terms in factorised form, the equation can
be written as
(.V - -)(+ )( x ± 2v )+ v(r - )+ I = 0.
516 Dli- ii kEN//A L CA I CuLl/S

Hence the possible asymptotes are parallel to - - Y = 0.. + y 0

and x ± 21 , = 0, and these asymptotes are respectively
y(xy)+10 (I)
ii
(x+y)(x+2y)

x+ y+ ...(2)
–y)v) = 0

and x + 2v + LI 0 .-• (3)

(x– y )(x +)')

The limit. involved in (l),= Li (x±x)(.v+2.)

- –x(x+x)+l –2x2+l
that in (2). = LI
Li =
+ .r )(—
x--27)
,---- -. 2x'

_ 1 X(X+ . )+ I X2 +1
and that in (3), TiT
3x2

1-lence the asymptotes arc

x--v0. x +ylO, x +2v-- 10.
1)2
Ex. 2. Find the asymptotes of 2x(y _5)i =3(y– 2 )(x--
As the curve is algebraic, arranging the terms in descending degrees
the equation can be written as
xv(2y-3x )+2x(3x--7y )+38x-3y+ 6=0 ... (I)
s and 2Y-3x=().
The po sible asymptotes are parallel to v =(), :r=0
The asymptote parallel to x.--O; i.e., to the v-axis, is (equating to zero the
in (1), since the let in
coefficient of y2, the highest available power of v
involving v 3 is absent here) 2x = 0, i.e.. ,v - 0, the '-axis itself.
The asymptote parallel to y =0, (in a similar manner, since term is

absenthere),is –3y+60,or, v=2.

The third asymptote is
2x(3x-7v )-i-38x-3y+6
2y--x+ Li --------------- =0
xv
 ASYMPTOTES 517

2x(3x_x)+(38.-)x+6
and since the limit involved = Li - ., - = - tO,

the asymptote is 2y-3x--I0=0.

Hence tile required asymptotes are x=O, y=2 , 2i'=3x+10.

Ex. 3. Determine the asymptotes of

X +x 2 y-xy 2 -y 3 +2xv+2v 2 -3x-i-y=O (C. P. 200SJ

Writing the equation as

(x+v ) (x-y )+2y( x + y )-3x+y = 0, ... (1)

we note that (here are presumably two parallel asymptotes parallel to
.v + y = 0, and one parallel to x - y = 0.
The asymptotes parallel to x + y = 0 are given by

(x+v +2(x+y).Lt_1__Lt.i_L=0 (2)

provided the limits involved exist.

Now, Li ...L = LI
,-=ox-v ..-+-,x+x

3x-i' 3x+x
and Lt --=LI —=2.
X- y . -. = x+x
=
Hence, the asymptotes from (2) are

(x-t-v) 2 -(x+),)-2=O, or, (x+),+l)(x+),-2)=O.

Again, the asymptote parallel to x - y = 0 is given from (I) by

Li 2v(x+v)-3x+y

(x+y)

2x.(x+x)-3x+x
I. e., x- v-t- Li =0,,.e., x-y+l=0.
+X)2
(X
Thus, the required asymptotes are
x+ y +1=0, x+y-2=0 and x-y+1=0.
518 DIFFERENT]. IL CALCULUS

Lx. 4. Find the asymptotes of the Folium of Descartes

x' + y 3 =3axy. [8.P 1993]

The equation can be written as (x + y) (2 - xv + y 2 ) = 3axj, and since
the highest degree terms have got only one real linear factor x +Y , (the
linear factors of x2 _N y+ v 2 being clearly imaginary), there is only one
possible asymptote here, which is parallel to x + y = 0. The asymptotes in
question is x+
3axj' 3av2
1 = . Li = Lt
2 - xy + 2
.= 2 + X2 2
y •' t +

i.e., x+y+a=0.

Ex. 5. Find the asymptotes ofx(x_v) 2 _3(x2._y )+8v=0.

The possible aymptotes here are one parallel to x = 0 and a pair
parallel to x - Y = 0.
The first one, which is parallel to y-axis, is found by equating the
coefficient ofy1 to zero, (the term involving y being absent, as it should be
under the circumstances), namely, x + 3 = 0.
The other two are given by
x+r V
(x-v Lt_-_-+8Lt- =0,

i.e.. (x-) 2 -3(x-v).2+8=0. or, (x-.'-4)(x-y-2)=0.

Thus, the required asymptotes are
x+3=0, x-'=4 and x-y=2.

Ex. 6. Prove that the asymptotes of the cubic

(2
-),2 )y_2m' +5x-7=0

form a triangle of area a2.

The equation to the curve may be written as

v(x2 _ y2 -2' )+5x-7=0

21
2
or -(y+a) j+a j'+5x-7=0,

i.e., y(x+,'+a)(x-v-a)+av2 +5x-70

which is of the form F3 + F1 = 0, F3 having three non-repeated linear
A SY MPTOTES 519

factors, and so the required asymptotes are given by equating these factors
to zero, namely,
y = 0, x+y + aO and x—y—a=0.
By solving in pairs, their points of intersection are easily seen to be
(—a,0), (a,O) and (0,—a).
The area of the triangle with these as vertices is
aOl
- —a 0 I =
. 2a 2
=a2
0 —a I

Ex. 7. Find the asymptotes, ,fanv, of the curve

y =a log sec: (x/a).
This is not an algebraic curve. To find its asymptotes, if any, which ale
not parallel to y-axis, we know that v = mx + c will be all asymptote, where

in L i I and c = Li (y — tax).
.V_3 A

a log sec (x/a)

Now in the curve, in = Li v = Ii
__ _V •-= N

which limit does not exist.

Hence there is no asymptote non-parallel to y-axis in this case.
To find if there he any asymptote parallel to y-axis, we notice that
Y —*00 when .x/a —* 2mm ± it, and accordingly the asymptotes parallel to
y-axis are (See § 16.3)

x=(2rnI±.n)a
which are the only asymptotes of the given curve.

Ex. 8. An asymptote is defined in the following two ways:

(A) An asymptote iv it hoe, the distance ofwhich from a point

on a curve diminishes without limit as the point on the curve moves to all
infinite distance froni the origin.

(B) An asymptote to a curve is the limiting position of the tangent

when the point of contact moves to an infinite distance from the origin,
Consider the two curves
C+SiflX
C
(ii)
(i) y=ax+b+—. y=ax+b+
X x
520 DIFFERENTIAL cALCuLI/S

Show that for the first curve, an asymptote exists according to both
the definitions, but for the second curve, an asymptote exists according to
the first definition, but not according to the second.
Let us consider the first definition. According to this it has been proved
( 16.2) that the straight line y = nzx+c will be an asymptote to a curve.
where in Lt I and c Lt(y— mx), (x, v) being a point on the curve.
.'.- x
provided the limits exist.
Now for the curve (i),
y qa+ h c"
m= El -= — +-- I=a,
X x V2

and c = Lt(y— mx)= Lt(— ax)= Lt.,(b+)= b.

Accordingly the asymptote exists, given by y = ax+b.

Again, for the curve (ii),
y qa b c+sinx"i I.'sin xI!5l
= +—+-------,— )I =a
X x X
S
c+sInx
and c=Ls(y— nix) =Lt(y— ax) = =b.
X)
Thus, the asymptote exists here also, given by y = ax+b.
Next, consider the send definition.

For curve (i), =a- and so the equation to the tangent line at
dx x2

(x,y)is Y _Y =(a— )(X _x)

c ( c
or, Y= — Ix+ y— xI )=(a—c.!X
a------- "i +Ib+ 2c
x )
(.— '\ I x x) X

As x - , the equation becomes Y = aX +b, which is then the

definite straight line towards which the tangent line approaches, as the
point of contact (x, y) moves to an infinite distance. Hence this is the
asymptote.
dy xcosx—(c-+- sin x
Forcurve(ii), — =a+
dx x2
ASYMPTOTES 521

and the equation to the tangent line at (x, ') is

• I xcosx—(c+sinx)1
• Y — y=a+ (x -x )
X2
or substituting the value of y from (ii),

( osx c+sinx'\ I 2(c-t-sinx)

Y= aI +--------- )IX -H •--cosx+b
X 2

Now, as x+oo • cosx does not tend to any definite limit. Hence the
tangent line at (x, y) does not tend to any definite limiting position and so
the asymptote does not exist in this case, according to the second definition.

Ex: 9. Find the aswnptotes, ifany, of the curve (r—a )sinO = b.

sinO
The equation can be written as u = 1' (0 )(sav).
b+asinO
o given by ii = 0, or sin = 0, giving
are
The directions in which r -4

0= nit. -
du cos0(b+ain0)-sin0.a.cos0
Now F'(0), 'f.. - =
dO (b+asin0)2
— bcosO
(b+asin0
bcos-nit cosnit
and for 0= nit , this =
b b

Hence, as in § 16.7, the required asymptote is given by

r sin(0 — mt ) = l/F'(nit) = b sec nit

which, whcther'n is even or odd, reduces to

rsinOb.

16.9 Miscellaneous Worked Out Examples

Lx. 1. Find the asymptotes of:

[C. P. 1982, 94 1
(') . t 2 -4y 2 = I.

(ii) x2— y2'9. [C. P 19981

(iii) x2y2 =a2(x2 +2). B . P 19901

I

(iv) xv-3x-4yO. I C. P 20001

 522 DII CI RENT/AL CALCULUS

Solution

(i) x2-4y2—l=0
or, ( x+2y Xx - 2y ) —l=o or, F,+i=0
where F2 = (x +2y)(x-2y) is of degree 2 and F0 = — I, which is
of degree 0.
So, asymptotes of the given curve are given by / = 0
i.e., (x + 2v)(x - 2v) = 0
Hence, the asymptotes are x+2y=0. x-2'=0
(ii) x2— v2-9=0
or, F,+J=0
where F, = (x-ty)(x—y) is of degree 2 in x and v and F0 = —9,
is of degree 0.
Hence the asymptotes are given by F, 0
i.e., x+v0 and x—y=O.
, , .7 , .7
(in) xm' — a x, — crv =0 ... (I)
This is a fourth degree equation in x and r So the curve
represented by (1) may have at ,nosz four asymptotes.
Here, terms containing x 4 and 114 are absent, so the curve has
asymptotes parallel to .v-axis and y-axis.
The coefficient of highest available power of x, i.e. of x 2 is
Y2 —a 2 and the coefficient of y2 is X 2—a2.
So, asymptotes parallel to the x-axis are : =0
i.e., y±a=0.
and asymptotes parallel to the y-axis are x 2 - a 2 = 0
i.e.. x±a=O.
(it) x'-3x-4v=0 ... (I)
As in the earlier problem, the asymptote parallel to the x-axis is
y —3 = 0 and the asymptote parallel to y-axis is x - 4 = 0.
Ex. 2. Find the asymptotes of:
(1) x3+2x?y+.y2_x+1=0 F C. P19921
(ii) (x+y)2(x+2v+2)=x+9y+2. IC. P. 20001
(iii) 4x3_3xv2_v3+2x2_xv_v2_l=0. 1 C. P 1998. 2001 1
ASYMPTOTES 523

Solution : (i) x 3 +2x 2 y+ 2 -x+1 =0 ... (I)

or, x(x+y+l)(x+y-1)+l=0

or, F+F0=0
where, F3 =x(x-t-)'+l)(x+y--l) is of degree 3 and it has three
non-repeated linear factors and F() = I, which is of decree 0.
The asymptotes of the curve (I) are given by
F3 =0, or, x(x+vi-l)(x+y-1)0
i.e., x=0, x+y+1=0. x+v-1=0

(i/) (x+i)2(x+2),+2)=x+9r+2

or, (x+v)2(x+2y+2) = (x+2y+2)+7y

or, (x+),)2(x+2v+2)-(x+2y+2)-7,'O
)2
or, (x ±2v + 2){(v+ i}_ 7y = 0

or, (. + 2v+ 2)(..v v+ + v - i) -- 7y = 0

or, F3 + i=0

where F3 =(x+2v-1-2)(x+v+i)(x+y-l) which is of degree 3

and it has three non-repealed linear factors, while F, = - 'Tv. which is
of degree I. Hence, the asymptotes of the curve are given by
x+2'i1+2=0. -+y+1=0 and x+y-1=0.

'
(in) 4x 3 -3xv -y +2x -,'-y-l=O
3 ... (1)

This is a third degree curve, so it may have three asymptotes at

most. Since the terms involving x3 and y3 are both present, it has no
asymptotes parallel to x-xis and y-axis. -
Equation (1) can be be written as
3 3 3 1 3
4x -4x +xv - v +2x - .' - y -1=0

or, 4x(x+y)(x-v)+y 2 (x-y)+2x 2 -.ty- y 2 -1=0

or, (x_y)(2x+y)3+2x2_2y+y_y2_l0

or, (x - ),)(2x +),)2 +2x(x- y)+ y(x- y )- I 0

or, (x-y)(2x+y)(2x+3+1)-1'0 or, F3 + F0 =0
524 DiFFIRENTIA 1. CALCULUS

where F3 is of degree three and it has three d i fferent linear factors,

while F0 =
-1, which is of degree 0.
Hence, the asymptotes of the curve (I) are given by
(x - y)(2x + ), )(2x + y + 1) = 0
i.e., x-y=O, 2x+y=O and. 2x+v+1=0.

EXAMPLES - XVI

Find the asymptotes of the following curves I Ex. 1-31]

1. v2-x2-2x-2)'-3=O.
2. 336xy+llx2y_6x3+v2_xI+2x3y_1=0.

3. x3+3x2y_xy2-3y3+x2-2.n+3v2+4x+5=r0.

4. 3x3+2x2y-7xy2+2y3-14xy+7y7+4x+5y=0.
I C.!' ]943]

5. x3+2x2y-xy2-2y3-f4y2+2xy-5y+6=0.

6. -=1. 7.x3-y3=3.y(x+y).
a 2b2
4 4 2 '
8. x -y +3xy+3xy+xy=0.

9. 4x4-5x2y2+y4+y3-3x2y+5x-8=0

10. (i) xy - - 3x = o. (ii) y 2 (x2 - a2 ) = x.

12
(in) xx
2\
^ y J+ a
I
x -
' Y2
) = 0.
11. x 2 y 2 -4(x- y)2 = 0.
+2v-3
12. x 2y2 -x 2 y- xy 2 + x + y + I = 0. [C.P 19371

13. y2 x
2
- 3
YX2
- 5X
V 2 + 2x + 6ç ' '
- x - 3y + 2 = 0.
14. 4
x - x 2 y 2 + x2 + y
' - ' = 0.
15. y3-yx2+y2+x2-4=0.
16. x 2 (x- y Y _a2(x2 + ),2 )= 0. IC.P 1945]

17. x 3 -4x y 2 -3x 2 +l2v-12y2+8x+2y+4=0.

18. x3 +3x2 y-4 3 - x+ y+3 =0. IC.? 2003]
 525
ASYMPTOTES

[Cr1939]
19. y3_xy2_x2y+x3+x2_Y21

20. y +x2 y + 2xy 2 - y + I = 0. [Ci'. 1941, '44, 87, '90, '961

21. (x2 - Y 2 )2 - 8(x 2 ±2 )+ 8x —16 = 0-

Z.2. y ( y _ x ) 2 (y_2x)+3x 2 (y_x)_2x 2 =0.

23. x2 (x+v)(x_y+2x 3 (x_) , ) 4Y 3 =0.

24. (x + y )( x - 2 )(x - y + 3x), ( x -)' ) + + y 20.

• (+v)3(x_v)2_2(x+y(xY)22(x22)(.+
- )2 +4(x ) , =
+2 ) O-

26. y3_5xy2+8x2y_4X34Y7+I2.V8+3Y_32O

27. (i) - 2 )= x 2 + y2.

(\
2 2 '
(it) xs'ix y )= x - v2
+

28. (v2_y2)(x2_9v2)+3vv_6x_5Y1.2=0.

29. (i) x' _6x),i +llxy 2 _6y 3 +2x—y+l=O-

2 +x ' —2v+2.+yf70
(a) x4 —5x2 y2 +4y4

30. (.y+1)(x+), +)(x_2Y+3)25Y+1.

31. (I) y = tan .x. (ii) y

(lii) y (iv) y = log x.
e-

32. Show that the asymptotes of the curve

2a.
x 2 v 2 = a 2 (x2 + y2 ) form a square of side

33. Show that the asymptotes of the curve

- 2 (2 + )2 )- a 3 (x 4 y )+ a 4 = 0
x2y2
form a squale. , two of whose angular points lie oil curve.
iCr /947]

34. Show that the finite points of intersection of the asymptotes of

xy(x2 - y )+a (X2 + y2 )- a 3 = 0 with the curve lie oil circle
2
whose centre is at the origin.
35. Find the equation of the cubic which has the same asymptotes as the
)2
curve 2x( v _3)2 = 3Y (x - 1 and which touches the axis of x at
 526 DI/FEI?LNIJAI, CALCULUS

the origin and goes through the point

36. If any of the asymptotes of the curve
11,2
+2hxy+by 2 +2gx+2fs'+c = o (/2 > ab)
passes through the origin, prove that
af 2 + bg 2 = 2fg/z.
37. If the equation of a curve can be put in the form

y = ax + h + 0 (. ), where 0 (x) 0 as
then show that y = ax + /, is an asymptote of the curve. Apply this
method in deterrninin2 the asymptotes of the cut .
- V 2 - 3x + 2 0.
38. 'An asymptote is sometimes dehncd as a stra:ltt line which cuts the
curve in two points at infinity svt(Iiout being it'.clfat infinity.' Comment
Ott his definition. Attempt a correct defi nit i tt and USC it to Obtain the
asymptotes of(x+v)2(x+7V.f7)+9,_2
39. Find the asymptotes of:

(I) r = a( cos Q + ,cc ()). (ii) r cos O 2a.cin0.

(iii) r = a see + b tan . (tv) rcosO = asinO
(v) r = a cosec O +1. (Vi) rsinnO = a.
(vii) rO = a .(viii) r" sin nO = a" (n > i).
40. Show that there is an infinite series of parallel asymptotes to the curve
a
= + h.
U sill O
41. Show that all the asymptotes ol the curve r tan nO = a touch the
circle r = a / n.
42. Show that the curve I = a sccnO + bout no has two Sets of
asymptotes, members of each set touching a fixed circle.

ANSWERS

1. 11 —x=2, 3, +x=O 2. vrx, 3'=2v+, v=3x-4.

3. 4x+12v+9=0 2x+2-3=O, 4x-41-41=0
4. 6r-6x+7=0, 2v6vf3=0. 61 1 +3x +5=0.
5. .v+2v=O,x+v_1()xv+l=o
ASYMPTOTES 527

6. Y X. 7.
x-y=2.
8. x-fy=O, 2x2y+30.
9. 3x-3y-I =0. 3x43y+1 =0, 12x-6y-1 =0, 12.x+6y+I = 0.
(ii) y=0, x=±a.
10. (i) x2, y=3.
x=a.
12. x=0,x=I,y0,y1.
11. x±2; v=±2.
13. x = 2.x = 3, y= t,v2.  14. x=±Iy±.r.
15. y=ly=x-J.y-x--I. 16. x 
x-y=±aJ.

17. x+3=O; x-2y=O; x+2y=6.
18. i=x; x+2v-!=0; x+2y+l=O.
.20. v=O, x+y=±I.
19. v=±x;y=x+I.
21. y =x±2 v=-x±2.
22. 2y+3=0x_y+1=0;x_y+20;4x-231_30.
23. x =±2; .r-y-f2=0x--y--1=O;x+y+l=O.
24.x-y _2=0;2x2+lr0;2x+2y-lO;x--2Y+2O.
25. x+y-20,x  +r=±I; x-y=±I.
26. y=.v; v=2v+I:y = 2x-f3.
27. (i) x=O.  v=O.x -+v=0. x - y (ii) V=O. v=O.
O.
28.x+y =0, .r-'=O, .'+3v=0.x-3yO.
29. (I) .i-v=0.x -2y=0.x-3v=0.
(ii) x+2v=0, x-2v=0,.s+y=O,  x-- y=O.

30. x-y+i0, x+v+I=0 .v-2v+3O.
31. (I) -v = ( 2,, + i )-- 7t. where n is zero or any integer positive or
negative.
(ii) Y=O. (iii) )'O. (iv) x=O.
35. 2xv 2 -3x7v-6xv+7v=o. 37.
38. x+2i'+2=0;x+v=±2.
39. (i) r cos O (ii)
a. r cos O (iii) r cos () a ±/
±2z.
(iv) r cos 8 = a. r sin U = a.
(v)

( v ii.. r sin 0- = - Sec rnir , where in is an integer.

('it) r Sifl 0 = a (yin) 0 = - . where m Is an Integer.

NQ SEC. A. ENVELOPE OF STRAIGHT LINES

17.1. Introduction.
Let us consider the equation
x cos a 4 y sin a = a.
This represents a straight line; by giving different values to a, we shall
obtain the equations of different straight lines but all these different straight
lines have one characteristic feature common to each of them, tic., each
straight line is at the same distance a from the origin. On account of this
property these stiaiht lines are said to constitute afamily and a which is
a constant for one but different for different lines, and whose different
values give different members of the family, is called the parameter oft/ic
family. It should be noted that the position of the line varies with a.
As we have a family of straight lines, we have a family of curve. Thus
the equation
(x—a+y2=r2
represents a family of circles for different values of a, all the individual
members of the family having the common characteristic, viz., they are of
equal radii and their centres lie on x-axis. Here is the parameter of the
family.
In general, 1/ic equation of a family of curves is represented by
F (x, . a ) = 0, when a is the parameter.

17.2. Definition of Envelope.

If each of the members of the family ofciirves C F ( x, y. a ) = 0
touches a fixed curve F, then F is called the envelope of the family of
curves C. The cuive F also, at each point, is touched by some member of
the family C.

Illustration
We know hat xcosa+ ysin a = ci touches the circle x 2 + y 2 =a 2 at
(a cos a, a sin a ). 'fhns, each of the members of the family of straight lines
C x cos a + \Sin (1 a (for different values of a) touches the fixed circle
X 2 + ),2 =
E x 2 + y 2 = a 2 , and hence the circle a 2 is the envelope of (he
2
family of straight hues x cos a + y sin a = a also the circle x 2 + y 2 = a at
each point (a cos a, a sin a ) obtained by varying values a, is touched by some
member 0/ i/ic fanuulv of straight lines.
 ENVELOPE OF STRAIGHT liNES

Fig 17.2. I
In the present section we shall confine ourselves to the determination
of the simplest type of envelope, i.e., the envelopes of straight lines.

17.3. Envelope of straight lines.

The equation of she envelope of the family of straight lines
x, y, a) M y - f ( a ).x - 0(a) = 0 (a being the parameter) is the

cz-elun,nan: of F = 0 and -W= 0.

aF
From F = 0 and 0, we have respectively
aa

• y =f(a).x+0(a)

and 0=f'(a).x+Ø'(a) . . .(2)

from (2), •=- gl,a) say . .. (3)

T- ) -
and from (l),y= f'(a)O(a)_f(a)'(a),,(à)say ...(4)
f(a)
Hence the curve (i.e., the envelope) whose equation is obtained by
eliminating a between (1) and (2) is the same as the curve whose equation
is given parametrically as
• x=g(a)1
• y=h(a)

Now, the equation of the tangent at the point 'a' on the curve (5) is

-
yh(a) =h'(a) {x—(a)}, . (6)
g(a)

34
530 DIFFERENTIAL CALCULUS

Substituting from (3) and (4) values of g (a), h(a ), g'(a ), h'(a )in (6)
and noting that h' (a )/g (a ) reduces to f (a ), and simplifying, we get
the equation (6), i.e., the equation of the tangent at 'cx' on the curve (5) as
y=f(a).x+.p(a)
which s the same as the equation of hte given family of straight lines.
Thus, every member of the family of straight lines F ( x, Y. a ) = 0
touches the curve whose equation is given as the a-eliminant of F = 0

and .- = 0, and hence the a-eliminant curve is the envelope of the family

of straight lines.

Cur. 1. From the definition, it at once follows that every curve, is the
envelope of its tangents.

Cot.2. Since we have seen that normals at different points on a curve

touch the e.volutc at the corresponding points, it follows that the evolute of
a curve is the envelope of its normals.
Thus, if N ( x, y, a) = 0 be the equation of the normal of a curve at a
point with parameter a, the evolute is obtained by eliminating a between
dN
N (, y, a )= 0 (1) and = 0 (2). Since the evolute is the locus of centres
aa
of curvature, the co-ordinates of the centre of curvature are obtained in
parametric form by solving the above two equations for x and yin terms of a.
The above methods of determiningihe evolute and centre of curvature
are much simpler than the methods already given in chapter XV (curvature).
[Sec Lx. 4 and Ex. 5, An. 17.4.1

17.4. Illustrative Examples.

Ex. 1. Find the envelope of the straight line y = ,nx+--,

in being the
in
variable parameter (,n * o). C.P 1994, 2008 V P '95]

Here, mx+--y=0
in
Differentiating with respect to ni.

10120 .. in =, ..
ENVEWPE OF STRAIGHT LINES 531

Substituting these values of in in (I),

.x+a/)_Y=O.
±[
i.e., ±2I = y. or. Y2 = 4ax (parabola)
which is the required envelope.
Ex. 2. Find the envelope of the famil y afstraight lines

Acx 2 +Ba+C=O'
where a is the variable parameter. and A , B, Care linear function of x, y

We have Aa 2 +Ba+C0 (I)

Differentiating this with respect to a. we have
2Aa+f1=0,i.e., a=-B/(2A).
Substituting this valie of a in (I). we get
B2 B
2
M B2=4A C.

Thus, the envelope of the family of straight lines Au 2 + Ba + C = 0 is

the curve B 2 = 4AC.

Note. When the parameter occurs as a quadratic in any equation the above
result is sometimes used in determining the envelope.
Ex. 3. Find the envelope of the straight lines

a
where a and bare variable parameters, connected In' the relation a + h = c
c being a non-zero constant. j C. P. 1998, 2006

Since a+b=c, :. bc - a.
the equation of the straight lines becomes
X V ( c-a)x+av=aC-a
\ I
or,
a c-a
or, a 2 +a(y-x-c)+cx"O.

Since it is in the form A ct 2 + Ba + C = 0, its envelope is

B2 = 4A C, i.e., (v - x- ' = 4cx,
which represents a parabola.
 532 DIFFERENTIAL CALCULUS

The above equation can be written as

x 2 +y 2 +c 2 =2xy+2cx+2cy.

, = .J (which represents a parabola).

Otherwise
The elimination ofa and b can also be performed thus:
Differentiating the equation of the line and the given relation With
respect to a, we get
x v db A
------=0 and l+—=O.
a2 b2 do do
dl,
On equating the values of - from these two, we get
do

a a-fl, c

TX Vy +JT+
c __
aTT.
Substituting these values of a and bin the equation of the line, we get

i.e.,
Ex. 4. Find the evolute of the parabola y 2 = 4ax (evolute being regarded

as the envelope of its normals).

The equation of the- normal to the parabola at any point 'm' is

y= nix –2am–am 3 . ... (1)

Let us find the envelope of th is, in being the parameter. Differentiating
(1) with respect to in,

0=x-2a-3a,n 2 . or, 12 =(x-2a.)/(3a). ...(2)

From (1), y = m(x-2a_arn2 )= ,n(am 2 Lam2 )= 2am3,

=4a2 (x-2a )3
y2 =4a2 m 6 from (2).
27a3

i.e., 27ay 2 =4(x-2a)3,

which is the envelope of the normals, i.e., the required evolute of the parabola.
ENVELOPEOFSTRAIGJITLINES 533

Ex.5. Find the centre of curvature of the ellipse x + = 1.

The normal at the point '4)' is

ax sec 4)-by cosec $=a 2 -b 2 . ,.• ( I)

Differentiating this partially with respect to 4),

axsec$ tan 4)+by cosec 4) cot $=O. ... (2)
Solving for and)' from ( I ) and (2), we easily get

a2-h2a2-b2
cos 4), y= sin 4).
a h

EXAMPLES- XVII (A)

Find the envelopes of the following families of straight lines

(Et i--9):

1. x Cos a + y sin a = a,parameter a.

2. aA sec a - bycoscca = a 2 - b2 . parameter a.

3. x Cos 30 + Y sin 30 = a( Cos 20 Y2' , parameter Q.

4. x Cos a + y sin a a Cos a sin a, parameter a. I V.P. 199tfl

5. y = ,nx+ all +,n 2 ,parameterm. CP. 1993,2004, '05VP 96}.,

6. v= mx + ..Ja2ni2 + b2 parameter m.
CP 1990, '97. B.? '86, '94. '96, VP 20011
7. x sec 2 0 + y cosec 2 0 = a, parametcr 0.

8. x,J j + y.J0 = a, parameter 0.

9. xcos' 0 + ysin 0 = a, parameter 0.

10. Find the envelopes of the straight line

a b
where the parameters a and b are connected by the relations
i) a2 + b2 = c2 Ic.? 1989, '96,200/, '03 B.? 1989, 95, VP '99, '971
(ii) ab = c 2 , [ 8.? 1988. '93 1
c being a constant.
34 DIFFERENTIA L CA LCULUS

11. Find the envelopes of the straight line

=1
I in
where I and ,n are parameters connected by the relation
I/a + rn/b = l,a and b being Constants.
12. Find the envelopes of straight lines at right angles to the radii of the
following curves drawn through their extremities:
(i) r=a(l+ COS O). (ii) 2 =a2 COS 2U. (iii) r=
13. From any point P on a parabola, PM and M arc drawn perpendiculars
to the axis and langcnt at the vertex : show that the envelope of MN i.s
another parabola. [ C. 1996, 99, 2005]
14. Show that the envelope of straight lines which join the extremities of a
pair ofconjugate diameters ofan ellip s e is a similar ellipse.
15. If PM, PN be the perpendiculars drawn I rom any point P on the curve
y = ax upon the co-ordinate axes, show that the envelope of MN is
27v + 4ax 3 = 0
16. From any point P on the ellipse x 2 / 2 + 1 ,2 /1, 2 = I. perpendiculars
PM and PNarc drawn upon the co-ordinate axes. Show that MNalways
touches the curve (x/a Y + ( v/I, ) = I.
17. Find the envelope, wheni varies, of
(2 \ (2 \ I
a1! + 2a2t +a, )x +b 1 s +2b2 1 +I,. )y+ic[ +2c2 1+c, j0.
18. Find the evolutes of the following curves (evolute being regarded as
envelope of normals):
(i) x=a COS Ø. y= b sin Ø. (ii) k(118
- . X2 2 + 21'
(ill) .v
' + y3 = . (iv) la v = I-
(v) x = a(0 —sinO ), y = a(l —cosO ).
(vi) x = a (cost + t sin , ), y = C (Sin I - i cost ).
19. Two particles P. Q move along parallel straight lines one with uniform
velocity u and the other with the same initial velocity ii but with uniform
accelerationf. Show that the line joining them always touches a fixed
hyperbola.
20. Show that the radius of curvature of the envelope of the line
xcosa+ysina=f(a) is f(a)+f'(cr).
 535
ENVELOPE OF STRAIGHT LINES

ANSWERS

1. x+y2a2. 2. (+(bv =(a2_b2

(2+y2)2=a2(x2_y2) 4. x+y=a.
3.

5. 2
x 2 +v2 a . 6. x 2 / 0 2 +y 2 1b2 I.

7. 8. x+ya.

1. X" +V 2 " = a'"•

10. (i) X , +y3c' (ii) 4xy=c2

11. I+f
Va
ly
12. (i) a circle through the pole (ii) a rectangular hyperbola;

(iii) an equiangular spiral.

17. (aj x + + c1 ax + i,v + C 3 ) = (u,x + I,2 y + c2

27a V 2 4 (2a.
)2
18. (I)(ii)

(ill) (.+y) +(-y) =2a• (iv) same as(i).

(v) o -eliminentof .v a(9 + sin 6), y = _ a(i - cosO ).

(vi) x2 +1 .2 = e1 2.
 536 DIFFERENTIAL CALCULUS

SEC. B. ENVELOPE OF CURVES

17.5. If a curve Eexists, which touches each member of a family of curves

C [ f(x, y, a)= OI. the curve Ejs called the Envelope of the curves
C. Since E is the locus of the points of contact of the family of curves
f(x, y, a)= 0 the point where the curve f(x,
y, a)= 0 for a
particular value of a. touches E, depends upon that value of a. Accordingly
the co-ordinates of any point on E are functions of the parameter a [being
of the forms x = 0(a), y = v(a)] and they satisfy the equation
f (x, y. a) = 0 of the enveloping curve which touches E at that point.
The equation of the tangent to a C-curve at (x, y) is

(x-x)-L+(y).)L=o
and that of the tangent to the E-curve at (x, y) is

'IX (
(2)
dy
da dot
[since equation of E is of the form .v = 0(a), y = V(a) I

or, yty
da 'da
Since the lines (1) and (3) are coincident, coefficients of X and Yin the
above two equations are proportional.

If If

i.e.,
'v dx axda ayda
da da

Now, differentiating f (r, y, a ) = 0 with respect to a, remembering

that x and y are now functions of a, we get

?_L+ aiaf dy
aa axda

from (4),-L=O
aa
Hence the equation of the envelope, in case an envelope exists, is to
be found by eliminating the parameter a between the equations
ENVELOPE OF CURVES 537

f(x, y,a)=0

and L=O.

oi It is shown in Art. 18.l (iii) that the circle on the radius vector of a curve
as diameter touches the pedal of the curve, so the pedal of a curve can be
obtained as the envelope of the circles described on the radius vectors of
the curve as diameters. F See Art. 17.9, Es. 6.

Note. It should be noted that the a-eliminant between f (, y, a )= 0

and ii = 0 may contain other loci, besides the envelope, for instance
ax
nodal locus, cuspidal locus, tac-locus, etc., in case the family of curves C
has singular points.
17.6. The envelope is, in general, the locus of the ultimate points of
intersection of neighbouring curves qf a family.

The co-ordinates of the point of intersection P of two neighbou, ing

curves of the family must satisfy

f(x, y,a)=.0 and f(x, y,a+Aa)=0,

i.e f(x, y,a)=Oand f(x, y,a+La)-f(.r,,a)- 0

ta
the second relation by Mean Value Theorem becomes

--f(x,y,a+Oa)=O, Whei0<O<j.
aa

Now as Aa -*0, P satisfies f = 0, = o.

Ocx

the required locus is the a-eliminant of f = o, .L = 0.

aa
Note l. Although the above theorem is generally true, but it is not always
true. For example, consider the family of semi-cubical parabolas
y = (x- a)3 . Here for different values of a, we have different semi-cubical
parabolas, no two of which intersect but every one of which touches the
x-axis. So here the x-axis is the envelope, although no two members of the
family intersect. From the graphs of the curves the whole thing becomes at
once clear.
538 DIFFERENTIA L CA LCULUS

Note 2. A lternative definition of Envelope.

The points of intersection of the curves f(x, y,cz)r0 and

characteristic points of the

-. --f (x, y, a) = 0 (a being given) are called
aa
family f (x, y, a) = 0 (for the given (x) if these points exist (i.e., if f = 0

and = 0 intersect) and if those points are not singular points of

of curl'es
f (x, y, a) = 0. The locus oft/ic characteristic points of afaniily
is sometimes called the envelope of the family.

117.7. Envelope of a special family.

If the curve f (x, v, a )= 0 be algebraic, the a-eliminant of f = 0,
af = 0 is the condition that .1 = 0 (considered as an equation in a) has
aa
equal roots. I
Theor y of Equation I
Thus, if f(x. a)= fl(x,y)a 2 + B(x.v )cr+C =0 (i.e., if
y.
f (.v, y, a ) = 0 be a quadratic in a, the parameter ], the envelope of the
family is given by

For illustration see Ex. 2 ofA rl. 17.91

17.8: Envelope of two-parameter family.

If the equation of a family of curves involves two parameters a and (3
connected by a given equation, then we can proceed by two methods in
finding out the envelope. Suppose the equation of the family is
J. ), a, i)=° ... (I)

where a and (3 are connected by the equation

First Method: Suppose we can solve 0 (a. fJ ) = 0 for 13 in terms of

a; then we substitute this value of 1 3 in (1) and now (I) reduces to one

parameter family and we eliminate a between f=0 and f

L = 0.
aa
Second Method: For a particular point (x, y) on the envelope

(3)
 ENV ELOPE OF c:u g vEs 539

and from (2), -du+d13=O .. .(4)

act
Eliminating da, d13 between these two equations (we may regard a
as the independent variable and j the dependent variable), we get

af f
da

If we eliminate a and 3 from (I), (2) and (5), we obtain the equation of
the envelope.

Note. This method can be extended to obtain the envelope of a family

depending upon it which are connected by ( - i ) equations.

17.9. Illustrative Examples.

Ex. 1. Find the envelope of ihe /wn,ly of ellipses

y ,=,
(a-a)
cx being the parameter
We have xcx 2 +y 2 (a-(X ) 2 =1. (I)
Differentiating with respect to a, we have

- i__ ... (2)

V2 )..2 •V2 -

by(i)J
a a-a a a

2 ' X, +

cz (a-ay (I a a-a a

X - 4-y' =a, which is the required a -eliminantbelween(l)and

(2), is the equation of the required envelope.

Ex. 2. Find the envelope of the s ystem of parabolas

X + ),2 v = I. A bei:ig the ,,aran:eleI:

540 DIFFERENT/AL CALCULUS

Since the equation of the family is

2y+Ax2-1 =0,

and since it is a quadratic in A, the parameter, by Art. 17.7, its envelope is

x4+4y=0.
Ex. 3. Prove that the envelope of the paths of projectiles in vacuum from
the same point with the same velocity in the same vertical plane is a
parabola with the point of projection as focus.
The equation of the path of the projectile with the point of projection
0 as origin and the horizontal and the vertical lines through 0 as axes of.
and y is
x2 c2
y =xtana--g 2 2 =xtana—g_.-(1+tan 2 a)
ucosu U

F See Authors' D namics: Art. 11.5] y

= mx_g (J+m2 ), where tan a=in,

-
2 I I
i.e., in
U -

Here a, and hence lana, i.e., ni being the variable parameter, the
equation of the envelope is, by Art. 17.7,

x2

2
2 (
I.e., ... .t
U2 =-- I
9 2g

Transferring the origin to the point 10, the equation of the

2g)

envelope is x2 ------ y
=
21?
8
, which is a parabola with its vertex on they-axis at

the point 1. .!L')

i 2j
and its concavity turned downwards and latus rectum

2u2 u2
4a = - and hence, a' being equal to —,the focus is at the origin.
9 9
 ENV ELOPE OF CURV ES
541

Ex. 4. Find the envelope of circles whose centres lie on the rectangular
hyperbola .y = c 2 and which pass through its centre.

Let the eqtiation of a circle having centre at (a, 3) and passing

through the centre of sy = c 2 , which is the origin here, be

x2 +)' 2 -2ax-2v0 ... (1) where (Z=c2 ... 2)

This is the case of two-parameter family, where the parameters are
connected by a given relation.
Following the first method of Art. 17.8, ie.,eliminating 0 between (I)
and (2), the equation of the circle becomes

x2+y2_2ax_2ivo, since from (2), =f

or, 2a2xa(x2+y2)+2c2y,,,0
by Art. 17.7, the required envelope is
(2 +
= 4.2x.2c 2 v = l6c2xv.
Note. By transformation to polars, this quaLion can be shown to be
transformed to r2 =8c2 cns2, where 7r—O: i.e., the required
envelope is a lemniscate. -
Ex. 5. Find the envelope oft/ic parabola

where ab = k 2 , a and b being variable parameters.

We have .

ab=k2.
We apply here the second method of Art. 17.8. Taking differentials of
both (1) and () with respect to a and b, we have

IrX da+J-idb=O,
. . .(3)

do db
—+—=o.
a b
 DIFFERENTIAL CALCULUS
542

From (3) and (4)

ifi+f
or Va '4bVn_]A.._! from (1)
I 2 2
1 I I
a .

?.e..

Squaring and using (2), we get (a,.', )-eiiminanL l6xv = k 2 and hence,
this is the required envelope. This obviously represents a hyperbola.

Ex. 6. Find the pedal of the cardioide , = a (I + cos 0 ) with respect to the

pole (origin).
We shall here find out the fi rst positive pedal by considering it as the
envelope of the circles described oil radii vectors as diameters.
I See Gor., Art. 17.51

Let (p, a) be the polar co-ordinates of any point on the cardiolde.

pa(l+Cos6). .. (I)
Then
Again, the equation of the circle oil radius vector j, as diameter is

r=pcos(0—a). .. .(2)
... (3)
or, r=a(l+cosa)cos(0_a)fr0mU>
Here a is the parameter.
Differentiating (3) with respect to a,
0= sinacos(0_a)+(1+cds(x )sin (0)

sin acos(0 .-a)—cosa sin (o a )= sin(0—a)

i.e., sin (2a_0) sin (O—a),

i.e., 2a-0=0—U, i.e., 0,

Substituting this value of a in (3), we have the required envelope as

r= a
l+cosa)cos0=2ac30.
( or, II

Ex. 7. Show that the pedal equation of the envelope of the line
xcos 2a+ y s i n 2a = 2a COS a,
_ ,2 )
where a is the para?neiel; is 1) 2 = (r
ENVELOPE OI CURVES 543

Let (., y ) be the co-ordinates of any point P on the envelope.

Then x, y satisfy the equations I ( .x, y, a )= o,

aa
=
i.e., xcos2a-i-vsin2a=2a COS a, .
COS
x sin 2a—y 2a=a sin a,
From the definition of the envelope, it follows that (I) is the tangent to
the envelope at P ( x. y).
Let p he the length of the perpendicular from the origin 0 upon the
tangent (I) to the envelope at P and r be the distance of P from 0,

= 4a2, COS 2 a. . . (3)

I
r 2 =x 2 +v 2 =4a 2 cos 2 a-f a 2 sin 2 a,

Squaring and adding (1) and (2)j

=3a2 COS ?a+a2.
Eliminating a between (3) and (4), the required pedal equation of the
envelope is obtained.

EXAMPLES - XVII (B)

1. Find the envelopes of the following curves, a being the parameter:
(I) circles (x —a )2 + y 2 — 4a = 0,
(ii) parabolas ay 2 =2x+12a7,
(iii) ellipses r 2 + a 2 y 2 =4a.
2. Find the envelopes of the family of curves, 0 being the parameter.
(0 x 2 cosO + y 2 sin = a2

(ii) P(x,v )cpsO + Q(x,y )sinO '= R(x.y ).

(iii) A (x, y )cos" 0 + B(x,y )sin"0= C(x,). ).

3. Find the envelope of the family of curves

LA3 +. 3M) 2 + 3NA + P = 0-

where A is a parameter and L, M, N, P are functions of x and y.
4. Show that the envelope of the family of ellipses, ((x being the parameter)

ax 2 sec 4 a+h 2 v 7 cos eca =(a 2 —b2).

is the evolute of the ellipse x 2 /a 2 + y2/a2 =

544 DIFFERENTIAL CALCULUS

5. Find the envelopes of the family of circles which arc described on the
double ordinates of
(i) the parabola y 2 = 4ax as diameters.
(ii) the ellipse x 2 1a 2 + y 2 /a2 = 1 as diameters.
6. Find in each case the envelope of circles described upon OP as
diameters, where 0 is the origin and I' is a point on
(i) the circles x 2 + = 2ax,
(ii) the parabolas y 2 = 4ax,
(iii) the ellipses b 2x 2 + a 2 y 2 a2b2

(iv) the rectangular hyperbolas y = c2

7. If the centre of a circle lies upon the parabola y 2 = 4ax and the circle
passes through the vertex of the parabola, show that the envelope of
the circle is y 2 (2a + x )+ f = 0.
x2 y2
8. Find the envelopes of the family of ellipses —s-2 + =
a
whose sum of semi-axes is constant =
(ii) whose area is constant (= c).
9. Show that the envelope ofthccircles x 2 1- y 2 - 2ar - 20y + = 0,
where a, P are parameters and whose centres lie on the parabola
4av,is x (x + y2 - 2ax ) = 6.
10. Find the envelopes of

(i) of the family of ellipses L + = I and

a b2

(ii) the family of parabolas +fy = I,

where a" + = c" (a, b being the parameters).
11. Show that the envelope of the ellipses
2+
(x - a )
a2 b2
where the parameters a, 0 are connected by the relation
P2 x
i s the ellipse - + L
a2
+ = I. is
a b2
ENV ELOPE OF CURVES 545

12. Find the envelope of the family of curves + = 1, where th

parameters a and b are connected by the equation a 1' + b" = c1'.

13. Find the pedal with respect to the pole of the curve r2 = a 2 cos 20
14. Find the envelope of the circles described on the radii vectors of the
curve = a cos mO as diameters.
15. Show that the pedal equation of the envelope of the line
xcosnux+ ysinmcr = acosna, (m n ),whemctistheparameter,
• , m2r2-n2a2
M2 2 — fl

16. Given that the_astroid x ,+ y = c 3 is the envelope of the family of

theIlipses - + = I, where a and bare parameters, show that a

a b
and bare connected by the relation a +b = c.
ANSW ERS

1. (i) Y 2 -4x-4 =0. (ii) y2 = ±lSx. (iii) xv = ±2.

2. (i) x4 + y 4 = a4 . ( ii) P2 + Q2 = R2.
-
(iii) A 2- "' +B 2- " = C2-".
3. (MN_LPy=4(MP_N2)(u.I_M2).

5. 0 2
( x + a). (ii)
a2+b2 +
)2
. (i) ( +Y 2 - a 2 (2 + y2 )
(ii) x(v2+y2)+ay2=o. clii) a2x2+b2y7=(x2+y2)2.

(he) (x + 2 )2 = 4c2 AY

8. (1 ) x 13 + Y 13 = L (ii) • 2xy = c2.

10. (i) x+y 2 =c. (ii) x+y' =c

"p Ir_
12. X T-
-, + Y . +,
C
.+p 13. r' =a 3 cos .-8

14. r"=a"cosnO where pim/(m+l)

35
 18.1. Pedal curve.
The locus of the foot of the perpendicular drawn from a fixed point on.
the tangent to a curve, is called the pedal of the curve with regard to the
fixed point.
(i) To find the pedal with regard to the origin of any curve whose cartesian
equation is given.
Let the equation of the curve be f(x, y ) = 0. . . .(l)
Let x cos a + y sin a = p be the equation of the tangent PT to the
curve at any point P.
Now, the condition that the line x cos a + y sin a p should touch
the curve is of the form Ø ( p,a ) = 0. . . . (2)

Fig 18.1.1
Since (p, a ) are the polar co-ordinates of the foot of the perpendicular
Non the tangent PT, hence in (2), ii r, 0 are written for p, a, the polar
equation of the locus of N, i.e., of the pedal curve will be obtained as
Ø(.r, o)=o
which can now be transformed into cartesian.
Alternative Method:
Let (x. y) be the co-ordinates of P; then the equation of the tangent
PTis
dv,
Y—y= ---tX — x)
 ASSOCIATED wci

and the equation of ON, which passes through the origin and is perpendicular
to PT, is

Hence the locus of N, (i.e., the pedal) which is the intersection of (I)
and (2) is obtained by eliminating x and y from (1) and (2) and from the
equation of the curvef(x, y) 0.

(ii) To find the pedal with regard to the pole of am' curve whose
polar equation is given.
Let the polar equation of the curve be f(r, o)= o,
and let (ii. o ) be the polar co-ordinates of the foot of the perpendicuhtrN
drawn from 0 o the tangent at P r, o ).

Nowf denoting LOPN, tan r dO ... (2)

I
-- Fig 18.1.2

A lso 0 = LX OP=LX ON+.LNOP=9 1 + .4,r -

and since ON = OP sin it,, = r sin 0, , .. (4)
If r, 0, be eliminated from (1), (2) (3) and (4), a relation between r and
0, will be obtained, and from this relation, by dropping the suffixes, we get.
the required polar equation of the pedal.

(iii) The circle on radius vector as diameter touches the pedal.

LX ON = LP1X - LONT. I See Fig. of(ii) J
 548 DIFFERENTIAL CALCULUS

Fie 18.1.3

O =#--,r; also p=ON=ij.

Again, dP = rcosØ. See E%. 7, § 13.171

dqi
1f 1 be the angle between the tangent NT, and the radius vector ON of
the pedal at any point N (i.e.,ZONT1 = 0 1 ) then

dO 1dO1 d VI dp
tanØ4=r—=r—•—.—
'di 1 4 djí dp dj

rsinØ = tan 4l
dp rcos4i

010,i.e.,LONT4=ZOPN.

TI N touches the circle passing through OPN.

Hence the result.
•(iv) If p 1 be the perpendicular from the pole on the tangent to the
pedal, then p 1r = p2.
Draw O7'4 perpendicular from 0 on the tangent NT, to the pedal.
ASSOCIATED LOCI 549

Since LOPN LONT1 , :. N OPN, ONT are similar.

• OP
_ ON . r p .
-, i.e., -p = -.. i.e.. p1 r = p 2
ON -

(v) To find the pedal of a curve when its pedal equation is given.

Let the pedal equation of the curve be p 1(r)

and let p, r denote the usual entities of the original curve and p 1 , r1 the
corresponding things of the pedal curve.

Then p ,j ;also from above, r= - zJ-.

Pt Pt
Hence from equation (I), we get

the pedal equation of the pedal curve is

(r2

"-p

Note. If there be a series of curves designated as

P,PI,P2, ...... .

such that each is the pedal of the one which immediately precedes it, then
PI , P2 ........ I, are called the first, the second,...., the n" positive pedal of

P. Also regarding any one curve of the series, say P3 , as the original curve,

the preceding curves P2 , P1 , P are called respectively thefirst, second and

third negative pedals of 1.

18.2. Inverse curve.

If on the radius vector OP (or OP produced) from the origin 0 to any
point P moving on a curve, a second point Q be taken such that OP. OQ =
a constant, say P. then the locus of Q is called the inverse of the curve
along which P moves, with respect to a circle of radius k and centre o, or
briefly with respect to 0.
 550 DIFFERENTIAL CALCULUS

(13 To find the inverse of a given curve whose cartesian equation is

given.

Fig 18.2.1

Let (x, y) be the co-ordinates of any point P on the curve

f(x,y)=O and let Q(x',y')bea point onOp such that OP. OQ=k2.
Draw PM, QNperpendiculars on OX.

Now, - = = (. A' OPM, OQN are similar )

x' ON OQ

_OP.OQ k2
0Q2x'2 +

2 ' 2'
k x .. ky
12 " Similarly,
x +y- = '2 '2
x

Since j( x, y ) = 0. f ( k 2x' k2y'

0.
+ '2 + ,12
J=
Hence, by dropping the dashes, i he equation of the inverse curve is

(_k2x k2y )
'1 2 2' 2 2 -
s x +y x +y

i.e., the equation of the inverse of a curve is obtained by writing

k24(x2 + y2 ), k 2 y/( x 2 + y2 ) for x, y in the cartesian equation of hte
curve.
ASSOCIATED LOCI . 551

(ii) To find the inverse of a given curve whose polar equation is

given.

Let f(r,9 )= 0 be the equation of the given curve and let (r,0 )be

the co-ordinates of P and (r, 0 ) be the co-ordinates of Q.

Since OP. OQ = k2, rr' = k 2. .. r = k2/r'.

( k 2"
Again, since f(r,0 )= 0, . 0 =0.
f --,
:1. r )
Hence the polar equation of the inverse curve is

( k2
f t—,
r
01=0.
)
Thus, the equation of the inverse of a Curve is obtained by writing
k 2 /r for r in the polar equation of the curve.

(iii) Tangents to a curve and its inverse are inclined to the radius
vector at supplementary angles.

Let denote the angles between the tangents and radius vetor at
the corresponding points of a curve and its inverse.
rdO
Then, tan = tan = r dO
-
dr dr

I1 2 \ dr' d (k2' k2 dr
Now, k;rr=k

dO k 2 ( ? )de dO
tan = r - =— i -- i—=–r—=–tanø
dr r( k 2 )dr dr

= tan (2r–).
i.e., 0+0' =7r.

(iv) To find the inverse of a curve when its pedal equation is given.
Let the pedal equation of the given curve be

p=f(r).
Let p, r, denote the usual entities of the original curve, and let
P', r', 0' denote the corresponding things of the inverse.
 552 ,DIFFERENTIA L CA LCULUS

Then rr'=k 2 , i.e., r=k2/r'.

A lso, = . p
7p =SInØ sin(2z—)=sinØ =-.

r .,
r;P--;1p
r. r -
from equation (1), we get
• k2, (k2
-p'=fI---).
r
Hence the pedal equation of the inverse curve is

r2 (k 2
Pf ('.

183. Polar reciprocal.

If on the perpendicular ON (or ON produced) from the origin on the
tangent at any point P on a curve, a second point Q be taken such that
ON. OQ = a constant (say k2), then the locus of Q is called the polar
reciprocal of the given curve with respect to a circle of radius k and
centre 0.
From the definition, it follows that the polar reciprocal of a curve is
the inverse of its pedal. Hence the equation of the polar reciprocal of a
curve can be obtained by the first finding the pedal of the curve and then its
inverse.
Let N/V1 be the tangent to the pedal at N and let QM be the tangent to
the polar reciprocal at Q meeting OP produced at M.

Fig lltJ.I
ASSOCIATED LOCI 553

Now, 4' = IOPN = ZONN 1 E by § 18.1 (iii)]. Since QM isthe tangent

to the inverse of the pedal, hence by § 18.2 (iii),
.LOQM = LON,V 1 = LOPN.
Hence the quadrilateral PMQN is cyclic.

OM. OP=OQ.ON=k2

Also, LPNQ 900 ZPMQ = 90 0 ,i.e., OM is perpendicular to QM.

Hence the locus of P. i.e., the original curve is the polar reciprocal of the
locus of Q, i.e., of the polar reciprocal. Thus, the polar reciprocal of the
polar reciprocal of a curve is the curve itself.

18.4 Illustrative Examples.

Ex. 1. Find the Pedal of the parabola y 2 = 4ax with respect to the vertex.
The condition that X cos a + Y sin a = p will touch the parabola
y 2 = 4ax is obtained by comparing the equation with the equation of the
tangent at (x, y) to the parabola, i.e., with
Y y = 2a (X + x), or —2aX+Yy=2ax.
Hence

y=-2atana, x=— pscca.

cosu sina p

Since . Y 2 =4ax, .. 4a2tan2a=-4a.pseca.

Hence the required condition of tangency is p + a sin a tan (x = 0.
the polar equation of the pedal is
r+asin01an0=0
or,- r2+arin0tan6r0.
Writing ,.2 =x 2 +y 2 , rsinO=y and tanOy/x, we get cancsian
equation of the pedal as
x(x2+v7)+ay2=o
Alternatively
y=m.c+a/rn ... (1) i.a tangent to the parabola y 2 =4ox
v = - (I / m )x ... (2) is a equation of the perpendicular from the
origin on the above tangent.
554 DIFFERENTIAL CALCULUS

the locus of the point intersection of (1) and (2), i.e., the locus of
the foot of the perpendicular, i.e., the equation of the pedal is obtained by
eliminating ni between (1) and (2).

From (2), 'n

X x2 a
= --; substituting in (1), Y=
Y y x
Le., xp2 -s-x 3 +ay 2 =0. Hence the result.

Ex. 2. Show that the pedal of the circle r = 2a cos 8 with respect to the
origin is the eardioide r=a(1+cos8).
Since the given equation is r = 2acosO, .. (1)
Idr 2a cos 0 I
tan=rI— = -O cot 8= tan L7t+9
2
I d9 -2asrn

Let (r1 , 01 ) be the co-ordinates of the foot of the perpendicular; then

as in (2)of 18.1.

0=0 1 +7t- = 0 +-(-n+0);

(3)

Again, r 1 =rsinq=2a cos0.sin(7t+8)

[front and ( 2) I
=2a COS 2O=2aCos2 _LO t=a(I+cos0i)

Hence, the required locus is r = a(1 + cosO ).

Ex. 3. Show that the inverse of the straight line ax + by + c = 0 is a circle.

k x
Writing and ' for x, y in the given equation of the
x- +y
2
.x 2 +y
straight line, the equation of the inverse is

k2 k
2
a i-b +c=O
x2 +y 2.z 2+y 2

Or, c(x 2 +y 2 )+ak2x+bk2y=O

which obviously represents a circle.

ASSOCIATED LOCI 555

Ex. 4. Find the inverse of the parabola r =1/ (i + cos o).

Writing k 2/r fort in the equation of the parabola, the equation of the
inverse is
2 2
-= COS ,or, r=—(l1-cos0)=a(1+cosO)
r 1+ O 1
[ k
where a = -

which represents a cardioide.

Ex. 5. Find the polar reciprocal of the parabola y 2 = 4ax with respect to
the vertex.

The pedal of the-parabola with respect to the vertex is

x(x2 +y2 ),y2_ø [SeeEx. 1]
Its inverse is

k2x I k 4 x 2k4y2 1 k4y2

H-a =0.
x2+3,21(2+y2T +(x2 +,V2 (x2+y2

or, k 2 x+ay 2 =0,i.e., y2(k2/a)x

which represents a parabola.

Ex. 6. Find the polar reciprocal of the curve p = I (r).

r' ) _
By Art. 18.1(v), the pedal equation of its pedal is r = I (
p

Now, to obtain its inverse, writing - for,' and k2 p' forp

r r'
see §1&2 (iv) J, we get
4
k 2 =f ( k r'2
r r'2 kp

Hence on simplifying and dropping the dashes, the pedal equation of

the polar reciprocal is
= f ( 2
k2
556 DIFFERENTIAL CALCULUS

EXANHUES-XVIU

1. Find the pedals of

(I) theeilipse- + = 1 with respect to the centre and focus.

(ü) the parabola y 2 = 4ax with respect to the focus.

2. Find the equations of the pedals of the following curves with respect
to the origin:

(i) X + y" = a". (ii) JI = i.

+.

a"
3. Show that the first positive pedal of the rectangular hyperbola
- )' = a 2 with respect to the centre is the lemniscate
r 2 = a 2 cos 29.

4. Find the pedals with' respect to the pole of the curves:

(i) r2 COS 2O=a2. (;I) r 2 =a 2 Cos 2O.
r=a(l+ cos 9). [C.P.2006]

(iv) r=ae!"°".
(v) r' = a" cos mU.
5. Show that the pedal of a circle with respect to any point is the curve
r a + b cos B ,where a is the radius of the circle and b the distance
of the centre from the origin.
6. Find the inverses of the following curves with respect to the origin.
(i) x 2 + y 2 =a 2 . (1i) x 2 /a 2 + y 2 /b2 = I.

(iii) r = a(l+ cos O). (iv) r = ae°".

7. Show that the inverses of the lines Zr + 3y = 4 and 3x - 2v = 6 are

a pair of orthogonal circles.

8. Show that the inverse of the conic r = with regard to the

1+e cos 0
focus is a curve of the form p = a+b Cos 8.
ASSOCIATED LOCI 557

9. Show that the inverse of a rectangular hyperbola is a lemniscate, and

conversely.
10. Find the polar reciprocals, with regard to a circle of radius k and Centre
at the origin, of the curves:
(i) x2/a2 + y- lb' = I. (ii) y 2 = 4ax.
(Iii) r=acosO. (iv)r=a(J+ Cos O).
(v) r" = a' cos mO.

ANSWERS
1. (i) 2+ 4- 22 +b 2 ' x 2 +v =a 2(ii) x=O.

2. (I) a" (x' + Y. ) ( + )", where rn F n/(n - i).

(ii) (XI +y2)(ax+by)=absy.

where i=m/(m-.1).

4. C') r 2 =a2 COS 2O (ii) r =a Cos O.

(iii)r 4 _-(2a)i Cos ..O.

(iv) r = a,e Oc0 ( , where a 1 = osin a (br-a )ro.a

(v) r"=a" Cos nO, Where n=rn/(m+I).

6. (i) (X2+2)k412

(ii) (2 + p2)2 = k4(X2/,2 + ),21,,2).

(iii) r=b where bPia.

Uv) r = aj e OG ,where a 1 = * 2/

10. c a 2 x 2 +b 2 y 2 =k4.

b
(ii) a 2 + k 2 x = 0. (iii) r = ,where b = 2* 21a.
I + cos B
(iv) rCos9=(*2/2a).

(v) rcosnO=(&2/a)',wheren ml (m i- i)
FIR. CoNcAmy AND COW40M POINT OF INFLEXION

19.1. Concavity and Convexity (with respect to a given point).

Fig (i) Fig(ii)

Fig 19.1.1

Let PT be the tangent to a curve at P. Then the curve at is said to

be concave or convex with respect to a point A not lying on PT),
according as a small portion of the curve in the immediate neighbourhood
of P (on both side of it) lies entirely on the same side of PTas A [ as in
Fig. (I) ], or on opposite sides of PTwith respect to A (as in Fig. (ii) 1.

Fig 19.1.2

Thus, in Fig. 19.1.2, the curve at P is convex with respect to A. and

concave with respect to B or C. The curve at Q is concave with respect
to A. Again, the curve at R is convex to Band concave to C.

Note. A curve at a point P on it is Convex or Concave with respect to a

given line according as it is convex or concave with respect to the foot
of the perpendicular from P on the line'T
 CO,VCAVITYAND CONVEX/TYPO/NT OFINFLEXION 559

19.2. PoipAoflnflexion.

r.

A
Fig 19.2.1

In some curves, at a particular point Pon it, the tangent line crosses
the curve, as in Fig. 19.2.1. At this point, clearly the curve, on one side of
P, is convex, and on the other side it is concave to any point A (not lying
on the tangent line). Such a point on a curve is defined to be a point of
inflexion (or a point of contrary flexure).

19.3. Analytical Test of Concavity or Convexity

(with respect to the x-axis).

Fig (i) Fig (ii)

Fig 19.3.1

Let P ( x. y) be a point on the curve y = j (x), Q a neighbouring

point whose abscissa is x + h (h being small, positive or negative). Let
PT be the tangent at P. and let the ordinate QM
of Q intersect PTat R.
The equation to PT is

Y-y= -
560 DIFFERENTIA L CA LCULUS

and abscissa XofR being .x + h, its ordinate

RM = V y+hf'(x).
Also the ordinate of Q is
QM = f(x4-h)

= f(x)+ hf'(x)+ f* (x + Oh), 0<0<1.

QM—RM=P(x+Oh).

Now assuming f" (x) to be Continuous at P and * o there,

x + Oh ) has the same sign as that of f"(x) when I h i is sufficiently
small.
Hence from (I), QM— RM has the same sign -as that of f'(x), for
positive as well as negative values of /i, provided it is sufficientiy small in
magnitude.
Firstly, let the ordinate PN or y be positive.

Then if f'(x) (orat) is positive, from (l)QM>RM for Qon

either side of P in its neighbourhood, and so the curve in the
neighbourhood of P (on either side of it) is entirely above the tangent,
i.e., on the side opposite to the foot Non the x-axis of the ordinate PN,
as in Fig. (i). Hence, the curve at P is convex with respect to the i-axis.

Again f f "(x) is negative, QM < RM on either side of P, and so

the curve near P is entirely below the tangent, on the same side of N, as
in Fig. (ii). Hence the curve at P is concave to the x-axis.
Secondly, let y or PN be negative.

FIg(i) Fig (ii)

Fig. 19.3.2
CONGA i'ITY.4ND CON VEXITY POINT OF 1NFLEKION 561

if f" (x) is positive, from (1), as before, QM> RMon either side of P,
and as both are negative, QM is numerically less than RM, as in Fig. () of Fig
19.3.2. The curve, therefore, at P lies on the same side as N with respect to
the tangent PT Hence, the curve at P is concave with respect to the x-axis.
If f" (9 is negative, we similarly get the curve at P convex with
respect to the x-axis,as in Fig. (ii) of Fig 16.3.2. Combining the two cases,
we get the following criterion for convexity or concavity of a curve at
point with respect to the x-axis:

If yis positive at P, the curve at P is convex to the x-axis.

dxz

1fY d2y is Negative at P, the curve at P is concave to the x-axis.

dx

Note. At a point where the tangent is parallel to the y-axis, is infinite. At

dx
such a point, instead of considering with respect to the x-axis, we investigate
cOnvexity or concavity of the curve with respect to the y-axis. The criterion,
as obtained by a method similar to above, is as follows:
The curve at P is'convex or concave with respect to they-axis according
d2 x .
as x W is positive or negative at P.

19.4. Analytical condition for Point of Inflexion.

Fig /9.4.!

In the above investigation, let /' (4 = 0 at P, and f (x) # 0.

36 Tn, QM=J(x +h)f (X )+hf ')f (x +9h)
562 DIFFERENTIAL CALCULUS

QM_RM=!f '(x+Oh).

and the sign of this for sufficiently small I h I is the same as that of

f ( x ), which has got opposite signs for positive and negetive values
of h, whatever be the sign of f (x) at P. Thus, near P the curve is above
the tangent on one side of P. and below the tangent on the other side, as in
the above figure. Hence, P is a point of inflexion.
Thus, the condition that P is a point of inflexion on the curve y = f (x)
is that, at P,
A2, d3
= 0 and t- 0.
dx 2dx3

Note. If is infinite at P. the condition that P isa point on inflexion is th.it,

dx
at P.

and
dv 2dy3

19.5. A more general criteriop.

= in_i (4=0 and
Suppose that at I 14= f - (X ) = ...... .

f"x)* 0.

Then, QM = f ( x + Ii ) .= f (4 + hf' (4 + n! " (x + B h).

[o<o <i}

QM - RM = f" (x + 0 h ) which, for sufficiently small

values of j hi. has the same sign as that of f" (x).

If n is even, h" is positive and the sign is the same as that of f" (

or at P for both positive and negative values of h. Considering both

the cases when y of P is positive and negative, we find that the curve at P is
convex or concave .vith respect to the x-axis according as y is positive
dx
or negative.
CONCAVITYAND CONVEXITY POINrOF INFLEXION 563

1f n is odd, (x ) will have opposite signs for gsil e.

n!'
negative values of h, whatever be the sign of (i). Hence, Q ls
opposite sides of the tangent for positive and negative values of/i. Thus, P
is a point of inflexion.
Note. Since from (1)Art.19.3, QM–RM =-- f(s+O/r),iff(x+O/r)
has opposite signs for opposite signs of It when I It is sufficiently small,
QM > RM on one side, and QM <RM on the other side of Pon the curve
in the immediate neighbourhood, and thus the tangent at P crosses the
culve at P. and so P is a point of inflexion. Thus, since q is positive and
numerically less than I, an alternative criterionfora pain, of in/le'.-ion is
that f' ( x + Oh ) should have opposite signs for opposite signs of It
Ii is numerically sufficiently small; in other words, f"( x ) change r igri
passing through P from one side to the other.

19.6. Illustrative Examples.

Ex. 1. Examine the curve y = sin x regarding its con vexitv or co,'v
the x-axis, and determine its point of inflexioir, if any

– dy
As y sin x, —=cosx and
dv
--sinx.lience, v– -2tn x
dx
which is negative for all values ofx excepting those which mak = 0,
i.e., for s = , k being any integer, positive or negative.
Titus, the curve is concave to the x-axis at every point, excepting at
points where it crosses the x-axis.

At these points, given by x=kit, =0, and '

Hence, those points Where the curve crosses the x-axis are points of
inflexion. -
Ex. 2. Show that the curve y3
= 8x 2 is concave to I/re foot oft/re ordinate
everywhere except at the irin.
From the given equation, v = 2x'.
1. 4
a"
dx 3 It ()

- 1
y----.
(L1 'I . - =--
564 DIFFERENTIA L CA LCULUS

Thus, excepting at the origin, x being positive for all values of x,

is negative.

Hence, the curve is concave everywhere to the foot of the ordinate

excepting at the origin.

Ex. 3. Prove that (a —2, - 2/e2 ) is a point of inflexion of the curve

I \
yx — a)e
Here, at points on the curve,

dy
dX

=e +(1+x— a)e =(2+x—a )e.

dy
and, similarly, _-3 -=(3+x--a )e

Hence, atx=a-2,(where — 2e 2 ), O and .e_2 #O.

dx dx3
Hence, the point (a —2. —2/e ) is a point of inflexion.

Ex. 4. Find if there is any point of inflexion on the curve

y-3=6(x-2)5.

Here, =3O(x-2) ,-

Thus, wheix=2 (and so y=3).

In the neighbourhood of this point, where x = 2 + I, (h numerically
small, positive or negative),

=120h 3 .which has opposite signs for positive and negative

values of h. Hence, changes sign in passing through x = 2.

Thus. (2,3) is the only point of inflexion.

CONCA VITY AND CONVEXITY POINT OF INFLEXION 565

Alternatively

Here, _236O(x_2)2, "y 720(x-2), .272O.

d3 dr4 dx5

Thus at x =2, = = O and * 0.

•• di? dx3 dx4
(which is of odd order)
Hencex = 2 gives the point of inflexion.
EXAMPLES -XIX
1. Prove that the curve y = e is convex to the x-axis at every point.
2. Prove that the curve y = cos' x is everywhere concave to the y-axis
excepting where it crosses the y-axis.
3. Prove that the curve y= log x is convex to the foot of the ordinate in
the range 0 <x < 1, and concave where x> 1. Prove also that the curve
is convex everywhere to they-axis.
4. Show that the curve (y—a) 3 = a 3 - 2a 2 x+ax 2 where a>0, is always
concave to the x-axis. How is ii situated with respect to they-axis?
5. Show that the origin is a point of inflexion on the curves:
(i) y=x 2 1og(l—x), (ii) yrxcos2x.
6. Find the points of inflexion, if any, on the curves:
X
(ii) y2 =x(x+1.) 2
(x+1) 2 +1
.
(iii) cy = (x— a ) 3 (iv ) y = ae

7. Show that the points of inflexion on the curve y 2 = ( x — a ).2 (x— b)

lieonthe line 3x+a=4b.
8. Show that the curve y(x 2 + a 2 )=a 2x has three points of inflexion
which lie on a straight line.
ANSWERS
4. Concave where 0 < x < a. convex everywhere else.

6. (I)

(ü) (, ± Jfl. (• iii) (a, 0),. (iv) (± 4, ae ).

 ON SOME WELL-KNOWN CURVES
FAI
20.1. We give below diagrams, equations, and a few characteristics of some
well-known curves which have been used in the preceding pages in obtaining
their properties. The student is supposed to be familiar with conic sections
and graphs of circular functions, so they are not given here.

20.2. Cycloid.
The cycloid is the curve traced out by a point on the circumference of
a circle which rolls (without sliding) on a straight line.

OL M D 0 X
t=a(O- sin O) v=a(1- cos O)
Fie 2O.2.1
Let P be the point on the circle MP called the generating circle, which
trêesoiit the cycloid. Let the line OMX on which the circle rolls be taken as
x-axis and the point 0 on OX, with which P was in contact when the circle
began rolling, be taken as origin.
Let a be the radius of the generating circle and C its centre, P the
point (x, y) orrit, and let zPCM = e . Then 0 is the angle through which the
circle turns as the point P traces out the locus.
OM = arc PM = all

Let PL be drawn perpendicular to OX.

x=OL=OM-LM=aO-PN = aO-a sin ll
= a(O - sinU ).
y= PL= NM =CM -CN =a - acosll

=a(l- cos ll).

Thus, the parametric equations of the cycloid with the starting point
as the origin and the line on which the circle rolls, called the base, as the
x-axis are
ON SOME WELL-KNOWN CURVES 1
567

x=a(O—sinO), y=a(1—cos0). .
The point . A at the greatest distance from the base OX is called the
vertex. Thus, for the vertex, y, i.e., a (i - cos 8 ) is maximum. Hence,
cos O=-1,i.e.,O=n.
A D=a(l— cos n)=2a.
vertex is (an ,2a).
For O and O', y=O. .. cosO = I. :. 0 = Oand 2n.
As the circle rolls on, arches like OA O' are generated over and over
again, and any single arch is called a cycloid.

x=a(O+ sin O) A T y=a(t— cos 9) x

Fie 20.2.2
Since the vertex is the point (an, 2a ) the equation of the cycloid
with the vertex as the origin and the tangent at the vertex as the x-axis can be
obtained from the previous equation by transferring the origin to an 2a )
and turning the axes through it, , i.e, by writing
air + x'cosn - y'sin ir and 2a + x'sinn + y'coslr
for and y respectively.
Hence, a (o - sin 0 ) = air -x'
or, x'=a(n— O )+a sin O
= a(0'+ sill O'), where 0' = n — U
and a(l— cos 0 )= 2a—v'
or, y' = 2a —a+a cos; 0 = a+a cos U
= a— a cos (ir -o )= a(l - cos 0').
Hence, (replacing 0' by 0) the equation of the cycloid with the vertex
as the origin and the tangent at ehe vertex as the x-axis are
x=a(O+ sin o), y=a(— cos O)
568 DIFFERENTIAL CALCULUS

In this equation, 8 = 0 for the vertex, 9 = ir forO, and 0 = - 7r for 0'.

The characteristic properties are
(i) For the cycloid x = a(0— sin ), y = a(1 — cosO ), radius of
curvature = twice the length of the normal, (the centre of curvature and the
x-axis being on the same side of the curvature).
(ii) The evolute of the cycloid is an equal cycloid.
(iii) Forthecycloid x = a(O+sin0 ), y = a(1—cosO ),
=- 0 and 2 = 8a y s being measured from the vertex.

Note. The above equation (2) can also be obtained from Fig. (i) geometrically
as follows
If (x', ') be the co-ordinates of!' referred to the vertex as the origin
and the tangent at the vertex as the x-axis.
= LD = OD - OL = - x = a — 0 )-- a sin 0,
y'= A D— PL = 2a - y 2a — a(l — cos 0 )= a(1 +cosO),
Hence, writing 0'(or0 )ibr it-0,etc.

203. Catenary.
The catenary is the curve in which a uniform heavy flexible string will
hang under the action of gravity when suspended from two points. It is also
called the chainetle.

OT N X

Fig 20.3.1
Its equation, as shown in books on Statics, is

y=ccosh[er+erJ.
ON SOME W ELL-KNOW N CURV ES 569

C is called the vertex, OC = c, OX is called the directrLr.

The characteristic properties are
(i) The perpendicular from the foot of the ordinate upon the tangent
at any point is of constant length.
(ii) Radius of curvature at any point= length of the normal at the point
(the centre of curvature and the .x-axis being on the opposite sides of the
curvc).
(id) y 2 = c 2 + s 2 $ being measured from the vertex C.
(iv) s = c tan 11, y = c sec V .

(v) .i=c log ( sec v+ tan yi).

20.4. Tractrix.
Its equation is

x = a2 —y2+—log
a
2 a+a2_y2

or, x=a(cost+Iogtant), y = asint.

Here, OA = a.

ox
Fig 20.4.1
The characteristic properties are
(I) The portion of the tangent intercepted between the curve and the
x-axis is constant.
(ii) The radius of curvature varies inversely as the normal (the centre
of curvature and the x-axis being on the opposite sides of the curve).
(iii) The evolute of the tractrix is the catenary
Y = a cosh (x/a ).
570 DIFFERENTIA L CA LCULUS

20.5. Four cusped Hypo-cyctoid.

V

Fig 20.5.1

2
(x' 3 y
Its equation is I - I + -1
a) [bJ -.

or. x = acos 3Ø, y = bsin'.

Hcrc, OA = OA ' =a; OB=OB'=b.
The astroid is a special case of this when a = b.
20.6. Astroid.

Fig 20.6.1

2 2 2
Its equation is x 3 + y' = a'
Or, x=acos38,y=a sin 'O
ON SOME W ELL-KNOW N CURVES 571

Here, OA =OB=OA '=OB'=a•

The whole figure lies completely within a circle of radius a and centre
0. The points A , A ', B, B' are called cusps. It is a special type of a
our-cusped hypo-cycloid. (See § 20.5 1
The characteristic property of this curve is that the tangent at any
point to the curve intercepted between the axes is of constant length.

20.7. Evolutes of Parabola and Ellipse.

(i) The equation of the evolute of the para' )la
= 4ax i

27ay 2 =4(x.-2a)3.
This curve is called a semi-cubical parabola.

Fig 20.7.1

• Transferring the origin to (2a, o), its equation assumes the form
2 = Ia where k 4 / (27a ), which is the standard equation of the
semi-cubical parabola with its vertex at the origin.
Hence, the vertex C of the evolute is (2a. 0 ).
(ii) The equation of the evolute of the ellipse

X 2/a 2 + y 2/b 2 = I is

(ax) 3 +(by) b2 )I
572 DIFFERENTIAL CALCULUS

Fig 20.7.2
which can be written in the form

I
(X) fl)
where a=(ä2_b2)/a, J3=(a2_b2)/b.
Hence, it is a four-cusped hypo-cycloid.

20.8. Folium of Descartes.

Fig 20.8.1

Its equation is x 3 +y 3 = 3axy

It is symmetrical about the line y = x.
The axes of co-ordinates are tangents at the origin, and there is a loop
in the first quadrant.
It has an asymptote x + y + a = 0 and its radii of curvature at the

origin are each = 1 a.

ON SOME WELL-KNOWN CURVES 573

20.9. Logarithmic and exponential curves.

(i) The equation of the Logarithmic curve is y = log x. x is always
positive; y = 0 when x = I and as x becomes smaller and smaller, y,
being negative, becomes numerically larger and larger. For x > 0, the curve
is Continuous.
PA

Y.

oil x C

(i) y=logx
Fig 20.9.1

(ii) The equation of the Exponential curve is y e. x may be positive

or-ncgative but y is always positive, and v becomes smaller and smaller.

I' rig zu. IV. i

as x, being negative, becomes numerically larger and larger. The curve is

continuous for all values of x.

20.10. Probability Curve.

The equation of the probability curve is y e
The x-axis is an asymptote.
The area between the curve and the asymptote is

= 2f e - " d, = 2. Ivrn=
574 DIFFERENTIAL CALCULUS

20.1I. Cissoid of Diodes.

Its cartesian equation is y2 (2a - x ) = x3.

VA

Fig 20.11.1
QA = 2a; x = 2a isanasymptote-
2asin2O
Its polar equation is r =
cos 0
20.12. Strophoid.
The equation of the curve is y2 =
U - X

OA =OB=a. OCBPisaloop.
x = a is an asymptote.

Fig 20.12A

The curve y 2 = . is similar, just the reverse of strophoid,

a +x
the loop being on the right side of the origin and the asymptote on the left
side.
ON SOME WELL-KNOWN CURVES 575

20.13. Witch olAgnesi.

2
The equation of the curve is .y 2 = 4a (2a -
Y

0 A X

Fig 20.13.1

Here, OA = 2a.
This curve was first discussed by the Italian lady mathcm1tician Maria
Gactaua Agnesi, Professor of Mathematics at Bologna.

20.14. Logarithmic (or Equiangular) spiral..

0),
Its equation is r = ae (or, r = ae"
where cot a or ?it constant.
Characteristic Properties
(i) The tangent at any point makes constant angle with the radii
vector (0 = a).
(ii) Its pedal, inverse, polar reciprocal and evolute area)] equiangular
sprials.
P

cI1)1X
Fig 20.14.1

(iii) The radi: of curvature subtends a right angle at the pole.

Note. Because of the property (i), the spiral is called equiangular
576 t. DIFFERENTIAL CALCULUS

20.15. Spiral of Archimedes.

Its equation r = aO.

Its characteristic property is that its polar subnormal is constant.
20.16. Cardioide.
Its equation is (i) r = a (1 + cos B ). or (ii) r = a (i - cos o ).
Jr. (i), B = 0 for A,and 0 = ir for 0.
In (ii), 0 = 2t for A, and O = ° for 0.
In both cases, the curve is symmetrical about the initial line, which
divides the whole curve into two equal halves and for the upper half. 0
varies from 0 to it, and On = 2a.
V V

K K

(i) r=a(l+ cos O) (i)r=a(I - cosO)

Fig 20.16.1
The curve (ii) is really the same as (i) turned through 180°.
The curve passes through the origin, its tangent there being the initial
line, and tangent at A is perpendicular to the initial line.
The evolute of the cardioide is a cardioide.
ON SOME WELL-KNOWN CURVES 577

Note. Because of its shape like human head, its is called a cardioide. The
cardioide r = a (i -,- cos 0 ) is the pedal of the circle r = 2a cos 0 with
respect to a point on the circumference of the circle, and inverse of the
parabola r=a/(l+ cos O)

20.17. Limacou.
The equation of the curve is r = a + b cos 0
When a> b, we have the outer curve, and when a < b, we have the
inner curve with the loop.
When a = b, the curve reduces to a cardioide. [Seefig. in § O. 161

Ri

Fig 20. 17. 1

Limacon is the pedal of a circle with respect to a point outside the
circumference of the circle.
20.18. Lemniscate.

2 2
Fig 20.18.1 r = a COS

Its equation is r 2 = a 2 cos 20

2 2
or, (.x2+y2)2
- =a (x -),).
578 DIFFERENTIA L CA LCULUS

It consists of two equal loops each symmetrical about the initial

line, which divides each loop into two equal halves.
Here, OA = OA' = a.
The tangents at the origin are y = ± x.
For the upper half of the right-hand loop, 0 varies from 0 to -1 it.
A characteristic property of it is that the product of the distances
of any point on it from each of the points (±a/ . fi, 0) is constant.

If

Fig 20.18.2

The lemniscate is the pedal of the rectangular hyperbola r.2 cos 28 = a2.
The curve represented by r 2 =a 2 sin 2O is also sometimes called
lemniscate or rose lemnisca:e, to distinguish it from the first lçmniscate
which is sometimes called lemniscate of Bernoulli after the name of the
mathematician J. Bernoulli who first studied its properties.
The curve consists of two equal loops, situH4ed in the first and
third quardants, and symmetrical about the line y = X . It is the first
curve turned through 45°.
The tangents at the origin are the axes of x and y.

20.19. Rose-Petals (r = a smug , r = a cosnO).

I)

r=a
B(2)
Fig 20. 19.1
ON SOME WELL-KNOWN CURVES 579

The curve represented by r = a sin 30, or. r = a cos 30 is called a

three-leaved rose, each consisting of three equal loops. The order in
which the loops are described is indicated in the figures by numbers. In
each case, OA = OB = OC = a, and LAOB = iBOC = ZCOA = 1200.
The curve represented by r = a sin 20, or, r = a cos28 is called
a four-leaved rose, each consisting of four equal loops. In each
case, OA=OBOC=aand LAOB= LBOC= LCOD =LDOA=90°.

D(4)
I A(i)

I B(2)
C(3) I
r=asin2O

Fig 20.19.2
The class of curves represented by r = a sin n 0, or, r = a cos n 0
where n is a positive integer is called rose-petal, there being t or 2n equal
loops according as n is odd or even, all being arranged symmetrically about
the origin and lying entirely within a circle whose centre is the pole and
radius a.

20.20. Sine Spiral (r' = a sin nO, or, r = a cos no).

The class of curves represented by (i) r" = a sin it or (ii)
= a' cos n 0 is called sine spiral and embraces several important and
well-known curves as particular cases.
Thus, for the values n -2, 2, - - ,and the sine spiral is
respectively a straight line, a circle, a rectangular hyperbola, a lemniscate, a
parabola and a cardioide.

For 0 = nO; for (ii)0 = 1 ir+n6

The pedal equation in both the cases is
p=r a
fl-I-I I n
DA . SINGULAR POINTS

21.1. Double Points.

If two branches of a pinc curve pass through a point P. that is, two
tangents at P can be drawn to the curve, then the point is called a Double
point on the curve.

21.2. Classification of Double Points.

Node : If the tangents at a double point P on a plane curve be real and
distinct, the double point is called a Node.

Fig 21.2.1
Cusp : If the tangents at a double point P on the plance curve be real
but coincident and the curve has real branches in the neighbourhood of P,
then the double point is called a Cusp.
Isolated Point: If the tangents at a double point P on a plane curve be
either non-real or real, coincident but the curve has no real branches in the
neighbourhood of P, then the double point is called an Isolated point.
21.3. Different types of cusps.
Single Cusp:

(ii)

Single cusp,first species Single cusp,second species

lfP be a cusp on a plane curve and both the branches of the curve lie on
the same side of the normal at P, then the cusp is called a Single Cusp.
SINGULAR POINTS 581

Double Cusp:

(ii)

Double cusp, first species Double cusp, second species

Fig 21.3.2
If a plane curve has branches on eiTher side of the normal at a cusp
P, then the cusp is called a Double Cusp.
21.4. Species of a Cusp.
Cusp of the Fist Species (or Keratoid Cusp):
If the branches of a curve at a cusp P (single or double) lie on either
side of thd tangent at F, then the cusp is of the first species or it is called
a Keratoid Cusp.
Cusp of the Second Species (or Ramphoid Cusp):
If the branches of a curve at a cusp P (single or double) lie on one
side of the tangent at P, then the cusp is of the second species or it is
called a Ramphoid Cusp.

Osculinflexion:
If a curve has double cusp at P and it is Keratoid on one side of the
normal at P and Ramphoid on the other, then P is called a Point of
osculinflexion.

Fig 21.4.1
Double cusp, change-of species
Osculinflexion
582 DIFFERENTIAL CALCULUS

21.5. Search for double points.

Let the given curve be represented by a rational algebraic equation.
Let (a, ) be a double point on the curve. We shift the origin to (a, fi)
through parallel axes. In the transformed equation the constant term and
the terms of the first degree must be absent in order that the new origin may
be a double point. We get three equations in a and 13. Take any two of them
to find a and 13 . If these va1ue of a and 13 satisfy the remaining equation,
then the point (a, fi) will be a double point.
If no such values of a, 13 are available to satisfy all these three equations,
then we say there is no double point on the curve.

21.6. CondItions for existence of double points on an algebraic curve.

Let f(x,y)=0
be a rational algebraic curve.
If we shift our origin to the point (a, p) through parallel axes, the
above equation transforms into
f(x+a,Y+)=o .
By Taylor's therom, we get

f(x+a, Y +fl)=f(a,fl)+x1.a—[)
ax)() t)(ap)
Y

+! X 2 I_L ) i-2XYI_
a2 Ia2f)
2 ( ax2 )(a.) axaY)(0fl)
Therefore, the equation (2) becomes

a ,p ) +
j( x1_L 1
dX )(ap)
^y1 L )
aY)(B)j 2
1+! f
a.x
(a.$)

2i) +y21)
+ 2XY l+...=o
aXaY) $)(aY2))]
...
(a, p ) will be a double point on (1) if origin is a double point on (3) which
requires

f(x,y)=o, [L ' =0and1- L 1 =0.

dXJ($)
SINGULAR POINTS 583

If these equations be consistent, the tangents at the new origin are -

given by
a + y2 I_L)
X 2 I_fI -i-2XYI _ ) =0
aX2 )(afi) axaY)(ap) ay2)0)
(4)
In general, the double point will be a node, cusp or an isolated point
according, as
( a2 j ' ( a2f (a2f
I >,=,or<I—I I-
aX a Y)(a p) (s ax aY2 (a.

21.7. fllustmve examples.

Ex. 1. Determine the existence and nature of the double points on the curve
(x-2) 2 =y(y-l)2.
The given curve is
f(x,y)n(x-2) 2 -y(y-1) 2 =0 ... (1)

=0 gives 2 (x - 2)= 0, i.e., x=2, ... (2)

dx)

.L1=o gives 2y(y-1)+(y-1)2 =01

1kaY)
i.e., (y-lX 3y.-1)=0,
i.e.. y=1,. ... (3)

Equations (2) and (3) say That the possible points are (2,1
The point (2,1) satisfies equation (1) but the point (2,1) does not.
So, there isadouble point at the point (2,1).
Let us shift the origin to (2,1) through parallel axes. Equation (1)
becomesX 2 (Y+1)Y 2 , i.e., Y3+Y2-X2=0.
The tangents at the new origin are given by
Y 2 - X 2 =0,i.e., Y =±X ,
which are real and distinct.
Hence, there is a node on the given curve.
EL 2. Examine the character of the origin on the curve
- y2 =2x2v+x3y+x3.
 584 DIFFERENTIAL CALCULUS

The given equation is

f(x,y)Ex3v+x322y._y2
(I)

Now, =3X2y+3X2+4y;
ax
af =X 3
+2x2-2y.
ay

At (0, 0), f (x, y)r 0, =0 and =0.

ax ay
So, there is a double point at (o, o) on the given curve.
Tangents at (0,0) are given by y 2 =0, that is, Y = O, y=O which
are real but repeated. The double point is either a cusp or an isolated point.
Take a point P(x, y) on the curve in the neighbourhood of (
0, 0).
Distance of P from the tangent y = 0 is given by
p=y (2)
Eliminating y between(1)and (2) we get
x 3 p+x 3 +2x 2 p—p 2 =0

i.e.. pz_(x3+2X2)PX3_o (3)

which is a quadratic in p whose discriminant is

(x3 +.2x2)2 +4x 3 =4x 3 +4x 4 +4x 5 +x6

4x 3 , since is small.
It is positive if x>0, that is, (3) can have two real roots ifx>0, that is,
(1)has two real branches on the right of they-axis in the neighbourhood of
(0, 0). So, there is a single cusp at the origin.
The product of the roots of (3) is - x 3 <0 for x >0, that is, the roots
of(3) are opposite signs, that is, the two branches of the curve lie on opposite
sides of the tangent y = 0. The cusp is of the first species.
Hence there is a single cusp of the first species at the origin.

Ex. 3. Is origin a double point on the curve y 2 = 2x 2 y + x4 y - 2x4 ? if so,

state its nature.
The given curve is
V2
f(x,v)= x 4y -2x+2
4 =0 (1)
aj
Now, +4.y;
 585
SINGULAR POINTS

+2x 2 —2y.
dy
At (o, o), we see that f (x, y )= 0 -0 and =0.
ax
So, there is a double point at (0, 0).
The tangents at origin are given by y 2 =0, that is, y = 0, y = 0
which are real and coincident. The double point may be a cusp or an isolated
point.
We take a point P(x, y) on the curve in the neighbourhood of the origin.
Distance 'of P from the tangent y = 0 is given by
p=y ... (2)
Eliminating y between (1) and (2) we get
x p . -2x+2x 2 p— p 2 =0,
i.e., 2_(2 X 2 +x 4 )p+2X 4 o ... (3)

which is a quadratic in p whose discriminant is

(2x2 +x 4 ) 2 —8x 4 =-4x 4 +4x 6 +x t —4x 4 <0 for all x,
that is, the roots of (3) are complex, that is, there is no real branch of the
curve in the neighbourhood of (0, 0).
Hence, the origin is an isolated point on the curve.
Ex. 4. Search for double points on the curve
x2y+x3y+5x4 ± y2.
The curve isf(x,y)5x4+x3y+x2y_y2=0

I32
= 0 gives 20x + 3x y + 2.y = 0,
ax
i.e., x(20x2 + 3xy + 2y) = 0: ... (2)

gives x 3 +x 2 —2y=O,

x2+x3 ... (3)

i.e., y —

Eliminating y between (2) and (3) we get

2O2+c2)+X2 +X3 0,

i.e., x3(40+3x(1+x)+2(l+x)}0
i.e., x 3 (3x 2 +5x+42) =0.
 586. DIFFERENT/A L CA LCULUS.

• —5±J25-504
i.e. x=Oor
6
-5±if
=Oor
6
Only x = 0 is acceptable as the other two values of x are complex.
When x=0,weget yO.
We see that (l)js satisfied with x=0, y=Q.
So (0,0) (0,0) is a double point on the given curve.
Tangentsat(0,0) are given by y 2 =O, that is, y=O, y0 which are
real but repeated, that is, the double point is either a cusp or an isolated
point.
Let us take a point P (x, y) on the curve in the neighbourhood of(0, 0).
Its distance from the tangent y = 0 is given by
py
Eliminating y between (1) and (4) we get
x2p+x3p+5x4=p2,

that is, p2 - (x2 + x3 )p — 5x 4 = 0 . . . (5)

which is a quadratic in p and so two branches of the curve exist in the
neighbourhood of (0, 0) depending on the roots of (5).
Its discriminant is (x2 +x3 )+2ox 4 =21x 4 >0 for positive as well as
negative values of x, that is, the curve has real branches on either side of
the y-axis. The origin is, therefore, a double cusp.
Product of the roots of (5) is -5x 4 <0 for all x. The roots are of
opposite signs implying that the branches lie on either side of the x-axis,
that is, the cusp is of the first species.
Hence, the origin is a double cusp of the first species.
EXAMPLFS - XXI
1. Examine the character of the origin on each of the following curves
(a>0):
(i) ay 2 = x 3 . (ü) a(y— ,)2 = x 3 . ( iii) (2a— .) y2 = x3.
(iv) x.4 - 412 y - 2xy2 + 4y 2 = 0.
(v) x - aA 2 y + axy 2 + a 22 = 0.
SINGULAR POINTS 587

(vi) x4 - 3x 2 y - 3.y 2 + 9y 2 = 0.
(vii) x4 - 2ax 2 y - axy 2 + a 2 y 2 = 0.
(viii) y 2 =2x2y+.4y+x4.
(1x)x4+x3y+5x32.x2y+x2_3xy+2y2=0

(x) 5x 2 y+x 3 y-5x4 y2.

(xi) (2x+y) 2 —6y(2x+y)-7x 3 =0.

2. Search for double points on each of the following curves:
2X I
(i) x4 + y + + 3y 2 = 0.
(d) y(y-6)=x2(x-2)3-9'.
(ii) x3—y2-7x2+4y+I5x-130.
(iv) x -2 Y3 - 3y2 - 2x2 + 1 = 0.

(v) X ( 2 y2):= ay 2 . (vi) x + y 3 - 3axy = 0.

(vii)y2=(x_l)(x_.2)2. (viui)y2=(x._2)2(x_5).

(ix) x 3 + 2x 2 + 2xy - y 2 + 5x - 2y = 0.
(x) (x+y)3=y(y—x+2)2.
()i) (2y++1)2.=4(1—x)5.
(xii) 1y 2 + 2a 2 y - ax 2 - 3a 2 X- 3a 3 0.
(xiii) x 4 + 4x3 + 2 Y3 t 4x 2 + 3y 2• —1 = 0.
(xiv) y 4 - 8y 3 12.y 2 + 16y 2 + 48xy + 4x 2 - 64x = 0.
3. Show that the curve (.zy+ 1)2 +(x—I)3 (x-2) = 0 has a single cusp of
the first species at (1, —1).
4. Prove that the curve ay2 = (i—a)2(x—b) has, at x = a, an isolated
point if a< b, anode if a>b, and a cusp if a= b.
5. Examine the nature of the point (0,—I) on the curve
- 2X2 - 2x2 - 2.y + y 2 - x + 2y + 1 = 0.

xI)
6. Examine the nature of the point on the curve
y —2 = x(l + x + where it cuts they-axis.
7. Examine the nature of the point (-1, —2) on the curve
x 3 +2x 2 +2.y—y 2 +5x-2y=0.
588 DIFFERENTIAL CALCULUS

8. Show that each of the curves

(xcosa — ysina . -b) 3 = c(xsina+ ycosa )2
where a is a variable, has a cusp and that all cusps lie on a circle.
9. Show that each of the curves
- 212x2 - y 2 + 1 4 x + 4 - 412 =
where t is a variable, has a node and that all these nodes lie on a
parabola.

ANSWERS
1. (i) Single cusp of the first species.
(ii) Single cusp of the first species.
(iii) Single cusp of the first spedes.
(iv) Single cusp of the second species.
(v) Isolated point. (vi) Isolated point.
(vii) Single cusp of the second species.
(viii) Double cusp of the first species. (ix) Node.
(x) Double cusp of the second species.
(xi) Single cusp of the first species.
2. (I) Isolated point at (0,0).
(ii) Isolated point at (0,3) and a single cusp of the first species at
(2,3 )
(iii) Node at (3,2 ).
(iv) Nodes at (0,— 1 ): (i,o ) and ( - 1,0
(v) Single cusp of the first species at (0,0 ).
(vi) Node at (0, 0). (vii) Node at (2, 0).
(viii)Isolated point at (2, 0).
(ix) Single cusp of the first species at (-1, —2).
(x) Single cusp of the first species at (I, - I).
(xi) Single cusp of the first species at (1, —1).
(xii) Single cusp of the first species at (—a, 0).
(xiii) Nodes at(0, —1),(—I,0) and (-2, —1).
(xiv) Node at (2, 2).
5. Single cusp of the second species.
6. Single cusp of the second species.
7. Single cusp of the first species.
INDEX
Corssptrison text, 145
Absolute convergency.149 CIocd intenal,16
Absolute value, 17 Complex numbers. 144
Actual crror.223 Concavity and convexity,558
Aggregate. 15 analytical tcst,359
Alternating series. 145 crileriorm,561
Angle bdwcen radius vector and Constant. 16
tangent,443 Cotstimmity,86.375
Angle between two curses Continuity of
Cartesian,439 elementary function,94
Polar.454 exponential function,95
Approximate inverse circular function,95
CalcuIaton.222 logarithmic function,96
value,223 polynomial.95
Archimedes' Spiral.576 retional aig. functiots,95
Artzument,18 trigonometric functiots,95
Arithmetical continuum,6 Continuity on one side,39
Associate Loci, 546 Continuous function properties,90
Astrmd,570 Continuous vanable,16
Asytnptote.505 Centinuum
not parallel to y-axis ,505 arithtnetic,6
of alg.curves.506,507 linear,6
of polar curves,5 13 Convergence
parallel to y-axis,506,510, 518 of geometric series, 141
special case.508.5 2, Convergent sequence.) IS
Average curvsture,470 theorems, 117
Bernoulli's lncquality.470 Converse of Euler's theorem,39 I
Bounded function,24 Critical value,320
Bounded sequence, I 13 Cusp
Bounds .24 different types of .580
Cardioidc,576 douhlc,581
Catenary.577 of the first specics,58 I
Cauchy's condition for the Existence of the second spccies.581
of limi(,67
simtgle,580
Cauchys General principle of species of a,581
- Convergence,122 Curvatstre.470
meats value Lheorem,284 a lheorem,476
root test, 147 at the origin,477
Cauchy's form of remainder,280 Cycloid.566
in binomial expansion281 D Aictabert's ratio test, 146
in logarithmic expansion,251 Darboux's theorem,282
in Maclaurin's expansion,280 Dcdckinds theorcm,6
in Taylor's expansion, Dc Moivres theorcm,247
Centre of curvature.47 1494496 Dependent variable .17
Chainrtte,568 Deriatise.171
Characteristic points.538 infinite. 172
Chord of corsatore,47 I left-hand, 172
through the origin (pole).480 partiul.376
Circle of corvature.47 1.492 right-band. 172
Cisnoid of L)iocics,574 sign.218.219
590 DIFFERENTIAL CALCULUS

Derivatives of arc length Exact diflerential,389

Cartcsian.441 Expansion in powers of
polar,453 binomial lunction.296
Derivatives of the coeff. of cosine fonction,294
andf(x + h),296 exponential funclion.294
Different classes of logarithmic funclion,294
disconhinuity.88 sine fijnclion,293
infinite,89 Expansion of funsaion,273
Different classes of discontinuity in Infinite power series.292
ordinaiy,88 Exponential curve,573
oscillatory.89 Extrcmal.320
removable,89 EMrcmum,320
simple,89 Fermat's Lheorem,329
Different def. of asymptote,50.515 First principle, 181
Differential,22 1,388 Folium of Descaztcs,572
exact,389 Four.cusped hypocycloid.570
Differential coefficient,22 I Functions.l ,I 7,20
in some standard cases. 173 bounded,24
Differentiate, graphical representation,19
Differentiation, 170 hyperbolic,
from first principal.181 monotone,26
fundamental theorems on,178 multiple-valued. 18,375
of a function of a fonction,191 Functions,
of implicit ftrnction.395 of a function, 191
Discontinuity,88 single-valued. 18,375
Divergent scquence.116 several variables,375
Division by zero,18 Fundamental theorems
Domuin,20 on differentiation. 178
Double Cuvp,581 on Iimit.56
Envclope,523 Gauss's Test, 148
of curves.536 Generalised mean value theorcm,276
of normals,530 Geometrical Interpretation
of straight lines,529 of the denvalivea.220,435
of tangenis,530 of Rolle's Lheorem,274
of two - parameter fansily,538 Geometrical representation
of system of parabolas.539 of z =f(.r, y),375
a 2a2 Gradient, 221
Equality of --' = .377 Graphical representation oIfn,19
&uLion of Homogeneous function.385
Normal,437 Hyperbolic function, 194
tangent,433 Important
Eqia:guIar spiraL575 Inequality. 121
Eider's theorem limits,62
converse 01,391 limits on sequence,t17
on homogeneous fn., 391 Increment..! 70
Evaluation of certain limits.349 Independent variable, Ii
indeterminate forrns,349 - Indeterminate forms,349
using power serics.353 Infinite
Evolules.496.530 dcrivative,172
of ellipse.57 I discontinuity,90
of partboIa,57I limtt,55
properties.496 sequence, Ill
'INDEX i,I1

Integer.! • condition.3 16,420

Interval of convergence, 150,298 of function of two vanablee,420
determination. 150 working rule,319
Interval of function,16 Mean value theosem,275
cloed.l6 geometrical lnterpretation,276
length, lo dp
open, 16 Meaning of .379
dt'
Inverse Methods of expansion.298
of a parabola,559 Monge's notation,379
of astraight line, 554 Monotone function,26
of the pedal,553 Monotonic sequence, 114
Inverse circular function, 192 Monotonically
Inverse curve349 decreasing and Increasiag,26
from Cartesian eqn..550 Multiple-valued function,18,375
from pedal eqn.,551 Necessary and sufficient
from polareqn..551 condition for the
lnvolute,496 convergence of a sequence. 122
Irrational numbers,2 existence of a limit.67
Keratoid Cusp,58 I Newton's formula for curvaturc,477
Lagrange's form of remainder,277 nth derivatives of
Lagrange' method of undetermined some special function.247
mu!tipliers,425 Null sequence, 121
Law of refraclion.329 Numbers,!
Left-hand On some well-known curves,566
derivative, 172 Open lntrrva,16
limit,53 Ordinary ditcontintticy.88
Lejbnirz's theosem,256 Oscillatory discontinuity,89
Lemniscate,577 Osculinflexion,581
of Bernoulli.578 - Parameter,528
Length of an arc Partial
of an evolute,497 dervatives.376,390
L' Hospital's theorcm.283 differentiation,375
Limacon,349 Pedal curves,546
Limit.47 from Cartesian equation,546
of a function,48 from pedal equation-549
of product of two functions,57 from polar equation.347
of quotient of two fns,,62 properties,547
of sum of or difference of two Pedal equatiori.456
functions, 58 from Cartesian.456
Limit of a function of a function,60 from polar,456
chord PQ of ellipse,459
Limit of as Q - P.440
of parahola.458
arc PQ
Linear continuum,6 of sine spiral.451
Logarithmic curve.573 Percentage error,223
Logarithmic spirali75 Perfect differential.389
Lower bound of a sequence, 118 of two function,390
Maclaurin's Series generalisation,391
finite forms,297 Perpendicular from pole on tangcnt,45
Infinite forms,293,297 Point of Inflesion.559-
Maxima and minima.315.317 analytical condition.561
determination,3 17,420 - general criterioni62
necessary and-sufficient Point of Osculinflcxion.581
 US

Polar reciprocsl.552 Sigrso ,.-,.,ce ofderivativc,2I8

of parabola.555 Str,e Spira1.579
o.pbIar reciprocal.552 Single Cusp,580
of p f(r).555 Single-valued fn.18,375
Positive lntegers,1 Slope,22 I
Power series, 15 1 Small error in calculation,225
Properties, Some well-known curvcs,566
Prsngsheims theorem. 144 Species of a Cusp,581
Probabibility curve.573 Strophoid.574
Properties of cont. function.90 Subtangent, Subnormal
Rasbe's Test. 148 Cartesian,439
Radius of convergence, 151 polar,454
Radius of curvarure,472 Successive differenliation,246
from Cartesian equation,472 Successive partial derivatives,377
from Implicit equation,473 Sum of an infinite series,139
from parametric eqoations.473 Symbol olnfinity.55
from pedal equation,475 Symbolic operation.257
from polar equatior,.474 important resullts,257
from tangential polar eqn.,475 Tangent and normal.433
Radius of curvature Tangent at the origin.435
of evolute,498 Taylor's series
Ramphoid Cusp. finite form,276
fraction,249 infinite form,292
Rate-measurer.? 18 Term by Term
Rate of change. IS dillere ntiation, 152
Rational integration,
number, 1 Test for absolute convergence, 148
point, 1 Test of convergence, 145
section.4.5 Theorems on
Real number.5 sequence.117
propertics,6 curvature,477
Relative error,223 Total differential coeflicient,387
Remainder Tractrix,569
Cauchy's form.279 Transcendental number,3
Lagrange's form,278 Turning value.320
Remarks on functions,20 Variable
Removable discontinuity,89 dependent, 17
Right-hand independent, 17
den vative, 172 Witch of Agnesi.575
limit Working rule for finding
Rolle's theorem,273 asymptote of alg. curve.510
Rose petals.578 max.& min.,317
Section,3,4
of rational number,3,4
Semi-cubical parabola,57 I
Sequence, lll
convergent,! 15
divergent, 116
null, 121
oscillatory. 116
Set, 15
Sing of the slerivstis'es.2 I .2 9