This action might not be possible to undo. Are you sure you want to continue?
Structure
1.1 Introduction
Objectives
1.2 Basic Properties of R 1.3 Absolute Value 1.4 Intervals on the Real Line 1.5 Functions
Definition and Examples Inverse Functions Graphs of Inverse Functions
1.6 New Functions from Old
Operations on Functions Composite of Functions
1.7
Types of Functions
Even and Odd Functions Monotone Functions Periodic Functions
1.8 Summary 1.9 Solutions and Answers
1.1 INTRODUCTION
This is the first unit of the course on Calculus. We thought it would be a good idea to acquaint you with some basic results about the real number system and functions, before you actually start your study of Calculus. Perhaps, you are already familiar with these results. But, a quick look through the pages will help you in refreshing your memory, and you will be ready to tackle the course. In the next three sections of this unit, we shall present some results about the real number system. You will find a number of examples of various types of functions in Sections 5 to 7. You should also study the graphs of these functions carefully. It is very important to be able to visualise a given function. In fact, try to draw a graph whenever you encounter a new function. We shall svstematically study the tracing of curves in Unit 9.
Objectives
After reading this unit you should be able to: recall the basic properties of real numbers, derive other properties with the help of the basic ones, identify various types of bounded and unbounded intervals, define a function and examine whether a given function is oneoneJonto, investigate whether a given function has an inverse or not, define the scalar multiple, absolute value, sum, difference, product, quotient of the given functions, determine whether a given function is even, odd, monotonic or periodic.
1.2 BASIC PROPERTIES OF R
In the next three sections, we are going to tell you about the set R of real numbers, which is allpervading in mathematics. The real number system is the foundation on which a large part of mathematics, including calculus, rests. Thus, before we actually start learning calculus, it is necessary to understand the structure of the real number system. you are already familiar with the operations of addition, subtraction, multiplication and division of real numbers, and with inequalities. Here we shall quickly recall some of their properties. We start with the operation of addition:
,
Elements of Differential Calculus
A1 R is closed under addition. If x and y are real numbers, then x + y is a unique real number. A2 Addition is associative. x + (y + z) = (X+ y) + z holds for all x, y, z in R . A3 Zero exists. There is a real number 0 such that x+O=O+x=xforallxinR. A4 Negatives exist. For each real number x, there exists a real number y (called a negative or an additive inverse 06x, and denoted by x) such that x + y = y + x = 0. A5 Addition is commutative. x + y = y + x holds for all x, y in R .
Similar to these properties of addition, we can also list some properties of the operation of multiplication:
M1 R is closed under multiplication. If x and y are real numbers, then x.y is a unique reJ number. M2 Multiplicatimis associative. x . (y . z) = (X. y ) . ~ holds for all x, y, z in R . M3 Unit element exists. There exists a real number 1 such that x . 1 = 1 .x=xforeveryxinR. M4 Inversesexist. For each real number x other than 0,there exists a real number y (called a multiplicative inverse of x and denoted by x'! or by l/x) such that xy = yx,= 1. M5 Multiplication is commutative. xy = yx holds for all x, y in R. The next property involves addition as well as multiplication.
D Multiplication is distrihutiveover addition.
x(y + z ) = xy + xz holds for all x, y, z in R
You may have come acrossa "field"
in the course on Linear Algebra.
Remark 1: The fact that the above eleven properties are satisfied is often expressed by saying that the real numbers form afield with respect to the usual addition and multiplication operations.
In addition to the above mentioned properties, we can also define an order relation on R with the.help of which we can compare any two real numbers. We write x > y to mean that x is greater than y. 'The order relation '>' has the following properties:
0 1 Law of Trichotomy holds. For any two real numbers a, b, one and only one of the following holds: a>b,a=b,b>a. 0 2 5' is transitive. I f a > b a n d b > c , t h e n a > c , + a , b , c ~R . 0 3 Addition is monotone. Ifa, b, c, inR aresuch thata>b, t h e n a + c > b + c . 0 4 Multiplication is monotone id the following sense. Ifa,b,cinRaresuchthata>bandc>O,thenac>bc. Caution:a>bandc<~+ac<bc. Remark 2: Any field together with a relation > satisfying 01 to 04 is called an ordered field. Thus R with the usual > is an example of an ordered field. Notations: We write x < y (and read x is less than y) to mean y Y x. We wAte x I (and y read x is less than or equal to y) to mean either x < y or x = y. We write x 2 y (and read x is greater than or equal to y) if either x > y or x = y.
A number x is said to be positive or negative according as x > 0 or x < 0. If x 2 0, we shy that x is nonnegative. NOW, know that given any number x E W, we can always find a number y E R such that you y 2 X. (In fact, there are infinitely many such real numbers y). Let us see what happkhs when
we take any subset of R instead of a single real number x. Do yo11 think that, given a set
+i
Real NumMrs and Functions
SC R, it is possible tq find u E R such that u 2 x for ad x E S ?
Before we try to answer this question, let us look at a definition. Definition 1 Let S be a subset of R. An element u in R is said to be an upper bound of S if u 2 x holds fot every x in S. We say that S is bounded above, if there is an upper bound of S.
#I &
$li
ill
s t j ~
h S
Now we can reword our earlier questions as follows : Is it possible to find an upper bound for a given set ? Let us consider the set Z= 11,  2,  3,  4,
.,I
Now, each x E Zis negative. Or, in other words, x < 0 for all x E Z . Yb you see, itr this case we are able to find an upper bound, namely zero, for our set Zy
On the other hand, if we consider the set of natural numvers, N = { 1,2,3 &..f,obvimsly we will not be able to find an upper bound. Thus N is not bounded aborc.
Yett wiN, f course, realise that if u is an upper bound for a set S then u + 1, u + 2, u + 3, e ....,(in fact, w r , where r is any positive number) are all uppr$naqds nf S. For example, we have seen that 0 is an upper bound for Z. Check that 1,2,3,4.R. .. . are d upper bounds l of Z,
.
5
$'
d. $dt
2g
2:1
From among all the upper bounds of a set S, which is bounded above, we can choose an upper bound u such that u is less than or equal to every upper bound of S. We call this u the least upper bound or the supremum of S. For example, consider the set T=(x
E
'$1
R : x 2 1 4 )= { X E R :  2 < x < 2 )
Now 2,3,3.5,4,4 + ~r all upper bwnds for this set. are But you will see that 2 is less than any other upper bound. Hence 2 is the supremum or the least upper bound of T. You will agree that 1 is the 1.u.b. (least upper bound) of Z, Note that forboth the setsT and Z, 1.u.b. belonged to the set. This may not be true m the general. Consider the set of all negative real numbers R= {x : x < 0 ) . The 1.u.b. of this set is 0 But 0 ei R. . Working on similar lines we can also define a lower bound for a given set S to be a real number v such that v 5 x for all x E S. We shall say that a set is bounded below, if we can find a lower bound for it. Further, the lower bound of S which is greater than any other lower bound of S will be called its infimum or greatest lower bound (g.1.b.).
As in the case of l.u.b., remember that the g.1.b. of a set may or may not belong to the set.
,
We shall say that a set S c R is bounded if it has both an upper bound and a lower bound.
>.I
t
Based on this discussion you will be able to solve the following ekercise.
E
E 1) Give examples to illustrate the fouowing:
a) A set of real numbers havlng a lower bound, b) A set of real numbers without any lower bound, c) A set of real numbers whose g.1.b. does not belong to it, d) A bounded set of real numbers.
Now we are ready to state an important property of R.
C The order is complete.
Every nonempty subset S of R that is bounded above, has a supremum. (We shall use this property in Unit 10). Many more properties are either restatements or consequences of these sixteen properties. Here is a list of some of them. 1 2 3 Zero is unique. Ifx+Of=iforallxinR,thenOf=O. Additive inverse is unique. ForeachxinR,thereisauniquey i n R s u c h t h a t x + y = y + x = 0. Addition is cancellative. Ifx+y=x+z,theny=z. Unity is unique. Ifx.lf=xforallxinR,then I f = 1. Multiplicative inverse is unique. For each nonzero real number x, there is a unique y in R such that xy = yx =l. Multiplication is cancellative. Ifxy=xzandx#O,theny=z.
I
4
5
6
Definition 2: If x and y are any two real numbers, the result of subtraction of y from x is
denoted by x  y and is defined as x + (y). Similarly, the division x + y (also denoted by x/y) is defined as xyI, provided y # 0. Now we are ready to list a few more properties. You are already aware of these. But let us quickly recall them. 7
 (X+ y) = (  x) + (  y) for all x, y in R.
If~y=O,theneitherx=Oory=O.
(x')I = x for all x # 0 in R.
8
9
10 If x and y are non zero numbers such that x' = yI, then x = y. 11 I f a i b a n d c > O , t h e n a c < b c . 12 a is positive if and only if a is negative. 13 I f a < b a n d c < d , . t h e n a + c < b + d .
14 Ifa>bandc<O,thenac<bc.
15 a2is nonnegative for all a in R.
16 If a and b are positive, then i) a2= b2 a = b. (The symbol w is read as 'if and only if") ii) a2> b2 a > b iii) a2< b2a a < b
*
*
17 Ifb>0,thena2<b2*b<a<b. You are,also familiar with the following subsets of R:
1)
The set N of natural numbers. Note that it is the smallest subset of R possessing the following properties: i) EN ii) k~ N * k + l c N
2) The set Z of integers. It is the smallest subset of R possessing the following properties: i) Z x N ii) If x, y E Z, then x  y E Z. 3) The set Q of rational numbers. We observe that it is the smallest subset of R possessing the following properties: i) Q 2 Z ii) If x, y E Q and y + O, then xy' E Q. You must have also studied the following properties of these sets. 1) k E N if and only if k is a positive integer, that is, k E Z and k > 0. 2) The operations of addition and multiplication on N satisfy A 1, A2, A5, M1, M2, M3, M5 and D. They do not, however, satisfy A3, A4 and M4.
Real Numbers and Functions
.,; , .
i
3) The operations on Q satisfy A 1 to A5, M1 to M5 and 0 1 to 04. Therefore Q is an ordered field. But C is not satisfied, that is, Q is not order complete. We list here some more properties of these sets which you will find useful in your study of calculus: 4) Archimedean Property :If a and b are any real numbers and if b > 0, then there is a positive integer n such that nb > a. 5) If a is any real number, there is a positive integer n such that n > a (Archimedean property applied to a and 1). 6) A real number s is the supremum of a set S c R if and only if the following conditions are satisfied. i) s 2 x for all x in S. I ii) For each E > 0, there is a y in S sucn chat y > s  E. For example, consider the set A = ( x E R : 8 Ix c 10). 10 ig the supremumafrhis set. Now, if we are given any E, say, E = 0.01, we should be able to find some y E X such that y > 10 0.01 = 9.99. As you can see, y = 9.999 serves our purpose. Now 10.01 is a h q n upper Bound for A. But 10.01 is not the supremuA of A. For E = 0.001, we cannot find any y E A such that y > 10.01  0.001 = 10.009.
is; :
f!'
g1 .i P.! r::
, I..
e (epsilon) is a Greek letter used to denote small real numbers.
:*& $Fb
s.,
ii .a+
b. !
<+$;
I
"
,;;,r i
:I. t?I
a1
7) Every nonempty set of real numbers that is bounded below, has an infimum. The exercise below can now be done easily.
E
$#l
L . .
' i.;,
E2) a) Show that the set of positive real nurpbers is bounded below. What is its infimum? b) Write the characterisation of the infimum of a subset of R, wlich corresponds to 6) above. Give an example.
Z,/I~L
k*, , r"t
8
.O ,I .
.I!
s3,:
..:
;*
i;lif
'L
@I,
~'i, J ;; :; ,,.,;
i ' : ,
% ,
,
1.3 ABSOLUTE VALUE
In this qection we shall define the absolute value of a real number. You will realise the importdnce of this simple concept as you study the later units. .
Elements o Differential f Calculus
Definitio~3: x is a real number, its absolute value, denoted by ( x I (read as modulus of If x, or mod x), is defined by the following rules: Ixl = x, i f x 2 . 0
For example, we get 151=5, 151=5, 11.7(=1.7, 121=2, 1 0 ) = 0 It is obvious that ( x I is defined for all x E important properties of I x I.
R.The following theorem gives some of the
4
Theorem 1: If x and y be any real numbers, then
I*
h
E
a) By the law of trichotomy (01) applied to the real numbers x and 0, exactly one of the following holds: i) x > 0, ii) x = 0, or iii) x < 0.
Lei us consider these one by one.
i)
Ifx>O,thenIxI=xandx>x,sothat max ( x, x ) = x and hence I x I = max (x, x )
ii) If x = 0, then x = 0 = x, and therefore, (x,x) max (x,x)=O.AlsoIxI=O,sothatIxI=max iii) I f x c O , @ e n I x I =  x , a n d  x > x , s o @ a t . rnax (x, x ) = x. Thus, again, I x I = & (x, x ) .
From this it follows that x I I x I
b) I x I =max ((x),x) =max (x,x) Amax (x, x ) = I x I .
d) We shall consider two different cases accokding as
i)
x + y 2 r ) or i i ) x + y c O . 1~tx+~~0.~hen1x+~1=x+~.~bwx11x1and~~1~1b~(a).Therefo Ix+yl =x+ySlxl+lyl. Let x + y < 0. Then (x + y) > 0, that i$, (X) + (y) > 0 and we can use the resfflt of (i)
forxandy.NowIx+yI=I(x+y)Iby(b~. Thus, I x + y I = I (x) + (y) 1 5 Ix I + I y I, by i) .
=Ixl+lyl,
Therefore,wegetIx+yIIIxI'+Iyl.
'
by@).
Thuswefindthatforallx,y~R, I x + y I S l x I + I y I . e) By writing x = (x  y) + y and applying the triangle inequality to the numbers x  y @y, we have
Since ( I ) holds for all x and y in R,
therefore by interchanging x and y in (I), we have IyIIx I I I y  X I = \  ( x  y ) I = I x  y l , sothat (\XIly I)<lxy I.
Real Numbers end Fundions
... (2)
From (1) and (2) we find that I x I  I y I and its negative  (I x I  I y I) are both less than or at themostequaltolxyI.Therefore,max { I x I  I y I ,  ( 1 x 1  I y I ) ] S I x  y l . But the left hand side of the above inequality is simply 1 I x I  I y I I. Therefore, we habe 1 1 x 1  l y l ls l x  y l
R. Now you should be able to prove some easy consequences of this theorem. The following exercise will also give you some practice in manipulating absolute values. This practice will come in handy when you study Unit 2.
E
That is, I x  y I 2 I I x I  I y I I for all x, y
E
E 3) Rove the following:
.e) and f) can b extended to any number of reds. Now if a E R and 6 > 0 then , I x 3 I < 6 3 x  a / 6, or  (X a) < S according as x  a 2 0, or x  a < 0, reqdively.

If x  a < 6, this means that x c a + 6 If (%  a) 5 6, this means that a  6 < r Thus,wegetthat I x  a I < & * a  & < x < a + 6 . This means that the difference between x and a is not more than 6. In the next s&tion.we shall see how the set (x : I x  a I < 61 can be represented
I
Elements of DifTerential Calculus


1.4 INTERVALS ON THE REAL LINE
Before we define an interval let us see what is meant by a number line. The real numbers in the set R can be put into onetoone correspondence with the points on a straight line L. In other words, we shall associate a unique point on L to each real number and vice versa. Consider a straight line L (see Fig. l(a)). Mark a point 0 on it. The point 0 divides the straight line into two parts. We shall use the part to the left of 0 for representing%egative real numbers and the part to the right of 0 for representing positive real numbers. We choose a point A on L which is to the right of 0.We shall represent the number 0 by 0 and 1 by A. OA can now serve as a unit. To each positive real number x we can associate exactly one point P lying to the right of 0 on L, so that OP = I x I units (= x units). A ?ega$ve real number y will be represented by a point Q lying to the left of 0 on the straight line L, so that OQ = I y I units. (= y units, since y is negative). We thus find that to each real number we can associate a point on the line. Also, each point S on the line represents a unique real number z, such that I z I = 0s. Further, z i~'~ositive is to the right of 0 , and is negative if if S S is to the left of 0. This representation of real numbers by points on astraight line is often very useful. Because of this onetoone correspondence between real numbers and the points of a straight line, we often call a real number " a point of R". Similarly L is called a "number line". Note that the absolute value or the modulus of any number x is nothing but its distance from the point 0 on the number line. In the same way, I x  y I denotes the distance between the two numbers x and y (see Fig. l(b)).
Distance is always nonnegative.
F i p r e 1 :(a) Number line
b)DtPtrncebehveenxandy IsIxy (
Now let us consider the set of all real numbers which lie between two given real numbers a and b. Actually, there will be four different sets satisfying this loose condition. These are:
ii) [ a , ~ = ( x : a I x I b ) b) iii) la, b] = { x : a < x l iv) [a,b[= { x : a l x < b )

a
b
The representation of each of these sets is given alongside. Each of these sets is called an interval, and a and b are called the end points of the interval. The interval ]a, b[, in which the end points aremot included, is called an open interval. Note that in this case we have drawn a hollow circle around a and b to indicate that they are not included in the graph. The set [a, b] contains both its end points and is called a closed interval. In the representation of this closed interval, we have put thick black dots at a and b to indicate that they are incIuded in the get. The sets [a, b[ and ] a,b] are called halfopen (or halfclosed) intervals or semiopen (or semiclosed) intervals, as they contain only one end point. This fact is also indicated in their geometrical representation.
Each of the.se intervals is bounded above by b and bounded below by a. Can wt represent the set1 = ( x : I x  a I < 6 ) on the number line? Yes, we can. We know that I x 1a I can be thought of as the distance between x and a. This means I is the set of all numbers x, whose distance from a is less than 6. Thus,
is the open interval ]a  6, a + 6[. Similarly, I, = {x : I x  a I 2 6 ) is the closed interval [a  6, a + 61. Sometimes we also come across sets like I, = {x : 0 < I x  a I < 6 ) .This means if x E I,, then the distance between x and a is less than 6, but is not zero. We can also say that the distance between x and a is less than 6, but x # a. Thus,
Real Numbers and Functions
Apart from the four types of intervals listed above, there are a few more types. These are: ]a, m [ = {x : a < x ) [a, m [ = {x : a I x ) (open right ray) (closed right ray) 4
?
a a
v
> >
b b
1 m,b[ = {x : k < b ) (open left ray)
1 m, b = {x : x 5 b ) (closed left ray) j
1,,[=R (open interval)
<

4 > As you can see easily, none of these sets are bounded. For instance, ]a, [ is bounded below. but is ndt bounded above. I , b] is bounded above, but is not bounded below. Note that does not denote a real number, it merely indicates that an interval extends without limit.

E
We note further that if S is any interval (bounded or unbounded) and if c and d are two elements of S, fhen all numbers lying ktwe'en c and d are also elements of S. i E 4) State whether the followink are true or false. a) 0 E 11,81, C) 1~ [1,21 b) 1 E I,2[ d) 5 E I % +
E
E 5) Represent the intervals in E 4) geometrically.
1.5 FUNCTIONS
Now let us move over to functions. Here we shall present some basic facts about functions which will help you refresh your knowledge. We shall look at various examples of functions and shall also define inverse functions. Let us start with th6tMinitkm of a f t m c h .
1.5.1 Definition and Examples
Definition 4 If X and Y are two sets, a function f from X to Y, is a rule or a correspondence which connects every member of X to a unique member of Y. We write f: X + Y (read as "f is a function from X to Y). X is called the domain and Y is called the codomain off. We shall denote by f(x) that unique element of Y which is associated t0XE X. The following examples will help you in understanding this definition better. Example 1 f: N + R, defined by f(x) = x. f i s a function since the rule f(x) = x associates a unique member (x) of R to every member x of N. The domain here is N and the codomain is R. Example 2 The rule f(x) = x/2 does not define a function from N + Z as odd natural numbers like 1,3,5 ... from N cannot be connected to any member of Z.
Elements of Differential Calculus
ExampW 3 Every natural number can be written as a product of some prime numbers, Consider the rule f(x) = a prime factor of x, which connects elements of N. Here since 6 = 2 x 3, f(6) has two values: f(6) =.2 and f(6) = 3. This rule does not associate a unique number with 6 and hence does not give a function from N to N.
Thus, you see, to describe a function completely we have to specify the following three . things:
>
a) the domain b) the codomain and C) the rule which associates a unique member of the codomain to each member of the domain. The rule which defines a function need not always be in the form of a formula. But it should clearly specify (perhaps by actual listing) the correspondence between X and Y. Iff: X + Y ,then y = f(x) is called the image of x under f o r the fimage of x. The set of fimages of all members of X, i.e. (f(x) : x E X ) is called the range off and is denoted by f(X). It is easy to see that f(X)G Y.
Remark 3 a) Throughout this course we shall consider functions whose domain and codomain are both subsets of R. Such functions are often called real functions or realvalued functions of a real variable. We shall, however, simply use the word 'function' to mean a real function.
b) The variable x used in describing a function is often called a dummy variable because it can be replaced by any other letter. Thus, for example, the rule f(x) = x, x E N can as well be written in the form f(t) = t, t E N or as f(u) =  u, u E N. The variable x (or t or u) is also called an independent variable, and f(x) is dependent on this independent ~ariable.
Graph of a function A convenient and useful method for studying a function is to study it . through its graph. To draw the graph of a function f :X + Y, we choose a system of coordinate axes in the plane. For each x E X, the ordered pair (x, f(x)) determines a point '.I the plane (see Fig. 2). The set of all the points obtained byconsidering all possible values of x (remember that the domain off is X) is the graph off. The role that the graph of a function plays in the study of the function will become clear as we proceed further. In the meantime let us consider some more examples of functions and their graphs.
Fig. 2
1) A constant function: The simplest example of a function is a constant function. A constant function sends all the elements of the domain to just one element of the codomain.
For example, let f: R 4 R be defined by f(x) = 1. Alternatively, we may write f:x + l V x E R The graph off is as s h ~ w n Fig. 3. in
Itistheliney= 1. In general, the graph of a constant function f : x + c is a straight line which is parallel to the xaxis at a distance of I c I units from it.
Real Numbers and Functions
#
*I
7'
fit
8,
2 The identity function: Another simple hut important example of a function is a function which sends every element of the domain to itself.
Let X be any nonempty set, and let f be the function on X defined by setting f(x) = x *X E X I This function is known as the identity function on X and is denoted by ix.
'
The graph of i, the identity function on R, is shown in Fig. 4. It is :he line y = x. 3 Absolute valuehnction: Another interesting function is the absolute value function (or modulus function) which can be defined by using the concept of the absolute value of a real number as: x if x 2 0 f ( x ) t IxT= x ifx<O,
m.4
1
.
E
The graph of this function is shown in Fig. 5. It consists of two rays, both starting at the origin and making angles n/4 and 3d4, respectively, with the positive direction of the xaxis. E 6) Given below are the graphs of four functions depending on the notion of absolute value. The functions are x +  I x I, x + I x I + 1, x + I x + I I, x + I x  1 I, though not necessarily in this order. (The domain in each case is R). Can you identify them3
4 The Exponential Function: If a is a positive real number other than 1,.we can define a
function f as:
f:W+R
re
+
*q+ t
3,
L.l.
:3:.
'4. L
This function is known as the exponential function. A special case of this function, where a=e, is often found useful. Fig. 6 shows the graph of the function f : R + R'such that f(x) = ex.This function is also cdled the natural exponential hnction. Its s i the set R+of positive real numbers. 5 The Natural Logarithmic Function: This function is defined on the set'R+of positive real numbers f: R+ + R such that f(x) = In (x). The range of this function is R. Its graph is shown in Fig. 7.
6 The Greatest Integer Function: Take a real number x. Either it is an integer, say n (so that x = n) or it is not an integer. If it is not an integer, we can find (by the Archimedean property df real numbeisj an integer n, such that n < x < n + 1. Therefore, for each reid number x we can find an integer n such that n S x < n + 1. Further, for a given real number x, we can find only one such integer n. We say that n is the greatest integer not exceeding x, and denote it by [XI.For example, [3] = 3 and [3.5] = 3,[3.5]=  4. Let us consider the function defined on R hv setting f(x) = [XI.
~lcmcntsf Mfferentlsl o Cakulw
This function is called the greatest integer function. The graph of the function is as shown in Fig. 8, (It resembles the step? of an infinite staircase). Notice that the graph consiits of infinitely many line segments of unit length, all parallel to the x  axis.
M
4
.2
1
. , 
CO
M
l ~ l l I
7 Other Functions The following are some important classes of functions.
I *
I
4
c0
1
I
1
I
l
x
4
I
2
I
2
a) Polynomial Functions f(x) = aoxn+ a,xnI+  + anwhere ao, a I , , a,, are given real numbers (constants) and n is a positive integer. b) Rational Functions f (x) = g (x)/k(x), where g (x) and k (x) are polynomial functions of degree n and in. This is defined for all real x, for which k(x) # 0. C) Trigonometric or Circular G c t i o n s f(x) = sin x, f(x) = cos x, f(x) = tan x, f(x) = cot X, f(x) = sec x, f(x) = cosec x. (ex  ex d) Hyperbolic Functions f(x) = coshx =.~(e" + e" ) , f(x) = sinhx = 2 2 ).we shall study these in detail in Unit 5.
T2
T4
m8
1.5.2 Inverse Functions
of In this subsection we shall see what is meant by the inve~se a function. But before talking about the inverse, let us look at some special categories of functions. These special types of functions will then lead us to the definition of the inverse of a function.
Oneone and Onto Functions Consider the function h : x 4 x2, defined on the set R. Here h(2) = h(2) = 4. 2 and 2 ark distinct members of the domain R, but their himages are the same. (Can you find some more numbers whose himages Bre equal?) his may be expressed by saying that 3x, y such that x # y but h(x) = h(y).
Now, consider the function g: x + 2x + 3 Here you will be able to see that if x, and x2are two distinct real numbers', then g(x,) and g(x2)are also.distinct. For, x, # x,
* 2x,
#
2x2* 2x, + 3 # 2x, + 3* g (x,) # g(x2)
We have considered two functions here. While one of them, namely g, sends distinct members of the domain to distinct members of the codomain, the other, namely h, does not always do so. We give a special name to functions like g above.
Definition 5 A function f: X + Y is said to be a oneone function ( a (1 1) function or an injective funciion) if the images of distinct members of X are distinct members of Y.
Thus the function g above is oneone, whereas h is not oneone.
1
II
1
Remark 4 The condition "the images of distinct members of X are distinct members of Y in the above definition can be replaced by either of the following equivalent conditions:
f(x) # f(y) a) For every pair of members x, y of X, x # y b) For every pair of members x,y of X, f(x) = f(y) x = y. We have observed earlier that for a function EX + Y, f(X)E Y. This opens two possibilities:
i) f (X) = Y, or ii) f(X) 9 Y, that is, f(X) is a proper subset of Y.
The function h. : x + x2 + x E R falls in the second category. Since the square of any real number is always nonnegative, h(R) = R+u { O ) , the set of nonnegative real numbers. Thus h(R) $ R. i On the other hand, the function g : x + 2x + 3 belongs to the first category. Given any y E R (codomain) if we take x = (112)~ 312, we find that g(x) = y. This shows that every member of the codomain is a gimage of some member of the domain and thus, is in the range g(R). From this we get that g(R) = R. The following definition characterises this property of the function.
Definition 6 A function f: X + Y is said to be an onto function (or a surjective function) if every member of Y is the image of some member of X. Iff is a function from X onto Y, we often write: f : x * > Y (or f : X + + Y).
Thus, h is not an onto function, whereas g is an onto function. Functions which are both one
I 'Li
1.i
I
I I
4 :I
Consider a function f:X+ Y which is both oneone and onto. Since f is an onto function, each y E Y is the image of some x E X. Also, since f is oneone, y cannot be the image of two distinct members of X. Thus, we find that to each y E Y there corresponds a unique x E X such that f(x) = y. Consequently, f sets up a onetoone correspondence between the members of X and Y. It is this onetoone correspondence between members of X and Y which makes a oneone and onto function so special, as we shall.soon see. Consider the functiorrf : N + E defined by f(x) = 2x, where E is the set of even natural numbers. We can see that f is oneone as well as onto. In fact, to each y E E there exists y/2 E N, such that f(y12) = y. The correspondence y + y/2 defines a function, say g, from E to N such that g(y) = y/2.
Real Numbers and Functions
,
t
The function g so defined is called an inverse off. Since, to each y E E there corresponds, a unique x E N such that f(x) = y, only one such function g can be defined corresponding to a given function f. For this reason g is called the inverse off. As you will notice, the function g is also oneone and onto and therefore it will also have an inverse. You must have already guessed that the inverse of g is the function f. From this discussion we have the following: Iff is a oneone and onto function from X to Y, then there exists a unique function g : Y+X such that for each y E Y, g(y) = x e y =f(x). The function g so defined is called the inverse off. Further, if g is the inverse off, thtn f is the inverse of g, and the two functions f and g are said to be the inverses of each other. The inverse of a function f i, usually denoted by f'. . To find the inverse of a given function f, we proceed as follows: Solve the equation f(x) = y for x. The resulting expression for x (in terms of y) defines the inverse function.
X Thus, if f(x) =  + 2, we solve X + 2 = y for x. 54 I 5 1 This gives us x = (5(y  2))'. Hence f' is the function defined by f'(y) = (5(y  2 ) f
5
5 .

.
1.5.3 Graphs of Inverse Functions
.
There is an interesting relation between the graphs of a pair of inverse functions because of which, if the graph of one of them is known, the graph of the other can be obtained easily. Let f : X + Y be a oneone and onto function, and let g : Y+ X be the inverse off. A point (p, q) lies on the graph o f f 9 q.= f(p) 6 p = g(q) (q, p) lies on the graph of g. Now the points (p, q) and (q, p) are reflections of each other with respect to (w.r.t.) the line y = x. 'Therefore, we can say that the graphs o f f and g are reflections of each other w.r.t. the line y = X.
'
Therefore, it follows that, if the'graph of one of the functions f and g is given, that of the Other can be obtained by reflecting it w.r.t. the line y = x. As an illustration, the graphs of the functions y = x3 and y = x1I3are given in Fig.9. Do you agree that these two functions are inverses of each other? If the sheet of paper on which the graphs have been drawn is folded along the liie y = x, the two graphs will exactly coincide.
Ekments d Differential
C.lfulus
E
E 7) Compare the graphs of In x and ex given in Figs. 6 and 7 and verify that they are inverses of each other.
\
If a given function is not oneone op its domain, we can choose a subset bf the domain on which it is oneone, and then define if8 inverse functioil. For example. consider the function f : x + sin x. Since we know that sin (x + 2n) = sip x, obviously this function is not oneone on R.But d we resmct it to the interval [ n/2,7/2], we find that it is meone. Thus if f(x) = sin x +X E [42, W2], then we can define fA'(x) = sin' (x) = y if siny = x. Similarly, we can define cos' and tan' functions as Inverse of cosine ahd 'angent functions if we reiitiict the domains to [0, n] and]d2, n/2[, respectively.
I
I
E
E 8) Which of the followlllg functions are oneone? 9 f: R + R defined by f(x) = I x I ) b) C R + R defied by f(x)%3x  1 C) f: R + R defined by f(x) = x' dS. f: R + R defined by f(x) = I
E
E 9) WKch of the following functions are onto? a) f: R + R defined by f(x) = 3 ~ 7 b) ?A&+ R define&by f(x) =.\/I; 4 C) f: R 4 R defined by f(x) = x2+1 d) f: X + R defined by f(x) = l/x ,
where X stands for the set of nonzero real numbers.
, where X is the set of all E 10) Show that the function f: X + X such that f(x) = E? x1 real numbers except 1, is oneone and,onto. Find its inverse.
E
E
E 11) Give one example of each of the following: a) a oneone function which is not onto. b) onto function which is not oneone. c) a function which is neither oneone nor onto.

Real Numbers end Functions
1.6 NEW FUNCTIONS FROM OLD
In this section we shall see how we can construct new functions from some given functions. This can be done by operating upon the given functions in a variety of ways. We give a few such ways here.
1.6.1 Operations on Functions
ScalarMultiple of a Function Consider the function f : x + 3x2+1 +x E R. The function g : x + 2(3x2+1) +x E R is such that g(x) = 2f(x) +x E R. We say that g = 2f, and that g is a scalar multiple off by 2. In the above example there is nothing special about the number 2. We could have taken any real number to construct a new f"nction from f. Also, there is nothing special about the particular function that we have considered. We could as well have taken any other function. This suggests the following definition: Let f be a $unction with domain D and let k be any real number. The scalar multiple o f f by k is a function with domain D. It is denoted by kf and is defined by setting (kf)(x) = kf(x). Two special cases of the above definition are important. i) Given any function f, if k = 0, the function kf turns out to be the zero function. That is, 0.f = 0.
.
ii) If k = 1, the function kf is called the negative off and is denoted simplyby f instead of the clumsy If.
Absolute Value Function (or modulus function) of a given function Let f be a function with domain D. The absolute value function of f, denoted by I f I and read
as mod f is defined by setting
(I f I) (x) = I f(x) I, for all x ED.
Since I f(x) I = f(x), if f(x)> 0, and I f I have the same graph for those values of x for which f f(x)? 0. Now let us consider those values of x for which f(x) < 0. Here I f(x) I = f(x). Therefore, the graphs off and I f l are reflections of each other w.r.t. the xaxis for those values of x for which f(x)< 0.

Fig. 10
Elements of Differential Calculos
As an example, consider the graph in Fig. 10(a). The portion of the graph below the xaxis (that is, the portion for which f(x) c 0 ) has been shown by a dotted line. To !raw the graph of I f I we retain the undotted portion in Fig. 10(a) ai it is, and replace the dotted portiori by its reflection w.r.t. the xaxis (see Fig. lob)).
Sum, difference, Product and Quotient of two functions
If we are given two functions with a common domain, we can form.severa1 new functions by applying the four fundamental operations of addition, subtraction, multiplication and division on them. i) Define a function s on D by setting The function s is called the sum of the functions f and g, and is denoted by f + g. Thus, (f + g) (XI=f(x)
1
ii)
Define a function d on D by setting d(x) = f(x)  g(x), The function d is the function obtained by subtracting g from f, and is denoted by fyg. Thus, for all x E D (f  g) (x) = f(x)  g(x).
I
Y
8 I
iii) Define a function p on D by setting p(x) = f(x) g(x). The function p, called the product of the function f and g, is denoted by fg. Thus, for all XED iv) Define a function q on D by setting q(x) = f(x)/g(x), provided g(x) # 0 for any x E D. The function q is called the quotient off by g and is denoted by f/g. Thus,
.
(f/g) (x) = f(x)/g(x) (g(x) # 0 for any x E D) .
E
Remark 5 In case g(x) = 0 for some x E D, we can consider the set, say D , all those ' of
j x values of x for which g(x) # 0, and define f/g on D* by setting (f/g)(x) = f(x)/g(x) + D*.
Example 4 Consider the functions f : x + x2and g : x + x3. Then the functions f + g, f  g, fg are defined as
(f + g)(x) = x2 + x3, (f  g)(x) = x2  x3' (fg) (x) = x5 Now, g(x) = 0 G x3 = 0 G x = 0. Therefore, in order to define the function f/g, we shall consider only nonzero values of x. If x # 0, f(x)/g(x) = x2/x3= 1/x. Therefore f/g is the function f/g : x
+ l/x,
whenever x # 0.
All the operations defined on functions till now, were simihr 10 the corresponding operations on real numbers. In the next subsection we are going to introduce an operation which has no parallel in R. Composite functions play a very important role in calculus. You will realise this as you read this course further.
1.6.2 Composite of Functions
We shall now describe a method of combining two function<which is somewhat different from the ones studied so far. Uptil now we have considered functions with the same domain. We shall now consider a pair of functions such that the codomain of one is the domain of the other. Let f : X
+ Y and g : Y + Z be two functions. We define a function h : X + Z by setting .
h(x) = g(f(x)) To obtain h(x), we first take the fimage, f(X), of an element x of X. This f(x) E Y, which is the domain of g. We then take the gimage of f(x), that is, g(f(x)), which is an element of Z. This scheme has been shown in Fig. 11.
z
Real Numbers and Functions
   , g
/
'
Fig. 11
The function h, defined above, is called the composite off and g and is written as gof. Note the'order. We fkst find the fimage.and then its gimage. Try to distinguish it from fog, which will be defined only when Z is a subset of X.
E R and g :.x + 8x + I +x ER. g f i s a . ' function from R to itself, defined by (gof)(x) = g(f(x)) = g(x2)= 8x2+I+ x E R. fog is a function from R to itself defined by (fog)(x) = f(g(x)) = f(8x + 1) = ( 8 ~ + 1 )Thus gof and fog ~. are both defined, but are different from each other.
Example 5 Consider the functions f : x + x2 +x
s used not only to combine functions, but also to look two simpler functions. For example, consider the
We can think of it as the composite (gd) of the functions f : x + 3x + 7 +x g : u + sinu +u E R. Now let us try to find the composites fog and gof of the functions: f:x+2x+3+x~ ~,and~:x+(1/2)~3f2$x~ R
E
R and
f(x)) = g(2x+3) .
Similarly, fog(x) = f(g(x)) = f(xl2  312) =2(x/2  312) + 3 ='x. Thus, we see that gof(x) = x and fog(x) = x for all x E R. Or, in other words, gof and fog are the identity function on R. Wh% we have observed here is true for any two functions f and g which are inverses of each other. Thus, iff : X 2Y and g :Y + X are inverses of each other, then @and fog are identity functions. Since the domain of gof is X and that of fog is Y, we can write this as : gof = ix, fog = i,.
.
This fact is often used to test whether two given functions are inverses of each other.
1.7 TYPES OF FUNCT~ONS
In this section we shall talk about various types of functions, namely, even, odd, increasing, decreasing and periodic functions. In each case we shall also try to explain the concept through graphs.
1.7.1 Even and Odd Functions
We shall firstintroduce two important classes of functions: even functions and odd r functions. ~ o d i d l the function f defined on;R by setting f(x) = x2
+X
A
Y
R.
..
You will notice that f(x) = ( x12 = x2 = f(x) +X
E
R
This is an example of an even function. Let's take a look at the graph (Fig 12) of this function. We find that the graph (a parabola) is symmetrical about the yaxis. If we fold the 'paper along the yaxis, we shall see that the parts of the graph on both sides of the yaxis completely coincide with.each other.. Such functions are called even functions. Thus, a function f, defined on R is even, if, for each x E R, f(x) = f(x). The graph of an even function is symmetric with respect to the yaxis. We also note that if
.
'
0
Fig. 12
X
C
23
Elements of Differential
Calculus
:f 4 the graph of a function is symmetric with respect to the yaxis, the function must be an even function. Thus, if we are required to draw the graph of an even function, we can use this property to our advantage. We only need to draw that part of the graph which lies to the right of the yaxis and then just take its reflection w.r.t. the yaxis to obtain the part of the graph which lies to the left of the yaxis.
E
"t
E 12) Given below are two examples of even functions, alongwith their graphs. Try to
convince yourself, by calculations as well as by looking at the graphs, that both the functions are, indeed, even functions. a) The absolute value function on R f : x + Ixl The graph off is shown alongside. b) The function g defined on the set of nonzero real numbers by setting g(x) = l/xZ,x # 0. The graph of g.is shown alongside.
4
Y
5 i  i o
i
2
3 X
Now let us consider the function f defined by setting f(x) = x3 +x E R. We observe that f(x) = = x3 = f(x) +X E R. If we consider another function g given by g(x) = sin x we shall be able to note again that g(x) = sin(x) = sinx = g(x). The functions f and g above are similar in one respect: the image of x is the negative dY the image of x. Such functions are called odd functions. Thus, a function f defined on R is said to be an odd function if f(x) = f(x) +x E R.
L
If (x, f(x)') is a point on the graph of an odd function f, then (x, f(x)) is also a point on it. This can be expressed by saying that the graph of an odd function is symmetric with respect to the origin. In other words, if you turn the graph of an odd function through 180"about the origin you will find that you get the original graph again. Conversely, if the graph of a function is symmetric with respect to the origin, the function must be an odd function. The above facts are often useful while handling odd functions.
E
A
E 13) We are giving below two functions alongwith their graphs. By calculations as well as
by looking at the graphs, find out whether each is even or 6dd. a) The identity function on R: f:*tx b) The function g defined on the set cf nonzero real numbers by sejting g(x) = ltx, x f '0.
Y
Y4
4 3 2 1
321
.x
(a)
3!>
4
*(b)
1
While many of the functions that yoti will come across in this course wil! turn out to be eithereven or odd, there will be many more which will be neither even nor odd. consider, for ex'ampie, the function f:x+(x+1)2 = i22x + 1. IS f(x) = f(x) +x E R? Here f( x) = (x The answer is 'no'. Therefore, f is not hn even function. Is f(x) = f(x) +x E R? Again, the answer is 'no'. Therefore f is not an odd function. The same conclusion io'uld have been drawn by considering the graph o f f which is given in Fig.13. ' You will observe that the graph is symmetric neither with respect to the yaxis, nor with respect to the origin. Now there should be no difficulty in solving the exercise below. E 14) Which of the following functions are even, which are odd, and which are neither even nor odd? a) b) C) d) x+x2+1 +XE R x+x3I, + x ~ R X~COSX, ER +X x + x 1x1 + X E R e) f ( x ) =
Real Numbers and Functions
Fig. 13
t
0, if x is rational 1, if x is irrational
1.7.2 %on%one Functions
In this subsection we shall consider two types of functions: i) Increasing and ii) Decreasing Any function which conforms to any one of these types is called a monotone function. Does the profit of a cornpany increase with production? Does the volunie of gas decrease with increase in pressure? Problems like these require the use of increasing or decreasing 'functions. Now let us see what we mean by an increasing function. Consider the functions g and h defined 1 if x l O g ( x ) = x3 and h ( x ) = if > 11
/
{
Note that whenever x, > x,, we get x,D > xI3,that is, g(x2) > g(x,). In other words, 9s x increases, g(x) also increases. This fact can also be seen rrom the graph of g shown in Fig. 14.
Let us find out how h(x) behaves as x increases. In this case we see that if x, > x,, then h(x,) 2 h(x,). (You can verify this by choosing any values for x, and x?). Equivalently, we ean say that h(x) increases (or does not decrease) as x increases. The same can be seen from the graph of h in Fig.! 5.
Fig. 14
Functions like g and h above are called increasing or nondecreasing functions. Thus, a function f defined on a domain D is said to be increasing (or nondecreasing) if, for every pair of elements x, x2 E D, x, > x, a f(x2) 2 f(x,). Further, we say that f is strictly increas?ng if x, > x, f(x2)> f(x,) (strict inequality). Clearly, a e function g : x + x3 discussed above, is a s&ctlyincreasing function; while h is ndt a strictly increasing function. We shall now study another concept which is, in some sense, complementary to that of an increasing function. Consider the function f, defined on R by setting
Fig.15
25
Elements of Differential Calculus '
The graph o f f , is as shown in Fig. 16.
:+:
1
~ r o m graph we can easily see that as x increases f, does not increase. the That is, x, > x, 3 f,(x,) 5 f,(x,) or f,(x,)
I=f,(x,)
Now considerlhe function f,: x + x3 (x E R) The graph of f, is shown in Fig. 17.
 x13 f,(x,) < f,(x,), we find that as x increases, f,(x) Since x, > x , x,S > xI3* x:< decreases. Functions like f, and f, are called decreasing or nonincreasing functions. The above two examples suggest the following definition:
A function f defined on a domain D is said to be decreasing (or nonincreasing) if for every pair of elements x,, x,, x, > x, f(x,) 5 f(x,). Further, f is said to be strictly decreasing if x, > X I f(x,) c f(x,).
*
*
Fig.16
*
*
We have seen that, of the two decreasing functions f, and f,, f, is strictly decreasing, while f, is not strictly decreasing. A function f defined on a domain D is said to be a monotone function if it is either increasing or decreasing on D.
\
All the four functions (g, h, f,, f,) discussed above are monotone functions. The phrases P 'monotonically increasing' and 'monotonically decreasing' are often used for 'increasing' and 'decreltsing', respectively.
While many functions are monotone, there are many others which are not monotone. Consider, for example, the function
f : x + x2 (XE R). You have, seen the graph o f f in Fig 12. This function is neither increasing nor decreasing. If we find that a given function is not monotone, we can still determine some subsets of the domain on which the function is increasing or decreasing. For example, the function f(x) = x2 is strictly decreasing in ], 0 and is $.rictly increasing in [O,[. 1
\
\
\
E
E 15) Given below are the graphs of some functions. Classify them as nondecreasing. strictly decreasing, neither increasing nor decreasing:
173 Periodic Functions ..
In this section we are going to tell you about yet another important class of functions, known as periodic functibns. Periodic functions occur very frequently in application of mathematics to various branches of science. Many phenomena in nature such as propagation of water waves, sound waves, light waves, electromagnetic waves etc. are periodic, and we need periodic functions to describe them. Similarly, weather conditions and prices can also be described in terms of periodic functions. Look at the following patterns:
+
:aa~aaaaa


.... .. ... ../.. ..*....... *.....
~oob'ddoddb&@~bi
(a)
0))
Fig. 18
You must have come across patterns similar to the ones shown iR Fig. 18 on the borders of sarees, wall papers etc. In each of these pattems~a design keeps on repeating itself. A similar situation dwurs in the graphs of periodic functionstLook at the graphs in Fig. 19.
Y4 '
*
.
(b)
In each of the figures shown b b v e the graph consists of a certain pattern repeated infinitel; many times. Both these graphs represent periodic functions. To understand the situation, let us examine these graphs closdly. Consider the graph in Fig. 19kilj. The portion of the graph lying between x = 1 and x = 1 is the graph of the function x + I x I on the domain 1 I x I 1. This portion is being repeated both to the left as well as to the right, by translating (pushing) the graph through two units along the xaxis. That is to say, if x is any point of [I, I], then the ordinates at x, x f 2, x f4, x f 6, .... are all equal. The graph therefore represents the function f defined by f(x) = I x I, if 1 I x I1 andf(x + 2) =f(x). The graph in Fig.l9(b) is the graph of the sine function, x + sin x, w x E R. You will notice that the portion of the graph between 0 and 2nis r eated both to the right and tb the left. You know already that sin (x + 2n) = sin x, +x E . We now give aprecise meaning to the term "a periodic function".
f
A function f defined on a domain D is said to be a periodic function if there exists a positive real number p such that f(x + p) = f(x) for all x E he he number p is said to be a period off. If there exists a smallest positive p with the property described above, it is called the period off. As you know, tan (x + nn) = tan x +An E N. This means that nn, n E N are all periods of the tangent function. The smallest of these, that is n, is the period of the tangent function. See if you can do this exercise.
27
Elements of Differential Calculus
E
E 16) a) What are the periods of the functions given in Fig. 19(a) and (b)?
b) Can you give one other period of each of these functions?
*
As another example of a periodic function, consider the function f defined on R by setting f(x) = x [XI Let us recall that [XI stands for the greatest integer not exceeding x.. The graph of this function is as shown in Fig. 20. From the graph (as also by calculation) we can easily see that f(x + n) = f(x) y x E R, and for each positive integer n.
Fig. 20
The given function is therelore periodic, the numbers 1,2, 3,4 being all periods. The smallest of these, namely 1, is the period. Thus the given function is periodic and has the period 1.
Remark 6 Monotonicity and periodicity are two properties of functions which cannot coexist. A monotone function can never be periodic, and a periodic function can never be monotone.
In general, it may not be easy to decide whether a given function is periodic or not. But sometimes it can be done in a straight forwarp manner. Suppose we want to find whether the function f : x + x2 f f  xE R i s periodic or not. We start by assuming that it is periodic with period p. Then we must have p > 0 and f(x + p) = f(x) +x *(x+p)2=x2 ffx * 2 x p + p 2 = o . +tx
* p(2x + p) =.o
ff x
Considering xi # p/2, we find'that 2x + p # 0. Thus, p = 0.This is a contradiction. Therefore, there does not exist.any positive number p such thatf(x+p) = f(x), % E consequently, f is not periodic.
R and,
E 17) Examine whether the following functions are periodic or not. Write the periods of the periodic functions. a) x + cos x b)x+x+2 4 C) x + sin 2x d)x+tan3x f) x + sin x + sin 2x e) x + cos (2x + 5 )
peal Numbers and Functlm
E
E 18) The graphs of three functions are given below: classify the functions as periodic and nonperiodic:
E ~ d Calculus
~
~
n
t
i
a
E
l
E 19) Is the sum of two periodic functions also periodic?Give reasons for your answer.
We end with summarising what we have discussed in this unit.
1.8 SUMMARY
In this unit we have 1 briefly revised the basic properties of real numbers, 2 defined the absolute value of a real number x as Ixl=xifx2O =xifxcO
3 discussed various types of intervals in R Open : ]a,b[ = ( X ER:acxcb) closed: [a,b] = ( X E R:alxlb J semiopen: ]a,b] = { x E R : a < x I b) or La.bl = { X ER:alxcb), where a, b E R,
4 5
defined a function and discussed various types of functions along with their graphs: oneone, onto, even, odd, monotone, periodic. defined composite of functions and discussed the existence of the inverse of a function.
1.9 SOLUTIONS AND ANSWERS
E 1) a) The set (1,2, 3. ...) has a lower bound, e.g., G. D) The set (. . . . . . . . . . , 3, 2, 1,0, 1,2, . . . . . . . . ) does not have a lower bound C) ~ h e g . l . b o f t h e s e t ~ = ( 1 . ~ / ~ , ~ / ~ /, , . . . ) is0,andOe S. ........1, d) (x : x E R and 1 S x 5 2 ) is a bounded set as it is bounded above by 2 and below
by 1.
E 2) a) A real number p is positive if p > 0. Hence 0 is a lower bound for the set P of
positive real numbers. Thus the set P is bounded below. Its infimum is 0. b) A real number,r is the infimum of a set S C R if and only if the following conditions are satisfied: i) r  x for all x E S. 5 ii) For each E > 0, there is y E S such that y c r + E The set Pin a) above has infimum 0, since i) 0 c p for all p E P and ii) foreachea.0 there is el2 E Psuch that ~ / < O + E = E 2 E3) a) I x I = m a x { x ,  x ) . H e n c e x = O = , I x I = O a n d Ixl=O+max. (x,x) = O = , x = O . b) There are three cases 1) x 2 0, y 2 0 2) x and y have opposite signs
3)xSq.ySO. We take 2). Suppose x > 0, y c 0, then,
1x1 =max ( x .  x ) = x , I y I = m a x ( y ,  y ] =  y .
xy<0~IxyI=xy=xx(y)=IxIIyI.
Real Numbers and FuncUonr
1) and 3) can be proved similarly. c) Ifx>O,IxI=xandI l/xI=l/x=l!IxI If x < 0, I x I = x and Il/x I = l/x = 1/ I x I d) Ixyl = Ix+(y)lS I x l + l  y I = l x l t l y l e) I x + y + z l = I ( x + y ) + z l S l x + y l + l z l l l x l + l y l + l z l f) I x y ~ I = l ~ y l i z l = l x l l y l l z l
E 4) a) False
b) True
c) True
d) False
E 8) b) is oneone E 9) a) is onto
Hence f is oneone. IfyE X,putx=Y1
* . Then r E X and y = f(x). Hence, f is onto.
E 12) a) f(x) = l x I * f( x) = 1 x I = I x I = f(x). Hence, f is even. b) g(x) = 1/x2a g(x) = l / (  ~ ) ~g(x) Hence, g is even. = E 13) a) f(x) = x a f( x) =  x =  f(x). Hence, f is odd. b) g(x) = l/x g(  x) =  l/x =  g(x). Hence, g is odd. E 14) a), c) and e) are even d) is odd b) is neither even nor odd E 15) a) neither increasing, nor decreasing b) nondecreasing c) strictly decreasing E 16) The period of the function in Fig.19 a) is 2. Other periohs are 4,6,8, .......... The period of the function in Fig. l9b) is 2x. Other periods are 4x, 6 ~.......... , E 17) a) Periodic wi$ period 2x , Since cos (x + 2z)= cos x for all x. b) not periodic c) Periodic with period x. d) Periodic with period d3. e) Periodic with period x. f) Periodic with period 2x. E 18) a) apd b) are periodic, c) is not. E 19) No. For example, r  [XIand lsin XI are periodic, but their rum is not.
UNIT 2 LIMITS AND CONTINUITY
Structure
2.1 2.2 Introduction
Objectives
32
Limits
Algebra of Limits (or Limits as x Onesided Limits
32
+
 )
43
2.3
Continuity
Definitions and Examples Algebra of Continuous Functions
2.4 2.5
Summary Solutions and Answers
48 48

The last unit has helped you in recalling some fundamentals that will be needed in this course. We will now begin the study of calculus, starting with the concept of 'limit'. As you read the later units, you will realise that the seeds of calculus were sown as early as the third century B.C. But it was only in the nineteenth century that a rigorous definition of a limit was given by Weierstrass. Before him, Newton, d' Alemben and Cauchy had a clear idea about limits, but none of them had given a formal and precise definition. They had depended,%more less, pn intuition or geometry. or The introduction of limits revolutionised the study of calculus. The cumbersome proofs which were used by the Greek mathematicians have given way to neat, simpler ones. You may already have an intuitive idea of limits. In Sec.2 of this unit, we shall give you a precise definition of this concept. This will lead to the study of continuous functions in Sec.3. ~ o ' sof the functions that you will come across in this course will be continuous. We t shall also give you some examples of discontinuous functions.
Objectives
After ieading this unit you should be able to: calculate the limits of functions whenever they exist, identify points of continuity and discontinuity of a function.
2.2 LIMITS
In this section we will introduce you to the notion of 'limit'. We start with considering a situation which a lot of us are familiar with, such as train travel. Suppose we are travelling from Delhi to Agra by a train which will reach Agra at 10,00 a.m. As the time gets closer and closer to 10.00 a.m., the distance of the train from Agra gets closer and closer to zero (assuming that the train is running on time!). Here, if we consider time as our independent variable, denoted by t and distance as a function of time, say f(t), then we see that f(t) approaches zero as t approaches 10. In this case we say that the limit of f(t) is zero as t tends to 10. Now consider the function f : R + R defined by f(x) = x2 + 1. Let us consider Tables liij and l(b) in which we give the values of f(x) as x takes values nearer and nearer to 1. In Table l(a) we see ;dues of x which are greater than 1. w e can also express this by saying that x approaches 1 from the right. Similarly, we can say that x approaches 1 from the left in
Tahla 1 fhl
Table l(a)
Limits and iontinulty
Table l(b)
We find that, as x gets closer and closer to 1, f(x) gets closer and closer to 2. Alternatively, we express this by saying that as x approaches 1 (or tends to I), the limit of f(x) is 2. Let us now gi.ve a precise meaning of 'limit'.
Definition 1 Let f be a function defined at all points near p (except possibly at p). ~ ebe t a real number. We say that f approaches the limit L as x approaches p if, for each real number E >O, we can find a real number 6 > 0 such that
As youknowfromunit l , I x  p l < 6 m e a r . s t h a t x ~ ] p  6 , p + 6 [ a n d O < I x  p I means that x # p. That is, 0 < I x  p I < 6 means that x can take any value lying between p  6 and p + 6 except p. The limit L is denoted by lim f (x ), We also write f(x) +L as x + p..
x
t
This definition of limit was fust ~ stated by Karl Weierstrass, around
1850.
(E epsilon) and 6 (delta) are Greek leners used to denote real numbers.
'P
'+ ' denotes 'rends to'
The E G definition do& not give us the value of Lilt just helps us check whether a given number L is' the limit of f(x).
Note that, in the above definition, we take any real number E >O and then choose some 6 > 0, so that L  E < f(x) < L + E , whenever I x  p I < 6, that is, p  6 < x < p + 6. x In Unit 1 we have also mentioned that I  p I can be thought of as the distance between x and p. In the light of this the definition of the limit of a function can also be interpreted as: Given E > 0, we can choose 6 > 0 such that if we choose x whose distance from p is less than 6, then the distance of its image from L must be less than E. The pictures in Fig. 1 may help you absorb the definition.
'

for each
&
>0
3 a 6 > 0 s.t.
O.<lxcks
Remember, the number & is given' first and the number 6 is to be produced.
An important point to note here is that while taking the limit of f(x) as x + p, we are concerned only with the values of f(x) as x takes va!ues closer and closer to p, but not when x' 1 . x = p. For example, consider the function f(x) = . This function is not defined for x = 1, but is defined ford1 other x G R. However, we can still talk about its limit as x + 1. This is because for taking the limit we will have to look at the values of f(x) as x tends to 1, but not when x = I . Now let us take the following examples:
x1
Example 1Consider the function f : R + R defined by f(x) = x3 How can we find lim f (x ) ? x ro
Ekments of Mfferential Calculus
Look at the graph off in Fig.2. You will see that when x is small, x3is also small. As x comes closer and closer to 0,x3also comes closer and closer to zero. It is reasonable to expect that lirn f(x) = 0 as x 0. Let us prove that this is what happens. Take any real number E > 0 Then,I f(x)  0 1 < E e . 1 x3 I <E w I x I < & I n . Therefore, if we choose 6 = &'I3 we get I f(x)  0 I< E whenever ..0 < I x  0 I < 6. This gives us lim f(x) = 0.
x+O
A useful general rule to prove lirn f (x ) = Lis to write down f(x)  L and then express it in
x+a
terms of (x  a) as much as possible. Let us now see how to use this rule to calculate the limit in the following examples. x2 1 Example 2 Let us calculate lirn x+, X1 x2 1 We know that division by zero is not defined. Thus, the function f ( x ) =   is not x1 defined at x = 1. But, as we have mentioned earlier, when we calculate the limit as x approaches 1, we do not take the value of the function at x = 1. Now, to obtain ,

lirn
X+l
x2 1 x21  , we first note that x2 1 = (X 1) (X + I), SO that, x + x 1 x 1
.
X+l
Therefore lim
= lirn (x + 1) x1
ILL
xt,
1
1 for x # 1.
As x approaches 1, we can intuitively see that this limit approaches 2. To prove that the limit is 2, we first write f(x)  L = x + 1  2 = x  1, which is itself in the form x  a, since a = 1 in this case. Let us take any number E > 0 Now, . I(x+~)~~<E&Ix~~<E Thus, if we choose 6 = E, in our definition of limit, we see that
Example 3 Let us prove that lirn (x
xt 3
+ 4) = 13.
That is, we shall prove that $ c > 0, 36 > 0 such that I x2+ 4 13 1 < E whenever ~ Ix3L<6. Here,f(x) L = ( x Z + 4 )  13 =x29, a n d x  a = x  3.
Now we writeIx291 in termsoflx3 1 Thus, apart from I x  3 1 we have a factor, namely I x + 3 1. To decide the limits of I x + 3 1 , , let us put a restriction on 6. Remember, we have to choose 6. So let us say we choose a 6 5 1. What does this imply?
Ix31<6aIx31<1a31<~<3+1
I ;wit\ .ind Continuity
2 < x < 4 a 5 < x + 3 <7. RecallSec.4, Unit 1.
For this we shall try to make 7 1 x  3 1 < E. Now when will this be true? It will be true when I x  3 1 < ED. So this &/7 the value of 6 we were looking for. But we have already chosen is 6 5 1. This means that given E > 0, the 6 we choose should satisfy
r
In other words, 6 = min ( 1, ~ / 1, should serve our purpose. Let us verify this: 7 I i x  3 1 < 6 a I x  3 1 < 1 and Ix31< & / 7 a I ~ ~  9 1 = I ~ + 3 1 I ~  3 1 < 7 . ~ / 7 = & .
r
Remark 1:Iff is a constant function on R,that is, if f(x) = k +x E R,where k is some fixed real number, then lim f (x ) = k .
X ' P
Now please try the following exercises.
E
E 1) show that 1 a) lim g = 1
x+ 1
b) lim
x, ,

x3 1 =3 x1
Before we go further, let us ask, 'Can a function f(x) tend to two diffixent l i d as x tends to p'?
I
The answer 15 NO, as you can see from the following :
Eiements ot Differential
Calculus
Theorem 1 If lim f (x ) = L and lim f(x) = M, then L = M.
X+P
xt P x+ P
Proof : Suppose L # M, then I L  M I > 0. Since lirn f(x) = L , if we take&= then 3 6 > 0 such that Ix~I<~,*I~(x)LI<& Similarly, since lim f (x ) = M , 3 6, > 0 such that
x +P
ILMI
2
,
'
IxpI<6,~lf(x)MI<~. If we choose 6 = min {a,,a,}, then Ix  pl < 6 will mean that Ixpl < 6 and lx  pi < 6,. In this case we will have both I f(x)  L I < &, as well as, I f(x)  M I < &. So that I L  M I = I L  f(x) + f(x)  M I 5 I f(x)  L I
la+blSI al+lbl
,
+ I f(x)  M I <& + &
=2&=lLMI.
That is, we get I L  M I < I L  M I, which is a contradiction. Therefore, our supposition is wrong. Hence L =M. We now state and prove a theorem whose usefulness willbe clear to you in Unit 4. Theorem 2 Let f, g and h be functions defined on an interval I containing a, except possibly at a. Suppose i) f(x)lg(x)Ih(x)+x
x+ a x +a
E
I\{a}
ii) lim f(x) = L = lim h (x ) Then lim g (x) exists and is equal to L.
x+ a
Proof: By the definition of limit, given e > 0 , 3 6, > 0 and 6, > 0 such that I f ( x )  L I < & for O < l x  a I < 6 , a n d I h ( x )  L I < E #or O < I x  a I < 6 , Let 6 = min {a,,6,}. Then, O<lxal<6*lf(x)LI<& andlh(x)LI<E
L&Ih(x)lL+& We also have f(x) I I h(x) +X g(x)
E
I \ {a)
Thus,wegetO<IxaI<6*L&< C(x)Sg(x)Ih(x)IL+& In other words, 0 < I x  a I < 6 3 I g(x)  L I < & Therefore lim g (x) = L .
x+ a
Theorem 2 is also called the sandwich theorem (or the squeeze theorem), because g is being sandwiched between f and h. Let us see how this theorem can be used. Example 4 Given that I f(x)  1 I S 3(x + 1)' + X E R, can we calculatelim f (x ) ?
x+1
We know that 3(x +I), I f(x)  1 I + +%.This means that 3(x 1 I f(x) 1 3 ( x +1), + 1 +x. Using the sandwich theorem and the fact that 3(x + 2 2 lim [3 (x + 1) + 1 ] = 1 = lim [3(x + 1) + 11, we get lim f ( x ) = 1 , ,
x+1 x+1 x+1
In the next section we will look at the limits of the sum, product and quotient of functions.
2.2.1 Algebra of limits
Now that you are familiar with limits. Let us state some basic proofs are beyond the scope of this course.) Theorem 3 Let f and g be two functions such that lim f(x) and lim g(x) exist. Then
x+ P x+ P x+p
of limits. (Their
i) ii) iii)
lirn [f(x)+ g(x)] = lim f ( x ) + lim g(x)
x+ P x+ P
Sum rule Product rule
lirn [f(x)g(x)]=
x+ P
1 lim  = X+P g ( x )
lilipg(x)
, provided lim g (x ) + 0
P ' X
~eci~&al rule
We can easily prove two more rules in addition to the thsee rules given in' Theorem 3. These are: Constant function rule iv) lirn k = k
x+
Limib nnd Continuity
P P
V)
lim x = p
x+
Identity function rule
We shall only indicate the method of proving iv) and v): iv) Here I f(x)  L I = I k  k I = 0 <
E,
whatever be the value of 6.
Using the properties that we have just stated, we will calculate the limit in the following example. Example 5 Let us evaluate lim 3 x 2 + 4x x+ 2 2x + 1 Now lim 2 x + 1 = lirn 2 x + lirn 1 by using i)
x+ 2 x+ 2 x+2
= lirn 2 lim x + lirn 1 by using ii)
,x+2 x2 x+2
= 2 x 2 + 1 = 5 # 0 by using iv) and v)
'. We can use (iii) of Theorem 3. Then the required limit is
lim ( 3 x 2 + 4 x )
x +2
lim ( 2 x + 1)

lirn
x+ 2
3 x 2 + lim 4x
lirn 2x
x+2
+ lim
x+2
1
by using i )
lirn 3 lim x lim x + lim 4 lim x
x+2 x+2 xi2 x+2
lim 2 lim x
+ lim
.
1
by using ii)
E
You can similarly calculate the limits in the following exercises. 3' E 2) Show that lim =3
x11
E
E 3) Calculate lim 2x + 5
x11
2.2.2 Limits as x +
> o in Fig.3. This is a decreasing function of x. In fact, we see from Fig. 3 that f(x) comes closer and closer to zero as x gets larger and larger. This situation is similar to the one where we have a function g(x) gett~ng closer and closer to a value L as x comes nearer and nearer to,some number p, that is when lim g ( x ) = L.
XiP
(or  00 ) = 1 Take a look at the graph of the function f ( ) ~ T ,
The only difference is that in the case of f(x), x is not approaching any finite value, and is just becoming larger and larger. We express this by saying that f(x) + 0 as x,+ m, or lim f(x) = 0. Note that, is not a real number. We write x + w merely to indicate that
xi

x becomes larger and larger. We now formalise this discussion in the following definition.
Oefinition 2 A function f is said to tend to a limit L as x tends to m if, for each C > 0 it is possible to choose K > 0 such that I f(x)  L I < E whenever x > K. In this case, as x gets larger and larger, f(x) gets nearer and nearer to L. We now give another example of this situation.
Example 6 Let f be defined by setting f(x) = 1/x2for all x E R \ ( 0 ) .Here f is defined for
all real values of x other than zero. Let us substitute larger and larger values of x in f(x) = 1/x2and see what happens (see Table 2).
Table 2
We see that as x becomes larger and larger, f(x) comes closer andcloser tozero. Now, let us choose any E > 0. If x >I/&, then 1/x2< E. Therefore, by choosing K = I/&, we find that x > K ~ I f ( x ) l < ~ . T h u s , l i m x)=O. f(
X+=
Fig.4 gives us a graphic idea of how this function behaves as x + .
Fig. 4
Sometimes we also need to study the behaviour of a function f(x3, as x takes smaller and smaller negative values. This can be examined by the following definition.
Definition 3 A function f is said to tend to a limit L as x +  m if, for each E > 0, it is possible to choose K > 0, such that I f(x)  L I < E whenever x < K, The following example will help you ir, understanding this idea.
Example 7 Consider the function f: R + R defined by
The graph off is as shown in Fig.5.
Limits and Continuity
What happens to f(x) as x takes smaller and smaller negative values? Let us make a table (Table 3) to get some idea.
Table 3
L _ _ _ _   J 
I
I
I
We see that as x takes smaller and smaller negative values, f(x) comes closer and closer to zero. In fact 1/(1 + x2)< E whenever 1 + x2> l/&,.thatis, whenever x2 > (I/&) 1, that is, whenever either x < 
K =  1
I. /
I
1'12
1; 
'12
1 or x > 1
:1
'12
 11. Thus, we find that if we take
x+00
,then x <  K=, I f(x) I< E. Consequently, lim f (x ) = 0 .
x+
In the above example we also find that lim f(x) = 0. Let us see how lirn f (x)can be interpreted geometrically.
x +
In the above example we have the function f(x) = 1/(1+ x2),and as x + =,,or x+ f(x) + 0. From Fig. 5 you can see that, as x + = or x +  =, the curve y = f(x) comes nearer and nearer the straight line y = 0,which is the xaxis. Siyilarly, if we say that lirn g(x) = L, then it means that, as x + = the curve y =g(x)
X+
.
comes closer and closer to the straight line v = L. x2 = 1. Example 8 Let us show that lim X+oo (1 + x2) Now, I x2/(1+ x2) 1 I = 1/(1 + x2).In the previous example we have shown that
I 1/(1 + x2)1< E for x > K, where K = ( 1 /  1 I'D. Thus, given E > 0, we choose ~ K = I 1 /  1 Ill2,SO that ~

x>K+

x 1 1 < & T h i s m e a n s t h a txlim = + x2 $1 I
.
We show this geometrically in Fig.6.
Y
t
Elements o Differential f
Calculus
You must be wondering ifall the properties given in Theorem 3 also hold when we take limits as x + m. Y s they do. e, You can use them to solve this exercise.
E 4) Show that a) lirn I/x = 0
x+
b) lirn ( l / x + 3 / x 2 + 5 ) = 5
x+
Sometimes we cannot use Theorem 2 directly, as is clear in the following example. Let us see how to overcome this problem.
.
Example 9 Suppose, we want to find lirn
xt
3x + 1 ' 2x+5
We cannot apply Theorem 3 directly since the limits of the numerator and the denominator, as x + =, cannot be found.' Instead, we rewrite the quotient by multiplying the numerator and denominator by l/r, for x # 0.
Then,  3t ( I h ) , f o r x +O.NOW we use 3x = 2x+5 2+(5/x) Theorem 3 and the fact that:L% I h = 0 (see E 4 a)), to get
+
'
'
? c .
.x+lirn
X
lirn (3 + l / x )
+
(2+5/x)
= . 3 + 0
2+0
3
2
By now you must be used to the various definitions of limits, so can try this exercise.
E 5 ) a) If for some E > 0, and for every K, 3 x > K s.t. I f(x)  L I > E , what will you infer? b) H lirn f (x ) # L ,how can you express it in the E  6 form?
X+P
1
We end this section with the following important remark.
Limits and Continuity
Remark 2 In case we have to show that a'function f does not tend to a limit L as x approaches p, we shall have to negate the definition of limit (also see E5(b)). Let us see what this means. Suppose we want to prove that lim f (x ) # L. Then, we should find some
E > 0 such that for every
6 > 0, there is some x E ]p  6, p + 6[ for which I f(x)  L I > E.
Xt
P
Through our next example we shall illustrate the negation of the definition of the limit of f(x) as x + rn.
> 0 such that for any K (howsoever large) we can always find an x > K such that I l/x  1 I > E Take E = 114. Now, for any K > 0, if we take x = m a (2, K + 1 ), we find that x > K and I Ilx  1 I > 114. This clearly .shows that lirn l l x # 1.
Example YO To show that lirn I/x
x+m #
I , we have to find some
E
x+
b
2.2.3 Onesided Limits
If we consider the graph of the function f(x) = [x], shown in Fig. 7, we see that f(x) does not seem to approach any fixed value as x approaches 2. But from the graph we can say that if x approaches 2 from the left then f(x) seems to tend to 1. At the same time, if x approaches 2 from the right, then f(x) seems to tend to 2. This means that the limit o f f exists if x approaches 2 from only one side (left or right) at a time. This example suggests that we introduce the idea of a onesided limit.
I
Fig. 7
Definition 4 Let f be a function defined for all x in the interval ]p, q[. f is said to approach a limit L as x approaches p from right if, given any E > 0, there exists a number 6 > 0 such thatp<x<p+ 6+lf(x)11~6.
In symbols we denote this limit by lirn f (x ) = L
x +p+
Similarly, the function f : ]a, p[ + R is said to approach a limit L as x approaches p from the left if, given any E > 0, 3 6 > 0 such that p  6 < x < p a I f(x)  L I < E. This limit is denoted by lirn f (x ) .
x + p
N ~ t that in computing these limits the values of f(x) for x lying on only one side of p are e taken into account. Let us apply this definition to the function f(x) = [XI.We know that for x E [I, 2[, [x] = 1. That is, [XIis a constant func!ion on [I, 2[. Hence lirn [ x ] = 1 .Arguing similarly, we find x+ 2 that since [x] = 2 for all x E [2,3[, [x] is, again, a constant functidn on [2,3[, and' lim, [x] = 2.
x 2
'
Let us improve our understanding of the definition of onesided limits by looking at some more examples.
Example 11 Let f be defined on R by setting
f(x)=
y,
when x'#O
f(0) = 0 We will show that lim f ( x ) equals  1.
x +o
When x < 0, I x I =  x, and therefore, f(x) = ( x)/x =  1. In order to show ta !ht ? x
f(x)
exists and equals 1, we have to start with any E > 0 and then find a 6 > 0 such that+if. . 6 s x < 0, then I f(x) (1) I < E. ,: Since f(x) = 1 for all x < 0, I f(x)  ( 1) I = 0 and, hence, any number 6 > 0 will work. .Therefore, whatever 6 > 0 we may choose, if  6 < x < 0, then I f(k)  (1) I = 0 < E . Hence, lim f ( x ) = 1.
x+
0
Example 12 f is a function defined on R by setting ' f(x) = x  [x], for all x E R. Let us examine whether lim f ( x ) exists.
x1
Recall (Unit 1) that this function is given by f(x) = x, if 0 S x < 1. f(x) = x 1 if 1 < x < 2, and, in general f(x) = x  n if n Ix < n + 1 (see Fig.8)
Llements of Dit'feranlinl Calculus
Y
A
Fig. 8
Since f(x) = x for values of x less than 1 but close to 1, it is reasonable to expect that lim f(x) = I .Let us prove this by taking any E > 0 and choosing 5 = min ( 1, €1. We find
x41
1 6 < x < 1 *
x11
f(x)=xandIf(x)1 I = I x  1 I<65&.
Therefore, lim f(x) = 1. Proceeding exactly as above, and notlng that f(x) = x  1 if 1 I x < 2, we can similarly prove that lim, f (x)= 0. +I
E
E 6 ) Rovathat a)
b)
i5yX,40+

[ X I =1
1x1 lim  = 1 x (x + 2)lx(
C) limx4 0
= 2
I
Before going further, let us see how the concepts of onesided l~mit li.t~itare connected. and
Theorem 4 The following statements are equivalent.
Limiti and Continuity
i) ii)
lirn f (x) exists
X+P
lirn f (x ) and lim f (x ) exist and are equal.
x+ 'p X+ F
Proof :To show that i) and ii) are equivalent, we have to show that i) + ii) and ii) 3 i). We first prove that i) 3 ii). For this we assume that lirn f(x) = L. Then given E > 0, x +P 36>OsuchthatIf(x)LI<&forO< I x  p I < 6 . Now,O< I x  p I < 6 3 p < x < p + 6 a n d p  6 < x < p . T h u s , w e h a v e I f ( x )  L I < ~ f o r p < x < p + 6 a n d f o r p  6 < x < p . T h i s m e a n s t h a t l i m f ( x ) = L = lim f(x).
x+ px+ p+
x3 p
We now prove the converse, that is, ii) 3 i). For this, we assume that lirn f (x ) = lirn f (x ) .= L. Then, given E > 0 , 3 6,, 6, >O
x+ p+
such that
This means that I f(x) L I < E whenever 0 < I x  p I < 6. Hence lirn f(x) = L.
x 'P
i), proving that they are equivalent. From Theorem 4, we can conclude that if lirn f(x) exists, then lim f (x ) and lirn f (x ) X'P x+ p+ x+ F also exist and further lirn f ( x ) = lirn f(x) = lirn f (x )
X+P x rp+ x + p
Thus, we have shown that i)
ii) and ii)
+
Remark 3 If you apply Theorem 4 to the function f(x) = x  [x] (see Example 12), you will see thatljm (x [x] ] does not exist as lim+ ( x 7 [x] } ?t lim ( x  [x] ),
x+ 1 x+l x+l
We shall use this concept of onesided limits to define continuous functions in the next section.
2.3 CONTINUITY
A continuous process is one that goes on smoothly without any abrupt change. Continuity of a function can also be interpreted in a similar way. Look at Fig. 9. The graph of the function f i n Fig.9 (a) has an abrupt cut at the point x = 3, whereas the graph of the function g in Fig.9 (b) proceeds smoothly. We say that the function g is continuous, while f is not.
(a)
Fig. 9 :(a) Graph o f f (b) Graph of g
(b)
Continuous functions play a very important role in calculus. As you proceed, you will be able to see that many theorems which we have stateqin this course are true only for continuous functions. You will also see that continuity is a necessary ~ondition the for derivability of a function, and that it is a sufficient condition for the integrability of a function. But l h usgive a precise meaning to "a continuous function" now.
I
2.3.1 Deflnitlons and Examplcs
In this section we shall mive you the definition and some examples of a continuous functiur~. We shall also give you a short list of conditions which a function must satisfy in order to be continuous at a point.
kflnltlon S k t f be a function defined on a domain D, and let r be a positive real number such that the interval ]p r, p t r [ c D. Pis said to be ~ntinuous x = p if lirn f (x ) = f (p) . at

X) 1)
By the definition of limit this means that f is continuous 'at p if given E > 0, 3 6 > 0 such that I f(x) f(p) I < € whenever I x p I < 6. To clarify this concept let us look at an example,


Example 13 k t us check the continuity of the function f :R + R such that f(x) = x atthe point x = 0. . Now, f(0) = 0. Thus we want to know if lirn f (x ) = 0 .
This is aue because given e z 0,we can choose 6 = E and verify that I x I < 6 3 I f(x) I < E Thus f is continuous at x = 0.
x 0
Remark 4 f is continuous at x s p provided the following two criteria are met:
i)
*+ P
I+ P
lirn f (x ) exists.
ii) lirn f(x )= f @ ) .
Fil. 10 : (a) Graph d f @I *rph dl/@ Fig. 10 shows two discontinuous functions f and g. Criterion i) is not met by f, whereas g .fails to meet criterion ii). If you read Remark 3 again, you will find.that f(x) = x  [XI not is continuous at x =l. But we have seen that we can calculate onesided limit8 of f(x) = x  [XI at x = 1. This leads us to the following definition.
Definition 6 A function f : ]p, q[& R is said to be continuous from the right at x =p if lim+f(x )= f @). We say that f is continuous from the left at q if lirn f(x) = f(q) . Thus, f(x) =x  [x] is continuous from the right but not from the left at x =1 since lim f(x)# f(1) and lim+ f(x) = f(1) = 0.
Xrl
x+ 1
x9 P
x9 q 
E
E 7) Give E  6 definition of continuity at a point from the right as well as from the left.
E 8) Show that a function f: ]p  r, p + r[ + R is continuous at x = p if and only i f f is continuous from the right as well as from the left at x = p (Use Theorem 4).
1
Now that you know how to test the continuity of a function at a potnt, let us go a step further, and define continuity of a function on a set.
Definition 7 A function f defined on a domain D is said to be continuous on D, if it is continuous at every point of D. Let us see some more examples. Example 14 Let f(x) = xn for an x 6 R and any n E ZC.Show that f(x) is continuovs at x = p f o r a l l p ~ R. : We know that lirn x = p for any p E R. Then, by F e product rule in Theorem 3, we get
x+ P
&I\ttw set 0 , IIO~I~IVL Integers.
lim x " = (lim x) (lim x ). .
x+p x+p
(n times) ... (limP x ) *+ = pp . . . . . p (n times) = d . Therefore , lim f(x] exists and equals f(p). Hence f is
X
x+a
+D
continuous at x = p. Since p was any arbitrary number in R, we can say that f is continuous on R.
Remark 5 Using Example 14 and Theorem 3, we can also. probe that polynomial a, + a,x + . . . . . . + anxn,where a,, a, . . . . . . an E R, is continuous on R, that is, lirn (a, + a,x + . . . . . . + anxn)= a, + alp + . . . . . . + anpnfor all p E R.
x+ P
Example 15 The greatest integer fu~ction : R + R : f(x) = [x] is discontinuous at x = 2. f To prove this we recall our discussion in Sec. 2 in which we have proved that lime f(x) = I
x+ 2
and lirn f (x ) = 2. Thus, since these two limits are not equal, lim f (x ) does not exist.
x+2+ x+ 2
Therefore, f is not continuous at x = 2 because the first criterion laid down in Remark 4 is not met.
Example 16 Let f(x) = I x I for all x e R. This f is continuous at x = 0. Here f(x) = x, if x 2 0, and f(x) =  x if x < 0. YOU can s that m
lirn f ( x ) = lirn x = O=f(O)and
r+g+
x+ 
x + ~ +
x+ 0 x+ 0
'lim f(x ) = lim ( x ) = 0 = f(0). Thusr lirn f (x ) exists and equals f(0). Hence f is
0
continuous at x = 0 .   Note :f is also continuous at every other point of R. (Check this statement). x2 I Example 17 Suppose we want to find whether f(x) = is continuous at.x = 0. x1 In Fig. 11 you see the graph of f. It is the line y = x + 1 except for the pointi(l,2).

We can write fix) a x + I for all
x + 1.
Elements of Uiflewntial Calculus
f(O) = (02 l)/(O  I) = 1 and lirn f ( x ) = lirn ( x 2  1)/(x  I ) = lirn ( x + I ) = 1
x 4
x 4
0 x +o xi 0 lirn f (x )= f (0), so that f is continuous at x = 0.
0
The exponential function f(x) = exand the logarithmic function f(x) = In x are continuous functions. You can check this by looking at their graphs in Unit 1. Similarly, x + sinx and x + cosx are continuous. x + tanx is continuous in 1 p/2, n/2[. This fact is quite obvious from the graphs of these functions (We have given their graphs in Unit 1). We shall not attempt a rigorous proof of their continuity here.
Caution :Checking the continuity of a function from the smoothness of its graph is not a
foolproof method. If you look at the graph (Fig. 12) of the function x + x sin (l/x), you will find that it has no breaks in the neighbourhood of x = 0. But this function is not continuous. Observe that the graph oscillates wildly near zero.
Fig.12
E
E 9) Show that the function f :R .+ R given by f(x) . l/(xz  9) is continuous at all points of R except at x = 3 and x = 3.
Now that we know how to check whether a function is continuous or not, let us go further, and talk about'the continuity of some combinations of functions.
Limits and Continuity
2.3.2 Algebra of Contihuous Functions
Let f and g be functions defined and continuous on a common domain D E R, and let k be any real number. In Unit 1, we defined the functions f + g, fg, f/g (provided g(x) + 0 any where in D), kf and I f t The following theorem tells us about the continuity of these functions.
Theorem 5 Let f and g be functions defined and continuous on a common domain D, and let k be any real number. The functions f + g, kf, I f I and fg are all continuous on D. If g(x) # 0 anywhere in D, then the function f/g is also continuous on D.
rl
We shall not prove this theorem here. In Unit 1, you have studied the important concept of composite functions. In Theorem 6, we will talk about the continuity of the composite of two continuous functions. Here again, we shall state the theorem without giving proof as it is beyond the level of this course.
Theorem 6 Let f : Dl + D, and g : D, continuous on D,. (Dl, D,, D, C R).
+ D, be continuous on their domains. Then gof is
Example 18 To prove that f : R + R : f(x) = (x2 + is continuous at x = 0, we'consider the functions g : R + R: g(x) = x' and h : R + R :h(x) = x2 + 1. You can check that f(x) = goh (x). Further. by Remark 5, h is continuous on R, and g is also continuous on R. Thus, goh = f is continuous on R.
Let us see if the converse of the above theorems are true. For example, iff and g are defined on an interval la, b] and iff + g is continuous on [a, b], does that mean that f and g are continuous on [a. b] ? No. Consider the functions f and g over the interval [0,1] given by
Then nelther f nor g is continuous at x F 112. (Why?) But (f + g) (x) = 1+x E [O, I]. Therefore, f + g is continuous on [0, I]. Now. if I f I is continuous at a point p, must f alsp be continuous at p? Again, the answer is
NQ. 'Take, for example, the function f :R + R given by
Then I f(x) I = 1 in R and hence I f I is continuous. But f is not continuous at x = 0 (Why?)
E
E10) I f f : . R + R i s d e f i n e d b y f ( x ) = is f continuous at a) x = 1 b) x =  3/2?
Element#of DUlemtlal Clleuhfi
Before we end this unit. we shall state an important theorem concerning continuous functions. Once again, we won't prove this theorem here. But try to understand its statement because we shdll be using it in subsequent units.
Theorem 7 (Intermediate Value Theorem) Let f be continuous on the closed interval [a,b].
Suppose c is a real number lying between f(a) and f(b). (That is, f(a) < c < f(b) or f(a) > c > f(b)). Then there exists some xo E ]a.h[, such that f(x0)= C. How can we interpret this geometrically? We have already seen that the graph of a continuous function is smooth. It does not have any breaks or jumps. This theorem says that, if the points (a, f(a)) and (b, f(b)) lie on two opposite sides of a line y = c (see Fig.13). then the graph off must cross the line y = c.
Note that this theorem guarantees only the existence of the number x,. It does not tell us how
to find it.'Another thing to note is that this xo need not be unique.
That brings us to the end of this unit.
2.4 SUMMARY
We end this unit by summarising what we have covered in it.
I
Tlw limi: of a function f a t a point p of its domain is L if given E > 0,36 > 0, such that I f(x) I, I < e whenever ! x  p I < 6.

2 Onesided limits
3 lirn f (x )exists if and only if
X+P
X4P
lim f (x ) and x rim f (x ) h t h exist and are equal. +:
4
A function f is continuous at a point x = p if
lim f(x) = f(p)
X+P
If the function f and p are continuous on D. then so are the functions f + g, fg, ! f I, kf (where k E R)flg (where g(x) ;t 0 in D) 6 The Intermediate Value Theorem : If f is continuous on [a, b] and if f(a) < c < f(b) (or f(a) > c > f(b) ), then 3x, E la, 'b[ such that f(x,) = c.
5
#
2.5 SOLUTrONS AND ANSWERS
E 1) a) Given any E > 0, if we choose S = min ( ~ 1 2112).then ,
IxlI<6$1/2~~>112
Llmlts and Contlnultj3
Given E > 0, if we choose 6 = min ((2/7)~, 1/21, then Ix1 I<l/2 x<3/2 x+2<7/2and
.
Thatis.1~1IcG Hence, lirn
x+l x+l
: : : 1 Ice
3
xJ 1 = 3. 1
X 
E 2) lirn 3/x =
x31
= 311 = 3 lim x
x31
lirn 3
5 lirn x 2
x+ 1 x3 1 x +1
E 4) a) Given E > 0 if we choose K = I/&,then x > K a I l/xOI=I l/xI<I/K=&. Thus, lirn l/x = 0
x+
b) Given E > 0, if we choose K = 11 fi, then x > K * I 1 / ~ ~  0 1 11/x1l< 1/K2=E = Hence. lirn l/x = 0
x+
Now, lirn (1 /x
x+

+ 3/x2 + 5)
= lim l / x + 3 lirn l / x 2 + lirn 5 = 0 + 3 x 0 + 5 = 5 X +x+x.+E 5 ) a). lirn f(x) # L
X+"
b)
3 e > 0, s.r. 4 6 > 0 3 x s.t. I x  p I c 6 and I f(x)  L I > e.
E.
Ii
.I)
Since x  [x] = x  2, 2 5; x < 3, l i m x [ X I = lirn x  2 = 1
"
+3

x 3x+p+
h) .lim (x(/x = lirn x/x = 1,
K+
o+
]XI=
x for x
> 0.
sincelxl=x forx<O
C,
x+o
lim
( x 2 + 2)Ixl (x2 + 2)(x) = lim x x+oX

= lim
x0I
 (x2 + 2 ) ~  2 s
E 7) f is continuous from the right at x = p if +be 5 0 there exists a 6 > 0 set. p c x c p + 6 *lf(x)f(p)l<e f is continuous from the left at x = p if +E > 0 there exists a 6 0 s.t. E
f is continuous at x = p a lim f(x) = f(p)
x 'P
3
lim+ f (x ) = f @) and
X+P
xhp
lirn f (x ) = f @) by Theorem 4
+ f is continuous from right and from left at x = p. I f f is continuous from right and left, a lim f ( x ) = f @ ) = lirn f(x)
x+p+
x'
x' p
a lim f (x ) exists and = f (p)
P
I
Elements of Differential Calculus
E9)
K+
lim f ( x ) = a
1  =  lim x 2  9 a2  9  f ( a ) f o r a l l a e x c e p t a = + 3
x+ a
Hence f is continuous at all points except at f 3, f is not defined at f 3. E 10) a) f is not continuous at x = 1. For E = 1 and any 6 > 0, if x is any noninteger E 11 6, 1 + 61,then I f(x)  f(1) I = I 1 1 I = 2 > E.
b) f is continuous at 312. Since given E > 0, if we choose 6 < 112, then I x  (312) 1 < 112 3  2 < x <  1 x e Z and hence I f(x)  f( 3/2)1= 0 < E.
UNIT 3 DIFFERENTIATION
Structure
3.1 3.2
Introduction
Objectives
51 52
The Derivative of a Function
Slope of a Tangent Rate of Change The Derivative
3.3 3.4
Derivatives of Some Simple ~uncti0r.s Algebra of Derivatives
~erivative a Scalar Multiple of a Function of Derivative of the Sum of Two Functions Derivative of the Product of Two Functions Derivative of the Quotient of Two Functions The Chain Rule of Differentiation
C
3.5 3.6 3.7
Continuity versus Derivability Summary 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 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
,

Newton (16421727)
.
Objectives
After studying this unit, you should be able to:
*
Leibniz (16461716)
;O
a
d.raw a tangent to a given curve at a given paint 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 discuss the relationship between continuity and derivability of a function.
.51
Elements of Differential Calculus
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?
The tangent of the angle which a ,line makes with the positive :direction of the xaxis is called the slope of the line.
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 xaxis, 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 f(x
~ h t indicates that the tangent line will exist only if the limit of s 6x + 0.
+ 6x)  f(x) . exists as
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 Cakulua
Let Q (x, xZ)be any other point on the parahfa. The slope of
Pa =
'
y coordinateof Q y coordirlateof P x coordinate of Q  x coordinate of P

I
Equation of a lime passing through a point (x,,y,)and having slope m is Y y, = m(x x,)


= x2 4 x,2 The tangent at P (2,4) is the limiting position of PQ as x + 2. Therefore, the slope of the tangent at P is x2  4

:2 5 '
I
7GT
(X +2)(k  2) = lim (X+ 2 ) = 4 'x2 x2 Xa The equation of the tangent line will be (y  4) = 4 (x  2).
= lim
Now, how do we draw this tangent? Just mark the point P'by moving a distance 1 unit from P, parallel to the xaxis to the right, and then, moving a distance equal to 4 units parallel to the yaxis 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 = m =e 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?
I
i
I
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]) for smaller and smaller values of 6t. t If 6t is very small. then t + 6t is very near t and so the average velocity during the time interval 5t would be very near the velocity at t. It seems reasonable, therefore, to define the f(t + 6t) f(t) instantaneous velocity at time t to be lim &+o 6t Thus, we have f ( t + 6t) f(t) v = lim 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
I
L
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].
,
I
1
This average rate of change off in [O,6x] is
 6x+55.=&=]
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.
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 xaxis and,distance along yaxis. Let P and Q be points on the curve which correspond to $ = 2 and t, = 4. 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
1
The curve represented by s = (1/2)t2is a parahoh as shownin Fig. 5. P and Q correspond to 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,). 1 2 Further, to calculate the slope of the tangent at P, we choose a point R(2 + 6t, (2 + 6t) )on the curve, near P. Then the required slope is lim (slope of PR),
61 +O
55
Elements of Differential
Calculus
6t(4 + 6t) =2 8to 26t And, what is the velocity at t,? It is lim W t , which is again equal to 2. Thus ihe velocity at lim
i (2 + 6t)? 2 2
Fig. 5
 = sto lim (2+6t)2
t, is the same as the slope of the tangent at P.
8t 0
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~cceleration due 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).
E
E 4) Find the average rate of change of the function f defined by f(x) = 2x2+ 1,+x the interval [l,1 h] and hence evaluate the rate of change of f at x = 1.
E
R in
Differentiation
I
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 realvalued function whose' domain is a subset D on R. Let x E D. If f(x + 6x) f(x) exists, then it is called the derivative off at x. lim 6x +o 6x Now, if we write f(x + 6x) = y + 6y, then derivative off = lim 6yISx. Here 6y denotes the , . &+O change in y caused by a change 6x in x. 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,

f (x,)=
Sx+O
lim
f(xo + 6x) f(x,) 8x
6x0
Thc notation dy/& is due to Leibniz and f(x) is due to Lagrange (1 736 1813).
f(xo + a x )  . f v , ) 6~ we write
If, in the expressinrl P(xo) = lim f(x)  f(xo
0
xo+6x=x,weget6x=xx,,and6x+Owx+x,.
f (x, )= lim
x+x
xXo
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. quantities (lefigth, time, distance, velocity, area, volume) then the If x and y are dir~tensional 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, !: . 'd' fro& dyldx to get ylx. The noation only suggests the fact that the deril aei~r: otl~iil~, 1s ; as a ratio.
, t ~
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 t write down the derivatives of some o 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 Calculus
function, [he power function and the absolute value function. We shall illustrate the m e t h ~ d 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 = lim = lim 0 = 0 SX+ o 6x SX+O 6~ 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 .
Y
4
4
C
*
Example 5 We now show that, if n is a positive integer, then D(xn)= nxnI.
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. lirn
h+ 0
(X + h)n  xn h
= lirn
h+O
( X" + nhxPI + . . . . . . . + h n,
h
xn
(by bionomial theorem)
= nx
ni
' d The result of the above example is very useful. We shall show later, that  (xn) = nxndx for all nonzero x E R even when n is a negative integer. The result also holds for all x >#O if n is any nonzero 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"',
dx
and this we do in the following example.
6 is not defined for x < 0 .
Example 6 We shall show that the function f defined by f(x)=, fi,0 is differentiabbe. x>
We have, lim
h+O
f(x
+ h)h
fix)
= lirn
h+O
mfi
h
Differentiation
 lim
(x
 h+o
h
(
+ h ) x ~ + hf ) i
1
i
1
I
L I
The result of our next example is of great significance. Recall that, in Sec. 2 we mentioned that there are functions that have no tangents at some point (or equivalently. have no derivative there). This example will illustrate this fact. Before giving the example we give some definitions. ( a ) , if it exists, is called the right hand derivative of f(xj at f (a + h )  f(a) x = a and is, written as Rf'(a). Likewise. Lf'(a) = lirn is called the left I h h+Ohand 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).
~f
(a) = lim .f(a
~,
+
)
h
rot
h
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, have to show that we
+o h they are not equal.
ti
lim
10
+
I  101
does not exist. In fact as we shall see, Rf(0) and Lfl(0) both exiht. but
10 + I1 h
Xow Rf '(0)= lirn
tlrOt
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 i Now, 0. f(x i)  f ( x )  lim h + h I  1x1 lim tl +() h b; 0 h x+hx = lim h+Q h = lim h/h = I
IX
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
1
I
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
x41
'
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~s these functions. Can we use this knowledge to find the derivative of f(x)? In of 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~,~~l;l~. m\i!tip;e
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
d But. in view of Exain@lc when x = 0,  (I x I) does not exist. When x 10. d (I x I) = 1 7, dx dx andwhenxc,@, (IxO=1. dx d d Therefore, ;i;(71x1)=7i:xl) dx

 7 when x < 0
and d (7 1x1) does not exist at x = 0.
7when 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) h+O h (f(x + h ) + g(x + h )  f ( x )  g ( x ) } = lim h f(x + h )  f ( x ) g(x + h )  g ( x ) . = h+o lim h + h lim
= lim
h +O
f(x
+ h)  f(x)
h
g(x + h )  g ( x ) t lim 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,
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.
{O+h
d Now. (3x2) dx
2
= 3dX (in view of the theorem)
dx
E
E 9) Differentiate the following:
a) 5x3+ 2 b) a,+a,x + a 2 x 2 +....... + anxn,wherea, E R f o r i = 1, 2,
Differentiation
......., 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. Now, lim fg(x + h )  f g ( x ) 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)) f(x+h)f(x) ' g(x + h )  g ( x ) lim h40 g ( x + h ) ) + l i mo{ hf (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
h 40

{
(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 d & Therefore,  (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, lim
h +o
@(x+ h) @(X) lim (f/gXx + h )  ( f / g ) ( x ) = h h 0 h
= lim = lim
h +o

f(x+h) g(x+h)
f(x)] g(x)
g(x) f (x + h)  f(x) g(x hg(x) g ( x + h ) gix)
+h)
= lim
h 0
f(x + h )  f(x) g(x+ h) g(x)  f(x) h h g(x + h) g(x)
(by adding and subtracting f(x)g(x) from the nutfierator) f(x + h )  f(x) lim g ( x )  f ( x ) g(x + h) &!(XI h h  h+u lim gdx 4 h) , g ( x )
i

1 
h+O

iim g (X
h 0
+ h) hlim0 g (X ) 4
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,
1 Example 1 We 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 . f '(x)=  g ' (XI (~(~11 (from Corollary 1)
 .  mxm' (g ' (X) = mx rn ' by using Example 5)
(xmf
  mxm' =

mxrnl
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 a) x +5
b)
 1 a+bx
+ cx2 + d B
where a, b, c, d are fixed real numbers
E
E 12) Obtain the derivative of l/f(x) by differentiating from first principles, assuming that f(x) # 0 for any x.
Differentiation
E
E .13) Differentiate f(x) =
2 + 5x
+ 7xI
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 dy du given by  = x du dx 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 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 oi. fs x ,. 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
b+O
This means that 6u + 0 as 6x + 0 We assume howgver that 6u # 0.
Elements of Differential Calculus
6Y dy This implies that lirn = lirn SY exists and is equal to 80 SU 6u+O 8u du

. 6~ =dy and lirn   Su du hm 6n10 6~ du ‘5.40SX dx Hence, we get
SY lirn = lirn  . lim SU sy
&+O~X
SX+O~U 6xO~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 dy du and du/dx = 2. Hence we can use the chain rule to get dy = . dx du dx 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 ~ ~ +4x) (3x2 . = 1 o 0 ( ~ ~ ~  1) +2 2
99
(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 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
1
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:
If x #xothen .we may write f(x)  f(xo) = Since f is derivable at x ,
f ( x )  f(xO) ( x x0)  ,o f(x)  f(xo) exists and equals f'(%).
 xo
Elemer?tsof DilPerenlial Calculus
Thus, taking limits as x + x,, we have,
xlX
lim [f (x )  f (xo)]
0
= lim
X+XO
.
f ( x )  f(xo) lim ( X  xO) xxo x+xo
Therefore, lim f (x )  lim f (xo) = 0
X+ Xo
x +wo
That is, x+ Xo f(x ) = X lim~ f (xo) = f (xo) lim +X, 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 attributed the discovery to w 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.
b Sometimes we use Theorem 6 to prove that a g ~ er, function is continuous at a given point. We prove that it$ derivative exists at that point. By Theorem 6 then, the continuity automatically follows.
E
E 15) Is the function f : [0,1] + R :f(x) = (2x
+ 3)
50
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)
 (if it exists) is called the derivative off at x, denoted by ff(x). 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.
lim
f(x + a x )  f(x)
6x40
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' (fg)'=ffg' (fg)' = fg' + gf' (flg )' = g' f  f'g 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.
b)dxx=l
dyI
Equation : (y  112) = (  114) (x  2) orx+4y=4 Equation : (y  1) = 3(x  1)
=3.
E 3) E 4)
area = A = rc$ .

average rate of change off in [I, 1 + h] =
rate of change off at x = 1 = lim
f(1 + h )  f(1) where h may be positive or negati h = lim [ 4 + 2 h ) = 4

h+O
h3 0
Elements of Differential Calculus
E 5)
If x < 0, choose 0 < h < I x I then x+ h < 0 and Ix+hIlxI   ( x + h )  (  X ) h == 1 lirn . h h +o h h Thus f'(x) = I if x < 0. f is derivable f o x c 0. ~
E 6)
a) b)
f' (2)= lim
h+O
.f ( 2 + hh) = a .
ah h
2+h2=1 h a(2+h)+b(ax2+b)
= lim
h+O
f(2)
f' (2) = lim
h+O
= lim
h+O
.
E7)
dy a)  = dx
lim (X + h $  x 3 h h+O
= lirn (3x2 + 3xh + h2) = 3x2
h+O
b) If x > 1, x + 1 > 0, choose h > 0 s.t. h < 1.x + 1 1. Ix + h + 11lx'+ 11 t h e n x + h + 1 >Oand lirn h+O h (X + h ' + 1 )  ( X + 1 ) = lirn = 1. h h+O I x + h + 11lx+ 11 = 1. 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. I f x <  1, lirn
h +O
Cl
9
*
= l i m ~ z ( ' x + h) + I h+O h
 J 2 x + l x J2(x
d2(;(+ 1 2 r 1=
+ h)
.
+I
+4
2 +~1
h)+l+ J2*+ 1
 lim
2
J2(x + h ) , + i
+J
~
i
1 = lim  f ( x )  f ( x + h ) 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  OX5  4 2 ~  7 b) f' ( x ) = =x
2 (5  7
10
~  ~5x4 (2 + 5 x )
X
+ 7xI)
10
(  ~ 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
Elements of Differential
Calculus
UNIT 4, DERIVATIVES OF TRIGONOMETRIC FUNCTIONS
Structure
4.1 4.2
Introduction
Objectives
Derivatives of Trigonometric Functions
Some Useful Limits Derivatives of Sin x and Cos x Derivatives of other Trigonometric Functions
4.3 4.4 4.5 4.6 4.7
Derivatives of Inverse Functions
The Inverse Function Theorem
Derivatives of Inverse Trigonometric Functions
Derivatives of Sin' x and Cos' x Derivatives of Sec' x and Cosec' x
Use of Transformations Summary Solutions and Answers
4.1 INTRODUCTION
In Unit 3 we have introduced the concept of derivatives. We have also talked about the algebra of derivatives and the chain rule which help us in calculating the derivatives of some complex functions. This unit will take you a step further in your study of differential c~lculus.
h this unit we shall first find the derivatives of standard trigonometric functions. We shall then go on to study the inverse function theorem and its applications in finding the derivatives of inverses of some standard functions. Finally, we shall see how the use of transformations can simplify the problem of differentiating some functions.
Objectives
After reading this unit you should be able to: find the derivatives of trigonometric functions state and prove the inverse function theorem use the inverse function theorem to find the derivatives of inverse trigonometric~functions use suitable transformations to differentiate given functions.
4.2 DERIVATIVES OF TRIGONOMETRIC FUNCTIONS
In this section we shall calculate the derivatives of the six trigonometric functions: sinx, cos x, tan x, cot x, sec x and cosec x. You already know that these six functions are related to each other. For example, we have: i) sin2x + cos2x = 1 ii) tan x = sin x/cos x, and many more identities which express the relationships between these functions. As you will soon see, our job of finding the derivatives of all trigonometric functions becomes a lot easier because of these identities. But let us first evaluate some important limits which will prove to be very useful later.
4.2.1 Some Useful Limits
In the next subsection we shall come across lim
t+O
and lim sin t. So let's try to calculate t t 0 these. For this, we first assume that 0 < t < 6 2 and consider a circle with radius 1 unit, given by xZ+ y2 =1 as shown in Fig. 1. The line OT passes through the origin and has slope =tan t. Thwefore, we can write its equation as y = x tan t. This means that the ycoordinate of the p0int.T is tan t, since its
umnrdinate i~ 1
*
Derivatives of Trigonometric
Fullc:ic,ns
F .1 i
From the figure we can see that area of AOPA < area of sector OPA < area of AOTA . . . . . ( 1 ) 1 I Now, the area of A0PA = X 1 x PB =  sin t, 2 1 1 The area of sector OPA = 2 x 1 x t = 2 t'
5
If the sectorial angle is 0, the area of a sector of a circle of radius r is (1/2)P0
The a of oTA = x 1 x tant t t ; Thus, inequality (I) can be written as:
sint < t < tan t
. . . . . (2)
sin Since 0 < t < 1~12, t > 0, therefore, from the lefthand inequality in (2) we get
I
I
O < s i n t < t . . . . . (3) Now, if 1~12< t < 0, then 0 < t . z/2, and applying (3) to  t, 0 < sin (  t) <  t or : 0 <  sin t < t since sin(t) = sin t. This means that if  1~12 t < 0, then t < sin t < 0. . .(4) <
1
I
0
sin t
I
I
t
I ~ / 2
I
XI2
I t
I
sin t
I
0
In Fig. 2(a) and (b) you can see the representation of (3) and (4), respectively.
1!
We can combine (3) and (4) and write  I t I < s i n t < I t I f o r  1 ~ / 2 < t < 1 ~ / 2# 0 . t, You have seen in Unit 2 that lirn I t I = 0. From this we can also say that lirn  I t I = 0. Now applying the sandwich theorem (Theorem 2 of Unit 2) to the functions  I t I, sin t and I t I, we get that lirn sint = 0 We shail use this result to calculate lim cost. As you know, cost = 1  2 sin2t/2. This means
110 t 10 r0
Here we are using various.results about the order relation from Unit 1
v0
t+ 0
lirn cost = lirn (1  2 sin2 t/2)
We can prove that lirn sin2t/2 = by
t+O
I
110
i
= I  2 lirn sin2 t/2
tto
using Theorem 3 of Unit 2 and by noting that t 4 0 t 2 + 0 /
=1
4.
Thw, we get lirn Cost = 1.
10
Now, let's get back to inequality (2) : sint < t < tant for 0 < t< x/2. Since 0 < t < x/2, sint : O ,
" . IL .   . . ,+  . C . 
" * . ,I:.:,I:,. :* c. L. .
( 9 1 I . , .
.
Elements of Diflerentlal Calculus
1 < tlsint < llcost orcost<sint/t< 1 ... (5),O<t<n/2 Now, since sin( t) =  sint, we see that sin ( t)/ ( t) = sintlt. This, alongwith the result COS(t) = cost, shows that the inequality (5) holds even when  nL? < t < 0.Thus,
cost<sint/t< I,  n / 2 < t < d 2 ,
t#O
Now, let us apply the sandwich theorem to the functions cost, sinttt and 1, and take the limits as t +0 This give us: . lim sin t/t = 1
t
.
+o
Example 1. Suppose we want to find out sin 5x sin 3x lim y and lim t+o t0 SIR 7x sin 3x. sin3x Let us first calculate lim . For this we shall write sin3x  x3.1fwe t+ 0 X 3x replace 3x by t in the right hand side, and take the limit as x + 0, we find that t = 3x also sin 3x sin t tends to zero, and lirn  lim x 3


x+O
sin t
t  O t
= 3 lirn
t+O
t 
(See Theorem 3 of Unit 2)
=3 To calculate lirn sin 5 x we start by writing X+O sin 7x sin 5x sin 5x 7x 5 = lim lirn . xx x + ~ s ~ n 7 x +O X 5x sm7x 7 7x 5 =  lim sin 5x lim 7. 7x+0 5x X + O s1n7x 1 = 2 since lim 7 x = = 1 by Theorem 3 of Unit 2 7 X+O slni Iirn (sin7xii~)
f



= 1 or lim cost = 1, the angle t is measured in radians. If in a t +o particular problem, the angles are measured in degrees, we have to first convert these Into radians before using these formulas. Thus,
Remark 1 In lim
t+O
sin t
sin (rct/l80) . sin lim sin to lim   (nt/l80) =  lim . 0 t t+O t 180 t + O ~tn80 180 See if you can solve this exercise now.

E
E 1)Provethata) l i m c o s ( a + x ) = c o s a
x +O
b) lim sin (a + x ) = sin a
x +o
4.2.2 Derivatives of Sin x and Cos x
We shail now find out the derivative of iinx from the first principles. If y = f(x) = sin x, then by definition
dy    lim
dx
h+O
h 2 sin(h/2) cos (X+ h/2) = lirn h+O h
.
sin (x
+ h)  sin x
Derivatives of Trigonometric Functions
Remember the formula sinA  sin B {AB) (A+B) cos 7 = 2sin 2 2

Thus, we get
I
I
dx Now, let us consider the function y= f(x) = cos x and find its derivative. In this case,
d  (sin x) = cos x
dy,  = lim dx
h+O

cos(x + h )  c o s x h 2 sin (h/2) sin (x.+ h/2) = lim h h+O
= 1im
h+o
=  sin x Thus, we have shown that
' h/2 ho ' 0 lim "
sin (x
+ h/2)
d (wsx)'=sinx dx
d (sin x ) ,we could have found out the derivative of cos x Actually, having first calculated by using the formula: cos x = sin (x + This gives us,
a).
d . d (COS )=  (sin (x + ~ 1 2 ) ) X
dx
= cos (X + ~ / 2 ) = sin x 
dx
+ a ) ) = cos (x + a ) can be proved by u s i q the chain rule.
,& (sin (x dx
II
I
In the next subsection we shall find the derivatives of the other four trigonometric functions by using similar formulas. But before that it is time to do some exercises.
E
E 2) Find the derivatives of the following: a) sid? x bJ cosZx c) 5sin7 x sin3x e) cos (sin x).
d) x3cos 9x
Elements o Differential f
Calculus
4.2.3 Derivatives of other Trigonometric Functions
We shall now find the derivatives of i) tan x
i)
ii) cct x iii) sec x iv) cosec x. sin x Suppose y = f(x) = tan x. We know that tan x =
COS X
dy Hence,  = d dx dx

cos x d/dx (sin x )  sin x d/dx (cqs x ) 0 s2
2
2
X
d
dx
(u/v) =
vdu
/dx  u
dv
/dx
v . ~
 ,COS x + sin x
ii)
cos2 x GO? x =sec2 x Now, suppose y = f(x) = cot x. Since cot x  l/tan x, we get

1
 tan  x d/dx (1 2 2 1 d/dx (tan x) till1 X
 _ sec_ x   C O W2 X _ _ _iii) tan2 x Now, let y = f(x) = sec x. Since we know that sec x = l/cos x, proceeding as in ~ i ) ,wt get sin x
2
If you have followed i), ii) and iii) above, you should not have any difficulty in finding the
derivative of cosec x by using cosec x = l/sin x.
Let us summarise our results.
Table 1
Function sin x
COS X
Derivative
COS X
'1
 sin x
 c0secLX
sec x tan x
tan x cot x sec x cosec r;
1
 cosec x cot x
Remark 2 Here again we note that the angle is measured in radians. Thus,
 (sin xq =   sin
d dx
d dx
:8:,) ( 
71
= l ( cos lo
(s)
RX
=  cos x0 180
We shall now see how we can use these results to find the derivatives of some more complicated functions. The chain rule and the algebra of derivatives with which you must have become quite familiar by now, will come in handy again.
Example 2 Let us differentiate i) sec3 x ii) ser: x tan x + cot x i) Let y = sec3 x. If we write u = sec x, we get y = u3. Thus,
ii)
= 3 set:3 x tan x If y = sec x tan x + cot x, then,
dy dx d (secx t a n x ) +  ( c o t x ) dx d d =S=x  (tan x ) + tan x  (secx)  cose2 x dx dx
c'x

= secx (se2 x
+ tan2 x ) 
cosec2 x
Remark 3 sin x, cos x, sec x ,cosec x are periodic functions with period 2n.Their derivatives are also periodic with period 2n.tan x and cot x are periodic with period n. Their derivatives are also periodic with period n.
We have been considering variables which are dimensionless. Actually, in practice, we may have to consider variables having dimensions of mass, length, time etc.., and we have to be careful in interpreting their derivatives. Thus, we may be given that the distance x travelled by a particle in time t is x = a cos bt. Here, since bt is dimensionless (being an angle), b must have the dimensior, that a must have the same dimension as x. That is dimension of a is L. Now dxldt =  ab sin bt has the dimension of ab =L x 1/T = L/T, which is not unexpected, since dx/dt is nothing but the velocity of that particle. See if you can do these exercises now.
T
. Similarly, xla = cos bt has to be dimensionless. Tbis means
E 4) Find the derivatrres of: a) cdsec 2x b) cot x t
C)Scot 9x
Elements o Difleryttlal f Calcplllll,
4.3 DERIVATIVES OF INVERSE FUNCTIONS
We have seen in Unit 1 that the graphs of a function and its inverse are very closely related to each other. If we are given the graph of a function, we have only to take its reflection in the line y = x, to obtain the graph of its inverse. In this section we shall establish a relation between the derivatives of a function and its inverie.
'4.3.1 The Inverse Function Theorem
Let us take two functions f and g, which are inverses of each other. We have already seenjn Unit 1 that in this case, gof(x) = g(f(x)) = x, for all x for which f is defined, fog(y) = f(g(y)) = y for all y for which g is defined. Now, suppose that both f and g are differentiable. Then, by applying the Chain rule to differentiate g(f(x)) = x, we get gf(f(x)).f'(x) = 1 or g'(y).fl(x) = 1, where y = f(x). This means that if ff(x) # 0, we can write gf(y) = l/ff(x). So we have been able to find some relation between the derivatives of these inverse functions. Let us state our results more precisely. Theorem 1(The Inverse Function Theorem) Let f be differentiable and strictly monotonic on an intervd I. If ff(x)# 0 at a certain x in I, then fI is differentiable at y = f(x) and
The strict monotonicity cond~tion in this theorem implies that f is oneone and thus ensures the
Thus, we have the inverse function rule:
existence off
'
is The derivative of the inverse functia~ the reciprocal of the derivative of the given function. Sobn we shall see that this rule is very useful if we want to find the derivative of a function when the derivative of its inverse function is already known. This will become clear when we consider the derivatives of the inverses of some standard functions. But first, let us use this rule to find the derivative of f(i) = xr, where r is a rational number. In Unit 3 we have d already proved that  (xh)= nxn' when n is an integer. We shall use this fact in proving dx Ehe general case. : Theorem 2 If y = f(x) = x where r is a rational number for which xr and x d (xr) = rxrT' defined, then dx
x' may not be always defined.For example, ifx = 1 and r = lR,
''are both

Proof: Let us first consider the case when r = l/q,q being any nonzero integer. In this case,
y = f(x) = x1/q.Its inverse function g wilrbe given by x = g(y) = yq. This means
xr=
fi"isnotdefinedinR.
Thus, by thg inverse function rule, we get dy ='= dx
1 dx/dy, 1
9 Y
q1
SOfar, we have seen that the theorem is true when r is of the fonn l/q, where q is an integer. Now, having proved this, let us take the general case when r = plq, p, q E Z(q is, of course, nonzero). Here, y = f(x) = xr = Xdq
Now,
&( X
d
119 P
) =P(
xllq)pl
. (X dx
d
14 1
) , by chain rule
&civativ& o Trigonometric f Functions
Thus,
This completes the proof of the theorem. The usefulness of this theorem is quite clear from the following example.
Example 3 Suppose we want to differentiate
We write u = x5I6+fi gives us, y = u"". This By chain rule, we get
Thus,
Why don't you try these exercises now?
4.4 DERIVATIVES OF'INVERSE TRIGONOMETRIC FUNCTIONS
In the last section we haveisem how the inverse function theorem helps us in finding the derivative of xn where n is a rational number. We shall now use that theorem to find the derivatives of inverse trigonometric functions. We have noted in Unit 1, Section 5, that sometimes when a given function is not oneone, we can still talk about its inverse, provided we restrict its (jomain suitably. Now, sin x is
Elements of Differential Calculus
neither a oneone, nor an onto function from R to R. But if we restrict its domain to [ ~ 1 2x/2], and codomain to [  I , 11, then it becomes a oneone and onto function, and , hence the existence of its inverse is assured. In a similar manner we can talk about the inverses of the remaining trigonometric functions if we place suitable restrictions on their domains and codomains. Now that we are sure of the existence of inverse trigonometric functions. let's go ahead and find their derivatives.
I
4.4.1 Derivatives of Sin' x and Cos' x
/
II
Let us consider the function y = f(x) = sin x on the domain [ 6 2 , x/2]. Fig. 3(a) showsthe graph of this fur.ction. Its inverse is given by g(y) = sin' (y) = x. We can see ckarly that I sin x is strictly increasing on [  ~ 1 2R2]. ,
We also know that the derivative,
d
(sin x) = cos x exists and is nonzero for all
This means that sin x satisfies the conditions of the inverse function' theorem. We can, therefore, conclude that sin' y is differentiable on.]  1 , 1 [, and d  (sin'
Sin~simx=y,cOsx= for  r c n < x < l d 2 .
(y )) =
4
dy
1 1 = ff(x ) cosx
Thus. we have the result 1 d 3 (sin' t) = at
Remember, sin'x is not the same as (sin x)' = llsin x or sin xI = sin llx.
JZ
Fig.3(b) shows the graph of sin' x. We shall follow exactly the same steps to find out the derivative of the inverse cosine function.
, Let's stkt with the function y = f(x) = cos ; and restrict its domain to [0, R ] and its codomain to [1,1]. Its inverse function g(y) = cos' y exists and the graphs of cos x and cos x are shown in ~ i ~ . 4 ( and 4(b), respectively. a)
'
As in the earlier case, we cap now check that the conditions of the inverse function $orem are satisfied and conclude that cos y is differentiable in 11, 1[. Further
'
Derivatives of Trigonometric Functions
d d ( g ( y ) ) S  (cos dy dy .
1
9 ) = .  1 f'(x )
1  sin x
Since cos x = y, sin x = * o<x<n. '
for
This gives us the result
Yoti can apply these two results to get the derivatives in the following exercise.
E
E 6) Differentiate a) sin' (5x)
b) cosI
&
C)sin x cos' (x3+ 2)
E
E 7) a) By looking at the graph of tan x given alongside, indicate the interval to which the domain of tan x should be restricted so that the existence of its inve~se ' is guaranteed.
,
b) What will be the domain of tan' x? d 2 c) Prove tdat (tan' x ) = 1/(1 + x )in its domain. dx

In this section we have calculated the derivatives of sin' x and cos' x and if you have done E 7). you will have calculated the derivative of tanm'xalso. Pn>ceedingalong exactly simila lines. we shall be able to see that d 1  (cotI x) = 7 dx l+x
4.4.2 Derivatives of Sec' x and Cosec' x
Let's tackle the inverses of the remaining two mgonomemc functions now. To find sec' x, we proeaed as follows: If y = secI x, then sec y = x or I/cos y = x, which means that Ilx = cos y. This gives u
y = cos' (I/x), where, I x I 2 1.
Thus, y =setI x =cos' (l/x), I x I 1 2
Remember, we have seen that cos' t is defined in the interval I1. I].
From this we get
=dy
dx
1  1 1/x . d z  7 (ltX)
dx
(~0s'(ltx 1)
  Ixl
4n
( llx2)
lx 1
Nore tnat dthough see'x is defined for 1 x I 2 I,the derivative o eec' x f
does nor exist when x =
1
,Ixl>l
1
Thus. we have I d (sec' x) = dx l x 1

, Ixl > I
E 8) Following exactly similar skps, show that
d  (cosecI x dx
)=
1
l x l d z i *
1x1 > I .
Exan)ple 4 Suppose we want to find the derivative of y = sec I 2 By chain rule, we get
6
Now, you will be able to solve these exercises using the results about the derivatives of inverse trigonometric functions.
Derivatives of Trigonometric Functions
E 9) Differentiate,
a) c o t ( ~ 1 2 ) c) cos (5x
'
b)
cot' (x + 1)
~
tan (x
' + 1)
'
+ 4)
4.5 USE OF TRANSFORMATIONS
Sometimes the ~IWXSS of finding derivatives is simplified to a large extent by making use of some suitable transformations. In this section we shall see some examj&ss vvbidt will illustrate this fact.
Example 5 Suppose we want to finuthe derivative of
y = cos' (4x3  3x)
I
I
Elements d Differential Calculus
As you know, we can differentiate this function by using the formula for thederivative of cos' x and the chain rule. But suppose we put x ='cos 8 ,then we get y = cos' (4 cos38  3 cos 8) (COS 38 = 4 c0s38  3 cos 8) = &s' (COS 38) = 38 = 3dos' x. Now this is a much simpler expression, and can be differentiated easily as:
Example 6 To differentiate y = tan' This gives us,
I \
,we use the transformation x = tan 8.
y = tan
l[dltan 8  I J= t a n  l ( ~ e c8~1,) + tan2e tan
1(12sin28/2) 2 sin 812 cos 812
sin 8 = tan' (tan 012) tan x = 812 = . 2
1
 respect to sin)with
O r aim is to find dyldz. We u
.
'
'
dy 1 Now, we can write  = dx 2(1+x2) Let's lackle another probl&n.
Example 7 Suppose we want to differentiate tan'
For this, let y = tan
shall use the transfo&ation x'= tan 8. This gi;es us y=m y 1
'(%) z = s i n  '1(+2x)2. and \ 1
2:2
[  )
2tan8 = tan (tan 28) = 20, and 1 tan 8
Now if we differentiate y and z with respect to 8, we get dy/d8 = 2 and dz/d8% 2. Therefore, dy dyId0  =. = 1. dz dz/d8
Alternatively, we have y = z. Hence, dyldz = 1. So, you see, a variety of complex.problems tan be solved easily by using transformations. The key to a successful solution is, however, the choice of a suitable transformation. We a e giving some exercises below, which will givtyou the necessary practice in choosing the right transformation.
E
E .lo) Find the derivatives of the following functions using suitable transformations :
Now let us summarise the points covercd in this unit.
4.6 SUMMARY
In this unit we have
,
1 calculated the derivatives of trigonometric functions:
1
Function sin x cos x tan x cot x sec x
COS~C x
1
Derivative cos x sin x sec2x cosec2 x sec x tan x cosec x cok x
I
:
2 discussed the inverse function theorem and
used the rule
toprove that d/dx(xr) = rx", where r is a rational number.
3  used the inverse function theorem to find the derivatives of inverse trigonometric
.functions:
Elements 01Differential f~lculus
I function tan'x
Derivative
'
!
Ixl > 1
1 x 4 4 7 7'  1
Ixl > 1
1xldZl'
4
used transformations to simplify the problems of finding the derivatives of some functions.
4.7 SOLUTIONS AND ANSWERS
E 1)
a) cos (a + X)= cos a cos x  sin a sin x lim cos (a + x ) = cos a lim cos x  sin a lim sin x
x+ 0 x+ 0 x+o
b) similar
= cos a
E 2)
a) 2 cos 2x
C)
d x b) 2 cos xcdx (COS ) =  2 sin x cos x
5(3sin7x cos 3x + 7sin6x cos x sin 3x) = 5 sin6 x (3 sin x cos 3x + 7cos x sin 3x)
d) 3x2 cos 9x  9x3sin 9x
e)
 sin (sin x) cos x
sin x x O  lcosx
d
d d E3) (cosecx)= dx =COS X
= cosecx sin x
2
cotx.
E 4)
a)  2cosec 2x cot 2x b)  cosec2x
+
1
d i z z x
(  cosecx cotx)
3x2 c1 sin
+ cos x cosI (x3 + 2)
E 7)
a) tan x restricted to ]  n/2, n/2[ is a strictly increasing oneone function of x. Thus, its inverse exists when restricted to ]  n/2, n/2[. b) The domain of tan' x is ]  w, [.
I
c) ~f y = f(x) = tan x,
Deri~atives Trigonometric of Functions
Hence d/dx (tan' x) =
1 . 1 + x2
3
I
I
E 8)
y = cosec ' x a cosec y = x a sin y = l/x
1
y = sin' (l/x) where I x 1 2 1.
 '"I ( I/&) fi
1
d ) ~ l i ; n ~ g /x~ s i n 2 8 2 2 (1  x cos 8)
1
[ sin 8 (1  x cos 8) + x sin 8 cos 8 2 (1  x cos 8 )
1
e)
1 I x + l I d ( x + 1)2 1
+
y = sin' (3x  4x3)
1 Ix lId(x1121
E 10) a) Put x = sin 8
= sin' (3sin 8  4sin3 8)
= sin' (sin3 8) = 38 = 3sin' x.
b) x = cos 812
* y = COS' (1  2x2)
c) Put x = tan 8
dy  2 
* y = sin'  = 2 tanlx 2 ) ::
(,
dx
1+x2
d) Put x = tan 8
e ) Put x = tan (3
UNIT 5 DERIVATIVES OF SOME STANDARD FUNCTIONS
Structure
5.1 5.2 5.3 5.4
Introduction.
Objectives
Exponential Fynctions
Definition of an Ewnential Function Derivative of an Exponential Function
Derivatives of Logaritlpc Functions
Diffe~ntiating Natural Log Function the Differentiating the General Log Function
Hyperbolic Functions
Definitions and Basic Properties Derivatives of Hyperbolic Functions Derivatives of Inverse Hyperbolic Functions
5.5
Methods of Differentiation
Derivatlve of x' Logarithmic Differentiation Derivatives of Functions Defined in Terms of a Panupeter Derivatives of Implicit Functions
101
5.6 5.7
Summary Solutions and Answers
1 W
107,
Exponential functions occupy an important place in pure and applied science. Laws ofi growth and decay are very often expressed in terms of these functions. In this unit we s w l s t a y the derivatives of exponential functions. The inverse function theorem which we stated in Unit 4 will then help us to differentiate their inverses, the,logarithmicfunctions,In particular, you will find that the natural exponential function is its own derivative. ~ u h e rwe shall introduce and differentiate hyperbolic functidns and their inverses since , they hold special significance for physical sciences. We shall demonstrate the method of finding derivatives by taking logarithms, and also that of differentiating implicit functions. Witb this unit we come to the end of ollr quest fot the derivatives of some standard, frequ&tly used functions. In rhe next block we shall see the geometrical significance of derivatives and shall use them for sketching graphs of functions.
Objectives
After stvdying this unit you should be able to: find the derivatives of exponential and log&thmic functions define hyperbolic functions and discuss the existence of their inverses differentiate hyperbolic functions and inverse hyperbolic functions 7 use the method of logarithmic differentiation for solving some problems differentiate implicit functions and also those functions which are defined with the help of a parameter.
5.2 EXPONENTIAL FUNCTIONS
Our main aim, here, is to find the derivatives of exponential functions. But let us first recall the definition of an exponential function.
I
5.2.1Definition of an Exponential Functioh
A function f defined on R by f(x) = ax,wherea > 0,is known as an exponential functiun. Now to find the derivative off, we shal have to take the limit:
h+Oi
lim
ax+" ax
h
= a x . lim
h+O
, 
So, if we put lim
h+O

ah 1 h
............... (1)
Derivauvar m some Standard hndlone
a' may not be defined fqr all x if a< 0. For example, (2)Gs not
ah 1 = k, we get h
dx
d ax = k. ax. We can also interpret k as the derivative of axat x = 0. In Fig.4
defined in R.
you can see the graphs of exponential functions for various values of a.
All these curves pass through (0, 1) as a0= 1 for all a. Now from all these curves, we shall choose that one, whose tangent at (0, 1) has slope = 1. (We assume that such a curve exists). The value of a comsponding to this curve is then denoted by e. Thus, we have singled out the exponential function :x+ ex, so that its derivative at x = 0 is 1. Thus,
h+O
lim eh 1 = 1 h This also means that dex s+h ex = Pm dx h+O h

zeX. lim h+O
&I  ex h
That is, the derivative of this function is the function itself. . This special exponential function is called the natural exponential f'unction.
5.2.2 Derivative of an Exponential Function
In Unit 1, we compared the graghs of the natural exponential 'function exand the natural logarithmic function lnx and found that they are reflections of each other w.r.t. the line y = x. (see Fig.2) We concluded that = e" and lnx are inverses of each other. This also means that eh" G x ++x > 0.
From this we can write ax = e n u s d a~ =
ln.
, or a x = eh,
where a > 0 .
dx

 e x ~ n r d(X ln a ) by chain rule,
dx
,d dx
lnah = blna
Elements o Differential f
Cdculurr
Remark 1 If we compare this result with (1) which we derived at the beginning oflhis
section, we find that ah  I h a = lim 7
h +O
Thus, we have d ex = e x , and
dx
dx
d ax=a'ln
a
Ewmpk 1Let us use these formulas to find the derivatives of
i)
Lety=e
(x 2+2x'
. Then, by chain rule
( x +2x)
2
Hence  ( J ' ~ + ~ " ' )=2(x + l ) e dx
iii) wx6pply the chain mle again.to differentiate asinlx d (asin'x)= lna asin' x d (sin' x) dx dx

,

= lna
4
+
a sin],
See if you can solve these exercises now.
E
E 1) Find the derivatives of:
a)
sW7 bl e(x
I)/X
c)(x+2)eG
d)emtmlx
e)2h
97'
Derivatives of Some Standard Functions
I
! E
I
E 2 ) How ,much ta\ter
ir
f(x) = 2' increasing at x = I/?
ihim
x = 0':
t
'3
In this ~ C C L ~ Ub Ie hdve defined e as that real number for which lim I Alternatively. e can also be defined as a limit:
h+O
h
h
= 1. ,
#
1 e = ~ i m ( ~ + l / n ) ~ o r a s t h e s u m o f a n i n f i n i t e s e r i e s : e = 1 +  1  +...... + n+m l! 2! But all these definitions give the same value, e = 2.7 18281828....e is an irrational number. In many situations the rate of growth (of human beings, or bacteria or radioactive particles) . is proportional to the present population. That is, if x(t) is the population at time t. dx d oc x . In such situationb the exponential function is of great relevance since (et = el. then dt dt


Now let us turn our attention to logarithmic functions.
5.3 DERIVATIVES OF LOGARITHMIC FUNCTIONS
In Unit 4, we studied the inverse function theorem, (Theorem 1, Unit 4) and used it to find the derivatives of various functions such as sin'x, cos'x, and so on. Here, we shall, yet function. again, apply this theorem to calculate the derivative of the n$ura~b~arithrnic
5.3.1 Differentiating the Natural Log Function
,
Consider the function y = lnx. This is the inverse of the natural exponential function, that is, y = lnx if and only if x = ey.
I
From Fig. 2, you can see that the natural exponential function is a strictly increasing function. (You will be able to rigorously prove this result by the end of this course). Further, the derivative of the function x = ey is
Y dx =(e d ) = e Y> o f o r all E R. dy dy Thus, all the conditions of the inverse function theorem are satisfieb. This means we can conclude that the derivative of the natural logarithmic function (which is the inverse of the natural exponential function) exists, and .
Inx
17 defined on
]O,

[.
1  1 1 (Inx ) =   dx dx/dy ey Thus, we have
dy  d    dx

Let's see how we can use this result.
Example 2 Suppose we want to differentiate y = . ln (x2 2x +2).
93
Elements of Differential
Calculus
1 d 2 (x  2x + 2) dx x 2  2 x + 2 dx  2x  2 x2 2x + 2 Note that x2  2x + 2 = (x + 1 and hence, is nonzero for d l X. Therefore, ln(x2  2x + 2) is welldefined. dy =

,
"
Example 3 If we want to differentiatey = In two cases: i) l x I > 1 and ii) I x I < 1
i) If I x I > ~ , 1+x2 =
, I x I # 1 , we will have to consider
x2+1
since I x I > 1 makes lxZ negative. So in this case,

4x 1x
4
' after simplification.
ii) when lx l'<l,
l.~l=
1x
.
1X
and

 4x 
1 x4
So, we see that dy = 4x4,forallxsuchthat I x I # l . dx l  x Now, let us turn our attention to logarithmic functions with arbitrary bases.
 
5.3.2 Differentiating the General Log Function
a' is a constant function for a = 1. H ~ does not ~ it ~ have any , ~
The log functions are thus defM only for a # 1.
Let us consider any positive number a # 1. We say logax= y if and only if x = ay. Obviously, the natural logarithmic function lnx Can be written as logex. Further, we know that logax = logae. lnx. This rule gives a connection between the natural and general logarithmic functions. We shall use this relationship to find the derivative of logsx. So, if y = log,x = log,e lnx,
Thus, we arrive at d 1 (log, x) = log,e r dx If we put a = e in this, we get our earlier result: d 1 .  lnx = y , logee= 1 since dx

Example 4 Let us differentiate y = log, tan3x
dy   log
dx d (tan3 x ) e. 1 tan3 x d~

se2 x tan x If you have followed the solved examples in this section you should have no difficulty in solving ttiese exercises. =310g7e
E
E 3) Find the derivatives of:
a\ lrrn
7v
h\ l l n n
I 5 v 2 I 71
"\
nxlnv
Derivatives of Some Standard Functions
5.4 HYPERBOLIC FUNCTIONS
In applications of mathematics to other sciences, we, very often, come across certain combinations of ex and e'. Because of their importance, these combinations are given special names, like the hyperbolic sine, the hyperbolic cosine etc. These names suggest that they have some similarity with the higonomemc functions. Let's look at their precise definitions and try to understand the points of similarity and dissimilarity between the hyperbolic and the higonomemc functions.
5.4.1 Definitions and Basic Properties
Definition 1 The hyperbolic sine function is defined by sinh x = The range of this function is also R.
ex  e'
2
for all x E R.
Definition 2 The hyperbolic cosine function is defined by cosh x = The range of this function is [l, 1.
You will notice that
+'' l l x ~ fora R. 2
=  sinh (x ), and
Elements of Differential Calculus
In other words, the hyperbolic sine is an odd function, while the hyperbolic cosine is an even function. Fig 3(a) and (b) show the graphs of these two functions.
I
(a)
(b)
Fig. 3 : Graph # (a) sinh x (b) cosh x
We also d d n e four other hyperbolic functions as: ex  e x ex + e' tanh x = , coth x =  , , e"  e" ? i + e"

E 4) Verify that a) cosh x  sinh x = 1 sinh x b) tanh x = cosh x
'
.2
2
E
E 5) Derive an identity connecting coth x and cosech x.
Derivatives of Some Standard
Flrnctlons
You must have noticed that the identities involving these hyperbolic functions are similar to those involving trigonometric functions. It is possible to extend this analogy and get some more formulas: sinh (x f y) = sinh x cosh y f cosh x sinh y cosh(x f y) = cosh x cosh y f sinh x sinh y tanhx f tanh y tanh(xf y ) = 1 f tanh x tanh y Since we have seen that coshZ  sinh2t = 1, it is obvious that apoint with coordinates t (cosh t, sinh t) lies on the unit hyperbola: x2 y2=l. (Hence the name, hyperbolic functions). We have a similar situation in the case of trigonometric functions. The point (cost, sint) lies on the unit circle: x2+ y2 = 1. That is why trigonometric functions are also called circular functions. There is one major point of difference between the hyperbolic and circular functions, though. While t in sint, cost, etc. is the measure of an angle, the t which appears in sinht, cosht, etc. cannot be interpreted as the measure of an angle. However, it is sometimes called the hypberbolic radian.
5.4.2 Derivatives of Hyperbolic Functions
Since the hyperbolic functions are defined in terms of the natural exponential function, whose derivative we already know, it is very easy to calculate their derivatives. For example, sinh x =
.
 eX
2
This means,
d d  (sinh x) = dx dx
(
ex e"
e x + e' )=T=coshx
Similarly, cosh x = "
eX
.
d  (cosh x) =  sinh x = dx 2 Inthecaseof tanhx =eX
'" gives  us 2
+
ex + e'
, we get
= 1  tanh x = sech x
2
2
We can adopt the same method for finding the derivatives of coth x, sech x and cosech x. In Table 1 we have collecm all these'results.
Table 1
sinh x cosh x tanh x coth x cosech
*
 sech x tanh x  cosech x coth x
Ekrnents o Mfferentlal f
Calculus
Example 5 Suppose we want to find dyldx when y = tanh (1  x2) .
See if you can solve these exercises on your own.
E 6) Find f'(x) when f(x) = 4x + 1 a) tanh
5 d) sexh (lnx )
b) sinh e2"
c ) coth (llx)
e ) e" cosh x
5.4.3 Derivatives of Inverse Hyperbolic Functions
We shall try to find the derivatives of the inverse hyperbolic functions now. Let us start with the inverse hyperbolic sine function. From Fig. 3(a) you can see that the hyperbolic sine is a strictly increasing function. This means that its inverse exists, and
y = sinh x O X = s i n h y =
1
2 ey
2
We have used the formula for finding the roots of a quadratic equation here. Note that if e =x y then ey < 0, which is impossible. Therefore we ignore this At.
c x , + '
Thus sinhlx = In (x sinhlx. Now,
+
m),I
E ]
.o. . o
[. In Fig. 4, we have drawn the graph of
Derivatives of Some Standard
Functions
d Thus.  (sinhlx) = dx K
1
2 =
1 c
1
I
: I
!
In the case of the hyperbolic cosine function, we see from Fig. 3(b), that its inverse will exist 2 if we restrict its domain to [0, m[. The domain of this inverse function will be [l, , and its range will be [0, m[ 1 ey + e' 1 Now y =cosh x e x =coshy = 2 e e 2 ' 2xey + I = O
4 Y

.
1 1
0

2
3
4)
$1
*!
e e Y= x + G
1
1)
Fig. 5 Again we igno the root
!
o y =h(x
+ dx2X
Thus rnshl x = In ( X + Fig. 5 shows the graph of cosh' x. Further d  (cosh' dx x) =
x
I
Jx), 2 1.
11
=x x'  1 , because then ey< 1. which is imwssible since
e y
?
+ vx21
d (x + v x2 1) dx
1
Note that the derivative of cosh'x does not exist at x = 1. Fig.6 (a), (b) and (cj show the graphs of tanh x, coth x and cosech x. You can see that each of these functions is oneone and smctly monotonic. Thus, we can talk about the inverse in each case.
Y
2
   '  _ _ _ , 4:~,x _
4
2 . 2 w
=.
44
(a)
(b)
(c)
Arguing as for sinh' xand cosh' x, we get
y = cosec h x u x = cosec hy e y = In
4
1
7+
Ix l
, x#O
Since sech x =
,we shall have to restrict the domain of sech x to [0, m[ before talking cosh x abouf its inverse, as we did for coshx. Sech'x is defined for all x E 10, 11, and we can write
Now, we can find the derivatives of each of these inverse hyperbolic functions. We p k e e d exactly a we did for the inverse hyperbolic sine and cosine functioris and get s
Elements of Differential Calculus
d  (sech1 x ) = dx
1
x
,O e x c l
&F
Let us use these results to solve someproblems now.
Example 6 Suppose we want to fnd the derivatives of a) f(x) = sinh'(tanx), and b) g(x) = tanh'(cos eX).
Let's start with f(x) = sinh'(tan x). f (x)=
.J
I secx I
1 tan2 x
+1
d &(tan x)

sec2x =I sec x I
Now if g(x) = tanh'(cos ex),this means that g'(x )=
1 d  (cos ex) I  cos2ex dx
I (  sin ex).ex sin e = = 8 excosecex sin 8 We are now listing some functions for you to differentrate.
_
E
E 7) Differentiate the following functions bn their respective domains. b) [sechI (cos2x)] a) cosech'f5&) (x2+5x6) c) coth' (e )
Derivstlves of Some Standard Functions
5.5 METHODS OF DIFFERENTIATION
In this section, we shall study different methods of finding derivatives. We shall also see that the problem of differentiating some functions is greatly simplified by using these methods. Some of the results we derived in the earlier sections will be useful to us here.
5.5.1 Derivative of xr
d In Unit 4 we have seen that  (xr) rxr' when r is a rational number. Now, we are in a ' dx position to extend this result to the case when r is any real number. So if y = xr, where x > 0 and r E R, we can write this as

Y = elnx'= erlnx ,since the natural exponential and logarithmic functions are inverses of each
olher.
rlnx d Thus 9 = d ( e r l n x ) = e (rlnx) dx dx dx
As we have mentioned in Unit 4, if x < 0, xr may not be a real number. For example, 3 m = R.
n,
This proves that
d (xr=xr for x > O , r e R. dx We are sure, you will be able to solve this exercise now.
E 8) Differentiate,
a) x fi b) X=
. 5.5.2 Logarithmic Diffepentiation
Sometimes we find that the process of taking derivatives becomes simple if we take logarithms before differentiating. E this section we shall illustrate this point through son
Elements d Differential
Calculus
By J a we mean the positive square root o a. f
examples. But to take the logarithm of any quantity we have to be sure that it is nonnegative. To overcome this difficulty, let us first try to find the derivative of In (1 x 1). Now you can check easily that I x I = . Therefore, ln(l x I ) = ln and
c2,
c2
We get, d 1 ln(lx I)= 7 dx d 1 du ln(l u I ) = . . Using chain rule we can now conclude that if u is any function of x, then dx u dx Let us see how this result helps us in simplifying the differentiation of some functions.


Example 7 To differentiate .
(x2 + 1f (x  3pI4
213
(x5) we start by taking y = (x Thus, I y I=
213
(x2 +
17 (X  3QI4
+1
 7jI3(x2 + 2x + 1)113
lx2+ l ( 9 1 x  3 1 ~ ~ ~ l x 51 lx2 + 2 x
Then taking logarithms of both sides, we get
In(e/b)=lMInb h(ab) = h a + lnb ln(ab)= blna
Differentiating throughout we get,
Example 8 Suppose we want to differentiate xu'"",x > 0. Let us write y = xma.Then y > 0 and so we can take logarithms of both sides to the base e
and write lny = lnx" = sin x.ln x Differentiating throughout, we get,
' A 4 = sin x 1 + cosx inx Y dy
i .
sin x = 7+ cosx lnx
Example 9 TO differentiate xco" + (cosx); let f(x) = xcos;and g(x) = cosxx. To ensure that f(x) and g(x) are well defined, let us restrict their domain to [O, nl21.. y = xCDSX + (COSX)~ + g(x) > 0 for X E [O, ~ 1 2 1 = f(x)
,
hrivatives of Some Standard Functions
Let us differentiate both f(x) and g(x) by taking logarithms. We have, Therefore lnf(x) = cosx Inx.
 x sinx Inx + cosx XCOEX
1
(cosx  x sinx Inx)
Similarly, g(x) = (cos x)" and sa lng(x) = x In cos x
cosx In cosx  x sin x
= (cosx )
x !
1
(cosx In cosx  x sin x )
(cosx  xsinx lnx) + cosx"' (cosx lncosx xsinx) If you have followed these examples you should have no difficulty in solving these exercises by the same method.
E
E'9) Differentiate,
a) (x2!
1) (x2 + 2)6 (x3 
b)
1
c) ( ~ i n x+ ~ ) (cosx)'""" d) (xX)" x +
(
(x  1) (x  2 ) (x  3) (sinx)'"" + xx
5
6
7
xX)
e!
Elements of Differential Calcl~lus
5.5.3 Derivatives of Functions Defined in Terms of a Parameter
Till now we were concerned with f k t i o n s which were expressed as y = f(x). We called x an independent variable, and y, a dependent one. But sometimes the relationship between two variables x and y may be expressed in terms of another vatiable, say t. That is, we may have a pair of equations x = ~ ( t )y = ~ ( t )where the functions 0 and y~ have a common , , domain. For example, we know that the circle x ~ = a2 is ~ ~ + also described by the pair of equations, x = acost, y = asint, 0 < t 5 2n. In such cases the auxiliary variable t is called a parameter and the equations x = B(t), y = ~ ( t are called parametric equations. Now, suppose a function is defined in terms of a ) parameter. To obtain its derivative, we need only differentiate the relations'in x and y separately. The following example illustrates this method.
Example 10 Let us try to find  if x = acos 8 and y = bsin 8.
(Here the parameter is 8) We differentiate the given equations w.r.t. 8, and get dy d  = b c o s 8 , and  =i  a s i n 8 d8 d8 dy/de b Now, dy =  bcos 8 =  , c o t e . dx dxJd8  asin.8 Try to apply this method now.
dy dx

E
E 10) Find dy ~f . dx a) x = acos 8, y = asin 8
b) x = at2, y = 2at c) x = acos78, y = bsin3 8 d) x = a(8  sin 8), y = a(l  cos 8)
Derlvatlvea of Some Standard . Functions
5.5.3 Derivatives of Implicit Functions
It is not always necessary to express y explicitly in terms of x(as in y = f(x)) to find its derivative. We shall now see how to differentiate a function defined implicitly by a relation in x and y (such as, g(x, y) = 0). dx ax2+ 2hxy + by2+ 2gx + 2fy + c = 0.
Example 1 Let us finddy if x and y are related by 1
Differentiating throughout with respect to x, we get
or
=
dy dx
(ax+hy+g) (hx + b y + f )
E
dy See if you can find . for 111cIbllowing implicit functions. dx dy . E 11) Find  ~fx and y are related as follows: dx

+ c) x3y3 x2y2 + xy + 1 = 0
d) cosx cosy  y2sin'x + 2x2 tanx = 0
Elements of Differential
Calculus
56 SUMMARY .
In this unit we have
1
obtained derivatives of the exponential and logarithmic functions, hyperbolic functions and their inverses. We give them in the following table.
unction
Derivative ex
1 X
Function sinhI x coshI x tanhI x , 1
Derivative
ex
lm
ax
1
aqna
1 loga x
JK
x>l
log, x sinh x cosh x tanh x coth x sech x cosech x
e
1 1  x2 1 1 x2
. 1x1 < 1
7
cosh x coth' x sinh x sech2x cosech2x sech x tanh x cosech x coth x cosech' x sech' x
(XI>1 , O<x<l
x
, z /
 1
 1'
I x 1&7
, x*o.
2
extended the result d (xr = ,,rl
dx
to all r E R and x > 0.
3
illustrated logarithmic differentiation, differentiatibn of functions involving parameters and
diff~rpntiatinn fiinrtinnc oiven hv imnlirit relatinn~ nf
5.7 SOLUTIONS AND ANSWERS
Derivatives of Some Standard Functions
E 2)
f'(x) = 2% 2 f(0) = In 2 ft(1/2) = 2ln ln2 = d 1 n 2 Hence f increasesfl times faster at x = l/2 than at x = 0.
c) e'(11~) e'lnx ~ 4 ) a) W& x =
2
e)
E 5). E6)
e2x .+ e2x  2 +e2x + 2 , '4  4 2 2 wsh x  sinh x = 1 2 e2" + e" + 2  = 4 2 coth x  1 = = cosec h x e2" + e'"  2 eZx+ ep2'  2
2'
sin4 x Sinh 2 =
1  sin3x cos x 4
a) ;sedT7j4x C) 1
X
+1
b) 2e2"cosh 2'
wseclh
d)  sech Inx mnh h x
1
e) ex(sinhx4 wshx)
b)
1 7 [ s ~ m  '( C O ~)I" 'f X
laa2x,~LX
1
1
2 cosx sin x
dl
 cosech x 
2
2 4x  1
2
1cothx
2
I
Elements of Differential Calculus
C)
Let f(x) = (sinx)" and g(x) = (cos x) . Then f (x) = sin xX (In sin x + x cot x )and gf(x) = cos xtm
(scc2x In cos x  tan2x)

dy = f(x) dx
+ gf(x)
(xL)
d) Let f(x) = (xa)l ,g(x) = x If y = xx, Iny = xlnx dy * dx = x x ( l + l n x ) 
. x >0
.
* ff(x ) = ( x  ) ~[1nxh'+ x (1 +
Ing(x) = xx Inx
IX K
)I
*(1 ) g'(x) g x
Answer = f(x)
=
x
+ lnx x X ( I + I n x)
X(I+
* gf(x) = x(x"lxx' + x
i n x)]
+ g'(x) ( x") [xa' + xXl.nx(l+ lnx)] = (xa)" [lnxx + x ( l + lnx)] +x
d Inx e)   ( s i n x ) dx =(sin ~ ) ' ~ ~ ( h x c o t 7 + xlnsin x
( x X ) = x X ( l+ ~ n x ) dx
~ n s w e =(sin x. r )' (Inx cot x +
d
7
I n sin x
3 bsin2 8 cos 8 13aco2 8 sin 8
dl
x=
dy
asin 8 a(1cos8)

sin 8 (lcose)
d)
dy cosx sinysinx dx + 2 x sec x
2
2
dy 2 cosy  2 y dxs i n  ' x  2 
W)
t4xtanx
=o
sinx cosy dy *dx
+ Y  4x t a n r
2
G)
 2x2 se'c'x
 (cosx siny + 2 y s i n  ' x )
UNIT 6 HIGHER ORDER DERIVATIVES
Structure
6.1
6.2 6.3 6.4 6.5 6.6 6.7
'
Introduction
Ohlcc~~\c\
Second and Third Order Derivatives nth Order Derivatives. Leibniz Theorem Taylor's Series and Maclaurin's Series Summary Solutions and Answers
6.1 INTRODUCTION
r
In the first block you have differentiated a number of functions. You know that the derivative f' of a differentiable function f is again a function and is called the derived function off.
We have already seen in Unit 3 that the concept of differentiation was motivated by some physical concepts (like the velocity of a moving particle) and also by geometrical notions (like the slope of a tangent to a curve). The second and higher'order derivatives are also s~milarly motivated by some physical considerations (like the acceleration) and some geometrical ideas (like the curvature of a curve). which we shall study in the remaining units of this block. We shall introduce higher order derivatives in Sec.1 and 2. Leibniz Theorem which is glven in Sec.4 gives us a formula for finding the higher derivatives of a product of twn functions. In the later sections, we shall study some useful formulas, called seties expansions. The significance of these expansions will become clearer in Unit 14.
Objectives
After reading this unit you should be able to. calculate higher order derivatives of a given function f use the Leibniz formula to find thc nth derivatives of products of functions expand a function using Taylor's Maclaurin's series.
6.2 SECOND AND THIRD ORDER DERIVATIVES
Consider the function f(x) = x4. You know that f'(x) = 4x3. Now, this f' is again a polynomial function and hence, can be differentiated (see Example 5, Unit 3). We shall denote the derivative off' by f". Thus, f"(x) = 12x2.This f"(x) is called the second derivative of the function f a t the point x. It d2y is also.A denoted by  (read as d square y by d x square) or yz or f(2'or D ~ ~ . 7 . dx Let us differentiate f". We get f'"(x) = 24x, where f"' denotes the derivative off", or the d3y third derivative off. Other notations for f"'(x) are  or y3 or f'3' or D3y.Differentiating dx d4y f"', we get the fourth derivative of f, f'4' (x) =  = y4 = 24. dx4 Thus, repeatedly differentiating (if possible) a given function f, we get the second, third. fourth, ..... derivatives off. These are called the higher order derivatives off. dny If n is any positive integer, then the nthderivative off is denoted by f'"' or by  (read as dxn d n y by d x n) or bv v, or Dny.
1
1
Drawing Curves
Note that in the notation f'"' the bracket is necessary to distinguish it from f", that is, f raised to the power n.'~his process of differentiating again and again, in succession, is called successive differentiation. We have already seen that there are functions f that are not differentiable. In other words f' need not always exist. Similarly even when f' exists it is possible that f" does not exist (see Example 3 near the end of this section). In general, for each positive integer n there are functions f sbch that f'" exists, but f'""' does not exist. However, many functions that we consider in these sections possess all higher derivatives.
A twice differentiable function is a function f such that f" exists. Let n be a positive
integer. A function f such that f'"' exists is called an ntimes differentiable function. Iff'") exists for every positive integer n, then f is said to be an infinitely differentiable function. Now we give some simple examples of higher defivatives.
Example 1 If we are given that the third derivative of the function
f(x) = ax3
+ bx + c has the value 6 at the point x = 1, can we find the value of a'! Here, f(x) = ax3 + bx + c
Differentiating this we get
Differentiating this again, we get ft'(x) = 6ax Differentiating once again, we get F3) (x) = 6a Taking the value at x = 1, P3)(I) = 6a It is given that fO' (1) = 6. Thus 6a 2'6. Therefore a = 1.
Such equations involving the derivativesare known as differential equations.
C
Example 2 If y = 2 sin x 3 cos x 5, let us prove that y2 y = 5. HOW,y = 2 sin x 3 cos x f 5 y2 =  2 sin x  3 cos x yl = 2 cos x  3 sin x . y2+y=2sinx3cosx+2sinx+3cosx+5=5
+
+
+
+
The example below gives a function f for which f' exists but
does not exist.
.Recall (Unit I ) that
Example 3 Consider the function f(x) = x 1x1 for all x in, R. The function f(x) can.be rewritten as
At points other than 0 we have f'(x) = 2x if x > 0 f'(x) =  2x if x < 0 At x = 0, the right derivative off,
the left derivative off,
Therefore ((0) = 0. Thus, f'(x) = 2 1x1 for all x in R. We have already seen in Ejtample 7, Unit 3 that the absolute value function 1x1 fails to be A:FFas.r+;nC.ln nt CI T h ~ r o f n r nf' i c nn+rliffprentiahle a t x = 0. That is. ft2)(0) does not exi
'
Try some exercises before going any further.
Higher Order'Derivatives
E
E 1) Find the second derivatives of the following functions.
a) f(x) = x3  4 b) y = e2'
1
*
E
E 2) Find f ( s/4) for the following functions. a) f(x) = sec x b) f(x) = sin 2 x cos 2 x
"'
+
E
E 3) Prove that thekollowing functions satisfy the differential equations shown against
them. a) y = sin x ;
(b) y = c ~ xs ;
Y4
=Y
(y212 + (yd2 = 1
I
Drawing Curves
E
E 4) Find the value of integer k in each of the following a) f(x) = sin k x and f'*' ( r / 6 ) = 2 b) f(x) = xk+ kx2 1 and f'?' (1) = 12
+
0
6.3 nth ORDER DERIVATIVES
Let n be a natural number. We have already defined the nth derivative of a function in Sec.2. When a function f is given by a formula, it is often necessary to express its nthderivative also by a formula using f and n. Usually, one can guess q"' after working out f"', f"' and fI3'. However, a rigorous proof would require an application of the principle of mathematical induction. In the examples below we shall derive formulas for the nthderivative of various functions. Study them carefully as we shall be using them in later sections. But, first let us recall the principle of mathematical induction. If (P,) is a sequence of propositions (sthtements) satisfying i) PNis true (usually N = 1).
ii) The truth of P, implies the truth of P,+I,i 2 N.
then P, is true for all n 2 N.
We shall apply this principle in the examples that follow.
Example 4 We shall prove w e that the nthderivative of the plityntxnial function xmis
dnxm  = (m(m  1) .....(m  n + l ) x m  " , i f n Im,and 0, if n > m dyn Let us denote by P, the statement dnxm (m  = (m if n  1) ..... (m  n + 1) xm"if n Im 0, > m.
dy" Note that the product m (m  1) ..... (m factor, namely, m is taken.
 n + 1) has n factors. When n = 1, only one
dxm Thus P I  = mxm' dy Suppose we have proved for some n that P, is true. This means that the nth derivative of xmis
(
=I
m (m  1) ..... m (m  n Oifn
+ l):xm" if n Im, and is

> m.
Then the n lth derivative of xmis = the derivative of the nthderivative of xm m (m  1) ..... (m  n 1) (m  n) xm"I if n = Oifn=m Oifn>m.
+
Higher Order Derivatives
I
+
<m
When n = m, the n"' derivative is constant, because xm " = x" = I; therefore the (n 1)'"erivative is
+
zero.
I
Check that the conditions "n = m orn>rn"and"n+ I 2 m " a r e equivalent.
I
t
I
This means that the truth of P implies the truth of P.+I. Therefore, by the principle of , , mathematical induction, P is true for all n L 1. Hence our result is true for all natural numbers n.
If
I
1
I
I
Remark 1 When n = m, the nthderivative of xmis = m (m  1) .....(m  n 1) xm"= m (m  1) .,...3.2.1. This is the same as m!.
.
+ Example 5 If f(x) = In (1 + x), let us find f'"' (x). 1 Differentiating f(x) = In (1 + x), we get f'(x) = . l+x
I
Differentiating again, f'2' (x) = 
1
+ x12 2 Differentiating once again, f(3'(x) = (1 + x)3
(1 i) The denominator off'"' (x) is (1
Can you guess f'"' (x) now? If you have guessed correctly, you must have arrived at these conclusions.
+ x)".
ii) Its sigh is positive or negative according as n is odd or even. iii) Its numerator has (n  l)! Do not think that it is merely (n  1). There is a factorial symbol too. To be convinced of this, calculate f'4' (x) and see. Therefore our guess is fn'(x) =
( I)"' X (n
 I,)!
(1 x)" This guess remains t o be proved. A proof is necessary because there could exist many other formulas for f'"'(x) that coincide with the correct answer when n = 1 , 2 or 3. For example, if we omit the factorial symbol, we get one such formula. But we have already mentioned that this formula does not hold for f'4'(x). So, let's try to prove (1). We first note that it is true for n = 1, 2 and 3 as we have seen in the beginning. Assume that it is true for n = m, that is,
+
..... (1)
Differentiating this we get, ffm+l'(x) ( ~ ) ~  l . (  l)! = m
(1
+ x)""
(1
m

(
I)"'' ( 1) X m (m  l)! (1 x)m"
+

( 1)"' (m)!
(1
+ x)""

( l)m+il(m 1
+ x)""
+
 I)!
This proves the guess for n = m f 1. Thus, assuming the truth of the formula (I) for n = m, we arrive at the truth of this formula for n = m 1. Therefore by the principle of mathematical induction, our suessed answer is correct for all positive integers n.
+
Thus, when f(x) = In (1
+ x).
Drawing Curves
E E 5)
If y = (1 (n < r).
+ x)', where r is a real number, find y, where n is a natural number
Example 6 If f(x) = cos 2x, let us use the principle of mathematical induction to find a formula for f'"' ( ) 0.
We first find f'"' (x) when n = 1,2, 3,4. We have f(x) = cos 2x. Diffe~entiating successively, we get this f'''(x) =  2 sin 2x f'='(x) =  4 COS 2x f'3)(x)= . 8 sin 2~ f ' 4 ) ( ~ ) 16 cos 2~ = We see that in the formula for f '"'(x), we have to have . i) a sign (positive or negative), ii) a coefficient (some power of 2), and

iii) a trigonometric function (sin 2x or cos 2x) We observe that the first two terms carry negative sign, the next two carry positive sign, the next two negativeand so on. We also observe that sin and cos occur alternately. Theretc..re our guess is
 2" sin 2x if n is of the form 4k
 2" cos 2x
+1
2
if n is of the form 4k
f ' n ) ( ~= )
\
Real\ that cos (B + r/2)=  sln 0 T ) =  cos e cos ( 1 a/2)= sin 0 cos (b
2" sin 2x if n is of the form 4k 2" cos 2x if n is of the form 4k
+3
..... (2)
We can also wrlte this in a compact form as f'")(x) = 2" cos (2x n7r/2) ..... (3) You can easily check that (2) and (3) are equivalent by putting n = 4k
+
u

d l I' 1
A L
:..
/l\ 1x7
Lall
..
+ 1,4k + 2,
Fn..ln
/l\
We have already seen that it is true for n = 1,2, 3 and 4. Suppose it is true for n = m, that is, '3 f'"'(x) = 2" cos (2x mn/2). Differentiating this we get, f ( m + ~ )=~ y + l sin (2x m ~ / 2 ) ( )
Higher Order Derivatives
+
+ or, f("'+" (x) = 2"'+' cos [2x + (m + 1) n/2]
cos (Zx (m I ) n/2) = cos (Zx mn/2 + n/2) = sin (2x mn/2)
+ + + +
So here again we see that the truth of the formula (3) for n = m implies its truth for n=m+l. Therefore by the principle of mathematical induction, we see that the guessed,formula for fIn'(x) is true for all natural numbers n. Now substitute x = 0. We obtain f'"'(0) = 2" cos n r / 2 This is the required answer. We can also use this method to Drove a general result about the nth derivative of a s3m of two functions.
Example 7 If fand g are two functions from R to R and if both of them are ntimes differentiable, we can prove that
(f + g)'"' = f'"'
+ g'n'.
We shall prove this result by induction. When n = 1, this means (f g)' = f' g' This has already been proved in Unit 3. Suppose (f g)'"' = f'"' g'"' is true.
+
+
+
+
Differentiating this we get [(f + g)('n)]' = [f(") + g'"'l' = [f(""]' + [g("'y This is the same as (f + g)(m+l) f(m+l) + g ( m + l ) =
\
~ h u the result is true for n = m 1. Therefore by the principle of mathematical s induction g'"' holds for all natural numbers n. (f g)'n' = f'"'
+
+
+
Remark 3 Similarly one can prove that
..
(cf)'"' = c. f'"' holds for all natural numbers n and all scalars c. This fact combined with Example 7 can be restated in the "linear algebra terminology" as : The collection of ntimes differentiate functions is a vector space under usual operations. Try to solve these exercises now.
'
E 6) Find the nlh derivative of the following functions :
a) x ( f b)
+ b)' f(x) = (ax + b)"
c) f(x) = ex d) f(x) = ekx
Drawing Curves
E
E 7) If f(x) = sin x, prove that f'"'(x) = sin [x n.
+ nn/2] holds for every natural number
E
E 8) If y = sin (ax b), prove that for every positive integer n, we have y, = an sin [(nn/2) ax b]
+
+ +
E E 9)
Prove that the n'h derivative of the polynomial hdction f(x) = ao a ~ x azxZ ... a d is the constant function n! an.
+
+
+ +
Higher Order Derivatives
I
E
E 10)If y = cos x and if n is any positive integer, prove t6St [ynlZ
+ [yn+112= 1.

6.4
LEIBNIZ THEOREM
In Unit 3 we have proved some rules regarding the derivatives of the sum, scalar multiple, product and quotient of two differentiable functions. These were (f g)' = f' g' (cf )' = cf' (fg)' = fg' gf'
+
+
+
(f/g)' =
(g(x) # 0 anywher in the domain) g2 In the last section we have seen (Example 7 and Remark 3) that the first two rules can be extended to the nthderivatives iff and g are ntimes differentiable functions. In this section we are going to extend the product rule of differentiation. We shall give a formula for the nth derivative of the product of two functions.
gf'  f i
The product rule for two functidns u and v can also be written as (uv), = U l V uv, Now we look for a similar formula for (uv)z, (uv),, etc.
+
u dQlotu tbe # order , derivatl"~f u o .
But first let us recall the meaning of the notation C(n,rJ, where n and r E 2 and r I This ' n. C(n,r) stands for the number of ways of choosing r objects from n objects. Sometimes it is
. .
..
n
,".
.
Drawing Curves
Also recall the formulas i) C(n,r) = n! r! (n

I
r)!
ii) C(n,O) = C(n,n) = 1 iii) C(n,r) = C(n,n  r) iv) C(n,r)
I
+ C(n,r + 1) = C(n + 1, r + 1)
I
These are combinatorial identities, true for all positive integers r and n with r I n. Theorem 1 (Leibniz Theorem) Let n be a positive integer. If u and v are n times differentiable functions, then (uv), = C(n,O) u.v C(n,l) U,IVI C(n,2) u.~vz ..., C(n,n) uv,
+
+
+ +
1
I
Leibniz had stated this result in his first article on differential calculus ...h:' .v nublishr$ ;q 1684.
I
The pattern in the formula for (uv), can be compared with the expansion of (X y)". The coefficients are binomial coefficients and they appear in the same order as those in the expansion of (x y)". The order of the derivative of u goes on decreasing one at a time, and the order of the derivative of v goes on increasing one at a time. The number of terms i s n + 1.
+
+
Remark 4 We omit the proof of this theorem and merely indicate how this can be pfoved by induction on n. Firstly, when n = 1, the above formula is the same as the already known product formula, and therefore is true. Assuming that it is true for n = m, we can prove it for n = m 1, by applying the product rule for each term of the expansion of (uv), and by using the combinatorial identities mentioned. (See E 17) for more details.
+
We start with a simple and direct application of the formula.
Example 8 If f(x) = x sin x, let us find the fourth derivative of f, using Leibniz Theorem.
We first observe that for n = 4, the Leibniz Theorem states ( u . v ) ~ C(4,O) U ~ V C(4,l) U ~ V I C(4,2) ~ 2 ~ C(4,3) u1v3 = + 2 = U ~ V 4 ~ 3 ~ 6 ~1 2 ~ 4 U I V ~ U V ~ . 2
+
+
+
+
+
+
+ C(4,4) UV4
In this problem we take u = x and v = sin x, so that f = uv We have u = x UI = 1 u2 = 0 = u3 = u4
v = sin x v1 = COS X v2 =  sin x v3=  c o s x v = sin x 4
,
Substituiing these in the above formula, we get
f'4' = ( u v ) ~ 0 0 0 = = x sin x  4 cos x
+ + + 4 (
cos X)
+ 1.x.sin x
What happens if we attack the same problem directly without the use of Leibniz Theorem? We have f(x) = x sin x Differentiating this, we get f'(x) = x cos x sin x (by product rule) Differentiating once again, we get fO'(x)= x ( sin x) 1 cos x cos x =2~0sxxsinx Differentiating once again, we get f'"(x) =  2 sin x  (x cos x sin x) =  3 sin x  x cos x Differentiating once again, f'4'(~)  3 cos x  [x ( sin x) cos x] =
+
+
+
+
+
=rcinr 4rncr
I
I
w e notice that we obtain the same answer. In this direct method, we had to apply the product formula four times, once for each differentiation. It is clear that when we want the nthderivative for bigger values of n, Leibniz theorem provides an easier method to write down the answer, avoiding the dificulty of repeatedly applying the product formula.
Higher Order Derivatives
Example 9 If y = (sin'x)', prove that
(1  x2) yn+2 (2n
+ 1) xyn+1 nZy, = 0 for each positive integer n.
Differentiating both sides of y = sin'^)^, we get 2 sin'x Y' =
, g /
i
I
Squaring and crossmultiplying we get (1  x2) y 1 2 ="4 sin'^)^ = 4y Differentiating once again, we get 2 (1  x2) y,y,  2xyi2  4 y, = 0 Dividing throughout by 2 y I gives us (1  x )2 y 2  x y l  2 = 0 Differentiating n times, using Leibniz Theorem for each of the first two terms we get (1  x2) yn+2  C(n,l) ~xY,,  C(n,2) 2 y. That is, (I  M ~ )  (2n I1) X Y ~ + I nZyn= 0 yn+2
b
L
 {xyn+l C(n,l) yn) = 0 +
t
The following exercises will give you some practice in applying Leibniz Theorem;
E
E 11)State Leibniz Theorem when n = 5. That is, (u.vJ5 = ?

E
E 12)Prove that when n = 1, Leibniz Theorem reduces to the product rule of
differentiation.
.it
$11 E
..y. . I
?$jj
*
,c
<!:
. .~. r , ,
8
I
E 13)Find the third derivative of sin'x using Leibniz Theorem. Find the same dlrectly also
and verify that you obtain the same answer.
.
.
El
E 14)If f(x) = x ex,find the sixth derivative of f, using Leibniz formula.
E
E 1S)Find the nthderivative x31nx
E
E 16)If y = eaxxa prove that y'") = ea" [anx2 2 nan'x
+
+ n(n  1) a"']
E
E 17)a) Write down Leibniz formula for (u.v), b) Differentiate it term by term and obtain (uv),+I = C(m,O) u,+lv C(m,l) (u,vl
~ i ~ h Order Derivatives er
+
+ urnIVZ) ..... + C(m,m) uv,+~. +
+
+
c) Deduce that ( u v ) ~ += C(m,O) u,+lv I [C(m,O) C(m,l)] u,vl [C(m,l) C(m,2)] urnlvz ..... [C(m,m  1) C(m,m:)] u ~ v , C(m,m)
+ +
+
+
+
+
UV,+I.
d) Deduce from (c) the Leibniz formula for (uv),+~.
E
E 18) Using Leibniz Theorem and induction, prove that (xn)'"' = n! for all natural numbers n.
6.5 TAYLOR'S SERIES AND MACLAURIN'S SERIES
In this section we obtain series expansions for many important functions. For this, we use higher derivatives. You must have come across the following series : i) Exponential Series :
.
ii) I .ogarithinic Series :
Drawing Curves
iii) Ge0metri.c Series :
iv) ~inomial'series : (1
..
x2
+ x)'= 1 + rx + r(r  1)
+
+ ..... provided 1x1 < 1 +
We observe that each of them is of the form
+ a, x + a2x2 a3x3 ..... a.xn ....,where 80, al, a2, ....,an,.... are some constants. We ask ourselves the questions; Is there anything else common to these four examples? Is it possible to express a, ..., a,, ... in terms of the function f? Our answer is : Yes. In all a,, these examples,
f(x) = a.
+ +
In other words, the series is d the form
Brook Taylor (1685173 1) and Colin Maclaurin ( 1698 1746) were both disciples of Newton, Taylor first published his series in a paper in 1715. Maclaurin used Taylor's series as a fundamental tool in his work on calculus,
We shall prove this for the above four instances, in the examples worked out below. This expansion is called Taylor's series for f around zero. It is also known as Maclaurin's series for f. The name "Taylor's Series for f around zero" suggests that there may be a Taylor's series for f around xo(xo # 0). But in this course we shall restrict ourselves d l y to the series around zero. This series expansion makes sense only when f is infinitely differentiable at zero. It is valid for many important functions (though not for all functions). You will learn more about the validity of these series in the course on real analysis. In this section, you should train yourselves to write down ~aclaurin'sseries for many functions. We have said above that the function f should be infinitely differentiable, that is, it should some functions it is not have derivatives of all orders. How do we check this condition? difficult. For example, we have dn nlr dn n' l ~  (sin x) = sin (x   (cos x) = cos (x  )' dxn 2 dxn 2
or
+
+
 (coshx)
dn dxn
=  (ex
:)
ex
=
1 y [ex + (
l)nex]

Icos hx, if n is even (sin hx, if n is odd

In these cases, we can say that the derivatives of all orders exist for all values of x and we can confidently use Taylor's series.
~ p f 10 Let us verify that the known series expansion of ex is the same as its e Maclaurin's series. Maclaurin's series for ex is
Where f(x) = ex NOW.f(0) = eO= 1 ~ l s o((x) = e. ((0) = 1. , " In fact, we know that f'"'(x) = ex for all natural numbers n, which means f'"'(0) = e0 = 1 for all natural numbers n. Substituting these values of f(O), ((o), ....., f'"'(0) in Maclaurin's series we get
which is the known expansion for ex. Exsmple 11 Obtain Taylor's series for In (1
+ x) around zero. Let f(x) = In (1 + x)
Then we have already seen in Example 5, that (n  I)! f ( y x ) = ( I)"' (1 x)"
+
Therefore, f'"'(0) .. f'"'(0) n!
= (  1)"' (n  I)!


( l)"l
n
's Therefore ~ a ~ l o rseries around zero is
We note that this is the same as the already known logarithmic series. Example 12 Now let us write down Maclaurin's series (or Taylor's series around zero) for l / ( l  x). 1 1 Let f(x) = lx . Then f(0) = 1  0 = 1

We can prove by induction that n! f(n)X = and therefore f'"'(0) = n! (1  x)"+~ l'herefore Maclaurin's series is
i
k
i
t
Note that this agrees with what we already know, namely that the sum of the geometric wries 1 x + xZ ..... xn ..... is 1/(1  x).
+
+ + +
Exrunple 13 Suppose we want to wnte down Taylor's series for (1 where r is a fued real number. Let f(%) = (1 f'(x) f'?(x) = r(r
+ x)' around zero,
+. ' ) x = r (1 + x)"
Then f(0) = 1
;((0) = r
 1) (1 t x)"
;f"'(0)
= r (r  1)
Drawing Curves
We can prove by induction that
Therefore Taylor's series around zero is
Note : This is the same as the binomial series that we already know. This expansion is valid only when I x 1 < 1. The reason for this will be clear when you study the course "Real Analysis". When r =  1, this bihomial series becomes
Note that Example 12 follows from this on replacing x by  x throughout. So far we have seen that the four known series occur as Taylor's series. In the next two examples, we find that we can write down similar series even for functions like sin x and
COS X.
Example 14 Let us write down Maclaurin's series for the function sin x. Let f(x) = sin x. Then we have already seen in E 7) 0 if n is even 1 if n is divisible by 4  1 otherwise
f(n)
(01
 sin 
n 5r
We see that, as n varies over 0, 1,2, 3, 4, $ 6 , 7 ...., f'"'(0) takes the values 0, 1,0,  1,O. 1, 0,  1, ..... Therefore Maclaurin's series for sin x is 1 0 ( 1) sin x = 0  x  x2 x3 I! 2! 3!
+
+
+
0 +  x4 + ... 4!
Example 15 To find Taylor's series for cos 2x around zero, let us write f(x) = cos 2x. We have already seen in Example 6 that
Therefore, Taylor's series around zero is
Example 16 Suppose we want to a) write down the first four terms of Maclaurin's series for tan x. b) write down the first three non zero terms of this series. Let f(x) = tan x. Then f(0) = 0 f8(x) = sec2 x, ((0) = 1 f"(x) = 2 sec2 x tan x, ('(0) = 0 f'3' (x) = 2 sec4x 4 sec2 x tan2 x = 2 sec2 x (sec2x 2 tan2 x) ~ " ( 0 )= 2 (1 0) = 2
+
+
+
Therefore the first four terms of Maclaurin's series for tan x are given by
Higher Order Derivatives
Hence Maclaurin's series for tan K is x
+
3
+ ""'
Now, we want the next nonzero term. ) We have f ( 4 ' ( ~= 16 sec4 x tan x f ( 4 ) ( ~ ) 0. =
+ 8 sec2x tan3 x and + 24 sec4x tan2 x + 16 sec2x tan4 x
~ e xf("(x) = 16 sec6 x 64 sec4 x tan2 x t This means f'"(0) = 16. Thus we have tanx=x+x3
+
3
+%+
2~~
Example 17 If in Maclaurin's series for sin kx, the coefficient of x3 is given to be  6 k, let us find all possible values of k.
for Maclaurin s~ries sin kx is given by k3x3 kx sin kx =   ....., since f (x) = k cos kx l! 3! f(*)(x)  k2 sin kx = f ' 3 ) ( ~ )  k3 cos kx, where f(x) = sin kx =
+
The coefficient of x3 is  (k3/6) Therefore  (k3/6) =  6k This gives the equation k (k2  36) = 0 The roots are k = 0 , 6 or  6. Thris, 0, 6 and  6 are all the possible values of k such that the coefficient of x3 in Maclaurin's series for sin kx is  6k.
E
E 19)Write down Maclaurin's series for the following: 1
c) cos X d) 1/(12x)
Drawing Curves
I:,
I"
E
E 20) Write down the first three nonzero terms in Maclaurin's series of tbe following.
a) sin 3 x b) In (1  x)
I
1 ,

E
E 21)Find the coefficient of x9 in Maclaurin's series for the functions.
a) cos 2 x b) sin x
( +1)
I
E
E 22)If Maclaurin's series for sin x is differentiated term by term, do you get Maclaurin's
series for cos x?
Higher Order Derivatives
E
E 23)If Maclaurin's series for ex is differentiated term by term, we get the same series again. Prove this.
E
E 24)Consider the function y = a tan' bx where a and b are fixed real numbers. We are given that its Taylor's series around zero is 2 3 x  9 x3 ...
+
+
+
Find the values of a and b.
Drawing Curves
E
E 25)Find the coefficient of x3 in Taylor's series around zero for the function sin' x.
Some General Remarks o n Taylor's and Maclaurin's Series
Though we have obtained infinite series for many functions, it is necessary to give a note of caution. These infinite series need not be valid for all values of x, and as such, these have to be used with care. In the course on real analysis,you will be able to study the conditions under which these series are valid.
6.6 SUMMARY
In this unit, we have 1) introduced higher order derivatives, 2) derived a formula (Leibniz's Theorem) for the nthderivative ui a product of two functions. (Uv), = C(n, 0) u, v C(n, 1) u,I V I C(n, 2) un2 VI ..... C(n, n) u v,.
+
+
+ +
3) written Taylor's series around zero/Maclaurin's series of a number of functions by using the formula
6.7 SOLUTIONS AND ANSWERS
E 3) a) y = sin x
= y 3 4
= sin x = y
b) y = cos x yl =  sin x, y =  cos x, y3 = sin x 2 (y2)2 (y3)2= cos2 x sin2 x = 1.
*
* +
+
E 4) a) f(x) = sin kx
k Z 0, since k0 *0=2fi which is impossible.
* f ' 2 ' ( ~=  k2 sin kx )
3
f1"(7r/6) =  k2sin k7r/6
Now,  k2 sin kn/6 = 2 6 sin k n/6 =  2 6 / k 2 Since  1 < sin k 7r/6 < 0,  T < k 7r/6 < 0 = 3 k=lor2 Out of these, k =  2 is the value which satisfies sin k 7r/6 =  2 f i / k 2
*
6 , a)
(ax  (3  n)! 
3! a"
+ b)3n;n 5 3
Higher Order Derivatives
b,
(
m! a" (ax (m  n! )
+ b)"" if n I m
cos x if n = 4 k + 1 E 7) f(x) = sin x = > f (x) = cos x, f" (x) =  sin x, sin x if n = 4 k + 2 fI3'(x) =  cos X,f ' 4 ' ( ~= sin x and so on. So Our guess is that pP"'(x)= cos x if n = 4k + 3 ) Now use tho principle of mathematical induction to prove that P) = sin I sin x i f n = 4k (x) (x + "x ) as in Example 6. 2 E 10)y = cos x S y~ =  sin x, y2 =  cos x, y3 = sin x, y4 = cos x and so on. y, = cos (x n.rr/2) +yn+l=  sin(x nn/2) yn2 y.+12 = cos2 (X n ~ / 2 ) sin2(x n?r/2) = 1.
:t
+
A
if
+
+
+
+
+
E 12)(u v)l = ul v
+u
VI
which is the product rule of differentiation.
d3(sin2x) d3 = J (sin x. sin x)  cos x sin x  3 sin x cos x dx3 dx  3 sin x cos x  cos x sin x =  8 sin x cos x d = 2 sin x cos x (sin2x) dx d2 a  (sin2x) = 2(cos2 x  sin2 X) dx2 d3 =  (sin2x) =  8 sin x cos x + dx3
( iln+lx3

E 15)
xn
[(n  l)!  3 C(n, 1) (n
 2)! + 6C(n, 2) (n  3)!
E 24) We take that tan' bx a!w$ys takes values between
b=3
~
/ and +/2. Then, a = 2, 2
UNIT
Structure
7.1 Introduction
Objectives
THE UPS AND DOWNS
.

7.2 MaximaMinima of Functions 7.3 Mean Value Theorems
Rolle's Theorem Lagrange's Mean Value Theorem
Definitions and Examples A Necessary Condition for the Existence of Extreme Points
7.4 Sufficient Conditions for the Existence of Extreme Points
First Derivative Test Second Derivative Test
7.5 More Information from the Second Derivative
Concavity/Convexity Points of Inflection
7.6 Summary 7.7 Solutions and Answers
7.1 INTRODUCTION
Why do drops of oil on the surface of water coalesce? Why do honeycombs have hexagonal cells? Why is a drop of water spherical? Why does a red corpuscle in blood have the shape of a biconcave disc? The answers to these questions are closely related to minimum and maximum values of some functions. Drops of oils tend to coalesce so as to minimise total surface tension. The scheme of hexagonal cells enables bees to store a fixed .amount of honey by using the minimum amount of wax for sealing. A drop of water is spherical because a sphere is the shape which encloses a given volume with minimum surface area. The oxygen carrier red corpuscle, on the other hand, i s in the shape of a biconcave disc so as to maximise the surface area. It enables our system to carry the maximum amount of oxygen on the surface of a fixed amount of blood. In this unit we shall discuss an important technique involved in solving the problem of maximising or minimising various functions. This technique, as you will soon see, involves the use of derivatives which you studied in Units 36. We shall also discuss Rolle's theorem and the mean value theorem which have very important applications as you will see further in your study of calculus.
Objectives
After having gone through this unit, you should be able to obtain the maximum and minimum values of some functions solve practical problems of maximaminima state Rolle's theorem and the mean value theorem find the points of inflection and the curvature of a curve, determine whether a given function is concave or convex or neither in a given interval.
7.2 MAXIMAMINIMA OF FUNCTIONS
Look at the points P and Q in the graphs in Fig. 1. How are they different from other points on the graphs? We could describe Q's as the peaks or hilltops and the P's as the valleybottoms. Using the language of mathematics, we could say that each P has the propert) that the value of the function f a t P is smaller than the value of f a t neighbouring points. Similarly, the peaks Q are distinguished by f having a maximum value there when compared to the values at nearby points. But before proceeding any further, let us define maximum and minimum points of a
filnctinn nrcci~elv
Fig. 1
7.2.1 Definitions and Examples
Defiqition 1 : A point c is said to be a maxiinum point for a function f if there exists a 6 > 0, such that f(c) 2 f(x) for all x such that ( x  c ( < 6 . f is said to have a maximum at c, and f(c) is called a maximum value off.
We can similarly define a minimum value off. f is said to have a minimum at a point c $fits domain provided there exists a 6 > 0 such that f(c) 5 f(x) whenever lx  cl < 6.
IX  6, x 6C, that is. the set o f points whose distance from x 1s less than a positive number 6 is called a neighbourhood of x.
+
A maximum or a minimum value is known as an extreme value.
Example 1 The function f(x) = Ix 1 1 has a minimum value at x = 0. You can see from Fig. 2 that f(0) = 1 and f(x) > 1 for all x in any every neighbourhood 10  6,O 6[ of 0. The minimum value o f f is thus f(0) = 1. D o you agree that this function has no
+
+
maximum value?
Fig. 2 :Graph of qx) = 1x1 + 1
In this example we see that f(x) L f(0) in every neighbourhood of 0. But to prove that f(0) is a minimum value, we need to firld only one such neighbourhood. Our next example shows that Definition 1 cannot be applied to functions defined on closed intervals if their extrema occur at either of the end points of the interval.
Example 2 Consider the function f(x) = x 2, Y x E [O, 21. From'its graph (Fig. 3), it is clear that x = 0 is a minimum point and x = 2 is a maximum point of this function. In fact, you will see that 2 = f(0) 5 f(x) l = 4 4 x E 10, 21 f(2)
+
i
I
I
I
Drawing Curves
Y
L
Note that here we cannot find a 6 > 0 such that f(0) I f(x)Vx E 1 6, 6[ for the simple reason that f(x) is not defined for x < 0. The same argument holds good for the maximum. Since f(x) is not defined for x > 2, we cannot find any 6 > 0, such that f(x) 5 f(2)Vx E 12  6, 2 6[. This means f does not have extrema at 0 or 2 in the sense of Definition 1. How do we resolve this paradox? We modify Definition 1 to suit such cases.
+
i
I
I
Let f be defined on [a, b]. We shall say that f has a maximum (minimum) at a if we can find 6 > 0, such that f(a) l (f(a) I f(x) f(x)) for all x E [a, a 6[.
+
Similarly, we shall say that f has a maximum (minimum) at b, if we can find 6 > 0 such that f(b) 2 f(x) (f(b) S f(x)) for all x E ]b  6, b].
I
I
I
,
I
Example 3 The function f(x) = sin x has a maximum value at several points. Fig. 4 shows the graph of !his function. Can you see that these points are lr/2, 5lr/2, ..... 3xI2, 7~12,....
I I
I
w
U
2
Fig. 3
fig. 4:
Graph of f(x) = sin x
In general, the sine function has a maximum at each of the points 2nx + (7~12). being any n integer. The maximum value is sin [2nlr (lr/2)] = 1. This function has a minimum value too at s e v ~ a points. What are these points? D o you agree that the minimum value l at these points is l? We can now say that this function has an extremum at all points nlr r / 2 , n being any integer.
+
+
The functions in the examples considered so far were all continuous. For such functions, a valley in the graph of the function indicates a minimum and a peak points to a maximum. But we can talk about maximum and minimum values of noncontinuous functions too. Here, we may find extreme values which are neither at a peak nor at a valleybottom, as in Example 4.
Example 4 Consider the function f defined hs follows
f(x) = (x2+ l ) i f x I l , (x 3)2ifx> I
This function has three extreme points, P, Q and R (see Fig. 5). You should be able to see for yourself that P is a maximum and R is a minimum point off. The function also has a minimum at Q, that is, at x = 1. Let us see why. Consider the 1 1 3 i that is,] 2 9 [of 1. The function in this interval
+
[,
is defined as follows : f(x) =
<x 5 1 [(x  3)2, when 1 < x < 3 / 2
(  1  x2, when,
L
1
The Ups and Downs
Fig. 5
For
1 < x < 1,x2< 1 2.
* x2>
1
Now, when 1 < x < 3/2, f(x) is positive, and therefore is greater than f(1). Hence f(r) > f(1) whenever x E ]1/2,3/2 [ . Thus f(1) =  2 is a minimum value of the function.
Remark 1 We found in the above examplethat f(0) =  1 was a maximum value o f f whereas f(3) = 0 was a minimum value. Did you find it hard to swallow that a maximum
value of the function is smaller than a minimum value? If yes, recall our definition of an extreme value. We were concerned with the values of the function only in a neighbaurhood (that is, points nearby) of the extreme point. Thus, the concept of maximaminima is essentially a local phenomenon. What happens globally or elsewhere was not under consideration. For this reason, some people use tht terms local (or relative) maximum and local (or relative) minimum instead of maximum and minimum.
A value f(c) such that f(c)
> f(x) for all x in the domain of the function, is then called an absolute (or global) maximum. Similarly, if f(b) I for all x in the domain, then f{b) is f(x) called the absolute (or global) minimum. Therefore, a function may have many local minima or maxima, but it can have only one absolute minimum or maximum. In the light of this we see that f(0) = 1 is the absolute minimum value for the function in Example 1.
E
E 1) Find the maxima, minima of each of the following functions. If a function has no maximum/minimum say so.
i) f(x) = 5, for all x E R. ii) f(x) = x, for all x E R. iii) f(x) = x, for 0 < x < I . [Be careful! 0 and 1 are NOT in the domain of the function!] iv) f(x) = x2. for all x E R.
.
\

.
fnr
all u F 14 161
Drawing Curves

7.2.2 A Necessary Condition for the Existence of Extreme Points
So far, we have used the graph of a function as an aid to discovering extreme points. Clearly, this is not going to be feasible in all circumstances. Drawing graphs may be timeconsuming and cumbersome. Certainly there must exist quicker and neater techniques. In this subsection we describe an analytic method of finding extrema. But first, a definition :
Definition 2 A point c of a set'A 6>0,]~6,c+6[CA.
C R is said to be an interior point of A if for some
The point 1 E A = [O, 21 is an interior point, since 11  6, 1
+ 6 [ C [O, 21,
1 if we choose 6 =  . But neither 0, nor 2 is an interior point of A. 2 The following theorem gives us a necessary condition for the existence of an extremum.
Theorem 1 Let f be a function which is derivable at an ifiterior point c of its domain D. If f has an extremum at c, then f(c) = 0.
Proof : Since f is derivable at c ((c) exists. This means that ~ ( ( c ) ~ f ( c exist, and , and ) are equal. Su pose, first, that f has a maximum at c. This means. f(x) 5 f(c), V.x E ]c  6, c 6[, where 6 is some positive number.
P
+
Similarly, if c < x < c fierefore
+ 6, then x  c > 0, and
f(x)  f(c)
.X
C
LO
I (c) = lim f XC '
[
X
 f(c) C
] 2 0, and
But we know that Lt'(c) = ~ f ' ( c ) , Hence the above inequalities yield Lft(c) = kf'(c) = 0. that is, ft(c) = 0. Proceeding exactly as above, we can prove that ((c) = 0 even when f has a minimum at c (see E 2)).
The Ups and Downs
Remark 2 The condition f'(c) = 0 is only a necessary condition for f to have an extremum at c. This means that iff has an extremum at c, then we must~~ecessarily have t'(c) = 0. But it is by no means a sufficient condition. In other words, if we are given that ((c) = 0, this information is not sufficient for us to conclude that f has an extremum at c. That is, a function may not have anextreme value at c even though ((c) is zero. For example, take the function f(x) = x3shown in Fig. 6. We know that f'(0) = 0. Now, f(0) = 0, whereas f(x) c 0 for 6 c x c 0, and f(x) 5 0 for 0 cx c 6, whatever pos~tive number 6 rnay be. rhus there is no 6 for whlch either f(x) 5 f(O), for all x E ] 6,6[or f(x) 2 f(O), tor all x E I 6,6[.
This indicates that f has neither a maximum, nor a m~nimum x = 0, even though t'(o) = 0. at
>2 0 2 1

i
/
Remark 3 The condition t'(c) = 0 applies in the m e when f'(c) exists. But a function may have an extremum at x = c even though it is not derivable at c. For instance, the function f of Example 4 is not derivable at x = I . Yet it has a minimum there. Similarly, function f(x) =  1x1 is not derivable at x = 0 but has a maximum there. Can you point is out another function from the examples above, wh~ch not derivable but has an exiremum?
From this discussion we arrive at the following corollary: Corollarr 1 If a function f has an extreme value at a point x = c, then oneof the following conditions is satisfie'd : i)
i
1 Fig. 6
f is n ~ derivable at c. t
ii) ' f is derivable at c, and <(c) = 0.
Definition 3 A point at which either ((x) does not exist or is zero is known as a critical point.
If you study Definition 3 and Cor. I carefully, you will realise that any extreme point is a critical point. But the converse is not true. That is, a critical point need not be an extreme point. You will find an example to illustrate this in Remark 2 above.
Example 5 Consider the function f(x) = x2I3for all x E R. 2 ((x) = 7 x 'I3. Thus, this function is not derivable at x = 0.
This means that f has a critical point at x = 0. Further, at other points t' exists, but ((x) = (2/3) x'I3 # 0. Hence, the function has only one critical point, namely 0. For every 6 > 0, and every x in L 0 and f(0) = 0. Thus x = 0 is a minimum point and ]  6, 6[, f(x) = x2I3= f(0) = 0 is the minimum value (also see Fig. 7).
Flg. 7 : Graph of f(x) =x*,
31
Drawing Curves
Example 6 A farmer has a certain length of fencing wire. He wants to make a fence in a rectangular shape to keep his goats. Now what dimensions should he choose to ensure maximum enclosed area?
It is clear that in order to get a maximum enclosed area, the farmer should use the whole length of wire. Suppose it is 4k metres. Then 4k is the perimeter of the enclosure. Thus'if x and y are the length and the breadth, respectively, of the enclosure (in metres), then 2(x y) = 4k, or y = 2k  x, 0 < x < 2k
+
Thus, we can regard the enclosed area xy (in m2) as a function f of its length x. Thus, we have the function f defined as f(x) = x(2k  x), for all x > 0 (length is always positive).Our task is to find that value of x, for which f(x) becomes maximum. Now f(x) is a polynomial in x, and therefore f is derivable for all x. Hence its critical points (among which we should seek the extreme values) are obtained from the equation f(x) = 0, which in this case gives 2k  2x = 0, that is x = k. This means that the function f(x) = x(2k  x) has a critical point at x = k. Let us see whether it has an extremum at x = k. Now, f(k) = k2, and
Since there is only one maximum value, namely f(k). of f(x). O < < 2k. f(k) 4 ~for )
It is obvious from (1) that for values of x less than or.greater than k, (x  k)' > 0 and therefore, f(x) < kZ = f(k). Hence, for any 6 > 0, f(x) I for all x E ]k  6, k 6[. f(k) This means f(k) is a local maximum. But since there is only one extreme point, f(k) is the maximum value of the area. Now, when x = k, y = 2k  x = k. So, the farmer should . have a square fence made, taking the entire length of his wire.
+
In this example it was easy to compare the neighbouring function values with f(k). Are you wondering how you would tackle this problem when the functions under consideration would not be so nice? There is no need for speculatisn. We have suficient conditions which tell us whether f(x) is an extreme value or not, and if it is, whether it is a minimum or a maximum value, without actually having to compare values. Section 6 is, in fact, devoted to such explorations. Before that, in the next section we shall discuss the mean value theorems. But now it's time to do some exercises.
E
E 2) Prove that f(c) = 0 iff has a minimum at c.
E
E 3) Find the critical points of each of the following functions.
The Ups and Downs
c) f(x) = sin x
+ 3,Vx E R.
d) f(x) = ex,ffx E R. e) f(x) = 2 lxl,ffx ER.
;.
L
$ti
7.3 MEAN VALUE THEOREMS
I
*, I
In this section we shall study the mean value theorems. These theorems have proved to be very handy tools in proving other theorems not only in calculus, but also in other branches of mathematics, such as Numerical Analysis. Their importance lies in their wide applicability and tremendous usefulness. We shall first consider Rolle's Theorem, which is a special case of Lagrange's mean value theorem. We shall not attempt the proofs of these theorems here, but you will agree that both are intuitively obvious. We sball discuss their geometrical significance and illustrate their usefulness through some tramples.
7.3.'1 R~lle'sTheorem
Rolle's Theorem was not actually proved by Rolle. He had only stated it as a remark. In
k e t Miehel Rnlle / I A<?1 7 1 01 u r m a Gnntun
tn
he
0
er;t;e nf the natulu fn.nnAd thenrr~ nf
Drawing Curves
Newton and Leibniz. It is ironical, t h p , that one of the most important theorems of this theory is known after h h . Now iet us see what this theorem is.
Fig. 8
In Fig. 8 we see the graphs of two continuous functions defined on the closkd interval [a,b]. Here we observe the following features common to both of them.
Rough Statement
1 The curve is drawn without breaks or gaps. 2 There are no corners in the curve, except possibly at the end points.
Precise Statement
The function f
IS
continuous on [a,b].
The function is differentiablein the open interval la,b[. f(a) = f(b) f (c) = 0 for some c in ]a, b[.
3 The two end points of the curve lie on the same horizontal line.
4 The curve admits a horizontal tangent (drawn as a dotted line) at some point.
The line joining the two end points may be imagined to be pushed upward or downward, keeping it always horizontal, and keeping the curve unmoved. Then there is a position, shown by the dotted line, where it touches the curve. This makes us believe that the fourth property holds for all the functions satisfying the first three properties. This is what Rolle's I Theorem states.
Theorem 2 (Rolle's Theorem) Let f be a function continuous on the closed interval [a, b] and differentiable in the open interval ]a, b[. Further. let f(a) = f(b). Then there is some c in ]a, b[ such that ((c) = 0.
We give some examples below to illustrate this theorem.
Example 7 Consider f(x) = sin x on the interval [O, 2 ~ 1All the assumptions of Rolle's . theorem are satisfied here. f(0) = 0 = f(27r)
Therefore according to Rolle's theorem, there should exist c in ]O,27r[ such that f (c) = 0. Here f (c) = cos c. Can we find an element c such that cos c = O ?
V n o
1 ,
f n ~ thntn n t n t u r n o n ~ n n i n t o t h
P
in
ln
3r
namnlrr
/3
rind 2  / 3
The Ups and Downs
Fig. 9
At ~ / 2 the function sin x attains its maximum value. , , At 3 ~ / 2 the function sin x attains its minimum value. Both these belong to the interval 10, 274. Rolle's theorem asserts that there is at least one c in ]a, b[ such that f'(c) = 0. Example 7 shows us that there may be more than one points in ]a, b[, at which f'(x) = 0. In Rolle's theorem, a function f on [a, b] has to satisfy three conditions. i) .f is continuous on [a, b] ii) f is differentiable on ]a, b[ iii) .f(a) = f(b) Now, we shall see through some examples that each of these conditions is essential. We cannot drop any one of them and still prove the theorem.
Example 8 Let f(x) = x  [x] = fractional part of x, be defined on [0, 11. This can also be described as
I
/
Fig. 10
Here : = f(1) = 0. f is differentiable in the open interval 10, l[. Thus, two of the three ) . ' conditions of Rolle's theorem are satisfied by fi. The derivative off is 1 at every point of 10, I [. There is no point of 10, 1[ where the derivative is zero. What happens to Rolle's Theorem in this example? Obviously, its conclusion does not hold here. The reason is that f is not continuous ,on the closed interval [0,1], since it fails to be I continuous at 1. In the next example, we see that the assumption of differentiability in ]a, b[ cannot be omitted.
Example 9 Consider f(x) = 1x1 on [ 1, I]. There is no c in j 1, 1[ such that ((c) = 0. Actual computation shows that
ft=[
ifl<x<O, 1 I ifo<x< 1 does not exist at x = 0.
f is continuous on [ 1, I]. Also, f( 1) = f(1). But f is not differentiable in 1 1, l[. Our next example shbws that the assumption f(a) = f(b) is essential in Rolle's Theorem.
Example 10 Let f(x) = x3 on [0, 11. Then f is continuous on [0, I], and is yfferentiable in 10, l[. But f(0) # f(1).
Fig. I I
35
<
I
; " /
+Wp
; i" bi
tY
.
,I In this case fP(x)= 3x2 # 0 for any x 610, l[. Thus, we see that the conclusion of Rolle's theorem may not hold when f(a) # f(b).
Lastly, we give an example where Rolle's Theorem is applicable and yields a unique c. Example 11 Let f(x) = x2on [ 1, I]. Then ((x) = 2x. Here all the three conditions of Rolle's Theorem are satisfied. There is only one c, namely c = 0, suchthat f (c) = 0. You will noy be able to solve these exercises.
,
\I I
E
.
Fig. 12
E 4) Can Rolle's Theorem be applied to each of the following functions? Find 'c' in case il can be applied.
] a) y = sin2xon the interval [0, n .
b) f(x) = x2 C) d)
+ 1 on [ 2,2]. f@) = x3 + x on [0, I]. f(x) = sin x + cos x on [0, rr/2].

e) f(x) = sin x  cos x on [O,21cl.
E 5) Consider the function f(x) = x2  3x 2. Prove that f( 1) = f(4). Find a point c between  1 and 4 such that the derivative of f vanishes at c. Is this point the midpoint of  1 and 4? 
+
T&e Ups and Down
E
E 6) For the same function f(x) = x2  3x r1,21.
+ 2, verify Rolle's Theorem on the interval


t

E
E 7) Let f(x) = ax2
+ bx + c be the given function. If p aad q are two real numb=
P+ [ 2
] = o m
that f(p) = f(q), prove that f
E
E 8) Consider the curve y = ax2 bx c. Let xo be the unique real number such that yo) the tangent at (XO, to this curve is horizontal. Prove that the function y is oneone on the interval [xo, [ (Hint : If f(p) = f(q), apply Rolle's theorem in the interval [P,Q1).
+ +
Drawing Curves
I
E
E 9) Let I be an open interval of R. Let f : I R be a differentiable functiin such that f does not vanish on I. Prove that f is oneone on I.
7.3.2 Lagrange's Mean Value Theorem
Now we shall discuss the mean value theorem. It was proved by Joseph Louis Lagrange, a towering mathematician of the eighteenth century. We have already mentioned that Rolle's theorem is a special case of the mean value theorem. Let us recall the statement of Rolle's Theorem in the following form. Let f be a continuous function on the closed interval [a, b], and differentiablein the open interval lab[. The graph of f is a curve in the plane. If the endpoints of this curve lie in the same horizontal line, (that is, f(a) =. f(b)) there is a point c on the curve where the tangent to the curve is horizontal (((c) = 0). The last sentence can be restated as follows. If the endpoints of this curve lie in the same horizontal line, there is a point on the ctirve, where the tangent to the curve is parallel to the line joining its endpoints. The mean value theorem asserts the same conclusion, even without the assumption of horizontality of the line joining the endpoints of the curve. Fig. 13 illustrates this difference. Here P and Q are the end points of the curve. The line PQ is horizontal in Fig. 13(a), but not in Fig. 13(b). But in both the cases the point R on the curve has the property that the tangent to the curve at R is parallel to the line PQ. The number c is the xcoordinate of R.
Fig. 13 (a) illustrates Rolle's theorem, whereas Fig. 13(b) illustrates Lagrange's mean value theorem. The two end points of the curve are (a, f(a)) and (b, f(b)). The line joining these two points has the slope [f(b)  f(a)]/(b  a). Any line parallel to this line will also have the same slope. Therefore, the conclusion of the mean value theorem is f(c) = [f(b)  f(a)J/(b  a) for some a < c < b. This is because, we already know that ((c) is the slope of the tangent to the curve at (c, f(c)). Now we are ready to give the precise statement of the theorem.
The Ups and Downs
'"* 6
Theorem 3 (Lagrange's Mean Value Theorem) Let f be a continuous function on a closed interval [a, b]. Let f be differentiable in the open interval ]a, q.Then there is a point c in f(b)  f(a) the open interval ]a, b[ such that ((c) = ba '
Rolle's Theorem has three assumptions :a continuity assumption, a differentiability assumption, and the assumption f(a) = f(b). The mean value theorem has only two assumptions. These are the same as the first two assumptions of Rolle's Theorem. Suppose in addition to the two assumptions of the mean value theorem, f(a) = f(b) also holds. Then what does the mean value theorem yield? It says that f'(c)
I
=
 f(a) for some a < c < b. ~ u f(b)  f(a) = 0. t b a
Therefore, we get ((c) = 0 for some a < c < b. This isthe same as the conclusion of Rolle's theorem. This proves our contention that Rolle's theorem can be deduced from the mean value theorem. But whjl the name mean value theorem? What is the mean value here? f(a) is the initial value off. f(b) is the final value off. Therefoie f(b)  f(a) is the total change in the value off. Tliis change has occurred when the xcoordinate has changed from a to b. For a change of b  a in the domain, there is a change of f(b)  f(a) in the value off. Therefore, the mean value, that is, the average value of the rate of change is [f(b)  f(a)]/(b  a). The mean value theorem asserts that this average value of the rate of change of f is assumed at some .pint c by derivative f. We shall illustrate the same thing by means of an example. Consider a car moving from time a to time b; let f(t) be the position of the car at time t. Then the average speed of the car is distance/time = [f(b)  f(a)]/(b  a) According to mean value theorem, the speedometer of the car would have shown this [f(b)  f(a)J/(b  a) at some time between a and b. For instance, if the car has travelled 100 kms. in two hours, at some point of time, its speed would have been actually 50 kmph. (which is its average velocity over the span of two hours).
Example 12 Let us verify the truth of Lagrange's mean value theorem for the function f(x) = x2  2x on the interval [l, 21.
This is a polynomial function. Therefore it is continuous on [1,2] and differentiable in ]1,2[. (We have seen this in Units 2 and 3). Here
f(b)f(a)  = ba
0(1)
21
=1
We want to check that ((c) = 1 for some c such that 1 < c < 2.
Drawing Curves
Now, 2x  2 = 1 when x = 3/2 and 2 3 1 '
E ]1,2[. Thus, we see that
Now consider the function f : [a, b]  R which satisfies the assumptions of mean value theorem. Let p and q be any two points such that a 5 p < q I b. Is there some c between p and q such that
f (c) = [f(q)  f(p)]/(q  p)? To answer this, consider the restriction off to the interval [p. q]. It satisfies the assumptions of the meanvaluetheorem. Therefore such a point c exists.
This result can be geometrically interpreted as follows. (p, f(p)) and (q, f(q)) are two points on the curve y = f(x). f(q)  YP) is The line joining them is called a chord of the curve. 'IP the slope of this chord. What we have shown is that the slope of this chord is the same as the slope of the tangent at the point (c, f(c)). This means, that the tangent at (c, f(c)) is parallel to the chord (see Fig. 14). Thus, for any chord of the curve, there is a point on the curve where the tangent is parallel to the chord.
L ! i i ;
a 0
;
!
V
;A/
P
~ i g 14 .
Example 13 i) Let us find the point c in ]  r / 4 , 7r/4[ such that the tangent to f(x) = f(x) = cos x at <c,ftc)) is parallel to the chord joining (n14, f(~14))and (n14, f(nl4)).
ii) We shall prove further that for the same c, the tangent at (c, g (c)) to the curve g(x) = cos x xZ x is parallel to the chord joining ( 7r/4, g( 7r/4)) and ( ~ 1 4g(7r/4)). ,
.
+ +
i)
The slope of the chord joining ( n/4, f( 7r/4)) and (7r/4, f(s/4)) is f(7r/4)  f( ~ 1 4 ) 1 1 6  l / f i = 0. 7r/4  ( 7r/4) 7r/2 Therefore we seek c such that f'(c) = 0. We have f(x) =  sin x. The only point in 1 7r/4, 7r/4,[ where this vanishes is at c = 0. The corresponding point on the curve is (O,f(O)) = (0,l).
ii) g( 7r/4) = ( l / n )
+ (7r2/16)  (7r/4)
The slope of the chord joining (( 7r/4), g( 7r/4)) and (7r/4, g(a/4)) is
When c = 0, we want to prove that the tangent at (c, g(c)) to the curve g(x) also has the same slope 1. In other words, we must prove that g'(0) = 1. ,
, '
I
Y !
I
Now, i ( x ) =  sin x 2x 1. ; g'(o)=O+o+l=l. This proves that, for both the furlctions f(x) and g(x) over 1 d 4 , 7r/4[, it is the same point c where the conclusion of the m;ean value theore  holds.
+ +
Example 14 For the curve y = In x, suppose we want to find rr point on the curve where the tangent is parallel to the chord joining the points (1, 0) and (e, 1).
/1
e
X
Since in 1 = 0 and ln e = I , these two points (1.0) and (e, 1) lie on the curve y = In x. Consider this function on the closed interval [I, el (see Fig. 15). It is continuous there.It is also differentiable on 11, e[. Therefore by the mean value theorem, there is a point c between I and e such that the tangent at (c, In c) is parallel to the chord joining (1,O) and (e, 1). We have to find this
nn;n+ N r \ r . r
r.'
~ i g 1s .

1 I..
The required point is given by
The Ups and Down.
: c=e1 . The required point on the curve is (e.  1, In (e  1)).
Remark 4 Let f : [a, b]  R satisfy the assumptions of the mean value theorem. Then 3 0 , 0 < 8 <1, suchthatf(b) =f(a) + ( b  a ) f P ( a + 8 ( b  a ) ) .
This is because any point c between a and b is of t:~eform a 8 (b  a) for some 0 < 8 < . I . Note t h a t a = a O(b a ) a n d b = a l(b a).
+

+
+
. , !
Are you ready for these exercises?
E 10)Verify the mean value theorem for f(x) = xZ
a) [ 1 , l I b)
+ 1 on the following intervals
1 4 2 1

E 1l)Verify the mean value theorem on the interval [O, 21 for the following functions
a) f(x) = sin rrx
I
Drawing Curves
E
E 12)a) Let f(x) = x3 on [0, I]. Find a point c in 10, I[ as in the mean value theorem.
b) Let f(x) = x3 on [ 1, 01. Find a point c in 1 1, O[ as in the mean value theorem. c) Let f(x) = x3 on [ 1, 1E Show that there are two points c in 1 1, l[such that
E
E 13)Let f be a function on [a, b] satisfying the assumptions of the mean value theorem, Let c be a point guaranteed by the mean value theorem. Prove that if gdx) = f(x) 1 and g2(x) = f(x) x for all x in [a, b], then the same point c satisfies '  gl(a) = g'l (c) ba
+
+
and
 g4a) = g'2(c) also. b a
E 14)At what point is the tangent to the curve y = xn parallel to the chord from
a) (0,O) to (2,2")7 b) (O,O)to(l, 1)7
'
The U p s and Downs
Just as in the case of Rolle's theorem, there may be more than one points at which the tangents may be parallel to the chord joining the end points of a curve represented by a function which is continuous at every point in the closed interval and is differentiable at every point in the open interval (see Fig. 16).
I
Fig. 16
Both Rolle's theorem and Lagrange's mean value theorem are existence theorems. They tell us that there exists at least one point whcre the tangent is parallcl to the chord joining the end points. But they do not tell us how many such points are there, nor how to find these points. For example, consider the function f(x) = x3  sin x on [O, 5 .rr] It satisfies the conditions of the mean value theorem. S o there is at least one value c at which 3c2  cos c = 25z2 .
Drawing Curves
The mean value theorem assures us that the equation 3x2 r cos x = 25 .rr2 has atleast one solution, c. But it does not enable us to find the value or values of C.You can study methods of solving such equations in the course in nurrlerical analysis. In the next section we shall see how the mean value theorem helps us to derive sufficient conditions for the existence of extreme points of a function.
7.4
SUFFICIENT CONDITIONS FOR THE EXISTENCE OF EXTREME POINTS
In Sec. 2 we have seen that a necessary condition for the existence of an extreme point of a given derivable function is that the derivative is zero at that point. We have also seen that the condition is not a sufficient one. In this section we shall discuss some tests which give sufficient conditions for the existence of extreme points. Lagrange's mean value theorem which we have studied in Sec. 3 is used in deriving these tests.
7.4.1 First Derivative Test
The following theorem gives a sufficient condition for a function f to have an extreme value at an interior point c of its domain. It also tells us whether the extreme value is a minimum or a maximum.
Theorem 4 (FIRST DERIVATIVE TEST) Let c be an interior point of the domain of a function f. Suppose f is derivable on ]c  6, c 6[ for some 6 > 0 and that f'(c) = 0. Then
+
a) if ft(x) > 0 when c  6 < x ft(x) < 0 when c < x < c
+ 6, then f has a maximum at c, a i d
< c and
b) if ft(x) < 0 when c  6 < x < c and ft(x) > 0 when c < x < c 6, then E has a minimum at c.
+
Proof a) We have to prove that f has a maximum at c. In other words, we have to show that f(c) 2 f(x) for all x in some neighbourhood of c.
Now, let x E ]c  6, c[. Then f is differentiable on ]x, c [ (in fact on [x, c]), and continuous on [x, c] because differentiability at a point implies continu~ty there (Theorem 6, Unit 3). By Lagrange's theorem, there exists a E ]x, c[ such that f(c)  f(x) = f'(a). That i$
C X
f(c)  f(x) = (C  x) f'(a). Since f(x) > 0 on ]c  6, c[ and a E ]x, c[C]c  6, c[, therefore, ft(a) > 0. Also, c  x > 0 sincax E ]c  6, c[. Hence, from (1) \ f(c)  f(x) > 0 or f(c) > f(x) .
..... (1)
..... (2)
Similarly, if x E ]c, c 6[, then by Lagrangqs theorem, there exists a point b E ]c, x[ such that .....(3) f(x)  f(c) = (x  c) ft(b). Since ft(x) < 0, when x E ]c, c @, ft(b) < 0. Further, since x E ]c, c 6[, x  c > 0.
+
+
+
(3) now gives
.....(4) f(x)  f(c) < 0 or f(c) > f(x) . c[ Thus, it follows from (2) and (4) that whether x E ]c 6, or to ]c, c 6[, f(c) > f(x). In other words, f(cJ 2 f(x) V x E ]c  6, c 6[. Hence, ff$) is a maximum value off and f has a maximum at c.
+
+
.
The proof of part (b) proceeds on similar lines. See if you can write it yourself.
E
E 15) Prove part (b) of Theorem 4.
The Ups and Downs

We can also interpret Theorem 4 as follows. If while crossing a point, moving from left to right, the derivative changes sign from positive to negative, then that point must be a maximum. Similarly, while crossing a point, if the derivative changes sign from negative to positive, then that point must be a minimum.
Remark 5 From Theorem 4 we can deduce that when ((x) > 0 in an interval, the function f'is strictly increasing in that interval. Similarly, f'(x) < 0 in an interval would imply that f is a strictly decreasing function in that interval. f' = 0 in an interval, likewise, means that the function is a constant function in that interval. We shall talk about this in Block 4.
You should note that the conditions stated in Theorem 4 are sufficient and are by no means necessary for the existence of extreme points. Thus, a function may have an extreme value at c even though the conditions given in the theorem are not satisfied as shown in the . following example.
Example 17 Let f(x) = x2[5 sin (l/x)], x # 0, and f(0) = 0. Thenf(x)> Oifx # 0,asx2 > O a n d 5 +sin(l/x) >Osincesin(l/x) 2 1.Thusf has a minimum at x = 0. ((x) = 2x[5 sin(1/x)]  cos (1 /x) V x # 0, and f(0 h)  f(0) r' (0) = lim h0 h h2[5 sin ( 1/h)] = lim h0 .h
+
+
+
+
= lim h[(5
sin (1 /h)] h0 = 0 since 4 I sin(l/h) I 5 6
+
+
Thus, f'(x) exists for all x. However, neither in ihe interval 1 6, 0[, nor in 10, 6[ does f'(x) keep the same sign, no matter how small 6 is (you may check this yourself). Thus, the question of ((x) being positive or negative on two sides of 0 does not arise.
Example 18 Among all rectangles having a given area, is there one which makes the perimeter a minimum?
Suppose that the length and the breadth of the rectangle are x and y, respectively. Then x > 0, y > 0. The area xy being constant, kZsay, we have y = k2 x'. Then, the perimeter of this rectangle = 2 (length breadth) = 2 (x kZx'), x > 0. This is a function of x alone. Let us denote it by f(x). Thus f(x) = 2(x k2 xI), x > 0.
+
+
+
Drawing Curves
We want that value of x which makes f(x) a minimum. The function f is derivable at all points of its domain. The derivative is given by f'(x) = 2 [1  (k2/x2)] Now, f'(x) = 0 =3 x2 = k2 =3 x = k. If we assume that k > 0, then  k is not in the domain off. Thus the only critical point is x = k. Let 6 be any positive number such that k  6 > 0. Then f is derivable at all points of ]k  6, k 6[.
+
+
ii) x E ]k, k
+ 6[
> k, * k2/x2 < 1, = f'(x) > 0. 3
= x 3
Hence by Theorem 4, f has a niinimum at x = k. Also, when h = k, y = k2 X' = k. This means that a square shape yields a minimum perimeter among all rectangles with a fixed area. You should be able to solve the following exercise now.
E
E 16) Find all possible extreme values of each of the following functions by applying the first derivative test.
a) f(x) = x5  5x4 5x3  1 for all x E R [It may be helpful to factorize f'(x)] b) f(x) = 2x4 c)
+
+ 8x3  4x2  24x + 15 for all x E R f(x) = (X  1)' (X + 1) for all xE R
7.4.2
Second Derivative Test
ThKUps and Downs
We now investigate another condition which, if satisfied, does away with the need to examine the sign of f'(x) in ]c  6, c[ and ]c, c + 6[ as in the first derivative test. This condition also is only sufficient, but very functional and hence, useful.
Theorem 7 (Second Derivative Test) Let f be derivable in ]c  6 , c f 6 [ for some 6 > 0 and suppose ((c) = 0.
Then a) 'f has a maximum at x = c provided f"(c) exists and is negative.
b) f has a minimum at x = c provided f"(c) exists and is positive.
Proof: a) Since ('(c) < 0, f is a strictly decreasing function in the neighbourhood of c (see Remark 5). Thus there exists an E > 0 (and we may take it smaller than 6) such that ((x) > ((c) %hen x E ]c  E , C[(since x < c) and ((x) < ((c) when x E ]c, c E [ (since, here, x > C)
+
Since ((c) = 0, this means that ((x) E.[. x E ]c, c
+
> 0 when x E ]c 
E,
c[ and ((x)
< 0 when
By the first derivative test it follows that f has a maximum at c. The proof for (b) follows along similar lines.
Remark 6 You must have observed that this theorem says nothing about the case when r ( c ) is also zero. In this case the function may have a maximum or a minimum value or neither as the following examples show.
i) f(x) 7  x4, for all x E R. Here ((0) = 0 = ('(o), but the function has a maximum at 0 (see Fig. 17(a)). .
ii) f(x) = x4, for all x E R. Here ((0) = 0 = r(O), but the function has a minimum at 0. (see Fig. 17(b)). iii) f(x) = x', for all x E R. Here ((0) = 0 = f"(0) and the function has neither a maximum nor a minimum at 0 (see Fig. ,17(c)). Thus,the first derivative test does have some merit.
I
(a)
T
4
(b)
Fig. 17
(4
*r
".
Example 19 Let us tind the extre&e values of the function f defined by ' f(x) = 2x 3/x, for all x f ir Here, ((x) = 2  3/x2, and therefore f)x)=0 r = f
+
Also, r ( x )= 6x', This means = 2,. . d* ,

&
 .
, , '
Drawing Curve2
Since, f ' (   a ) = 0 = f ' ( a ) . it follows from the second derivative test that f ha4 a and a maximum at The minimum value is minimum at
m.
f ( m ) = 2 f i , and the maximum value is f(
a) Jy.
=
2
Example 20 From each corner of a square paper of side 24 cm, suppose we remove a square of side x cm and fold the edges upward to form an open box. Let us try to find that value of x which will give us a box with maximum capacity.
Fig. 18
,
Clearly, 0 < x < 12 for a box to be formed. Also, the box thus formed has dimensions (24 '2x), (24  2xj and x (see Fig. 18). The volume f(x) is a function of x given by f(x) = (24  x12 x, 0 < x < 12. = 4x3  96x2 242x. ((x) = 12x2  192x 242 = 12(x  4) (x  12) Now, <(x) = 0 =3 x = 12 or x = 4. Since 12 is not in the domain of our function f, 4 is the only critical point. Also, r ( x ) = 24x  192. Therefore, ('(4) = 96  492 < 0. Hence x = 4 is a maximum point off. The maximum value f(4) of f (that is, the maximum ciipacity of the box) is 1024 cm3.
+
+

Are you surprised that the box is not a cube for maximum capacity? But, had it been a cube, four squves each of side8 cm (the removed portions) would have been wasted, Ghereas, now four squares each of side only 4 cm have been thrown away. There had to be a compromise between the waste material and making the box as near a cube as possible! Moral : Mathematics does not fail even though intuition may! Here are some exercises for you to solve.
E
E 17) Find the extreme points for each of the following functions. Point out which of them
are maximum, which areminimum and which are nei1l.s. Also find the extreme values off.
d) f(x) = ao a1x2 a2x4 ... a,x2", where x E R and each a, is positive. [Do not get bogged down by the degree of f(x)!] e) f(x) = x/(x2
+
+
+ +
00
+ I), 0 < 4 <
The Ups and Downs
I
E
E 18)Show that 7r/3 is a critical point of f, where f(x) = sin x (1 cos x), x E R.Does f have a rnaximuni or a minimum at this point?
+
E
E l%)~howihat the rectangle of maximum area which can be inscribed in a circle, is a square. [Notice that the,diagonals of the rectangle must be the diameters of t h circle].
D M w i n g Curves
E
E 20)A man wants a nameplate with display area equal to 48cm2 bordered by a white strip 2 cm along top and bottom and 1 cm along each of the two remaining sides. What dimensions should the plate have so that the total area of the plate is a minimum?
7.5
MORE INFORMATION FROM THE SECOND DERIVATIVE
In the previous section we used the sign of the second derivative at a critical point to discover whether the function has a maximum or a minimum at that point. For drawing the necessary conclusion we regarded as the derivative of the function and then applied the first derivative test. W e shall now use the monotonicity considerations off' to draw another conclusion, which would give us an idea about the shape of the graph of the function f. Given that a function f increases in [a, b], its graph may be anything like (a), (b) or (c) in Fig. 19
r,
In this section we shall use the second derivative to determine the type of graphthe given function f has. We shall assume that the function is twice differentiable in it$ domain.
Fig. 19
I f f " > 0 in some interval [a, b], then I increases in this interval (Remark 5 applied to I ) In other words, as x increases, the slope of the tangent to the graph at (x, f(x)) increases, so that the tangent turns counterclockwise. This results in the graph bending upward or bulging downward. Such graphs and functions are known as convex. We also use the terms concave upward or convex downward. Such graphs lie below their chord (linesegment joining the points on graph which correspond to x = a and x = b) and above their tangents. See Fig. 20(a). Similarly if f" C O in [a, b] so that f' decreases, and hence slope decreases, the tangent turns clockwise,. Such functions are known as concave (equivalently, concave downward and convex upward). The graph in this case lies below the tangents and above the chord. (see Fig. 2qb)).
Fig. 20
We say that a function f is concave (convex) at a point, if it is concave (convex) in a neighbourhood of that point.
Remark 7 : i) The concavity is towards the chord. If the chord lies above the graph, the graph is concave upward. When the chord lies below the graph, the graph is concave downward.
ii) Every convex function is of the form  f,where f is concave. iii) Ohly concave and convex functions have the property that each of their tangent lines intersect their graph exactly once. Till now we were concerned only with the manner of bending of a graph.*let us now discuss the measure of the bending 6f a graph at a point also known as the curvature at that point. We measure t$e curvature at fc. f(c)) by the ratio
Drawing ; L..
k(c) =.
f "(c)
[I
+ ?'(C)~]'/~
The radius of curvature at (c, f(c)) is denoted by p(e) and is defined by P(C) = k(c) f 0. You will note that the radius of curvature is positive if the function is convex at that point and is negative if the function is concave there. How does one get this ratio? Oh well! Surely, you don't think this course covers the why of everything about functions?
7.5.2 Points of Inflection
Suppose f"(c) = 0 for some c E ]a, b[, and f" charjges sign in passing through c. That is, f" has one sign on the left of c, in ]c  6, c[ say, and the other on the right of c, in c 6 [ say. Then on one side of (c, f(c)) the tangents are above the graph and on the other side of (c, f(c)) these are below the graph. This means that the tangent\at (c, f(c)) must k crossing the graph (Also see Fig. 2 1). '
+
Such points (c, f(c)) are known as the points of inflection of (the graph of) f.
Fig. 21
Note that we did not really use the fact that f"(c) is zero. In fact f"(c) = 0 is neither a necessary nor a suffitient condition as you will find out from Examples 21 and 22. A function may have a point of inflection at x = c without f"'(c) existing at all. For the tangent to exist, the existence of f'(c) is enough. However, note that if f"(c) exists in ]c  6, c 6[ and changes sign in passing through c, then ft'(c) must be zero.
+
Example 21 Consider the function f(x) = x 1x1, for all x E R shown in Fig. 22. This function can be rewritten as
Hence f'(x) = 2xifx20 2xifx<O
But f"(0) does not exist and
Fig. 22
Thus I' changes sign at 0 and the tangent crosses the graph at 0. x = 0 is thus a point'of inflection even though f"(0) # 0 (it does not exist!).Hence f"(c) = 0 is NOT a necessary condition for f to have a point of inflection at x = c.
Example 22 Let f(x) = x4.The first derivative test shows that f has a minimum at x = 0. Thus f does not have a point of inflection at x = 0 even though f"(0) = 0 (also see Fig. 17(b)). This happens because f"(x) = 12x2> 0 for all x # 0, and accordingly it <oes not change sign in passing through 0. Thus, f"(c) = 0 (by itself) is not sufficient either, for f to have a point of inflection at x = c. Example 23 Suppose we want to find the values of x for which the graph of the function f(x) = 1/x, x E R \ (0) is convex (i.e., concave upward) and concave (i.e.'concave downward). We shall also find if the graph (Fig. 23) has any points of inflection, and the curvature at x = 1.
Here, f'(x) =  l/x2, f"(x) = 2/x:
Fig. 23
Clearly, i) f"(x) > 0 if x > 0 ii) f"(x) # 0 for any x E R \ (01, and iii) ff'(x) < 0 if x < 0 .
It follows from (i) that the graph in question is convex (i.e., concave upward) in 10, a[. From (ii) and the fact that f" exists for all x in the domain off, we conclude that the graph has no points of inflection. From (iii) we deduce that the graph is concave (i.e., concave 2 downward) in]  =,0[Further, curvature at x = 1, is f"(1)  2  1 [I f'(1 )2]3'2 [l 11312 fi
The Ups and Downs
+
+
Example 24 Suppose we want to examine the function f for points of inflection if f(x) = x2'+', for all x E R, n E N . ((x) = (2n 1) xZn,f"(x) = (2n 1). 2n. x2"'
If f '
+
+
( c ) e x i s ~ sthen f (c) = (0) , ' is a necessary condition for (c, f(c)) to be a point of inflection
Therefore, f"(x) =.O x = 0. Since fP'(x) exists for all x, there can be at most one point of inflection, namely, (0, f(O)), or (0,O). Now to the left of x = 0, that is; for x < 0, f"(x) < 0, and to the right of x = 0, c ( x ) 2.0. Hence, changes sign (from negative to positive) in passing through the origin. Thus the origin is a point 6f inflection of f(x) = xZn+I.
*
E 21)Examine each of the following functions for concavity, convexity and points of inflection. 
.
Drawing Curves
E 22)Find the curvature at an arbitrary point of the graph of the function
a) f(x) = x5, x b) f(x) = x2
E
R
+ 9, x E R
 1<x < 1
c) f(x) = sin x, x E R d) f(x) = d3,
SUMMARY
In this unit we have discussed the following points. 1) A function f is said to have a maximum (resp., minimum) value at a point c of its domain if there exists a positive number 6 such that for all x E ]c  6, c 6[, f(x) 5 f(c) (resp. f(x) 2 f(c)). Maximum and minimum values are known as the extreme values of the function,
+
2) At an extreme point c of a function f, either the derivative does not exist or is zero.
3) Critical points for a function are those where either the derivative does nct exist or else has the value zero. All extreme points are critical points. A critical point may fail to be an extreme point.
4) Rolle's and Lagrange's mean value theorems, and their geometrical interpretation. 5) A sufficient condition for a function f to have an extreme value at x = c is that f is .f continuous at c and the derivative f' changes sign m passing through c. If the change is from positive to negative, c is a maximum point. In the other event, c is a minimum point. This test is known as the first derivative test.
6) Second derivative test, another sufficient condition for the existence of extreme points asserts that if ((c) = 0 then r ( c ) > 0 implies f has a minimum at x = c and r ( c ) < 0 guarantees a maximum value at c.
7) If r ( x ) it.
> 0 on some interval then f is convex on it. If f"(x) < 0 then f is concave on
8) If r ( c ) = 0 or does not exist and changes sign in passing through c, then f has a point of inflection at x = c. This means the tangent at (c, f(c)) crosses the graph off at this point.
r
9) The radius of curvature =
[ 1 + 'f'(c:1213" ,f,#(c)
+
o.
f"(c)
7.7 SOLUTIONS AND ANSWERS
E 1) a) all point of R are maxima as well as minima.
b) no maxima or minima on R .
c) no maxima or minima on ]O, 1[ , d) 0 E R is a minimum. No maxima. e) x = 4 is a minimum, x = 16 is a maximum.
E 2) f has a minimum at c 3 6 such that f(x) 2 f(c) Yx E]c  6, c a[ *f(x)f(c)LOYxE]c6,c+6[
* +
and
0 )  f(c) I0,c6<x<c (X  c) Rf'(c) L 0 and Lf'(c) I 0. But since f is differentiable at c, f'(c) exists and Lf'(c) = Rf'(c) +each is equalro zero. f'(c) = 0.
E 3) a) f'(x) = (x f'(x) = 0
 *
: .
x =4 x = 4 is a critical point.
 3) + (x  5) = 2x  8
d) no critical points g) All points x for which 0 I x I 1 are critical points because the function is defined as
Drawing Curves
E4) a) Yes.yl = 2sinxcosx =sin2x = Oifx = n/2E[O, ST] C) No. f'(x) = 3x2 4 1 # 0 for x E [0, 11. Rolle's theorem does not hold as f(0) # f(1) d) Yes. f'(x) = cos x  sin x = 0 if x = n/4 E [0, n/2] e) Yes. f'(x) = cos x
3 + sin x = 0 if x = 4
ST
7 or  ST E [O,~ST] 4
E 5) 'c = 3/2. Yes. E 7) ap2
+ bp + c = aq2 + bq + c
f'(x) = 2ax
+b
+ bp = aq2 + bq 3  q2) + b(p  q) = 0 2a(p.+ q) + b = 0 (since p # q)
3 ap2
E 8) Suppose f(x) is not (1  1) on [XO, m[ a p, q €'fxo, m[, such that p # q and f(p) = f(q)
= xo as xo is the unique point with ft(xo) = 0 2 Therefore either p < x or q < xo, since p and q both cannot be equal to xo. o This is a contradiction as we have taken p, q E [xo, m[ .
E 9) Suppose p, q E I s.t. p # q and f(p) = f(q) If p < q we have [p, ql c I, f is differentiable on [p, q] and f(p) = f(q).
Thus f.satisfies the conditions of Rolle's theorem oa [p, q] 3 ft(xo) = 0 for some xo E [p, q] C I. But this is a contradiction. Therefore f is oneone
E 1l)a) f(0) = 0 = f(2) I(x) = ST cos STX
+
b) similar

1(1/2) = 0.
3 1 E [0, 21 s.t. f'(1) =
f(2)  sf(0)
_
I(x) = 3x2 = 1 : c=l/,h .
X
= 1 / 6 € [0, 11
f(1)  f( 1) =1 1  ( 1) c=1 / a ,are two points in [ 1, 11 f(1)  f( 1) set. f(c) = 1  ( 1)
1/0
~ l ( b  gda)  f(b)  f(a) = ft(c) = glt(c) ) b a ba
E 14)a) y~ = nxn'. Slope of the chord from (0,O) to (2,2") is Y2  Y l = 2"I x  XI 2 2 Xnl = 2"I nxnl = 2111 n X = 7"I'
b) Slope of the chord from (0,O) to (1, 1) is 1
,
: .
nxn' = 1
3
= x"l = l/n
' 3 x
1 =1nAI)
E 15)Let x E ]c  6, c[. Applying the mean value theorem to [x,c], 3 a E ]x, c[, such that f(c)  f(x) = f'(a) (c  x) < 0, since f'(a) < 0 and c  x > 0. Hence f(c) > f(x) . Similarly, if x E ]c, c 6[,then by the mean value theorem 3 b E ]c, x[, such that f(x)  f(c) = f(b) (x  c) > 0 since f'(b) > 0 and x  c > 0. f(x) > f(c), or f(c) < f(x). c is a minimum.
1
+
*
E 16)d) Max. at x = I, min. at x = 3. Extreme values are 0 and  28. x = 0 is not an extremum because there is a neighbourhood of 0 in which ((x) has the same sign on either side of 0.
b) Min. at x = 1, x = 3 Max. x =  l Extreme values are  3,285,29
C) Min. at x = 1, max. at x =  1/3, extreme values are 0 and 32/27. E 17)a) x = 0; min. poipt. extreme value = 0 b) There are no extreme points. c) x =. 7/6; min. point; E.V =  37/12 d) f'(x) = xg(x) where g(x) is a polynomial in x2 with all coefficientspositive. Hence g(x) > 0 for all x + 0. Therefore the only extreme point off is x = 0. Clearly, f(0) = ao and f(x) > a,,, x # 0. Hence 0 is a min. and E.V is ao. e) g(x) = and E 18)max. 1
=x
+, 1
X
. Extreme points of g are k 1, 1 being a min.
= f 1/2.
 1, a max. Hence 1 is a max. and  1 is a min. for f. E.V.
Drawing Curves
i
E 19)If a and b are the sides of the inscribed rectangle,
Area = A = ab = ad=' A' =O if a = d / f i A" < 0 for a = d / f i a = d/ f i b =d / f i
.
E 20)'Suppose the display area is a rectangle with sides a cm and b cm. Then the dimensions of the name plate are a 2 cm and b 4 cm. b = 48/a. ab = 48 A = (a 2) (b 4) = (a 2) (48/a 4)
+


* a = d / d is a max.
3 the
rectangle is a square.
+
+
+
+
+
> 0 if a = 2
Dimensions of the plate: 2(1
6
 ' This .
is a minimum, )
+ &),
4(1
+6
E 21)a) Cocvex in 10, [; concave in 1 m, O[; point of inflection (P.1) =(O, 0).
b) Convex in 1 , O[; concave in 10, [ : P.I. : 0.0) c) Convex in
1 ,

1[ U 12, [; concave in
1
1,2[, P.I.:  1, (1, 12). ( 2 . 4 7 )
d) Convex if x > 3; concave if x < 3; no P.I. e) Concave everywhere; no P.I.
f)
Convex in ] ~ / 2 3 ~ / 2 [ concave in 10, ~ / 2 [U ] 3 ~ / 2 , 2 ~P.I.;(1t/2,0) and , ; [; (31t12.0) . 2 (1 +4x2)312
 sin x
E 22)a) 0,
b,
c)
(1
+C O S ~ X ) ~ ~ ~
UNIT 8 GEOMETRICAL PROPERTIES OF CURVES
Structure
8.1 Introduction
Objectives
8.2 Equations of ~anaents Normals and 8 ; f A ~ g l o Intersection of Two Curves ef 8.4 Singular Points
Tangents a the Origin t Classifying Singular Points
8.5 Asymptotes
Asymptotes Parallel to the Axes Oblique Asymptotes
8.6 Summary 8.7 Solutions and Answers
8.1 INTRODUCTION
We started our study of Calculus by stating two problems. One of them was the problem of finding a tangent to a curve at a given point. In Unit 3 we have seen that the solution of this problem was instrumental in the development of differential calculus. Now having studied various techniques of differentiation, we shall once again take up this problem. The study of the tangents of a curve will then lead us to normals and asymptotes of curves, which we shall study i,n Sec.2 and Sec.5, respectively. In the last unit we discussed some other geomettic features of functions, like maxima, minima, points of inflection and curvature. You will see that all these will prove very useful when we tackle curve tracing in the next unit.
Objectives
After studying this unit you should be able to obtain the equations of the tangent and the normal to a given curve at a given point calculate the angle of intersection of two curves at a given point of intersection obtain the angle between the radius vector and the tangent at a point o p a given curve 6 ' define and identify a singular point, a node and a cusp define asymptotes and obtain their equations.
8.2
EQUATIONS OF TANGENTS AND NORMALS
In Unit 3 we have seen how a tangent can be defined precisely with the help of derivatives. We have noted that the slope of a tangent to the curve y = f(x) at (XO, is yo) whenever it exists. In fact, we had also obtained the equations of the given by ~ ( x o ) , tangents of some dimple curves. Once we know how to find the equation of a tangent, it is yo) easy to find one for a normal too. A normal to a 'curve, y = f(x) at (XO, is a line which passes through (xo, yo) and is perpendicular to the tangent at that point. This means that the slope of this normal will be
A line LI.is perpendicular to a line Lz iff mlmz =  I , where m~ and mz are the slopes of LIand L2, respective1y.
Now, what happens when f(xo) = O? f(x6) = 0 implies that the slope of the tangent at (xo, yo) is zero, that is, this tangent is parallel to the xaxis. In this case, the normal (which is perpendicular to the tangent) would be parallel to the yaxis. The equation of this normal, would then be x = XO. Now let us look at various curves and try to ob&n the equations of their tangents and normals.
Drawing Curves
a = 0 gives us the trivial case when the curve is the line y = 0, or the xaxis.
Example 1 Consider the curve y = 2 6 a # 0, shown in Fig. 1: dy 2a Here,  = a = . Thus, exists and is nonzero for all y # 0 Now y will . Y dx dx be zero only if x is zero. Thus, we can find the equations of tangents and normals to this curve at any point, except the origin (0, 0). We know that the slope of the tangent at any point (xo, yo) will be 2a/yo. The slope of the normal will, therefore, be  y0/2a. Thus, the equation of the tangent at (xo, yo) is 2a  xo) y  yo = (x
dk
Recall that the equation of a line through (xo, yo) having a slope m is y  yo = m(x  xn) .
* yyo  yo2 = 2ax  2axo
+yyo = 2ax +
yyo = 2a(x XO), since yo2 = 4 axo. The equation of the normal at (XO, is yo) Yyo=
Yo
*
+
 2axo
2a (x  xo)
Now let us see an example where the equation of the curve is given in the parametric form. In Unit 4 we have already seen what a parameter is.
Example 2 To find the equations of the tangent and the normal at the point 8 = 7r/4 to (see the curve given by x = a cos38, y = a sin3@, Fig. 2), we first note that e dy dy/dO  3a sin28eos8  tan8   =dx/d8 dx  3a cos28sin8

I
Fig. 1
Hence, the slope of the tangent at 8 = 7r/4 is  tan 7r/4 =  1. The slope of the normal at this point thus comes out to be .I. Now, if 8 = 7r/4, cos8 = 1 / 0 and sin Thus, x = a / 2 6 and y = a / 2 f i The equation of the tangent at (a/2\fi; a / 2 f i ) a   l(x a

is
27%
That is, x
a + y =
6
or 6(x
+ y) = a
The equation of the normal at ( a / 2 n , a / 2 6 ) is
Example 3 illustrates the method of finding the equations of tangents.and normals when the equation of the curve is given in the implicit form.
I
Fig. 2
Example 3 Let us find the equations of the tangent and the normal to the curve defined by x3 y3  6xy = 0 at a point (XO, on it. yo)
+
Fig. 3 shows this curve.
Fig. 3
.
In Unit 4 we have seen how to calculate the derivative when the relation between x and y is expressed implicitly. We shall follow the same procedure again. Differentiating the given equation throughout with respect to x, we get dy dy 3x2 3y2 6y  6x = 0, which means dx dx
Geometrical Properties of Cur
+
yo  2x0 Hence, the e q u a t i ~ the tangent at (xo, yo) is of .. 2yo 2 xo yyo= (x  xo) yo2  2x0 Simplifying, and using the relation xo3 yo3 = 6x0~0, reduces to this
Thus, the slope at the point (XO, is yo)
2yo  xoL
+
(2yo  xo2)x (2x0  yo2)y 2xoyo = 0 Now the normal at (XO, is a line passing through (XO, and having slope yo) yo)  (yo2  2x0) . Hence, the equation of the normal at (XO, is yo) , 2yo  xo
+
+
Simplifying, we get (yo2  2xo)x (2yo  x o 2 ) ~ (xo  yo) (2x0 xoyo 2yo) = 0 If you have followed these examples, you should have no problem in solving the following exercises.
+
+
+
+
E
E 1) Find the equations of the tangent and the normal to each of the following at the
specified point.
b) x = a cos t, y = b sin t at the point given by t = 7r/4 c) x2
+
y2
= 25 at ( 3,4).
Drawing Curves
Vertical Tangents
By now, you are quite familiar with the fact that ((x) or dy/dx may not exist at some points. At such points either the tangent does not exist, or else, is parallel to the yaxis: that is, vertical. To examine the existence of vertical tangents at (xo, yo), we examine dx / = 0, then, we conclude that there is a vertical taqgent at
,
(XO, In such cases the equation of the tangent can be written as x = XO. yo). The normal corresponding to a vertical tangent will obviously be horizontal or parallel to the xaxis. This means we can write its equation as y = yo, as its passes through (xo, yo). dy If you take the curve in Example 2, you will find that does dx not exist whea 8 = n/2.
dx dx Let us examine at this point.  =  cot 8 = 0 if 8 = ~ / 2 . dy dy This means that the curve has a vertical tangent and, consequently, a horizontal normal at , this point. Now, when 8 = ~ / 2 x = 0 and y = a. Thus the equation of the tangent at (0, a) is x = 0 and that of the normal is y = a. See if you can solve this exercise now. E 2) Are there any points on the following curves where the tangent is parallel to either axis? If yes, find all such points.
b) y = sin x
Let us now look at another example.
Example 4 To find the equations of those tangents to the curve y = x3, which are parallel to the line 12x y  3 = 0, we first observe that the'slope of the line 12x  y  3 = 0 is 12. Thus, the slope of any line parallel to this :;fie should also be 12. Now, the slope of the tangent to the curve y = x3 at any point (x, y) is ((x) = 3x2.

If we equate8f'(x) to 12, we will get those points on the curve where the tangent is parallel to 12x  y  3 = 0. Thus, 3x2 = 12, or x2 = 4, that is, x = f 2. Ifx=2,y=x3 =8.Ifx=2,y=x3=2. Thus, the points in question are (2, 8) and ( 2,  8). The equations of the tangents at these points are y  8 = 12(x  2) and y 8 = 12(x 2), respectively.
Geometrical Properties of Curves
+
+
The following exercises will give you some more practice in applying the concepts learned in this section.
E
I
E 3) Find the equations of the tangent and the normal to each of the following curves at the point 't' :
b) x = a(t
I
+ sin t), y = a(l  cos t)
E
\
E 4) Find the equation of the tangent to each of themowing curves at the point (xO,yo).
a) x Z  k Y 2 + 4 x + 6 y  1 = O
b) x y = a
 
Drawing Curves
E.
E 5) Prove that the line 2x xcoordinate is zero.
+ 3y = 1 touches the curve 3y = e'ht a poinh whose
E
. E 6)
Prove that the equation of the normal to the hyperbola x 2    I at a point ( a f i b) is ax b f l y = (a2 Y' a2 b2
+
+ b2) fl
Geometrical Properties of Curves
8.3 ANGLE OF INTERSECTION OF TWO CURVES
The concept of a tangent to a curve has proved very useful in describing various geometrical features of the curve. In this section we shall look at one such feature. When two curves intersect at a point, their angle of intersection at that point can be defined with the help of their tangents there. In fact, we say that if two curves intersect at a point P, the angle of intersection ofthese two curves at P is an angle between the tangents to these curves at P, such that 0 5 I9 5 7r/2 (see Fig. 4).
I
Fig. 4
We now prove a theorem which gives us the angle of intersection at a point when the equations of the two curves are known. Theorem 1 If two curves y = f(x) and y = g(x) intersect at a point P(xI, y ~ )then the , angle I9 of intersection of these curves at P(xI,y ~ is given by ) tan I9 = I
+ f'(x,)g'(x,)
Proof From Fig. 4, tan I9 = tan (P @) ! tan @  tan V 1 tan T tan @
+

Drawing Curves
Fig. 4 shows V'  to be an acute angle. But if the curves f and g were as in Fig. 5, then angle 8 = rr  ( V  a ) , since wk. take the acute angle as the angle of intersection.
Fig. 5
In this case, tan 0 = tan [T  ( V  a ) ] =  tan ( V  @) But we are not in a position to decide whether we should take tan0 as tan ( V  @) or as  tan ( y  @), unless we have drawn the curves. Since it would be tedious to first draw ' the curves and then decide, we think of an alternate scheme. We observe that since 8 lies I between 0 and rr/2, tan 8 is nonnegative. Thus, we take tan 0 to be Itan ( V  @)I. Hence, tan 8 = Having proved this theorem, we can easily deduce the following corollaries.
Corollary 1Two curves y = f(x) and y = g(x) touch each other at (XI, that is, have 1 yl), common tangents at (XI, iff 0 = 0, that is, iff yl), ((XI) = g'(x1). Corollary 2 Two curves cut each other at right angles, or orthogonally, at (XI,y ~ iff )
f'(x1) g'(x,) =  1. If you study Example 5 carefully, you will have no difficulty in solving the exercises later.
Example 5 Let us find the angle of intersection af the parabola y2 = 2x and the circle
x2
+
y2
= 8.
First we find the paints of intersection of these curves, if there are any. The coordinates of these paints will satisfy both the equation to the parabola and the equation to the circle. So substituting y2 = 2x in x2 y2 = 8, we get x2 2x = 8, or x =  4 or 2.
+
+
It is clear from Y2 = 2x that the abscissa x (= y2/2) of any common point must be nonnegative. So we reject the value  4 of x. When x = 2, y = f 2. Hence the common points are P(2,2) and Q(2,  2). Since both curves are symmetric about the xaxis (see Fig. 6) and since P and Q are reflections of each other w.r.t. the xaxis, it is sufficient to find the angle at one point, say P; the angle at Q being equal to the angle at P. Differentiating the two equations w.r. to x, we get dy dy 2y = 2 and 2x 2y  = 0 dx dx
+
Geometripal Properties of Curves
Fig. 6
Hence the values of <x and g'(x), that is, the slopes of the tangents to the two curves at () (x,y) are l/y and  x / y . Hence the slopes of the tangents at (2,2) to the two curves are 1/2 and  1. Hence if 8 is the required angle, then
You can try these exercises now.
E 7) Find the angle of intersection of the parabolas y2 = 4x and x2 = 4y.
E 8) Show that
a) the ellipse x2 4y2 = 8 and the hyperbola x2  2Y2= 4 cut each other . orthogonally (at right angles) at four points.
b) the curves xy = a2 and x2 tangent) at two points.
+
+
Y2
= 2a2 touch each other (have a common
Drawing Curvea
You know that given a pair of axes in a plane, the position of a point in the plane can be determined if we know its distances frbm the xand yaxes. There is one more way in which we can determine its position. Suppose we are given an initial line O X in a plane (see Fig. 7(a)). Then a point P can be located if we know i) r, its distance from 0 and ,
'
ii) 8, the angle made by OP with OX.
(b)
Fig. 7
r, and 8 are called the polar coordinates of P.r is always taken to be nonnegative, and 8 takes values between 0 and 27r. From Fig. 7(b) it is clear that if x and y are the artesian

9
1
.
1

t?; '' . .;I.The equation of a curve is sometimes expressed in polar coordinates by an equation r = f(8). For example, the equation of a circle with centreoand radius r is r = a. Now let us turn once again to the problem of finding the angle of intersection of two curves. The method that we have been following till now, cannot be used if the equation of the curve is given in the polar form. In this case we follow a somewhat indirect method.
Geometrical Properties of CU
Fig. 8
Take a look at Fig. 8. It shows a curve whose equation is given in the polar form as r = f(8). P(r,B) and Q(r 6r, 8 60) are two points on this curve. PT is the tangent at P and OPR is the line through the origin and the point P. We shall now try to find 8, the angle between PT and OR.
+
+
We note here; that the tangent PT is the limiting position of the secant PQ. If we denote the angle between PQ and OR by a , then we can similarly say that 4 is the limit of a as QP along the curve. Now from AOPQ we have OQ sin LOPQ = OP sinLOQP r 6r.  sin (n  a) or r sin ( a  68) 6r sin (a  a)

Remember the sine rule for a A ABC? sin A sin B sin C
A

B

C
+
6r or = r or
7
sin a  sin ( a  68) (since sin ( a a) lr sin a) sin ( a  68)
'
sin A

sin B
1 6r r 68
=
2 cos(a  68/2).sin (60/2) sin ( a  68).68 sin (68/2) 68/2
= 2 Sin
(?)A  B
COS
tT)
A+B
 2 cos(a  68/2) sin ( a  68)
As QP, a4.680 and 6r0. Hence as QP, we get t d   cos C#I = cot 4 r d8 sin4
stn ( S 8 / 2 )
Recall lim 660

6,912

I
de so that tan 4 = r. dr
This formula helps us to find the angle between O P and the tangent at the point P on the
Drawing Curies
We shall use this r.esult to find the angle between two curves C I and C2 which intersect at P (say). If the angles between O P and the tangents to C I and CZat P are 41and 42, respectively, the angle of intersection of C I and C2 will be 141 4 2 1 (see Fig. 9).
Fig. 9
This can be easily calcu1ate.d as we know tan Thus, tan 141  4 2 1 = tan 41 tan 4 2 1 k tan 41tan $2
&I
and tan 4 2 .
Further, if the curves intersect orthogonally, tan 41.tan 4 =  1. The following examples 2 will help clarify this discussion.
~ x a m p l e Suppose we want to find the angle of intersection of the curves r = a sin 28 6 and r = a cos 28 at the point P ( a / a rr/8). The coordinates of P satisfy both the equations r = a sin 28 and r = a cos 28.
If 4, is the angle between OP and the tangent to r = a sin 28,then d8 a sin 28  a sin 28 1 1   tan 28 = tan 41= r = dr dr/d8 2a cos 28 2 2 Similarly, if $iis the angle between O P and the tangent to r = a cos 28, then
Thus, tan (41  $2) = Thus, $r
tan 41 tan 4 2  1/2 1/2 1 tan $I tan 4 2 1  1/4
+
+
4 
3
 4 2 = tan'
(4/3) =53.13", which is the required angle.
NOWtry to do a few exercises on your own.
E
E 9) Find the angle between the line joining a point P(r,8) on the curve to the origin and the tangent for each of the following curves. F
a) r2 = a2 cos 28 C) rm = amcos m8 b) I/r = 1
+ ecos8
b) rm = am(c,s mf3  sin me)
E
E 10)Check whether the following two curves intersect orthogonally.
and a) r = aeO re0 = b b) r = a(l
+ sin 8) andr = a(l  sin 8)
Drawing Curves
8.4
SINGULAR POINTS
In this section we shall study a category of points on curves, called singular points. But to properly classify singular points we have to find the nature of tangents at these points. So let us first study on easy method of finding the tangents to a curve at the origin. This knowledge will then help us tofind the tangents at any point of the curve easily.
8.4.1 Tangents at the Origin
We shall now give you a simple method of finding the tangent to a curve at the origin, when the equation of the curve is given by a polynomial equation. That is, the equation is of the form f (x,y) = 0 where f is a polynomial in x and y. You will agree that the constant term in this polynomial is zero since the curve passes through the origin. For such a curve the equation of the tangent at the origin can be found out by equating to zero the lowest degree terms in x and y (we shall not prove this here). Thus, if x3 3xy 2x y = 0 is the equation of a curve, the equation of the tangent at the origin is 2x y = 0.
+
+
+ +
Similarly, if the equation of a curve is x4 x2  y2 = 0, then the equation of the tangent to this curve at (0,O) is x2  y2 = 0, or x2 = y2,or x = f y. Hence we get two equations x = y and x =  y. This means that the curve X ' + X ~  ~ ' = 0 has two tangents at the origin. We shall consider such eventualities in the next subsection.
When the origin is shifted to (h, k) the coordinates a point P(x, y) in the new coordinate system are given by x'=xh
+
Now, consider a curve given by g(x,y) = 0, where g(x, y) is some polynomial in x and y. Suppose we want to find the equation of the tangent to this curve at some point, say (h, k), on it. What we do is, we shift the origin to (h, k). Then with respect to this new origin, the equation of the curve will be g ( i h, y k) = 0. We can also express it by G(x ,y ) = 0, where G(;, y') = g(x h, y k).
+
+
+
+
Note that this change of origin does not change the shape of the curve. Now the tangent to the curve G(x,y) = 0 at the origin will be the tangent to the curve g(x,y) = 0 at the point (h, k) (see Fig. 10(a) and (b)).
Fig. LO
(b)
The method for finding the equation of the tangent at any point of a curve will be clear to you when you read our next example.
Example 7 Consider the curve defined by the equation ay2 = x(x equation of the tangent to this curve at the point ( a, 0).
+ a)'. Let us find the
For this, we first shift the origin to ( a, 0). The equation of the curve then becomes ay2 = (x  a) (x  a a12 2 or ay2 = x (x  a).
+
We can also write this as a(x2 y2) = x3
+
Geometrical Properties of Curves
Now, the equation of the tangent to this curve at the origin will be given by a (x2 y2) = o * X ~ + ~ ~ = O *x2=y 2
+
This is impossible, since the write this as x = f iy, x
real number has to be nonnegative. But we can is an imaginary number.
Thus the equations of the tangents to the given curve at the point ( a, 0) are
+ a = f iy (shifting back the origin).
In such cases we say that the curve has imaginary tangents at the point (  a, 0) . Now that you have seen how to find the tangents to curves given by polynomial equations, let us try and categorise the points on a given curve with the help of the tangents at those points. Applying the procedure used in the Example 7, you should be able to solve this exercise. E 11)Find the equations of the tangents at the origin to each of the following curves a) 16y2=x2(16x2) b) (y  x212= x4 c) x3
+ 6xZy 8y2 = 0.
+ 3x3.
8.4.2 Classifying Singular Points
An equation of the type y = f(x) determines a unique value of y for a given value of x. Ti means, every straight line parallel to the yaxis meets the curve y = f(x) in a unique hs point. However the equation of a curve is often given as f(x, y) = 0. If f(x, y) is not a linear expression in y, then it may uot be possible to write f(x, y) = 0 in the form y = F(x) uniquely. For example, if f(x,y) = y7  x', then f(x, y) = 0 gives
This gives us two relations y = d
m and y =  d
nof the type y = F(x).
Drawing Curves
The curve has 2 branches, as you can see from Fig. 11.
4
The origin is common to the t%u branches. Put, differently we can say that two branches of the circle x2 + v* = a' pass through polnts A and B. We have a generic name, A singular points, for points like 0. precise definition is as follows.
Definition 1 If k branches of a curve pass through a point P on the curve f(x, y) = 0 and k > 1, then P is said to be a singular point or a multiple point of order k.
Singular points of order two are known as double points. Thus the points A and B in Fig. 1l(c) are double points. Obviously, a curve will have more than one taagent at a singular point (one corresponding to each branch). Depending upon whether tangents at double points are distinct, coincident or imaginary, we shall give special names to such points.
Definition 2 A double point is known as
i) a node if the two tangents at that point are real and distinct,
ii) a cusp if the two tangents are real and coincident, iii) a cocjugate (or isolated) point if the two tangents are imaginary. In Fig. 12 we show an example of each. For the curve f(x; y) = 0, the origin is a node. while the point Q on the For the curve g(x,y) = 0, the points PI, P2, P3 and Pq are CUSPS, curve h(x, y) = 0 is a conjugate point.
(b)
Fig. 12
I
I
I
I
In Example 7 we have seen one more example of a conjugate point. See if you can solve this exercise now.
Geometrical properties of Curves
I
E
E 12)Show that ( 1,  2) is a singular point on each of the following curves. a) x3 b)
1
!
+ 2x2 + 2xy  y2 + 5x  2y = 0 x3 + xy2 + 2(x2 + Y2)4 4xy 5x + 8y + 8 = 0.
I
Also deteimine the nature (cusp, node etc.) of this singular point in both cases. (Hint : Shift the origin to ( 1,  2) and check the tangents at the new origin.
8.5
ASYMPTOTES
In this section we shall study another feature of curves which will prove very useft1 in tracing curves as you will see in the next unit. This involves taking limits as x =  for y* .
Yuu fnmefnmecome such limits in Unit 2. Let us define an asymptote now. across
Definition 3 A straight line is said to be an asymptote to a curve, if as a point P moves to infinity along the curve, the per#ndicular distance of P from the straight line tends to zero.
Example 8 Consider the rectangular hyperbola xy = c shown in Fig. 13. xy = c implies y =C/X and this implies that as x or  , y  0. Now ly? is the distance of a point
DIP
..\.
rr
thm
h.mar)v\lm
frr.m thn

.
vav;a
Cr.
7, 1
con
*
naf na v

s nr ,
m
thp
Drawing Curves
distance of a point, P(x, y) on the hyperbola from the xaxis approaches zero. In other words, this means that the xaxis is an asymptote of the hyperbola.
4
Y
Fig. 13
Writing xy = c as x =c/y, and repeating the arguments exactly as above, we can prove that the yaxis is also an asymptote of the hyperbola. Example 9 Let us prove that the xaxis is an asymptote of the curve y =10 shown l+x
3 2 1
,
,
1 2 3
,
.
,
X
.
in Fig. 14. From the equation of the curve. it is quite clear that y  0 as x  = or  =. Again, this means that the distance of the point P(x, y) on the curve from the xaxis tends to zero as x  = or 00. This proves that the xaxis is an asymptote of the curve.
I
I
Fig. 14
8.5.1 Asymptotes Parallel to the Axes
Here we shall derive tests to decide whether a given curve has asymptotes parallel to the x and y axes. For this we shall consider a curve given by f(x, y) = 0, where f(x, y) is a polynomial in x and y. Theorem 2 A straight line y = c is an asymptote of a curve f(x, y) = 0 iff y  c is a factor of the coefficient of the highest power of x in f(x; y). Proof Arrange f(x, y) in descending powers of x so that the equation of the curve is written as go(y)xn g,(y)xn' ..... gn(y) = 0
+
+ +
P(x,y) tends to infinity means
at'east One of infinity.
and y must lend to
76
The perpendicular distance PM of P(x, y) from the line y = c is 1 y c 1. (Check this by drawing a suitable figure). Now according to Definition 3, y = c is an asymptote iff PM tends to zero as P tends to infinity, that is iff y  c as P tends to infinity. If the y coordinate y of P  c (a finite number) as P tends t o infinity, then its xcoordinate x must tend to infinity. Now since P $ q point on ihe curve, its coordinates satisfy the equatio$ 0 , Y)= 0 4"

So, as P tends to infinity along the curve, we get lim f(x, y) = 0 ='D From this we can sfly that y = c is an asymptote iff lim f(x, y) =0
X W '
Geometrical Properties of Curves
Hence, y s c is an asymptote of (1) iff 1 1 ] lim [go(y)+gl(y)y+ .....+g " ( y ) ~= O
I
I
I
X= C Y '
t
go(c) y
I C
 c is a factor of gb(y), the coefficient of the highest power of x in f(x, y).
=0,
This theorem can also be interpreted as follows. Asymptotes parallel to the xaxis are obtained by equating to zero the real linear factors of the coefficient of the highest power of x in the equation of the curve. We can also state a theorem, similar to Theorem 2, giving a test to decide whether a given curve has an asymptote parallel to the yaxis or not. Theorem 3 Asymptotes parallel to the yaxis are obtained by equating to zero the real linear factors ax b of the coefficient of the highest power of y in the equation of the curve.
+
Proof : Similar to that of Theorm 2 Example 10 Let d n d the asymptotes parallel to either axis for the curve y = x
1 +;
Writing the given equation in the form f(x, y) = 0, we have x2  xy 1 = 0. You can see the grabh of this curve in Fig. 15. The coefficient of the highest power of x is 1. It has no factors of the form y  c. Hence there are no asymptotes parallel to the xaxis. The coefficient of the highest power of y when equated to zero gives x = 0. Hence there is one asymptote parallel to the yaxis and moreover, it is the yaxis itself.
+
I
Fig. 15
See if you can do these exercises on your own.
E
E 13)For each of the following curves, find asymptotes parallel to either axis, if there are
any, a) xXy= 2 c) (x
+
+y
= x2
b) xy2 = 16x2
Y)2
+4
d)
+ 20y2 xZy2= 9(x2 + y2)
8.5.2
Oblique Asymptotes
You may be wondering whether an asymptote must always be parallel to a coordinate axis. No, there are many curves having asymptotes which are not parallel to either axis. Such asymptotes are generally referred to as oblique asymptotes. We shall now learn how to find oblique asymptotes y = mx c to rational aigebraic curves f(x, y) = 0. The problem is to determine m and c so that y = mx c may be an asymptnte to f(x, y) = 0.
+
+
The perpendicular distance of a point P(x1, yj) from the line ax+by+c=Ois
Suppose that the line y = mx c is an oblique asymptote to the curve f(x, y) = 0. This means that m = 0. The perpendicular distance PM of a point P(x, y) on this curve from ly  mx  c J Now, since m f 0, P can be at infinity y = mx c is given by PM =
+
+
dG=T
on the curve only when x (as also y), tends to OQ. h u sas XOQ, ~ , as X00, (y  mx  c) 0..

PM0. This means that
or, lim (y '  c) = 0 mx x00 That is, c = iim (y  mx). x00 Thus c would be known as soon as m is known. Now, (Y  mx) Y lirn (  m) = lim xOQ X xOQ x 1 = lim (y  mx) lim () x00 x00, = c.0 = 0, using (1) Hence m = lim (Y ) x00 X
,' a
.... (1)
4
Thus, given any curve f(x, y) = 0, we first find lim x4= calculate c = lim (y  mx) . x00 The following example will clarify this procedure.
Y
= m and then use this m to
Example 11 Let us examine the curve x3  y3 = 3xy for oblique asymptotes.
Suppose that the given curve has an oblique wymptote y = mx curve can be written as
"3
+ c. The equation of the
 .,) QY..  n
Dividing throughout by x3 we get
I
I
Geometrical Properties of Curves
Thus, lirn x*
[ l  ;J ;
x
1
* 1  lim x00
I
(Y
x3
) = 0,since lim
x00
1 = 0
x
I
* m3 = 1 * m = 1, the other roots of m3  1 = 0 being complex numbers. Rewriting the equation of the curve as (x  y) (x2 + xy + y2) = 3xy, we have  3xy c = lim (y  x) = lim x00 [ x2 + xy + y2 x*
I
=
Iim x 00

3
1+1+1
, since lirn
x*
X  =  lirn Y
 1
Hence the required asymptote is y = x
I
 1.
Try to solve these exercises now.
E
E 14)Find Oblique asymptotes to each of the following curves.
Drawing Curves


8.6 SUMMARY
In this unit we have covered the following points. 1) The equation of the tangent at (xo, yo) to the curve y = f(x) is y  yo = f'(x0) (x  xo) dx yo) 2) The curve has a vertical tangent at (XO, if  . = 0 at this point. dy 3) The angle 8 of intersection of two curves v = f(x), y = g(x) is the acute angle between the tangents at that point to the curves. It is given by the relation f'(x)  g'(x) 1 tan 0 = 1 + f'(x) g'(x) ( 4) y = f(x) and y = g(x) cut each other orthogonally at (XO, if ff(xo) i(xo) =  1. yo)
5) The angle c between the tangent and the radius vector of the curve r = f(H) at the b de point 8 is given by tan cb = r
clr
6 ) The tangents at the origin to any curve (which passes through the origin) are obtained by equating to zero the lowest degree terms in the equation of the curve.
7) If k branches of a curve pass through a point P on the curve f(x, y) = 0 and k > I , then P is said to be a singular point or a multiple point cf order k. Singular points of order two are known as double points. A double point is known as a node, a cusp or a conjugate (isolated) point according as the two tangents at t h a ~ point are real and distinct, real but coincident, or imaginary.
8) A straight line is said to be an asymptote to an infinite branch of a curve, if, as a point P on the curve moves to infinity along the curve, the perpendicular distance of P from * the straight line tends to zero.
9) Asymptotes parallel to the coordinate axes are obtained by equating ta zero the real linear factors in the coefficients of the highest power of x and the highest power of y in the equation of the curve.
10) If y = mx
+ c is an oblique asymptote of the curve f(x, y) = 0, lim x00
Y  = m and X
lim (y  mx) = c x00
8.7 SOLUTIONS AND ANSWERS
= 4. Equation of the tangent at (1, 4) is E l ) a) dx dy . = I (y  4) = 4(x  1) Slope of the normal at (1,4) =  114 Equation of the normal at (1, 4) is (y  4) = ( 1/4) (x  1).
I
b) Slope of the tangent   b/a Slope of the normal a/b at t = ~ / 4 x, = a / f i y = b.!\/2 Equation of the tangent : ,.. (y b / \ g ) =  b/a ( X  a/\/ 2 ) c) Slope of the tangent = 3/4 Slope c>f ?he ~ztlrma! r  4/3

Equation of the tangent : y  4 = (3/4) (x 3) Equation of the normal : y  4 =( 4/3) ( x 3)
+ +
Ceomelrical Properties of Curves
E 2) a) Tangents are parallel to the xaxis at x = (1 f n ) / 3
b) Tangenfs are parallel to the xaxis at all points where x = n s integer n. There are no tangents parallel to the yaxis.
+ s / 2 for some
E 3) a) Tangent : ty = x at2 Normal : y tx = at(2
b)
+ + + t2) Tangent : (1 + cos t)y = sin t(x  at). Equivaltr:!ly, sin (t/2)x  cos (t/2)y = at sin (1/2) Normal : sin (t/2)y + cos (t/2)x = 2a sin (t/2) + at cos (tJ12)
E 5 ) 3y = e2'
dy dx at (0, 1/3) is given by
1
.=o
 2 . (0, 1/3) is.a point on this curve. The tangent 3

Slope of the normal =  a / b o Equation of the normal is y  b =
*
E 7)
Y2
= 4x
Y4
= x 3
= y2/4
* x2 = y4/16 = 4y at the point of hlersection.
* * *
 64y = 0
' 3 Y(Y3 
64) = 0 *y(y  4)(y2 4y 16) = O y = 0 , 4 (other roots are complex) x =Oor4.
+ +
Slope of the tangent to y2 = 4x at (4,4) = 1/2 Slope of the tangent to x2 = 4y at (4,4) = 2 angle of intersection = tan' (3/4) The tangent at (0,O) to y2 = 4x is vertical, and the tangent at (0,O) to x2 = 4y is horizontal. Hence the angle of intersection at (0,O) is s/2.
*
E 8)
a)
The four points are ( 4 / a f 5 3 , ( 4 / f i
f
6
dy for x2  2y2= 4 is x/2y dx
J3
: .
They cut orthogonally.
1
Drawing Curves
* angle = tan' (r )d8 dr
= tan' (tan (
(24 2
= tan'
 r2
+ 1) n + 28)

b) tan' (
1
+ e cos 8
sin 8
)
= tan 3
tan 4 =  1 2
=3
the curves cut orthogonally.
b) The curves cut orthogonally.
E 1l)a) 16y2 = 16x2 x4 The equation of the tangent is 16y2 16x2= 0 =3 y2  x2 = 0
b) y2  2x2y x4 = x4 3x3. The equation of the tangent is y2 = 0 c) . Equation : y2 =O or y = 0.
= y 3
= f x.
= 0.
+
+
= y 3
E 12)a) Change the origin to ( 1,  2).
Then the equation of the curve is (X 113 2 (  i)2 2 (  I ) ( Y 2)  (Y 2)2 S(X 1) ~ ~  2(y  2) = 0. x3  x2 2xy  y2 = 0 The equation of the tangents at the origin is x2  2xy y2 = 0 (x  y)2 =o:x = y. There are two real and coincident tangents at this point. Hence it is a cusp.

+
+
+
+
+
b) After shifting the origin we get the equation : (X 1)3 (X I ) ( Y 212 2 [ ( ~ 112 (y  212] 4 (  I ) ( Y 2) ~ +5(x l ) + 8 ( y  2 ) + 8 = 0 . The equation of the tangents at the origin is y2  x2 = 0, that is, C3 y2 = x2 OF y = X. There are two real and distinct tangents at this point. Hence it is a node.
+
+
+
+
*
E 13)a) x'y = 2
y a x2y  y  2 = 0 Highest power of x is 2. The coefficient of x2is y. Hence y = 0 is an asymptote. Highest power of y is 1. The coefficient of y is x2  1 = (x  1) (x 1). Hence x =  1 and x = 1 are two asymptotes.
+
+
b) No asymptotes parallel to the xaxis. x = 20 is an asymptote. c) No asymptotes parallel to the yaxis. y = 0 is an asymptote.
d) y = f 3 are asymptotes. x = f 3 are asymptotes.
Ceomciricml Properties of Curves
e) y = 0 is an asymptote. f) y = 0 is an asymptote.
* 1 + lim
XQ O ' XQ O '
( ; 
Y
= lim
XQ O '
3a y
*1+m3=0 *m3=l'*m_l. c = lim (y  mx) = lim (y x)
XQ O '
+
= lim
X '
OQ
3ax2 x2  xy
+ y2
= lim,
XQ O '
3a 1  y/x
+(y/~)2

Hence the equation of the asymptote is y
+ x = a.
UNIT 9 CURVE TRACING
Structure
9.1 Introduction
Objectives
9.2 Graphing a Function and Curve Tracing 9.3 Tracing a Curve : Cartesian Equation 9.4 Tracing a Curve : Parametric Equation 9.5 'Tracing a Curve : Polar Equation 9.6 Summary 9.7 Solutions and Answers
9.1 INTRODUCTION
A picture is worth a thousand words. A curve which is the visual image of a functional relation gives us a whole lot of information about the relation. Of course, we can also obtain this information by analysing the equation which defines the functional relation. But studying the associated curve is often easier and quicker. In addition to this, a curve which represents a relation between two quantities also helps us to easily find the value of one quantity corresponding to a specific value of the other. In this unit we shall try to understand what is meant by the picture or the graph of a relation like f(x, y) = 0, and how to draw it. We shall be using many results from the earlier units here. With this unit we come to the end of Block 2, in which we have studied various geometrical features of functional relations with the help of differential calculus.
Objectives
After studying this unit you should be able to list the properties which can be used for tracing a curve trace some simple curves whose equations are given in Cartesian, parametric or polar forms.
9.2 GRAPHING A FUNCTION AND CURVE TRACING
Recall that by the graph of a function f : D  R we mean the set of points {(x, f(x)) :x ED). Similarly, the set of points {(x, y) : f(x, y) = 0] is known as the graph of the functional relation f(x, y) = 0.Graphing a function or a functional relation means showing the points of the corresponding set in a plane. Thus, essentially curve tracing means plotting the points which satisfy a given relation. However, there are some difficulties involved in this. Let's see what these are and how to overcome them. It is often not possible to plot all the points on a curve. The standard technique is to plot some suitable points and to get a general idea of the shape of the curve by considering tangents, asymptotes, singular points, extreme points, inflection points, concavity, monotonicity, periodicity etc. Then we draw a free hand curve as nearly satisfying the various properties as is possible. The curves or graphs that we draw have a limitation. If the range of values of either (or both) variable is not finite, then it is not possible to draw the complete graph. In such cases the graph is not only approximate, but is also inconiplete. For example, consider the simplest curve, a straight line. Suppose we want to draw the graph o f f : R R such that f(x) = c. We know that this is a line paraiiel to the xaxis. But it is not possible to draw a

complete graph as thls line extends infinitely on both sides. W e indicate this by arrows at both ends as in Fig. 1.
Curve Tracing
Fig. I
In the next section we shall take up the problem of tracing of curves when the equation is given in the Cartesian form.
9.3 TRACING A CURVE : CARTESIAN EQUATION
Suppose the equation of a curve is f(x, y) = 0.We shall now list some steps which, when taken, will simplify our job of tracing this curve.
1) The first step is to determine the extent of the curve. In other words we try to find a region or regions of the plane which cannot contain any point of the curve. For example, no point on the curve y2 = x, lies in the second or the third quadrant, as the xcoordinate of any point on the curve has to be nonnegative. This means that our curve lies entirely in the first and the fourth quadrants.
A curve is symmetr~cal b t ~ ~ . t a line if, when we fold the culve cn the lice, the two portions of the ,o,nclde.
A point to note here is that it is easier to determine the extent of a curve if iis equation can be written explicitly as y = f(x) or x = f(y).
2) The second step is to find out if the curve is symmetrical about any line, or about the
origin. We have already discussed symmetry of curves in Unit 1. Fig. 2, shows you some examples of symmetric curves.
.4 curve is symmetrical about the origin if we get the same curve after rotating it through 180°
(a) Symmetric ?bout the xaxis. Fig. 2 :
(b) Symmetric about the origin.
(c) Symmetric about the line y = x.
Here we give you some'hink to help you determine the symmetry of a curve. a) If all the powers of x occurring in f(x, y) = 0 are even, then f(x, y) = f( x, y) and the curve is symmetrical about the yaxis.
Drawing Curves
In this case we need to draw the portion of the graph on only one side of the yaxis. Then we can take its reflection in the yaxis to get the complete graph. We can similarly test the symmetry of a curve about the xaxis.
f( x,  y) = 0, then the curve is symmetrical about the origin. In b) If f(x, y) = 0 such cases, it is enough to draw the part of the graph above the xaxis and rotate it through 180' to get the complete graph.
*
c) If the equation of the curve does not c$ange when we interchar.de x and y, then the curve is symmetrical about the line y = x. Table 1 illustrates the application of these criteria for different curves.
Table 1
Equation
Symmetry
I
I x' +
+ y4 = 0 I x4 + + y2 = 0
y2
y3
I
I
About the xaxis (even powers of y) About the yaxis (even powers of x) About the origin (f( x,  y) = 0 f(x, y) = 0) Aboui both axes (f(x, Y) = f( x, Y),f(x, Y) = f(x,  Y)) About the line y = x (f(x, Y)= f(y, x)) About both axes, (even powers of x and y) but not about y = x. (f(x, Y)# f(y, x))

i
I
3) The next step is to determine the points where the curve intersects the axes. If we put y = 0 in f(x, y) = 0, and solve the resulting equation for x, we get the points of. intersection with the xaxis. Similarly, putting x = 0 and solving the resulting equation for y, we can find the points of intersection with the yaxis.
4) Try to locate the points where the function is discontinuous.
5) Calculate dy/dx. This will help you in locating the portions where the curve is rising (dy/dx > 0) or faifing (dy/dx < 0) or the points where it has a corner (dy/dx does not exist).
6) Calculate d2y/dx2.This will help you in locating maxima (dy/dx = 0, d2y/dx2< 0) 2 and minima (dy/dx = 0, d 2 y / d ~> 0). YOUwill also be able to determine the points of inflection (d2y/dx2= 0). These will give you a good idea about the shape of the curve.
7) The next step is to find the asymptotes, if there are any. They indicate the trend of the branches of the curve extending to infinity.
8) Another important step is to determine the singular points. The shape of the curve at these points is, generally, more complex, as more than one branch of the curve passes through them.
9) Finally, plot as many points as you can, around the points already plotted. Also try to draw tangents to the curve at soqe of these plotted points. For this you will have to calculate the derivative at these points. Now join the plotted points by a smooth curve (except at points of discontinuity). The tangents will guide you in this, as they give you the direction of the curve.
We shall now illustrate this procedure through a number of examples. You will notice, that it may not be necessary to take all the nine steps mentioned above, in each case. We begin by tracing some functions which were introduced in Unit 1. Example 1 Consider the function y = ( x 1. Here y can take only positive values. Thus, the x graph lies above the xaxis. Further, the function y = 1 1is symmetric about the yaxis. On the right of the yaxis, x > 0 and so 1 x I = x. T'nus the graph reduces to that of y = x and you know that this is a straight line equally inclined to the axes (Fig. 3(a) below).
The curve meets the yaxis only at the origin. Taking its reflection in the yaxis, we get the complete graph as shown in Fig. 3(b). We have drawn arrows at the end of the line segment to indicate that the graph extends indefinitely.
Curve Tracing
~ i g 3 : (a) Graph to the right of the yaxis. .
(h) Complete graph.
Example 2 The greatest integer function y = [x] is discontinuous at every integer point. Hence there is a break in the graph at every integer point n. In every interval In, n 1[ its value is constant, namely n. Hence the graph is as shown in Fig. 4. Note that a hollow circle around a point indicates that the point is not included in the graph.
+
Fig. : 4 Graph
tl(x) =
I*]
Example 3 Consider the curve y = x3. Now (x, y) lies on the curve y = x 3 e  y = ( x)'( x,  y) is on the curve. This means that the curve is symmetric about the origin. Thus, it is sufficient to draw the graph above the xaxis and join to it the portion obtained by rotating it through 180°.
Above the xaxis, y is positive. Hence x =fi be positive. Thus, there is no must portion of the graph in the second quadrant. The curve meets the axes of coordinates only at the origin and the tangent there, is the xaxis. dy = 3x2 which is always nonnegative. This means that as x increases, so does y. Thus dx the graph keeps on rising.
1
This implies that there are no extreme points, and that (0,O) is a point of inflection. The Y graph has no asymptotes parallel to the axes. Further lim = lim x2 and X ' x

Drawing Curves
obviously, this does not exist. This means that the curve does not have any oblique asymptotes. You can also verify that it has no singular points. The graph is shown in Fig. 5.
Fig. 5 : Graph of y =x3
1 Example 4 Consider y =  . The ycoordinates of any point on the curve cannot be x2 negative. So the curve must be above the xaxis. The curve is also symmetric about the yaxis. Hence we shall draw the graph to the right of the yaxis first. The curve does not intersect the axes of coordinates at all. 2 d2y 6 dy dy    and  = 7. Since < 0 for all x > 0, the function is ds x3 dx2 x dx nonincreasing in 10, m[, that is, the graph keeps on fatling as x increases. Further, since dy is nonzero for all x, there are no extreme points. dx d2y Similarly, since  is nonzero, there are no points of inflection. Writing the equation of dx2 the curve as x~~ = 1, we see that both the axes are asymptotes of the curve. There are no singular points. Therefore, the curve does not fold upon itself. The curve is shown in Fig. 6.
Fig. 6 :Graph of y = l/x2
Example 5 Let us try to trace the curve given by the equation xy = 1.
i
\
Here we can see that either x and y both will be positive or both will be negative. This means that the curvo.Iies in the first and the third quadrants. 'Further, it wymmetric about the origin and hence, it is sufficient to trace it in the first quadrant an*tate   this through 180°to get the portion of the curve in the third quadrant
1
I
I
I
(1, 1) is a point on the curve and x = l/y means that as x increases in the first quadrant, y decreases. 1 Now the distance of any point (x, y) on the curve from the xaxis = J y1 = y = X the yaxis is also an asymptote. dy     I # 0 for any x. That is, there are no extrema. dx X2 dy At the point (1, 1) we have, =  1, which implies that the tangent at (1, 1) makes dx
Curve Tracing
I
 0 as
x  =. This means that the xaxis is an asymptote. Arguing on the same lines we see that
I
an angle of 135' with the xaxis. Considering all these points we can trace the curve in the first quadrant (see Fig. 7(a)). Fig. 7(b) gives the complete curve.
A
I
(a)
Fig. 7 (a) Graph of xy = 1 in thc tlrst Qudranf (b) complete graph
The curve traced in Example 5 is a hyperbola. If we cut a double cone by a plane as in Fig. 8(a), we get a hyperbola. It is a section of a cone. For this reason, it is also called a conic section. Figs. 8(b), (c) (d) and (e) show some other conic sections. You are already familiar with the circle in Fig. 8(d) and the pair of intersecting lines in Fig. 8(e). The curve in Fig. 8(b) is called a parabola and that in Fig. 8(c) is called an ellipse.
Fig. 8
The earliest mention of these curves is found in the works of a Greek mathematician Menaechmas (fourth century B.C). Later Apollonius (third century B.C.) studied them extensively and gave them their current names. In the seventeenth century Rene Descartes discovered that the conic sections can be characterised as curves which are governed by a second degree equation in two variables. Blaise Pascal (16231662) presented them as projections of a circle. (Why don't you try this? Throw the light of a torch on a wall at different angles and watch the different conic sections on the wall). Galileo (15641642) showed that the nath nf rr nrni~t;ln+h*n=,
Drawing Curves
obliquely (Fig. 9) is a parabola. Paraboloid curves are also used in arches and suspension bridges (Fig. 10). Paraboloid surfaces are used in telescopes, search lights, solar heaters . and radar receivers.
Fig.
L /
C2
=c
Fig. 10

In the seventeenth century Johannes Kepler discovered that planets move in elliptical orbits around the sun. Halley's comet is also known to move along a very elongated ellipse. A comet or meteorite coming into the solar system from a great distance moves in a hyperbolic path. Hyperbolas are also used in sound ranging and navigation systems. Let's look at the next example now. Example 6 Consider the curve y = g3
+ x2.
There is no symmetry and the curve meets the axes at (0,O) and ( 1,O). dy dy = 3x2 + 2x. The xaxis is the tangent at the origin as = 0, at x = 0. Since dx dx dy = dx
I when x =  1,tangent at (  1,O) makes an angle of 45' with the xaxis (Fig. 1 l(a)).
in Fig. 1l(b),O is a valley and P is a peak.
A
Y
/
I
A
Y Y
I
,'I
/I
P
I
d2y = O a t x =  dx2 through  1/3.
1
Curve Tracing
3
and changes sign from negative to positive as x passes
Hence ( 1/3,2/27)js a point of inflection. dy dy 2 = x(3x 2). Hence > 0 when x <  or x > 0. dx dx 3 2 dy If  '3 < x < 0, then < 0. Thus the graph rises in ]  *, 2/3 [ and 10, oo[, but dx falls in 1 2/3,0[.
+
As x tends to infinity, so does y. As x 
 *, so does y. There are no asymptotes.
Hence the graph is as shown in Fig. 1l(c). So far, all our curves were graphs of functions. We shall now trace some curves which are not the graphs of functions, but have more than one branch.
Example 7 To trace the semi cubical parabola y2 = x3, we note that x3 is always ponnegative for points on the curve. This means x is always nonnegative and no portion of the curve lies on the left of the yaxis.
There is symmetry about the xaxis (even powers of y). The curve meets the axes only at the origin. The tangents at the origin are given by y2 = 0 so that the origin is a cusp. (see Sec. 4 in Unit 8). In the first quadrant y increases with x and y  00 as x
 oo.
Thete are no asymptotes, extreme points and points of inflection. Taking.reflection in the xaxis we get the complete graph as shown in Fig. 12.
Example 8 Suppose we want .to trace the curve
If x < 2, we get a negative value for Y2 which is impossible. So, no portion of the curve lies to the left of the line x = 2. For the same reason, no portion of the curve lies bebween the lines x = 3 and x = 4. . sin& y occurs with even powers alone, the c&ve is symmetrical about the xaxis. We may thus trace it for points above the xaxis and then get a reflection in the xaxis to complete the graph. The curve meets the axes in points A(2, O), B(3,O) and C(4, Q). At each of these points, the curve has a vertical tangent (see Sec. 2 of Unit 8). Combining these facts, the shape of .the curve near A, B, C must be as shown in Fig. 13(a).
Fig. 12 :Semi cubical parabola, y' =xJ
Fix. 13
Drawing Curves
Let us take y > 0 (i.e., consider points of the curve above the xaxis). Then
dydx
3x218xt26 * This is zero at x = 3 24(x  2) (x  3) (X 4)
.
* 1 / 6 . If cu = 3 + 1 / a
and p = 3  1 / 6 then a lies between 3 and 4, and can therefore be ignored. Also, 3x2  18x 26 = 3(x  p) (x  a) and 2 < fl < 3 < a.For x E]2,3[, x  a remains
+
dy negative. Hence for 2 < x < fl,  > 0 since (x  a ) and (x  p) are both negative. dx Similarly for j3 < x < 3, dy
< 0. Hence the graph rises in 12, p[ and falls in ]P, 3[. Thus
the shape of the curve is oval above the xaxis, and by symmetry about the xaxis, we can complete the graph between x = 2 and x = 3 as in Fig. 13(b). Now let us consider the portion of the graph to the right of x = 4. Shifting the origin to (4, O), the equation of the curve becomes Y2 = X(X 1) (x 2) = x3 3x2 2x.
+
+
+ +
As x increases, so does y. As x  w, so does y (considering points above the xaxis). When x is very small, x3 and 3x2are negligible as compared to 2x, so that near the (new) origin, the curve is approximately of the shape of y2 = Zx. For large values of x, 3x2and 2x are negligible as compared to x3,so that the curve shapes like y2 = x3 for large x. Thus, at some point the curve changes its convexity. This conclusion could also be drawn by showing the existence of a point of inflection. There are no asymptotes or multiple points. Considering the reflection in the xaxis, we have the complete graph as shown in Fig. 13 (c).
Example 9 Let us trace the curve (x2  1) (y2 4) = 4 .
.
There is symmetry about both axes. We can therefore sketch the graph in the first quadrant only and then take its reflection in thiyaxis to get the graph above the xaxis. The reflection of this graph in the xaxis will give the complete graph. Notice that the origin is s point on the graph and the tangents there, are given by 4x2 y2 = 0. These being imaginary, the origin is an isolated point on the graph. The curve does not meet the axes at any other points.
+
For x > 0, y > 0, the equation (x2  1) (Y2 4) = 4 shows that x should be greater than 1 and y should be greater than 2.
Equating to zero the coefficients of the highest powers of x and y, we get y = f2 and .x = f 1 as asymptotes of the curve. Thus, the portion of the curve in the first quadrant approaches the lines x = 1 and y = 2 in the region far away from the origin. In the first quadrant, as x increases, so does x2  1, and since x2  1 = y decreases as x increases. There are no extreme points, singular points or points of inflection. As x  =, y  2 and as y a, x
Curve Tracing
4 (Y* 4)
,
 1. Hence the graph is as shown in Fig. 14.
Example 10 To trace the curve y2 = (x  I) (x  2)2 we note that there is symmetry about the xaxis.
No portion of the curve lies to the left of x = 1. Points of interseition with the axes are A(l, 0) and B(2,O) and the tangent at (1.0) is vertical. Shifting the origin to B(2, O), the curve transforms into y2 = x2(x 1). The tangents at the new origin B, are given by y2 = x2. This means that B is a node, and the tangents ataBare equally inclined to the axes. Let us try to build up the graph above the xaxis between x = 1 and x = 2. Differentiating the equation of the curve with respect to x, we get 2yy' = (x  212 2(x  I ) (x  2), = (x  2)(3x  4).
+
+
when L < x < 2, (x  2) < 0. If y is positive, then > 0 provided 3x  4 < 0. Thus y' > 0 when x E ]1,4/3[ and y' < 0 which x E ]4/3, 2[. The tangent is parallel to the xaxis when 3x  4 = 0, that is, when x = 4/3 (see Fig. 15(a)). Hence, for 1 < x < 2, the curve shapes as in Fig. 15(b). Now for x > 2. As x  = g  00, y ?r in the first quadrant. Note that when B(2.O) is taken as the origin, the equation of the curve reduces to Y2 = x2(x I ) = x3 x2

+
+
This shows that when x > 0 and y > 0, the curve lies above the line y = x (on which Y2 = x2). Hence the final sketch (Fig. IS (c)) shows the complete graph.
Drawing Cllrves
*If you have gone through Examples 110 carefully, you should be able to do the following exercise. E 1) Trace the curves given by
(Graph paper is provided at the end of this unit.)
9.4 TRACING A CURVE :PARAMETRIC EQUATION
Sometimes a functional relationship mav be defined with the help of a parameter. In such cases we are given a pair of equations wEich relate x and y with the parameter. You have already come across such parametric equations in Unit 4. Now we shall see how to trace a curve whose equation is in the parametric form. We shall illustrate the p r d through an example. Example 11Let us trace the cycloid x = a(t sint), y = a(l . from  n to T. dx dy a(l cos t), a sin t, so that = dt
+
 cos t) as t varies
+
x=
dx Since > 0 for all t E ]  n, n[, x increases with t from  a n (at dt t =  n) to 0 (at t = 0) to a n (at t = n). dy Also, is negative when t E ]A, 0[ and positive when t E 10, T[. Hence y decreases dx from 2a to 0 in [ T, 01 and increases from 0 to 2a in [0, n]. Let us tabulate this data. t E [T,0]
dy = tan (t/2). dx
t E 10, TI
i) x increases from 0 to a
i) x increases from  a to 0 ii) y decreases from 2a to 0 iii) Hence the curve falls Also, at the terminal points  n, 0 and following. t
(x, Y)
T
ii) y increases from 0 to 2a iii) Hence the curve rises of the intervals [T,
01 and [O,,n], we have the
dy dx not defined 0 not defined
dx dy 0 not defined 0
Tangent vertical horizontal vertical
n
0
T
( an, 2 a)
(0,o) (am, 2a)
On the basis of the data tabulated above, the graph is drawn in Fig. 16.
Remark 1 If t is increased by 27r, x is increased by 2na and y does not change. Thus the complete graph can be obtained in intervals ..... [ 5n,  3 ~ 1[ 3n.  n], [n, 3 ~ 1 , , [3n, 5n] .....by mere translation through a proper distance. The cycloid is known as the Helen of geometry because it was the cause of many disputes among mathematicians. It has many interesting properties. We shall describe just one of them here. Consider this question : What shape should be given to a trough connecting two points A and B, so that a ball rolls from A to B in the shortest possible time? Now, we know that the shortest distance between A and B would be along the line AB (Fig. 17). But since we are interested in the shortest time rather than distance, we must also consider the fact that the ball will roll quicker, if the trough is steeper at A. The Swiss mathematicians Jakob and Johann Bernoulli proved by exact calculations that the trough should be made in the form of an arc of a cycloid. Because of this, a cycloid is also called the curve of the quickest descent. . A
Curve Tracing
The cycloid is used in clocks and in teeth for gear wheels. It can be obtained as the locus of a fixed point on a circle as the circle rolls along a straight line. See if you can do this exercise now.
=,' + .';\ :
F.3.
\'I\\
,\ :
\\

\ : \
\:q* '5:.
* \ :
E 2) Trace the following curves on the graph paper given at the end of this unit.
a) x = a(t
\. \:. . \\..  >= \..I : =>: & B
2w : \ \
a
+ sin t),
y = a(l
+ cos t),  n 5 t 5
T.
Fig. 17
9.5
TRACING A CURVE : POLAR EQUATION
In this section we shall consider the problem of tracing those curves, whose equations are given in the polar form. The following considerations can be useful in this connection. Symmetry : If the equation remains unchanged when 8 is replaced by  8, then the curve to is symmetric with re~pect the initial line. If the equation does not change when r is replaced by  r, then the curve is symmetric about the pole (or the origin). Finally if the equation does not change when 8 is replaced by n  8, then the curve is symmetric with respect to the line 8 = n/2. Extent : (i) Find the limits within which r must lie for the permissible values of 8. If r (r > a) for some a > 0, then the curve lies entirely within (outside) the circle r = a. (ii) If rZis negative for some values of 8, then the curve has no portion in the corresponding region. Angle between the line joining a point of the curve to the origin and the tangent : At suitable points, this angle can be determined easily. It helps in knowing the shape of the de curve at these points. Recall that angle 4 is given by the relation tan 4 = r . dr We shall illustrate the procedure through some examples. Study them carefully, so that you can trace some curves'on your own later. Example 12 Suppose we want to trace the cardioid r = a (1 following observations.
<a
+ cos 8). We can make the
,
Since cos 8 = cos ( 8), the curve is symmetric with respect to the initial line. Since  1 5 cos6 I 1, the curve lies inside the circle r = 2a.
 =  a sin 6. Hence  < 0 when 6< 8 < n. Thus r decreases as 8 increases in d8 de
dr
dr
I
Drawing Curves
the interval 10, a[. Similarly, r increases with 8 in ]a/2, a[. some corresponding values of r and 8 are tabulated below.
d8 a(1 cos 0) . This shows that the angle =  cot (0/2) = tan asin8 dr = between the line joining a point (r, 8) on the curve to the origin and the tangent is 0 or 7r/2 according to 8 = a or 0. Hence the line joining a point on the curve to the origin is orthogonal to the tangent when 8 = 0 and coincides with it when 8 = a . Combining the above facts, we can easily draw the graph above the initial line. By reflecting this portion in the initial line we can completely draw the curve as shown in Fig 18. Notice the decreasing radii 2a, cl, c2, c3 etc. This curve is called a cardioid since it resembles a heart.
+
u
Fig. 18
Example 13 Let us trace the equiangular spiral r = aee CO"u. We proceed as follows.
When 8 = 0, r = a. dr  = r cot a , which is positive, assuming cot a d8
> 0. Hence as 8 increases so does r.
d8 r  = tan a . Thus, at every point, the angle between the line joining a point on the, dr curve to the origin and the tangent is the same, namely a. Hence the name. Combining these facts, we get the shape of the curve as shown in Fig. 19. is also known as the curve of pursuit. The equiangular (or logarithmic) spiral r = aee Suppose four dogs start from the four corners of a square, each pursues the dog in front with the same uniform velocity (always following the dog in front in a straight line), then each will describe an equiangular spiral. Several shells and fossils have forms which are quite close to equiangular spirals (Fig. 20). Seeds in the sunflower or blades of pine cones are also arranged in this form. This spiral was first studied by Descartes in 1638. John Bernoulli rectified this curve and was so fascinated by it that he willed that an equiangular spiral be carved on his tomb with the words 'Though changed, I rise unchanged' inscribed below it.. The spiral r = a8 is known as the Archimedean spiral. Its study was, however, initiated by Conan. Archimedes used this spiral to square the circle, that is, to find a square of area equal to that of a given circle. This spiral is widely used as a cam to produce uniform linear motion. It is also used as casings of centrifugal pumps to allow air which increases uniformly in volume with each degree of rotation of the fan blades to be conducted to the outlet without creating backpressure.
Fig. 20
Fig. 19
The spiral r8 = a, due to Varignon, is known as the reciprocal or hyperbolic (recall that xy = a is a hyperbola) spiral. It is the path of a particle under a central force which varies as the cube of the distance. Now let's consider one last example.
#
Example 14 To trace the curve r = a sin 38, a > 0, we note that there is symmetry about the line 8 = a/2, since the equation is unchanged if 8 is replaced by a  8.
*
The curve lies inside the circle r = a, because sin 38 5 1. The origin lies on the curve and this is the only point where the initial line meets the curve. r =0 8 = n7r/3, where n is any integer. Hence the origin is a multiple point, the lines 8 = 0, a/3, 27r/3, 7r, 4a/3, 5a/3, 27r etc. being tangents at the pole. dr  = 3 cos 38. Hence r increases in the intervals 10, 7r/6[, ]a/2, 5a/6[, and d8 ]7i~/6,37r/2[, and decreases in the intervals ]7r/6, a/2[,]5a/6, 77r/6[ and ]3a/2, 5a/3[ !'Jotice that r is negative when 8 E ]a/3,27r/3[ or 8 E IT, 4 ~ / 3 or 8 E ]5a/3, 2 ~ [ . [
*
(0a)
Fig. 21
Hence the curve consists of three loops as shown in Fig. 21. The function is periodic and the curve traces,itself ai 8 increases from 2n on. Now try to trace a few curves on your own.
Curve Tracing
E 3) Trace the following curves on the graph paper provided. a) r = a(l  cos 8),a > 0. b) r = 2 4 cos 8. C) r = a cos 38, a > 0. $) r = a sin 28, a > 0 (Graph paper is provided at the end of this unit.)
+
9.6 SUMMARY
'
In this unit we have covered the following points. 1) Tracing a curve y = f(x) or f(x, y) = 0 means plotting the points which satisfy this relation. 2) Criteria for symmetry and monotonicity, equations of tangents, asymptotes and points 'of inflection are used in curve tracing.
3) Curve tracing is illustrated by some examples when the equation of the curve is given
in a) Cartesian form b) Parametric form c) Polar form
9.7 SOLUTIONS AND ANSWERS
i
Dotted lines represent tangents or asymptotes throughout.
b) Shifting the origin to (2,O) we get yZ = x3 which you know how to draw.
Drawing Curves
C)
y = x is the tangent at the origin (0. 0).( 1. (\Ti 3 arc point of infections.iaxis is asasymptote. E~ther y areioth positive or both negative. x, funct~on rlses in 1 1, I[ and falls elsewhere. Graph is shown alongside.
6 $5
7
Y= 1  x2 shows that the entire curve lies within the lines x = f 1. d) xTangents at the origin are y = x. Tangents at x J f 1 are vertical.
7
+
Maxima at (f 1 / f i
1/4), symmetry about both axes.
t
Curve Tracing
I
~
I
I
Drawing Curves
c)
UNIT 10 DEFINITE INTEGRAL
Structure
10.1 Introduction
Objectives
10.2 Preliminaries
Partitions of a Closed Interval Upper and Lower Product Sums Upper and Lower Integrals
10.3 10.4 10.5 10.6
Definite Integral Fundamental Theorem of Calculus Summary solutions and Answers
10.1 INTRODUCTION
We have seen in Unit 3 of Block 1that one of the problems which motivated the concept of a derivative was a geometrical one  that of finding a tangent to a curve at a point. The concept of integration was also similarly motivated by a geometrical problem that of finding the areas of plane regions enclosed by curves. Some recently discovered Egyptian manuscripts reveal that the formulas for finding the areas of triangks and rectangles were known even in 1800 B.C. Using these one could also find the area of any figure bounded by straight line segments. But no method for findingthe area of figures bounded by curves had evolved till much later. In the third century B.C. Archimedes was successful in rigorously proving the formula for the area of a circle. His solution contained the seeds of the present day integral calculus. But it was only later, in the seventeenth century, that Newton and Leibniz were able to generalise Archimedes' method and also to establish the link between differential and integral calculus. The definition of the definite integral of a function, which we shall give in this unit was first given by Riemann in 1854.1.nUnit 11, we will acquaint you with various methods of integration. You have probably studied integration before. But in this unit we shall adopt a new approach towards integration. When you have finished the unit, you should be able to tie in our treatment with your previous knowledge.
objectives
After reading this unit you should be able to define and calculate the lower and upper sums of some simple functions defined on [a,b], corresponding to a partition of [a,b], define the upper and lower integrals of a function, define the definite integral of a given function and check whether a given function is integrable or not, state and prove the Fundamental Theorem of Calculus, use the Fundamental Theorem to calculate the definite integral of an integrable function.
'
10.2 PRELIMINARIES
We have mentioned in the introduction that Archimedes was able to find the formula for the area of a circle. For this he approximated a circle by an inscribed regular polygon (See Fig. 1 (a)). ~ . Further, we can see from ~ il(b) that this approximation becomes better and better as we increase the number of sides of the polygon. Archimedes also tried to .approximate the area of the circle by a number of circumscribed polygons as in Fig. l(c). The area of the circle was thus compressed between the inscribed and the circumscribed polygons.
Integral Calculus
Fig. I
Y =
f
(x)
We shall follow a similar procedure for finding the area of the shaded region Shown in Fig. 2. We begin with the concept of a partition.
10.2. l Partition of a Closed Interval
Let us consider the closed interval [a,b]
R. Then we have the following definition
0
X
D d f ~ ~ i t1o ~ ~ g x , , x2,....,xIF],X, be numbers in [a,b] such that i Let , a=x,<~~<x,<...<x~~<x,=b.
Then the ordered set P = {x,),xl,x2,.... ,x,} is called a
I
I
+
of [a,b].
Example 1 P I = {0,..114,112,314, 1) and P2 = { O , 113, 112,213, 617, l), bcith are partitions of [0,1]. FIB.2 Moreover, ~yanordtrsdsetwemeanaset.in' P I U P2 = (0, 114, 113, 112,213, 314,617, 1) and PI n P2 = (0, 112, 1) arealso the in which its partitions of [O,I]. See Fig. 3 (a). (b). (c) and (d).
ments occur is fixed.
I
A set J iscalled asubinterval of an interval I, if i) J is an interval, and ii) J ~ I ..
A partition P ='{xo,xI ,x2..... ...xn) of [a;b] dihdes [a,b] into n closed subintervals,
.
'A' is read as delta.
[ x I ~ x z ] * . . . . . . [ x ~xn].. I with the n + 1 partitioningpoints as endpoints. The interval [x,, ,xi] is called the ith subinterval of the partition. The length of the ith subinterval, denoted by A x i , is defined by A x I. = x.  x. I 11It follows that
[XO,XI],
d2  x l
We call partition P regular if every sibinterval has the same I'ength, that is;if xl  xO, ,....,x, are all equal. In this case, the length of [a,b], that is ba, is equally divided into h parts, and we get XIXo = xzXI = ......  X n  X n  ~  .ba .. n Thus, a regular partition.of [a,b] may be written as
,
{a, a + h, a + 2h,
1
....., a + nh), where a + nh = b: We shall denote this partition by
{a + ~h);=() F o r P = 11,312, 2,5/2,3,7/2,4), A x , = x ,  x o = 312  1 = 112, A x 2 = x2  x, = 2  312 = 112. If you calculate Ax3, Ax4, A x s and Ax,, you will see that P is a regular partition of [1,4].
E
E l ) See Exiinple 1. Whichpartitionsamong PI, P2, P I U P2 and PI t P2are regular? l What are the lengths of the third subintervals in P, and in P,?
$1
1'
E
L2) Write down a regular partit~on each ut the lulluw~ng for mtervals. a) [O, 21 with 7 partitioning points. b) [2,9] with 11partitioning points.
:.A
Definition 2 Given two partitions P, and P2 of [a,b], we say that P2is a refinement of PI (or P2 is finer than PI) if P2 3 PI. In other words, P2 is a refinement of P I if each subinterval of P, is contained in some subinterval of P I . Example 2 Consider the partitions p1 = {i,5/4,3/2,7i4,2), P2 = {1,6/5,5/4,3/2, 19/10,2), P3 = {1,5/4,3/2,2) P, and P2 are both finer than P3, as P, 3 P3 and P2 3 P3. However, neither is P, a refinement of P i nor is P2 a refinement of P I . If P, and P2 are partitions of [a,b], then from Definition 2 it follows that
$1;i
'7.
;'!I
:i,r
i) PI U P2 is a refinement of both P, and P,. ii) PI and P2 are both finer than P I n P,. Now, suppose for every n E N we define P, as P,={a+i ban )Ln o 2 l= This means P, has 2" + 1 elements. We can see that P, is a regular partition, with each subinterval having length =
Ax, x,  x,_, =
=a
h.
2"
+ib~!(~+(il)b2!}
2"
2"
ba 2"
,,4
#
&f
1
This means that the length of the subintervals corresponding to P,, is half the length of those corresponding to P,. We can also see that P,, 13 In other words, P,, P,. is finer than P, (also see E3)). Thus, we have defined a sequence of partitions {P,) of
,
,
:2 r',
[a,b], such that P,, is a refinement of P, for all n. Such a sequence {P,) is called a seqdence of refinements of partitions of [a,b].
E
E3) From the sequence of partitions {P,) defined above, Pl = {a,7 a + b b}. a) Find P2 and P3. b) Verify that P3 3 P 23PI. c) What are the lengths of the subintervals in each of these partitions?
E4) Let {Pdm 1 be a sequence of partitions of [a,b], and let P; = PI, n
E = P1 U P2, P3 = P1 U P2 U PJ, and in general, PE = PnU PL1. Show that { P ~ } ~is a, sequence of refinements of [a,b]. =
1
10.2.3 Upper and Lower Product Sums
By now, we suppose you are quite familiar with partitions. Here we shall introduce the concept of product sums. It is through this that we shall be in a position to probe the mwe~ubtle concept o a definite integral in the next section. f
Let f: [a,b] + R be a bounded function, and let P = {xo, xl ,xz,....,xJ be a partition of [a,b].
We have defined 1.u.b. (supm w n ) and g.1. b. (infimum) of a Since f is a bounded function, Si must be a bounded subset of R. This means, in it has a supremum (or least upper bound) and infimum (or p a t e s t lower bound). rot uf Unit1.
1
I
I'
I
Now for any subinterval [xi_,, x,], consider the set Si = {f(x) : x € [xi4, 4 1. 1 We write Mi sup& = sup{f(x) i x t [ X C ~ , ~and, = ]} mi * i d Si 6 i d {f(x) : x € [kl, xi]).
.
We now define the opper product amn U( , ) the lower prwl~ct L P f by P iand sum ( , )
You must have come across this S .notation earlier. But let us state clearly what (1) means : U(P,f) = M1 (xl  ~ g ) M2 (x2  xl) + .... + Mn (xn  x ~  ~and + ), L(P,f) = m1 (xl  ~ g ) m2 (x2  xl) + .... + mn (x,  x,~). + Thus, to get U(P,f) we have multiplied the supremum in each subinterval bf the length of that subinterval, and have taken the sum of all such products. Similarly, L(P,£) is obtained by summing the products obtained by multiplying the infimum in each subinterval by the length of that subinterval. U(P,f) and L(P,f) are also called Rlemcurn s m m after the mathematician George Friedrich Bernhard Riemann . Riemann gave a definition of definite integral that, to this day, remains the most convenienl and useful one. We started.this unit saying that we wanted to find the area of the shaded region in Fig. 2. Then what are we doing with partitions, U(P,f) and L(P,f)? Fig. 4 will give you a clue to the path which we are going to follow to achieve our aim.
G.F.B. Risnuna (18261866)
I ,t
I
1'
Fig. 4(a) and 4(b) give the geometricview of MiAxi and miAxi as areas of rectangles with base Axi and heights Mi and mi, respectively. The shaded rectangles in Fig. 4(a) are termed as outer rectangles, while the shaded rectangles in Fig. 4(b) are called in.ner rectangles. Thus, when f is a nonnegative valued function (f(x) 2 0 V x), U(P,f) = sum of the areas of outer rectangles as in Fig. 4(a). L(P,f) = sum of the areas of inner rectangles as in Fig. 4(b), and u(P,@  L(P,f) = sum of the areas of the shaded rectangles along the graph of f in Fig. qc). As you see from Fig. 5, U(P,f) and L(P,f) depend upon the function f:[a,b] + R (compare Fig, 5(a) and (b) ), and the partition P of [a,b] (compare Fig. 5(c) and (d) ).
L
I
mg. s (a) U(P ,f) where y = x :
(b) U(P,f) where (c) U(P,f) when P = {0,1,2,3) (d) UP,f) when P
9=x
{0,112,1,3/2,2,5/2,3)
If we denote the area between the curve given by y = f(x), the xaxis, and the lines x = a and x = b, (the shaded area in Fig. 2) by A, then it is also quite
clear from Fig. 4(a) and (b), that L(P,f) 5 A U(P,f). The geometric view suggests the following theorem:
Theorem 1Let f:[a,b]  R be a bounded function, and let P be a partition of , [a,b]. If M and m are the supremum and the infimum off, respectively, in [a,b], then m(ba) s L(P,f) 5 U(P,f) 5 M(ba).
Prod Now M = Sup {f(x) : x E [a,b]), and
aa
. .  (st\
r
11
11

If X C Y , then eupX~supY,oad inf X inf Y,
am
am
~rlcdcdm
mi
Further, m = inf {f(x) : x c [a,b]), and = inf {f(x) : x c [x,, , xi]). Thus, m 5 mi. This means m ~ m ~ s M , s M ...... (2) Once we have the inequalities (2), we can complete our proof in easy steps. (2) implies that m Ax, 5 mi Axi 5 Mi Axi c: M a x i This implies that if we take the sum over i = l , 2 ,....., we get n, m
i=l
2 Axi s L(P.0
d
n
c U(P.0
S
M
S
i=l
2 Ax,
n
* m(ba).
n
L(P,f)
U(P,f)
M(ba),
since
2 Axi =
i=l
the sum .of the lengths of all subintervals
= the length of [a,b] = ba.
Fig. 6 will help you understand this theorem better. Let us verify this theorem in the case of a given function.
Example 3 Let f:[1,2] + R be a function defined by f(x) = x2. and let P = {1,5/4,3/2,5/3, 2) be a partition of [1,2]. .The subintervals associated with P are [I, 5/41, [5/4, 3/21, [3/2, 5/31 and [5/3, 21.
The function f is a bounded function on [1,2]. In fact. the image set of f is [1,4], which is obviously bounded.
ne. 7
, 
Since f is an increasing function on each subinterval (see Fig. 7) the supremum of f in [xi_,, xi] will be attained at xi and the infimum will be attained at x,,. That is, Mi = f(xi) and m, = f(x,,). Therefore, we can write U(P,f) = Z Mi Axi = z f(xi) Axi = s X (x,x~I) : (~3x;) + 4 (&~3) = $ ( X I  3 + d &XI) +
6
:;I
Now, the supremum of f(x) in [1,2] = M = f(2) = 2'= 4, and the infimum = m = f(1) = 1. Thus, M(ba) = (4) (21) = 4, and m(ba) = (1) (21) = 1. Thus, m(ba) 5 L(P,f) 5 U(P,f) 5 M(ba). We have noted that the upper and lower product sums depend on the partition of the given interval. Here we have a theorem which gives us a relation between the lower and upper sums corresponding to two partitions of an interval.
Y
i t
Theorem 2 Let f:[a,b] + R be a bounded function, and let P1 and P2 be partitioris of [a,b]. If P2 is finir than PI, then
L(P,,f) 5 L(P2,f) 5 U(P2,f) 5 U(Pl,f). Pro& For proving this theorem we look at Fig. 8(a) and (b). Let PI = {xO,XI, x2,....,x,) and P2 = {x0, vl, x1,x2,.....,x,) be two partitions of [a,b]. P2 contains one element more than PI, namely, vl. heref fore, P2 is finer than P,. In fact, P2 can be rightly called a simple refinement of PI. We shall prove the theorem for this simple refinement here. PI divides [a,b] into n subintervals : [xo, XI], [XI, xzI,.....,[x,+1,
~"1.
Fig. 8
Fig. 8(a) clearly shows that L(Pl,f) 5 L(P2,f) (by an amount represented by the area of the shaded rectangle). Similarly, Fig 8(b) shows that U(P2,f) s U(P,,f). Since L(P,;f) 5 U(P,,f), the conclusion of the theorem follows in this case. Now if P2 is not a simple refinement of PI, then suppose P2 has m elements more than P,. Then we can find (m1) partitions P;, P;, P;, ......., such that ) P, c pi c c P, c ..... c py: c p2 and each partition in this sequence is a simple refinement of the previous one.
C
r1
~ n t c gC d d m d
Theorem 2 then holds for each pair of successive refinements, and we get L(Pl,f) 5 L(P$,~) U(P2,f)
5
L(@,f)
5
5
.......
5
I .
L(PT',~)
5
I
L(P2,f) and
u(pT1,f)
....... 5
5
u(P$,~)
u(P:,~) sU(P1,f)
Thus, L(P1,f)
=
L(P2,f)
U(P2,f)
U(P1,f).
From Theorem 2 we conclude the following : Let f:[a,b] + R be a continuous and nonnegative valued.function, and let {P,JL1 be a sequence of refinements of [a,b]. Then we have L(P,,f) 5 L(P2,f) 5 ....... 5 L(Pn,f) ...... s A s . : . c . . g U(Pn,f) 5 ...... 5 U(P2,f) 5 U(Pl,f), where A is the area bounded by the curve, the xaxis and the lines x = a and x = b.
=
E
ES) Find the upper product sum and the lower product sum of the function f
relative to the partition P, when a) f(x) = 1
*
+ x2,
P = (0, 112, 1, 312, 2)
b) f(x) = llx, P = (1, 2, 3, 4)
E
E6) Verify Theorem 2 for the function f(x) = llx, 2 C x 5 3, and the partitions PI = (2, 512, 3) and P2 = (2, 914, 512, 1114, 3) of [2,3].
In this subsection we have seen that the area A in Fig. 2 can be approximated by means of the lower and upper sums corresponding to some partition of [a&]. Further, Theorem 2 tells us that as we go on refining our partition, the lower and upper sums approach A from both sides. The lower sums underestimate A (L(P,f) I A), while the upper sums overestimate A (U(P,f) r A). Let us go a step further in the next subsection, and define lower and upper integrals.
10.2.3 Upper and Lower Integrals
Let f:[a,b] + R be a n~n~negative bounded function. Then to each partition P of [a,b], there correspond the upper product sum U(P,f) and the lower product sum L(P,f). Let P be the set of all partitions of [a,b]. Then the set u = {U(P,f) : P e P) is a subset of R and is bounded below since A s U(P,f) 4 P e P. Thus, it is possible to find the infimum of u. Similarly, the set u' = {L(P,f) : P E P} is bounded above, since L(P,f) s A 4 P E P. Hence we can find the supremum of u'. The infimum of u and the supremum of ' u ' are given special names as you will see from this definition.
R ~nit ,)aatCveryICtWhich ( ~ is h.udd bDla hrn smum, and every set which is bounded above has a supremum.
Defiiition 3 If a function f is defined on [a,b] and if P denotes the set of all partitions of [a,b], then infimum of {U(P,f) : P E P) is called the upper integral of f
b
on [a,b], and is denoted by /f(x) dx.
a ' t
XI
The symbol arb read as integral.
,
The supremum of {L(P,f) : P E P ) is called the lower integral of f on [a,b], and is denoted by

%
f(x) dx.
From Theorem 2 it follows that Thus we have
/ f(x) dx
b
b
1
r A and

[ f(x) dx
I
A.
[ f(x) dx s A s 1 f(xj dx. 
Example 4 Let us find
j:
+
l1
f(x) dx and

f(x) dx.
.
1
for the function f, defined by f(x) =
0 if x is rational 1 if x is irrational.
*+
I
+
Suppose P= {xo, xl, xz,.....,x,) is a partition of [0,1]. Each subinterval [xil, xi] contains both rational and irrational numbers. This means, Mi = 1 and mi = 0 for each i.
24
and
Since P was any arbitrary partition of [0,1], this means that U(P,f) = 1 and L(P,f) = 0 4 P E P.
. Thus,
u = {U(P,f) : P E P ) = (1)
E
::.I rt
4
and u = {L(P,f) : P
P ) = (0)
Hence inf u = 1 and sup u' = 0. That is, f(x) f(x) dx 0. dSee if dx can1doand jo exercises now. you these
=
=
Integral Calculus
E
E7) Find
[ f(x) dx and b
1

f(x) dx. for the function f defined as
E8) If the functions f and g are bounded nonnegative valued functions in [a.b]
and if f(x)
h
5
I)
g(xi in ,a,b], prove that
(
flx)dx Lfg(x)dx
a
a.1
f(x) dx
jr(x) dx.
10.3 DEFINITE INTEGRAL

In the last section we had restricted our discussion t o nonnegative valued function\.
But we can easily extend o u r definitions of L(P,f), U(P,f) and the lower and upper
integrals t o all bounded functions, However, we shall have t o modify our
internretatinn nf these w m s a 1 areas Fnr this ntlmnse we introduce the conrent
of signed area. If R is any region, its signed area is defined to be the area of its
Definite Integral
Fig. 9
With this definition then, we can interpret L(P,f) as the signed area of a polygon inscribed inside the given region, and U(P,f) as the signed area of a polygon circumscribed about the region. Thus, for any bounded function on a closed interval [a,b]. we can define

la dx f(x)
b
= sup {L(P.f) : p
r P), and
Now we are in a position to discuss the definite integral for a bounded function on a closed interval. (The adjective 'definite' anticipates the study of indefinite integral later). Definition 4 Let f:[a,b] + R be a bounded function. f is said to be integrable over [a,b] if, and only if,
This common value is called the definite integral of f over the interval of integration [a,b]. and is denoted by
1
I
h
f(x) d r
In this notation for the definite integral, f(x) is called the integrand, a is called the lower limit and b is called the upper limit of integration. The symbol dx following f(x) indicates the independent variable. Here x is merely a dummy variable, and we may replace it by t or v , or any other letter. This means,
The symbol
I
reminds us of S which is appropriate. because a definite integral is.
in some sense, the limit of a sum. In fact it is the common value (when it exists) of the lower and upper integrals which are themselves infimum and supremum sums. The use of f(x) dx reminds us that we do not take the sum of function values, rather we take the sum of terms. each of which is the product of the supremum or infimum of the function in an interval multiplied by the length of the subinterval. The definition of definite integral above, applies only if a < b, but it would be appropriate to include the cases a = b and a > b as well. In such cases we define ff(x) dx = O
L
h
and
' I
f(x) dx =
1
b
a
f(x) dx
provided the right hand integral exists.
~ntqp.I c8bhn
In Example 4, we have seen that if
f(x) =
I
0 if x is rational 1 if x is irrational, then
I,
E
i(x) dx = 0, and
[ f(x) dx
1
= 1.
Since the lower and upper integrals for this function are not equal, we conclude that it is not integrable.
E9) Check whether the function given in E7) is integrable or not.
~t
Now we shall list some basic properties of definite integrals. I Intqp1 o a constant function f(x) = c f
I
jcdr
a
=
~(b~)
.
I1
This is intuitively obvious since the area represented by tha integral is simply a rectangle with base ba and height c, (see Fig 10.). Now let us consider a function f which is integrable over [a, b]. Constant Multiple Property
I
a
k f(x) dx = k f(x) dx.
a
1
I11 Interval Union Property
b
a c
b
.
If a < c < b, then /f(x) dx = /f(x) dx
a
+
/t(x) dx
c
Its geometrical interpretation is shown in Fig. ll(a).
09 .
Fig. 11
IV Compruison Propeity If c and d are constants such that c
b
I
f(x)
Id
for all x in [a,b], then
c(ba)
I lf(x)
dx
5
d(ba)
Fig. l l ( b ) makes this statement clearer. Note that c and ,d are not necessarily the minimum and maximum values of f(x) on [a,b]. c may be less than the minimum, and d may be greater than the maximum. The following theqrep given a criterion for a function to b e integrable.
Definite Integrql
Theorem 3 A bounded function f is integrable over [a,b] 'if and only if, for every E > 0, thcre exists a partition P of [a,b] such that 0 5 U(P,f)  L(P,f) < e . Proof : We know that for any partition P of [a,b],
If the function f has the property that for every P of [a,b] such that U(P,f)  L(P,f) < E, we conclude that
05
E
> 0 there exists a partition
1
d

b
f(x) dx 

6
b
f(x) dx <
b
E
for every
E
> 0.
= 0 and hence f is integrable over
From this it follows that tab].
jf(x) dx  [ f(x) dx

On the other hand, if f is integrable over [a,b],
jb
a
f(x) dx =
SUP
{L(P,f) : P E P} = inf {U(P,f) : P E P}. Thus. for every
E
>0
we can find partitions P' and P" of [a,b], such that.
05
1
I
b
f(x) dx
 L ( I ,f) < ED, and
0 s U(P".f) 
1
b
f (x) dx < ED (see Sec. 2 of
Unit 1). Taking some partition P which is finer than both P' and P , and adding the two inequalities, we have
0 5 U(P,f)  L(P,f) < E. This completes the proof.
Now arises a natural question : Which are the functions which satisfy the above criterion? The following theorems provide a? answer.
Theorem 4 A function that is monotonic (increasing or decreasing) on [a,b], is integrable over [a,b]. Proof Let the function f:[a,b] + R be increasing. Then x, < x, =% f(x,) 5 f(x2). ba For eachpositive integer n, let Pn = {a, a+h ,.. .., a+nh = b}, where h =  be a n regular partition of [a,b]. h e n
since the supremum of f(x) in [a+(i1) h, a+ih] is f(a+ih). and L(Pn,f) = h
2 ml = ba
I=
n
n
f(a+(i1)h)
i=l
I
Therefore U(Pn,f)  L(Pn,f) =
ba [f(a+h) + f(a+2h) + .... + f(a+nh) n f(a)  f(a+h) .... f(a+(nl)h)]
n [f(a+nh)  f(a)]
 ba
> 0. Can we choose an n which will make U(Pn,f)  L(Pn,f) < E ? Yes, we can. Try some n > (ba)  f(a)l . If we substitute this value of .n
Let
E
E
in Ill, we get
+h=
r
ba <
n
E
f (b)f (a)
Thus, applying
heo or em 3, we can conclude that
,  . .
f is integrable.
Theorem 4 leads us to the following useful result.
Corollary 1 If f is increasing or decreasing on [a,b], then
1 f(x) dx
1
= lim h[f(a)
'h0
+ f(a+h) + .... + f(a+(nl)h)] + f (a+2h) + ..... + f(a+nh)],
where h =
= lim h[f(a+h)
b 0
ba .' n
We shall illustrate the usefulness of Corollary 1 through some examples. But before that we state another theorem, which identifies one more class of integrable functions. Theorem 5 If a function f:[a,b] + R is continuous, then f is integrable. .The proof of this theorem is beyond the scope of this course. We shall prove it in a later course on real analysis. In Sec. 5 in Unit 3, we have; seen that differentiability implies continuity. Now we can write integrability differentiability s continuity Now, let us evaluate some definite integrals with the help of Corollary 1.
Example 5 T o evaluate
f ;x
/COsx
I
1 cosx dx, 0
I
b
5
a
5
b c n12, we observe that
+ cosx
is a decreasing function on [a,b]. Therefore, by Corollary 1
dx = lim h[cos(a+h)
h.O
+ cos(af2h) +
. . . + cor(a+nh)], a+nh
= b.
Now 2 sin (hl2) [cos(a+ h) + cos(a+2h) + .. .. + cos(a+nh)] = 2 sin (hi2) cos(a+ h) + 2sin (h/2) cos (a+2h) + .... . f 2 sin.(h/2) cos (a+nh). 5h  ~ i n ( a + ~ + ].... + 3h = [sin(a+) 3h  s i n ( a +  )h . + I [ s i n ( a + ) ) 2 2 2 2n+ 1 ) h)  sin (a+( 2n' ) h)] [sin (a+(2 2 h .2n+ 1 = sin (a+(2 ) h)  sin (a+ ) 2 h = sin (b+  ) 7 sin (a + since a + nh = b 2 sin (b+h/2)  sin (a+h/2) cos(a+h) + cos(a+2h) + .... + cos(a+nh) = 2 sin h/2
p),
*
Thus, cosx dx = lim [sin (b
ho
h 2 = sin b  sin a .
+
) 
sin (a
h +  )] 2
h/2 sin (h/2)
L
Example 6 Suppose we want to evaluate
Here, f : x + x Therefore,
+ xZ
is an increasing function on [1,2].
/
2 
(x+x2) dx = lirn h
h0
f; ,=, f(l + ih), h = lln
n
= lirn h
b 0
'
[(l+ih)
i=l
n
+ (l+ih)*]
Recall that
= lim h
i=l
2 (2 + 3hi + h2 i2)
i=l
=
lirn [Zh
b.0
2 1 + 3h2 2 i + h3
,
n
n
i=l
i=l
iZ]
= lim
h0
[2nh +* h b ( n + l )
+
h3n(n+l) (2n+l)]
=
lim [2 + $ ( l
h40
+ h) +*
( l + h ) (2+h)]. since nh = 1.
In this section we have noted that a continuous function is integrable. We have also proved that a monotone function is integrable. Corollary 1 gives us a method of finding the integral of a monotone function. One condition which is very essential for the integrability of a function in an interval, is its boundness in that interval. If a function is unbounded, it cannot be integrable. In fact, if a function is not bounded, we cannot talk of M, or m,, and thus cannot form the upper or lower product sums. Now on the basis of the criteria discussed in this section you should be able to solve this exercise. E10) State whether or not each of the following functions is integrable in the giyen interval. Give reasons for each answer. a) f(x) = x2  2x + 2 in [1,5j b) f(x) = fiin [l,a] c) f(x) = l/x in [1,1] d) f(x) = [x] in [0,4] e) f(x) = (x11 in [0,3]
X' + f) f(x) = in [4,0] 2x + 1 x + 1 whenx<O in [1,1] 1  x when x 2 0,
h) f(x) =
x+lwhenx<l in [0,3] 2x + 1 when x L 1,
I

\
Intugral Calculus
E
E l l ) Use Corollary 1 to evaluate the following definite integral.
(I +x) dx
10.4
FUNDAMENTAL THEOREM OF CALCULUS
As you have already read in the introduction, the basic concepts of definite integral were used by the ancient Greeks, mainly Archimedes (287212 B.C.), d o r e than 2000 years ago. This was long before calculus was invented. But in the seventeenth century Newton and Leibniz developed a procedure for evaluating a definite integral by antidifferentlation. This procedure is embodied in the Fundamental Theorem of Calculus (FTC). Before we state this theorem, we introduce the notions of the average value ot function and the antiderivative of a function.
A
Definition 5 Let f be integrable over [:i.b]. The average value over [a,b] is
h
of y =: f ( x )
The following theorem tells us that every continuous function on a closed interval attains its average value at some point of the interval. We shall not give its proof here.
Theorem 6 (Average Value Theorem) If f : [i~,bl R is continuous, then t
f(X) =  f(x) dx ba ;, for some % E [a,b]. We shall now dcfinc the antiderivative of a function. Definition 6 Let f : [a.b] + R and F : [a,bl + R be two functions such that d  (F(x)) = F f ( x ) = f(x)for each x E ]a,b[. We call F(x) an antiderivative (or dx inyerse derivative) of f ( x ) .
1
2 IS :in iintiderivative of x 2 , s i n c e L ( 2 ) = x'; 3 dx 3 d  cds x is an antiderivative of sin x, since  (cos x) = sin x. dx xJ Is  x an antidcrivativc of x"  2x6?
4
~l;r example,
3
.
Consider the two functions f(x)
.
. .
C C.
=
. .
x2 and g(x)
1 . .
=
x2+5. Both these are
AL.
a>.*.::..:.
..
1
TI.:..
 c 
function is not unique. In fact, if F(x) is an antiderivative of f(x), then F(x) + c is also an antiderivative of f(x). This follows from the fact that d d  (F(x)) =  (F(x)+c)=~(x). dx dx We can also say that any two antiderivatives of a function differ only by a constant. Because, if F(x) and G(x) are two antiderivatives of f(x), then Ff(x) = Gf(x) = f(x). That is, [F(x)  G(x)]' = 0. We have noted in Unit 7 that if the derivative of a function is zero on an interval. then that function must be a constant. Thus (F(x)  G(x) = c. Now having defined the average value and the antiderivative, we are in a position to state the Fundamental Theorem of Calculus. We shall give this theorem in two parts. Theorem 7 (FTC) : Let f : [a, bl + R be a continuous function. Part 1 If the function F : [a,b] i , R is defined by
r X
Dellnlte Integral
The interval [a, b] on which f and its antiderivative are defined, so that F (x) = f(x) v x E la, b[, 1s irnplici~n our i diseussiar here.
i
r
F(x) = ) f(t) dt,
a
.....(3)
then F is an antiderivative of f, that is, Ff(x) = f(x) for all x in ]a,b[.
Part 2 If G is an antiderivative of f in ]a,b[, then
Proof of Part 1. By the definition of derivative, F1(x)= lim
hro
F(x+h) h
 F(x)
by the interval union property of definite integrals. But, by the Average Value Theorem (Theorem 6) f(t) dt = f(i) for some T E [x;.x+h]. hx Therefore, F1(x) = lim f(i). We know that i E [x,x+h]. This means that as h80 h + 0, i + x. Therefore, F1(x) =Jim f(i) = f(x), since f is a continuous function.
t+x
1
Hence, F as defined' by (3), is an antiderivaiive of f.
Proof of Part 2 G is given as an antiderivative of f in ]a,b[. Also, as shown in Part 1, F defined by (3) is an antiderivative of f in ]a,b[. Therefore, G(x) = .F(x) + c on ]a,b[ for some constant c.
To evaluate c, we substitute x = a, and obtain c = G(a)  F(a) = G(a)  0 = G(a). Hence G(x) = F(x) + G(a), or F(x) = G(x)  G(a) If we put x = b, we get
The Fundamental Theorem of Calculus tells us that differcnt~i~t~on intcgri~tion and are inverse processes, because Part 1 may be rewritten as f(t) dt) = f(x), if f is continuous. dx n That is, if we first integrate the continuous function f with the variable x as the upper'limit of integration and then differentiaie with respect to x , the result is the function f Gain. So differentiation offsets the effect of integration. On the other hand, if we assume that G' is'continuous, then Part 2 of FTC may be written as
(1
X
'
jot
1
(1) dt = G(x)
 C(a)
Here we can say that if we first differentiate thebfunctim G and then integrate the resulx from a to x, the result can differ from the originttl function G(x) only by the constant G(a). If G is so chosen that G(a) = 0, then integration offsets the effect of biffere@i&n. Till now vvs b?d evaluated the integralsof iome functions by first finding the lowcr and upper sums. and then taking their suprcrnum and infimum, respectively. This is a tedious prucedure and we cannot apply it easily to ail functionSs.But 'irow. FTC gives us an easy method of evaluating definite integrals. We shall illustrare this through some examples.
Since f :x+
G(x) =
.3
'.
ax' +.bx +/c is continuous on 12.31. it is integrable over [2,3)
+ + cx is an ilntiderivi~tivuof f(,x). 3 Hcncc. by FTC (Part 1)
= G(3) = (Oa =
It)"
 G(2)
+ 9b/2 + .k)  (Kid3 +
+ 5k +
alJ
3
,
a
2h
+ ') c
E u m P 8 Let us uvi~lu;ltc c t r l r dx
/
Then
x
.
=
 J'sint dt dx ,,
d
 J sill1 tit
dx
d

,,
Example 9 suggests the following formula.
If you have followed these examples, you should be able to solve the exercises below. Remember that the main thing in evaluating a definite integral is to find an antiderivative of the givep function.
E
E12) The second column in the table below consists of some functions which are antiderivatives of the functions given in column 1. Match a function with its antiderivative by pairing appropriate numbers. For example, we can match xn with n + ,We shall indica'te this by iii) + viii).
Xn+ 1
since.
~ n + l.
n+l
IS an
antiderivative of xn
Function
Antiderivative
i) ii) iii)
sinx cosx xn
eax
i) ii) iii) iv)
 In cosx
In coshx sechx cosx sinx
1
iv) v)
tanx
1
)
1
vii) tanh x viii) sechx tanhx
vii) ax viii)
n+ 1
Xnf l
E13) Evaluate the following integrals by using FTC.
j)
!, (sinh x
 cmh x) dx
E
E14) Find
&[F(x)] when F(x) is defined by thc following definite integrals.
In .this unit we have covered the following points:
1) A partition P of a closed interval [a,b] is'a set {a = XO, XI, x2 ,,...,, x,
x, = b)
2) A partition PI of [a,b] is finer than a partition P,, if PI 2 P2. 3) If M and m are the supremum and the infimum of a bounded function f i n [a,b], then, given any partition P of [a,b], m(ba) IL (P,f) a U (P,f) 5 M(ba). 4) The lower integral of a bounded function ik less than or equal t&ts upper integral. 5) A bounded function f is integrable over [a,q.if and only if its lower and upper integrals are equal, In such a situation the lower (or upper) integral is called the definite integral off over [a,b], denoted by
b
1
b
a
f (a) dx.
6) If f is monotonic or continuous on [a,b], then f is integrable over [a,b]. 7) If f is continuous on [a,b], then /f(x) dx represents the signed area of the
a
region bounded by the curve y = f(x), the xaxis and the lines x = a and x = b.
8) If f is monotonic on [a,b], then
b
.lf(x)dx = lim h
a
b 0
i=1
2
f(a
+ ih),
=
lim h $f(a
L O
+ (il)h),
ba where h = n 9) The Fundamental Theorem of Calculus: i) If f is continuous on [a,b]; then for x E ]a,b[
x
d ( 1 f(t) dt) = ~ x ) dx .a ii) If f is continuous on [a,b] and F'(x) = f(x) for x ff(x)dx = F(b)

]qbl, then
 F(a).
10.6 SOLUTIONS AND ANSWERS
E 1) PI, P, n P2 are regular. A x3 = 114 in PI, Ax3 = 116 in P2.
1 2 1 2 E 2 ) a) ~ o , ~ , T , l , 1 3 , 1 T , 2 } 7 2 1 1 &,6 9 7,8, 3 9) 1 3 b) {2,2,3,4,4T,5T, 4 10 5 10 5 10 ' 5 10 3a+b a + b a + 3 b E 3 ) a) P2 = (a, 4 2 4  ,b) 7a+b 3a+ b 5a+3b , P3 = { a , g ,   ...,bl . ba c) Ax in P2 is 4 *
Ax
in Pg is=
ba 8 '
= 3
E 4) P,*+
Pn+,U P;
P; c Pz+
+P, : +
is a refinement of P,' 4 n.
E 5) a) f(x) is an increasing function on [0,2]. Hence
5 1 and U(P,f) =  t 2. 1 1 3 . 1 + 5 .  1 4 2 2 ' 4 2 2
b) f(x) is a decreasing function on [1,4]. Hence 1 1 1 L(P,f) = .1 + .1 + .1 2 3 4 1 U(P,f) = 1.1 +7.1 + 1
i
I
;
U(P,,f)
=1.1+ 9 2:1=
2 2
L(P2, f) = 4.1 + 9 4
5 2 20 1 1 2 1 4 1  .  + .  +  . 5 4 1 1 4 3 4
 2289 5940 1 1 u(p2,f)=.+4.1+2. 2 4 9 4 5
1+4.1
4 1 1 4
 1691
3960 L(P1,f)
5
L(P2,
f) 5
U(P2,
f) 5
U(P1,
f).
E 7) If P = {xo, ......,x,} is a partition of {0,1), L(P,f) = U(P.f) = Z m,Ax, = 22Ax, = 22Ax, = 21 = 2 Hence

1,
I
*
f(x) dx =
[ f(x) dx
I

= 2.
E 8 ) I f P = {a = x,,.x,. ........., x, = b} is any partition of [ a h ] , then L(P,f) = Zmi.f Axi 5 2 mi, Axi = L(P,g),' where mi.f = inf {f(x) : x E [xi,. xi]} and mi, = inf {g(x) : x E [xi,. xiD and mi., 5 mi., since f(x) 5 g(x) for all x. Similarly, U(P.f) i U(P.g) for all 1'. The result follows. E 9) f(x) = 2 is integrable E 10)
4, b),
e) and g) are integrable as these are continuous.
c), f) are not integrable as they are not bounded. d), h) are integrable as these are increasing functions.
2
E l l ) I(l+x)dx=lim h [ 2 + 2 + , + 2 +
I
110
I
2 3 B+2+T
1 = lim Ti [2n
n41
1 +.Ti ( I
+ 2 + ....... + nI)]
'
E 12)
i) + iv) ii) + v) iii) + viii) iv) + vi) v) + i) vi) + _vii) vii) + ii) viii) + iii)
E 13) a)
*
2
is a n antiderivative of 2x3. Hence
1
IntegralCskulw
b)
$ j jtdt
o
xz
xz
=
dx2 o
j tj
dx2 dt.=
d) dx
d
I
x

d cos t2dt =  [ dx
.
I
0
cos t2dt x2
I
0
cos t2dt]
= 2xcos x4 xz
COS
e)
dx
(It
J
dt) =
dx
d
[I J F F
t
0
xz
J;dt) 0
I Jn
t dt]
UNIT 11 METHODS OF INTEGRATION
Structure
11.1 Introduction
Objectives
11.2 Basic Definitions
Standard Integrals Algebra of Integrals
11.3 Integration by Substitution
Method of Substitution Integrals using Trigonometric Formulas Trigonometric and Hyperbolic Substitutions Two Properties o f Definite Integrals
11.4 Integration by Parts
Integral of a Product of Two Functions Evaluation of /eux sinbx dx and l e a cosbx dx Evah@onot/ ~air'dx. m d x . and
49
/
m
d
x
Integrals of the Type jex[f(x)
+ f (x)] dx
11.5 Summary 11.6 Solutions and Answers
11.1 INTRODUCTION
In the last unit we have seen that the definite integral f(x) dx represents the signed
a
/
b
area bounded by the curve y = f(x), the xaxis and the lines x = a and x = b. The Fundamental Theorem of Galculus gives us an easy way of evaluating such an integral, by first finding the antiderivative of the given function, whenever it exists. Starting from this unit, we shall study various methods and techniques of integration. In this unit, we shall consider two main methods: the method of substitution and the method of integration by parts.'The next two units will cover some special integrals, which can be evaluated using these two methods.
.
Objectives
After reading this unit you should be able to define the indefinite integral of a function evaluate certain standard integrals by finding the antiderivatives of the integrands use the rules of the algebra of integrals to evaluate some integrals use the method of substitution to simplify and evaluate certain integrals integrate by parts a product of two functions.
11.2 BASIC DEFINITIONS
I
t
We have seen in Unit 10, that the antideri;ative of a function is not unique. More
>:I.
L
." *I*
:s s
.A: 
c :"
a
*:AA..a*:..
c
s &L
n
,
s notation here: We shall use the symbol f(x) dx to denote the class of all
antiderivatives of f. We call it the indefinite integral or just the integral off. You must have noticed that we use the same sign 1 , here that we have used for definite integrals in Unit 10. Thus, if F(x) is an antiderh stive of f(x) , then we can write
l
I
f(x) dx = F(x) + F.
This c is called the constant of integration. As in the case of definite integrals, f(x) is called the integrand and dx indicates that f(x) is integrated with respect to the variable x. For example, in the equation
I
(av+ b14 dv =
(av+ b)' 5a
+ c,
( a ~ + bis the integrand, v is the variable of integration, and (av+b)' )~ 5a integral of the integrand ( a ~ + b ) ~ .
+c
is the
You will also agree that the indefinite integral of cosx is sinx c, since we know that sinx is an antiderivative of cosx. Similarly, the indefinite integral of e2" is
+
I
1 e" dx =  e2' 2
x4 + c, and the indefinite integral of x' + 1 i s l (x3f 1)dx =  + x + c. 4 dx is a uniquely defined
You have seen in Unit 10 that the definite integral
red number whose value depends on a, b and the functidn f.
On the other hand, the indefinite integral f(x) dx is a class of functions which
differ from one another by constants. It is not a definite number; it is not even a definite
l
. function. We say that the indefinite integral is unique upto an arbitrary constant.
Unlike the definite integral which dependson a, band f, the indefinite integral depends only on f. All the symbols in the notation f(x) dx for the definite integral have an interpretation.
a
I
b
The symbol reminds us of summatioh, a and b give the limits for x for the summation.
j
f ( x ) dx shows that we are not considering the sum of just the function values, rather we are considering the sum of function values multiplied by small increments in the value of x. f In the case o an indefinite integral, however, the notation f(x) dx has no similar
I
interpretation. The inspiration for this notation iomes from the Fundamental Theorem of Calculus. Thus, having defined an indefinite integral, let us get acquainted with the various techniques for evaluating integrals.
11.2.1 Standard Integrals
Integration would be a fairly simple matter if we had a list of integral formulas, or a tabledintegrals, in which we could locate any integral that we ever needed to evaluate. But thk diversity of integrals that we encounter in practice, makes it impossibleto have such a table. One way to overcome this problem is to have a short table of integrals of elementary functions, and learn the techniques by which the range of applicability of thisshort table can be extended. Accordingly, we build upa table (Table 1) of standard types of integral formulas by inverting formulas for derivatives, which you have already studiqd in Block 1. Check the validity of each entry in Table 1, by verifying that the derivative of any integral is the given corresponding function.
Table 1
I
Methods zf lntegratior
S.No.
Function
Integral
xn sinx
3.
4.
COSX
sec2x cosec2x
5. 6.
7.
8.
1
secxtanx cosec x cotx
,
1
+c tanx + c cotx + c secx + c
sinx
,/GF
tan'x cot'x
1
+ c or +c
sin hx coshx sech2x cosech2x sechx tanhx cosechx cothx
coshx + c sinhx c tanhx c
+ +
Now let us see how to evaluate some functions which are linear combinations of the functions listed in Table 1.
11.2.2 Algebra of Integrals
You are familiar with the rule for differentiation which says [ af(x) + bg(x) 1 = a d [f(x)l+ b$ [g(x)l dx dx There is a similar rule for integration :'
Rule 1 [af(x) + bg(x)]dx = a /f(x)dx + b lg(x)dx This rule follows from the following two theorems. Theorem 1 Iff is an integrable function, then so is kf(x) and
/
/kf(x)dx = k /f(x)dx F(X) + c. d [F(x)+c] = f(x) Then by definition,dx [k{F(x) + c)] = kf(x) ' dx Again, by definition, we have
~ o o Let f(x)dx f
]
=
"
I
I
Vc"c'
Theorem 2 Iff and g are two integrable functions, then f+g is integrable, and we have
Roof Lct f(x)dx = F(x) + c, g(x)dx = G(x) Then, [{F(x) + c} + {G(x) + c}] = f(x) + g(x) dx
Thus, [f(x) + g(x)]dx = [F(x)
1
\
+c
\
+ c] + [G(x) + c]
+
= f(x)dx
\
1g(x)dx
Rule (1) may be extended t o include a finite number of functions, that is, we can write
= kl
1fl(x)dx + k21f2(x)dx + ............. + k, Ifn(x)dx
\
+
We can make use of Rule (2) to evaluate certain integrals which are not listed in Table 1. 1 Example 1 Let us evaluate (x + x)3 dx We know that (i + ;l3 = x3 3x +
4+ 1.
x3
Therefore,
......... 2 Rule
Using integral formulas 1 and 11from
able 1, we have
+ 3% + 3c3 + c4 has been replaced by a single arbitrary constant c. Example 2 Suppose we want to evaluate (2 + 3sinx + 4ex) dx
Note that cl This integral can be written as
Note that \dx = \ldx = \xOdx = x
+c
1 0
Example3 To evaluate the definite integral fix + 2x )2 dx, we first find the indefinite
integral (x Thus, /(x
j + 2x2)2 dx.
+ 2x2)' dx = \(x2 + 4x3 + 4x4) dx
According to our definition of indefinite integral, this gives an antiderivative of ( x + 2 ~ ~ ) ~afgiven value of c. By using the Fundamental Theorem of Calculus we can or now evaluate the definite integral.
1
Note that for the purpose of evaluating a definite integral, we could take the antiderivative corresponding to c = 0,that is,
I
m says that if G (x) is an C h antiderivative of f(x), then

f (XI=G dx
(41:
= G (b)  G(a).
4 1 x3 + x4 +  x 5 ,as the constant cancels out. 5 3 See if you can do these exercises now.
E
E l ) Write down the integrals of the following using Table 1 and Rule 2. a) (i) x4 (ii) xP3l2 (iii) 4x' b) (i) 1  2x + x2 c) (i) ex + e* (ii) (x  1 l 2 x (iv) 3 (iii) (1 + x ) ~
'
+4
(ii) 4cosx  3sinx
+ ex + x
(iv) 4sech2x + ex
 8x
Integral Cakulns
E2) Evaluate the following definite integrals.
6
a) (i) \x4 dx
5
(ii)
1 5
1
2
dx
You have seen that with the help of Rule 2 we could evaluate a number of integrals. But still there are certain integrals like sin2x dx which cannot be evaluated by using Rule 2. The method of substitution which we are going to describe in the next section will come in handy in thes%cases.
I
11.3 INTEGRATION BY SUBSTITUTION
In this section we shall study the first of the main methods of integration dealt with in this unit: the method of substitution. This is one of the most commonly used techniques of integration. We shall illustrate its application through a number of examples.
11.3.1 Method of Substitution
The following theorem will lead us to this method. Theorem 3 If \f(v)dv = F(v)
+ c , then on substituting v by g(x), we get
/4g(.)l g' ( 4 dx = If(v)dv Roof We shall make use of the chain rule for derivatives (Unit 3) to prove this theorem. Since If(v)dv = F(v)
+ c, we can write
= f ( v ) Now if we write v as a
function of x, say v = g(x), then
= fIg(x)]

since v = g(x)
= f[g (x>l . g'(x) This shows that F[g(x)] is an antiderivative of f[g(x)gl(x). This means that
The statement of this theorem by itself may not seem very useful to you. But it does dx. simplify our task of evaluating integrals. For example. to evaluate J s i n ~ x we could take v
= g(x) = =
sin 2x dx
1 2
I sin2x (2) dx
2x and get
=
I 2
I
2
sinv dv by Theorem 3. since g(x) = 2r and g'(x) =2.
+
  cosv  
2 We make a special mention of the following three cases which follow from Theorem 3.
Case i) If f(v) = v", n f 1 and v = g(x),, then, by Formula 1 of Tilble I.
cos2x
+
Case ii) If f(v) = l/v and v = g(x).
then, by Formula 11 of Table 1
Case iii) 1f f(x) dx = F(r)
I
+ c. then
1
h
g(h)
flg(x)l gt(x)dx=l f(v) dv, where v = g(x) (The limits of integration are g(a) and g(b)
g( ;I)
1~ ( h )
since x = a = > v = g(x) = g(a), and x = b = > v = g(x) = g(b). ] We shall be using these three cases very often. Their usefulness is evident from the
Example 4 Let us integrate (2x + l)(x2+ x + 1) For this we ob&rve that d (x2 +x+ 1) = 2x+ 1 dx
'

Thus, (2x+1) (x2+x+1)\x evaluated as in i) above.
1
is of the form Mx)]" gr(x)dx and hence can be
I
Therefore, /(2x+l) ( ~ ~ + x + l ) ~ d1x(x2+r+l)6 + c. = Alternatively, to find [(2x+l) (x2+x+l)' dx, we can substitute x2+x+l by u. This means
 = 2x + 1 d" dx
Therefore, \(2x+ 1) (x2+x+ 1)' dx = us du by Theorem 3
=1 u6
I
6
+c
by Formula 1 from Table 1.
E *
s k t us evaluate k a x + by dx.
f(ax + b)" dx =
$ I a(ax+b)" dx
I
d (ax+b) dx
= a l(ax+b).
Therefore, when n f
 1,
and when n =  I ,
(ax+b)" dx =
dx
= In a
1
Jax+bl+ c
Example 6 Suppose we want to evaluate the definite integral
We put x2+2x+3 = u. his implies
a = 2(x+1). Further, dx
II
I1
when x = 0 , u = 3, and when x = 2, u = 11. Thus,
d x = l 0I lddx d x = T1I T idui l n l u l ] U iX2+2x+l 2 ~
x+'
3
2
2
.
3
du Example7 Toevaluate xe2"dx, we substitute 2x2 = u. Since dx = 4x, we can write,
\

E
Methods oflntepration
E3) Evaluate a) J',~x3 dx
b) J(2x+l)"x
C)
J&
1
3
Now we shall use the method of substitution to integrate some trigonometric func6ons. Let's start with sin ax.
Example 8 To evaluate
I
I
sin ax dx, we proceed in the same manner as we did for
s n 2x h.WL make the substitution ax = u. i
This gives
a = a. Thus, dx
E
a E4) Proceeding exactly as in Example 8, fill up the blanks in the table below.
=  cosax+c 1
S.No.
1.
f(x)
I
f(X) dx
sinax
COS~X
1  cosax
2.
3.
4. 5.
sec2ax
. cosec2ax
secax tanax cosecax cotax eax amx
6.
7. 8.
........ ........ ........ ........ ........
+c  sinax + c 1 a
a
........
Example 9 Suppose we want to evaluate
i) lcotx dx, i) ii) ltanx dx and iii)
I
cosec2x dx
We can write
I I
,
cotx dx =
I
cosx dx. Now, since d sinx = COSX., integral falls in the category this
of case ii) mentioned earlier, and thw, ii) T o evaluate tanx dx =
= In
I
cotx dx = In s i n x
+c
I
tanx dx, we write
secx tanx dx
,,,
d Isecx) + c, as  secx = secx tanx dx iii) T o integrate cosec2x we write Jcosec~x = dx 2
1
J 2cosec2x(cosec2x cot2x) dx cosec2x  cot2x
d Here again,  (cosec2x cot2x) = 2cosec2x(cosec2x cot2x) dx 1 This means jcosec2x dx =  ln (cosec2x  cot2xl + c 2
In this example we have used some 'tricks' to put the integrand in some standard form. After you study various examples and try out a number of exercises, you will be able to
decide on the particular substitution or the particular trick which will reduce the given integrand to one of the known forms. Let's look at the next example now.
Example 10 Let us evaluate
1eGn2'sinZx
1eudu
dx
If we put sin2x = u, then
Therefore. eanZx sin2rdr =
1
a = 2sinx cosx = sin2x dx
= eU
+ c = esin2xCc
See if you can solve this exercise now.
E
E5) Evaluate the following integrals
a) j s e c x dr
b)
1sin2x cosx dx
0
lrl2
c)
1etanxsec2xdr
I
I
lntegrrlCakuhw

'
11.3.2 Integrals using Trigonometric Formulas
In this section, we shall evaluate integrals with the help of the fauowing trigonoq@!ric formulas:
'1 cos2x =  (1 2 sin"
+ cos2x)
3 1 =  sinx  sin 3x 4 4
4
4
3 cos3x =  cosx +1 cos 3x
1 sinmx cosnx =  [sin(m+n)x
2
+ sin(mn)x]
1 cosmx cosnx =  [cos(m+n)x 2
1 sinmx sinnx =  [cos(mn)x 2
+ cos(mn)x]
 cos(m+n)x]
In each of these formulas vou will find that on the left hand side we have either a Dower of a trigonometric functi6n or a product of two trigonometric functions. And dn the l right hand side we have a sum (or difference) of two trigonometric fuslctians. You d realise that the functions on the right hand side car& easily'integrated by making suitable substitutions. The f o M n g examples d l illustrate how we make we of theabove formulas in evaluating certain integrals.
,
Example 11 To evaluate cos3ax &. We write
Example 12 Let us evaluate i ) jsin3r c d x dx and ii) lsinx s i d x sin3r dx
Here the integrand is in the fonn of a product of trigonometric functions. We shall write it as a sum of trigonometric functions so that it can be integrated easily.
i)
jsin3x d
x dx =
I
(sin7x  sinx) dx
ii) To evaluate Jsim sin211 d x dx, again we express k c pmduct sinx sb2x &X as a sum of trigonometric functions. sinx sin2x sin3x = 1sinx (cosx  codx)
2
1 =  sinx cosx  sinx codx I 2. 2
Therefore, lsinx sin2xsin3x d r
1 1 cos2x  cos4x + cos6x + c 16 , 24 Try to do some exercises now. You will be able to solve them either by applyirig the trigonometric formulas mentioned in the beginning of this section or by using the method of substitution. Don't be scared by the number of integrals to be evaluated. The more integrals you evaluate, the more skilled you will become. You have to practise a lot to be able to decide on the best method to be applied for evaluating any given integral.
= 3
E
E6) Evaluate each of the tollowing integrals, a) i) jsin5xcosx dx ii) cosx
ax
iii)
j cot2x coxc22x dx
rr16
iv) Isin20 ec0s2ed0
v) 7:in0(1 +cos40) d0 ii)
b) i) J ( l + c o s ~ ) ~ s de ~ in
1114
I
sec2ede
iii)
0
I sec0 tan0(l+sec0)~d0
1112
o (1  ~ t a n 0 ) ~
c) i) ls!n40 do iii)
0
ii) !sin30 cos0 d0
rrl2
cos50 cos0 d0
iv) jcos0 cos20 cos40 d0
O
Integral Calculus
E7) The cost of a transistor radio is Rs. 7001. Its value is deoreciating with time
= where Rs.v i~ its value t years after its dt (l+t2) purchase. What will be its value 3 years after its purchase? (Don't forget the constarit of integration! Think how you can find It with the help of the given information.)
'
according to the formula
*
a,

1
I
11.3.3 Trigonometric and Hyperbolic Substitution
Various trigonometric and hyperbolic identities like sin28 + cos20 = 1,
1 + tan2@ sec26, tanhe = =
and so on, prove very useful while evaluating certain
i
1
integrals. In this section we shall see how.
/
1
A trigonometric or hyperbolic substitution is generally used to integrate expressions
involving
Methodsof Integration
d m ,, & &?
or a2+x2. We suggest the following substitutions. Substitution
Expression involved
JzF
a2+ x2 ~ h u sto evaluate ,
3
&z
x = a sine
x = a tan0 0r.a sinhe x = a sece or a coshe
x = a tan0
Jz7
dx
, put x = a sine. Then we know that
$$ = a cose. This means we can write
1
= sin' (xla)
+c
we shall put x = atan0
similarly to evaluate
I & ,
a2+x2
i
Since
= asec2e do, we get
C
t
= 1 tan' (xla)
+c
dx
We can also evaluate This eives
, by substituting x = atane
dX = asec20
We can also evaluate this integral by putting x = asinhe. With this substitution we get,
1
i
J Ja'+xL
[ , dx 
= sinh'(xla) = In.
+ c, and we know that
a (see "nit 5)
sinh'(xla)
+
J JxLaL
= In.
I
x
1
+ J Z F +c a
+C
and
dx x&2
=  KC'(x/li)
a
Integral Calculus
Let us put these results in the form of a table.
Table 3
Sometimes the integrand does not seem to fall in any of the types mentioned in Table 3, but it is possible to modify or rearrange it so that it conforms to one of these types. We shall illustrate this through some examples. dx Example 13 Suppose we want to evaluate I
I
2
J2xx2
Let us try to rearrange the terms in the integrand
2 2
j z 7
to suit us. You will see that
dx

dx
! J 2 2 
!
J
m
1 and
Ifweputx1 = v , & = dx
2
. Note the new limits of integration.
This integral is finally in the form that we want and using the first formula in Table 3 weget
2
 0 = iF . 2 2 x2 dx does not again Example 14 The integrand in l+xh fall into the types mentioned in Table 3. But let's see what wetan do.
= sin'l
 sinI0 =
I
7 Ifwe put x = u,
*
dx
1
=
3x2, T ~ U S ,
I
x2
1
I
 1
1:u2
du, by Theorem 3
Here the integrand Thus, we get
can be evaluated using formula 2 in Table 3
1+u2
If you have followed this discussion, you will certainly be able to solve this exercise.
E
t
E8) Integrate each of the following with respect to the corresponding variables.
i)
1
1
t2 
iii)
1
1+4x2
iv)
1 
J9X29X2
X
2x2+5
vi)
t6+ 16
vii)
u2 ,/=
1
viii)
JzP X' xi) (Hint:. &
I +x2'
ix)
1
= I  _) 1
I+x2
I +x
r
Integral Calculus
11.3.4 Two Properties of Definite Integrals
We have already derived some properties of the definite integrals in Unit 10. These are the i) Constant Function Property : [cdx = c(ba)
ii) Constant Multiple Property : iii) Interval Union Property iv) Comparison Property
:
/ k f(x) dx k If(x) dx / f(x) dx / f(x)dx +'If(x) dx
=
c a
a b
h
a
b
b
=
a
u
c
:~fcsf(x)sd+x~[a,b],
then, c(ba) r f(x) dx s d(ba)
/
b
.
Now we shall use the method of substitution to derive two more properties to add to this list. Let's consider them one by one. v) \f(x) dx = jf(x) dx + f(ax) dx for any integrable function f.
0
1
0
0
We already know that
a
d2
/f(x) dx = f(x) dx + f(x) dx .
0 0
d2
\
0
/
a
Now if we put x = ay in the second integral on the right hand side, then since
/ f(x) dx =  / f(ay) dy = /f(ay)
a12
a12
a
ai2
a12
dy = f(ax) dx,
1)
/
'
0
since x is a dummy vaiiable. ~ h u r / f ( x ) d x= /f(x)dx
0
+ / f(ax)
0
dx.
(I
Example 15 Let us evalozte
i)
i)
I
0
Met h a s of Inlel(rution
sin4xcos5xdx and
ii) lcos3x dx
0
Using property v), we can write
sin ( a  X ) = b i n x. and cos ( a X ) = COSY

=2
cos'x dx
0 n12
=
2 [ I CO$X dx + cos'fnx) dx]
0
n12
0
=
2 [ I cos'x dx
I '
I I
n12
nll
cos'x dx ] = 0
0
Our next property greatly simplifiessome inregrals when the integrands are even or odd functions. vi) I f f is an even function of x, then
and i f f is an odd functicm, then,
This is also obvious from Fig. I (a) and (b)
Pig. I
We shall piove the result for even functions. The result for odd functions follows easily and is left to you as an exercise. (see E 9) c)).
So let f be an even function of x in [a,a], that is, f(x) = f(x) V x c [a,a]. Then,
If we put x = y in the first integral on the right hand side, we get
0
0
a
f(x) dx
a
 lf(Y)
a
a
d = jf(y) dy = ff(x) dx. Y
0
0
Thus,
a
f(x) dx = 2 lf(x) dx .
0
Using this property we can directly say that
Ty to gdve this exercise now. r
E9j a) Evaluate
t
b) show that sin& in (tan x) & = 0
U
l
C)
Prove that laf(x) dx = 0 iff is an odd function of r.
a
In this section we have seen how the method of substitution enables ys to substantially increase our list of integrable functions. (Here by "integrable function" we mean a function which we can integrate!) We shall discuss another techniaue in the next section.
11.4 INTEGRATION BY PARTS
In this section we shall evolve a method for evaluating integrals of the type
1
u(x)v(x) dx, in which the integrand u(x)v(x) is the product of two functions. In other
words, we shall first evolve the integral analogue of d d  [u(x)v(x)] = u(x) d v(x) + v(x) u(x) dx dx and then use that result to evaluate some standard integrals.
11.4.1 Integrals of aProductof Two Functions
We can calculate the derivative of the product of two functions by the formula d d  [u(x)v(x)] = u(x) d v(x) + v(x)  u(x) dx dx Let us rewrite this as d d d / u(x)  v(x) =  [u(x)v(x)]  v(x) u(x) dx dx Integrating both the sides with respect to x, we have
a
I
d d U(X); (v(x) ) dx = u(x)v(x)  v(x) ((ux) ) dx ........... (1) T o express this in a more symmetrical form, we replace u(x) by f(x), and put
1
d  V(X)= g(x). This means v(x)
= g(x) dx. dx As a result of this substitution, (1) takes the form
I
This formula may be read as : The integral of the product of two functions = First factor X integral of second factor  integral of (derivative of first factor x integral of second factor) It is called the formula for integration by parts. This formula may appear a little complicated to you. But thqsuccess of this method depends upon choosing the first factor in such a way that the second term on the righthand side may be easy to evaluate. It is also essential to choose the second factor such that it can be easily integrated. The following examples will show you the wide variety of integrals which can be evaluated by this technique. You should carefully study our choice of first and second functions in each example'. You may also try to evaluate the integrals by reversing the' order of functions. This will make you realise why we have chosen these functions the way we have.
Example 16 Let us use the method of integration by parts to evaluate xe' dx.
In the integrand xeXwe choose x as the first factor and ex as the second factor. Thus, we get .
I

Example 17 To evaluate x2cos? dr, w e shall take x2 as the first factor and emx as the second. Let us first evaluate the corresponding indefinite integral.
0
j
I
x2cow dx = x2 cosx dx G
1
I{
d  (x2) C O S X ~dx ) X dx
I
x2 sinx  2xsinx dx
= x2 sinx  2
1
xsinx dx
We shall again use the formula of integration by parts to evaluate xsinx dx. Thus, jxsinx dx
= x(cosx)
= xcosx = xcosx
I
 j (1) (cosx)
dx (f(x) = x, g(x) = sinx)
+ cosx dx + sinx + c
I
Hence, x2 cosx dx = x2sinx + 2xcorx  2rinx + c Note that we have written the arbitrary constant as c instead of 2c.
nl2
NOW
1
0
x2cosx dx = (x2sinx + 2xcosx  2sinx + e )
Example 18 Let us now evaluate x in 1x1 dx Here we take In(xl as the first factor since it can be differentiated easily, but cannot be integrated that easily. We shall take x to be the second factor.
I
While choosing Inlxl as the first factor, we mentioned that it cannot be integrated eqsily. The method of integration by parts, in fact, helps us in integrating lnx too.
Ex
19 We can find lnx dx by taking lnx as the first factor and 1as the semad
j
factor. Thus,
I
lnx dx =
1
(lnx)(l) dx
 J dx = xlnx  x + c = xlnx  xlne + c since lne = 1
= xlnx = xln(x1e)
+c
The trick used in Example 19, that is, considering 1 (unity) as the second factor, helps us to evaluate many integrals which could not be evaluated earlier. I You will be able to solve the following exercises by using the method of integration by parts.

E10) Evaluate
a)
b)
Ix21nxdx
'ake Take f(x) = 1 +x and g(x) = ex
j (1 +x)ex dx
r
d)
I x2sinx cosx dx
Take f(x) = x2 and g(x) = sinx cosx sir
I
E
E l l ) Evaluate the following integrals by choosing 1 as the second factor
E12) Integrate
a) xsin'x
b) 1n(l+x2)w.r.t.x.
11.4.2 Evaluation of
.' dx and S eaxcosbxdx I eaxsinbx
h l r lhedr or 111tt%riltih1
To evaluate re" sinbx dx a n d l eaxcosbx dx, we use the formula for integration by parts.
1 1 J' e?"sinbx dx = (eax)(   cosbx)  I (aea") (   cosbx) dx b b
=
1   e a xcosbx
b
1 b
+ 2 I e'"cosbx b
dx
=   eaXcosbx
1 + 2 [(em) ( b sinbx) b
I
(ea"
:
sinbx) dx
b b2 b2 Therefore, you will notice that the last integral on the right hand side is the same as the integral on the left hand side. Now we transfer the third term on the right to the left hand side, and obtain, (1 +
)

=
1 a2   e"xcosb~ aeaxsinbx   edxsinbxdx +
I
b2
1ea'sinbx dx
= e'"
1 ( a sinbx   cosbx) b2 b
This means,
I I
enX eaxsinbxdx =  (asinbx  bcosbx) a2+ b2
+c
We can similarly show that eaxcosbxdx =  (acosbx + bsinbx) eax a2+b2
+c
If we pht a = rcos0, b = rsin0, these formulas become
1
eaxcosbxdx =
J
m
ea"cos (bx0)
+ c, where 0 = tan I . b
i l
Example 20 Using the formulas discussed in this subsection, we can easily check that i) l e x s i n x d x = 1 e ' s i n ( ~  ~Tr + ~ )
and
J
Example LI To evaluate c" sinx cos2x dx, wc shall first write 1 sinx cos2x =  (sin3x  sinx) as in Sec. 3. 2 Therefore,
I
I
e2"sinx cos2x dx
=I e"sin3x 2
dx 
e" sinx.dx
Mow the two integrals on the right hand side can be evaluated. We see that .
1
eixsin3x d r = e2"sin (3x  tan'
and Jezxsinx dx = Hence
fl
2 )+ c
2
fi
e2"sin (x  tan'
f) + c'

\e2%inx cos2xdx = e2"[
D
sin (3x  tan';)
1 (X tan' sin
Js
t )1 +
c
Epie
22 Supp0se.w want to evahate x3 sin (alnx) dx.
I
Let lnx = u. This implies x = eU dutdx = llx. and
Then, x3 sin (alnx) dx = x4 sin (alnx) (llx) dx
=
1
I
J eb sinau du
J~GF
e4' sin (au
=
 tan'(al4))
+c
Why don't you try some exercises now.
E
E13) Evaluate the following integrals a) e 2 ' d x dx
b)
1
e3'sin3x dx
c)
1
$'COSX
cos2x dx
d)
f)
I
e2'cos% dx xeasinbx dx
e ) lcashax sinbx dx (write cashax in terms of the exponential function)


11.4.3 ~valuntion of
S
dx,
$
dx, and
dx dr
In this subsection, we shall see that integrals like and
1
1Ja2I 1
dr,
dx can also be evaluated with the help of the formula for integration by
parts and Table 3.

JaLxL
Shifting the last term on the right hand side to the left we get, 21dx
=I+
dx
a2j 1 dx2 4 a x2
Using the formula,
,I
= s i n 1( f )
+ c, we obtain
Similarly, we shall have,
Example23 ~ eustevaluafe
1 0
dx
Let x
+ 3= u, Then,
3/2
/,/=dx
=
1du
E14) Verify that
a n a) I J Z 2 d x = ~ x J Z G ? +2 ~ ~ X + Ja'+x'+c
In the next subsection we shall consider another type of integrand which occurs quite frequently in mathematics.
11.4.4 Integrals of the ~~k [ f ( x ) ex
We first prove the formula ex [f(x)
I
+ f'
(x) ] dx 
+ fr(x)]dx = ex f(x) + c and see how it can be used in integrating some
functions. By the formula for integration by parts
J exf(x) dx = J f(x) ex dx = f(x) ex  J r(x) ex dx + c
This implies
Example 24 Let us evaluate the following integrals.
Methods of Integration
We take up (i) first,

1  ex+ c,
2+x
since
Now ii)
* shall evaluate ii)
1+cosx
X X cos   sin 2 ed2 dx
'
Now
1
sec
e&dx = 2
(set?) (2eR) 
I
(? sec
1
x tan ?) (2e") X
dx
= 2secA
e"Q + sec
J
r
tan
$ e'"
dx
Thus,
= set X e*
2
+z
1%
2
tan
$ dx  I 2
sec
X
tan
$ ed2cIx
In this unit we have exposed you to various methods of integration, You have also had a fair amount of practice in using these methods. We are now giving you some additional exercises. You may like to try your hand at these too. T o solve these you will have to first identify the method which will suit the particular integrand the best. This is the crucial step. The next step where you apply the chosen method to get the answer is relatively easy. If you have studied this unit thoroughly, neither of these steps should pose any problem. So, good luck! E15) EvaluateJhe following integrals
f, I
(x2+2)" dx
g)
1
sinx ecoSx dx
IntegralCalculus
Methodsof lntcgretlon
E16) Prove that
and use it to evaluate x3 ninx dx
Before we end this unit, here are some general remarks about the existence of integrals. The result
b
where F(x) is the antiderivative of f(x), will make sense only if f(x) exists at every point of the interval. Hence we have to be careful in using this result. Thus,
But llx is not defined at x = 0, and In 1x1 is also not differentiable at x = 0. As such, at this stage, we should use the result only if the interval [a,bl does not include x = 0 Thus,
2

l+dx=ln
11 1
IZ1 
In 2 is not valid.
Again, consider
1
We have used this in Example 3. However
JTsr;Tr
does not exist at x = 1, and sin'x is
not differentiable at x = 1. sin'x)' exists at x = 1,but sinL'x itself does not exist when x > 1.
sin'x)' does not exist, since
However, the above result is true in some sense. This sense will be elear to you in your
course on analysis.
The antiderivative of every function need not exist, i.e., it need not be any of the functions we are familiar with. For example, there is no function known to us whose
b
derivative is eeX2. However, the value of the definite integral f(x) dx of every
a
j
function, where f(x) is continuous on the interval [a,b], can be found out by numerical methods to any degree of approximation. You can study these methods in detail if you take the course on numerical analysis. You will study two simple numerical methods in 2 Block 4100. Thus, we cannot find the antiderivative o e" ,but still, we can find f the approximate value of
b
e" dx, for all real.values of a and b. In fact, this integral is every important in
a
probability theory and you will use it very often if you take the course on probability and statistics. That brings us to the end of this unit. Let us summarise what we have studied so far.
1.5
SUMMARY
In this unit we have covered the following points. 1) If F(x) is an antiderivative of f(x), then the indefinite integral (or simply, integral) of f(x) is
I
f(x) dx = Fix)
+ e, where e is an arbitrary constant.
k,
b
I
f,(x)dx + k2 f2(x)dx + ..... + kn fn(x)dx
I
1
3) 'The method of substitution gives :
]f[g(x)] g'(x) dx =gr:(u) du. if u = g(x).
a
,
da)
In particular, \[f(x)l.ff(x) dx = 'f(x)ln+' n+l
+ c, n +  1, and
x) dx, iff is even
a
0, iff is odd.
4) Standard formulas :
5) Integrationof a product of two functions (integration by parts) :
I
U ( X ) V ( X )=X ~ y(x)
I
v(x)dx 
{u'(x) v ( x ) ~ xdx )
I
This leads us to:
I
eaxsinbx d r =
vL2+ b2
,
eaxsin (bx  tan'
z)+
c
\ eaxcosbx dx =
1
a2+ b2
e" cos (bx  tan' b) + c a
11.6 SOLUTIONS AND ANSWERS
E 11 a\ i1
&+r
ii\ 3x'I2
A c
iiii
A
, 
1

iii) x + x + x 3 2
+3+c 4
c) i) ex  e"
iii) d) i) 2sin'x
+ 4x + c ii) 4sinx + 3cosx + ex + 2 + L 2 4 tanhx + ex  4x2 + c
+ 51nlxl + c
= cotx
+ tanx
 2x
+c
3x2 4 iii) 6x +   x3/2   xsn 2 2 3 5
+c
b) i)
12
ii) 15
S.No.
1.
f(x)
sin ax
cos ax
1Rx)
a a
dx
 cos ax + c 1 a
2.
3.
4.
 sin ax + c 1
1  tan ax + c
secZax cosec2 ax sec ax tan ax
COS~C ax
l a
cot ax
+c
5.
1  sec ax +s. a
6
7.
8.
cot ax
 1 cosec ax + c
1  eax + c a
e ax amx
1 amx  +
m
Ina
b)
0
7
sin" sin2x cosx dx = ]=I2  1
C) I f u = tanx,
I
a dx

0
secZx
=
+ J eta"" sec2x dx
E6) a ) i ) M + c 6
n/3
I
e" du = e"
+ c = eta"' + c
i
I
I
I
ii) 2 + c sin2x
n/3
1 iii) cot2x cosec22x dx =  cot2x (2cosec2 2x) dx 2 .n/6 n/h
I
I
iv) Put cos20 = u. Then
*
de
= 2sin20
*
1 ii) 10 (I5tan0)
= ~ / d 0  ~ 1 c o s 2 0 d 8  ~ 1 d0 8 0 cos20
1 + 8/ c o d 0
d0
1 = 2 . g  T T sin20 8
=1 0 (2
, 8 sin40 4
1
+
4
2
1  sin 20 +  sin40) + c 8
ii)
1
sin30 cos0 d0 =
1[lsir~0 d0
+ Jsin20 do]
iii) lcos50 core dB = sin40
0
WR
sin60 12
. iv) lcose coS20 c o d 0 dB  19
0
v(0) = 700 = 500 tand10+ c = c *c=700. v(3) = 700  500 tan'3.
E8) For solution, see P. 104
w WR 12
b) lsin2x In tanx dx = fsin2x h tanx dx
0 0
= jrin2x In tanx dx
0
+ j sin2x In c o b dx
0
I
I
EIO) a) Ix21nxdx
=
lnxIx2dx  ] ( + I
x2 dx) dx
) I,
EIZ) a) lxsin'x dx =
1 x2 2sin'x   I dx 2
./COSU
Put x = sinu in
x2 I JI;;T d x = I S B m S u d u
 7 u   s i n 1u + c = 1 2 4
= 1 [sin',
1 u  1 sinu cosu 2
+c
2
 xcos(sin'x)]
+c
.'. lxsin:'x
dx
1  x2  2sin'x   [sin' x  x m ] 4
+c
b) I1n(l+x2) cix =
I 1.1n (1+x2)

dx
= xln (l+x2) = xln (1+x2) = xln (l+x2)
1+x2
cix
 ]2[1
1l +1 2  x
dx
 2[x  tan'x] + c
El3
a) 1 2 x(2cos4x
+ 4sin4x) + c
1 2
c)
I
e4' cosx m s h dx =
1
e4x (COS~X coax) cix +
=L{Le2. (2cos~x+ ~ s i n ~ + )1 e2'] 8 x 1
e)
+c
\ cosh ax sinbx dx \ ( 1[ \ eY sinbx dx + \ ea 2
= =
=
eax+eax
) sinbx dx
sinbx dx]
(a2+b2)
f)
[em (asinbxbcosbx)
=
+ e (asinbxbcosbx)] + c "
\ xeaxsinbx dx
4 eax sinbx dx
eY (asinbx
eU (asinbx
\  \
 beosbx) dx

a2+b2
 bcosbx)
beY(acosbx+bsinbx) ]
+c
h)
! tan'
3
(3x)
+c
=
i, i
sinx cosx dx ~ s
i
[
dt~
=
I[ j t7 dt  2 j t6 dt] 2
=
 0 cos0 + j
a
c o d d0 (integration by parts)
m) Put x = tan0 in
1 j ms' ()1x2 +x2
dx ] +c
Answer = 2 [ 0 tan0 n)
+ lnlcos0l
=
1ex (lnsinx + cotx) dx I ex In sinx dx + ex cotx dx = In sinx ex  I cot; ex dx + ex cotx dx
= ex In sinx.
d2 j x 3 s i n x d x = l x3 (sinx) dx dx2
=
=
 x3cosx + 3x2 sinx
 x3cosx
6
1x sinx dx
xcou
 x3cosx + 3x2 sinx  6 [
+
1cosx dx]
=
+ 3x2sinx + 6(xcosx  sinx) + c
UNIT
Structure
REDUCTION FORMULAS
12.1 Introduction 12.2 Reduction Formula 12.3 Integrals Involving Trigonometric Functions
Reduction Formulas for Reduction 'Formulas for
\
sinnxdr and
I
tan? dx and secnxdx
/ /
cos"x dx
12.4 Integrals Involving Products of Trigonometric Functions
Integrand of the Type sinmxcos"x Integrand of the Type en' sinnx
75
79 80 81
12.5 Integrals Involving Hyperbolic Functions 12.6 Summary 12.7 Solutions and Answers
In the first two units of this block we have introduced the concept of a definite integral and have obtained the values of integrals of some standard forms. We have also studied two important methods of evaluating iv.!,e.&ak, namely, the method of substitution and the method of integration by parts. In t h & d ~ p i o of many physical or engineering n problems, we have to integrate some integiands involving powers or products of trigonometric functions. In this unit we shall devise a quicker method for evaluating these integrals. We shall consider some stan'dard forms of integrands one by one, and derive formulas to integrate them. The integrands which we will discuss here have one thing in common. They depend upon an integer parameter. By using the method of integration by parts we shall try to express such an integral in terms of another similar integral with a lower value of the parameter. You will see that by the repeated use of this technique, we shall be able to evaluate the given integral.
Objectives
After reading this unit you should be able to derive and apply the reduction formulas for
jsinnxdx , /cosnxdx, /tannxdx ,etc. \sinmx eosnxdx
12.2
REDUCTION FORMULA
Sometimes the integrand is not only a function of the independent variable, but also depends upon a number n (usually an integer). For example, in /sinnxdx, the integrand i n n x depends on x and n. Similarly, in excosmx dx, the integrand exmsmx depends on x and m. P,c numbsh n and m in these two examples are called parameters. We
L11
f
>:

I.
:*
 
4
LA
On integrating by parts we sometimes obtain the value of the given integral in termsof another similar integral in which the parameter has a smaller value. Thus, after a number of steps we might arrive at an integrand which can be readily evaluated. Such a process is called the method o successive reduction,and a formula connecting an f integral with parameter n to a similar integral with a lower value of the parameter, is called a reduction formula. Definition I: A forn1ul:r ot the form jr(s.n)dx = g(x)
Reduction ~nrmul&
'Ihrough a reduction formula we duet the value of the parameter.
+ jf(r.~)dx.
where k < n, is called a reduction formula. the following ex;rmplc as an illustr:rtion. Example I The integrand in is the exponent of x. 1.ct I,, = rne'dx. Integrating this by parts, with sf'as the first function and ex as the second function
~ivc4 11s
/X~IC'JI
depends on x and also on the parameter n which
I
I,,
=
snlr'd\  /(n
\"I
( c'dr)dx
e'dx Note that the integrand in the integral on the right hand side is similar to the one we started with. The only difference is that the exponent of x is n I. Or, we can sa) that the exponent of x is reduced by 1. Thus. we can write I,, = xne'  n I,,, . .. .. . ...., .. (1) The formula (1 ) is a reduction formula. Now suppose we want to evaluate 14, that is, Ix4e'dx. Using (1) we can write I, = xJe'  41,
 xfle'  n 
J xnI
 4 [x3e'  312]using (1) for 13. = xJeX 4 x%' + 12 l2 = x4ex 4 u"e' + 12x2ex 24 I , , using (1) for I? a= x4e'  4 x'e' + 12x'eX  24xeX+ 24 I,,.
= xJe'
NOWI,, = /x"exdx
=
jexdx = e x + c.
Thus, the method, of successive reduction gives us
I
in five simple steps. You must have noted that we were saved from having to integrate by parts four times. This became possible because of formula (1). In this unit we shall derive many such reduction formulas. . These fall into three main categories according as the integrand i) is a power of trigonometric functions.
ii) is a product of trigonometric functions, and
iii) involves hyperbolic functions. We will take these up in the nexsthree sections.
12.3 INTEGRALS INVOLVING TRIGONOMETRIC FUNCTIONS
There are many occasions when we have to integrate powers o&.$rigonometric
f..ot;ne
1 
thir r ~ o t ; n nZ J D S
rholl
; n r l ; ~ h n t s r tn nrr\raerl ~ x ~
in rnnrh 0  e n r
I
Integral Calculus
12.3.1 Reduetion Formulas for ssinnxdx and Scosnxdx
In this subsection we will consider integrandswhich are powers of either s ~ n x r cosx. o Let's take a power of sinx first. For evaluating \sinnx dx. we write I, = /sinnxdx
= !sinn'
x
S~IIr dx,
if n> 1 .
Taking sinn'x as the first function and sin x as the second and integrating by parts, \ we get I, = sinn'x cosx  (nI) lsinn2x cos x (cos x) dx
= sinn'r cosx
+ (n+ (n
I ) ]sinn'x cos2x dx
!
= sinn'x cosx
= sinn'X cosx
+ (n  1) ~ l s i n ~ (I~sin2x) x
I ) [/sinn'x dx
1) [I,,
dx]
 fsinnx dx]
= sinn'x cosx
+ (n
 I,]
Hence, I, + (n  1) In = sin"'x cosx That is, nI, = sinn'x cosx
For n = 1
+ (n
1) In,
+ (n1)
In2. O r ,
1, =
n
x cosx + n
I,, 1 n
\sin" x dx = lsinx dx
= cosx
+C
This is the reduction formula for isinnxdx (valid for n s 2). . Example 2 We will now use the reduction formula for isinnxdx to evaluate the
n12
definite integral, sinsxdx. We first observe that
0
Let us now derive the reduction formula for J cosnxdx. Again, let us write
I, = Jcosnxdx = jcos'l
x cosxdx, n > 1.
Integrating this integral by parts we get
I, = cosn'x sin x  l(nl)cosw2x (sinx). sinxdx
= cos"x
sin x + (n 1) cosw2x sin%dx
= cosn'x sin x
+ (n  1) k o s w 2 x (1cos2x)dx
By rearranging the terms we get
cosn'xsinx n 1 n In2 This formula is valid for n r 2. What happens when n = Oor I? You will agree that the integral in each Ease is easy to evaluate. .As we have observed in Example 2, In = kosnxdx =
+
12
Tinnxdx = 'I
0
l ~ i n "  ~ x d x , 2 2. n
0 nl2
Using this formula repeatedly we get
nl2
52
n1 ..n n3 n2 n5 n4
""'
Isinxdx, if n is an odd number. n a 3.
0 nl2
0
2 .LJdx, if n is an even number, n 2 2.
0
This means
b!
We can reverse the order of the factors, and write this as
nl2
~ r ~ usimilarly for Imsnx<x we get i n ~
0
a12
2.4
3
5
""'
i f n isodd,andn r 3 , n
o
o
1 3
2 4
""'
. n1 w i .f n iseven, n 2 2.
n 2
We are leaving the proof df this formula to you as an exercise. See E 1).
El) Prove that lmsnxdx
o
r
73
1 3
2 4
...**
n 1 , if n isodd, n r 3 n
7 7 .....
n1
n
w , if n is even, n r 2.
2
E.2) Evalusc
a) I m ~ x d x .
b)
0
coPxdx,
using the reduction formula
S tannxdx and S secnxdx In this subsection we will take up two other trigonometric functions tanx and secx. That is. we will derive the reduction formulas for 1tannxdx and 1secnxdx. T o derive a
12.3.2 Reduction Formulas for
:
reduction formula for tannxdx, n > 2, we start in a slightly different manner. Insteadof writing tan" x = tanx t a n n  ' ~ ,as we did in thecase of sinnxand cosnx,we shall write tannx = tann'xtan'x. You will shortly see the reason behind this. So, we write 1, = Jtannxdx = Jtann'x tan'x dx.
1
= jtann"x sec2x dx
 I t a n n  fd r
... .. .. (2)
You must have observed that the second integral on the right hand side is Now in the first integral on the right hand side, the intigrand is of the form [f(x)lrn. f'(x) As we have seen in Unit 11.
Thus, Jtany2x sec'xdx = tann'x n 1

+
Reduction Furmuls
71hu* reduction Glrnlul;~ thc li,r tanllx dx i i
1
You must have
now rcaliscd uh?
we wrote
tannx = ran'x rann%
To derive the reductio~~ formula for jrecnxdx (n > 1).we first write
SeCllX
= , C C ~ l  zY (ecs. and then integrate by parth. Thus,

 secll'x tan x  (n2) /scc"~'s (sec'x 1 ) d s s,,cl.l
x tan s  (n2) (I,,  I,, _,)
t
, Thesc fornluli~~ jt:lnnr ds and 1 srcnr dx are valid for n > 2. For n = 0. I and 2. the tor
integrals jtanl'x dx and 1sec"xdx can be easily evaluated. You have come across them
After rearranging the tcrms wc get SeC'~12 x t a n s n2 /srrnxds = I,. = +  1 1112 n 1
iti Units 10 and 1 1
Exiunplc3 Let's calculate
' il
ni'
i)
II
7ri1 "' 1
tan'x dx
and
ii) bccbx dx
(I
ftanir dx
11
=
T
],)
tan's

1
l a d s dx
nil
   1+ 1
4
2
j
sinx dx cosx
nil

 I  ~ n ( c o s x ) ] j
1'.
nl4
ii) jsecOxdx =
0
secJx tanx 5
1,)+ 4
rill
"4 I
hec'xdx
,
'
4 = 
3
8 +  tanx 15
28
Integral Calculus
F.3) Derive the following reduction formulas for jcotnxdx and jcosecnxdx:
a)
jootn xdx = I,
=  I cotnlx
n1
 I,,
+
~x b) jcosecnxdx = I,, =  c ~ s e c ~ cot x n I
n2
n  2 In2
E4) Evaluate
12.4 INTEGRALS INVOLVING PRODUCTS OF TRIGONOMETRIC FUNCTIONS
In the last section we have seen the reduction formulas for the case where integrands were powers of a single trigonometric function. Here we shall consider some integrands involving products of powers of trigonometric functions. The technique of finding a reduction formula basically involves integration by parts. Since there can be more than one way of writing the integrand as a product of two functions, you will see that we can have many reduction formulas for the same integral. We start with the first one of the two types of integrands which we shall study in this section.
ReductionFormulas
12.4.1 Integrand of the Type sinmxcosnx
The function sinmxcosnxdepends on two parameters m and n. T o find a reduction formula for /sinmx cosnxdx, let us first write
Since we have two parameters here, we shall take a reduction formula to mean a formula connecting I,,, and Ip,q,where either p < m, or q < n, or both p < m, q < n hold. In others words, the value of at least one parameter should be reduced.
:I* k
pit
Hence we assume that n > 1. Now, 1m.n = lsinmx cosnx dx = Icos*'x Integrating by parts we get (sinmxcosx) dx
Therefore, I,,,
+
n1
Im,n
.  m+n m + l
I
m.n
= cosnlx sinm+'x
m+l
+
n1
+1
Im ,n2
This gives us,
But, surely this formula will not work if m n = 0. So, what do we do if m Actually we have a simple way out. If m + n = 0, then since n is positive, we write m = n.
+
+ n = O?
Remember we have taken n > l ?
Hence I  , , = /sin"x cosnxdx = Icotnxdx, which is easy to evaluate using the reduction formula derived in Sec. 3 (See E2)). T o obtain formula (3) we had started with the assumption that n > 1. Instead of this, if we assume that m > 1, we can write Im." = /sinmx cosnxdx = /sinm'x(cosnx sin x) dx. Integrating this by parts we get 1m.n
 sinmlx cosn+lx  (m l ) [sinm2

(cosn+lx) xcosx
 1 1
 
dxfornfI


Integral Cakulus
From this we obtain 1m.n 
  ~ i n ~  I x c o s ~ ++ xm1 (Im2.n) ~ m+n
,+,
........ (4)
If m or n is a positive odd integer, we can proceed as follows : Suppose n = 2p + I , p > 0, then 1m.n = lsinmxcos2P+'x dx = lsinmx (Isin2x)P cosx dx
= I t m (1 t2)p dt if we put t = sin x.
Expanding (1t21P by binomial theorem and integrating term by term, we get
If m and n are positive integers, by repeated applications of formula (3) o r formula (4), we keep reducing n or m by 2 at each step. Thus, eventually, we come to an integral of the form I or I,., or 11,, or I,,*,. In the previous section we have seen how these can be evaluated. This means we should be able to evaluate I,,, in a finite number of steps. We shall now look at an example to see how these formulas are used.
,,
nl2
E r a m p k 4 Let usevaluate lsin4x cos6x dx. Here m = 4 and n = 6. Since m is the
0
smaller of the two, we shall employ formula (4) which reduces m at each step.
using formula (4) again.
 3 x
80
96
(from~2)b)) =
$
Are you ready to solve some exercises now?
E
E5) In deriving formula (4) we had assumed that m> 1. How would you evaluate, I,,,, if m = I ?
Reduction~ o i h u l a s

E
I
I
E6) Formulas (3) and (4) fail when m+n = 0. We have seen how to evaluate I,,, if m+n = 0 and n is a positive integer. How would you evaluate it if m+n = 0 and n
is a negative integer?

E
E7) Evaluate
12.4.2 Integrand of the Type eaxsin"x
P
r',
In this subsection we will consider the evaluation of those integrals, where the integrand is a product of a power of a trigonometric function and an exponential function. That is, we win consider inregrands of the type ea*sinnx.Let us denote ~eaxsinnxdx L,, and integrate it by parts, taking rinnx as the first function and eU as by the second function. This gives us
We shall now evaluate the integral on the right hand side, again by parts, with sinn'x cosx as the first function and eaxasthe second one. Thus,
.
a
a l
a
a '
,
,'"
A , " '
"
Thismeans
.
+
L , = , eaxsinnx  neaxsinn'x cosx 
"("1)
n2 Ln2  7 n L
intrgrai Calculus
Rearranging the terms we get L, = aeaxsinnx  neaxsinnlx cosx n2+ a2 n2+a2
+
n(n 3)
n2+a2 Ln2a
Given any L,, we use this reduction formula repeatedly, till we get L, or Lo (depending on whether n is odd or even). Since L1 and Loare easy to evaluate, we are sure you can evaluate them yourself. (See E 8)). This means that L, can be evaluated for any positive integer n.
Remark I if we put a = 0 in L,; it reduces lo the integral isinnxdx. This suggests that
the reduction formula for jsinnxdx which we have derived in Sec. 3 is a special case of the reduction formula for L,. If you have followed the arguments in this subsection closely, you should be able to do the exercises below.
E
E8) Prove that
E
E9) Prove : If C, =]eaxcosnxdx, then
E10) Verify that the reduction formula for cosnxdx is a special case of the formula in E9).
12.5 INTEGRALS INVOLVING HYPERBOLIC FUNCTIONS
,
In this section we shall discuss the evaluation of integrals of the type jsinhnxdx, Jcoshnxdx, etc. Actually, you will find that the evaluation of these integrals does not involve any new techniques. In fact, the procedure we follow here isvery similar to the one we followed for integrating sinnx,cosnx etc. Let US find the reduction formula for, say, jtanhnxdx. We are sure you will be abie to follow this easily and derive the reduction formulas for the other hyperbolic functions (see E l l ) ). If I, = Itanhnxdx, we can write I, = ItanhnP2xtanh2x dx
= Itanhn2x (1 sech2x)dx
d  unhx = sec h2x dx
Integral Calculus
Don't you agree that the above method is similar to the one adopted for /tannr dx? The following exercises can be easily done now.
E
E l l ) Prove the following reduction formula:
Isinhnxdx = sinhnlx coshx n n 1 n
\sinhn2x
dx
E
E12) Derive a reduction formula for \coshnxdx
'
That brings w t o the end of this unit. We shail now s u m a r i s e what we have covered in it.
1 . SUMMARY 26
A reduction formula is one which links an in.egral dependent on a parameter with a similar integral with a lower value of the parameter.
In this unit we have derived a number of reduction formulas.
1
/rnex.dx = xnex n \xn'ex dx
5
!secnxdx
=
n
6
0
2
n
0
~ec'l~x tanx n1
n2
n 1
isecn
n n
Rrduction Fornlulao
2x
dx,
>2
2
24
Isinnx dx = Icosnxdx =
3 5 1 3 2 4
i f n i s o d d , n h 3 . ,
""'
''
""'
x, n is even, n 2 2. if 2
We have noted that the prime technique of deriving reduction formulas involves integration by parts. We have also observed that many more reduction formulas itlvolving other trigonometric and hyperbolic functions can be derived using the same technique.
'
12.7 SOLUTIONS AND ANSWERS
n/2 n/7
E 1) w e have Icosnx dx =
0
n
xdx, n
2
2

n
n2
n4
""'
2 1c o s 0 x d xif. n is even ~ 4 2
0
Integral Calculus
Therefore, Ill =
n 1

1112
  c ~ s e c ~  cotx  /(n2) ~ w c ~  ~ x c dxt ~ x ~x o
 c~sec"~x cotx

(n2)/c0sec"~x (cosec2x 1) dx
n12
E 4) a) cosec'x dx = cosecxcotx 2 n14
/
.
]
nlZ
+
n14
+ j4
n12
cosecx dx
 nl2
JZ
4
1 +  (In I  In tan
f)
. 2
l' In tan
1 '
b) JsinxXdx
0
,fi 2 8 = 1.J.l.l. n = 357~
8 6 2 2 256
z
E5) ifrn=l,I,.,=I,,,
= Jsin x cosnxdx
E 6) m + n = 0 =s n = m
+ m is a positive integer.
'
1 ,.n = lsinmxcos"xdx = 1
Je dx =Jtanm
COS~X
X ~ X
Now use the formula for tanmxdx
12
=
2
2 cos" Jsinxcos5x dx =  8 6 0

1 24
eaxsinx sinx d x =  a a leaxcosxdx
  eax 
1 +a2
(asirix  cosx) + c
 eaXcosnx +
a
fl
a
xsinx
+
 eaXcosnx +
a
4 e a ~ C O S n  I xsinx +
a
E 10) Puf a
=0
in t h e formula for C,.
which is t h e reduction formula for l c o s n x d x
=
sinhn Ix coshx  (n 1) lsinhn'x cosh2x d x sinhn'x coshx  (n 1 ) lsinhn'x (1 t s i n h 2 x ) d x sinhnlx coshx  (n1) In2  (nl)Il1
= 
I,
= = = =
l c o s h n x d x = l c o s h n l x coshx d x coshn lx sin hx  (n I ) lcoshn'x sinh2x d x c o s h n l x sinhx coshnlxsinhx. n .
+
(n 1) /coshn'x n 1
(cosh2x1) d x
coshnIx sinhx  I n 
I)I,, + (n 1)ln, I,,,
I,
=
UNIT 13 INTEGRATION OF RATIONAG
AND IRRATIONAL FLNCTIONS
Structure
13.1 Introduction
Objectives
13.2
Integration of Rational Functions
Some Simple Rational Functions Partial Fraction Decomposition Method of Substitution
13.3 13.4 13.5 13.6
Integration of Rational Trigonometric Functions Integration of Irrational Functions Summary Solutions and Answcrs
13.1' INTRODUCTION
In the previous unit you'have come across various methods of integration. This unit, which
is the last one in this block, will coniplete the discussion of method!, of irltcgration in in this course. Here we shall deal with the integration of rational fur~ctioris detail. The method, hhich we shall dcscrihe in Sec. 2 , depends upon partial fraction clccclinposition with which you might be already familiar., RLII Later on in the unit we shall consider somc siliiplc typcs of irrittional f~~nctions a full discussion of thc intcgration of irration;iI functions is beyond the scope of this coursc. We end the unit by giving you a clicck list of points to hc considered t>cforc deciding upon thc nietliod of integration for i\ny giver1 function. While going thrn~lgti this unit you will ncctl to recall scvcr;~l standard forms like
I
,
Jx+a
dx
,
,I
, = '
rlx etc. which r e h;lvc i~lrcady a,vcrccl in Viiit I I .
objectives
Aftcr reading this unit you shoul(l be ahlc to rccognisc proper and improper r;~tio~ii~l Functions intcgri~tc rational l't~nctions ;I vi~riahle using tlic mcrhocl of ~)arti:~l (IS by fractiorls + intcgratc certain typcs of r:ltion:ll i n ~ l c t i o ~ ~\inx ;11\(.1 cosx of s cvalu:ltc the integrals ol'so~nc specified tyl~cs irrntional I'unctiolis of dccidc upon the ~iicthod integration to he i~scdl'or. of integrating an! givcn functio~i
13.2 INTEGRATION OF RATIONAL FUNCTIONS
Wc know by now that it i s c ~ ~ tointcgrutc any polvnc)mi;~l sy function. t l i i ~ tis. n function ,.... + a,,. In thissection wcshnllsee howa rational fgivcn by f(x) = a,,xn+ a,, ] x " function is integrated. But Ict us first define :I rational function.
'+
Definition I A function R iscallcd :i rational function if it is givcn by K(x) = Q(x)/P(x). whcrc Q(x) i ~ n d P(x) arc po1ynomi;lls. I t is defined for ill1 x for which P(x) f 0.
x? + 3 x
+6
x2.7
x1 zx2
3x2 + 5 3 x 2  6x
If the dcgrcc of Q(x) is icss than the clcgrcc of P(x). we say that R(x) is n proper rational function. Othcrwisc. i t is called :in i~npropcr rational function. Thus,
f(x)
=
x2+x+2
X+
I
is n propel. rationill function. ; ~ n d an irnpropcr one.
g(x) = >,is x2
17
Rut g(x) c:~nalso written as I,c
g(x) = (i2+3x+6)
81
+ (by long divi5ion) x2
Here we have expressed g(x), which is an improper rational function, as the sum of a polynomial and a proper rational function. This can be done for any improper rational function. Thus, we can always write
Integration of Rational and lrratlonnl Functions.
p k q i
=
+
function
rational function
As we have already observed, a polynomial can be easily integrated. This means that the problem of integrating an improper rational function is reduced to that of integrating a proper rational function. Therefore, it is enough to study the techniques of integrating proper rational functions. But first let's see whether you can identify proper rational functions.
E
E l ) Which of the following functions are proper rational functions? Write the improper ones as a sum of a polynomial and a proper rational function.
x3+ 1 a) x4+x+l d) xJ+x'5
x2
I
, , I
x2+x3 x?+ 1
x+8 x1+5x+18
.
.
13*2.1 Some Simple Rational Functions
Now we shall consider some simple types of proper rational functions, like
1 I and
Xm . Later you will see that any proper rational function xa'(~b)~ ax2+bx+c can be written as a sum of these simple types of functions.
We shall illustrate the method of integrating these functions through some examples.
Example 1 The simplest proper rational function is of the type
we already know that
From Unit 10, .
(xa)
Integral Calculus
Example 2 Consider the function f(x) =
1 (x+ 214
T o integrate this function we shall use the method of substitution which we have studied in Unit 11.Thus, if we put u=x+2, dx
= 1 andwecan
write
The next example is a little more complicated. 2x+3 Example3 Consider the function f(x) = x24x+5 This has a quadratic polynomial in the denominator.
\
NOW/
dx , Perhaps you are wondering why we have split the integral into two parts. The reason for this break up is that now the integrand in the first integral on the
2x+3 dxcan bewritten as!, x24x+5
2x4 dx x24x+5
+
1
x24x+5
g'(x) ; right isof the form  and we know that g(x)
2X4 dx = ln lx24x+51 c, ~hus,\ x24x+5 To evaluatethe second integral on the right, we write
+
Now,ifweputx2 = u,
dx
=
1 and
This implies,
In the beginning of this subsection we said that any proper rational function can b e written as the sum of some fpnctions of the type we considered in the three examples above. In the next subsectionwe shall see how this is done. But try t o solve anexercise before reading the next section. It will give you some practice in evaluating integrals of the types mentioned in this subsection.
E
E2) Evaluate
.. 13.2.2 Partial Fraction Decomposition
*
In school you must have studied the factorisation of polynomials. For example, we know that x25x+6 = (x2) (x3) Here (x2) and (x3) are two linear factors of x25x+6. You must have also come across polynomials like x2+x+ 1, which cannot be factorised into real linear factors. Thus, it is not always possible to factorise a given polynomial into linear factors. But any polynomial can, in principle, be factored into linear and quadratic factors. We shall not prove this statement here. It is a consequence of the Fundamental theorem of algebra which has been stated in Unit 11 of the Linear Algebracourse. The actual factorisation of a polynomial may not be very easy to carry out. But, whenever we can factorise the denominator of a proper rational function. we can integrate it by employing the method of partial fractions. The following examples will illustrate this method.
Example 4 Let us evaluate
f,
x21
dx. Here the integiand
x21
is a proper rational
function. Its denominator x2 1 cari be factored into linear factors as : x2 1 = (x 1) (x+ 1) This suggests that we can write the decomposition of
xZ 1
into partial fractions as :
If we multiply both sides by (x1) (x+l), we get 5x1 = A(x+l) + B(x1). That is, 5x1 = (A+B)x + AB. By equating the coefficients of x we get A+B = 5. Equating the constant terms on both'sides we get AB = 1. Solving these two equations in A and B we get A = 2 and B = 3. 5x12 Thus',   + 3 x21 ' x1 x+1 Into6rating both sides of this equation, we obtain,
*
Integral Calculus
5x 1 dx was the 1 decomposition of the integrand into partial fractions. The procedure for finding the values of the two unknowns A and B, involved two simple simultaneous equations in two unknowns. But the higher the degree of the denominator, the more will be the number of unknowns, and it might be very tedious to find them. What can we do in such cases? There is a very simple way out. As you have seen, the most important step in the evaluationof In the equation 5x1 = A(x+l) B(x1), if we put x = 1, we get 6 = 2B, or B = 3. Similarly. if we put x = 1, we get 4 = 2A or A = 2. Isq't this a much simpler way of finding A and B?
1x2
+
Note that 1 and 1 are the zeros of the denominator x'  1
.
Let's go on to our next example now.
Example 5 Suppose we want to integrate 2x2+x4 x3x22x ' We first observe that the denominator factors as x(x+l) (x2). This means we can write
0.  1 and 2 are the zeros of 2x.
Multiplying by x3x22x we get 2x2+x4 = (x+l) (x2)A + Bx(x2) Cx(x+l) Now, if we put x = 0 in this equation, we get 4 = 2Aor A = 2. Putting x =  1 gives us 3 = +3B, or B = 1.
+
Puttingx = 2 , w e g e t 6 = 6 C , o r C = 1.
= 2 lnlxl  Inlx+ll + lnJx21 + c. Our next example illust;ates the use of this method when the denominator has' repeated linear factors.
Example 6 Take a look at the denominator of the inregrand in
It factors into (xl)'(x+2). decomposition of xL3x+ 2. In this case we write
X dx. xL3x+2 The linear factor (x 1) is repeated twice in the
from this point we proceed tis before to find A, B and C. We get x = A ( X  I ) ~ B(x+2) (x1.) + C ( X + ~ ) + We put x = 1 and x = 2 and get C = 1/3 and A = 2/9 Then to find B, let us put any other convenient value, say x = 0. This give us 0 = A  2B + 2C or,O =
2 2B 9
2 + . 3
This implies B = 219. Thus,
1 2dx x33x+2
=2 F ~ + A ldX 2j ~ ~ 9
x1 +  jI3 (x1)* dx
In our next example, we shall consider the case when the denominator of the integrand contains an irreducible quadratic factor (i.e. a quadratic factor which cannot be further
Example 7 To evaluate
\
Integrationof Rutimd und Irrutionsl Functions
we factorise x42x3+x22x as X(X2) (x2
+ 1). Then we write
t
Thus, 6x311x2+5x4 = A(x2) (x2+1) B X ( X ~ + ~ ) + (Cx+D)x(x2) Next, we substitute x=O and x=2 to get A=2 and B = l . Then we put x = l and x = 1 (some convenient values) to get C=3 and D = 1
+
Thus, you see, once we decompose our integrand, which is a proper rational function, into partial fractions, then the given integral can be written as the sum of some integrals of the type discussed in Examples 1,2 and 3.
i
II
All the functions which we integrated till now were proper rational functions. Now we shall take up an example of an improper rational function. Example 8 Let us evaluate
I x2x2
dx.
Since the integrand is an improper rational function, we shall first write it as the sum of a polynomial and a proper rational function. Thus
Therefore,
.J+2x 1 x2x2
dx = lxdx + dx
l
+
5x+2 x2x2
dx
Now let us decompose
into partial fraction as
5x 2 = A(x+l) + B(x2) I f x = 1,weget 3 = 3B, thatis, B = 1. If x= 2, we get 12 = 3A, that is A = 4.
+
Try to do the following exercise now. You will find that each integrand falls in one of the various types we have seen in Examples I to 8, E3) Evaluate
1
13.2.3 Method of Substitution
The method of partial fraction decomposition which we studied in the last subsection can be applied to all rational functions. We can say this kecapse as we have mentioned earlier, the Fundamental theorem of algebra guarantees the factorisation of any polynomial into linear and quadratic factors. But the actual process of factorising a polynomial is sometimes not quite simple. In such cases it would be a good idea to critically examine the integrand to check if the method of substitution can be applied. We will now give two examples to show how we can sometimes integrate a given rational function with the help of a suitable substitution.
Example 9 Suppose we want to integrate
x(x5+!)
with respect to x.
For this we write
1
dx x(x5+ 1) dx
x4 dx x5(x5+ 1)
= 5x4.
Now let us write x5 = t. Then
: 1 r
$<$I
ij
1
I
x4dx  1 x5(x% 1)  5
dt
Example10 Let usintegrate
x~+x~+ 1
x21
w.r.t. x (division by x2)
x21
x~+x~+ 1 dx
=
(1 l/x2) dx x2+l+l/x2
In Examples 9 and 10 you y u s t have noted that the denominators of the integrands were not easily factorisable. The method of substitution provided an easier alternative. See if you can solve this exercise now.
E
E4) Integrate the following functions w.r.t. x
rational functions. In the next section we take up the case of rational trigonometric functions.
'
13.3 INTEGRATION OF RATIONAL TRIGONOMETRIC FUNCTIONS
You know that a polynomial in two variables x and y is an expression of the form
k
u
Accordingly, a polynomial in sinx and cosx is an expression of the form
The integration of P(sinx, cosx) can be carried out easily as we have already integrated sinmx cosnx in Unit 12. An expression, which is the ratio of two polynomials, P(sinx,cosx) and Q(sinx, cosx) is called a rational function of sinx and cosx. In this section we shall discuss the integration of some simple rational functions in sinx and cosx. We shall first indicate a general method for integrating these functions.
Let f(sinx, cosx) be a rational function in sinx and cosx. The first step in the evaluation of the integral off is t o make the substitution tan = t.
we get, lf(sinx, cosx) dx =
=
2t 1t2 l f ( 1+t , i3 27 x4 /~(t) dt,
dt
is a rational function oft. Now we can use the method of partial fraction decomposition to integrate F(t). In principle then, we can integrate any rational function in sinx and cosx. But in actual practice we find that the rational function F(t) is often complicated, and it is not feasible to apply the method of partial fractions. In this unit, however, we shall restrict ourselves to a few simple rational functions only.
Example 1 Let us integrate a+ 1
l
lcosX
x x + b(cos2 + sin 2) 2
<+
Now a+ bcosx = a(sin2 5 +cost )2 2
2
= (a+ b)cos22 + (a b)sin2 2
2
(ab) If we put tan 2 = t, we get dx a+ bcosx = 2 / (ab)
Iab *
L
+ tan2=] X
J
dt a+b 7 +t2 (ab dt
)
ab If a >b > 0, then a+b > 0,and we get ab

d
m
tan'
( Jfi 5 tan a+b
1f 0 a < b, t h e n u < 0, and ab
<
+
.
, i + J'G 5 /G tan
Integral Calculus
Example 12 T o evaluate
I
sinx(1 +cosx) dx sinx(1 +cosx)
dx, we write
1 + cosx = 2 C O S ~
2
I
+
I
dx l+cosx
sinx = 2sin 5 cos 1 ! 2 2
dx sin  cos3 2 2
X
cos2 2
Thus,
1 1 X l+sinx dx =  lnltanxill +  t a n 2 x + tan  + c sinx(1 +cosx) 2 4 2 2 Now proceeding exactly as in Examples 11 and 12, you can do these exercises.
I
lntcgrslh~n Rallnnal and or Irratbnal Functlnns
By now you have seen and applied many different methods of integration. The crux of the matter lies in choosing the appropriate method for integrating a given function. 'OSX. Realising that 1+sin2x this is a rational function in sinx and cosx, you may put t a n 3 = t and droceed :
'InX
For example, suppose we ask you to integrate the function
2
*: ih
1
I
sinxcosx dx = 1+sin2x
t(lt2)dt(1+t2) (l+6t2+t4)
*,
'?
i
By this step you will realise that it is going to be a tough job. But don't worry. There is an easy way out. 1 n dx, if we make the substitutiqn l+sin2x = t, we ~ ~ ~ ~ get ~ ~
1'1
J
=
1 1n(l+sin2x) + c. 2
Thus, the choice of the method is very crucial. And only practice can help you make a good choice. We shall now illustrate some techniques used in integrating irrational functions.
13.4 INTEGRATION OF IRRATIONAL FUNCTIONS
The task of integrating functions gets tougher if the given function is an irrational one, Q(X) . that is, it is nor of the form  In thivsection we shall give you some t i p for P(x) evaluating some particular types of irratio~~al functions. In most cases our endea. ~ a r will be to arrive at a rational function through an appropriate substitution. Thitrational function can then be easily evaluated by using the techniques developed in Sec. 2.
I) Integration of functions containing only fractional powers of x : In this case we put x = tn, where n is the lowest common multiple (1.c.m.) of the denominators of po\ilrers of x. This substitution reduces the function t d a rational
function o f t . Look at the following example.
Example 13 ~ eustevaluate
1
2x1'2
+ 3xU3 dx.
1+~"3
We put x = t6, as 6 is the 1.c.m. of 2 and 3. We get
12
X7/6 + 3X  J 2 X5/6  2 X 2 / 3 + 4x''2 + gXU3 12X116 9 In 11+x1131+ . 7 5 2 12 tan' x1I6 c
+
II)Integral of the type
Here we shall have to consider two cases : (i) a > 0 and (ii) a < 0. In each case we will try to put the given integrand in a form which we have already seen how to integrate. i) a > O
S
dx
J
If we put t = x+b/2a, we get
I
dx
JZ =
+I"
dt ~ t z + ( d a b2/4a2) 
This is one of the standard types of integrals listed in Table 3 in Unit 11, and can thu! be evaluated. ii) a < 0 : If we put a = d , then d > 0, and we can write
dt , if t = xb/2d J ( c / ~ b2/4d2)  t2 This is again in one of the standard forms.

%I
+
111) Integration of (fx+e)
1
JSGX '
dx
We will illustrate the method through an example. Example 1.4 Suppose we want to evaluate Letusputx+l
=
1ly.Then
Y
dx
=
1.
Now we will try to express x2+4x+2 in terms of y . For this we write x2+4x+2 = (x+l)' 2(x+1)  1
+
Therefore,
A 'I"
This example suggests that in integrating (fx+e)
1
JiZGC
,we should make the
substitution fx + e = 1 and then simplify the expression. Y' Let us move over to the next type now.
I V ) Integrationof (Ax+B)
J
..
We break Ax+B into two parts such that the first part is a constant multiple of the differential coefficient of ax2+bx+c, that is, 2ax + b, and the second part is independent of x. Thus, Ax+B =
A (2ax+b) 2a
+B 
Ab and
Evaluation of the last integral has already been discussed in 11).
V) Integrationof(Ax+B) qax2
+ bx + r
..
We break Ax+B as we did iq tV), and obtain
We have already seen how to evaluate the integral on the right hand side (see Sec. 4, Unit 11). Let us use these methods to solve some examples now.
Example 15 To evaluate
1 we note that x+2 =  (2xb2) + 1 , and write 2
Example 16 To evaluate
x2+2x+3 / Jx2+x+1
,
dx
wenotethafx2+2x+3 = x2+x+l'+x+2 = x 2 + x + l + Hence (x2+2x+3) dx = l d x 2 + x + l dx
1(Zx+I) + 1
2 2
+l l
3
~ z x ~ ~
(2x+ 1)

dx J(x+l)2+J 2 4
.=
We have used two results from Unit I l here.
/ J(x++)'+$dX
+ 3 I
2
and
 O i + i ) j 2 + ~71 + 3 1 n 2 +  2 8
+
=
f
See, if you can solve this exercise.
E
E7) Integrate the following:
7
+
+
J z x d x
n  ( x + ; + ~ ) + c
J5
1
.
( x + L 2
+ JG)
J'+
3
~nl(x+'+ d'5 2
JG) +c
+ J m ) + c.
(2x+5)
+
$ln
% (x++
d
When you are facedkith a new integrand, the following suggestions furnish a thread through the labyrinthof methods.
Integra,lioh of Rational and Irralional Funclions
(1) Check the integrand to see if it fits one of the patterns
'
(2) See if the integrand fits any one of the patterns obtained by the reversal of differentiation formulas. (We have considered these in Unit 11). (3) If none of these patterns is appropriate, and if the integrand is a rational function, then our theory of partial fractions enables UStointegrate it. (4) If the integrand is a rational function of sinx and cosx, and simpler methods of previous units fail, the substitution t = t a n x will make the integrand into a rational
2
function o f t , which can then be evaluated. (5) If the integrand is a radical of one of the forms
JG', Jm, ,/,
then the trigonometric substitutions x = asin0, x = acos0 or x = asec0 will reduce the integrand to a rational function of sin0 and cos0. If the radical is of the
,
form
, /
a square completion Ja(x+ b/2a)2 + c b2i4a will reduce it
essentially to one of the above radicals. (6) If the integrand is an irrational function of x, try to express it as a rational function or an integrable radical through appropriate substitutions. (7) Inspect the integrand to see if it will yield to integration by parts. Finally, we would like to remind you again that a lot of practice is essential if you want to master the various techniques of integration. We have already mentioned that a proper choice of the method of integration is the key to the correct evaluation of any integral. Now let us briefly recall what we have covered in this unit.
13.5 SUMMARY

In this unit we have covered the following points : 1 A rational function f of x is given by f(x) = P(x)/Q(x), where P(x) and Q(x) are polynomials in x. It is called proper if the degree of P(x) is less than the degree of Q(x). Otherwise it is called improper. 2 A proper rational expression can be resolved into partial fractions with linear or quadratic denominators. 3 A rational function can be integrated by the method of partial fractions. 4 5 Integration of a rational function of sinx and cosx can be done by putting t = tan 3 2 Integration of irrational functions of the following typesis discussed. i) integrand contains fractional power of x. 1 iii) 1 ii' JGG~GTC (fx e) J '
+
6
A check list of points to be considered while evaluating any integral is given.
13.6 SOLUTIONS AND ANSWERS
E l ) a) and c) are proper.
.' . we have x2 + x  1 = x2x
Thus 1 = 1
+B
+ 1 + B X + (CR)x ~
C
(Coefficients of x2)
B =0 Also,  1 = 1  C (constant terms)
. :
: C =2 .
E 5 ) 1 a +b sinx dx 
2dt , if t = tan x/2 a(l+t2) + 2bt
Ifwe write
 At+B 1t2 (t2+1) (3t2+1) t2+ 1 .
1= B + D 0=A + C 1 = 3B D 0 = 3A + C
f
Ct+D 3t2+1 '
then 1t2 = (At
+ B) (3t2+1) + (Ct + D) (t2+ 1)
+
(constants) (coefficients oft) (coefficients of t2) (coefficients of t3)
.'. A = C = 0, B = 1, D = 2
.'. Answer
= 21
dt
+4
I &
4 = x + t a n  ' ( a
fi
tan f )
+c
=
dx
1 dx 1 put 2x = . Then  = t dt t2
dx
=
1
dx (2X) J s ( z  ~ ) ~ io(zX) 
+.9
Solutioa of E8, Unit I1
i) ii)
($1 cashI ($1
sin'
+c +c
1 =tan
1
iii)
Putting 2x = t, we get
t+c
iv)
hnl J+ + c
JG
vi)
tan'
12
(+)
+c
vii) viii) ix)
 sin
3 sin' (x  1) 1
+ c (as in Example
1
13).
Let x
1 +  = y, Then the given integral is 2
xi)
x  tan' x
+c
I
I
t
1
UNIT 14 APPLICATIONS OF DIFFERENTIAL CALCULUS
Structure
14.1 Introduction
Objectives
14.2 14.3 14.4 14.5 14.6
Monotonic Functions Inequalities Approximate Values Summary Solutions and Answeh
! I
f
14.1 INTRODUCTION
In this unit we see some immediate applications of the concepts which we have studied in Blocks I and 2. Differential Calculus has varied applications. You have already seen some applications t o geometrical, physical and practical problems in Units 8 and 9. In this unit, we shall study some applications t o the theory of real functions. There are some questions in mathematics, that can be asked even before knowing differentiation, but can be solved with ease by using the theorems on differentiation. In other words, differentiation is a useful tool in solving problems that arise independent of this nbtion, for example, a) Is the function sin x + cos x increasing in [0, u/4]? b) Do we have the inequality eX c) What is the limit of
1ex
> I + x for all x?
x tends to O? sin x d) What is the approximate value of cos I "? Do you notice that these questions do not involve any concept that yoll have not studied earlier? You could have asked them earlier. You could have even answered some of them: Here we shall see how the theorems of Unit 7 can be systematically applied to yield solutions t o such questions. Keep Block 2 ready with you since you will need to read the relevant portions from the units in that block.
 when
Objectives
0
After studying this unit you should be able to: a recognise the equivalence of some properties of functions (like monotonicity and positiveness or negativeness of its derivative), a prove some inequalities using the mean value theorems, a apply Taylor's series to obtain approximate values of certain functions at  certain points.
.

14.2 MONOTONIC FUNCTIONS
In this section we employ differentiation t o decide whether a given function is monotonic or not. First y e recall some terms from Unit I . We say that a function from an interval I to R is monotonic if it is either an increasing function on I or a decreasing function on I. It is said to be a n increasing functionon I, if x s. y implies f(x) c f(y). Increasing

.
.
I
I
I
\ pplications
of Calculus
Let us also recall that every constant function is an increasing function, the identity function is an increasing function, the function f(x) = 2x + 3 is an increasing function, the function f(x)= sinx is not an increasing function on R.
~:
I
h his is hecause even though 0
< 3n/2, we have sin O = O > 1 =sin 3a/2.
However, on the interval [0, ~ / 2 ] sin x is an increasing function. , Further, we know that a function f from I to R is said to be a decreasing function if x 5 y implies f(x) r f(y). Decreasing functions are the orderreversing functions. Here are some examples: The function f(x) = 42x is decreasing. The function f(x) = sin x is decreasing in the interval [ir/2, ~ 1 . Do you agree that each constant function is both increasing and decreasing? Warning: It is incorrect t o say that if a function is not increasing, then it is decreasing. It may happen that a function is neither increasing ngr decreasing. For instance, if we consider the interval [0, R],the function sin x is neither increasing nor decreasing. It is increasing on [O, r/2] and decreasing on [ r / 2 ,r].There are other functions that are even worse. They are not monotonic on any subintervaI also. But most of the functions that we consider are not so bad. Usually, by looking at the graph of the function one can say whether the function is increasing or decreasing or neither. The graph of a n iincreasing function does not fall as we go from left to right, while the graph of a decreasing function does not rise as we go from left to right. But if we are not given the grqph, how d o we decide whether a given function is monotonic or not? Theorem 1 gives us a criterion to do just that. Theorem 1: Let I be an open interval. Let f : I
 R be differentiable. Then
a) f is increasing if and only if f'(x) r 0 for all x in I. b) f is decreasing if and only if f'(x) Proof: a) Let f be increasing. Let x
E
I0
for all x in I.
I. Then,
Since f is increasing, if h > 0, then x + h > x and f(x + h) r f(x). Hence f(x + h)  f(x) r 0. If h < 0, x + h < x, and f(x h) s f(x). Hence f(x+ h)  f(x) I 0. So either f(x+ h)  f(x) and h are both nonnegative or they are both nonpositive. Therefore,
+
f(x+ h, h Therefore,
lim ho
is ndnnegative for all nonzero values of h. f(x+ h)  f(x) must also be nonnegative. h
Thus, ff(x) r 0. Conversely, let f'(x) r 0, forall x in I. Let a < b in 1. We shall prove that f(a) 3, Unit 7), f(b) ba
= ft(c) for' some c
(_
f(b). By mean value theorem (Theorem
c 1 a, b I c1.
Since f'(c) r 0, we have f(b)  f(a) a 0 . l o b  a > 0. It follows that f(b)  f(a) r 0, or f(b) r f(a). :~s ba Thus, a < b implies f(a)
If(b).
Therefore, f is increasing.
. .
b) This can be proved similarly. We leave it as an exercise. It can also be deduced by applying part a) to the function, f.
Applications of 1)ilferentirl C?lculu~
E E l ) Prove Part b) of Theorem
1.
.
From the class of increasing functions we can separate out functions which are strictly increasing. The following definition gives the precise meaning of the term "strictly increasing function". /
Definition 1: f : I f(a) < f(b).
 R is said to be strictly increasing if a < b implies that
~e can similarly say that a function defined on I is strictly decreasing if a < b implies f(a) > f(b). For example, a constant function is not strictly increasing, nor is it strictly decreasing. The function f(x) = [x] too,is increasing, but not strictly increasing. increasing, whereas the function f(x) = x is stric~ly Fig. 1 shows the graphs of these three functions. In Fig. l(a) the graph is horizontal. In Fig. l(b) there are parts of the graph which are horizontal. But the graph in Fig. l(c) has no horizontal portions, and rises as we go from left to right.
(a)
(b) Fig. 1: (a) graph o f fix)= 1 (b) graph of f(x)= 1x1 (c) graph o f f(x)=x.
NOW let us see whether strict monotonicity of a function is reflected by its derivative. We have the following theorem.
Theorem 2: a) Let f ' be positive on I. Then f is strictly increasing on I.
b) Let f' be negative on 1. Then f is strictly decreasing on I.
Proof: a) By Theorem 1, we know that since f ' > 0 on 1, f must be increasing on \. That is, x < y L. f(x) r f(y). We have only to prove that if x < y, then f(x) :annot be equal t o f(y). Let, if possible, f(x)= f(y), where x < y in 1.
1
Appllcatlons of Calculus
Then, by Rolle's Theorem (Theorem 2, Unit 7) applied to the function f on the interval [x,y], we have f (c) = 0 for some x < c < y. But this contradicts the assumption that f ' is strictly positive on I. Hence f(x) cannot be equal to f(y) for any x.<y in I. Thus, x < y =, f(x) < f(y). That is, f is strictly increasing.
I
b) We indicate two different proofs for this part. One way is to imitate the proof of part a) by changing the symbol ' < ' to the symbol ' > ', wherever it occurs. Another way is to consider f, apply Theorem 2 a) to it and use the facts: (0' =  f l . Th'erefore f ' is negative if and only if (0' is positive. Also f is strictly decreasing if and only if f is strictly increesing. Combining the results for increasing and decreasing functions we get the foliowing corollary.
Corollary 1: f is strictly monotonic on the interval I if f ' is of the same sign throughout I.
You may have noticed that there is a difference between the statements of Theorem 1 and Theorem 2. We Said: "f is increasing if and only if f ' isnonnegative".
"If f ' >0, then f is strictly increasing"
It is natural to ask: Can we have "if and only if" in Theorem 2 also? That is, if f is strictly ? increasing, does it follow that f ' > O Unfortunately, we have a negative answer as shown in Example 2' below. But before that, our first example shows how among many methods available to prove the monotonicity, the one using differentiation is the simplest (provided the function is differentiable, of course).
Example 1: Let S(,x)= x3 for all x in R. We shall prove that the function f is increasing. First Method: Ler; x < y. We want to prove that x3
I y3.
Consider t y o cases.
Case 1: x and y are of the same sign. (Either both are positive or both are negative.) In this case xy > 0. 'Now,
y3  x3 = (y  x) (Y2+xy+x2)1 0 since both y  x and y 2 + y ~ +are2 ~ nonnegative.
Case 2: Let x and )r be not of the same sign. Since x < y, this means that x < 0 < y. (Note that if either x or y is zero, it comes under Case 1.) Therefore, x3 < 0 < y3. Because, the cube of a negative number is negative, and the cube of a positive number is positive.
Thus, in both cases x3 I 3. In fact, we have the strict inequality (x3 < 3), indicating that x x3 is a strictly increasing function.

~ e c o " dMethod: Let x <' 9. We want to prove that x3
Now (y3  x3) = (y  X) (Y2 + yx
I
~3.
+
x2)
Here y  x > 0 (since x < y). Also, y 2
+
yx
+ x2 r
Ci because
since the square of any number is nonnegative. Thus nonnegative numbers'and, Qence is nonnegative.
y3
 x3 is a product of two
Third Method: (Using Differentiation): Let f(x) = x3. Then f'(x) = 3x2. This is always nonnegative. Therefore using Theorein 1 we can say that f is an increasing function.
.
Example 2: Here we give an e)rarnple of a strictly increasing function whose derivative is not strictly positive.
Let f: R R be the function defined by f(x) increasing because, x < y =. y  x > O a n d ~ ' + ~ ' > 0

Applications of Uiffercntial Calculus
=
x3 (see Fig. 2). It is strictly
"
=
Y1
 x3
=
(y  x) ( y 2
t;
+ +
yx (x
+ +
x2) y)2] > 0
1
2
(y  x ) [(x'
y2)'
Its derivative is not strictly positive because f'(0) = 0. Our next example describes two different ways in which nonmonotonicity of a function can be proved.
Example 3: To prove that the function f : x sin x on the interval [O, ~ / 4 ] ,we can proceed as follows:

+
cos 2x is not monotonic
First Method: We shall consider three points, 0, 7r/6 and n/10 belonging to [O, */4]. Then, f(0) f[a/6)
=

sin 0
T

+
cos 0
7r
= 1 = 
Fig. 2
sin
6
+
cos 3 cos 5
a
1
2
+
 = 1
1
2
.
f(a/lO)
=
r sin 10
+
=
0.3090
+
0.8090 > 1
We have 0 < ~ / 1 0 a16 and f(0) < f(a/lO) > f(r/6). < Therefore, f is neither increasing, nor decreasing on [O, 1~/4].Or, We can say that f is not monotonic on [0, ~ / 4 ] .
,
Second Method: (Using Differentiation)
sin x + cos 2x Let f(x) Then f'(x) = cos x  2 sin 2x Now, f'(0) = 1  0 = 1 a n d f f ( r / 4 )
= 2
a
1
x l < 0.
Thus f ' is of different signs at 0 and a/4. Therefore f is not monotonic on [o, a/4]. Our next example warns us in dealing with functions not defined at some points.
Example 4: If the function f : x  tanx is defined on an interval, we can prove that it is a n increasing function there. See Fig. 3.
Now consider the intervql [O, TI. Can we prove that tanx is increasing on this interval? Suppose we argue as follows: f'(x) = secZx, and sec2 x 2 0 v x since the square of any quantity is nonnegative. Hence by Theorem 1, f is an increasing function on [0, TI. a 2T But if we take two points  and  in [0, a], 4 3 a 2~ =  6. then f = tan 7 = 1, and f = ian
):(
($)
1 2~ Thus,  < , 4 3
K 2~ but tan  S tan , 4 3
This indicates that tanx is not an increasing function on [O.. T]. So where did we go wrong? We can explain it as follows: u In the interval [0, TI, there is a point, namely , where tan is not defined. Hence 2 its derivative does not exist at that point, and therefore we can not apply Theorem 1. What we proved is that this function is increasing in an interval, provided it is defined throughout this interval. It is only when it is defined, that we can differentiate it and apply our theorems. In the next example we use an additional property of continuous functions in the first method, and repeated differentiation in the second method. Example 5: Let us prove that the function x  sin x+cos x is increasing on [O, u/4] and decreasing on [u/4, '1~/2]. First Method: Recall that for all x in thc first quadrant (i.e. if 0 s x 'sin x, cos x and tan x are nonnegative.
4
u/2),
sin Let f(x) = sin x + cos x, then f '(x) = cos x : x We note that f '(0) = 1, f '(u/4) = 0, f1(s/2)= 1. ] x = u/4 is the only point in [O, ~ / 2 at which f '(x) = 0. Because f ' ( 3 = 0 * cos x = sin x =, tan x = 1. But tan x is strictly increasing on [O, ~ 1 2 1by Example 4. So. it cannot take the value 1 at any point other than , a/4. Also, f ' is a.continuous function because the sine and cosine functions are continubus. Therefore, F cannot take negative values in [O, ~ / 4 ] . Explanation: If a continuous function takes a positiye value at' 0 and a negative value at so~nt: then it must be zero somewhere in between 0 and x. In h i s x, ~roblem,rhe continuous function f ' cannot take the value zero between 0 and x if x < a/4. Therefore, since f ' is nonnegative on [O, ~ / 4 ] ,f is increasing on.[O, u/4]. Similarly, f ' cannot take positive values in [s/4, ~ / 2 ] ,because its value at 7r/2 is negative. It follows that f is decreasing on [a/4, 7~121. Second Method: Let f(x) = sin x + cos x, and f '(x) = cos x  sin x. To prove that f is increasing on [O, u/4], we have to prove that f ' is no.. negative bn [O, '1~/4].We first note that f '(0) = 1 and f ' (u/4) = 0. It is enough to prove that f ' is decreasing on [0, a/4] (for then, all values of f ' in this interval will be between 0 and 1). For this purpose we consider fN(x)=  sin x  cos x. Also see Fig. 4(a) and (b). We note that fff(x) I0 for all x in the first quadrant, and in particular for all x in [0, s/4]. Therefore f ' is decreasing on [0,~/4].Therefore, (since f '(0) = 1 and f1(r/4)=O), f ' is nonnegative on [O, u/4]. Therefore f is increasing on [O, u/4]. Next, we shall prove that f ' is nonpositive on [a/4, u/2]. First, we note that f '(7r/4) = 0 and f ' (u/2) = 1. Also, f ' is decreasing on [u/4, a/2], since f " (x) 5 0 on l?r/4. r/21 Therefore all t h e values of f ' o n this interval are between 0 and 1
Fig. 4: (a) Graph of f(x) = sin x
+ cos x,
(b) Graph of f(x) =cos x  sin x
If you have followed the arguments in these examples, you should not have any difficulty in solving these exercises.
E
E '2) Assuming that e' never takes negative values, prove that it is an increasing . function o n R.
E 3) Prove that lnx is an increasing function o n ] 0, a [.
Applications of Calculus
Using the fact that sin x and cos x are never negative in the first quadrant, prove by differentiation that sin x is an increasing function and cos x is a decreasing function on [O; */?I.
Which of the following functions are increasing o n the interval given? Which of them are decreasing? a ) x 2  1 o n [O, 21
b) 2 x 2 + 3x o n [ 1/2, 1/21
c) e
' o n [O,

I]
d ) x (x
1 ) ( x + I ) o n [ 2,  11
e) x sin x o n [0, a / 2 ]
f) tan x + c o t x o n [O, ~ / 4 ]
Applicalions of 1)ifferenlial CIlculus
Applications of Calculus
Example 6: Suppose we want to prove ex 2 I + x for all x in R.
Let f(x)=eX x, Then f'(x)=eY  1, f"(x)=eX We know that ex > 0 for all x. Therefore, f '(x) is a strictly increasing function. Also f '(0) = 0. Therefoie there is no other point where f ' vanishes. Also, f'(x) > 0 if x > 0 f'(x)
< 0 if x < 0.
Therefore f(x) is increasing on ] 0, oo [. So x > 0 implies f(x) > f(0). This means ex  x > e0  0 = 1. This proves ex > 1 + x, if x > 0. It remains to prove this for negative values of x also. For this purpose we let g(x) = e x  xeX. Then gl(x) = ex  (xex+ ex)=
= , = ,
 xeX <
0 wherever x > 0.
g is strictly decreasing on ] 0, oo [
< g(0) whenever x > 0. ex  xeX < e0  O.eO= 1 for all x > 0.
gKx)

1 for all x > 0. 1x 1 Putting y'=  x, we get eY <  Or , l+y eY 5 l + y f o r y < 0. I,notherwords,eX> l + x f o r a l l x < 0. In other words, ex < When x0, e x = l = 1 +x. Thus, the inequality ex r 1 + x is true for all values of x. In the next example we give an inequality that is still better. x2 x3 Example 7: We prove that e x > 1 + x +  +  for all x > 0. 2 6 We have seen in Unit 6 that the Taylor's series expansion is vcllid for all x. This proves the inequality x2 x3 ex > 1 + x +  +whenever x > 0. 2 6 Fig. 5 represents the results of Examples 6 and 7.
I I
Example 8: T o prove b"  a " C n b n  ' ( b  a), wherever 0 < a < b and n > .1, we consider the function f : x x n o n the interval [a, b]. It is continuous there. It is also differentiable in ] a , b [. Therefore, by the mean value theorem, there is some c, a < c < b such that

Applications of Differential Calculor
Crossmultiplying we get, b "  a n = nc"' Therefore, it suffices t o prove that c"'
\
I
( b  a).
< bn'.
'
This is true because 0 < c < b, a n d n > 1. You can try these exercises now.
E E 8) Prove the following inequalities using the methods indicated alongside in
brackets. a ) In ( I + x) < x for all positive x (first prove that x  In ( 1 + x) is increadng). b) tan.'x < x for all positive x (by mean value theorem on [0, x] for tan' x).
$t
?! I
c)
d)
+ e  X > 2 v x. (writing 'ex+ eX2 a s a perfect square). ex + e  * > 2 v x . (using the already proved result e x r 1 + x).
ex
e) ex
 ex r 2
x
v x > 0. (using the inequality in d ) and differentiation).
I
Applications of Calculus
.
14.4 APPROXIMATE VALUES
In the previous two sections we have seen how ttie concept of.derivatives can be used in proving the monotonicity of a given differentiable function; and how this knowledge can be applied to prove some inequalities. In this section we shall see how Taylor's series can be used to find approximate values bf some functions at some points. We use the symbol = to mean approximately equal to. Example 9: Ta'king the first two nonzero terms in Maclaurin's series for sinx, we shall prove that sin 20" is approximately equal to 0.342 (in symbols, sin 20" = 0.342). Remember that in the formula sin x = x x3 +
3!
 . . . . . . .
5!
x5
the angle x i s always measured in radians. The same holds for cos x, tan x, and so on. a Now, 20" =  radians. Therefore,
. sin . sin
9
=  
a
a
9
9
 + ........ 
(*/9)3
=9
a
a

9
6
3! (a/9)3
Taking
a = 3.142,
we get this quantity to be
If you lo& into a table of sines, you will find siin 20" =0.342. This shows that our approxiination is really a good approximation. In fact, the tables are written by using precisely these methods. Example 10: Let us find the approximate value of (0.99)"~ by taking three terms of Maclaurin's series for (1  x ) ~ ' ~ Maclaurin's series for (1  x)'I2 is
5 1x+ 2 (5/2) (3/2) 2 x2
+ .......
W c can write (0.99)5'2 as (1  0.01)~'~.
So when x=0.01, taking the first three terms of M~claurin'sseries, we get
That is, (0.99)~'~0.975.
The error is the difference between Example 11: We know that COS  = 6 / 2 . If the first two nonzero terms of 6 the actual value and the approxiMaclaurin's series for cos x are tqken to approximate it, let us calculate the error, nate value. /
if
rounded off to two decimal places. Maclaurin's series for cos x is
If we take the first two terms alone and put x = n/6, then
The actual value is cos to thrce dccirual placcs.
6
=\/j/2. We know that f i / 2 =0.866 whcn roundcd o S T
We have found that 1   ~0.863 when r o u ~ d e doff to three decimal places.
(~/6)~
,
.
I'he error i:, 0.866  0.863 = 0.003.
Applications of Differential Calculus
When rcli~tidedo f f t o two decimal places, the error is.O.00. (This means that the error is so negligible that there is no error at all when rounded off t o two decimal places.)
Sce if you can find the approximate values in the following exercises.
E
E 9) Find the approximate value of sin 31" by taking the first two nonzero terms of its Maclaurin's series.
E
E: 10) I f the first three nonzero terms of Maclaurin's series for cos x are used t o
approximate cos , show that the error is less than 1/50.
'Ir
E :E
11) Find the value of cos 59", rounded o f f t o one decimal place
E E 12) Find (I . O l ) ' I
upto two decimal places.
That brings us to the end of this unit. Let us summarise what we have studied in it.
14.5 SUMMARY
h:this unit we have studied the following results.
1)
I
L
1ffis increasing decreasing constant monotonic If f ' is nonnegative nonpositive (strictly) positive (strictly) negative identically zero of same sign throughmt
I
then f ' is nonnegativenonpositive identically zero of same sign throughout
I
then f is increasing decreasing strictly increasing strictly decreasing constant monotonic
3)
Differentiation can be 'used t a test wnether a function is monotonic or not, to prove some inequalities, and t o find some approximate values.
44.6 SOLUTIONS AND ANSWERS
E I ) Let f be a decreasing function. If x 6 I, then f'(x) exists, and
So.
f(x + h)  f(x)
lim
Hence f '(x) = Now, let fl(x)
h+,o
f(x+h)f(x) h
5
0.
I
Oin I, and a, b E I s.t. a < b. Then 3 c E I s.t.
r
i
E 2) Let f(x) = e x . f'(x)=eX 2
= ,
ovx
R.
Applications of Differential Calculus
1
f(x)=eXis an increasing function on R (by Theorem 1). 1 1 E 3) If f(x) = In x, f'(x) = . Now  > 0 V x E 1 0, oo [.
X
X
!
Hence In x is an increasing function o n ] 0, m [.
E 4) If f(x) = sin x, then f '(x)
cos x 2 0 v x E [0, a/2].
= cos
x.
Hence sin x is an increasing function on [O, a/2]. Similarly for cos x.
E5) a) f1(x)=2x r O on [O, 21: f is increasing on [O, 21.
f is increasing on
1 / T, f1
e) increasing
f) f'(x)=sec2 x
,'I
 cosec2 X =
 cos 2x
sinZ x cos2 x
r 0 since cos 2x 2 0
i., !,I
vx
E
(0, a/4].
: .
f is decreasing.
E 6) a ) f ' ( x ) = 4 ~ + 3
fl(x)
I0
if x E [ 1, 3/4] and fl(x) r 0 if x f 13/4,0].
b) f'(x)=3x2  1. ff(x) 5 Oif x E 0, and f '(x) r 0 if x E
[+9
I;
21
C) f'(x)=sin x + x cos x d) similar argument
r 0 for x € [O, n/2] 5 0 for x 6 [5~/4, a].
E 7) Let f ( x ) = a x 3 + b x 2 + c x + d
ft(x) = 3ax2 + 2bx + c f has'extreme at x = 2 and x = 3. f'(2)=0 =, 1 2 a + 4 b + c = 0 and f'(3)=O 27a+6b+c=O .'. 1 5 ~ + 2 b = O

Suppose a =  2 and d = 0: Then f(x) =  2x3 + 1 5x2  36x satisfies the given conditions. 1 E 8) a) f(x)=x  In ( I +x). f l ( x ) = 1   r 0 for all x > 0 1+ x x  In (1
.r .
+ x) is increasing o n 10, oo [
In ( I + x ) < x V x t 10, m[.
b) f(x) = tan' x. Let x > 0. By the mean value theorem 3 y' E 10, x[ s.t. tan' x  tan' 0 = f'(y) x0 tan' x 1 or = < 1 for y € 10, x[. X 1+Y2
d) ex r 1 + x and e'=eY ; 1 + y = 1  x (if y=x) . . e q e " 2 1+ x + 1  x=2.
:.
e) Let f(x)=eX ex  2x. Then f r ( x )= e X + e  '  2 r 0. = f is an increasing function = f(x) r f(0) for a l l x > 0 => e'  e'  2x 2 0 for all x > 0. = ex  e' r 2x for all x 0.
3 1 ~ 31"~radians.
180
= 1  (3.142)~  + (3.142)~ 8 16x24 = 1  1.234+0.235
=0.001. We know that cos  = 0
2
?r
. Error =O.OOI < :
50
1
UNIT 15 AREA UNDER A CURVE
Structure
15.1 Introduction
Objectives Cartesian Equation Polar Equations Area Bounded by a Closed Curve
15.2 Area Under a Curve
15.3 Numerical Integration
Trapezoidal Rule Simpson's Rule
I
15.4 15.5
Summary Solutions and Answers

15.1 INTRODUCTION
When we introduced you t o integration, we mentioned that the origin of the method of integration lies in the attempt t o estimate the areas of regions bounded by plane curves. In t h ~ s unit we shall see how t o calculate the area under a given curve, when the equation of the curve is given in theCartesian o r polar or parametric form. This process is also called quadrature. We shall also study two methods of numerical integration. These are helpful when the antiderivative of the integrand cannot be expressed in terms of known functions, and the given definite integral cannot be exactly evaluated.
Objectives
After reading this unit you should be able to: use your knowledge of integration t o find the area under a given curve whose equation is given in the Cartesian o r polar o r parametric form, recognise the role of numerical integration in solving some practical problems when some values of the function are known, but the function, a s a whole, is not known, a use trapezoidal and Simpson's rules t o find approximate values of some definite integrals, compare the two rules of numerical integration.
15.2 AREA UNDER A CURVE In this section we shall see how the area under a curve can be calculated when the equation of the curve is given in the i) Cartesian form ii) polar form iii) parametric form . Some curves may have a simple equation in one form, but complicated ones in others. So, once we have considered all these forms, we can choose a n appropriate form for a given curve, and then integrate it accordingly. Let us consider these forms of equations one by one.
15.2.1 Cartesian Equation
We shall quickly recall what we studied in Sec. 2 of Unit 10. Let y = f(x) define a continuous function of x on the closed interval [a, b]. For simplicity, we make the assumptioi that f(x) is positive for x c [a, b]. Let R be the plane region in Fig. 1 (a) bounded by the graph!; of the four equations:
~pptiationa C&II!IW of
We divide the region R into n thin strips by lines perpendicular to the xaxis through the end points x = a and x = b, and through many intermediate points which we indicate by x,, xzi ..., xn,. Such a subdivision, as you have already seen in Sec. 2 of Unit 10, is referred to as a partition P of the interval [a, b] and is , indicated briefly by writing
Fig. I
7
We write
and take the set of n points such that xi,s ti 5 xi for i = 1, 2, ..., n. We now covtruct the n rectangles ( ~ i g :1 (b)) whose bases are the n subintervals [xi  xi], i = I , 2, ...,n induced by the partition P,, and whose altitudes are f(t,), f(tz). .... f(ti), ..., f(t,,), f(tn). The
,,
sum
C f(ti)
i=l
n
A
xi
of the areas of these n rectangles will be an approximation to the "area of R". Notice (Fig. 2(a) and (b)) that if we increase the number of subintervals, and decrease the length of each subinterval, we obtain a closer approximation to the "area of R"
Thus, we have Definition 1: Let f be a real valued function continuous on [a, b], and let f(x) a 0
x t [a, b]. If the limit of
A
Area Under a Curve
z
n
I =I
f(ti)
A
xi exists as the lengths of the
subintervals,
xi
 0,
then that limit is the area A of the region R. f(ti)
A
That is, A = lim
axi0
z
xi
i= l i = I , 2, ..., n
Compare this definition with that of a definite integral given in Unit 10. Over, there we had seen that the definite integral,
n
n
f(x) dx is the common limit of
I=
mi a xi and
I i=l
Mi A xi as the
A
xi's
 0.
Now since m, s f(ti)
IMi
vi, we have Mi A xi
i= 1
z
n
mi a x i 5
C f(ti) a x i
i= 1
n
5
i=l
Hence if the limit of each of these as A xi's Theorem in Unit 2. lim
AX^
'0
 0 exists, then by the Sandwich
A
z
i=l
mi
A
xi a
lirn
AXi0
z
i=l
f(ti)
xi
5
lim
A X i '0
z
i=l
Mi A xi
Now, if
[
a
f(x) dx exists, then the first and the third limits here are equal, and
b P
therefore we get A =
\
f(x) dx.
.:..(I)
J
a
The equality in ( 1 ) is a consequence of the definitions of the area of R and the definite integral
( f(x) dx. Since f(x) is assumed to he continuous on the interval
i
[a, b], the integral in (1) exists, and hence yields the area of the region R under consideration. From the Interval Union Property (Sec. 3, Unit 10) of definite integrals, we have f(x) dx, a
Ic I b.
....(2)
This means if A:, A, A t denote the areas under the graph of y = f(x) above the : xaxis from a to c, from c to b and fro= a to b, respectively, (Fig. 3) then, if c is in between a and b, then we have
If we define A: too.
=
0, A! = 0, then the abovt: equation is true for c = a and c = b
Mg. 3
Till now, we have assumed the function f(x) to be positive in the interval (a, b]. In general, a function f(x) mav assume both positive and negative values ih the interval [a, b]. T o cover such a case, we introduce the convention about signed areas.
II
I
8
Applicalions of Calculus
I
I
I
interval [a, b] also. In that case (3) is true even if c is beyond b, since according to b our convention of signed areas, Acwill turn out to be a negative quantity (Fig. 4). Thus, A: = A: Or, A:
:I
+
A: = A:  Ag.
+
Ab = A:.
b
Now, if f(x) r 0 for all x in some interval [a, b], then by applying the definition of "area of R" to the function :f(x), we get the area A = 
[ f(x) dx.
If we do.not take the negative sign, the value of the area will come out to be negative, since f(x) is negative for all x E [a, b]. To avoid a "negative" area, we follow .this convention. Thus, if f(x) ( 0 for x E [a, b] (Fig. 5 ) , then the area : between the ordinates x = a and x = b will be
Fig. 4
A=
 f(x) dx
i'
The following examples will illustrate how our knowledge of evaluating definite integrals can be used to calculate certain areas.
Example 1 Suppose we want to find the area of the region bounded by the curve : y = 16  x2, the xaxis and the ordinates x = 3, x = 3. The region R, whose area is
to be found, is s'hown in Fig. 6.
Fig. S
Fig. 6
The area A of the regi0n.R is given by
Example 2: Consider the shaded region R in Fig. 7.
Area Under n Curve
4
I
1
I
Fig. 7
1
1I
I
R is composed of two parts, the region R, and the region R2. We have Area R = Area R l + Area R2 The region R I is bounded above the xaxis by the graph of y = x 3 + x 2  2x, x=2 and x = 0 . Hence,
The region R, is bounded below the xaxis by the graph of y = x 3 + x 2  2x, x = O and x = I . Hence,
I
Area R 2 = 
[
(x3+x2
 2x1 dx
8 5 37 Therefore, Area R =  +  = . 3 12 12
In this example we had to calculate area R, and area R2 separately, since the region R2 was below &hexaxis. Therefore, according to our convention Area R2 = 
S
0
1 .
f(x) dx.
If we calculate
,I
2
f(x) dx, it will amount to calaulating
f(r) dx +
2
f(x) dx = area R l  area R2. which would be a wrong
~pplimions of
I
I
C&UIUS
Example 3: Let us find the area of the smaller region lying above the xaxis and included between the circle x2+ = 2x and the parabola y 2 sX.
On solving the equations x 2 + y2 = 2x and y2= x simultaneously, we get (0, O), (1, l), (1, 1) as the points of intersection of the given curves. We have to find the area of the region R bounded by OAPBO (Fig. 8).
Fig. 8
From the figure \he see that area of region OAPBO =area of region OCPBO  area of region OCPAO
Now
1
I
&iidx=
0
0
=
i
cos 8 (  cos 8) dB, on putting 1  x = sin 8
Also.
f
0
& dx = 
2 3
Therefore, the required area = Try to solve these exercises now.
E E 1) Find the area under the curve y = sin x between x = 0 and x = n.
.
A m Undy a Curve
1
E E 2) Find the area bounded by the xaxis, the curve y = ex, and the ordinates x = 1
and x = 2.
E E 3)
Find the area of the region bounded by the curve y = 5x  x2, x = 0, x * 5 and lying above the xaxis.
E E 4) Find the areacut off from the parabola y2=4ax by its latus rectum, x = a
E E 5 ) Find the area between the parabola
I
J
y 2 = 4ax
and the chord y = mx.
In this subsection we have derived a formula ( I , ~ ~ I I I I I (1)) to find the area under \~ a curve when the equation of the curve is given in the @artesian form. With slight modifications we can use this: formula to find the area when the curve is described by a pair of parametric equations. We shall take a look at curves given by parametric equations a little later. But first, let us consider the curves !+en by a polar equation.
1 . . Polar Equations 522
Sometimes the Cartesian equation of a curve is wry complicated, but its polar equation'is not so. Cardioids and spirals which ; haiff encountered in Unit 9 are c, examples of such curves. For these curves it is much simpler to work with their polar equation rather than with theCartesian ones. In this subsection we shall see how to find the area under a curve if the equation of the curve is given in the polar form. Here we shall try to approximate the g ven area through the areas of a series of circular sectors. These circular sectors will perform the same function here as rectangles did i n Cartesian coordinates. Let r = f(8) determine a continbous curve between the rays 8 = a! and 8 = P (0  a s. 27r). We want to find the area A(R) of the shaded region R in Fig. 9 (a).
Imagine that the angle AOB is divided into n equal parts A 0. Then A 0
=  a
'
Area Under a Curve
n
. This amounts to slicing R into n smaller regions,
R,, R2,..., R,, as shown in Fig. 9(b). Then clearly
Now let us take the ith slice Ri, and try to approximate its area. Look at Fig. 10. Suppose f attains its minimum and maximum values on [8i_l,oil at ui and vi. 1 Then  [f(ui)12 A 0 2 From this we get
IA(Ri) 5 
Fig. 10
The area of a wctor of a circlc of radius r and \cctorial angle AH is YZ rr? AO.
1 [f(vi)12 2
A
0.
.The first and the third sums in this inequality are the lower and upper Riemann sums (ref. ,Unit 10) for the same definite integral, namely,
S:
a
P
[f(0)l2dO.
Therefore, by applying the sandwich theorem as A 0
R
R
 0, we get
a
We shall illustrate the use of this formula through some examples. Study them carefully, so that you can do the exercises that follow later.
Example 4: Suppose we want to find the area enclosed by the cardioid r = a (I  cos 0).
We have r=O for O = O and r = 2 a for O=a. Sincc cos 0 = cos (O), the cardioid is symmetrical about the initia! lincs AOX (Fig. 11).
Fig. 11
Hence the required area A, which is twice the area of the shaded region in Fig. 11, is given by
= 4a2
0
0 0 0 sin4  do, since cos 8 = cos2   sin 2 2 2 2
Applications of Calculus
$ dr$, where r$ = I
0
0 2
=
3 1 a 8 a 2    by applying the reduction formula from Section 3 o f u n i t 12'. 4 2 2

In the case of someCartesian equations of higher degree it is often convenient to change the equation into polar form. Thc Collowing cxamplc gives one such situation.
~ x a r n ~ 5:e T o find the area of the loop of the curve l
we change the given equation into a polar equation by the transformation
x = r cos 13 and y = r sin 13. Thus, we obtain
.
which yields r = 0 for 0 = 0 and 0 = a/2. Hence, area A of the loop is that of a sectorial area bounded by the curve and radius vectors 0 = 0 and 0 = a/2, that is, the area swept out by the radius vector as it moves from 0 = 0 to 0 = a/2. See Fig. 12.
Fig. 12
Thus, 25a2 cos40 sin40
0
5 a2 2
I
dt
, where t = 1 + tan5 8.
Try. to d o these exercises now.
I
E E 6) Find the area of a loop of the curve r 4 a sin 38.
Area Under a Cune
E 7) Find the area enclosed b y the curve r = a cos 28 and the radius vectors4 e=o. e=?r/2.
E E 8) Find the area of the region outside the circle r = 2 and inside the lemniscate
r 2= 8 cos 219.
[Hint: First find the points of intersection. Then the reauired area = the area under the lemniscate  the area under the circle.]
~~plkrti0118 ot ~81cmlo.s
15.2.3 Area Bounded by a Closed Curve
.
Now we shall turn our attention to closed curves whose equations are givenin the parametric form. Let the parametric equations x=4(t), Y=$(t), t E [a,
Here we shall consider curves which d o not intersect themselves.
PI,
where 4(a) = $I(@, and $(a) = $(P), represent a plane closed curve (Fig. 13).
Fig. 13
This means t.hat as the parameter t increases from a value a to a value P, the point P(x, y) describes the curve completely in the counter clockwise sense. Since the . curve is closed, the point on it corresponding to the value P is the same as the point corresponding to the value a. This is reflected by the conditions &x)= 4(p) and #(a) = $(P). Suppose further that the curve is cut at most in two points by every line drawn parallel to the k or yaxis. We also assume that the functions 4 and $ are differentiable, and that the derivatives 4' and $' do not vanish simultaneously. Let the point R on the closed curve correspond to the values a and P, i.e., at R we have $(a) = 4(P) and $44 $(P). = Now suppose A is a point on the curve which has the least xcoordinate, a. Similarly, suppose B is a point on thc curve which has the greatest xcoordinate, b. Thus the lines x = a and x = b touch the curve in points A and B, respectively. Further let tl and t2 be the values of t that correspond to A and B, respectively. Then Let a point Q correspond to t = t3 such that t2 < t3 < t, The area of the region enclosed is S = S2  S,, where S2 and S, are the areas under the arcs AQB and ARB, 5spectively (see Fig. 13): Hence, S2 = I Y d x a n d S l =
a
.
(AQB)
(ARB)
*
Now, as a point P(x, y) moves from B to A along BQA, the value of the parameter increases from t2 to tl. Therefore, ydx =
b
1
12
Ydldf
dx
Note how we have modified formula (1) for a pair of parametric equations.
(BQA) Hence S2= 
t,
Now the movement of P from A to B along ARB, can be viewed in two parts: From A to R and from R to B. As P moves from A to R, the value of the parameter increases from t , to 0, and as P moves from R t o B, t increases from
Area Under a cube
CY
Remember the point R corresponds to t = a and also to t = 8.
Therefore S I = I.
Thus, we have
Note that the negative sign is due to the direction in which we go round the curve .as marked by arrows in Fig. 13. Similarly, by drawing tangents to the curve that are parallel to the xaxis, it can be shown that
0
.
....(ii)
From (i) and (ii), we get
Hence. the arca enclosed is
s
=
2
'
S
I
(xdy

ydx)
We can use any of the forniulas (i), (ii) and (5) above for calculating S. But in many caw5 vou will find that formula (5) is more convenient because of its sy~nliietry.
I
Ksamplc 6: L.ct us find the arca of the astroid
s  a cos3t. y = h sin3t. O 5 t 5 'n
L
Thc rcgion bou~idcdb. tlic astroid is shown in Fig. 14. !
Tlic arca A of thc ~.cgioni.s gi\.cn b. !
27r
1
F

2
[
;a cos't (3b sin't w.t) b sin't (  3a cos'l sin I ) dl
I '
Fig. 14
We have seen in Section 3 of Unit 11 that
= 2
Here cos2 ( 2 r  t) sin2 ( 2  t) = cos2t sin2t. ~ 2* * Hence
I
f(x) dx, if f(2a  x) = f(x).
[ cos2t sin2t dt
0
=
2
[ cos2t sin2t dt.
~hereforc = A Now, by a similar argument we can say that
A = 6 ab
S
0
r/2
cos2t sin2t dt,
3~   ab , by using the reduction formula from Section 4 of Unit 12. 8 You can solve these exercises now.
,
E E 9) Find the area of the curve
x = a (3 sine  sin3f3), y = a cos38, 0 I0
4
2 ~ .
Area Under a Curve
E E 10) Find the area enclosed by the curve
x = a c o s O + b s i n O+c y = a ' cos O + b ' in H+c' 0 r 19 r 2a
E E 11)
Find the area of one of the loops of the curve x = a sin 2t, y = a sin t.
(Hint: first find two values of t which give the same values of x and y, and take these as the limits of integration.)
15.3 NUMERICAL INTEGRATION
In many practical problems, the values of the integrand, that is, the function wlroseintegral is required, are known only at some chosen points. For example,
Applications of C~I&IUS
In some cases, no simple integral' is known for the given integrand. For example, sin x the function does not have a simple indefinite integral. In such a situation
X
f(x)dx canno1 bc evaluated cxactly. Bul we can find an approximate value of the integral by considering the sum of the areas of inscribed (inner) or circumscribed (outer) rectangles, as we have seen at the beginning of this unit. In this section, we shall describe two more methods for approximating the value of an integral: i) Trapezoidal Rule and ii) Simpson's Rule.
15.3.1 Trapezoidal Rule
We know that a given definite integral can be approximated by inner and outer rectangles. A better method is given by the trapezoidal rule in which we approximate the area of each strip by the area of a trapezium. Such an approximation is also called a linear approximation since the portions of the curve in each strip are approximated by line segments. As before, we divide the interval [a, b] into n subintervals, each of length
A
x =   , by using the points x, = a + (ba)
: 1
A
x, x2 = a + 2 A x,. ..,
x n  = a + (n  1.) ~
A
x between xo = a and xn = b. Then
Now, we approximate the first integral on the RHS by the area of the trapezium aPoPlxl (Fig. 15), the second by xlPlP2x2and so on, thus getting
The area of a trapezium = fi (the sum of 11s parallel sides) x heisht.
S
a
f(x) dx =  (yo 2 
1
+
yl) Ax
+
 (yl
1
2
+
y2)
AX
+ ... +
1  ( ~ ~  1yn) A X .
2
+
where
YO
= ~(xo), = XI), YI
. . . ,Yn
= f(xn)
0
Fig. 15
The formula (6) is known as the trapezoidal rule for approximating the value of a definite integral. See if you can solve this exercise now.
Area Under a Curve
E E 12) L(se the trapezoidal rule to estimate the fallowing integrals with the given
value of n. a)
Sx2
2

dx, n = 4
b)
1& T
4
d l , n = 6.
15.3.2 Simpson's Rule
In this method, instead of approximating the given curve by line segments, we approximate it by segments of a simple curve such as a parabola.
.
t
Yo
The area under the arc of the parabola y = ~ ~ B2 + C X+ becwecn x =  h and x = h (Fig. 16) is given by
h
I
Fig. 16
37
I I
Applications of CX~~CU~US
Since the curie passes through the points, (  h, yo), (0, yl) and (h, yz), substituting the coordinates of these ljoints in the equation of the parabola, we get
yo = Ah'
 Bh
+
C
from which we get
C = y,
~ h= = Bh  )1 Ah' + Bh = y,  y,
},
2 ~ h =? y , + y,  2y, Hence, o n substitution tve get
3
T o obtain Simpson's rule, we apply the above result t o successive pieces of the curve y = f(u) betmeen \ = a and \  b. For this n c divide [a, b] into n subintervals each 01. width h. Then we approximate the portion of the curve in each pair of subintervals by a n arc of a parabola passing through the end points of that portion of the curve and the point corresponding t o the common point (Fig. 17) of those subintervals. Now consider the first two subintervals [a, s,] and I\,.\,I. The points on the curve corresponding t o a , \ , and \ ? arc P,,, P I and P:, rcspectively. Let us draw a parabola passing through P,,, P I and P2. We will assume that the portion of the curve passing through P,,, P , and P, corncides with this parabola. See F!g. 17.
I
Fig. 17
Similarly we can approximate the portion of the curve in the interval [xz, x,] by a parabola passing through the points P2, P3 and P4. We shall repeat this process f o r the remaining pairs of intervals. Now the area under the parabola Po P I P2 is given by h A, =  (yo + 4y, + y,). (On using formula (7)).
.
3 Similarly, the area under the parabola passing through the points PZ, P3, P4 is given by h A3 =  [Y, + 4 ~ 3 Y,] + ! Next, we use formula (7) for the parabola passing through the points P4, P5, P6 and get the area
A, =
 [y, + 4y5
h 3
+
y6], and so on.
Note, that to approximate the whole area under the given curve in this manner, the number n of the subdivisions of the interval [a, b] must be even. Summing all the areas we obtain
Area Under a Curve
as the total area. The above formula can also be written in the form
where we have taken n =2m. Note that yo, y,, Y ~ , . . . ,are~the values of the Y ~ function, f, for x = a, a + h, a + 2h, ...,a + 2mh = b, respectively. Hence by Simpson's rule, we have
h
The following example will help us compare the accuracy of the trapezoidal and Simpson's rules. You will find that Simpson's rule is more accurate than the trapezoidal rule.
Example 7: Taking four subdivisions of the interval [ l , 31, let us find the
approximate value of
\ x2dx by the trapezoidal rule and also by Simpson's rule.
Division by 4 gives 0.5 as the width of each subinterval of [ l , 31. The values of the integrand at these points of subdivision are given in the following table:
Using the formula for the trapezoidal rule, we obtain
3
Using the formula for Simpson's rule, we get
3
.
...(b)
The actual value of the definite integral
On comparing (a) with the actual value in (c), we observe that the value given by the trapezoidal rule has an error of + 0.08. Comparison of (b) with (c) shows that the error in the value given by Simpson's rule is zero in this case.
Example 8: Let us use the trapezoidal rule with six subdivisions to evaluate
I
C
x
F
x dx. We shall also find the value of the integral by using Simpson's rule.
0
I
cmtioas of Calculus
We have
I
I
Using trapezoidal rule, we obtain
Using Simpson's rule, we have
T
Example 9: A river is 80 m wide. The depth d in meters at a distance x m from one bank is given by the following table:
Let us find, approximately, the area of crosssection. Applying Simpson's rule with n = 8, we obtain the area uT crosssection
As we have seen i'n this example, Simpson's rule is very useful in approximating the area of'irregular figures like the crosssections of lakes and rivers. See if you can do these exercises now.
E E 13) a) Use Simpson's rule to evaluate the following, taking the given value of n.
sin x dx, n = 4
b) From the formula
H
4
0
dx , calculate H, using Simpson's rule with h = 0.1. l+xZ
E E 14) A curve is drawn through the points (1, 2), (1.5, 2.4), (2, 2.27), (2.5, 2.8),
(3, 3), (3.5, 2.6) and (4, 2.1). Estimate the area between the curve, the xaxis and the ordinate x = 1 , x = 4.
Area Lnder a Curve
EE

 
 



15) A river has width 30 meters. I f the depth y meters at distance x meters from one bank be given by the table,
Find the a ~ n r o x i m a t earea of crosssection.
E E 16) T h e ,velocity of a train, which starts from' rest, is given by the following
table, t h e time being reckoned in minutes f r o m the start and the speed in
Min
2 10
4
6
8
10
Kms/hr
18
25
29
32
20
11
5
2
Fcti~bntennnrnximatelv the total distance run in 23 m i n u t e 1
41
Applications of Calculus
a
Now let us quickly recall what we have done in this unit.
15.4 SUMMARY
In this unit we have covered the following points:
1) The knowledge 01' integration is helpful in finding areas enclosed by plane curves when their equations are known in ,
.
f
2
b
a) Cartesian form: A
=
y dk
b) Polar form:
A
=I [r2
B
dB,
2) The area bounded by a closed curve given by parametric equations is
B
3) When the integral cannot be exactly evaluated, we can use the method of numerical integration. The two methods given here are: a) Trapezoidal rule:
b
where n is the number of subdivisions of [a, b], and LAthe length of each subinterval. b) Simpson's rule: h f(x) dx =  [ ( ~ 0 + ~ 2 m ) + 2 ( ~ 2 + ~ 4 +    + ~ ? t2 ;) i + 4 ( ~ 1 + ~ 3 + .   + ~ 2 m  l ) I 3
a
when [a, b] is divided into 2m subintervals of length h.
15.5 SOLUTIONS AND ANSWERS
EI)
1
0
sin x dx
= 
cos
XI;
=  C O S T
+cosO= 1
+
1 =2.
Area Under a cur";
E5) Points of intersection of
4a/m2
y2 =4ax
and y = mx are (0,O) and (4a/m2, 4a/m).
X3/2 4a/mz
4a/mz
: .
mX2
4a/m2
A = [Z&dx.~mxdx
0 0
=
2 C 3/2Io
TIo
E6) For a loop, we should find two distinct, consecutive values of 8, for which we get the same value of r. If r, = a sin 38,, and r2 = a sin 3d2, r l = r2 * sin 3d1  sin 302 = 0. or 2 cos +82) sin 3(@1 82) 2 2
=
0
Thus, 8, = 0 and O2 = uL3 will give the same value of r.
*/3
Applications of Calculus
E8) Points of intersection are given by 8 cos 28 = 4 ar cos 2 3 = 112, i.e., cos28 = 314 1 7C 5C 7C 117C 7 7 . . e=' 6 ' 6 ' 6 6 Because of symmetry w.r.t. the initial line, the required area A = 2 (the area under the lemniscate above the initial line from 8 =0 to 8 = 7c/6 and from 8 = 5W6 to 8 = 7~ minus the area under the wide from 8 = 0 to 8 = d 6 and from 8 = 5 s to B = ~1
E9) x = a(3 sine  sin38), y = a cos38
= 12a2 
5 3 1 . 6 4 . 2
.
u .  (reduction formula)
2

dx =  a sin 0 de
+
dy b cos 8,  =  a'sin dB
e
+ blcos
e
dy dx x  y = (ab'  ba') dB de
+
(ac'  ca') sin 0
+
(cb'  bc') cos 0
El 1) A loop of this curve lies between t = O and t = u
1$ 1
T T
5 '
x
=
a'
sin 2t cor t dt
=
2a2
0
0
S
0
sin t cos2t dt
*/2
=
4a2
S
0
sin t cos2t dt
Area Under a Curve
sinx X
I
1
/
I
1
I
I
4/\12~
2/r
4/\i~n
o
..
(
X
sin x
12
1 + 0 + 2(2/i~) 4(4/fi7r + 4/\IZ?r) +
I
Now 7r/4 = 0.7846
*
r = 3.1384.
= 0.5(15.12) = 7.56 (trapezoidal rule).
=
 (7.4+ 16.8)
5 3
=
 (24.2)
5 3
=
(Simpson's rule).
121 3
UNIT 16 FURTHER APPLICATIONS OF INTEGRALCALCULUS
Structure
1'6.1 .Introduction
Objectives
47
47
16.2 Length of a Plane Curve
Cartesian Form
Parametric Form Polar Form
16.3 16.4 16.5 16.6
Volume of a Solid of Revolution Area of Surface of Revolution Summary Solutions and Answers
56
63 69 70

6 . 1 INTRODUCTION
In the last unit we have seen how definite integrals can be used to calculate areas. In fact, this application of definite integrals is not surprising. Because, as we have seen earlier, the problem of finding areas was the motivation behind the definition o f integrals. In this unit we shall see that the length of an arc of a curve, the volume of a cone .and other solids of revolution, the area of a sphere and other surfaces of revolution, can all be expressed as definite integrals. This unit alsb brings us to the end of this course on calculus.
Objectives
After reading thie unit, you should be able to: find the length of an arc of a given curve whose equation is expressed in either the Cartesian or parametric or polar forms, a find the volumes of some solids of revolu"tion, find the areas of some surfaces of revolution.
  
16.2 LENGTH OF A PLANE CURVE 

In this section we shall see how defihite integrals can be used t o find the lengths of plane curves whose equations are given in the Cartesian, polar or parametric form. A curve whose length can be found is called a rectifiable curve and the process of finding the length of a curve is called rectification. You will see hereJhat to find the length of an arc of a curve, we shall have to integrate an expression which involves not only the given function, but also its derivative. Therefore, to ensure the existence of the Integral which determines the arc length, we make an assumption that the function defining the curve is derivable, and its derivative is also continuous on the interval of integration.
Let's first consider a curve whose equation is given in the Cartesian form.
16.2.1 Cartesian Form
Let y = f(x) be defiiled on the interval [a, b]. We assume that f is derivable and its derivative f ' is continuous. Let us consider a partition P, of [a, b],, given by P, = [a=xo < x, < x2 < ...... < x n = b J The ordinates x = a and x = b determine the extent of the arc AB of the curve y = f(x) [Fig. l(a)]. Let M, = 1, 2, ......, n1, be the points in which the lines x = xi meet the curve. Join .. the succcss;ve points A, M I , M2, M3, ......, Mn_,,B by straight line segments. . ..:
:&>
.L
:
L.
 r I:
Furlher Applications of lnlegral Calculus
I
(a)
Fig. 1
If we can find the length of each line segment, the total length of this series will give us an approximation to the length of the curve. But how do we find the' length of any of these line segments? Take M2 M3, for example. Fig. I(b) shows an enlargement of the encircled portion in Fig. ](a). Lookjng at it we find that
fi
M2"3=fiX3)2+(
A
Y3l2,
where A x3 = M2Q is the length (x3  x2), and A ~3 =M3Q = f(x3)  f(x2)=y3 . ~ 2 . In this way we can find the lengths of the chords AM,. M1M2, ......, Mn,B, and take their sum
S , gives an approximation to the length of the arc AB. When the number of division points is increased indefinitely, and the length of each segment tends to zero, we obtain the length of the arc AB as L A~ . = lim nw r ~ x ~ ) ~ + ( A y ~ ) ~ ,
i=l
2
provided this limit exists. Our assumptions that f is derivable on [a, bj, and that f ' is continuous, permit us to apply the mean value theorem [Theorem 3, Unit 71. Thus, there exists a point pi' (xi, y,:) between the points Mil and Mi on the curve, where the tangent to the curve is parallei to the chord Mi,Mi. That is,
Hence we can write (1) as
A LB
= nco
lim
d( A

[fl(x;)
A
xi12
This is nothing but the definite integral
Applications of Calculus
l'herefore
Remark 1 It is sometimes convenient to express x as a single valued function of y. : In this case we interchange the roles of x and y, and get the length
where the limits of integration are with respect to y. Note that the length of an arc of a curve is invariant since it does not depend on the choice of coordinates, that is, on the frame of reference. Our assumption that f ' is continuous on [a, b] ensures that the integrals in (2) and (3) exist, and their value L is the length : of the curve y = f(x) hetween the ordinates x = a and x = b. The folIowing example illustrates the use of the formulas given by (2) and (3). Example Is Suppose we want to find the length of the arc of the curve y =lnx intercepted by the ordinates x = 1 and x = 2. We have drawn the curve y =lnx in Fig. 2.
Fig. 2
Using (2), the required length L: is given by
=
J(l+
i )
dy cix, since dx
=
1 x
dx t If we put 1 + x2 = t 2 , we get  = , and therefore, x dt
Further Applicationh of Integral Cnlcul~~.,
We can also use (3) to solve this example. For this we write the equation y =lnx as x = ey. The limits x = 1 and x = 2, then correspond to the limits y = 0 and y = 1112, respectively. Hence, using (3), we obtain
u du, on putting 1 + e2y= u 2 u2  1
A
as we have seen earlier. This verifies our observation in Remark 1 that both (2) and (3) give us the same value of arc length. Now here are some exercises for you to solve.
E E f) Find the length of the line x=3y between the points (3, 1) and ( 6 , 2). Verify
your answer by using the distance formula.
I
I
E 3) Find the length of the arc of the catenary y = C cosh (xtc) measured from the vertex (0,c) to any point (x, y) on the catenary.
I
I
E E 4) Find the length of the semicubical parabola a y 2 = x 3from the vertex to the
point (a, a).
E
5 ) Show that thc length of the arc of the parabola y2=4ax cut off by the line
3y=8x is
Further Applications of Integyl Calculus
In the next subsection we shall consider curves whose equations are expressed in the parametric form.
16.2.2 Parametric Form
Somet~mes equation of a curve cannot be written either in the form y =f(x) or the In the form x = g(y). A common example is a circle x2t.y2 = a2. In such cases we try to write the equation of the curve in the parametric form. For example, the above circle can be represented by the pair of equations x = a cos t, y = a sin t. Here we shall derive a formula to find the length of a curve given by a pair of parametric equations. Let x = +(t), y = $(t), a 5 t I p be the equation of a curve in parametric form. As in the previous subsection, we assume that the functions 4 and $ are both derivable and have continuous derivatives and $ < on the interval [a, We 81. have t
+'
I
dy $'s) , , Hence,  =  a dx 4'0)
I
Now, using (.3), we obtain the length
Applications. of Calculus
The following example shows that sometimes it is more convenient to express the equation of a given curve i s t h e parametric form in order to find its length. Example 2: Let us find the whole length of the curve
By substitution, you can easily check that x = a cos3t, y b sin3t is the parametric form of the given curve. The curve lies between the lines x = + a and y = + b since 1 c cos t 5 1, and 1 I sin t .c 1 . The curve is symmetrical about both the axes since its equation remains unchanged if we change the signs of x and y. The value t = 0 corresponds to the point (a, 0) and t = ?r/2 corresponds to the point (0, b). By applying the curve tracing methods discussed in Unit 9 we can draw this curve (see Fig. 3).
I
Fig. 3
Since the curve is symmetrical about both axes, the total length of the curve is four times its length in the first quadrant. dx dy NOW,  = 3a cos% sin t ; dt dt
=
3b sin% cos t
Hence, the length of the curve is
*/2
L = 4
1
0
3 sin t cost Ja2 cos%
+

b2 sin2t dt
*/2
= 12
[ lint cost /wb= dt
,
Putting u2= a 2 cost%+ b2sin%, we obtain dt 2u = (2b2  2a2) sin t cost , du and the limits t = 0,t = r / 2 correspond to
u = a, u = b, respectively.
Thus, we have
b
Further Applications of Iptegral Calculus
12
b2a2'
b3  a 3  4(a2+ b 2 + ab) 3 a+b
Vow you can apply equation (4) to solve these exercises.
E E 6) Find the length of the cycloid
E E 7) Show that the length of the arc of the curve
x = etsin t, y = etcos t from t = 0 to t = r/2 is fi(e"I2  1).
.A
16.2.3 Polar Form
In this subsection we shall consider the case of a curve whose equation is given in the polar form. Let r = f(9) determine a curve as 9 varies from 9 = a to 9 = p, i.e., the function f is defined in the interval [a, (see Fig. 4). As before, we assume that the function f 01 is derivable and its derivative f' is continuous on [a, This assumption ensures B]. that the curve represented by r = f(9) is rectifiable.
Applialions of Calculus
Fig. 4
X
Transforming the given equations into Cartesian coordinates by taking x = r cos 8, y = r sin 8, we obtain x = f(8) cos 8, y = f(8) sin 8. Now we proceed as in the case of parametric equations, and get.
Hence, the length of the arc of the curve r = f(8) from 8 = a to 8 = P is given by
x = f(4) cos
p
P
,
changing the variable x to 8.
P
sin 912
a
+ [ft(8) sin 8 +
f(8) cos 812 dB
We shall apply this formula to find the length of the curve in the following example.
Example 3: T o find the perimeter of the cardioid r = a(l + cos 8) we note that the curve is symmetrical about the initial line (Fig. 5). Therefore its perimeter is double the length 'of the arc of the curve lying above the xaxis.
Fig. 5
dr Now,  = de
 a sin 8. Hence, we have
Further Applications of Integral Calculus
=
4a 2 sin 
In this section we have derived and applied the formulas for finding the length of a curve when its equation is given in either of the three forms: Cartesian, parametric or polar. Let us summarise our discussion in the following table.
Table 1: Length of ap arc of a curve
I :1:
=
8a.
Length L
1
Using this table you will be able to solve these exercises now.
E
E 8) Find the length of the curve r = a cos3 (0/3).
r
E
E 9) Find the length of the circle of rddius 2 which is given by the equations x = 2 c o s t + 3 , y = 2 sin t + 4 , 0 It 5 2a.
E E 10) Show that the arc of the upper ha!f
bv 8 = 2?r/3.
of the curve r = a(l  cos 8) is bisected
A

16.3 VOLUME OF A SOLID OF REVOLUTION
Until now, in this course. we were concerned with only plane curves and regions. In this section we shall see how our knowledge of integration can be used to find the volumes of certain solids. Look at the plane region in Fig. qa). It is bounded by x = a, x = b, y = f(x) and the xaxis. If we rotate this plane region about the xaxis, we get a solid. See Fig. Mb).
Further Applicatinns of Integral Calc~~lu!.
Fig. 6
Such solids are called solids of revolution. Fig. 7(a) and Fig. 7(b) show two more examples of solids of revolution.
Fig. 7
The solid in Fig. 7(a) is obtained by revolving the region ABCQ around the yaxis. The solid of revolution in Fig. 7(b) differs from the others in that its axis of rotation does not form a part of the boundary of the plane region PQRS which is rotated.
Wc see many examples of solids of revolution in every day life. The various kinds of pots made by a potter using his wheel are solids of revolution. See Fig. 8(a). Some objects manufactured with the help of !athe machines are also solids of revolution. See Fig. 8(b).
Fig. 8
f; 7
I
I
l
Applications of
C~ICU~US
Now, let us try to find the volume of a qolid of revolution. The method which we are going to use is called the method of slicing. The reason for this will be clear in a few moments. Let Tn = [a = x, < xl < x2 < .... < x..,, x,, = b] be a partition of the interval [a, b] into n subintervals.
I
~ P M MN. is the volume of the ~ disc with radius P M and thickness MN. ~ Q N ~ M N volume of the is the .disc with radius Q N and thickness MN. If f is continuous.on [a, b], f(a) = c and f(b) = d , and z lies between c and d, then 3xo € ] a , b [ s.t. f(x,)=z.
(b)
Fig. 9
Let Axi denote the length of the ith subinterval [xi,, xi]. Further, let P and Q be the points on the curve, y = f(x) corresponding to the ordinates x = xi, and x = xi, respectively. Then, as the curve revolves about the xaxis, the shaded strip PQNM (Fig. 9(a)) generates a disc o f thickness Axi. In general, the ordinates P M and QN may not be of equal length. Hence, the disc is actually the frustum of a cone with and n Q N ~ M N , that is, between its volume Avi, lying between r P M ~ M N (Fig. r [ ~ ( X ~ _ ~ and r [ ~~( X , ) ] ~ A ~ 9(b) and (c)) )]~AX If we assume that f is a continuous function o n [a, b], we can apply the intermediate value theorem (Theorem 7, Unit 2, also see margin remark), and express this volume as a v i = r[f(ti)) OX^, where ti is a suitable point in the interval [xi,, xi]. Now summing up over all the discs, w e ob~ain
n
n
Vn =
i= 1
A Vi =
1
7
[f(tj)12 AX^,
xi1 5 ti 5 xi, as an approximation
to the volume of the solid of revolution. As we have observed earlier while defining a definite integral, the approximation gets better as the partition Pn gets finer and finer and Ax, tends to zero. Thus, we get the volume of the solid of revolution as
n
V = lim Vn = lim
nco
noa.
1=1
r [f(ti)12Axi
We shall use this formula t o find the volume of the solid described in the following example. Example 4: Let us find the volume of the solid of revolution formed when the arc of the parabola y 2 = 4ax between the ordinates x = 0, and x = a is revolved about its axis. The solid of revolution is the parabolic cap in Fig. 10. The volume V of the cap is given by
Fig. 10
Our next example illustrates a slight modification of Formula (6) to find the volume of a solid obtained by revolving a plane region about the yaxis. x2 Y Z ,+ b = 1, (a > b) is revolved about the minor a axis, AB (see Fig. 11). Let us find the volume of the solid generated.
Example 5: Suppose the ellipse 7
In this case the axis of rotation is the yaxis. The area revolved about the yaxis is shown by the shaded region in Fig. 11. You will agree that we need to consider only the area to the right of the yaxis.
Fig. 11
.
To find the volume of this solid we interchange x and y in (6) and get
b
b
= 2n2
[(I$) 1
0
Y .s c~y since I  7 an even function of y? b
2
Applications o C~lculus f
We can also modify Formula (6) to apply to curves whose equations are given in the parametric or polar forms. Let us tackle these one by one.
Parametric Form
If a curve is given by x = 4(t), y = $(t), a 5 t 5 8, then the volume of the solid of revolution about the xaxis can be found by substituting x and y in Formula (6) by @(t)and $(t),respectively. Thus,
3
or V
=
a
i
[+(t)l2 +'(t) dt.
We'll now derive the formula for curves given by polar equations.
Polar Form
Suppose a curve is given by r = f(O), 0, 5 0 I 0'. The volume of the solid generated by rotating the area bounded by x = a, x = b , the xaxis and r = f(0) about the xaxis is
Y
=
a
S
9,
(r sin 0)l
d do
7
(r cos 0) dB
Thus, V = n
9
0,
[f(fl) sin 012 [ff(6) cos 8  f(8) sin P ] dB
Let's use this formula to find the volume of the solid generated by a cardioid about its iniiial line.
Example 6: The cardioid shown in Fig. 12 is given by r = a ( ] +cos 8).
Fig. I 2
The points A and 0 correspond t o 0 = 0 and 0 = a, respectively. Here, again, we need t o consider only the part of the cardioid above the initial line. Thus,
Y =
1
0
r (r sin 0)'
 (r cos
do
d
8 ) d0
=
1
(1 +cos 0)' sin3 8 (1 + 2
COP
P ) do, since r = a ( l
+ cos 8)
I
rr ?
t
=
256ra3
[ sini@ cosV$ d g  6 4 r a 3 [ sin3+ cos74 d+, where 4 ~ 6 1 2
5
a/2
 64.1ra3  '*a' 
IS
on applying a reduction formula from Unit 12.
In all the examples that we have seen till now, the axis o f rotation formed a boundary of the region which was rotated. Now we take an example in which the axis touches the region at only one point.
Example 7: Let us find the volume o f the solid generated by revolving the region bounded by the parabolas y = x 2 and y 2 = 8 x about the xaxis. We have shown thearea rotated and thc solid in Fig. 13 (a) and (b), respectively.
Here, the required volume will be the difference betyeen the volume of the solid generated by the parabola y 2 = 8 x and that of the solid generated by the parabola y =x 2.
Note the limits of integration. Here, we list some exercises which you can solve by applying the formulas derived in this section.
E E 12) Find the volume o f the right circular cone of height h and radius of the
circular base r.
(Hint: cone will be generated by rotating the triangle bounded by This
the Y  a x i ~ and the line v = (r/h)
X'I
I
I
Applications of Calculus
E 13) S h o ~ that the volume of the solid generated by revolving the curve x?,3 y 2 / 7  ,2 3 about the xaxis is 32*a3/105.
+
E 14j The arc of the cycloid x = a (tsin t), y = a (Icos t) in [0, 2x1 is rotated about the yaxis. Find the volume generated. (Caution:The rotation is about the yaxis.)
E E 15) Find the volume of the selid obtained by revolving the lirnacon
r = a + b cos 8 about thk initlal line.
Further. Applications of Inlegral Calculus
Il 16) The semicircular region bounded by y2 = h
z 2 the line Y = 2 is and
rotated about the xaxis. Find the volume of the solid generated.
16.4 AREA OF SURFACE OF REVOLUTION
Instead of rotating a plane region, if we rotate a curve about an xaxis, we shall get a surfaceof revolution. In this section we shall find a formula for the area of such a surface. Let us start with the case when the equation of the curve is given in the Carlcslan form.
Cartesian Form
Suppose that the curve y = f(x) [Fig. 141 is rotated about the xaxis. To find the , area of the geneiated surface, we consider a partition P of the interval [a, b], namely, P = [a=$ ixl < x2 < ...... < xnI < x, bj ,

Applications of Calculus
Fig. 14
Let the lines x = xi intersect the curve in points Mi, i = 1, 2, ..... n. If we revolve the chord Mi_lMiabout the xaxis, we shall get the surface of the frustum of a cone of thickness Ax, = x,  x,+,. Let As, be the area of the surface of this frustum. . Then the total surface aria of all the frusta is
n
Sn =
1=
As,
1
This Sn is an approximation to the area of the surface of revolution. The area of the surface of revolution generated by the curve y = f(x), is the limit of if it exists), as n  oo and each Qx,. 0.

Fig. 15
To find the area A of the curved surface of a typical frustum, we use the formula A=7r(rl+rz) 1, , where 1 is the slant height of the frustum and r l and r2 are the radii of its bases (Fig. 15). In the frustum under consideration the radii of the bases are the ordinates f(x,,) and f(xi), while the slant height MilMi is given by wheie A yi= f(xi)  f ( ~ ~ _ ~ )assume that f is derivable on [a, b] and f ' is We . continuous. Then by the meal1 value theorem (Theorem 3, Unit 7), we obtain A yi = f '(ti) A xi, for some ti E xi] . Therefore,
AS,
= T =
[ f ( ~ ~+ ~f(xi)l  )
Yi
+
Jm+ yi12
(A
A
2x Yil 2
m '(ti)12 f
yi
xi.
A
i= 1
2
Proceeding to the limit as n
S
=
2~
i
a
f ( ~ ) / W d x
= 27r f y d w d s z dx
We shall now illustrate the use of this formula.
+
and Sn = 2 7 r z yi'
xi.

oo, and each A xi
 0, we have
Example 8: Let us find the area of the surface of revolution obtained by revolving the parabola y2= 4ax from x = a to x = 3a, about the xaxis.
Further Applications of Integral Calculus
Fig. 16
The area of the surface of revolution
where
L
Y2
= 4ax,
dy 2a  = . dx
Hence
Y
J Instead of.revolving the given curve about the'xaxis, if .we revolve it about the yaxis, we get another surface of revolution. The area of the surface of revolution generated by the curve x = g(y), c Iy Id , as it revolves about the yaxis is given
j
L
Now let us look at curves represented by parametric equations.
Parametric Form
Suppose a curve is given by the parametric e~uations = +(t), y = $ (t), t E [a, x 01. Then, we know that
Applications of Calculus
Substituting this in formula (lo), we get the area of the surface of revolution generated by the curve as jt revolves about the xaxis, to be
S = 2~
\ J.0) * J
+
[ ~ . ~ ( tdt.] ~ )
Now we shall state the formula for the surface generated by a curve represented by a polar equation. Polar Form If r = h(8) is the polar equation ,of the curve, then the area of the surface of revolution generated by the arc of the curve for 8 , I 8 I 82, as it revolves about the initial line, is S
=
2~
6,
3
r sin 8
p+(dr/do)2 de
Study the following examples carefully before trying the exercises given at the end of this section.
Example 9: Suppose the astroid x = a sin3t, y = a cos3t, is revolved about the xaxis. Let us find the area of the surface of revolution. You will agree that we need to consider only the portion of the curve above the xaxis.
For this portion y > 0, and thus t varies from a/2 to r/2.

dx dy = 3a sin2t cos t,  =  3a cos2t sin t dt dt
Therefore.
($)i +
(%)2
= 9a2 sin2t cos2t
We therefore get,
a/2
S = 2a
a/2
*/2
S
a cos3tJ9a2 sin2t cos2t dt
, S
= 2r
5
a cos3t
*/ 2
I
3a sin t cos t
I
dt
*/2
= 6aa2
*/2
= 12aa2
S S
0
cos4t ( sin t
'
I
. dt
= 
s/2
cos4t sin t dt
127ra2 
L
':I
Example 10: Suppose we want to find the area of the surface generated by revolving the cardinid r = a ( l +cos 6 ) about its initial line.
Notice that the cardioid is symmetrical about the initial line, and extends above this line from 8 = 0 to 8 = a . The surface generated by revolving the whole curve about the initial line is the same as that generated by the upper half of the curve. Hence
= 2r
1
0
a (1 +cos 8) sin
e
Forther Applications of Integral Calculus
Since r r2
=
+
($)2
dr a (1 + cos 8). and  =  a sin 8, we have d8 8 = a2(l + cos e)2 + a 2 sin2 8 = 4a2 cos 2 2
Therefore. S
=
2r
0
S
8 a (1 +cos 8) sin 8 2a cos  d8 2
=
4ra2
S
8 8 4 sin  cos4  d8 2
=
32na2
1
sin+ cos4+ d+, where
+= 8/2
E E 17) Find the area of the surface generated by revolving the circle r = a about the
xaxis, and thus verify that the surface area of a sphere of radius a is 4na2.
E E 18) The arc of the curve y = sin x, from x = 0 to x = .r is revolved about the
xaxis. Find the area of the surface of the solid of revolution generated.

Applkatioos of Calculus
E E 19) The ellipse x 2 / a 2 +y 2 / b 2 = 1 revolves about the xaxis. Find the area of the
curface of the sqlid of revolution thuc obtained.
E 20) Prove that the surface of the solid generated by the revolution about the xaxis of the loop of the curve x = t2,
E E 21) Find the surface area of the solid generated by revolving the cycloid
x = a(0  sin 0). v = a(l  cos 0). about the line v =O.
'
Further Applications of Integral Calculus
Now let us quickly recall what we have covered in this unit.
16.5 SUMMARY
In this unit we have seen how to find
i). the lengths of curves
2) volumes of solids of revolution and
3) the areas of surfaces of revolution. In each case we have derived formulas when the equation of the curve is given in either the Cartesian or parametric or polar form. We give the results here in the form of the following tables.
Length of an arc of a curve
Equation
Length
Volume of the solid of revolution
I
I
~quation
I
Volume
1
Y = f(x) about xaxis
x = g(y) , about yaxis
x = 4(t), Y = rC/(t) about xaxis '
I
I
r = h(0) about the initial line
r
sin 01' [hl(0) cus 0
 h(0) sin 01 d0
Applicalions of Cakulus
Area of t h e surface of revolution
I
Equation
I
Area
x = g(y) about yaxis
x = tJ(t), Y = $(I) about xaxis
16.6 SOLUTIONS AND ANSWERS
El) L
=
1
d
C
n
jl+(dx/dy)'
dy
2
= 0
1
S
dy =
Fo.
By distance formula,
1 sec x
. sec x tan x
= tan x
40

i'
set
x dx = In
I sec x
+ tan x
(
s e c ~ / 3+ tan */3 ' sec 0 + tan 0
Funher Aaolications of lntegml Calculus
. c sinh (x/c) =
I:
= c sinh (xlc)
E5) 3y
= 8x = y =
. Substituting this in
64x2    4ax
9
8x 3
y2 =
4ax we get
i.e. 64x2  36ax = 0 9a * x = Oorx = 16 3a * y = Oory = 2
Hence (0, 0) and
are the points of intersection.
E6) 
dx
ria
= a(l
dy  cos O),  ne
a sin
o
..
($)' (2;)'
+
= a 2 [I
+
cos2e  '2 cos B
+
sin28]
... L
=
2a
0
5
0
sin (6/2) dB
= 4a
[ sin 4 d 9
J */2
= 8,
[ sin 4 d+
= 8a
E7)
dx dt
=
dy e' (cos t +'sin t), dt
=
e' (COS  sin t) t
E8) r
= a cos3
 =,
e

3
dr =  a cos2  sin 6 e dB 3 3
dx E9) dt
= 
2 sin t

dy = 2 cos t dt
Note that L ~ 1 0 r)
=
2ar since, here, r = 2. dr a(l  cos O),  = a sin 6 dB
=
.
J.2
+
($y
=
2a sin 2
e
The length of the curve in the upper half =
1
2a sin ( ! ) B2
dB
Furlher Applicalions of IntegraI~Cnlculus
the volume generated by
b15) r
=
a
+
bcos8
 =
dx (19
d(r cos 8) d8
! sc,s ,
( a + b cos C) ( sin 0)  k sin =  a sin 8  2b sin 8 cos 0.
3

= r
1
0
( a + b cos B ) sin28 (a sin 8 ~
+
2b sin 0 cos 8) dB
=
*
1
0
(a3sin38+ 4a2b sin3t9 cos 8 + 5ab2 rin38 cos28+ 2b3 sin* cos3 8) do
(~hk other
=
. [
two integrals are equal t o zero since cos
(T
4a 3
+
10ab2 .
2.
5
1
(1
 8) =
 cos 8.
sin 8 cos28 d~
. (using a reduction formula)
I
= 4ra'
( sin 8 dB
since sin (r
.
8)
=
sin 8
= 47ra2 cos 8
I:"
E18) y
=
sinx

dy dx
=
cosx
Further Applications of Integral Calculus
=
4r
i
J1
+
t2 dt, if t
=
cos x
 4rb
a2
= 2rb2
a 2 (a2  b2) 2ra2b
sin'
x
4757,
a2 .
+
7XiF
sin'
&2:72
a.
E20) The loop is between t =  0 and t = 0.Because,of symmetry, it is' enough to consider the curve between t = 0 and t = 6.
. :
S =. 2 r
J (t  ) :
0
A
~;i;(~t'? dt
Applications of Cskulus
a ( I  cos 8) da2(1  cos 8)2
8 4 m 2 ( ( 1  cos 0) sin  do 2
+
a2 sin2 1 dI9 9
=