UNIT 3 DIFFERENTIATION

Structure
3.1 Introduction 51
Objectives
3.2 The Derivative of a Function 52
Slope of a Tangent
Rate of Change
The Derivative
3.3 Derivatives of Some Simple ~uncti0r.s
3.4 Algebra of Derivatives
~erivativeof a Scalar Multiple of a Function
Derivative of the Sum of Two Functions
Derivative of the Product of Two Functions
Derivative of the Quotient of Two Functions
C The Chain Rule of Differentiation
3.5 Continuity versus Derivability
3.6 Summary
3.7 Solutions and Answers

3.1 ,INTRODUCTION

It was the seventeenth century. Some European mathematicians were working on two basic
problems.

i) Is it possible to find the tangent to a given curve at a given point of the curve?
.ii) Is it possible to find the area under a given curve?
Two mathematical giants, Newton and Leibniz, independent of each other, solved these
problems. The theory that they invented in the process was Calculus.
In this first unit on differentiation, we propose to introduce the concept of a derivative which ,

is a basic tool of calculus. Leibniz was motivated directly by the first problem given
above - a problem which was of great significance fo? scientific applications;.He
- Newton (1642-1727)

recognised the derivative as the slope of the tangent to the curve at the given point. Newton,
on the other hand, anived at it by considefing some physical such as determination
of the velocity or the acceleration of a particle at a particular instant; He recognised the
. derivative as a rate of change of physical quantities. We shall now show that both these
considerations lead to the concept of derivative as the limit of a ratio. Of course, to
understand what a derivative is, you should have gonethrough Sec. 2 thoroughly.
We shall first differentiale some standard functions using the definition of the derivative.
The algebra of derivatives can then be effectively used to write down the derivatives of.
severalfunctions which are algebraic combinations of these functions. We shall also discuss
the' chain rule of differentiation which offers an unbelievable simplificatiori in the process of
finding derivatives. We shall also establish a relationship between differentiable functions
and continuous functions which you have studied in Unit 2.
I

Objectives
Leibniz (1646-1716)
After studying this unit, you should be able to:
*

d.raw a tangent to a given curve at a given paint

;O determine the rate of change of a given quantity with respect to another
obtain the derivatives of some simple functions such as xn, I x I, fi etc. from the first
principles
Ond the derivatives of functions which can be written as the sum, difference, product,
quotient of functions whose derivatives you already know
derive and use the chain rule of differentiation for writing down the derivatiGes of a
composite of functions
a discuss the relationship between continuity and derivability of a function. .51
Elements of Differential
Calculus
The tangent of the angle which a
,line makes with the positive
:direction of the x-axis is called the
slope of the line.
3.2 THE DERIVATIVE OF A FUNCTION
Before defining a derivative, let us consider two problems in the next two subsections. The
'Eiist is to find the slope of a tangent and the second is to find the rate of change of a givcn
quantity in terms of another.
3.2.1 Slope of a Tangent
Let us consider the problem of finding a tangent to a given curve at a given point. But, what
do we mean by the tangent to a curve? Euclid (300 B.C.) thought of a tangent as a line
touching the curve at one point. This definition works fine in the case of a circle Fig.1 (a).
but it fails in the case of many other curves (see Fig. 1 (b).
Fig. 1 (b)
We may define a tangent to a curve at P to be a line which best approximates the curve near
P. But this definition is still too vague. Then how can we define a tangent precisely? The
concept of limit which you have studied in Unit 2 comes to our aid here.
Let P be a fixed point on the curve in Fig. 2 (a), and let Q be a nearby point on that curve.
The line through P and Q is called a secant. We define the tangent line at P to be the limiting
position (if it exists) of the secant PQ as Q moves towards P along the curve (Fig.2(b)).
It may not be always possible to find the lipiting position of the secant. As we shall see
later, there are curves which do not have tangents at some points. In fact, there are curves
which do not have a tangent at any point!
There is another question which we can ask here. Suppose we know that a tangent to a curve
exists at a point, how do we go about ktually drawing the tangent?
We have said earlier that the tangent at P is the limiting position of the secant PQ. With
reference to a system of coordinate axes OX and OY (Fig.3), we can also say that the
tangent at P is a line through P whose slope is the limiting value of the slope of PQ as Q
approaches P along the curve. The problem of determining the tangent is, then, the problem
of finding the slope of the tangent line,
 Differentiation

Suppose the curve in Fig.3 is given by y = f(x). Let (x,f(x)) be the point P and let
Q (X+ ax, f(x + &;)) be any other point on the curve. The prefix 6 before a variable quantity
means a small change in the quantity. Thus, 6x means a small change in the variable x.
(Caution : 6x is one inseparable quantity. It is not 6 x x). The coordinates (x + 6x, f(x + 6x))
indicate that Q is near P. If 8 is the angle which PQ makes with the x-axis, then the slope of
PQ = tan 8 = QMPM
- f(x + a x ) - f ( x )
6x
The limiting value of tan 8,as Q tends to P, (and hence 6x + 0) then gives us the slope of
the tangent at P. Thus,

i
I
the slope of the tangent line = lim
h+ o
f(x + 6x) - f(x)
6x

~ h t indicates
s that the tangent line will exist only if the limit of
f(x + 6x) - f(x) . exists as
6x + 0. 6x

Remark 1 In Fig.3.we have taken 6x to be positive. But our discussion is valid even for
negative values of 6x.
Let us take an example .
Example 1 Suppose we want to determine the tangent to the parabola y = x2at the point
P(2,4).
In Fig.4 we give a portion of the parabola in the vicinity of P(2.4).
 $lem?nts of Differential Let Q (x, xZ)be any other point on the parahfa. The slope of
Cakulua
-
y coordinateof Q y coordirlateof P
Pa = x coordinate of Q - x coordinate of P

= -x2 4 -
' x-,2
The tangent at P (2,4) is the limiting position of PQ as x + 2. Therefore, the slope of the
I Equation of a lime passing through a
tangent at P is
-
point (x,,y,)and having slope m is
-
Y y, = m(x x,)
:52'
x2 - 4
7GT
I
(X +2)(k - 2)
= lim = lim (X+ 2 ) = 4
'x-2 x-2 X- a
The equation of the tangent line will be (y - 4) = 4 (x - 2).
Now, how do we draw this tangent? Just mark the point P'by moving a distance 1 unit from
P, parallel to the x-axis to the right, and then, moving a distance equal to 4 units parallel to
the y-axis upward. Join P to P'as shown in Fig.4. The coordinates of p'are (2 t 1 , 4 + 4).
The resulting line will touch the parabola at P, and the slope of the tangent at P = tan 8 = 4.
E 1) Find the equation of the tangent to the following curves at the given points.
a) y = llx at (2, 112)
b) = x 3 a t ( l , 1)

In this gubsection we have given a pkcise ddinition of a tangent to a curve. We have also
seen how to draw the tangent to a given curve at a given point. Now let us consider the
Second problem mentioded at the beginning of this section.

3.2.2 Rate of Change

Suppose a particle is moving along a straight line, h d covers a distance s in time t. The
distance covered depends on the time t. Q a t is s = f(t), a function of time. When the time
changes to t + 6t. the distance covered 'changes from f(t) = s to f(t + 6t) = s + 6s. Therefore
we oan say that 6s is the distance covered fi~the time St. We want to know the average
velocity of the particle during the time interval t to t + 6t (or t + t to t, according as t > 0 ot
t < O).,'
Total distance travelled
Now ,'h e average velkity =
-=me taken
Therefore, the average velocity. in the time interval [t, t + 6t] (or [t + t, t]).
 f(t+6t)-f(t) - @ + a s ) - s
- - = -6s where
(t+6t)-t (t+6t)-t 6t'

But this does not give us the velocity of the particle at a particular instant t. which is called
the instantaneous velocity. How do we calculate this?
To f h d the velocity at a particular time t, we proceed to find the average velocity in the time
interval [t, t + 6t] (or It + ,t t]) for smaller and smaller values of 6t.
I If 6t is very small. then t + 6t is very near t and so the average velocity during the time
i interval 5t would be very near the velocity at t. It seems reasonable, therefore, to define the
I instantaneous velocity at time t to be lim
f(t + 6t)- f(t)
I
&+o 6t
Thus, we have f ( t + 6t)- f(t)
v = lim

L where s = f(t) is the distance travelled in time t. Comparing this box with the one given at
the end of the last subsection, we find that the concepts of the slope of a tangent and the
instantaneous velocity are identical. Further, velocity can be considered as the rate of change
of distance with respect to .time. So, extending our definition of velocity to other rates of
change, we can saythat if a quantity y depends on x according to the Ale y = f(x), then the
rate ofchange of y with respect to x canbe defined as

Exiimple 2 Suppose we want to find the rate of change of the function f defined by
f(x) = x + 5, +X E R, at x = 0.
We shall first calculate the average rate of change off in an interval [O,6x].
, This average rate of change off in [O,6x] is

I - 6x+5-5.=&=]
1 6x 6x
I Hence, the rate of change off at 0, which is the limiting value of this average rate as
! 6x -+ 0,
f(O+ 6 x ) - f ( @
, = lim = lim 1 = 1.

Example 3 Suppose a particle is moving along a straight line and the distance s covered in
time 1 is given by the equation s = (1/2)t2.Let us draw the curve represented by the function s
= (1/2)tZ,measuring time along x-axis and,distance along y-axis. Let P and Q be points on
the curve which correspond to $ = 2 and t, = 4.
1
We shall show that the average velocity of the particle in the time interval [2,4] is the slope
of the line PQ and the velocity at time t, = 2 is the slope of the tangent to the curve at t, = 2.
b

The curve represented by s = (1/2)t2is a parahoh as shownin Fig. 5. P and Q correspond to

1 the values t, = 2 and t, = 4 oft. Now s, = ( I n ) tI2=.2 and S, = (112) $2 = 8. Therefore, the
coordinates of the points P and Q are C2,2) and (4, 8), respectively.
8- 2
Theslopeof PQ= -=6/2=3.
4- 2
Also, the distance travelled by the particle in the time (t, - t,) is s, - s, = 8 - 2 = 6.
Therefore, the average velocity of the particle in the time (t, - t,) is
distance travelled 26/2 = 3.
time taken 6t may be positive or negative
Hence, the slope of PQ is the same as the average velocity of the particle in the time
0, - t,).
2
Further, to calculate the slope of the tangent at P, we choose a point R(2 + 6t, (2 + 6t) )on
1

the curve, near P. Then the required slope is lim (slope of PR),
61 +O 55
Elements of Differential
Calculus

Fig. 5
i2 (2 + 6t)?- 2 6t(4 + 6t)
lim
8t-o
- = st-o
(2+6t)-2
lim
26t
=2

And, what is the velocity at t,? It is lim W t , which is again equal to 2. Thus ihe velocity at
8t -0
t, is the same as the slope of the tangent at P.
Remark 2 i) Example 3 is a particular case of the general result: If the path of a particle
moving according to s = f(t) is shown in the ts - plane and if P and Q are points on the path
which correspond to t = t, and t = t,, then the average velocity of the particle in time (t, - t,)
is given by the slope of PQ and the velocity at time t, is given by the slope of the tangent atP.
ii) Distance is always measured in units of length (metres, centimetres, feet) and so velocity
v really means v units of distance per unit of time. The slope of the tangent is a
dimensionless number, while the velocity has the dimension of lengthltime.
Now you can try some exercises on your own.

E E 2) A particle is thrown vertically upwards in the air. The distance it covers in 'lime t is
given by s(t) = ut - (112) gt2where u is the initial velocity'and g denotes thc i~ccelerationdue
to gravity. Find the velocity of the particle at any time t.

E E 3) Find the rate of change of the area of a circle with respect to its radius when the radius
is 2 cm. (Hint: Express the area of a circle as a function of its radius first).
 Differentiation
E E 4) Find the average rate of change of the function f defined by f(x) = 2x2+ 1,+x E R in
the interval [l,1 h] and hence evaluate the rate of change of f at x = 1.

We have seen that the slope of a tangent and the rate of growth have the same basic concept
behind them. Won't it be better, then, to give a separate name to this basic concept, and
study it independently of its diverse applications? We give it the name "derivative".
Definition 1 Let y = f(x) be a real-valued function whose' domain is a subset D on R. Let
x E D. If

lim
f(x + 6x) f(x) - exists, then it is called the derivative off at x.
6x +o 6x
Now, if we write f(x + 6x) = y + 6y, then derivative off = lim 6yISx. Here 6y denotes the
change in y caused by a change 6x in x. , . &+O
The derivative is denoted variously by f(x), dyldx or Df. The value o f f (x) at a point x, is
denoted by f (xd. Thus, . Thc notation dy/& is due to Leibniz
and f(x) is due to Lagrange (1 736 -
f(xo + 6x)- f(x,) 1813).
f (x,)= lim
Sx+O 8x
f(xo + a x ) - . f v , )
If, in the expressinrl P(xo) = lim we write
6x-0 6~
xo+6x=x,weget6x=x-x,,and6x+Owx+x,.
f(x) - f(xo
f (x, )= lim
x+x
0
x-Xo
This is an altemapve expressiorr for the denvauve off at the point x,.
Remark 3 In this definition x and y are real numbers and are two dimensionlessnumbers.
If x and y are dir~tensionalquantities (lefigth, time, distance, velocity, area, volume) then the
derivative wll also'havea dimension. For convenience, we shall always treat x arlc y a.
dimensionless i'eal ntirnbqs. The appronrias dimensions can k addcd later.
Cautionz 'dy' and 'dx' in the e x p s i o n dyMx are not separate entities : :.rn.l, !: . , t ~

'd' fro& dyldx to get ylx. The noation only suggests the fact that the deril aei~r:1s otl~iil~,
;

as a ratio.
When f'(x) exists, we say that f is diflerentiale (or dkrivable) at x. When f is
differentiable at each wint'of i a domain D, then f is said to be a differentinble functiurr.
'E'he process of obtaining the derivative is called differentiation. The funchurl f kkuc;,
associates to each poifit x of D, the derivative f'(x) at x, is called the derived function off.
Tbus, the domain of the denved function is { x E D:fyx) exists).
The process of tinding the derivative of 8 function by actually calculating the limit of the
f(x +6x)-f(x)
ratio is called differentiating from first principle&
6x
As we shall see later, it is not always necessary to find a derivative from the first principles.
We shall develop certain rules which can k used to write down the derivatives of some
functions without actually finding the limit. Some such rules arecontained in the next
section.

3.3 DERIVATIVES OF SOME SIMPLE FUNCTIONS

Inthis section we shall find the derivatives of some simple functions like the constant
Elements of Differential function, [he power function and the absolute value function. We shall illustrate the m e t h ~ d
Calculus
of finding the derivative by the first principle method through some e x a p l e s . . .

Example 4 Let f : R + R be a constant function, that is, f(x) = c for all x E R, c being a
real number. We shall show that f is differentiable, and its derivative is zero.
f(x + 6 x ) - f ( x ) c - C.
Now, lim
SX+ o 6x
= lim
SX+O 6~
-
= lim 0 = 0
SX+O

Hence, a constant function is differentiable and its derivative is equal to zero at any point of
its domain.
The result of the above example can be seen more easily geometrically. The constant
function f(x) = c +x E R represents the straight line y = c which is parallel to the x - axis
(Fig.6). If we join any two points, P and Q, on it, the line PQ is parallel to the x - axis.
Hence, the angle made by PQ with the x - axis is zero. This means the slope of PQ is
tan 0 = 0.Since f'(x) is the limit of this slope as Q -+ P, we get f'(x) = 0 for all x in the
domain o f f .
4
Y

4 *
C

Example 5 We now show that, if n is a positive integer, then D(xn)= nxn-I.

In order to obtain D(xn), in case it exists, we have to deteirtine
(x+h)" -xn
lirn
h +O h
Notice that we have used the letter h (instead of our usual 6x) to denote the small change in
the variable x. We are, in fact, free to use any notation; but 6x and h are the more commonly,
used ones.
(X + h)n - xn
lirn
h+ 0 h

= lirn
( X" + nhxP-I + . . . . . . . + h n, --xn (by bionomial theorem)
h+O h

n-i
= nx

The result of the above example is very useful. We shall show later, that - d (xn) = nxn- '
'
dx
for all non-zero x E R even when n is a negative integer. The result also holds for all x >#O
if n is any non-zero real number. Of course, if n = 0,then xn = 1 +x and hence, Dxn = 0 for
all x E R. This means that the result is trivially true for n = 0.~ e v e h h e l e s sright
, now we
are in a position to prove this result for n = 112. That is,

-( &. = (+)x-"', and this we do in the following example.

dx
6 is not defined for x < 0 . Example 6 We shall show that the function f defined by f(x)=, -fi-,
x > 0 is differentiabbe.

We have, lim
f(x + h)- fix)
= lirn
m-fi
h+O h h+O h
 Differentiation

- lim (x + h ) -x
- h+o h ( ~ + hfi)

1 The result of our next example is of great significance. Recall that, in Sec. 2 we mentioned
i that there are functions that have no tangents at some point (or equivalently. have no
I derivative there). This example will illustrate this fact. Before giving the example we give
1

LI
some definitions.

~f (a) = lim .f(a ( a ) , if it exists, is called the right hand derivative of f(xj at
)-
f (a + h ) - f(a)
+
~,
h -rot h
x = a and is, written as Rf'(a). Likewise. Lf'(a) = lirn is called the left
I h+O- h
hand derivative of f(x) at x = a and is written as L f (a). I f f (a) exists, we must have
Rf(a) = L f (a) = f (a) (See Unit 2, Theorem 4).
Example 7 The function f : R + R defined by f(x) = I x I is not derivable at x = 0 but is
derivable at every other point of its domain.

Fig. 7

Fig. 7 shows the graph of this function.

To prove that the given function is not derivable at x = 0,we have to show that

lim 10 + I - 101 does not exist. In fact as we shall see, Rf(0) and Lfl(0) both exiht. but
ti+o h
they are not equal.
10 + I1
Xow Rf '(0)= lirn -
tl-rOt h

Therefore. RtY(O) = 1 # -1 = Lfl(0). Hence l'(0) does not e u \ t Wc shali now \how that the
function is derivable at every other point.
First, let x > 0. Choose h so that I 1 I < x. This will ensure that x + h > 0 whether h > O or
h i0.Now,

lim
f(x i- h ) - f ( x ) - lim + h I - 1x1 IX
tl +() h b; 0 h
x+h-x
= lim
h+Q h
= lim h/h = I
I1 -+ 0
Thus f is derivable at x , and f ( x ) = 1 for all x > 0.
You can now complete the solution by solving E 5 ) .
I Elements of Dlffcrcntlel
Calculus E E 5 ) Show that y = I x I is derivable arld f ( x ) = -1 at all points x < 0.

E E 6) Show that each of the following functions is derivable at x = 2. Find f'(2) in each case.
a) f : R + R given by f(x) = x
b) f :R +R given by f(x) = ax + b where a and b are fixed real numbers.

I
1

j
I
I
1

E E 7) Find dyJdx, w h v e r it exista. for each bf tho following functions:

@)ymrs b ) y - ) r + I 1 C ) ~ = ~ ~ X ~ - ~
r
1
I
So far we have obtained derivatives of certain functions by differentiating from the first
f(x + 6x)- f ( x )
principles. That is, each time we have calculated lim
SX+O 6x -
But the process of taking limits is not an easy operation. It can become a very lengthy and
complicated affair. In the next section we shall see how to simplify the process of
differentiation for some functions.

3.4 ALGEBRA OF DERIVATIVES

Consider the function f(x) 2x3

x4-1
' 3xf If we try to find the derivative of this function
from the first principles, we will have to do lengthy, complicated calculations. However, a.
close look at this function reveals that it is composed of several functions: constant
functions like 2 , 3 and - I. and power functions like x3, x2 and x4. We already know the
derivativ~sof these functions. Can we use this knowledge to find the derivative of f(x)? In
this section we shall state and prove some theorems which help us do just that.

3.4.1. Derivative o f a Scaiar hlultiple of a Funstion

Lct i': R -+ R be a differentiable func:ion and let c E R. 'Then, consider the function
\, ci'i
, ., ". 9 '7 c;!il !hi> f ~:i'+o:! B : i ~ , ~ ~m\i!tip;e
l;l~. o f f by c (see Unit 1). The derivative of y
w i l j l rt.<pt,5,,,. ., ..

= cf'(x)
Tlius. we h,n.e jubt proved the following tneorcrn.

Theorem I If f is a differentiable function and c 6 R then c i is differentiable and

(cf)'!x) = cff(u).
Example 8 To differentiate y = 7 1 x I we apply the scalar multiple rule obtained in Theorem 1
at all points where the function I x I is differentiable and get

But. in view of Exain@lc7, when x = 0, -d (I x I) does not exist. When x 10. d (I x I) = 1

dx dx
-
andwhenxc,@, (IxO=-1.
dx
d d
Therefore, -;i;-(71x1)=7-i:xl)
dx
7when x > 0
- 7 when x < 0
-
and d (7 1x1) does not exist at x = 0.
dx
Note: In example 8 we have used the fact that iff '(x) does not exist, at a point then fcf)'(x)
also does not exist at that point.
Try the following exercise cow.
Elements of Differential
Calculus E E 8) Differentiate the following, using Theorem 1.
a> (5/3>x3 b) 84;

3.4t2 Derivative of the Sum of Two Functions

Let f and g be differentiable functions from R to R. Let us examine whether f + g, the sum
of the functions f and g, is differentiable. Now,
( f + g ) ( x + h ) - (f + g)(x)
lim
h+O h
(f(x + h ) + g(x + h ) - f ( x ) - g ( x ) }
= lim

h+O{
lim
= h+o

f(x
f(x + h ) - f ( x )
h

+ h) - f(x)
h

+
g(x + h ) - g ( x ) .
h

g(x + h ) - g ( x )
?
= lim t lim -
h +O h h+O h

Thus, we htve proved the following

Theorem 2 The sum of two differentiable functions f and g is a differentiable function and
(f + g)' (x) = f'(x) + g'(x) -tf- x E R.
The above result can be easily extended to a finite sum, that is,

where f, , ........ , fn are differentiable functions.

Remark 4 From Theorems 1 and 2 it follows that i f f and g are differentiable functions, then
f - g is also a differentiable function (since f - g = f + ( - g)), and (f - g)'(x) = ff(x) - gf(x).
Let us see how Theorem 2 is useful in the following example.
Example 9 To differentiate 3x2+ 41x- 9, we apply Theorem 2, and get,

2
d
Now. -(3x2) = 3dX- (in view of the theorem)
dx dx

d
and -( - 9 ) = O(see Example 4).
dx
d
Thus, -(3x2 + 41x - 9) = 6x + 41
dx
You are now in a position to solve this exercise.
E E 9) Differentiate the following: Differentiation

a) 5x3+ 2
b) a,+a,x + a 2 x 2 +....... + anxn,wherea, E R f o r i = 1, 2, ......., n.

- -

3.4.3 Derivative of the Product of Two Functions

Let f and g be two differentiable functions on R. We want to find out whether their product
fg is also differentiable.
fg(x + h ) - f g ( x )
Now, lim
h -40 h
f(x f h)g(x + h ) - f(x)g(x)
= lim
h +0 h
{ f ( x + h ) - f ( x ) } g ( x + h ) + {g(x + h ) - g ( x ) } f ( x ) )
= lim
h -+O h
(We have added and subtracted f(x) g(x + h))

- lim
h-40
{ f(x+h)-f(x) '
g(x+h))+lim h - o{
g(x + h ) - g ( x )
f (x ))
f(x + h ) - f ( x )
= lim l i m g ( x + h ) + l i m g (x + h) - g ( x ) lim ( x )
h-+O h h -+\0 h-+o h h -+o

(Ref. Unit 2, Theorem 3)

= f'W g(x) + g'W f(x)
Thus, we get the following:
Theorem 3 The product of two differentiable functions is again a differentiable function and
its derivative at any point x is given by the formula.
(fg)'(x) = f'(x) g(x) + f(x) g'(x)
We can extend this result to the product of three differentiable functions. This gives us
(fgh)'(x) = f'(x)g(x)h(x) + f(x)g'(x)h(x) + f(x)g(x)h'(x)
You see, you have to differentiate only F e function at a time. This respllt can also be
extended to the product of any finite number of differentiable functions. Thus, iff,, ... f,, are
differentiable functions, then,
(f ,f2. ... fn)'(x) = f ,'(x)f2(x) ... fn(x) + f ,(x)fi (x)f,(x) ... fn(x)+ ... + f ,(x)f2(x) ... .fnP(x).
Theorem 3 is very useful in simplifying calculations, as you can see in the following
example.
Example 10 To differentiate f(x) = x2(x + 4). we take g(x) = x2, h(x) = x + 4. We have,
f(x) = x2(x + 4) = g(x)h(x)
d
Now, gt(x) =- (x2) = 2x and h t ( x ) - = L (x + 4) =1
dx dx
 Thus, f'(x) = g'(x)h(x) + h'(x)g(x)
= 2X(x + 4) + 1 x x2

Remnrk 5 You could also have differentiated x2(x + 4) without using Theorem 3, as
follows:
x2(x+ 4) = x3+ 4x2
Therefore, -& -
d (x2(x + 4) = d (x3 + 4x2)
dx

This shows that the same function can be differentiated by using different methods. You
may use any method that you find convenient. This observation should also help you to
check the correctness of your resuit. (We assume that you would not make the same mistake
while using two different methods!)
E E 10) Using Theorem 3, differentiate the following functions. Also, differentiate these
functions without using Theorem 3, and compare the results.
a)x fi b) ( x +
~ 2~~+ 512 c)(x+ l ) ( x + 2 ) ( ~ + 3 )

3.4.4 Derivative of the Quotient of Two Functions

Let 0 = flg, where f and g are differentiable functions on R, and g(x) # 0 for any x. Then,
@(x+ h)-- @(X)= lim (f/gXx + h ) - ( f / g ) ( x )
lim
h -+o h h -0 h
f(x+h) f(x)]
= lim -
g(x+h) g(x)
g(x) f (x + h) - f(x) g(x +h)
= lim
h -+o hg(x) g ( x + h )
f(x + h ) - f(x) g(x+ h)- g(x)
gix) h - f(x) h
= lim
h -0 g(x + h) g(x)
 (by adding and subtracting f(x)g(x) from the nutfierator)
f(x + h ) - f(x) -
- h+u i
lim g ( x )
h
- f(x)
g(x + h) &!(XI

lim gdx 4 h) ,- g ( x )
h 1-
h+O

- iim g (X + h) hlim g (X )
h -0 4 0

Thus, we get the following.

Theorem 4 The quotient f/g of two differentiable functions f and g such that g(x) # 0. for
any x in its domain, is again a differentiable function and its derivative at any point x is
given by the following formula :

This can also be written as

numerator -

(denominator) (derivative of numerator ) - (numerator) (derivative of denominator)

(denominator)*
We will obtain an important corollary to Theorem 4 now.
Corollary 1If g is a function such that g(x) # 0 for any x in its domain, then

Proof :In the result of Theorem 4, take f to be the constant function 1. Then f'(x) = 0 for
all x.
Therefore,

Example 11We shall now show that- d (x") = rixn-', where n is a negative integer and
dx
x # 0.We have already proved this result for a positive integer n in Example 5.
Consider the function f : R\ { 0)- R given by f(x) = x", where m E N. Then
f(x) = l/xmwx E R. Thus, f = l/g, where g(x) = xmfor all x E R,x # 0.g is a differentiable
function and g(x) # 0 if x # 0.So, except at x = 0,we find that
- g ' (XI
f '(x)= (from Corollary 1)
(~(~11
--. - mxm-' (g ' (X) = mx rn -' by using Example 5)
(xmf
-
-
- mxm-'
= - mx-rn-l
x2
Denoting -m by n, we get f(x) = xn, and f'(x) = nxn- I .
Example 12 Let us differentiate the function f given by f(x) = ( x - ~+. 2) / (x2+ 2x)
We can write f as the quotient g/h where g(x) = ( x - ~+ 2) and h(x) = xZ+ 2x.
d d
NOW, g.(~)= - - ( x - ~ ) + - - ( ~ ) = - ~ x - ~ + o = L ~
dx dx x3
Also hl(x) = 2x + 2.
Elements of Differential
Calculus Therefore, f' (X) = h(x)g' (x )- g tx )h' (x)
(h (x ))2

E E 11) Differentiate
2x+ 1 - 1
a) - b) where a, b, c, d are fixed real numbers
x +5 a+bx + cx2 + d B

E E 12) Obtain the derivative of l/f(x) by differentiating from first principles, assuming that
f(x) # 0 for any x.
 Differentiation

2 + 5x + 7x-I
E E .13) Differentiate f(x) = C by three different methods.

3.4.5 The Chain Rule of Differentiation

The chain rule of differentiation is a rule for differentiating a composite of functio. ..,
(Ref. Unit I). It is a remarkable rule which helps us to differentiate complicated functions in
an easy and elegant way.
We establish the rule in the following theorem.
Theorem 5 Let y = g(u) and u = f(x). If both dy/du and du/dx exist, then dy/dx exists and is
dy = -
given by - dy x -
du
dx du dx
Proof: We first note that y = g(u) =gf(x) = (gof) (x), so that y is the composite function gof.
We are given that y, regarded as a function of u, is differentiable. We want to prove that y,
regarded as a function of x, is also differentiable. To do this we must show.that lirn 6y/6x
b+O
exists, where 6y is the change in the variable y corresponding to a change 6x in the variable
x. Now. 6u. the change in the value of u corresponding to a change 6x in the value ofx
,is..
giGn by 6u = f(x + 6x) - f(x).

We have lim 6u = lirn ( $ 6 ~ )

&+O &+O
6u
= lirn - x lim 6 x
&+O 6x &-0

This means that 6u + 0 as 6x + 0

We assume howgver that 6u # 0.
Elements of Differential
Calculus This implies that lirn -
8-0 SU 6u+O 8u
-
6 Y = lirn SY exists and is equal to -
dy
du

. 6~ =dy and lirn --

hm -
Su - du
-
6n-10 6~ du ‘5.40SX dx
Hence, we get
SY = lirn -
lirn -
&-+O~X
sy . lim SU
SX+O~U 6x-O~X
-

Hence dy/dx exists and is equal to dy/du x du/dx.

You may find it more convenient to remember and use the rule in the following form:
If h(x) = g(f(x) ) is the composite of two differentiablefunctions g and f, then h is
differentiable and h'(x) = g'(f(x))f'(x).
To clarify this rule let us look at the following example
Example 19 Here we shall differentiate y = (2x + with respect to x.
Letu=2x+ l . T h e n y = ( 2 ~ + 1 ) ~ = u ~ .
Now y is a differentiable function of u and u is a differentiable function of x. dy/du = 3u2
-
and du/dx = 2. Hence we can use the chain rule to get dy = -
dx
dy . -
du dx
du
2
= 3 u 2 . 2 = 6u2 =6(2x + 1).
You might be thinking that there was really no necessity of using the chain rule here. We
could simply expand (2x + and then write the derivative. But the situation i s q t always
ai simple as in this example. You would appreciate the power of the chain rule after using it
in the next example.
Example 14 To differentiate (x3+ 2x2- 1)"'.
let y = (x3+ 2x2- l)lm and let u = (x3+ 2x2- 1). Then y = uIm
Since dy/du and du/dx both exist, and dyfdu = 100UW and du/dx = 3x2+ 4x, therefore, by
chain rule, dy/dx = dy/du. du/dx.
= 1 0 0 ~ 1(3x2
~~+ . 4x)
99
= 1 o 0 ( ~+ ~2 ~ -
2 1) ( 3 ~ 2+ 4 ~ )
Can you really attempt to solve the above example without using the chain rule? Don't you
think the rule has simplified matters a lot for you?
Instead of introducing u explicitly each time while applying the'chain rule, after a little
practice you would find it more convenient to do away with u and arrange the working in the
above example as follows:

Our next example illustrates that this rule can be extended to three functions.
Example 15 To differentiate ((5x + 2)2+314,we write y = {(Sx+ 2)2+ 314,u = (5x + 2)2 + 3
andv=5x+2
Then y = u4 and u = vZ+ 3. That.is, y is a function of u, u is a function of v, and v is function
of x. By extending the chain rule, we get

This gives,
d y / d x = 4 u 3 . 2 v . 5 = 4 0 u3 v
 This example illustrates that there may be situations in which we may go on using chain rule
for a function of a function of a function ..., and so on. This perhaps justifies the name
'chain' rule. Thus, if g, . . . g,, and h are functions such that h = (g,og,o . . . ogn)(x), then

h'(x)= g;(g,o Ogn(x)) g;(g30 ... Og") (x)...g;-,(g,)(x). gj,(x)

E E 14) Find dyldx for each of the followtng using the chain rule:

3.5 CONTINUITY VERSUS DERIVABILITY

We end this unit with the relationship of differentiability with continuity, which we have
studied in Unit 2. In Sec. 3 of Unit 2 we proved that the function y = I x I is continuous
1
t f x E R. We have also proved that this function is derivable at every point except at x = 0 in
Example 7. This means that the function y=l x I is continuous at x = 0, but is not derivable at
that point. Thus, this shows that a function can be continuous at a point without being
derivable at that point. However, we will now prove that if a function is derivable at a point,
then it must be continuous at that point; or derivability continuity.
Recall that a function f is said to be continuous at a point xo if lim f(x) = f(xo) .
x-+ Xo

Tbeorem 6 Let f be a function defined on an interval I. Iff is derivable at a point x, E I,

then it is continuous at x,.
Root: f ( x ) - f(xO)
If x #xothen .we may write f(x) - f(xo) = ( x -x0)
- ,o
f(x) - f(xo)
Since f is derivable at x , exists and equals f'(%).
- xo
Elemer?tsof DilPerenlial Thus, taking limits as x + x,, we have,
Calculus
lim [f (x ) - f (xo)]
x--lX
0

f ( x ) - f(xo)
= lim
X+XO
. x-xo
lim ( X - xO)
x+xo

Therefore, lim f (x ) - lim f (xo) = 0

X+ Xo x +wo

+ X , ~ f (xo) = f (xo)
That is, x+limXo f(x ) = X lim

Consequently, f is contiquous at x,.

As we have seen, the fuhction y = I x I is continuous but not derivable at only one point,
x = 0.But there are sofne continuous functions which are not derivable at infinitely many
points. For instance, lbok at Fig. 8.

Fig. 8

It shows the graph of a continuous function which is not derivable at infinitely many points.
Can you mark those infinitely many points at which this function is not derivable? You can
take your hint from the graph of the function y = I x I .
The situation is, in fact, much worse. There are functions which are continuous everywhere
but differentiable nowhere. The discovery came as a surprise to the nineteenth century
mathematicians who believed, till then, that if a function is so bad that it is not derivable at
any point, then it can't be so good that it is continuous at every point. The first such function
was put forth by Weierstrass (although he is said to have attributedw
the discovery to
Riemann) in 1812. He showed that the function f given by f(x) =C bn cos (an x x ) ,
n =O
where a is an odd integer and b is a positive constant between 0 and 1 such that
ab > 1 + 3x12, is a function which is continuous everywhere, but derivable nowhere. It will
not be possible for us to prove this assertion at this stage.
Sometimes we use Theorem 6 to prove that a g ~ er, b function is continuous at a given point.

We prove that it$ derivative exists at that point. By Theorem 6 then, the continuity
automatically follows.
50
E E 15) Is the function f : [0,1] + R :f(x) = (2x + 3) continuous at x = 0. l ?
9x+ 2
3.6 SUMMARY

We conclude this unit by summarising what we have covered in it.

1 For any function y = f(x)
f(x + a x ) - f(x)
lim - (if it exists) is called the derivative off at x, denoted by ff(x).
6x40 6x
The function f' is the derived function. The derivative ff(x) is the slope of the tangent to
the curve y = f(x) at the point (x, y). The derivative also gives the rate of change of the
function with respect to the independent variable.
2 The derivative of a constant function is 0.

where n is any integer (and x # 0 if n c 0).

3 The function y = I x I is derivable at every point except at x = 0.

4 (cf)' = cf', c a constant.
(f +g)'= f' + g'
(f-g)'=ff-g'
(fg)' = fg' + gf'
g' f - f'g
(flg )' =
g2
(gof)' = g'(f).ff
5 Every derivable function is contipluous. The converse is not true, that is, there exist
functions which are continuous but not differentiakle.

3.7 SOLUTIONS AND ANSWERS

= - 114. Equation : (y - 112) = ( - 114) (x - 2)

orx+4y=4
dyI
b)dxx=l
=3. Equation : (y - 1) = 3(x - 1)

E 3) area = A = rc$ . -
E 4) average rate of change off in [I, 1 + h] =

f(1 + h ) - f(1)
rate of change off at x = 1 = lim -
where h may be positive or negati
h-+O h
= lim [ 4 + 2 h ) = 4
h3 0
Elements of Differential E 5) If x < 0, choose 0 < h < I x I then x-+ h < 0 and
Calculus
Ix+hI-lxI - - ( x + h ) - ( - X ) -h
lirn . - =- =- 1
h +o h h h
Thus f'(x) = -I if x < 0. f is derivable f o x~c 0.

E 6) a) f' (2)= lim

h+O
.f ( 2 + hh) - f(2)
2+h-2=1
h
= lim
h+O
a(2+h)+b-(ax2+b)
b) f' (2) = lim
h+O

= lim
. ah
-= a .
h+O h

E7) a) -dy
= lim (X + h $ - x 3 = lirn (3x2 + 3xh + h2) = 3x2
dx h+O h h+O

b) If x > -1, x + 1 > 0,choose h > 0 s.t. h < 1.x + 1 1.

Ix + h + 11-lx'+ 11
t h e n x + h + 1 >Oand lirn
h+O h
(X + h ' + 1 ) - ( X + 1 )
= lirn = 1.
h+O h
I x + h + 11-lx+ 11
I f x < - 1, lirn =- 1.
h +O h
Thus dyldx exists when x > - 1 01 when x < -1. It does not exist when x = - 4
since R f (-1) = 1 and Lf (-I)= -1.

9 + h) +4 2 +~1
Cl
* = l i m ~ z ( ' x + h) + I
h+O h
- J 2 x + l x J2(x
d2(;(+
+I
h)+l+ J2*+ 1
- lim 2 1
J2(x + h ) , + i +J 2 r 1= ~ . i
 1 f ( x ) - f(x + h )
= lim -
h+o h f(x)f(x + h)
f ( x ) - f(x + h )
-
-.
f ( x ) lim f(x
h4 0
+ h)

E 13) a) f(x) =2xU5 + 5 ~ - 7xd ~ +

ff(x) = - 1 0 x 4 - OX-5 - 4 2 ~ - 7
2 (5 - 7 ) (2 + 5 x
~ - -~5x4 + 7x-I)
b) f' ( x ) = 10
X
-10
=x ( - ~ O X ~ - ~ ~ X ~ - I O X ~ )
= - 20x-' - -
c) f(x) = x-"2 + 5x + 7x-')
= -20~-5 - 4 2 ~ - 7- 1 0 ~ 4
E 14) a) u = 1 + 5x + 7x2, y = 5/u
- 25 - 7 0 ~

b) u = 2x + 3, y = u2/(1 + u3)

c) u = 9 x + S s v = u 3 + u - : y = v 7
dy/dx = 7v6 (3u2 - 3u4) x 9
= 63 [(9x + 5)3 + (9x + 5)-'I6 [3(9x + 5)2 -

exists at x = 0.1. Hence the function is continuous at x