Introduction to Calculus

Volume I
by J.H. Heinbockel
 The regular solids or regular polyhedra are solid geometric figures with the same identical regular
polygon on each face. There are only five regular solids discovered by the ancient Greek mathematicians.
These five solids are the following.
the tetrahedron (4 faces)
the cube or hexadron (6 faces)
the octahedron (8 faces)
the dodecahedron (12 faces)
the icosahedron (20 faces)
Each figure follows the Euler formula

Number of faces + Number of vertices = Number of edges + 2

F + V = E + 2
Introduction to Calculus

Volume I

by J.H. Heinbockel
Emeritus Professor of Mathematics
Old Dominion University
 c Copyright
2012 by John H. Heinbockel All rights reserved
Paper or electronic copies for noncommercial use may be made freely without explicit
permission of the author. All other rights are reserved.

This Introduction to Calculus is intended to be a free ebook where portions of the text
can be printed out. Commercial sale of this book or any part of it is strictly forbidden.

ii
 Preface
This is the first volume of an introductory calculus presentation intended for
future scientists and engineers. Volume I contains five chapters emphasizing funda-
mental concepts from calculus and analytic geometry and the application of these
concepts to selected areas of science and engineering. Chapter one is a review of
fundamental background material needed for the development of differential and
integral calculus together with an introduction to limits. Chapter two introduces
the differential calculus and develops differentiation formulas and rules for finding
the derivatives associated with a variety of basic functions. Chapter three intro-
duces the integral calculus and develops indefinite and definite integrals. Rules
for integration and the construction of integral tables are developed throughout
the chapter. Chapter four is an investigation of sequences and numerical sums
and how these quantities are related to the functions, derivatives and integrals of
the previous chapters. Chapter five investigates many selected applications of the
differential and integral calculus. The selected applications come mainly from the
areas of economics, physics, biology, chemistry and engineering.
The main purpose of these two volumes is to (i) Provide an introduction to
calculus in its many forms (ii) Give some presentations to illustrate how power-
ful calculus is as a mathematical tool for solving a variety of scientific problems,
(iii) Present numerous examples to show how calculus can be extended to other
mathematical areas, (iv) Provide material detailed enough so that two volumes
of basic material can be used as reference books, (v) Introduce concepts from a
variety of application areas, such as biology, chemistry, economics, physics and en-
gineering, to demonstrate applications of calculus (vi) Emphasize that definitions
are extremely important in the study of any mathematical subject (vii) Introduce
proofs of important results as an aid to the development of analytical and critical
reasoning skills (viii) Introduce mathematical terminology and symbols which can
be used to help model physical systems and (ix) Illustrate multiple approaches to
various calculus subjects.
If the main thrust of an introductory calculus course is the application of cal-
culus to solve problems, then a student must quickly get to a point where he or
she understands enough fundamentals so that calculus can be used as a tool for
solving the problems of interest. If on the other hand a deeper understanding of
calculus is required in order to develop the basics for more advanced mathematical
iii
efforts, then students need to be exposed to theorems and proofs. If the calculus
course leans toward more applications, rather than theory, then the proofs pre-
sented throughout the text can be skimmed over. However, if the calculus course
is for mathematics majors, then one would want to be sure to go into the proofs
in greater detail, because these proofs are laying the groundwork and providing
background material for the study of more advanced concepts.
If you are a beginner in calculus, then be sure that you have had the appro-
priate background material of algebra and trigonometry. If you dont understand
something then dont be afraid to ask your instructor a question. Go to the li-
brary and check out some other calculus books to get a presentation of the subject
from a different perspective. The internet is a place where one can find numerous
help aids for calculus. Also on the internet one can find many illustrations of
the applications of calculus. These additional study aids will show you that there
are multiple approaches to various calculus subjects and should help you with the
development of your analytical and reasoning skills.

J.H. Heinbockel
September 2012

iv
 Introduction to Calculus
Volume I
Chapter 1 Sets, Functions, Graphs and Limits ....................1
Elementary Set Theory, Subsets, Set Operations, Coordinate Systems, Distance Be-
tween Two Points in the Plane, Graphs and Functions, Increasing and Decreasing
Functions, Linear Dependence and Independence, Single-valued Functions, Paramet-
ric Representation of Curve, Equation of Circle, Types of Functions, The Exponential
and Logarithmic Functions, The Trigonometric Functions, Graphs of Trigonomet-
ric Functions, The Hyperbolic Functions, Symmetry of Functions, Translation and
Scaling of Axes, Inverse Functions, Equations of Lines, Perpendicular Lines, Limits,
Infinitesimals, Limiting Value of a Function, Formal Definition of a Limit, Special
Considerations, Properties of Limits, The Squeeze Theorem, Continuous Functions
and Discontinuous Functions, Asymptotic Lines, Finding Asymptotic Lines, Conic
Sections, Circle, Parabola, Ellipse, Hyperbola, Conic Sections in Polar Coordinates,
Rotation of Axes, General Equation of the Second Degree, Computer Languages

Chapter 2 Differential Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

Slope of Tangent Line to Curve, The Derivative of y = f(x), Right and Left-hand
Derivatives, Alternative Notations for the Derivative, Higher Derivatives, Rules and
Properties, Differentiation of a Composite Function, Differentials, Differentiation
of Implicit Functions, Importance of Tangent Line and Derivative Function f 0 (x),
Rolles Theorem, The Mean-Value Theorem, Cauchys Generalized Mean-Value The-
orem, Derivative of the Logarithm Function, Derivative of the Exponential Function,
Derivative and Continuity, Maxima and Minima, Concavity of Curve, Comments on
Local Maxima and Minima, First Derivative Test, Second Derivative Test, Loga-
rithmic Differentiation, Differentiation of Inverse Functions, Differentiation of Para-
metric Equations, Differentiation of the Trigonometric Functions, Simple Harmonic
Motion, LHopitals Rule, Differentiation of Inverse Trigonometric Functions, Hyper-
bolic Functions and their Derivatives, Approximations, Hyperbolic Identities, Eulers
Formula, Derivatives of the Hyperbolic Functions, Inverse Hyperbolic Functions and
their Derivatives, Relations between Inverse Hyperbolic Functions, Derivatives of the
Inverse Hyperbolic Functions, Table of Derivatives, Table of Differentials, Partial
Derivatives, Total Differential, Notation, Differential Operator, Maxima and Minima
for Functions of Two Variables, Implicit Differentiation

Chapter 3 Integral Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

Summations, Special Sums, Integration, Properties of the Integral Operator, Notation,
Integration of derivatives, Polynomials, General Considerations, Table of Integrals,
Trigonometric Substitutions, Products of Sines and Cosines, Special Trigonometric
Integrals, Method of Partial Fractions, Sums and Differences of Squares, Summary of
Integrals, Reduction Formula, The Definite Integral, Fundamental theorem of integral
calculus, Properties of the Definite Integral, Solids of Revolution, Slicing Method,
Integration by Parts, Physical Interpretation, Improper Integrals, Integrals used to
define Functions, Arc Length, Area Polar Coordinates, Arc Length in Polar Coordi-
nates, Surface of Revolution, Mean Value Theorems for Integrals, Proof of Mean Value
Theorems, Differentiation of Integrals, Double Integrals, Summations over nonrect-
angular regions, Polar Coordinates, Cylindrical Coordinates, Spherical Coordinates,
Using Table of Integrals, The Bliss Theorem

v
 Table of Contents

Chapter 4 Sequences, Summations and Products . . . . . . . . . . . . . 271

Sequences, Limit of a Sequence, Convergence of a sequence, Divergence of a sequence,
Relation between Sequences and Functions, Establish Bounds for Sequences, Addi-
tional Terminology Associated with Sequences, Stolz -Cesaro Theorem, Examples of
Sequences, Infinite Series, Sequence of Partial Sums, Convergence and Divergence of
a Series, Comparison of Two Series, Test For Divergence, Cauchy Convergence, The
Integral Test for Convergence, Alternating Series Test, Bracketing Terms of a Con-
vergent Series, Comparison Tests, Ratio Comparison Test, Absolute Convergence,
Slowly Converging or Slowly Diverging Series, Certain Limits, Power Series, Opera-
tions with Power Series, Maclaurin Series, Taylor and Maclaurin Series, Taylor Series
for Functions of Two Variables, Alternative Derivation of the Taylor Series, Remain-
der Term for Taylor Series, Schlomilch and Roche remainder term, Indeterminate
forms 0 , , 00 , 0 , 1 , Modification of a Series, Conditional Convergence,
Algebraic Operations with Series, Bernoulli Numbers, Euler Numbers, Functions De-
fined by Series, Generating Functions, Functions Defined by Products, Continued
Fractions, Terminology, Evaluation of Continued Fractions, Convergent Continued
Fraction, Regular Continued Fractions, Eulers Theorem for Continued Fractions,
Gauss Representation for the Hypergeometric Function, Representation of Functions,
Fourier Series, Properties of the Fourier trigonometric series, Fourier Series of Odd
Functions, Fourier Series of Even Functions, Options,

Chapter 5 Applications of Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363

Related Rates, Newtons Laws, Newtons Law of Gravitation, Work, Energy, First
Moments and Center of Gravity, Centroid and Center of Mass, Centroid of an Area,
Symmetry, Centroids of composite shapes, Centroid for Curve, Higher Order Mo-
ments, Moment of Inertia of an Area, Moment of Inertia of a Solid, Moment of Inertia
of Composite Shapes, Pressure, Chemical Kinetics, Rates of Reactions, The Law of
Mass Action, Differential Equations, Spring-mass System, Simple Harmonic Motion,
Damping Forces, Mechanical Resonance, Particular Solution, Torsional Vibrations,
The simple pendulum, Electrical Circuits, Thermodynamics, Radioactive Decay, Eco-
nomics, Population Models, Approximations, Partial Differential Equations, Easy to
Solve Partial Differential Equations

Appendix A Units of Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452

Appendix B Background Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454

Appendix C Table of Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466
Appendix D Solutions to Selected Problems . . . . . . . . . . . . . . . . . . . 520
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552

vi
 1
Chapter 1
Sets, Functions, Graphs and Limits
The study of different types of functions, limits associated with these functions
and how these functions change, together with the ability to graphically illustrate
basic concepts associated with these functions, is fundamental to the understanding
of calculus. These important issues are presented along with the development of
some additional elementary concepts which will aid in our later studies of more ad-
vanced concepts. In this chapter and throughout this text be aware that definitions
and their consequences are the keys to success for the understanding of calculus and
its many applications and extensions. Note that appendix B contains a summary of
fundamentals from algebra and trigonometry which is a prerequisite for the study
of calculus. This first chapter is a preliminary to calculus and begins by introducing
the concepts of a function, graph of a function and limits associated with functions.
These concepts are introduced using some basic elements from the theory of sets.
Elementary Set Theory
A set can be any collection of objects. A set of objects can be represented using
the notation
S = {x | statement about x}
and is read, S is the set of objects x which make the statement about x true .
Alternatively, a finite number of objects within S can be denoted by listing the
objects and writing
S = {S1 , S2 , . . . , Sn }

For example, the notation

S = { x | x 4 > 0}

can be used to denote the set of points x which are greater than 4 and the notation

T = {A, B, C, D, E}

can be used to represent a set containing the first 5 letters of the alphabet.
A set with no elements is denoted by the symbol and is known as the empty set.
The elements within a set are usually selected from some universal set U associated
with the elements x belonging to the set. When dealing with real numbers the
universal set U is understood to be the set of all real numbers. The universal set is
2
usually defined beforehand or is implied within the context of how the set is being
used. For example, the universal set associated with the set T above could be the
set of all symbols if that is appropriate and within the context of how the set T is
being used.
The symbol is read belongs to or is a member of and the symbol / is
read not in or is not a member of . The statement x S is read x is a member
of S or x belongs to S . The statement y / S is read y does not belong to S
or y is not a member of S .
Let S denote a non-empty set containing real numbers x. This set is said to be
bounded above if one can find a number b such that for each x S , one finds x b.
The number b is called an upper bound of the set S . In a similar fashion the set S
containing real numbers x is said to be bounded below if one can find a number

such that
x for all x S . The number
is called a lower bound for the set S .
Note that any number greater than b is also an upper bound for S and any number
less than
can be considered a lower bound for S . Let B and C denote the sets

B = {x | x is an upper bound of S} and C = { x | x is a lower bound of S},

then the set B has a least upper bound (
.u.b.) and the set C has a greatest lower
bound (g.
.b.). A set which is bounded both above and below is called a bounded set.
Some examples of well known sets are the following.
The set of natural numbers N = {1, 2, 3, . . .}
The set of integers Z = {. . . , 3, 2, 1, 0, 1, 2, 3, . . .}
The set of rational numbers Q = { p/q | p is an integer, q is an integer, q
= 0}
The set of prime numbers P = {2, 3, 5, 7, 11, . . .}
The set of complex numbers C = { x + i y | i2 = 1, x, y are real numbers}
The set of real numbers R = {All decimal numbers}
The set of 2-tuples R2 = { (x, y) | x, y are real numbers }
The set of 3-tuples R3 = { (x, y, z) | x, y, z are real numbers }
The set of n-tuples Rn = { (1 , 2 , . . . , n) | 1 , 2, . . . , n are real numbers }

where it is understood that i is an imaginary unit with the property i2 = 1 and

decimal numbers represent all terminating and nonterminating decimals.
 3
Example 1-1. Intervals
When dealing with real numbers a, b, x it is customary to use the following no-
tations to represent various intervals of real numbers.

Set Notation Set Definition Name

[a, b] {x | a x b} closed interval
(a, b) {x | a < x < b} open interval
[a, b) {x | a x < b} left-closed, right-open
(a, b] {x | a < x b} left-open, right-closed
(a, ) {x | x > a} left-open, unbounded
[a, ) {x | x a} left-closed,unbounded
(, a) {x | x < a} unbounded, right-open
(, a] {x | x a} unbounded, right-closed
(, ) R = {x | < x < } Set of real numbers

Subsets
If for every element x A one can show that x is also an element of a set B ,
then the set A is called a subset of B or one can say the set A is contained in the
set B . This is expressed using the mathematical statement A B , which is read A
is a subset of B . This can also be expressed by saying that B contains A, which is
written as B A. If one can find one element of A which is not in the set B , then A
is not a subset of B . This is expressed using either of the notations A

B or B

A.
Note that the above definition implies that every set is a subset of itself, since the
elements of a set A belong to the set A. Whenever A B and A
= B , then A is called
a proper subset of B .
Set Operations
Given two sets A and B , the union of these sets is written A B and defined

A B = { x | x A or x B, or x both A and B}

The intersection of two sets A and B is written A B and defined

A B = { x | x both A and B }

If A B is the empty set one writes A B = and then the sets A and B are said to
be disjoint.
4
The difference1 between two sets A and B is written A B and defined

AB = {x | x A and x / B }

The equality of two sets is written A = B and defined

A= B if and only if A B and B A

That is, if A B and B A, then the sets A and B must have the same elements
which implies equality. Conversely, if two sets are equal A = B , then A B and
B A since every set is a subset of itself.

AB AB AB

Ac A (B C) (A B)c

Figure 1-1. Selected Venn diagrams.

The complement of set A with respect to the universal set U is written Ac and
defined
Ac = { x | x U but x
/A}

Observe that the complement of a set A satisfies the complement laws

A Ac = U, A Ac = , c = U, Uc =

The operations of union and intersection satisfy the distributive laws

A (B C) = (A B) (A C) A (B C) = (A B) (A C)
1
The difference between two sets A and B in some texts is expressed using the notation A \ B .
 5
and the identity laws

A = A, A U = U, A U = A, A=

The above set operations can be illustrated using circles and rectangles, where
the universal set is denoted by the rectangle and individual sets are denoted by
circles. This pictorial representation for the various set operations was devised by
John Venn2 and are known as Venn diagrams. Selected Venn diagrams are illustrated
in the figure 1-1.

Example 1-2. Equivalent Statements

Prove that the following statements are equivalent A B and A B = A
Solution To show these statements are equivalent one must show
(i) if A B , then A B = A and (ii) if A B = A, then it follows that A B .
(i) Assume A B , then if x A it follows that x B since A is a subset of B .
Consequently, one can state that x (A B), all of which implies A (A B).
Conversely, if x (A B), then x belongs to both A and B and certainly one can
say that x A. This implies (A B) A. If A (A B) and (A B) A, then it
follows that (A B) = A.
(ii) Assume A B = A, then if x A, it must also be in A B so that one can say
x A and x B , which implies A B .

Coordinate Systems
There are many different kinds of coordinate systems most of which are created
to transform a problem or object into a simpler representation. The rectangular
coordinate system3 with axes labeled x and y provides a way of plotting number
pairs (x, y) which are interpreted as points within a plane.
2
John Venn (1834-1923) An English mathematician who studied logic and set theory.
3
Also called a cartesian coordinate system and named for Rene Descartes (1596-1650) a French philosopher who
applied algebra to geometry problems.
6
rectangular coordinates polar coordinates

x = r cos , y = r sin
x2 + y 2 = 2
r=
tan = xy x2 + y 2 = r
Figure 1-2. Rectangular and polar coordinate systems
A cartesian or rectangular coordinate system is constructed by selecting two
straight lines intersecting at right angles and labeling the point of intersection as
the origin of the coordinate system and then labeling the horizontal line as the
x-axis and the vertical line as the y -axis. On these axes some kind of a scale is
constructed with positive numbers to the right on the horizontal axis and upward
on the vertical axes. For example, by constructing lines at equally spaced distances
along the axes one can create a grid of intersecting lines.
A point in the plane defined by the two axes can then be represented by a number
pair (x, y). In rectangular coordinates a number pair (x, y) is said to have the abscissa


x and the ordinate y . The point (x, y) is located a distance r = x2 + y 2 from the
origin with x representing distance of the point from the y-axis and y representing
the distance of the point from the x-axis. The x axis or abscissa axis and the y axis
or ordinate axis divides the plane into four quadrants labeled I, II, III and IV .
To construct a polar coordinate system one selects an origin for the polar co-
ordinates and labels it 0. Next construct a half-line similar to the x-axis of the
rectangular coordinates. This half-line is called the polar axis or initial ray and the
origin is called the pole of the polar coordinate system. By placing another line on
top of the polar axis and rotating this line about the pole through a positive angle
, measured in radians, one can create a ray emanating from the origin at an angle
as illustrated in the figure 1-3. In polar coordinates the rays are illustrated emanat-
ing from the origin at equally spaced angular distances around the origin and then
concentric circles are constructed representing constant distances from the origin.
 7
A point in polar coordinates is then denoted by the number pair (r, ) where is
the angle of rotation associated with the ray and r is a distance outward from the
origin along the ray. The polar origin or pole has the coordinates (0, ) for any angle
. All points having the polar coordinates (, 0), with 0, lie on the polar axis.

Here angle rotations are treated the same as

in trigonometry with a counterclockwise ro-
tation being in the positive direction and a
clockwise rotation being in the negative di-
rection. Note that the polar representation
of a point is not unique since the angle can
be increased or decreased by some multiple
Figure 1-3.
of 2 to arrive at the same point. That is,
Construction of polar axes
(r, ) = (r, 2n) where n is an integer.

Also note that a ray at angle can be extended to represent negative distances
along the ray. Points (r, ) can also be represented by the number pair (r, + ).
Alternatively, one can think of a rectangular point (x, y) and the corresponding polar
point (r, ) as being related by the equations

= arctan (y/x), x =r cos

(1.1)
r = x2 + y 2 , y =r sin

An example of a rectangular coordinate system and polar coordinate system are

illustrated in the figure 1-2.

Distance Between Two Points in the Plane

If two points are given in polar coordinates
as (r1 , 1 ) and (r2, 2 ), as illustrated in the fig-
ure 1-4, then one can use the law of cosines
to calculate the distance d between the points
since
Figure 1-4.
d2 = r12 + r22 2r1 r2 cos(1 2 ) (1.2)
Distance between points
in polar coordinates.
8
Alternatively, let (x1 , y1 ) and (x2 , y2 ) denote two points which are plotted on a
cartesian set of axes as illustrated in the figure 1-5. The Greek letter (delta) is
used to denote a change in a quantity. For example, in moving from the point (x1 , y1 )
to the point (x2 , y2 ) the change in x is denoted x = x2 x1 and the change in y is
denoted y = y2 y1 . These changes can be thought of as the legs of a right-triangle
as illustrated in the figure 1-5.

Figure 1-5.
Using a right-triangle to calculate distance between
two points in rectangular coordinates.
The figure 1-5 illustrates that by using the Pythagorean theorem the distance d
between the two points can be determined from the equations


d2 = (x)2 + (y)2 or d= (x2 x1 )2 + (y2 y1 )2 (1.3)

Graphs and Functions

Let X and Y denote sets which contain some subset of the real numbers with
elements x X and y Y . If a rule or relation f is given such that for each x X
there corresponds exactly one real number y Y , then y is said to be a real single-
valued function of x and the relation between y and x is denoted y = f (x) and read
as y is a function of x . If for each x X , there is only one ordered pair (x, y),
then a functional relation from X to Y is said to exist. The function is called single-
valued if no two different ordered pairs (x, y) have the same first element. A way
of representing the set of ordered pairs which define a function is to use one of the
notations
{ (x, y) | y = f (x), x X } or { (x, f (x)) | x X } (1.4)
 9
The set of values x X is called the domain of definition of the function f (x).

The set of values {y | y = f (x) , x X} is called the range of the function or the image
of the set X under the mapping or transformation given by f. The set of ordered
pairs
C = { (x, y) | y = f (x), x X}

is called the graph of the function and represents a curve in the x, y-plane giving
a pictorial representation of the function. If y = f (x) for x X , the number x is
called the independent variable or argument of the function and the image value
y is called the dependent variable of the function. It is to be understood that the
domain of definition of a function contains real values for x for which the relation
f (x) is also real-valued. In many physical problems, the domain of definition X must
be restricted in order that a given physical problem be well defined. For example,

in order that x 1 be real-valued, x must be restricted to be greater than or equal
to 1.
When representing many different functions the symbol f can be replaced by
any of the letters from the alphabet. For example, one might have several different
functions labeled as

y = f (x), y = g(x), y = h(x), . . ., y = y(x), . . ., y = z(x) (1.5)

or one could add subscripts to the letter f to denote a set of n-different functions

F = {f1 (x), f2 (x), . . . , fn (x)} (1.6)

Example 1-3. (Functions)

(a) Functions defined by a formula over a given domain.
Let y = f (x) = x2 + x for x R be a given rule defining a function which can
be represented by a curve in the cartesian x, y-plane. The variable x is a dummy
variable used to define the function rule. Substituting the value 3 in place of x in
the function rule gives y = f (3) = 32 + 3 = 12 which represents the height of the curve
at x = 3. In general, for any given value of x the quantity y = f (x) represents the
height of the curve at the point x. By assigning a collection of ordered values to x
and calculating the corresponding value y = f (x), using the given rule, one collects
a set of (x, y) pairs which can be interpreted as representing a set of points in the
cartesian coordinate system. The set of all points corresponding to a given rule is
10
called the locus of points satisfying the rule. The graph illustrated in the figure 1-6
is a pictorial representation of the given rule.

x y = f (x) = x2 + x
-2.0 2.00
-1.8 1.44
-1.6 0.96
-1.4 0.56
-1.2 0.24
-1.0 0.00
-0.8 -0.16
-0.6 -0.24
-0.4 -0.24
-0.2 -0.16
0.0 0.00
0.2 0.24
0.4 0.56
0.6 0.96
0.8 1.44
1.0 2.00
1.2 2.64
1.4 3.36
1.6 4.16
1.8 5.04
2.0 6.00

Figure 1-6. A graph of the function y = f (x) = x2 + x

Note that substituting x + h in place of x in the function rule gives

f (x + h) = (x + h)2 + (x + h) = x2 + (2h + 1)x + (h2 + h).

If f (x) represents the height of the curve at the point x, then f (x + h) represents the
height of the curve at x + h.
(b) Let r = f () = 1 + for 0 2 be a given rule defining a function which
can be represented by a curve in polar coordinates (r, ). One can select a set
of ordered values for in the interval [0, 2] and calculate the corresponding
values for r = f (). The set of points (r, ) created can then be plotted on polar
graph paper to give a pictorial representation of the function rule. The graph
illustrated in the figure 1-7 is a pictorial representation of the given function
over the given domain.
Note in dealing with polar coordinates a radial distance r and polar angle can
have any of the representations

( (1)n r, + n )
 11
and consequently a functional relation like r = f () can be represented by one of the
alternative equations r = (1)nf ( + n)

r = f () = 1 +
0 1+0
/4 1 + /4
2/4 1 + 2/4
3/4 1 + 3/4
4/4 1 + 4/4
5/4 1 + 5/4
6/4 1 + 6/4
7/4 1 + 7/4
8/4 1 + 8/4
Figure 1-7.
A polar plot of the function r = f () = 1 +

(c) The absolute value function

The absolute value function is defined

x, x0
y = f (x) =| x |=
x, x0

Substituting in a couple of specific values for x one can

form a set of (x, y) number pairs and then sketch a graph
of the function, which represents a pictorial image of the functional relationship
between x and y.
(d) Functions defined in a piecewise fashion.
A function defined by

1 + x, x 1


1 x, 1 x 0
f (x) = xR

x2 , 0x2


2 + x, x>2

is a collection of rules which defines the function in a piecewise fashion. One must
examine values of the input x to determine which portion of the rule is to be used
in evaluating the function. The above example illustrates a function having jump
discontinuities at the points where x = 1 and x = 0.
12
(e) Numerical data.
If one collects numerical data from an experiment such as recording temperature
T at different times t, then one obtains a set of data points called number pairs. If
these number pairs are labeled (ti , Ti), for i = 1, 2, . . ., n, one obtains a table of values
such as

Time t t1 t2 tn
Temperature T T1 T2 Tn

It is then possible to plot a t, T -axes graph associated with these data points by plot-
ting the points and then drawing a smooth curve through the points or by connecting
the points with straight line segments. In doing this, one is assuming that the curve
sketched is a graphical representation of an unknown functional relationship between
the variables.
(e) Other representation of functions
Functions can be represented by different methods such as using equations,
graphs, tables of values, a verbal rule, or by using a machine like a pocket cal-
culator which is programmable to give some output for a given input. Functions can
be continuous or they can have discontinuities. Continuous functions are recognized
by their graphs which are smooth unbroken curves with continuously turning tangent
lines at each point on the curve. Discontinuities usually occur when functional values
or tangent lines are not well defined at a point.

Increasing and Decreasing Functions

One aspect in the study of calculus is to examine how functions change over an
interval. A function is said to be increasing over an interval (a, b) if for every pair
of points (x0 , x1 ) within the interval (a, b), satisfying x0 < x1 , the height of the curve
at x0 is less than the height of the curve at x1 or f (x0 ) < f (x1 ). A function is called
decreasing over an interval (a, b) if for every pair of points (x0 , x1 ) within the interval
(a, b), satisfying x0 < x1 , one finds the height of the curve at x0 is greater than the
height of the curve at x1 or f (x0 ) > f (x1 ).
 13
Linear Dependence and Independence
A linear combination of a set of functions {f1 , f2 , . . . , fn} is formed by taking
arbitrary constants c1 , c2, . . . , cn and forming the sum

y = c1 f1 + c2 f2 + + cn fn (1.7)

One can then say that y is a linear combination of the set of functions {f1 , f2 , . . . , fn}.
If a function f1 (x) is some constant c times another function f2 (x), then one can
write f1 (x) = cf2 (x) and under this condition the function f1 is said to be linearly
dependent upon f2 . If no such constant c exists, then the functions are said to
be linearly independent. Another way of expressing linear dependence and linear
independence applied to functions f1 and f2 is as follows. One can say that, if there
are nonzero constants c1 , c2 such that the linear combination

c1 f1 (x) + c2 f2 (x) = 0 (1.8)

for all values of x, then the set of functions {f1 , f2 } is called a set of dependent
functions. This is due to the fact that if c1
= 0, then one can divided by c1 and
express the equation (1.8) in the form f1 (x) = cc12 f2 (x) = cf2 (x). If the only constants
which make equation (1.8) a true statement are c1 = 0 and c2 = 0, then the set of
functions {f1 , f2 } is called a set of linearly independent functions.
An immediate generalization of the above is the following. If there exists con-
stants c1 , c2 , . . . , cn, not all zero, such that the linear combination

c1 f1 (x) + c2 f2 (x) + cn fn (x) = 0, (1.9)

for all values of x, then the set of functions {f1 , f2 , . . . , fn } is called a linearly dependent
set of functions. If the only constants, for which equation (1.9) is true, are when
c1 = c2 = = cn = 0, then the set of functions {f1 , f2 , . . . , fn } is called a linearly
independent set of functions. Note that if the set of functions are linearly dependent,
then one of the functions can be made to become a linear combination of the other
functions. For example, assume that c1
= 0 in equation (1.9). One can then write
c2 cn
f1 (x) = f2 (x) fn (x)
c1 c1

which shows that f1 is some linear combination of the other functions. That is, f1
is dependent upon the values of the other functions.
14
Single-valued Functions
Consider plane curves represented in rectangular coordinates such as the curves
illustrated in the figures 1-8. These curves can be considered as a set of ordered
pairs (x, y) where the x and y values satisfy some specified condition.

Figure 1-8. Selected curves sketched in rectangular coordinates

In terms of a set representation, these curves can be described using the set
notation
C = { (x, y) | relationship satisfied by x and y with x X }

This represents a collection of points (x, y), where x is restricted to values from some
set X and y is related to x in some fashion. A graph of the function results when the
points of the set are plotted in rectangular coordinates. If for all values x0 a vertical
line x = x0 cuts the graph of the function in only a single point, then the function is
called single-valued. If the vertical line intersects the graph of the function in more
than one point, then the function is called multiple-valued.
Similarly, in polar coordinates, a graph of the function is a curve which can be
represented by a collection of ordered pairs (r, ). For example,

C = { (r, ) | relationship satisfied by r and with }

where is some specified domain of definition of the function. There are available
many plotting programs for the computer which produce a variety of specialized
graphs. Some computer programs produce not only cartesian plots and polar plots,
but also many other specialized graph types needed for various science and engineer-
ing applications. These other graph types give an alternative way of representing
functional relationships between variables.
 15
Example 1-4. Rectangular and Polar Graphs
Plotting the same function in both rectangular coordinates and polar coordinates
gives different shaped curves and so the graphs of these functions have different
properties depending upon the coordinate system used to represent the function.
For example, plot the function y = f (x) = x2 for 2 x 2 in rectangular coordinates
and then plot the function r = g() = 2 for 0 2 in polar coordinates. Show
that one curve is a parabola and the other curve is a spiral.
Solution

x y = f (x) = x2 r = g() = 2
-2.00 4.0000
-1.75 3.0625
0.00 0.0000
-1.50 2.2500
-1.25 1.5625 /4 2 /16
-1.00 1.0000
-0.75 0.5625 /2 2 /4
-0.50 0.2500
-0.25 0.0625 3/4 92 /16
0.00 0.0000
0.25 0.0625 2
0.50 0.2500
0.75 0.5625 5/4 252 /16
1.00 1.0000
1.25 1.5625 3/2 92 /4
1.50 2.2500
1.75 3.0625 7/4 492 /16
2.00 4.0000
2 42

C1 = { (x, y) | y = f (x) = x2 , 2 x 2 } C2 = { (r, ) | r = g() = 2 , 0 2 }

Figure 1-9. Rectangular and polar graphs give different pictures of function.

Select some points x from the domain of the function and calculate the image
points under the mapping y = f (x) = x2 . For example, one can use a spread sheet
and put values of x in one column and the image values y in an adjacent column
to obtain a table of values for representing the function at a discrete set of selected
points.
16
Similarly, select some points from the domain of the function to be plotted in
polar coordinates and calculate the image points under the mapping r = g() = 2 .
Use a spread sheet and put values of in one column and the image values r in an
adjacent column to obtain a table for representing the function as a discrete set of
selected points. Using an x spacing of 0.25 between points for the rectangular graph
and a spacing of /4 for the polar graph, one can verify the table of values and
graphs given in the figure 1-9.

Some well known cartesian curves are illustrated in the following figures.

Figure 1-10. Polynomial curves y = x, y = x2 and y = x3

Figure 1-11. The trigonometric functions y = sin x and y = cos x for x 2 .

 17
Some well known polar curves are illustrated in the following figures.

b= -a b=2a b=3a
Figure 1-12. The limacon curves r = 2a cos + b

r = a sin 3 r = a cos 2 r = a sin 5

Figure 1-13. The rose curves r = a cos n and r = a sin n
If n odd, curve has n-loops and if n is even, curve has 2n loops.

Parametric Representation of Curve

Examine the graph in figure 1-8(b) and observe that it does not represent a
single-valued function y = f (x). Also the circle in figure 1-8(c) does not define a
single valued function. An alternative way of graphing a function is to represent
it in a parametric form.4 In general, a graphical representation of a function or a
4
The parametric form for representing a curve is not unique and the parameter used may or may not have a
physical meaning.
18
section of a function, be it single-valued or multiple-valued, can be defined by a
parametric representation

C = {(x, y) | x = x(t), y = y(t), a t b} (1.10)

where both x(t) and y(t) are single-valued functions of the parameter t. The re-
lationship between x and y is obtained by eliminating the parameter t from the
representation x = x(t) and y = y(t). For example, the parametric representation
x = x(t) = t and y = y(t) = t2 , for t R, is one parametric representation of the
parabola y = x2 .
The Equation of a Circle
A circle of radius and centered at
the point (h, k) is illustrated in the figure
1-14 and is defined as the set of all points
(x, y) whose distance from the point (h, k)
has the constant value of . Using the dis-
tance formula (1.3), with (x1 , y1 ) replaced
by (h, k), the point (x2 , y2 ) replaced by the
variable point (x, y) and replacing d by ,
one can show the equation of the circle is
given by one of the formulas

(x h)2 + (y k)2 = 2

(1.11) Figure 1-14.
or (x h)2 + (y k)2 =
Circle centered at (h, k).

Equations of the form

x2 + y 2 + x + y + = 0, , , constants (1.12)

can be converted to the form of equation (1.11) by completing the square on the
x and y terms. This is accomplished by taking 1/2 of the x-coefficient, squaring
and adding the result to both sides of equation (1.12) and then taking 1/2 of the
y -coefficient, squaring and adding the result to both sides of equation (1.12). One
then obtains
 19
2 2 2 2
   
x2 + x + + y 2 + y + = +
4 4 4 4
 2
 2
which simplifies to x+ + y+ = r2
2 2
2 2
where r2 = 4 + 4 . Completing the square is a valid conversion whenever the
2 2
right-hand side 4 + 4 0.
An alternative method of representing the equation of the circle is to introduce
a parameter such as the angle illustrated in the figure 1-14 and observe that by
trigonometry
yk xh
sin = and cos =

These equations are used to represent the circle in the alternative form

C = { (x, y) | x = h + cos , y = k + sin , 0 2 } (1.13)

This is called a parametric representation of the circle in terms of a parameter .

The equation of a circle in polar form can
be constructed as follows. Let (r, ) denote
a variable point which moves along the cir-
cumference of a circle of radius which is
centered at the point (r1 , 1 ) as illustrated
in the figure 1-15. Using the distances r, r1
and , one can employ the law of cosines
to express the polar form of the equation
of a circle as

Figure 1-15. r 2 + r12 2rr1 cos( 1 ) = 2 (1.14)

Circle centered at (r1 , 1 ).

Functions can be represented in a variety of ways. Sometimes functions are

represented in the implicit form G(x, y) = 0, because it is not always possible to
solve for one variable explicitly in terms of another. In those cases where it is
possible to solve for one variable in terms of another to obtain y = f (x) or x = g(y),
the function is said to be represented in an explicit form.
20
For example, the circle of radius can be represented by any of the relations
G(x, y) = x2 + y 2 2 = 0,



+ 2 x2 , x + 2 y 2 , y (1.15)
y = f (x) =  , x = g(y) = 
2 x2 , x 2 y 2 , y

Note that the circle in figure 1-8(c) does not define a single valued function. The

circle can be thought of as a graph of two single-valued functions y = + 2 x2

and y = 2 x2 for x if one treats y as a function of x. The other
representation in equation (1.15) results if one treats x as a function of y.

Types of functions
One can define a functional relationships between the two variables x and y in
different ways.
A polynomial function in the variable x has the form

y = pn (x) = a0 xn + a1 xn1 + a2 xn2 + + an1 x + an (1.16)

where a0 , a1 , . . . , an represent constants with a0
= 0 and n is a positive integer. The

integer n is called the degree of the polynomial function. The fundamental theorem
of algebra states that a polynomial of degree n has n-roots. That is, the polynomial
equation pn (x) = 0 has n-solutions called the roots of the polynomial equation. If
these roots are denoted by x1 , x2 , . . . , xn , then the polynomial can also be represented
in the form
pn (x) = a0 (x x1 )(x x2 ) (x xn )

If xi is a number, real or complex, which satisfies pn(xi ) = 0, then (x xi ) is a factor of

the polynomial function pn (x). Complex roots of a polynomial function must always
occur in conjugate pairs and one can say that if + i is a root of pn (x) = 0, then
i is also a root. Real roots xi of a polynomial function give rise to linear factors
(x xi ), while complex roots of a polynomial function give rise to quadratic factors
 
of the form (x )2 + 2 , , constant terms.
A rational function is any function of the form
P (x)
y = f (x) = (1.17)
Q(x)
where both P (x) and Q(x) are polynomial functions in x and Q(x)
= 0. If y = f (x) is
a root of an equation of the form

b0 (x)y n + b1 (x)y n1 + b2 (x)y n2 + + bn1(x)y + bn (x) = 0 (1.18)

 21
where b0 (x), b1(x), . . . , bn(x) are polynomial functions of x and n is a positive integer,
then y = f (x) is called an algebraic function. Note that polynomial functions and
rational functions are special types of algebraic functions. Functions which are built
up from a finite number of operations of the five basic operations of addition, sub-
traction, multiplication, division and extraction of integer roots, usually represent
algebraic functions. Some examples of algebraic functions are
1. Any polynomial function.

2. f1 (x) = (x3 + 1) x + 4

x2 + 3 6 + x2
3. f2 (x) =
(x 3)4/3

The function f (x) = x2 is an example of a function which is not an algebraic func-
tion. This is because the square root of x2 is the absolute value of x and represented

2
x, if x 0
f (x) = x = |x| =
x, if x < 0
and the absolute value operation is not one of the five basic operations mentioned
above.
A transcendental function is any function which is not an algebraic function.
The exponential functions, logarithmic functions, trigonometric functions, inverse
trigonometric functions, hyperbolic functions and inverse hyperbolic functions are
examples of transcendental functions considered in this calculus text.
The Exponential and Logarithmic Functions
The exponential functions have the form y = bx , where b > 0 is a positive constant
and the variable x is an exponent. If x = n is a positive integer, one defines
1
bn = b
b b and bn = (1.19)
bn
n factors
By definition, if x = 0, then b0 = 1. Note that if y = bx , then y > 0 for all real values
of x.
Logarithmic functions and exponential functions are related. By definition,

if y = bx then x = logb y (1.20)

and x, the exponent, is called the logarithm of y to the base b. Consequently, one
can write

logb (bx ) = x for every x R and blogb x = x for every x > 0 (1.21)
22
Recall that logarithms satisfy the following properties
logb (xy) = logb x + logb y, x > 0 and y > 0


x
logb = logb x logb y, x > 0 and y > 0 (1.22)
y
logb (y x ) = x logb y, x can be any real number

Of all the numbers b > 0 available for use as a base for the logarithm function
the base b = 10 and base b = e = 2.71818 are the most often seen in engineering
and scientific research. The number e is a physical constant5 like . It can not be
represented as the ratio of two integer so is an irrational number. It can be defined
as the limiting sum of the infinite series
1 1 1 1 1 1 1
e= + + + + + ++ +
0! 1! 2! 3! 4! 5! n!
Using a computer6 one can verify that the numerical value of e to 50 decimal places
is given by

e = 2.7182818284590452353602874713526624977572470936999 . . .

The irrational number e can also be determined from the limit e = lim (1 + h)1/h .
h0
In the early part of the seventeenth century many mathematicians dealt with
and calculated the number e, but it was Leibnitz7 in 1690 who first gave it a name
and notation. His notation for the representation of e didnt catch on. The value of
the number represented by the limit lim (1 + h)1/h is used so much in mathematics
h0
it was represented using the symbol e by Leonhard Euler8 sometime around 1731
and his notation for representing this number has been used ever since. The number
e is sometimes referred to as Eulers number, the base of the natural logarithms.
The number e and the exponential function ex will occur frequently in our study of
calculus.

5
There are many physical constants in mathematics. Some examples are e, , i,(imaginary component),
(Euler-Mascheroni constant). For a listing of additional mathematical constants go to the web site
http : //en.wikipedia.org/wiki/M athematical constant.
6
One can go to the web site
http : //www.numberworld.org/misc runs/e 500b.html to see that over 500 billion digits of this number have
been calculated.
7
Gottfried Wilhelm Leibnitz (1646-1716) a German physicist, mathematician.
8
Leonhard Euler (1707-1783) a famous Swiss mathematician.
 23
The logarithm to the base e, is called the natural logarithm and its properties
are developed in a later chapter. The natural logarithm is given a special notation.
Whenever the base b = e one can write either

y = loge x = ln x or y = log x (1.23)

That is, if the notation ln is used or whenever the base is not specified in using
logarithms, it is to be understood that the base b = e is being employed. In this
special case one can show

y = ex = exp(x) x = ln y (1.24)

which gives the identities

ln(ex ) = x, xR and eln x = x, x>0 (1.25)

In our study of calculus it will be demonstrated that the natural logarithm has the
special value ln(e) = 1.

Figure 1-16. The exponential function y = ex and logarithmic function y = ln x

Note that if y = logb x, then one can write the equivalent statement by = x since a
logarithm is an exponent. Taking the natural logarithm of both sides of this last
equation gives
ln(by ) = ln x or y ln b = ln x (1.26)
24
Consequently, for any positive number b different from one
ln x
y = logb x = , b
= 1 (1.27)
ln b
The exponential function y = ex , together with the natural logarithm function can
then be used to define all exponential functions by employing the identity
y = bx = (eln b )x = ex ln b (1.28)

Graphs of the exponential function y = ex = exp(x) and the natural logarithmic

function y = ln(x) = log(x) are illustrated in the figure 1-16.
The Trigonometric Functions
The ratio of sides of a right triangle are used to define the six trigonometric
functions associated with one of the acute angles of a right triangle. These definitions
can then be extended to apply to positive and negative angles associated with a point
moving on a unit circle.
The six trigonometric functions asso-
ciated with a right triangle are
sine tangent secant
cosine cotangent cosecant

which are abbreviated respectively as

sin, tan, sec, cos, cot, and csc .
Let and denote complementary angles in a right triangle as illustrated above.
The six trigonometric functions associated with the angle are

y opposite side y opposite side r hypotenuse

sin = = , tan = = , sec = =
r hypotenuse x adjacent side x adjacent side
x adjacent side x adjacent side r hypotenuse
cos = = , cot = = , csc = =
r hypotenuse y opposite side y opposite side

Graphs of the Trigonometric Functions

Graphs of the trigonometric functions sin , cos and tan , for varying over the
domain 0 4 , can be represented in rectangular coordinates by the point sets
S = { (, y) | y = sin , 0 4 },

C = { (, x) | x = cos , 0 4 }
T = { (, y) | y = tan , 0 4 }
and are illustrated in the figure 1-17.
 25
Using the periodic properties

sin( + 2) = sin , cos( + 2) = cos and tan( + ) = tan

these graphs can be extend and plotted over other domains.

Figure 1-17. Graphs of the trigonometric functions sin , cos and tan

The function y = sin can also be interpreted as representing the motion of a point P
moving on the circumference of a unit circle. The point P starts at the point (1, 0),
where the angle is zero and then moves in a counterclockwise direction about the
circle. As the point P moves around the circle its ordinate value is plotted against
the angle . The situation is as illustrated in the figure 1-17(a). The function x = cos
can be interpreted in the same way with the point P moving on a circle but starting
at a point which is shifted /2 radians clockwise. This is the equivalent to rotating
the x, yaxes for the circle by /2 radians and starting the point P at the coordinate
(1, 0) as illustrated in the figure 1-17(b).
The Hyperbolic Functions
Related to the exponential functions ex and ex are the hyperbolic functions
hyperbolic sine written sinh , hyperbolic cotangent written coth
hyperbolic cosine written cosh , hyperbolic secant written sech
hyperbolic tangent written tanh , hyperbolic cosecant written csch
26
These functions are defined
ex ex 1
sinh x = , csch x =
2 sinh x
ex + ex 1
cosh x = , sech x = (1.29)
2 cosh x
sinh x 1
tanh x = , coth x =
cosh x tanh x

As the trigonometric functions are related to the circle and are sometimes referred
to as circular functions, it has been found that the hyperbolic functions are related
to equilateral hyperbola and hence the name hyperbolic functions. This will be
explained in more detail in the next chapter.
Symmetry of Functions
The expression y = f (x) is the representation of a function in an explicit form
where one variable is expressed in terms of a second variable. The set of values given
by
S = {(x, y) | y = f (x), x X}

where X is the domain of the function, represents a graph of the function. The
notation f (x), read f of x , has the physical interpretation of representing y which
is the height of the curve at the point x. Given a function y = f (x), one can replace
x by any other argument. For example, if f (x) is a periodic function with least
period T , one can write f (x) = f (x + T ) for all values of x. One can interpret the
equation f (x) = f (x + T ) for all values of x as stating that the height of the curve at
any point x is the same as the height of the curve at the point x + T . As another
example, if the notation y = f (x) represents the height of the curve at the point x,
then y + y = f (x + x) would represent the height of the given curve at the point
x + x and y = f (x + x) f (x) would represent the change in the height of the curve
y in moving from the point x to the point x + x. If the argument x of the function
is replaced by x, then one can compare the height of the curve at the points x and
x. If f (x) = f (x) for all values of x, then the height of the curve at x equals the
height of the curve at x and when this happens the function f (x) is called an even
function of x and one can state that f (x) is a symmetric function about the y-axis.
If f (x) = f (x) for all values of x, then the height of the curve at x equals the
negative of the height of the curve at x and in this case the function f (x) is called
an odd function of x and one can state that the function f (x) is symmetric about the
origin. By interchanging the roles of x and y and shifting or rotation of axes, other
 27
symmetries can be discovered. The figure 1-18 and 1-19 illustrates some examples
of symmetric functions.

Figure 1-18. Examples of symmetric functions

Symmetry about a line. Symmetry about a point.

Figure 1-19. Examples of symmetric functions

In general, two points P1 and P2 are said to be symmetric to a line if the line is
the perpendicular bisector of the line segment joining the two points. In a similar
fashion a graph is said to symmetric to a line if all points of the graph can be
grouped into pairs which are symmetric to the line and then the line is called the
axis of symmetry of the graph. A point of symmetry occurs if all points on the graph
can be grouped into pairs so that all the line segments joining the pairs are then
bisected by the same point. See for example the figure 1-19. For example, one can
say that a curve is symmetric with respect to the x-axis if for each point (x, y) on the
curve, the point (x, y) is also on the curve. A curve is symmetric with respect to
the y-axis if for each point (x, y) on the curve, the point (x, y) is also on the curve.
A curve is said to be symmetric about the origin if for each point (x, y) on the curve,
then the point (x, y) is also on the curve.
28
Example 1-5. A polynomial function pn (x) = a0 xn + a1 xn1 + + an1x + an of
degree n has the following properties.
(i) If only even powers of x occur in pn (x), then the polynomial curve is symmetric
about the y-axis, because in this case pn (x) = pn (x).
(ii) If only odd powers of x occur in pn (x), then the polynomial curve is symmetric
about the origin, because in this case pn (x) = pn (x).
(iii) If there are points x = a and x = c such that pn (a) and pn (c) have opposite signs,
then there exists at least one point x = b satisfying a < b < c, such that pn (b) = 0.
This is because polynomial functions are continuous functions and they must
change continuously from the value pn (a) to the value pn (c) and so must pass, at
least once, through the value zero.

Translation and Scaling of Axes

Consider two sets of axes labeled (x, y) and (x, y) as illustrated in the figure 1-
20(a) and (b). Pick up the (x, y) axes and keep the axes parallel to each other and
place the barred axes at some point P having the coordinates (h, k) on the (x, y) axes
as illustrated in the figure 1-20(c). One can now think of the barred axes as being
a translated set of axes where the new origin has been translated to the point (h, k)
of the old set of unbarred axes. How is an arbitrary point (x, y) represented in terms
of the new barred axes coordinates? An examination of the figure 1-20 shows that
a general point (x, y) can be represented as

x = x + h and y = y + k or x = x h and y = y k (1.30)

Consider a curve y = f (x) sketched on the (x, y) axes of figure 1-20(a). Change
the symbols x and y to x and y and sketch the curve y = f (x) on the (x, y) axes of
figure 1-20(b). The two curves should look exactly the same, the only difference
being how the curves are labeled. Now move the (x, y) axes to a point (h, k) of the
(x, y) coordinate system to produce a situation where the curve y = f (x) is now to be
represented with respect to the (x, y) coordinate system.
The new representation can be determined by using the transformation equations
(1.30). That is, the new representation of the curve is obtained by replacing y by
y k and replacing x by x h to obtain

y k = f (x h) (1.31)
 29

Figure 1-20. Shifting of axes.

In the special case k = 0, the curve y = f (x h) represents a shifting of the curve

y = f (x) a distance of h units to the right. Replacing h by h and letting k = 0 one
finds the curve y = f (x + h) is a shifting of the curve y = f (x) a distance of h units
to the left. In the special case h = 0 and k
= 0, the curve y = f (x) + k represents a
shifting of the graph y = f (x) a distance of k units upwards. In a similar fashion,
the curve y = f (x) k represents a shifting of the graph of y = f (x) a distance of k
units downward. The figure 1-21 illustrates the shifting and translation of axes for
the function y = f (x) = x2 .

Figure 1-21. Translation and shifting of axes.

Introducing a constant scaling factor s > 0, by replacing y by y/s one can create
the scaled function y = sf (x). Alternatively one can replace x by sx and obtain the
scaled function y = f (sx). These functions are interpreted as follows.
(1) Plotting the function y = sf (x) has the effect of expanding the graph of y = f (x)
in the vertical direction if s > 1 and compresses the graph if s < 1. This is
30
equivalent to changing the scaling of the units on the y-axis in plotting a graph.
As an exercise plot graphs of y = sin x, y = 5 sin x and y = 51 sin x.
(2) Plotting the function y = f (sx) has the effect of expanding the graph of y = f (x)
in the horizontal direction if s < 1 and compresses the graph in the x-direction
if s > 1. This is equivalent to changing the scaling of the units on the x-axis in
plotting a graph. As an exercise plot graphs of y = sin x, y = sin( 13 x) and y = sin(3x).
(3) A plot of the graph (x, f (x)) gives a reflection of the graph y = f (x) with respect
to the x-axis.

(4) A plot of the graph (x, f (x)) gives a reflection of the graph y = f (x) with respect
to the y-axis.

Rotation of Axes
Place the (x, y) axes from figure 1-20(b) on top of the (x, y) axes of figure 1-20(a)
and then rotate the (x, y) axes through an angle to obtain the figure 1-22. An
arbitrary point (x, y), a distance r from the origin, has the coordinates (x, y) when
referenced to the (x, y) coordinate system. Using basic trigonometry one can find
the relationship between the rotated and unrotated axes. Examine the figure 1-22
and verify the following trigonometric relationships.
The projection of r onto the x axis produces x = r cos and the projection of r onto
the y axis produces y = r sin . In a similar fashion consider the projection of r onto
the y-axis to show y = r sin( + ) and the projection of r onto the x-axis produces
x = r cos( + ).
 31
Expressing these projections in the form
x
cos( + ) =
r (1.32)
y
sin( + ) =
r
x
cos =
r (1.33)
y
sin =
r
one can expand the equations (1.32) to obtain
Figure 1-22.
x = r cos( + ) = r(cos cos sin sin )
Rotation of axes. (1.34)
y = r sin( + ) = r(sin cos + cos sin )

Substitute the results from the equations (1.33) into the equations (1.34) to ob-
tain the transformation equations from the rotated coordinates to the unrotated
coordinates. One finds these transformation equations can be expressed
x =x cos y sin
(1.35)
y =x sin + y cos

Solving the equations (1.35) for x and y produces the inverse transformation

x =x cos + y sin
(1.36)
y = x sin + y cos

Inverse Functions
If a function y = f (x) is such that it never takes on the same value twice, then
it is called a one-to-one function. One-to-one functions are such that if x1
= x2 , then
f (x1 )
= f (x2 ). One can test to determine if a function is a one-to-one function by
using the horizontal line test which is as follows. If there exists a horizontal line
y = a constant, which intersects the graph of y = f (x) in more than one point, then
there will exist numbers x1 and x2 such that the height of the curve at x1 is the
same as the height of the curve at x2 or f (x1 ) = f (x2 ). This shows that the function
y = f (x) is not a one-to-one function.
Let y = f (x) be a single-valued function of x which is a one-to-one function such as
the function sketched in the figure 1-23(a). On this graph interchange the values of
x and y everywhere to obtain the graph in figure 1-23(b). To represent the function
32
x = f (y) with y in terms of x define the inverse operator f 1 with the property that
f 1 f (x) = x. Now apply this operator to both sides of the equation x = f (y) to obtain
f 1 (x) = f 1 f (y) = y or y = f 1 (x). The function y = f 1 (x) is called the inverse
function associated with the function y = f (x). Rearrange the axes in figure 1-23(b)
so the x-axis is to the right and the y-axis is vertical so that the axes agree with
the axes representation in figure 1-23(a). This produces the figure 1-23(c). Now
place the figure 1-23(c) on top of the original graph of figure 1-23(a) to obtain the
figure 1-23(d), which represents a comparison of the original function and its inverse
function. This figure illustrates the function f (x) and its inverse function f 1 (x) are
one-to-one functions which are symmetric about the line y = x.

Figure 1-23. Sketch of a function and its inverse function.

Note that some functions do not have an inverse function. This is usually the
result of the original function not being a one-to-one function. By selecting a domain
of the function where it is a one-to-one function one can define a branch of the function
which has an inverse function associated with it.
Lets examine what has just been done in a slightly different way. If y = f (x) is a
one-to-one function, then the graph of the function y = f (x) is a set of ordered pairs
(x, y), where x X , the domain of the function and y Y the range of the function.
 33
Now if the function f (x) is such that no two ordered pairs have the same second
element, then the function obtained from the set

S = { (x, y) | y = f (x), x X}

by interchanging the values of x and y is called the inverse function of f and it is

denoted by f 1 . Observe that the inverse function has the domain of definition Y
and its range is X and one can write

y = f (x) f 1 (y) = x (1.32)

Still another way to approach the problem is as follows. Two functions f (x)
and g(x) are said to be inverse functions of one another if f (x) and g(x) have the
properties that
g(f (x)) = x and f (g(x)) = x (1.33)

If g(x) is an inverse function of f (x), the notation f 1 , (read f-inverse), is used to

denote the function g . That is, an inverse function of f (x) is denoted f 1 (x) and has
the properties
f 1 (f (x)) = x and f (f 1 (x)) = x (1.34)

Given a function y = f (x), then by interchanging the symbols x and y there results
x = f (y). This is an equation which defines the inverse function. If the equation
x = f (y) can be solved for y in terms of x, to obtain a single valued function, then
this function is called the inverse function of f (x). One then obtains the equivalent
statements
x = f (y) y = f 1 (x) (1.35)

The process of interchanging x and y in the representation y = f (x) to obtain

x = f (y) implies that geometrically the graphs of f and f 1 are mirror images of
each other about the line y = x. In order that the inverse function be single valued
and one-to-one, it is necessary that there are no horizontal lines, y = constant, which
intersect the graph y = f (x) more than once. Observe that one way to find the inverse
function is the following.
34
(1.) Write y = f (x) and then interchange x and y to obtain x = f (y)
(2.) Solve x = f (y) for the variable y to obtain y = f 1 (x)
(3.) Note the inverse function f 1 (x) sometimes turns out to be a multiple-valued
function. Whenever this happens it is customary to break the function up into
a collection of single-valued functions, called branches, and then one of these
branches is selected to be called the principal branch of the function. That is,
if multiple-valued functions are involved, then select a branch of the function
which is single-valued such that the range of y = f (x) is the domain of f 1 (x).
An example of a function and its inverse is given in the figure 1-22.

Figure 1-22. An example of a function and its inverse.

Example 1-6. (Inverse Trigonometric Functions)

The inverse trigonometric functions are defined in the table 1-1. The inverse
trigonometric functions can be graphed by interchanging the axes on the graphs
of the trigonometric functions as illustrated in the figures 1-23 and 1-24. Observe
that these inverse functions are multi-valued functions and consequently one must
define an interval associated with each inverse function such that the inverse function
becomes a single-valued function. This is called selecting a branch of the function
such that it is single-valued.
 35

Table 1-1
Inverse Trigonometric Functions
Alternate Interval for
Function notation Definition single-valuedness
1 1
arcsinx sin x sin x=y if and only if x = sin y 2 y
2
arccos x cos1 x cos1 x = y if and only if x = cos y 0y
arctanx tan1 x tan1 x = y if and only if x = tan y 2 < y <
2
arccot x cot1 x cot1 x = y if and only if x = cot y 0<y<

arcsec x sec1 x sec1 x = y if and only if x = sec y 0 y , y
= 2
arccsc x csc1 x csc1 x = y if and only if x = csc y 2 y 2 , y
= 0

Figure 1-23. The inverse trigonometric functions sin1 x, cos1 x and tan1 x.

There are many different intervals over which each inverse trigonometric function
can be made into a single-valued function. These different intervals are referred to
as branches of the inverse trigonometric functions. Whenever a particular branch is
required for certain problems, then by agreement these branches are called principal
branches and are always used in doing calculations. The following table gives one
way of defining principal value branches for the inverse trigonometric functions.
These branches are highlighted in the figures 1-23 and 1-24.
36
Principal Values for Regions Indicated
x<0 x0
2 sin1 x < 0 0 sin1 x
2

2
cos1 x 0 cos1 x 2
2 tan1 x < 0 0 tan1 x <
2

2
< cot1 x < 0 < cot1 x 2
1
2 sec x 0 sec1 x < 2
1
2 csc x < 0 0 < csc1 x 2

Figure 1-24. The inverse trigonometric functions cot1 x, sec1 x and csc1 x.

Equations of lines
Given two points (x1 , y1 ) and (x2 , y2 ) one can plot these points on a rectangular
coordinate system and then draw a straight line
through the two points as illus-
trated in the sketch given in the figure 1-25. By definition the slope of the line is
defined as the tangent of the angle which is formed where the line
intersects the
x-axis.
 37
Move from point (x1 , y1 ) to point (x2 , y2 ) along the line and let y denote a change
in y and let x denote a change in x, then the slope of the line, call it m, is calculated
y2 y1 change in y y
slope of line = m = tan = = = (1.36)
x2 x1 change in x x
If (x, y) is used to denote a variable point which moves along the line, then one can
make use of similar triangles and write either of the statements
y y1 y y2
m= or m= (1.37)
x x1 x x2

The first equation representing the change in y over a change in x relative to the
first point and the second equation representing a change in y over a change in x
relative to the second point on the line.
This gives the two-point formulas for representing a line
y y1 y2 y1 y y2 y2 y1
= =m or = =m (1.38)
x x1 x2 x1 x x2 x2 x1

Figure 1-25. The line y y1 = m(x x1 )

Once the slope m = change in y = tan of the line is known, one can represent
change in x
the line using either of the point-slope forms

y y1 = m(x x1 ) or y y2 = m(x x2 ) (1.39)

Note that lines parallel to the x-axis have zero slope and are represented by
equations of the form y = y0 = a constant. For lines which are perpendicular to the
x-axis or parallel to the y -axis, the slope is not defined. This is because the slope
38
tends toward a + infinite slope or - infinite slope depending upon how the angle of
intersection approaches /2. Lines of this type are represented by an equation
having the form x = x0 = a constant. The figure 1-26 illustrates the general shape
of a straight line which has a positive, zero and negative slope.

positive slope zero slope negative slope slope

Figure 1-26. The slope of a line.

The general equation of a line is given by

Ax + By + C = 0, where A, B, C are constants. (1.40)

The slope-intercept form for the equation of a line is given by

y = mx + b (1.41)

where m is the slope and b is the yintercept. Note that when x = 0, then the point
(0, b) is where the line intersects the y axis. If the line intersects the y-axis at the
point (0, b) and intersects the x-axis at the point (a, 0), then the intercept form for
the equation of a straight line is given by
x y
+ = 1, a
= 0, b
= 0 (1.42)
a b

One form1 for the parametric equation of a straight line is given by the set of points

= { (x, y) | x = t and y = mt + b, < t < }

Another parametric form such as

= { (x, y) | x = sin t and y = m sin t + b, < t < }

gives only a segment of the total line.

1
The parametric representation of a line or curve is not unique and depends upon the representation.
 39
The polar form for the equation of a straight line can be obtained from the
intercept form for a line, given by equation (1.42), by using the transformation
equations (1.1) previously considered. For example, if the intercept form of the line
x y

is + = 1, then the transformation equations x = r cos and y = r sin change this
a b
equation to the form
r cos r sin
+ = 1 (1.43)
a b
Let d denote the perpendicular distance
from the line
to the origin of the rectan-
gular x, y coordinate system as illustrated
in the figure 1-27. This perpendicular line
makes an angle such that
d d
cos = and cos( ) = sin =
a 2 b
1 1
Solve for and and substitute the results
Figure 1-27. a b
into the equation (1.43) to show
Determining polar form for line.
r r
cos cos + sin sin = 1
d d
Use trigonometry to simplify the above equation and show the polar form for the
equation of the line is
r cos( ) = d (1.44)

Here (r, ) is a general point in polar coordinates which moves along the line
de-
scribed by the polar equation (1.44) and is the angle that the x-axis makes with
the line which passes through the origin and is perpendicular to the line
.

Perpendicular Lines
Consider a line
2 which is perpendicular to a given line

1 as illustrated in the figure 1-28. The slope of the line

1 is given by m1 = tan 1 and the slope of the line
2 is
given by m2 = tan 2 where 1 and 2 are the positive
angles made when the lines
1 and
2 intersect the x-axis. Figure 1-28.
Perpendicular lines.
40
Two lines are said to intersect orthogonally when they intersect to form two right
angles. Note that 2 is an exterior angle to a right triangle ABC and so one can
write 2 = 1 + /2. If the two lines are perpendicular, then the product of the slopes
m1 and m2 must satisfy


sin 1 sin(1 + /2) sin 1 cos 1
m1 m2 = tan 1 tan 2 = = = 1 (1.45)
cos 1 cos(1 + /2) cos 1 ( sin 1 )

which shows that if the two lines are perpendicular, then the product of their slopes
must equal -1, or alternatively, one slope must be the negative reciprocal of the other
slope. This relation breaks down if one of the lines is parallel to the x-axis, because
then a zero slope occurs.
In general, if line
1 with slope m1 = tan 1 inter-
sects line
2 with slope m2 = tan 2 and denotes the
angle of intersection as measured from line
1 to
2 ,
then = 2 1 and
tan 2 tan 1
tan = tan(2 1) =
1 + tan 1 tan 2
m2 m1
tan =
1 + m1 m2

Note that as m2 approaches 1/m1 , then the angle approaches /2.

Limits
The notation x is used to denote the value x increasing without bound
in the positive direction and the notation x is used to denote x increasing
without bound in the negative direction. The limits x
lim f (x) and lim f (x), if they
x
exist, are used to denote the values of a function f (x) as the variable x is allowed
to increase without bound in the positive and negative directions. For example one
can write
1 1
lim (2 + ) = 2 and lim (2 + 2 ) = 2
x x x x
The notation x x0 denotes the variable x approaching the finite value x0 , but
it never gets there. If x and x0 denote real numbers, then x can approach x0 from
any direction and get as close to x0 as you desire, but it cannot equal the value x0 .
For  > 0 a small real number, sketch on a piece of paper the difference between the
x-values defined by the sets

N = {x | | x x0 | < } and N0 = {x | 0 < |x x0 | < } (1.46)

 41
Observe that the set N contains all the points between x0  and x0 +  together with
the point x0 N , while the set N0 is the same as N but has the point x0 excluded so
one can write x0 / N0 . The set N is called a neighborhood of the point x0 while the set
N0 is called a deleted neighborhood of the point x0 . The notation x x0 emphasizes
the requirement that x approach the value x0 , but x is restricted to taking on the
values in the set N0 . The situation is illustrated in the figures 1-29(a) and 1-29(b).
The notation x x+ 0 is used to denote x approaching x0 from the right-hand side of

x0 , where x is restricted such that x > x0 . The notation x x 0 is used to denote x

approaching x0 from the left-hand side of x0 , where x is restricted such that x < x0 .
In general, the notation x x0 means x can approach x0 from any direction, but x
can never equal x0 .

Figure 1-29. Left and right-hand approaches of x to the value x0 .

Infinitesimals
You cannot compare two quantities which are completely different in every way.
To measure related quantities you must have
(i) A basic unit of measurement applicable to the quantity being measured.
(ii) A number representing the ratio of the measured quantity to the basic unit.

The concept of largeness or smallness of a quantity is relative to the basic unit

selected for use in the measurement. For example, if a quantity Q is divided up
into some fractional part f , then f Q is smaller than Q because the ratio f Q/Q = f
is small. For f < 1, quantities like f Q, f 2 Q, f 3 Q, . . . are called small quantities of
the first, second, third orders of smallness, since each quantity is a small fraction
f of the previous quantity. If the fraction f is allowed to approach zero, then the
quantities f Q, f 2 Q, f 3 Q, . . . are very, very small and are called infinitesimals of the
first, second, third, . . ., orders. Thus, if x is a small change in x, then (x)2 would
be an infinitesimal of the second order, (x)3 would be an infinitesimal of the third
order, etc.
42

In terms of limits, if and are infinitesimals and lim is some constant
0
different from zero, then and are called infinitesimals of the same order. However,

if lim = 0, then is called an infinitesimal of higher order than .
0
If you are dealing with an equation involving infinitesimals of different
orders, you only need to retain those infinitesimals of lowest order, since the higher
order infinitesimals are significantly smaller and will not affect the results
when these infinitesimals approach zero .
This concept is often used in comparing the ratio of two small quantities which
approach zero. Consider the problem of finding the volume of a hollow cylinder, as
illustrated in the figure 1-30, as the thickness of the cylinder sides approaches zero.

Figure 1-30. Volume of hollow cylinder.

Let V denote the volume of the hollow cylinder with r the inner radius of the
hollow cylinder and r + r the outer radius. One can write
V =Volume of outer cylinder Volume of inner cylinder
V =(r + r)2 h r 2 h = [r 2 + 2rr + (r)2 ]h r 2 h

V =2rhr + h(r)2

This relation gives the exact volume of the hollow cylinder. If one takes the limit
as r tends toward zero, then the r and (r)2 terms become infinitesimals and
the infinitesimal of the second order can be neglected since one is only interested in
comparison of ratios when dealing with small quantities. For example
V
lim = lim (2rh + hr) = 2rh
r0 r r0
 43
Limiting Value of a Function
The notation lim f (x) =
is used to denote the limiting value of a function f (x)
xx0
as x approaches the value x0 , but x
= x0 . Note that the limit statement lim f (x) is
xx0
dependent upon values of f (x) for x near x0 , but not for x = x0 . One must examine
the values of f (x) both for x+0 values (values of x slightly greater than x0 ) and for

x0 values (values of x slightly less than x0 ). These type of limiting statements are
written
lim f (x)
+
and lim f (x)

xx0 xx0

and are called right-hand and left-hand limits respectively. There may be situations
where (a) f (x0 ) is not defined (b) f (x0 ) is defined but does not equal the limiting
value
(c) the limit xx
lim f (x) might become unbounded, in which case one can write
0

a statement stating that no limit exists as x x0 .

Some limits are easy to calculate, for example lim (3x + 1) = 7, is a limit of the
x2
form xx
lim f (x) = f (x0 ), where f (x0 ) is the value of f (x) at x = x0 , if the value f (x0 )
0

exists. This method is fine if the graph of the function f (x) is a smooth unbroken
curve in the neighborhood of the point x0 .

The limiting value xx

lim f (x) cannot be calculated by eval-
0
uating the function f (x) at the point x0 if the function is
not defined at the point x0 . A function f (x) is called a
sectionally continuous function if its graph can be repre-
sented by sections of unbroken curves. The function f (x)
defined
Figure 1-31. 
2 (x x0 ), x < x0
f (x) =
Sectionally-continuous 5 + 2(x x0 ), x > x0
function
is an example of a sectionally continuous function. Note this function is not defined
at the point x0 and the left-hand limit lim f (x) = 2 and the right-hand limit given
xx0
by lim+ f (x) = 5 are not equal and f (x0 ) is not defined. The graph of f (x) is sketched
xx0
in the figure 1-31. The function f (x) is said to have a jump discontinuity at the point
x = x0 and one would write lim f (x) does not exit.
xx0
Some limiting values produce the indeterminate forms 00 or
and these resulting
forms must be analyzed further to determine their limiting value. For example, the
44
limiting value 00 may reduce to a zero value, a finite quantity or it may become an
infinite quantity as the following examples illustrate.
3x 3
3x2 3 lim 2
= lim = , an infinite limit
lim = lim x = 0, a zero limit x0 2x x0 2x
x0 2x x0 2 x2 x20 (x x0 )(x + x0 )
3x 3 3 lim = lim = 2x0 ,
lim = lim = , a finite limit xx0 x x0 xx0 (x x0 )
x0 2x x0 2 2
a finite limit

Example 1-7. (Geometry used to determine limiting value)

sin x
Consider the function f (x) = and observe that this function is not defined
x
sin x 0
at the value x = 0, because f (0) = = , an indeterminate form. Let us
x x=0 0
sin x
investigate the limit lim using the geometry of the figure 1-32 as the angle2 x
x0 x
gets very small, but with x
= 0. The figure 1-32 illustrates part of a circle of radius
r sketched in the first quadrant along with a ray from the origin constructed at the
angle x. The lines AD and BC perpendicular to the polar axis are constructed along
with the line BD representing a chord. These constructions are illustrated in the
figure 1-32. From the geometry of figure 1-32 verify the following values.
1 1
Area  0BD = 0B AD = r2 sin x
2 2
AD = r sin x 1
Area sector 0BD = r2 x
BC = r tan x 2
1 1
Area  0BC = 0B BC = r2 tan x
2 2

One can compare the areas of triangles 0BD, 0BC and sector 0BD to come
up with the inequalities
Area  0BD Area sector 0BD Area  0BC
1 2 1 2 1 2 (1.47)
or r sin x r x r tan x
2 2 2
x 1
Divide this inequality through by 12 r2 sin x to obtain the result 1 .
sin x cos x
Taking the reciprocals one can write
sin x
1 cos x (1.48)
x

2
Radian measure is always used.
 45

Figure 1-32. Construction of triangles and sector associated with circle.

Now take the limit as x approaches zero to show
sin x
1 lim lim cos x (1.49)
x0 x x0

sin x
The function is squeezed or sandwiched between the values 1 and cos x and since
x
the cosine function approaches 1 as x approaches zero, one can say the limit of the
sin x
function must also approach 1 and so one can write
x
sin x
lim = 1 (1.50)
x0 x

In our study of calculus other methods are developed to verify the above limiting
sin x
value associated with the indeterminate form as x approaches zero.
x

Example 1-8. (Algebra used to determine limiting value)

Algebra as well as geometry can be used to aid in evaluating limits. For example,
to calculate the limit
xn 1 xn 1 0
lim = = an indeterminate form (1.51)
x1 x 1 x1 x=1 0

one can make the change of variables z = x1 and express the limit given by equation
(1.51) in the form
(1 + z)n 1
lim (1.52)
z0 z
The numerator of this limit expression can be expanded by using the binomial
theorem
n(n 1) 2 n(n 1)(n 2) 3
(1 + z)n = 1 + n z + z + z + (1.53)
2! 3!
46
Substituting the expansion (1.53) into the equation (1.52) and simplifying reduces
the given limit to the form
 
n(n 1) n(n 1)(n 2) 2
lim n + z+ z + = n (1.54)
z0 2! 3!
xn 1
This shows that lim =n
x1 x 1

Example 1-9. (Limits)

The following are some examples illustrating limiting values associated with
functions.
1 tan
lim x2 = 9 lim = lim =1
x3 x0x 0
1

lim = + 1 x+11 1
x0+ x lim 3 + =3 lim =
x x x0 x 2

Formal Definition of Limit

In the early development of mathematics the concept of a limit was very vague.
The calculation of a limit was so fundamental to understanding certain aspects of
calculus, that it required a precise definition. A more formal  (read epsilon-delta)
definition of a limit was finally developed around the 1800s. This formalization
resulted from the combined research into limits developed by the mathematicians
Weierstrass,3 Bolzano4 and Cauchy.5

Definition 1: Limit of a function

Let f (x) be defined and single-valued for all values of x in some deleted neigh-
borhood of the point x0 . A number
is called a limit of f (x) as x approaches x0 ,
written xx
lim f (x) =
, if for every small positive number  > 0 there exists a number
0
such that 6
|f (x)
| <  whenever 0 < |x x0 | < (1.55)

Then one can write f (x)
(f (x) approaches
) as x x0 (x approaches x0 ). Note

that f (x) need not be defined at the point x0 in order for a limit to exist.
3
Karl Theodor Wilhelm Weierstrass (1815-1897) A German mathematician.
4
Bernard Placidus Johan Nepomuk Bolzano (1781-1848) A Bohemian philosopher and mathematician.
5
Augustin Louis Cauchy (1789-1857) A French mathematician.
6
The number usually depends upon how  is selected.
 47
The above definition must be modified if restrictions are placed upon how x
approaches x0 . For example, the limits lim+ f (x) =
1 and lim f (x) =
2 are called
xx0 xx0
the right-hand and left-hand limits associated with the function f (x) as x approaches
the point x0 . Sometimes the right-hand limit is expressed lim+ f (x) = f (x+ 0 ) and
xx0
the left-hand limit is expressed lim f (x) = f (x
0 ). The  definitions associated
xx0
with these left and right-hand limits is exactly the same as given above with the
understanding that for right-hand limits x is restricted to the set of values x > x0
and for left-hand limits x is restricted to the set of values x < x0 .

Definition 2: Limit of a function f (x) as x

Let f (x) be defined over the unbounded interval c < x < , then a number
is
called a limit of f (x) as x increases without bound, written as x lim f (x) =
, if for
every  > 0, there exists a number N1 > 0 such that |f (x)
| < , whenever x N1 .
In a similar fashion, if f (x) is defined over the unbounded interval < x < c,
then the number
is called a limit of f (x) as x decreases without bound, written
lim f (x) =
, if for every  > 0, there exists a number N2 > 0 such that |f (x)
| < ,
x
whenever x N2.

In terms of the graph { (x, y) | y = f (x), x R } one can say that for x sufficiently
large, larger than N1 or less than N2 , the y values of the graph would get as close
as you want to the line y =
.

Definition 3: Limit of a function becomes unbounded

In the cases where the limit xx
lim f (x) either increases or decreases without bound
0

and does not approach a limit, then the notation xx lim f (x) = + is used to denote
0
that there exists a number N3 > 0, such that f (x) > N3 , whenever 0 < |x x0 | < and
the notation xxlim f (x) = is used to denote that there exists a number N4 > 0,
0

such that f (x) < N4 , whenever 0 < |x x0 | < .

In the above notations the symbols + (plus infinity) and (minus infinity)
are used to denote unboundedness of the functions. These symbols are not numbers.
Also observe that there are situations where to use the above notation one might
have to replace the limit subscript x x0 by either x x+
0 or x x0 in order to

denote right or left-handed limits.

48
In using the  methods to prove limit statements, observe that the statement
f (x) is near
is expressed mathematically by the statement |f (x)
| <  where
 > 0 is very small. By selecting  very small you can force f (x) to be very near
,
but what must be in order that f (x) be that close to
? The selection of , in most
cases, will depend upon how  is specified. The statement that x is near x0 , but x
is not equal to x0 is expressed mathematically by the statement 0 < |x x0 | < .
The real number which is selected to achieve the smallness specified by , is not
a unique number. Once one value of is found, then any other value 1 < would
also satisfy the definition.

Figure 1-33.
(a) Graphical sketch of  limit.
(b) Function having jump discontinuity at the point x0

A sketch of the  definition of a limit is given in the figure 1-33(a). Here

x0 ,
,  > 0, > 0 are all real numbers and the given function y = f (x) is understood to
be well defined for both x < x0 and for x > x0 , while the function value f (x0 ) may or
may not be defined. That portion of the graph inside the shaded rectangle is given
by the set of values

G = { (x, y) | 0 < |x x0 | < and y = f (x) }

which is a subset of all the points inside the shaded rectangle.

 49
The shaded rectangle consists of the set of values

S = { (x, y) | 0 < |x x0 | < and |y
| <  }

Note that the line where x = x0 and
 < y <
+  is excluded from the set. The
problem is that for every  > 0 that is specified, one must know how to select the
to insure the curve stays within the shaded rectangle. If this can be done then

is defined to be the lim f (x). In order to make |f (x)
| small, as x x0 , one must
xx0
restrict the values of x to some small deleted neighborhood of the point x0 . If only
points near x0 are to be considered, it is customary to always select to be less than
or equal to 1. Thus if |x x0 | < 1, then x is restricted to the interval [x0 1, x0 + 1].

Example 1-10. (
proof )

Use the  definition of a limit to prove that lim x2 = 9
x3
Solution
Here f (x) = x2 and
= 9 so that

|f (x)
| = |x2 9| = |(x + 3)(x 3)| = |x + 3| |x 3| (1.56)

To make |f (x)
| small one must control the size of |x 3|. Recall that by agreement
is to be selected such that < 1 and as a consequence of this the statement x is
near 3 is to mean x is restricted to the interval [2, 4]. This information allows us to
place bounds upon the factor (x + 3). That is, |x + 3| < 7, since x is restricted to the
interval [2, 4]. One can now use this information to change equation (1.56) into an
inequality by noting that if |x 3| < , one can then select such that

|f (x)
| = |x2 9| = |x + 3| |x 3| < 7 <  (1.57)

where  > 0 and less than 1, is as small as you want it to be. The inequality (1.57)
tells us that if < /7, then it follows that

|x2 9| <  whenever |x 3| <

Special Considerations
1. The quantity  used in the definition of a limit is often replaced by some scaled

value of , such as , 2 , , etc. in order to make the algebra associated with
some theorem or proof easier.
50
2. The limiting process has the property that for f (x) = c, a constant, for all values
of x, then
lim c = c (1.58)
xx0

This is known as the constant function rule for limits.

3. The limiting process has the property that for f (x) = x, then lim x = x0 .
xx0
This is sometimes called the identity function rule for limits.
Properties of Limits
If f (x) and g(x) are functions and the limits xx
lim f (x) =
1 and lim g(x) =
2 both
xx0
0

exist and are finite, then

(a) The limit of a constant times a function equals the constant times the limit of
the function.
lim cf (x) = c lim f (x) = c
1
xx0 xx0
for all constants c
(b) The limit of a sum is the sum of the limits.
lim [f (x) + g(x)] = lim f (x) + lim g(x) =
1 +
2
xx0 xx0 xx0
(c) The limit of a difference is the difference of the limits.
lim [f (x) g(x)] = lim f (x) lim g(x) =
1
2
xx0 xx0 xx0
(d) The limit of a product

of functions

equals the
 product of the function limits.
lim [f (x) g(x)] = lim f (x) lim g(x) =
1
2
xx0 xx0 xx0
(e) The limit of a quotient is the quotient of the limits provided that the denom-
inator limit is nonzero.
lim f (x)
f (x) xx0
1
lim
xx0 g(x)
= = , provided
2
= 0
lim g(x)
2
xx0
(f) The limit of an nth root is the nth root of the limit.

if n is an odd positive integer or

 

n n
lim f (x) = n lim f (x) =
1
xx0 xx0 if n is an even positive integer and
1 > 0
(g) Repeated applications of the product rule with g(x) = f (x) gives the extended
product rule.
n
n
lim f (x) = lim f (x)
xx0 xx0

(h) The limit theorem for composite
functions is as follows.

If xx
lim g(x) =
, then lim f (g(x)) = f
xx
lim g(x) = f (
)
xx0
0 0
 51
Example 1-11. (Limit Theorem)
Use the  definition of a limit to prove the limit of a sum is the sum of the
limits xx lim (f (x) + g(x)) = lim f (x) + lim g(x) =
1 +
2
xx0 xx0
0

Solution By hypothesis, lim f (x) =
1 and lim g(x) =
2 , so that for a small number
xx0 xx0
1 > 0, there exists numbers 1 and 2 such that

|f (x)
1 | < 1 when 0 < |x x0 | < 1

|g(x)
2 | < 1 when 0 < |x x0 | < 2

where 1 > 0 is a small quantity to be specified at a later time. Select to be the

smaller of 1 and 2 , then using the triangle inequality, one can write
|(f (x) + g(x)) (
1 +
2 )| =|(f (x)
1 ) + (g(x)
2 )|

|f (x)
1 | + |g(x)
2 |
1 + 1 = 21 when 0 < |x x0 | <
Consequently, if 1 is selected as /2, then one can say that

|(f (x) + g(x)) (
1 +
2 )| <  when 0 < |x x0 | <

which implies
lim (f (x) + g(x)) = lim f (x) + lim g(x) =
1 +
2
xx0 xx0 xx0

Example 1-12. (Limit Theorem)

Use the  definition of a limit to prove the limit of a product of functions
equals the product of the function limits. That is, if xx
lim f (x) =
1 and lim g(x) =
2 ,
xx


 0 0

then lim f (x)g(x) = lim f (x) lim g(x) =
1
2 .

xx0 xx0 xx0
Solution
By hypothesis, lim f (x) =
1 and lim g(x) =
2 , so that for every small number
xx0 xx0
1 , there exists numbers 1 , 1 and 1 , 2 such that

|f (x)
1 | < 1 , whenever 0 < |x x0 | < 1

and |g(x)
2 | < 1 whenever 0 < |x x0 | < 2

where 1 > 0 is some small number to be specified later. Let equal the smaller of
the numbers 1 and 2 so that one can write

|f (x)
1 | < 1 , |g(x)
2 | < 1 , whenever 0 < |x x0 | <

52
To prove the above limit one must specify how to select associated with a given
value of  such that

|f (x)g(x)
1
2 | <  whenever |x x0 | <

One can select 1 above as a small number which is some scaled version of . Observe
that the function f (x) is bounded, since by the triangle inequality one can write

|f (x)| = |f (x)
1 +
1 | < |f (x)
1 | + |
1 | < 1 + |
1 | < 1 + |
1 |

where 1 is assumed to be less than unity. Also note that one can write
|f (x)g(x)
1
2 | =|f (x)g(x)
2 f (x) +
2 f (x)
1
2 |

|f (x)(g(x)
2 )| + |
2 (f (x)
1 )|

|f (x)||g(x)
2 | + |
2 ||f (x)
1 |
(1 + |
1 |)1 + |
2 |1 = (1 + |
1 | + |
2 |)1

Consequently, if the quantity 1 is selected to satisfy the inequality (1+|
1|+|
2|)1 < ,

then one can say that

|f (x)g(x)
1
2 | <  whenever |x x0 | <

so that


lim f (x)g(x) = lim f (x) lim g(x) =
1
2
xx0 xx0 xx0

Example 1-13. (Limit Theorem)

1 1 1
If
2
= 0, prove that if lim g(x) =
2 , then lim = =
xx0 xx0 g(x) lim g(x)
2
xx0
Solution By hypothesis xx
lim g(x) =
2 , with
2
= 0. This means that for every 1 > 0
0
there exists a 1 such that |g(x)
2 | < 1 whenever |x x0 | < 1 . How can this
information be used to show that for every  > 0 there exists a such that
 
 1 1 
 g(x)
 < 
  whenever |x x0 | < ? (1.59)
2

The left-hand side of the inequality (1.59) can be expressed

 
 1 1  |
2 g(x)| |g(x)
2 | 1
 g(x)
2  = |g(x)
2| = (1.60)

|
2 | |g(x)|
 53
For a given 1 , one can find a 1 such that the quantity |g(x)
2 | < 1 whenever
|x x0 | < 1 . What can be constructed as an inequality concerning the quantity
1
? If
2
= 0 one can employ the triangle inequality and write
|g(x)|
|
2 | = |
2 g(x) + g(x)| |
2 g(x)| + |g(x)|

By the definition of a limit, one can select values 3 and 3 such that
|g(x)
2 | < 3 when |x x0 | < 3
This gives the inequalities
1 1
|
2 | < 3 + |g(x)| or |
2 | 3 < |g(x)| or < (1.61)
|g(x)| |
2 | 3
provided |
2 | 3 is not zero. Recall that the values of 1 and 3 have not been
specified and their values can be selected to have any small values that we desire.
The inequality given by equation (1.60) can be expressed in the form
 
 1 1  |g(x)
2 | 1 1 1
 g(x)
 < < (1.62)

2 |
2 | |g(x)| |
2 | |
2 | 3
and is valid for all x values satisfying |x x0| < , where is selected as the smaller of
the values 1 and 3 . Let us now specify an 1 and 3 value so that with some algebra
the right-hand side of equation (1.62) can be made less than  for |x x0 | < . One
way to accomplish this is as follows. After  is selected, one can select 1 above to go
with 1 = (1 )|
2 |2 and select 3 above to go with 3 = |
2 |, where is some small
fraction less than 1. Then is selected as the smaller of the values 1 and 3 and the
product on the right-hand side of equation (1.62) is less than  for |x x0 | < .
The above result can now be combined with the limit of a product rule
1
lim f (x)h(x) = lim f (x) lim h(x) with h(x) = g(x) to establish the quotient rule
xx0 xx 0 xx 0


 lim f (x)
f (x) 1 xx0
1
lim
xx0 g(x)
= lim f (x) lim
xx0 g(x)
= = , provided
2
= 0
xx0 lim g(x)
2
xx0

The Squeeze Theorem

Assume that for x near x0 there exists three functions f (x), g(x) and h(x) which
can be shown to satisfy the inequalities f (x) g(x) h(x). If one can show that
lim f (x) =
and lim h(x) =
,
xx0 xx0

then one can conclude

lim g(x) =

xx0

This result is known as the squeeze theorem.

54
Continuous Functions and Discontinuous Functions
A function f (x) is called a continuous function over the interval a x0 b if for
all points x0 within the interval
(i) f (x0 ) is well defined
(ii) lim f (x)
xx0
exists (1.63)
(iii) lim f (x) = f (x0 )
xx0

Polynomial functions are continuous functions. Rational algebraic functions,

represented by the quotient of two polynomials, are continuous except for those
points where the denominator becomes zero. The trigonometric functions, exponen-
tial functions and logarithmic functions are continuous functions over appropriate
intervals.
Alternatively, one can assume the right-hand limit lim f (x) =
1 and the left-
xx0 +
hand limit lim f (x) =
2 both exist, then if
1 =
2 and f (x0 ) =
1 , then the function
xx0
f (x)is said to be continuous at the point x0 . A function continuous at all points x0
within an interval is said to be continuous over the interval.
If any of the conditions given in equation (1.63) are not met, then f (x) is called
a discontinuous function. For example, if lim f (x) =
1 and lim f (x) =
2 both exist
xx + 0
xx0
and
1
=
2 , the function f (x) is said to have a jump discontinuity at the point x0 .
An example of a function with a jump discontinuity is given in the figure 1-33(b).
If xx
lim f (x) does not exist, then f (x) is said to be discontinuous at the point x0 .
0

The Intermediate Value Property states that a function f (x) which is continuous
on a closed interval a x b is such that when x moves from the point a to the point
b the function takes on every intermediate value between f (a) and f (b) at least once.
An alternative version of the intermediate value property
is the following. If y = f (x) denotes a continuous function
on the interval a x b, where f (a) < c < f (b), and the
line y = c = constant is constructed , then the Intermediate
Value Theorem states that there must exist at least one
number satisfying a < < b such that f () = c.
 55
Example 1-14. (Discontinuities)
2
x 1
(a) f (x) = x1
is not defined at the point x = 1, so f (x) is said to be discontin-
(x 1)(x + 1)
uous at the point x = 1. The limit lim = 2 exists and so by defining
x1 x1
the function f (x) to have the value f (1) = 2, the function can be made continuous.
In this case the function is said to have a removable discontinuity at the point
x = 1.
(b) If xx
lim f (x) = , then obviously f (x) is not defined at the point x0 . Another
0

way to spot an infinite discontinuity is to set the denominator of a function equal

1
to zero and solve for x. For example, if f (x) = , then f (x) is
(x 1)(x 2)(x 3)
said to have infinite discontinuities
 at the points x = 1, x = 2 and x = 3.
1, x<2
(c) The function f (x) = is not defined at the point x = 2. The
5, x>2
left and right-hand limits as x 2 are
not the same and so the function is said
to have a jump discontinuity at the point x = 2. The limit limx2 f (x) does not
exist. At a point where a jump discontinuity occurs, it is sometimes convenient
to define the value of the function as the average value of the left and right-hand
limits.

Asymptotic Lines
A graph is a set of ordered pairs (x, y) which are well defined over some region
of the x, y-plane. If there exists one or more straight lines such that the graph
approaches one of these lines as x or y increases without bound, then the lines are
called asymptotic lines.

Example 1-15. (Asymptotic Lines)

1
Consider the curve C = { (x, y) | y = f (x) = 1 + , x R } and observe that the
x1
curve passes through the origin since when x = 0 one finds y = f (0) = 0. Also note
that as x increases, lim f (x) = 1 and that as x approaches the value 1, lim f (x) = .
x x1
If one plots some selected points one can produce the illustration of the curve C as
given by the figure 1-34. In the figure 1-34 the line y = 1 is a horizontal asymptote
associated with the curve and the line x = 1 is a vertical asymptote associated with
the curve.
56
Consider a curve defined by one of the equations

G(x, y) = 0, y = f (x), x = g(y)

If a line
is an asymptotic line associated with one of the above curves, then the
following properties must be satisfied. Let d denote the perpendicular distance from
a point (x, y) on the curve to the line
. If one or more of the conditions

lim d = 0, lim d = 0, lim d = 0, lim d = 0,

x x y y

is satisfied, then the line
is called an asymptotic line or asymptote associated with
the given curve.

1
y =1+
x1
y=1

x=1
1
Figure 1-34. The graph of y = f (x) = 1 +
x1

Finding Asymptotic Lines

One can determine an asymptotic line associated with a curve y = f (x) by ap-
plying one of the following procedures.
1. Solve for y in terms of x and set the denominator equal to zero and solve for x.
The resulting values for x represent the vertical asymptotic lines.
2. Solve for x in terms of y and set the denominator equal to zero and solve for y.
The resulting values for y represent the horizontal asymptotic lines.
3. The line x = x0 is called a vertical asymptote if one of the following conditions
is true.
lim f (x) = , lim f (x) = , lim f (x) =
xx0 xx xx+
0 0

lim f (x) = , lim f (x) = , lim f (x) =

xx0 xx xx+
0 0
 57
4. The line y = y0 is called a horizontal asymptote if one of the following condi-
tions is true.
lim f (x) = y0 , lim f (x) = y0
x x

5. The line y = mx + b is called a slant asymptote or oblique asymptote if

lim [f (x) (mx + b)] = 0

x

Example 1-16. Asymptotic Lines

1
Consider the curve y = f (x) = 2x + 1 + , where x R. This function has the
x
properties that
1
lim [f (x) (2x + 1)] = lim
x x x
=0 and lim f (x) =
x0

so that one can say the line y = 2x + 1 is an oblique asymptote and the line x = 0 is
a vertical asymptote. A sketch of this curve is given in the figure 1-35.

y = 2x + 1

x=0
1
Figure 1-35. Sketch of curve y = f (x) = 2x + 1 +
x

Conic Sections
A general equation of the second degree has the form

Ax2 + Bxy + Cy 2 + Dx + Ey + F = 0 (1.64)

where A, B, C, D, E, F are constants. All curves which have the form of equation (1.64)
can be obtained by cutting a right circular cone with a plane. The figure 1-36(a)
illustrates a right circular cone obtained by constructing a circle in a horizontal plane
and then moving perpendicular to the plane to a point V above or below the center
of the circle. The point V is called the vertex of the cone. All the lines through the
point V and points on the circumference of the circle are called generators of the
58
cone. The set of all generators produces a right circular cone. The figure 1-36(b)
illustrates a horizontal plane intersecting the cone in a circle. The figure 1-36(c)
illustrates a nonhorizontal plane section which cuts two opposite generators. The
resulting curve of intersection is called an ellipse. Figure 1-36(d) illustrates a plane
parallel to a generator of the cone which also intersects the cone. The resulting
curve of intersection is called a parabola. Any plane cutting both the upper and
lower parts of a cone will intersect the cone in a curve called a hyperbola which is
illustrated in the figure 1-36(e).

Figure 1-36. The intersection of right circular cone with a plane.

Conic sections were studied by the early Greeks. Euclid7 supposedly wrote four
books on conic sections. The Greek geometer Appollonius8 wrote eight books on
conic sections which summarized Greek knowledge of conic sections and his work
has survived the passage of time.
Conic sections can be defined as follows. In the xy-plane select a point f , called
the focus, and a line D not through f . This line is called the directrix. The set of
points P satisfying the condition that the distance from f to P , call it r = P f , is some
multiple e times the distance d = P P
, where d represents the perpendicular distance
from the point P to the line D. The resulting equation for the conic section is
obtained from the equation r = e d with the geometric interpretation of this equation
illustrated in the figure 1-37.
7
Euclid of Alexandria (325-265 BCE)
8
Appollonius of Perga (262-190 BCE)
 59

The plane curve resulting from the equation

r = e d is called a conic section with eccentricity
e, focus f and directrix D and if the eccentricity e
satisfies
0 < e < 1, the conic section is an ellipse.
e = 1, the conic section is a parabola.
e > 1, the conic section is a hyperbola.

Figure 1-37.
Defining a conic section.

In addition to the focus and directrix there is associated with each conic section
the following quantities.
The vertex V The vertex V of a conic section is the midpoint of a line from the
focus perpendicular to the directrix.
Axis of symmetry The line through the focus and perpendicular to the directrix
is called an axis of symmetry.
Focal parameter 2p This is the perpendicular distance from the focus to the
directrix, where p is the distance from the focus to the vertex or distance from
vertex to directrix.
Latus rectum 2
This is a chord parallel to a directrix and perpendicular to a
focus which passes between two points on the conic section. The latus rectum
is used as a measure associated with the spread of a conic section. If
is the
semi-latus rectum intersecting the conic section at the point where x = p, one
finds r =
= ed and so it follows that 2
= 2ed.
Circle
A circle is the locus of points (x, y) in a plane equidistant from a fixed point called
the center of the circle. Note that no real locus occurs if the radius r is negative or
imaginary. It has been previously demonstrated how to calculate the equation of a
circle. The figure 1-38 is a summary of these previous results. The circle x2 + y2 = r2
has eccentricity zero and latus rectum of 2r. Parametric equations for the circle
(x x0 )2 + (y y0 )2 = r 2 , centered at (x0 , y0 ), are

x = x0 + r cos t, y = y0 + r sin t, 0 t 2
60
When dealing with second degree equations of the form x2 + y2 + x + y = ,
where , and are constants, it is customary to complete the square on the x and
y terms to obtain

2 2 2 2
(x2 + x + ) + (y 2 + y + )= + + = (x + )2 + (y + )2 = r 2
4 4 4 4 2 2
2 2
where it is assumed that r2 = + 4 + 4 > 0.
This produces

the equation of a circle

with radius r which is centered at the point 2 , 2 .

Figure 1-38.
Circle about origin and circle translated to point (x0 , y0 )

Parabola
The parabola can be defined as the locus of points (x, y) in a plane, such that
(x, y) moves to remain equidistant from a fixed point (x0 , y0 ) and fixed line
. The fixed
point is called the focus of the parabola and the fixed line is called the directrix of
the parabola. The midpoint of the perpendicular line from the focus to the directrix
is called the vertex of the parabola.
In figure 1-39(b), let the point (0, p) denote the focus of the parabola symmetric
about the y-axis and let the line y = p denote the directrix of the parabola. If (x, y)
is a general point on the parabola, then

d1 = distance from (x, y) to focus = x2 + (y p)2
d2 = distance from (x, y) to directrix = y + p
 61
If d1 = d2 for all values of x and y, then

x2 + (y p)2 = y + p or x2 = 4py, p
= 0 (1.65)

This parabola has its vertex at the origin, an eccentricity of 1, a semi-latus rectum
of length 2p, latus rectum of 2
= 4p and focal parameter of 2p.

Figure 1-39.
Parabolas symmetric about the x and y axes.

Other forms for the equation of a parabola are obtained by replacing p by p

and interchanging the variables x and y. For p > 0, other standard forms for the
equation of a parabola are illustrated in the figure 1-40. In the figure 1-40 observe
the upward/downward and left/right opening of the parabola depend upon the sign
before the parameter p, where p > 0 represents the distance from the origin to the
focus. By replacing x by x and y by y one can verify the various symmetries
associated with these shapes.
Using the translation of axes equations (1.30), the vertex of the parabolas in the
figure 1-40 can be translated to a point (h, k). These translated equations have the
representations
(x h)2 = 4p(y k) (y k)2 = 4p(x h)
(1.66)
(x h)2 = 4p(y k) (y k)2 = 4p(x h)
Also note that the lines of symmetry are also shifted.
62

Figure 1-40.
Other forms for representing a parabola.

One form for the parametric representation of the parabola (x h)2 = 4p(y k) is
given by
P = { (x, y) | x = h + t, y = k + t2 /4p, < t < } (1.67)

with similar parametric representations for the other parabolas.

Use of determinants
The equation of the parabola passing through the three distinct points (x1 , y1 ),
(x2 , y2 ) and (x3 , y3 ) can be determined by evaluating the determinant9

y x2 x 1
 
 y1 x21 x1 1
 
= 0
 y2 x22 x2 1

y3 x23 x3 1
 

provided the following determinants are different from zero.

   
 x2 x1 1   x1 y1 1 
 12 
 x2
 2 x2 1 
= 0, and  x2
 y2 1 
= 0
x x3 1  x3 y3 1
3

9
Determinants and their properties are discussed in chapter 10.
 63
Ellipse
The eccentricity e of an ellipse satisfies 0 < e < 1 so that for any given positive
number a one can state that
a
ae < , 0<e<1 (1.68)
e

Consequently, if the point (ae, 0) is selected as the focus of an ellipse and the line
x = a/e is selected as the directrix of the ellipse, then in relation to this fixed focus
and fixed line a general point (x, y) will satisfy


d1 = distance of (x, y) to focus = (x ae)2 + y 2
d2 = distance of (x, y) to directrix = |x a/e|

The ellipse can then be defined as the set of points (x, y) satisfying the constraint
condition d1 = e d2 which can be expressed as the set of points

E1 = { (x, y) | (x ae)2 + y 2 = e|x a/e|, 0 < e < 1 } (1.69)

Applying some algebra to the constraint condition on the points (x, y), the ellipse
can be expressed in a different form. Observe that if d1 = e d2 , then

(x ae)2 + y 2 =e2 (x a/e)2

or x2 2aex + a2 e2 + y 2 =e2 x2 2aex + a2

which simplifies to the condition

x2 y2 x2 y2
+ = 1 or + = 1, b2 = a2 (1 e2 ) (1.70)
a2 a2 (1 e2 ) a2 b2

where the eccentricity satisfies 0 < e < 1. In the case where the focus is selected as
(ae, 0) and the directrix is selected as the line x = a/e, there results the following
situation


d1 = distance of (x, y) to focus = (x + ae)2 + y 2
d2 = distance of (x, y) to directrix = |x + a/e|
64
The condition that d1 = e d2 can be represented as the set of points

E2 = { (x, y) | (x + ae)2 + y 2 = e|x + a/e|, 0 < e < 1 } (1.71)

As an exercise, show that the simplification of the constraint condition for the set
of points E2 also produces the equation (1.70).

x2 y2
Figure 1-41. The ellipse + =1
a2 b2

Define the constants

c = ae and b2 = a2 (1 e2 ) = a2 c2 (1.72)

and note that b2 < a2 , then from the above discussion one can conclude that an
ellipse is defined by the equation
x2 y2
+ = 1, 0 < e < 1, b2 = a2 (1 e2 ), c = ae (1.73)
a2 b2

and has the points (ae, 0) and (ae, 0) as foci and the lines x = a/e and x = a/e as
directrices. The resulting graph for the ellipse is illustrated in the figure 1-41. This
 65
ellipse has vertices at (a, 0) and (a, 0), a latus rectum of length 2b2 /a and eccentricity

given by 1 b2 /a2 .
In the figure 1-41 a right triangle has been constructed as a mnemonic device to
help remember the relations given by the equations (1.72). The distance 2a between
(a, 0) and (a, 0) is called the major axis of the ellipse and the distance 2b from (0, b)
to (0, b) is called the minor axis of the ellipse. The origin (0, 0) is called the center of
the ellipse.

Figure 1-42. Symmetry of the ellipse.

Some algebra can verify the following property satisfied by a general point (x, y)
on the ellipse. Construct the distances

d3 = distance of (x, y) to focus (c, 0) = (x c)2 + y 2
 (1.74)
d4 = distance of (x, y) to focus (c, 0) = (x + c)2 + y 2

and show
 
d3 + d4 = (x c)2 + y 2 + (x + c)2 + y 2 = 2a (1.75)

One can use this property to define the ellipse as the locus of points (x, y) such
that the sum of its distances from two fixed points equals a constant.
The figure 1-42 illustrates that when the roles of x and y are interchanged, then
the major axis and minor axis of the ellipse are reversed. A shifting of the axes so
that the point (x0 , y0 ) is the center of the ellipse produces the equations
(x x0 )2 (y y0 )2 (y y0 )2 (x x0 )2
+ = 1 or + =1 (1.76)
a2 b2 a2 b2
These equations represent the ellipses illustrated in the figure 1-42 where the centers
are shifted to the point (x0 , y0 ).
66
(x h)2 (y k)2
The ellipse given by + = 1 which is centered at the point (h, k)
a2 b2
10
can be represented in a parametric form . One parametric form is to represent the
ellipse as the set of points
E = { (x, y) | x = h + a cos , y = k + b sin , 0 2 } (1.77)

involving the parameter which varies from 0 to 2.

Hyperbola
Let e > 1 denote the eccentricity of a hyperbola. Again let (ae, 0) denote the focus
of the hyperbola and let the line x = a/e denote the directrix of the hyperbola. The
hyperbola is defined such that points (x, y) on the hyperbola satisfy d1 = ed2 where
d1 is the distance from (x, y) to the focus and d2 is the perpendicular distance from
the point (x, y) to the directrix. The hyperbola can then be represented by the set
of points


H1 = { (x, y) | (x ae)2 + y 2 = e|xa/e|, e > 1 }

A simplification of the constraint condition on the set of points (x, y) produces

the alternative representation of the hyperbola
x2 y2
= 1, e>1 (1.78)
a2 a2 (e2 1)
Placing the focus at the point (ae, 0) and using as the directrix the line x = a/e,
one can verify that the hyperbola is represented by the set of points


H2 = { (x, y) | (x + ae)2 + y 2 = e|x + a/e|, e > 1 }

and it can be verified that the constraint con-

dition on the points (x, y) also simplifies to the
equation (1.78).

10
The parametric representation of a curve or part of a curve is not unique.
 67
Define c = ae and b2 = a2 (e2 1) = c2 a2 > 0 and note that for an eccentricity
e > 1 there results the inequality c > a. The hyperbola can then be described as
having the foci (c, 0) and (c, 0) and directrices x = a/e and x = a/e. The hyperbola
represented by
x2 y2
= 1, b2 = a2 (e2 1) = c2 a2 (1.79)
a2 b2
is illustrated in the figure 1-43.

x2 y2
Figure 1-43. The hyperbola =1
a2 b2

This hyperbola has vertices at (a, 0) and (a, 0), a latus rectum of length 2b2 /a and

eccentricity of 1 + b2 /a2 . The origin is called the center of the hyperbola. The line
containing the two foci of the hyperbola is called the principal axis of the hyperbola.
Setting y = 0 and solving for x one can determine that the hyperbola intersects
the principal axis at the points (a, 0) and (a, 0) which are called the vertices of the
hyperbola. The line segment between the vertices is called the major axis of the
hyperbola or transverse axis of the hyperbola. The distance between the points (b, 0)
and (b, 0) is called the conjugate axis of the hyperbola. The chord through either
focus which is perpendicular to the transverse axis is called a latus rectum. One
can verify that the latus rectum intersects the hyperbola at the points (c, b2/a) and
(c, b2/a).
Write the equation (1.79) in the form

b a2
y = x 1 2 (1.80)
a x
68
and note that for very large values of x the right-hand side of this equation ap-
proaches 1. Consequently, for large values of x the equation (1.80) becomes the
lines
b b
y= x and y= x (1.81)
a a
These lines are called the asymptotic lines associated with the hyperbola and are
illustrated in the figure 1-43. Note that the hyperbola has two branches with each
branch approaching the asymptotic lines for large values of x.
Let (x, y) denote a general point on the above hyperbola and construct the dis-
tances 
d3 = distance from (x, y) to the focus (c, 0) = (x c)2 + y 2
 (1.82)
d4 = distance
from (x, y) to the focus (c, 0) = (x + c)2 + y 2
Use some algebra to verify that
d4 d3 = 2a (1.83)

This property of the hyperbola is sometimes used to define the hyperbola as the
locus of points (x, y) in the plane such that the difference of its distances from two
fixed points is a constant.
The hyperbola with transverse axis on the x-axis have the asymptotic lines
y = + ab x and y = ab x. Any hyperbola with the property that the conjugate axis has
the same length as the transverse axis is called a rectangular or equilateral hyperbola.
Rectangular hyperbola are such that the asymptotic lines are perpendicular to each
other.

Figure 1-44. Conjugate hyperbola.

 69
If two hyperbola are such that the transverse axis of either is the conjugate axis
of the other, then they are called conjugate hyperbola. Conjugate hyperbola will
have the same asymptotic lines. Conjugate hyperbola are illustrated in the figure
1-44.
The figure 1-45 illustrates that when the roles of x and y are interchanged, then
the transverse axis and conjugate axis of the hyperbola are reversed. A shifting
of the axes so that the point (x0 , y0 ) is the center of the hyperbola produces the
equations
(x x0 )2 (y y0 )2 (y y0 )2 (x x0 )2
= 1 or =1 (1.84)
a2 b2 a2 b2

The figure 1-45 illustrates what happens to the hyperbola when the values of x and
y are interchanged.

If the foci are on the x-axis at If the foci are on the y-axis at
2
x y2 y2 x2
(c, 0) and (c, 0), then 2 =1 (0, c) and (0, c), then =1
a2 b a2 b2
Figure 1-45.
Symmetry of the hyperbola .
x2 y2
The hyperbola 2 2 = 1 can also be represented in a parametric form as the
a b
set of points
H = H1 H2 where
H1 = { (x, y) | x = a cosh t, y = b sinh t, < t < } (1.85)

and H2 = { (x, y) | x = a cosh t, y = b sinh t, < t < }

which represents a union of the right-branch and left-branch of the hyperbola. Simi-
lar parametric representations can be constructed for those hyperbola which undergo
70
a translation or rotation of axes. Remember that the parametric representation of
a curve is not unique.
Conic Sections in Polar Coordinates
Place the origin of the polar coordinate system at the focus of a conic section
with the y-axis parallel to the directrix as illustrated in the figure 1-46. If the point
(x, y) = (r cos , r sin ) is a point on the conic section, then the distance d from the
point (x, y) to the directrix of the conic section is given by either

d = p + r cos or d = p r cos (1.86)

depending upon whether the directrix is to the left or right of the focus. The conic
section is defined by r = ed so there results two possible equations r = e(p r cos ) or
r = e(p + r cos ). Solving these equations for r demonstrates that the equations

ep ep
r= or r= (1.87)
1 e cos 1 + e cos

represent the basic forms associated with representing a conic section in polar coor-
dinates.

Figure 1-46. Representing conic sections using polar coordinates.

If the directrix is parallel to the x-axis at y = p or y = p, then the general forms for
representing a conic section in polar coordinates are given by
ep ep
r= or r= (1.88)
1 e sin 1 + e sin

If the eccentricity satisfies e = 1, then the conic section is a parabola, if 0 < e < 1,
an ellipse results and if e > 1, a hyperbola results.
 71
General Equation of the Second Degree
Consider the equation
ax2 + bxy + cy 2 + dx + ey + f = 0, (1.89)

where a, b, c, d, e, f are constants, which is a general equation of the second degree. If

one performs a rotation of axes by substituting the rotation equations
x = x cos y sin and y = x sin + y cos (1.90)

into the equation (1.89), one obtains the new equation

a x2 + b x y + c y 2 + dx + e y + f = 0 (1.91)

with new coefficients a, b, c, d, e, f defined by the equations

a =a cos2 + b cos sin + c sin2 d =d cos + e sin
b =b(cos2 sin2 ) + 2(c a) sin cos e = d sin + e cos (1.92)

c =a sin2 b sin cos + c cos2 f =f

As an exercise one can show the quantity b2 4ac, called the discriminant, is an
invariant under a rotation of axes. One can show b2 4ac = b2 4ac. The discriminant is
used to predict the conic section from the equation (1.89). For example, if b2 4ac < 0,
then an ellipse results, if b2 4ac = 0, then a parabola results, if b2 4ac > 0, then a
hyperbola results.
In the case where the original equation (1.89) has a cross product term xy, so
that b
= 0, then one can always find a rotation angle such that in the new equations
(1.91) and (1.92) the term b = 0. If the cross product term b is made zero, then one
can complete the square on the x and y terms which remain. This completing the
square operation converts the new equation (1.90) into one of the standard forms
associated with a conic section. By setting the b term in equation (1.92) equal to
zero one can determine the angle such that b vanishes. Using the trigonometric
identities
cos2 sin2 = cos 2 and 2 sin cos = sin 2 (1.93)

one can determine the angle which makes the cross product term vanish by solving
the equation
b = b cos 2 + (c a) sin 2 = 0 (1.94)

for the angle . One finds the new term b is zero if is selected to satisfy
ac
cot 2 = (recall our hypothesis that b
= 0) (1.95)
b
72
Example 1-17. (Conic Section) Sketch the curve 4xy 3y2 = 64
Solution
To remove the product term xy from the general equation
ax + bxy + cy 2 + dx + ey + f = 0 of a conic, the axes must be rotated through an angle
2

determined by the equation cot 2 = ac b

= 34 . For the given conic a = 0, b = 4, c = 3.
This implies that cos 2 = 3/5 = 1 2 sin2 or

2 sin2 = 2/5 giving sin = 1/ 5 and cos = 2/ 5.
The rotation equations (1.90) become
1 1
x = (2x y) and y = (x + 2y)
5 5

The given equation then becomes



2
2x y x + 2y x + 2y
4 3 = 64
5 5 5

x2 y 2
which simplifies to the hyperbola 2 2 = 1 with
8 4
respect to the x and y axes.

Example 1-18. The parametric forms for representing conic sections are not
unique. For a, b constants and , t used as parameters, the following are some repre-
sentative parametric equations which produce conic sections.
Parametric form for conic sections
Conic Section x y parameter
Circle a cos a sin
Parabola at2 2at t
Ellipse a cos b sin
Hyperbola a sec b tan
Rectangular Hyperbola at a/t t
The symbol a > 0 denotes a nonzero constant.
The shape of the curves depends upon the range of values assigned to the pa-
rameters representing the curve. Because of this restriction, the parametric repre-
sentation usually only gives a portion of the total curve. Sample graphs using the
parameter values indicated are given below.
 73

Computer Languages
There are many computer languages and apps that can do graphics and math-
ematical computations to aid in the understanding of calculus. Many of these pro-
gramming languages can be used to perform specific functions on a computing device
such as a desk-top computer, a lap-top computer, a touch-pad, or hand held calcu-
lator. The following is a partial list11 of some computer languages that you might
want to investigate. In alphabetical order:
Ada, APL, C, C++, C#, Cobol, Fortran, Java, Javascript, Maple, Mathcad, Math-
ematica, Matlab, Pascal, Perl, PHP, Python, Visual Basic.

11
For a more detailed list of programming languages go to
en.Wikipedia.org/wiki/List of programming Languages
74
Exercises

1-1. Find the union A B and intersection A B if

(a) A = {x| 2 < x 4} and B = {x| 2 x < 7}
(b) A = {x| 2 < x 4} and B = {x| 4 < x 7}
(c) A = {x| x3 < 8} and B = {x| x2 < 16}

1-2. Sketch a Venn diagram to illustrate the following statements.

(a) AB (c) B A (e) (A B) (A C)
(b) AB = (d) A (B C) (f ) A Bc

1-3. If any set operation involving , U, , is an identity, then the principle of

duality states that the replacements U, U , , in the identity
produces a dual statement which is also an identity. Determine the dual statements
associated with the given identities.

(a) (U A) (B A) = A (b) (B A) ( A) = A (c) A (A B) = A

1-4. Show that the following are equivalent.

(a) AB if and only if A B = B
(b) AB if and only if A B c =
(c) AB if and only ifAc B = U

1-5. Prove the absorption laws

(a) A (A B) = A (b) A (A B) = A

1-6. Shade the Venn diagram to represent the statement underneath.

ABC A (B C) Ac B c C
 75

1-7. Sketch a Venn diagram to illustrate the following set operations.

(a) A (B C) (b) (A B)c (c) (A B C)c

1-8. Determine if the given sets are bounded. If a set is bounded above find the
least upper bound (
.u.b.), if the set is bounded below, find its greatest lower bound
(g.
.b.).
(a) Sa = {x| x2 < 16} (c) Sc = {x| x < 5}

(b) Sb = {x|x3 < 27} (d) Sd = {x| 3 x > 3}

1-9. Find the general equation of the line satisfying the given conditions.
(a) The line passes through the point (2, 4) with slope -2.
(b) The line has zero slope and passes through the point (2, 4)
(c) The line is parallel to 2x + 3y = 4 and passes through the point (2, 4)
(d) The line is parallel to the y-axis and passes through the point (2, 4)

1-10. Express the line 3x + 4y = 12 in the following forms.

(a) The slope-intercept form and then find the slope and y-intercept.
(b) The intercept form and then find the x-intercept and y-intercept.
(c) Polar form.
(d) The point-slope form using the point (1, 1)

1-11. Determine conditions that x must satisfy if the following inequalities are to
be satisfied.
12
(a) x < 0 (c) x+1 <0
x
2x + 3 x+2
(b) <0 (d) 0
x+4 x3

1-12. For each function state how the domain of the function is to be restricted?

(a) y = f (x) = 8x
1
(b) y = f (x) = , a, b, c are real constants.
(x a)(x b)(x c)
(c) Area of a circle is given by A = f (r) = r2
x1
(d) y = f (x) =
x3 + 1
4
(e) Volume of a sphere V = f (r) = r3
3
76

1-13. Sketch a graph of the given functions.
1
(a) y= x, y = x, y = 2x, 4 x 4
2
1
(b) y = x2 , y = x2 , y = 4x2 , 4 x 4
4
1
(c) y = sin x, y = sin x, y = 2 sin x, 0 x 2
2
1
(d) y = cos x, y = cos x, y = 2 cos x, 0 x 2
2

1-14. Sketch the graphs defined by the parametric equations.

(a) Ca = { (x, y) | x = t2 , y = 2t + 1, 0t4 }

(b) Cb = { (x, y) | x = t, y = 2t + 1,2 t 2 }
3
(c) Cc = { (x, y) | x = cos t, y = sin t, t }
2 2
(d) Cd = { (x, y) | x = sin t, y = cos t, 0 t }

(e) Ce = { (x, y) | x = t, y = 9 t2 , 3 t 3 }

Note that the part of the curve represented depends on (i) the form of the parametric
representation and (ii) the values assigned to the parameters.

1-15. Sketch a graph of the given polynomial functions for x R.

(a) y = x1 (b) y = x2 2x 3 (c) y = (x 1)(x 2)(x 3)

(d) Show the function y = (x 1)(x 2)(x 3) is skew-symmetric about the line x = 2.

1-16. The Heaviside12 step function is defined


0, <0
H() =
1, >0

Sketch the following functions.

(a) y = H(x) (d) y = H(x 1) H(x 2)

(b) y = H(x 1) (e) y = H(x) + H(x 1) 2H(x 2)

1
(c) y = H(x 2) (f ) y = [H(x x0 ) H(x (x0 + ))] , >0 is small.

12
Oliver Heaviside (1850-1925) An English engineer.
 77

1-17. Sketch the given curves.
(a) { (x, y) | y = x2 , 2 x 2} (d) { (x, y) | y = (x 1)2 , 1 x 3}
(b) { (x, y) | y = 1 + x2 , 2 x 2} (e) { (x, y) | y = (x + 1)2 , 3 x 1}

(c) { (x, y) | y = 1 + x2 , 2 x 2} (f ) { (x, y) | y = 1 + (x 1)2 , 1 x 3}

1-18. In polar coordinates the equation of a circle with radius and center at the
point (r1 , 1 ) is given by
r 2 + r12 2rr1 cos( 1 ) = 2

Write the equation of the circle and sketch its graph in polar coordinates for the
following special cases.
(a) r1 = , 1 = 0 (d) r1 = , 1 = 3/2

(b) r1 = , 1 = /2 (e) r1 = 0, 1 = 0
(c) r1 = , 1 = (f ) r1 = 3, 1 = /4 in the cases < 3, = 3, > 3

1-19. In rectangular coordinates the equation of a circle with radius > 0 and
center (h, k) is given by the equation

(x h)2 + (y k)2 = 2

Write the equation of the circle and sketch its graph in the following special cases.
(a) h = , k=0
(d) h = 0, k =
(b) h = 0, k=
(e) h = 3, k = 4, in the cases < 5, = 5 and > 5
(c) h = , k=0

1-20. Show that each trigonometric function of an acute angle is equal to the
co-function of the complementary angle = 2 .

sin = cos tan = cot sec = csc

cos = sin cot = tan csc = sec

1-21. If f (x) = x, 0 x < 1, and f (x + 1) = f (x) for all values of x, sketch a graph of
this function over the domain X = { x | 0 x < 5 }.
78

1-22. If f (x) = x2 and g(x) = 3 2x, calculate each of the following quantities.
f (x + h) f (x)
(a) f (3) (c) f (x + h) (e) (g) f (g(x))
h
(b) g(3) (d) g(x + h) g(x + h) g(x) (h) g(f (x))
(f )
h

1-23. Sketch the given curves.
(a) { (x, y) | y = sin x, 0 x 2 } (c) { (x, y) | y = sin(2x), 0 x 2 }


1
(b) { (x, y) | y = sin x , 0 x 2 } (d) { (x, y) | y = sin(x ), 0 x 2 }
2

1-24. Sketch the given curves.
(a) { (x, y) | y = cos x, 0 x 2 } (c) { (x, y) | y = cos(2x), 0 x 2 }


1
(b) { (x, y) | y = cos x , 0 x 2 } (d) { (x, y) | y = cos(x ), 0 x 2 }
2

1-25. Graph the functions and then find the inverse functions.
(a) y = f1 (x) = x2 ,
3
(d) y = f4 (x) = x+4
(b) y = f2 (x) = x3 2x 3
(e) y = f5 (x) =
(c) y = f3 (x) = 5x 1 5x 2

1-26. Test for symmetry, asymptotes and intercepts and then sketch the given
curve.
1
(a) y =1 (d) y 2 x2 y 2 = 1
x2
1 (e) x2 y 2y = 1
(b) y =1+
(x 1)(x 3)2
(f ) xy = x2 1
(c) x2 y = 1

1-27. Sketch the given curves.

(a) { (x, y) | x = 3 cosh t, y = 4 sinh t, 0t3 }
(b) { (x, y) | x = 3 cos t, y = 4 sin t, 0 t 2 }

(c) { (x, y) | x = 3 + t, y = 4 + t2 /4, 3 t 3 }

4
(d) { (r, ) | r = , 0 2 }
1 + 2 cos

1-28. Test for symmetry and sketch the given curves.
x2 y2 x2 y2
(a) y=x 2
(d) + =1 (g) =1
9 16 9 16
y2 x2 y2 x2
(b) x = y2 (h) =1
(e) + =1 4 16
4 16
(c) y 2 = x y2 x2
(f ) x2 + y 2 = 25 (i) =1
9 4
 79

1-29. Translate axes, then sketch the given curves.
(x 1)2 (y 2)2
(a) 2
(x 2) = 4(y 1) (a) + =1
9 16
(x 1)2 (x 2)2
(b) (y 2)2 = 4(x 1) (b) =1
9 16
(c) (x 3)2 = 8(y 2) (y 1)2 (x 2)2
(c) + =1
9 16

1-30. The number e
Consider two methods for estimating the limit e = lim (1 + h)1/h
h0
Method 1
n
1
(a) Make the substitution n = 1/h and show e = lim 1 + and then use the
n n
binomial theorem to show
1 n1 (1 n1 )(1 n2 ) (1 n1 )(1 n2 )(1 n3 )
 
1
e = lim 1 + + + + +
n 1! 2! 3! 4!

(b) Show that as n increases without bound that

1 1 1 1
e=1+ + + ++ +
1! 2! 3! n!

Find the sum of 5,6 and 7 terms of the series to estimate the number e.
Method 2

1/h 1
Use a calculator to fill in the given h e (1 + h) 1 h ln(1 + h)
table to estimate both (1 + h)1/h and 1.0
1 0.5
ln(1 + h) for small values of h. Your
h 0.1
results should show 0.01
e = lim (1 + h)1/h 0.001
h0 0.0001
1
and ln e = lim ln(1 + h) = 1 0.00001
h0 h
0.000001

x
1

1-31. If lim 1 + = e, then make appropriate substitutions and find the fol-
x x
lowing limits.
x

x
(a) lim 1 + (b) lim 1+
x x x x
where and are positive constants.

1-32. Use the  definition of a limit to prove that lim (4x + 2) = 14

x3
80

1-33. Sketch a graph of the following straight lines. State the slope of each line
and specify the x or y-intercept if it exists.
x y
(a) x=5 (e) + =1
3 4
(b) y=5 (f )
= { (x, y) | x = t + 2, y = 2t + 3 }
(c) y =x+1 (g) r cos( /4) = 2
(d) 3x + 4y + 5 = 0 (h) 3x + 4y = 0

1-34. For each line in the previous problem construct the perpendicular bisector
which passes through the origin.
x2 1

1-35. Consider the function y = f (x) = , for 2 x 2.
x1

(a) Is f (1) defined?

(b) Is the function continuous over the interval 2 x 2?
(c) Find lim f (x)
x1
(d) Can f (x) be made into a continuous function?
(e) Sketch the function f (x).

1-36. Assume xx
lim f (x) =
1 and lim g(x) =
2 . Use the  proof to show that
xx
0 0

lim [f (x) g(x)] =
1
2

xx0

1-37.
(a) Find the equation of the line with slope 2 which passes through the point (3, 4).
(b) Find the equation of the line perpendicular to the line in part (a) which passes
through the point (3, 4).

1-38. Find the following limits if the limit exists.


2 1
(a) lim x + f (x + h) f (x)
x1 x (d) lim , f (x) = x2

 h0 h
x 1 x2
(b) lim 2 + x (e) lim 2
x0 2 x2 x + x 6

x+h x x2
(c) lim  , x
= 0 (f ) lim 2
h0 h x(x + h) x2 x 4
 81

1-39. Find the following limits if the limit exists.
3x2 + 2x + 1 1
(a) lim (d) lim x sin
x x2 + 3x + 2 x0 x
sin x 7
(b) lim (e) lim
x x x4 x4
1 x3 1
(c) lim sin (f ) lim
x0 x x1 x 1

1-40. Show that

sin2 x



1 cos x sin x sin x
lim = lim = lim lim =0
x0 x x0 x(1 + cos x) x0 x x0 1 + cos x

1-41. Evaluate the following limits.

sin 4h sin(h/2)
(a) lim (d) lim
h0h h0h
sin2 (2h) tan x
(b) lim (e) lim
h0 h2 x0 x

1 cos h 1+x1
(c) lim (f ) lim
h0 h2 x0 x
f (x + h) f (x)

1-42. Evaluate the limit lim , if f (x) = x and x
= 0.
h0 h

1-43. Determine if the following limits exist. State why they exist or do not exist.
1 cos(mx)
(a) lim sin x (d) lim
x x sin(nx)
sin(mx)
(b) lim sin2 x
x0 sin(nx) (e) lim
x0 1 cos x
1 x sin x
(c) lim x sin( ) (f ) lim
x0 x x0 1 cos x

1-44. Given the line 3x + 4y + 5 = 0

(a) Find the slope of the line.
(b) Find the x and y-intercepts.
(c) Write the equation of the line in intercept form.
(d) Find the line perpendicular to the given line which passes through the point (0, 1).

1-45.
1
(a) Show that if p > 0, and r = , then r is such that 0 < r < 1.
1+p
1
(b) Write rn = and use the binomial theorem to show that if 0 < r < 1, then
(1 + p)n
lim r n = 0.
n
(c) Show that if p > 0, and x = (1 + p), then x is such that x > 1.
(d) Write xn = (1 + p)n and use the binomial theorem to show that if x > 1, then
lim xn =
n
82

1-46. Use the  method to prove that if xx lim f (x) = f (x0 ), then
0
lim cf (x) = c lim f (x) = cf (x0 ) where c is a constant.
xx0 xx0

1-47. The equation of a line passing through two points on a curve is called a
secant line.
(a) Given the parabola y = x2 find the equation of the secant line passing through
the points (1, 1) and (2, 4). Sketch a graph of the curve and the secant line.
(b) Find the equation of the secant line which passes through the points (1, 1) and
(3/2, 9/4). Sketch this secant line on your graph from part (a).
(c) Discuss how one can determine the equation of the tangent line to the curve
y = x2 at the point (1, 1).
(d) Can you find the equation of the tangent line to the curve y = x2 at the point
(1, 1)?

1-48. Sketch the given parabola and find (i) the focus (ii) the vertex (iii) the
directrix and (iv) the latus rectum.

(a) y 2 8y 8x + 40 = 0 (d) y 2 6y + 12x 3 = 0

(b) x2 = 12y (e) y 2 = 8x

(c) y 2 8y + 4x + 8 = 0 (f ) x2 6x + 12y 15 = 0

1-49. Sketch the given ellipse and find (i) the foci (ii) the directrices (iii) the latus
rectum and (iv) the eccentricity and (v) center.

(a) 4y 2 + 9x2 16y 18x 11 = 0 (d) 25y 2 + 16x2 150y 64x 689 = 0
x2 y2 x2 y2
(b) + =1 (e) + =1
25 9 4 9
(c) 16y 2 + 25x2 64y 150x 111 = 0 (f ) 4y 2 + 9x2 + 8y + 18x 23 = 0

1-50. Sketch the given hyperbola and find (i) the foci (ii) the vertices (iii) the
directrices (iv) the eccentricity and (v) the asymptotes.

(a) 9x2 4y 2 36x + 24y 36 = 0 (d) x2 4y 2 + 32y 2x 67 = 0

x2 y2 y2 x2
(b) =1 (e) =1
4 9 4 9
(c) 4y 2 9x2 16y + 54x 101 = 0 (f ) 4x2 y2 + 4y 24x + 28 = 0

1-51. Given the parabola y2 = 4x and the line y = x + b. What condition(s) must
be satisfied in order for the line to be a tangent line to the parabola?
 83

1-52. Examine the general equation of the second degree, given by equation (1.89).
When this equation is transformed using a rotation of axes there results the equation
(1.91) with coefficients defined by equation (1.92).
(a) Show that the quantity a + c is an invariant. That is, show a + c = a + c.
(b) Show that the discriminant is an invariant. That is, show b2 4ac = b2 4ac
Note that these two invariants are used as a check for numerical errors when one
performs the algebra involved in the rotation of axes.

1-53. Show that the equation xy = a2 , with a constant is a hyperbola.

1-54. Find the parabola symmetric about the x-axis which passes through the
points (1, 0), (0, 1) and (0, 1).

1-55.
y2 x2
(a) Sketch the hyperbola = 1 and label the y-intercepts, and the asymptotes.
4 9
(b) Find the equation of the conjugate hyperbola.

1-56. Normal form for equation of line

Let p = 0N > 0 denote the perpendicular distance
of the line from the origin. Assume the line 0N makes
an angle with the x-axis and let (x, y) denote a
variable point on the line. Write p = p1 + p2 , where
p1 = projection of x on ON and p2 = projection of y
on ON .
(a) Show p1 + p2 = p = x cos + y sin p = 0 which is called the normal form for
the equation of a line.
Ax + By + C
(b) Show that the line Ax + By + C = 0 has the normal form 2 2
= 0 where
A +B
the correct sign is selected so that p > 0.
(c) Find the normal form for the line 3x + 4y 5 = 0
(d) Show that the distance d of a point
 (x0 , y0 ) from
 the line described by
 Ax0 + By0 + C 
Ax + By + C = 0 is given by d =  
A2 + B 2 

1-57. The boiling point of water is 100 Celsius or 212
Fahrenheit and the freezing

point of water is 32 Fahrenheit and 0 Celsius. If there is a linear relationship
between degrees Celsius and degrees Fahrenheit, then find this relationship.
84

1-58. Graph the given equations by selecting such that r is well defined.

(a) r cos = 3 (b) r sin = 3 (c) r = 4 cos (d) r = 4 sin

1-59. Graph the given equations by selecting so that r is well defined.

3 3 6
(a) r= (b) r= (c) r=
1 cos 1 + cos 2 cos

1-60. Graph the following equations by selecting so that r is well defined.

3 3 3
(a) r= (b) r= (c) r=
1 sin 2 3 cos 2 cos

1-61. Find the horizontal and vertical asymptotes of the curve

xy 3x + 2y 10 = 0

1-62. Verify the following graphs by any method you wish.

Folium of Descartes Conchoid of Nicomedes with b > a

Witch of Agnesi
x3 + y 3 3axy = 0 (x a)2 (x2 + y 2 ) = b2x2
x2 y + 4a2 y 8a3 = 0

Limacon of Pascal with a = b Limnescate of Bernoulli

Trisectrix of Maclaurin
(x2 + y 2 2ax)2 b2 (x2 + y 2 ) (x2 + y 2 )2 = 2a2 (x2 y 2 )
y 2(a x) = x2 (3a + x)

Hints: Try substitutions r2 = x2 + y2 or x = r cos , y = r sin or let y = tx and try to

obtain a parametric representation of the given curves.
 85
Chapter 2
Differential Calculus
The history of mathematics presents the development of calculus as being ac-
credited to Sir Isaac Newton (1642-1727) an English physicist, mathematician and
Gottfried Wilhelm Leibnitz (1646-1716) a German physicist, mathematician. Por-
traits of these famous individuals are given in the figure 2-1. The introduction of
calculus created an explosion in the development of the physical sciences and other
areas of science as calculus provided a way of describing natural and physical laws
in a mathematical format which is easily understood. The development of calculus
also opened new areas of mathematics and science as individuals sought out new
ways to apply the techniques of calculus.

Figure 2-1. Joint developers of the calculus.

Calculus is the study of things that change and finding ways to represent these
changes in a mathematical way. The symbol will be used to represent change.
For example, the notation y is to be read The change in y .
Slope of Tangent Line to Curve
Consider a continuous smooth1 curve y = f (x), defined over a closed interval
defined by the set of points X = { x | x [a, b] }. Here x is the independent variable, y
1
A continuous smooth curve is an unbroken curve defined everywhere over the domain of definition of the
function and is a curve which has no sharp edges. If P is a point on the curve and
is the tangent line to the point
P , then a smooth curve is said to have a continuously turning tangent line as P moves along the curve.
86
is the dependent variable and the function can be represented graphically as a curve
defined by the set of points

{ (x, y) | x X, y = f (x) }

The slope of the curve at some given point P on the curve is defined to be the same
as the slope of the tangent line to the curve at the point P .

Figure 2-2. Secant line approaching tangent line as Q P

One can construct a tangent line to any point P on the curve as follows. On
the curve y = f (x) consider two neighboring points P and Q with coordinates (x, f (x))
and (x + x, f (x + x)), as illustrated in the figure 2-2. In this figure x represents
some small change in x and y = f (x + x) f (x) denotes the change in y in moving
from point P to Q. In the figure 2-2, the near points P and Q on the curve define a
straight line called a secant line of the curve. In the limit as the point Q approaches
the point P the quantity x tends toward zero and the secant line approaches the
tangent line. The quantity y also tends toward zero while the slope of the secant
line ms approaches the slope of the tangent line mt to the curve at the point P .
y
The slope of the secant line ms is given by ms = and the slope of the tangent
x
line mt at the point P with coordinates (x, f (x)) is given by the derived function f
(x)
calculated from the limiting process
dy y f (x + x) f (x)
mt = = f
(x) = lim = lim (2.1)
dx x0 x x0 x

if this limit exists.

 87
The Derivative of y = f (x)
The secant line through the near points P and Q approaches the tangent line in
y
the limit as x tends toward zero. The slope of the secant line ms = represents
x
the average change of the height of the curve y with respect to changes in x over
dy
the interval x. The derived function f
(x) = represents the slope of the tangent
dx
line mt at the point P and is obtained from equation (2.1) as a limiting process. The
derived function f
(x) is called the derivative of y = f (x) with respect to x and
represents the slope of the curve y = f (x) at the point (x, f (x)). It also represents
the instantaneous rate of change of y = f (x) with respect to x at the point (x, f (x)) on
the curve y = f (x). The derived function f
(x) or derivative represents the slope mt
of the tangent line constructed through the point (x, f (x)) on the curve. The limit,
defined by the equation (2.1), represents a process, called differentiation, for finding
dy
the derivative = f
(x). A function y = f (x), where the differentiation process
dx
dy
is successful, is called a differentiable function. The derived function = f
(x)
dx
obtained from the differentiation process is called the derivative function associated
with the given function y = f (x). By agreement, when referencing the derivative of
an explicit function y = f (x), it is understood to represent the ratio of changes of the
dependent variable y, with respect to changes of the independent variable x, as these
changes tend toward zero. There are alternative equivalent methods for calculating
the derivative of a function y = f (x). One alternative method is the following
dy y f (x + h) f (x) f () f (x)
= f
(x) = lim = lim = lim (2.2)
dx x0 x h0 h x x
if these limits exist. If y = f (x) for x X , then the domain of definition of the






derivative f (x) is the set X defined by X = { x | f (x) exists }. In general X X .

Example 2-1. Tangent line to curve

Given the parabola y = f (x) = 16 x2 . Find the tangent lines which touch this
curve at the points (2, 12) , (3, 7) and at a general point (x0 , y0 ) on the curve.
dy
Solution The derivative function = f
(x) associated with the parabolic function
dx
y = f (x) = 16 x2 represents the slope of the tangent line to the curve at the point
(x, f (x)) on the curve. The derivative function is calculated using the limiting process
defined by equation (2.1) or equation (2.2). One finds using the equation (2.1)
dy f (x + h) f (x) 16 (x + h)2 (16 x2 )
= f
(x) = lim = lim
dx h0 h h0 h
2 2 2
16 (x + 2xh + h ) (16 x )
= lim = lim (2x + h) = 2x
h0 h h0
or if one uses equation (2.2)
88
dy f () f (x) 16 2 (16 x2 ) (x )(x + )
= f
(x) = lim = lim = lim
dx x x x x x ( x)
= lim ( + x) = 2x
x

Here the derivative function is f
(x) = 2x and from the derivative function
the slope of the tangent line at (2, 12) is mt = f
(2) = 2(2) = 4
the slope of the tangent line at (3, 7) is mt = f
(3) = 2(3) = 6

Using the point-slope formula y y0 = m(x x0 ) for representing the equation of a

line, one finds
tangent line through point (2, 12) is y 12 = 4(x + 2)
tangent line through point (3, 7) is y 7 = 6(x 3)

dy
Knowing that a function y = f (x) has a derivative function = f
(x) which is
dx
defined and continuous for all values of the independent variable x (a, b) implies
that the given function y = f (x) is a continuous function for x (a, b). This is because
the tangent line to a point P on the curve is a continuous turning tangent line as
the point P moves along the curve. This is illustrated in the figure 2-3 where the
tangent line to the curve is continuously turning without any interruptions, the slope
moving continuously from a positive value, through zero to a negative value.

In general, at each point (x0 , y0 ), where

y0 = f (x0 ) = 16x20 , the slope of the curve at that
point is also the slope of the tangent line at that
point and this slope is given by f
(x0 ) = 2x0 .
The equation of the tangent line to the curve
y = f (x) = 16x2 which passes through the point
(x0 , y0 ) is given by the point-slope formula

y y0 = (2x0 )(x x0 ), where y0 = f (x0 ) = 16 x20

and x0 is some fixed abscissa value within the

Figure 2-3.
domain of definition of the function. The
Tangent lines to curve y = 16 x2
parabolic curve and tangent lines are illustrated
in the figure 2-3.
 89
Also note that for x < 0, the slope of the curve is positive and indicates that as x
increases, y increases. For x > 0, the slope of the curve is negative and indicates that
as x increases, y decreases. If the derivative function changes continuously from a
positive value to a negative value, then it must pass through zero. Here the zero
slope occurs where the function y = f (x) = 16 x2 has a maximum value.

dy d
Example 2-2. If y = f (x) = sin x, then show = f
(x) = sin x = cos x
dx dx
dy sin(x + x) sin x
Solution By definition = f
(x) = lim . Use the results
from
dx x0 x
example 1-7 together with the trigonometric identity for the difference of two sine
functions to obtain
dy x sin( x
2 )
= f
(x) = lim cos(x + ) lim = cos x
dx x0 2 x0 x 2

dy d
Example 2-3. If y = g(x) = cos x, then show = g
(x) = cos x = sin x
dx dx
dy cos(x + x) cos x
Solution By definition = g
(x) = lim . Use the results
from
dx x0 x
example 1-7 together with the trigonometric identity for the difference of two cosine
functions to obtain
dy x sin( x )
= g
(x) = lim sin(x + ) lim x
2
= sin x
dx x0 2 x0 2

Right and Left-hand Derivatives

If a function y = f (x) has a jump
discontinuity at the point x = x0 , then
one can define the right-hand derivative
of f (x) at the point x = x0 as the following
limit
f (x+ +
0 + h) f (x0 )
f
(x+
0 ) = lim , h>0
h0+ h
if this limit exists.

The left-hand derivative of f (x) at the point x = x0 is defined as the limit

f (x
0 + h) f (x0 )
f
(x
0 ) = lim , where h is restricted such that h < 0
h0 h
if this limit exists. If the left-hand derivative is different from the right-hand deriva-
tive, then there exists a tangent line through the point (x+ +
0 , f (x0 )) and a different

tangent line through the point (x

0 , f (x0 )). If the left-hand derivative equals the

right-hand derivative then f (x) is said to have a derivative at the point x = x0 .

90
Alternative Notations for the Derivative
Some of the notations used to represent the derivative of a function y = f (x) are
dy d d df
= y = f
(x) = y
= f (x) = = Df (x) = Dx f (x) (2.3)
dx dx dx dx
d d d
where ,D= and Dx = are called differentiation operators. One can think
dx dx dx
d
of as a derivative operator which operates upon a given function to produce the
dx
y
differences x and y and then evaluates the limit lim to produce the derivative
x0 x
function.

Figure 2-4. Differentiation performed by an operator box.

The figure 2-4 illustrates an operator box where functions that enter the operator
box get operated upon using the differentiation process defined by equation (2.1)
and the output from the box represents a derivative of the input function. Observe
that a derivative of a derivative function is called a second derivative function. In
dy
general, a derivative, such as = f
(x), is a measure of the instantaneous rate of
dx
dy
change of y = f (x) with respect to a change in x. The notation for the derivative
dx


was introduced by Gottfried Leibnitz. The prime notation f (x) for the derivative
of a function f (x) was introduced by Joseph-Louis Lagrange2. If y = y(t), Sir Isaac
Newton used the dot notation y , y, . . . for representing the first, second and higher
derivatives. The operator notation Dx y was introduced by Leonhard Euler3.
Higher Derivatives
If the derivative function f
(x) is the input to the operator box illustrated in
the figure 2-4, then the output function is denoted f

(x) and represents a derivative
of a derivative called a second

derivative
 . Higher ordered derivatives are defined in
d dn1 y dn y
a similar fashion with = which states that the derivative of the
dx dxn1 dxn

2
Joseph-Louis Lagrange (1736-1813) an Italian born French mathematician.
3
Leonhard Euler (1707-1783) A Swiss mathematician.
 91
dy
(n 1)st derivative is the nth derivative. The function f
(x) = is called a first
dx
d2 y d3 y
derivative, f

(x) = is called a second derivative, f


(x) = is called a third
dx2 n dx3
d y
derivative,. . . f (n) (x) = n is called a n-th derivative. Other notations for higher
dx
ordered derivatives are as follows.
The first derivative of y = f (x) is denoted
dy
= f
(x) or Dx y or Dy or y

dx
The second derivative of y = f (x) is denoted
d2 y
 
d dy
2
= = f

(x) or Dx2 y or D2 y or y


dx dx dx
The third derivative of y = f (x) is denoted
d3 y d2 y
 
d
= = f


(x) or Dx3 y or D3 y or y



dx3 dx dx2
The n-th derivative of y = f (x) is denoted


dn y d dn1 y
= = f (n) (x) or Dxn y or Dn y or y (n)
dxn dx dxn1

Rules and Properties

The following sections cover fundamental material associated with the introduc-
tion of different kinds of functions and developing techniques to find the derivatives
associated with these functions. The following list contains fundamental rules and
properties associated with the differentiation of sums, products and quotients of
functions. These fundamental properties should be memorized and recognized in
applications. Note that most proofs of a differentiation property use one of the pre-
vious definitions of differentiation given above and so the student should memorize
the definitions of a derivative as given by equations (2.2).
d
The derivative of a constant C is zero or C = 0. That is, if
dx
dy d
y = f (x) = C = constant, then = f
(x) = C =0
dx dx

Proof
Sketch the curve y = f (x) = C = constant and observe that it has a zero slope
everywhere. The derivative function represents the slope of the curve y = C at the
92
d
point with abscissa x and consequently C = 0 for all values of x since the slope is
dx
zero at every point on the curve and the height of the curve is not changing. Using
the definition of a derivative one finds
dy f (x + h) f (x) CC 0
= f
(x) = lim = lim = lim = 0
dx h0 h h0 h h0 h

The converse statement that if f
(x) = 0 for all values of x, then y = f (x) = C is a
constant also holds and will be proven later in this chapter.
dy d
The derivative of the function y = f (x) = x, is = f
(x) = x=1
dx dx
Proof
Sketch the curve y = f (x) = x and observe that it is a line which passes through
the origin making an angle of /4 with the x-axis. The slope of this line is given by
d
m = tan = 1 for all values of x. Consequently, f
(x) = x = 1 for all values of x
4 dx
since the derivative function represents the slope of the curve at the point x. Using
the definition of a derivative one finds
dy d f (x + h) f (x) x+hx
= f
(x) = x = lim = lim =1
dx dx h0 h h0 h

for all values of x.

The derivative of the function y = f (x) = xn , where n is a nonzero integer, is
dy d n
given by = f
(x) = x = n xn1 or in words one can say the derivative of x to
dx dx
an integer power n equals the power n times x to the (n 1)st power.
Proof
Using the limiting process which defines a derivative one finds
dy d n f (x + h) f (x) (x + h)n xn
= f
(x) = x = lim = lim
dx dx h0 h h0 h

One can employ the binomial theorem to expand the numerator and obtain
n(n1)
dy d xn + nxn1 h + 2! xn2 h2 + + hn xn
= f
(x) = xn = lim
dx dx h0 h
 
n(n 1) n2
= lim nxn1 + x h + + hn1
h0 2!
= nxn1

d
Consequently, one can write xn = nxn1 where n is an integer.
dx
 93
d
Later it will be demonstrated that xr = rxr1 for all real numbers r which
dx
are different from zero.
The derivative of a constant times a function equals the constant times the deriva-
tive of the function or
d d
[Cf (x)] = C f (x) = Cf
(x)
dx dx
Proof
Use the definition of a derivative applied to the function g(x) = Cf (x) and show
that
d g(x + h) g(x) Cf (x + h) Cf (x)
g(x) = lim = lim
dx h0 h h0 h
 
f (x + h) f (x)
= lim C
h0 h
It is known that the limit of a constant times a function is the constant times the
limit of the function and so one can write
d f (x + h) f (x)
g(x) = C lim = Cf
(x)
dx h0 h
or
d d
[Cf (x)] = C f (x) = Cf
(x)
dx dx
The derivative of a sum is the sum of the derivatives or
d d d du dv
[u(x) + v(x)] = u(x) + v(x) = + = u
(x) + v
(x) (2.4)
dx dx dx dx dx
This result can be extended to include n-functions
d d d d
[u1 (x) + u2 (x) + + un (x)] = u1 (x) + u2 (x) + + un (x)
dx dx dx dx

Proof
If y(x) = u(x) + v(x), then
dy y(x + h) y(x)
= lim
dx h0 h
u(x + h) + v(x + h) [u(x) + v(x)]
= lim
h0 h
   
u(x + h) u(x) v(x + h) v(x)
= lim + lim
h0 h h0 h
or
dy d d d
= [u(x) + v(x)] = u(x) + v(x) = u
(x) + v
(x)
dx dx dx dx
This result follows from the limit property that the limit of a sum is the sum of
the limits. The above proof can be extended to larger sums by breaking the larger
sums into smaller groups of summing two functions.
94
Example 2-4. The above properties are combined into the following examples.
(a) If y = F (x) is a function which is differentiable and C is a nonzero constant, then
d dF (x) d  3 d
[CF (x)] =C 5x =5 x3 = 5(3x2 ) = 15x2
dx dx dx dx
d dF (x) d dF (x) d  3  d 3 d
[F (x) + C] = + C= x +8 = x + 8 = 3x2
dx dx dx dx dx dx dx
since the derivative of a constant times a function equals the constant times
the derivative of the function and the derivative of a sum is the sum of the
derivatives.
(b) If S = {f1 (x), f2 (x), f3 (x), . . ., fn (x), . . . } is a set of functions, define the set of deriva-
dS df1 df2 df3 dfn
tives ={ , , , . . ., , . . . }. To find the derivatives of each of the func-
dx dx dx dx dx
tions in the set S = {1, x, x , x , x4 , x5 , . . . , x100 , . . . , xm, . . . }, where m is a very large
2 3

integer, one can use properties 1 , 2 and 3 above to write

dS
= {0, 1, 2x, 3x2, 4x3 , 5x4, . . . , 100x99 , . . . , mxm1 , . . . }
dx
(c) Consider the polynomial function y = x6 + 7x4 + 32x2 17x + 33. To find the
derivative of this function one can combine the properties 1, 2, 3, 4 to show
dy d 6
= (x + 7x4 + 32x2 17x + 33)
dx dx
dy d d d d d
= x6 + 7 x4 + 32 x2 17 x + (33)
dx dx dx dx dx dx
dy
=6x5 + 7(4x3 ) + 32(2x) 17(1) + 0
dx
dy
=6x5 + 28x3 + 64x 17
dx
This result follows from use of the properties (i) the derivative of a sum is the
sum of the derivatives (ii) the derivative of a constant times a function is that
constant times the derivative of the function (iii) the derivative of x to a power
is the power times x to the one less power and (iv) the derivative of a constant
is zero.
(d) To find the derivative of a polynomial function

y = pn (x) = a0 xn + a1 xn1 + a2 xn2 + + an2 x2 + an1 x + an , (2.5)

where a0 , a1 , . . . , an are constants,with a0
= 0, one can use the first four proper-
ties above to show that by differentiating each term one obtains the derivative
function
dy
= a0 nxn1 + a1 (n 1)xn2 + a2 (n 2)xn3 + an2 [2x] + an1 [1] + 0
     
dx
 95

(e) The polynomial function pn (x) of degree n given by equation (2.5) is a linear
combination of terms involving x to a power. The first term a0 xn , with a0
= 0,
being the term containing the largest power of x. Make note of the higher
derivatives associated with the function xn . These derivatives are
d n
(x ) =nxn1
dx
d2 n
(x ) =n(n 1)xn2
dx2
d3 n
(x ) =n(n 1)(n 2)xn3
dx3
.. ..
. .
dn n
(x ) =n(n 1)(n 2) (3)(2)(1)x0 = n! Read n-factorial.
dxn
dn+1 n
(x ) =0
dxn+1
This result demonstrates that the (n+1)st and higher derivatives of a polynomial
of degree n will all be zero.
(f ) One can readily verify the following derivatives

d3 3 d5 5
(x ) =3! = 3 2 1 = 6 (x ) =5! = 5 4 3 2 1 = 120
dx3 x5
d4 3 d6 5
(x ) =0 (x ) =0
dx4 dx6
The derivative of a product of two functions is the first function times the deriva-
tive of the second function plus the second function times the derivative of the first
function or
d dv du
[u(x)v(x)] =u(x) + v(x) = u(x)v
(x) + v(x)u
(x)
dx dx  dx
(2.6)
u
(x) v
(x)

d
or [u(x)v(x)] =u(x)v(x) +
dx u(x) v(x)
Proof
Use the properties of limits along with the definition of a derivative to show that if
y(x) = u(x)v(x), then
dy y(x + h) y(x)
= lim
dx h0 h
u(x + h)v(x + h) u(x)v(x)
= lim
h0 h
u(x + h)v(x + h) u(x)v(x + h) + u(x)v(x + h) u(x)v(x)
= lim
h0 h
96
Where the term u(x)v(x + h) has been added and subtracted to the numerator.
Now rearrange terms and use the limit properties to write
   
dy u(x + h) u(x) v(x + h) v(x)
= lim v(x + h) + lim u(x)
dx h0 h h0 h
   
dy v(x + h) v(x) u(x + h) u(x)
= lim u(x) lim + lim v(x + h) lim
dx h0 h0 h h0 h0 h
or
dy d dv du
= [u(x)v(x)] = u(x) + v(x) = u(x)v
(x) + v(x)u
(x)
dx dx dx dx
The result given by equation (2.6) is known as the product rule for differentiation.

Example 2-5.
(a) To find the derivative of the function y = (3x2 + 2x + 1)(8x+3) one should recognize
the function is defined as a product of polynomial functions and consequently
the derivative is given by
dy d  2 
= (3x + 2x + 1)(8x + 3)
dx dx
dy d d
=(3x2 + 2x + 1) (8x + 3) + (8x + 3) (3x2 + 2x + 1)
dx dx dx
dy
=(3x2 + 2x + 1)(8) + (8x + 3)(6x + 2)
dx
dy
=72x2 + 50x + 14
dx
(b) The second derivative is by definition a derivative of the first derivative so that
differentiating the result in part(a) gives
d2 y d dy d
72x2 + 50x + 14 = 144x + 50

2
= =
dx dx dx dx
Similarly, the third derivative is
d3 y d d2 y d
= = (144x + 50) = 144
dx3 dx dx2 dx
and the fourth derivative and higher derivatives are all zero.

Example 2-6.
Consider the problem of differentiating the function y = u(x)v(x)w(x) which is a
product of three functions. To differentiate this function one can apply the product
rule to the function y = [u(x)v(x)] w(x) to obtain
 97
dy d dw(x) d
= ([u(x)v(x)] w(x)) = [u(x)v(x)] + w(x) [u(x)v(x)]
dx dx dx dx
Applying the product rule to the last term one finds
dy d dw(x) dv(x) du(x)
= [u(x)v(x)w(x)] = u(x)v(x) + u(x) v(x)w(x) w(x) +
dx dx dx dx dx
dy d
= [u(x)v(x)w(x)] = u(x)v(x)w
(x) + u(x)v
(x)w(x) + u
(x)v(x)w(x)
dx dx

A generalization of the above procedure produces the generalized product rule for
differentiating a product of n-functions
d dun (x)
[u1 (x)u2 (x)u3 (x) un1 (x)un (x)] =u1 (x)u2 (x)u3 (x) un1 (x)
dx dx
dun1 (x)
+u1 (x)u2 (x)u3 (x) un (x)
dx
+
du3 (x)
+u1 (x)u2 (x) un1 (x)un (x)
dx
du2 (x)
+u1 (x) u3 (x) un1 (x)un (x)
dx
du1 (x)
+ u2 (x)u3 (x) un1 (x)un (x)
dx

This result can also be expressed in the form

u
1 u
2 u
3 u
n
 
d
[u1 u2 u3 un1 un ] = u1 u2 u3 un + + + +
dx u1 u2 u3 un

and is obtained by a repeated application of the original product rule for two
functions.

The derivative of a quotient of two functions is the denominator times the deriva-
tive of the numerator minus the numerator times the derivative of the denominator
all divided by the denominator squared or

du dv
v(x) u(x)
dx dx v(x)u
(x) u(x)v
(x)
 
d u(x)
= = (2.7)
dx v(x) v 2(x) v 2(x)
98
Proof
u(x)
Let y(x) = and write
v(x)

u(x + h) u(x)

dy y(x + h) y(x) v(x + h) v(x)
= lim = lim
dx h0 h h0 h
v(x)u(x + h) u(x)v(x) + u(x)v(x) u(x)v(x + h)
v(x + h)v(x)
= lim
h0 h
   
u(x + h) u(x) v(x + h) v(x)
v(x) u(x)
h h
= lim
h0 v(x + h)v(x)
   
u(x + h) u(x) v(x + h) v(x)
v(x) lim u(x) lim
h0 h h0 h
=
lim v(x + h)v(x)
h0

or
v(x)u
(x) u(x)v
(x)
 
dy d u(x)
= = , where v 2(x) = [v(x)]2
dx dx v(x) v 2(x)
This result is known as the quotient rule for differentiation.
A special case of the above result is the differentiation formula
 
d 1 d 1 1 dv 1
v
(x)

v(x) = = 2
= (2.8)
dx dx v(x) [v(x)] dx [v(x)]2

3x2 + 8 dy
Example 2-7. If y = , then find
x3 x2 + x dx
Solution
Using the derivative of a quotient property one finds
d d
dy d

3x2 + 8
 (x3 x2 + x) dx (3x2 + 8) (3x2 + 8) dx (x3 x2 + x)
= =
dx dx x3 x2 + x (x3 x2 + x)2
(x3 x2 + x)(6 x) (3x2 + 8)(3x2 2x + 1) 3x4 21x2 + 16x 8
= =
(x3 x2 + x)2 (x3 x2 + x)2

Differentiation of a Composite Function

If y = y(u) is a function of u and u = u(x) is a function of x, then the derivative
of y with respect to x equals the derivative of y with respect to u times the derivative
of u with respect to x or
dy d dy du
= y(u) = = y
(u)u
(x) (2.9)
dx dx du dx
 99

This is known as the composite function rule for differentiation or the chain rule
for differentiation. Note that the prime notation
always denotes differentiation with
dz
respect to the argument of the function. For example z
() = .
d
Proof
If y = y(u) is a function of u and u = u(x) is a function of x, then make note of
the fact that if x changes to x + x, then u changes to u + u and u 0 as x 0.
Hence, if u
= 0, one can use the identity
y y u
=
x u x
together with the limit theorem for products of functions, to obtain
     
dy y y u dy du
= lim = lim lim = = y
(u)u
(x)
dx x0 x u0 u x0 x du dx

which is known as the chain rule for differentiation.

An alternative derivation of this rule makes use of the definition of a derivative
given by equation (2.2). If y = y(u) is a function of u and u = u(x) is a function of x,
then one can write
dy y(x + h) y(x) y(u(x + h)) y(u(x))
= lim = lim (2.10)
dx h0 h h0 h
In equation (2.10) make the substitutions u = u(x) and = u(x+h) and write equation
(2.10) in the form  
dy y() y(u) ( u)
= lim
dx u ( u) h
dy y() y(u) u(x + h) u(x)
= lim lim
dx u u h0 h
dy dy du
= = y
(u)u
(x)
dx du dx
Here the chain rule is used to differentiate a function of a function. For example, if
y = f (g(x)) is a function of a function, then make the substitution u = g(x) and write
y = f (u), then by the chain rule
dy dy du
= = f
(u)u
(x) = f
(u)g
(x) (2.11)
dx du dx
The derivative of a function u = u(x) raised to a power n, n an integer, equals the
power times the function to the one less power times the derivative of the function or
d du
[u(x)n ] = n u(x)n1 = n u(x)n1 u
(x) (2.12)
dx dx
100
This result is known as the power rule for differentiation.
Proof
This is a special case of the previous property. If y = y(u) = un is a function of
u and u = u(x) is a function of x, then differentiation of these functions with respect
to their independent variables gives the derivatives
dy d n du d
= u = nun1 and = u(x) = u
(x)
du du dx dx
Using the chain rule for differentiation one finds
dy d dy du du
= u(x)n = = nun1 = nun1 u
(x)
dx dx du dx dx
or
d du
u(x)n = nu(x)n1 = nu(x)n1 u
(x)
dx dx

The general power rule for differentiation is

d du
u(x)r = ru(x)r1 (2.13)
dx dx
where r is any real number. Here it is understood that for the derivative to exist,
then u(x)
= 0 and the function u(x)r is well defined everywhere. A proof of the general
power rule is given later in this chapter.
dy 
Example 2-8. Find the derivative of the function y = 3 x2 + x
dx
Solution Let u = x2 + x and write y = u1/3 . These functions have the derivatives
du d 2 dy d 1/3 1
= (x + x) = 2x + 1 and = u = u2/3
dx dx du du 3
By the chain rule for differentiation
dy dy du 1 2x + 1
= = 2/3 (2x + 1) =
dx du dx 3u 3(x2 + x)2/3
Using the general power rule one can write
dy d 2 1 2x + 1
= (x + x)1/3 = (x2 + x)2/3 (2x + 1) =
dx dx 3 3(x2 + x)2/3

3
x2 1

dy
Example 2-9. Find the derivative of the function y =
dx x4 + 1
x2 1
Solution Let u = and write y = u3 so that by the chain rule for differentiation
x4 + 1
dy dy du dy d 3
one has = where = u = 3u2 and
dx du dx du du
 101
x2 1 (x4 + 1)(2x) (x2 1)(4x3) 2x5 4x3 + 2x
 
du d
= = =
dx dx x4 + 1 (x4 + 1)2 (x4 + 1)2
This gives the final result
 2 2 
2x5 4x3 + 2x

dy dy du 2 du x 1
= = 3u =3
dx du dx dx x4 + 1 (x4 + 1)2
dy 6(x 4x5 + 4x7 x9 )
=
dx (x4 + 1)4

Differentials
f (x + x) f (x)
If y = f (x) is differentiable, then the limit lim = f
(x) exists.
x0 x
The quantity x is called the increment given to x and y = f (x + x) f (x)
is called the increment in y = f (x) corresponding to the increment in x. Since the
derivative is determined by a limiting process, then one can define dx = x as the
differential of x and write

y = f (x + x) f (x) = f
(x)x + x = f
(x) dx +  dx (2.14)

where  0 as x 0. Define the quantity dy = f
(x)dx as the differential of y

which represents the principal part of the change in y as x 0. Note that the
differential dy does not equal y because dy is only an approximation to the actual
change in y . Using the above definitions one can write

dy f (x + x) f (x) y
= f
(x) = lim = lim
dx x0 x x0 x

Here dx = x, but dy is not y because y = dy +  dx, where  0 as x 0.

Using the definition dy = f
(x) dx one can verify the following differentials

df (x) =f
(x) dx d(uv) =u dv + v du

 u v du u dv
dC =0 C is a constant d =
v v2
d(Cu) =C du df
df = du
du
d(u + v) =du + dv d(u ) =nun1 du
n
102
Differentiation of Implicit Functions
A function defined by an equation of the form F (x, y) = 0, where one of the
variables x or y is not explicitly solved for in terms of the other variable, then one
says that y is defined as an implicit function of x. For example, the equation of the
circle given by
F (x, y) = x2 + y 2 2 = 0, is a constant

is an example of an implicit function, where a dependent variable has not been

defined explicitly in terms of an independent variable. In general, when given an
implicit function F (x, y) = 0, there are times where it is possible to solve for one
variable in terms of another and thereby convert the implicit form into an explicit
form for representing the function. Note also that there are times where the implicit
functions F (x, y) = 0 cannot be converted into an explicit form. Given an implicit
function F (x, y) = 0, where it is not possible to solve for y in terms of x, it is still
dy
possible to calculate the derivative by treating the function F (x, y) = 0 as a function
dx
F (x, y(x)) = 0, where it is to be understood, that theoretically the implicit function
defines y as a function of x.One can then differentiate every part of the implicit
function with respect to x and then solve the resulting equation for the derivative
dy
term .
dx

Example 2-10. Given the implicit function F (x, y) = x3 + xy2 + y3 = 0, find the
dy
derivative .
dx
Solution
Differentiate each term of the given implicit function with respect to x to obtain
d 3 d d 3 d
(x ) + (xy 2 ) + (y ) = 0 (2.15)
dx dx dx dx

The derivative of the first term in equation (2.15) represents the derivative of x to
a power. The second term in equation (2.15) represents the derivative of a product
of two functions (the function x times the function y2 (x)). The third term in equa-
tion (2.15) represents the derivative of a function to a power (the function y3 (x)).
Remember, that when dealing with implicit functions, it is understood that y is to
be treated as a function of x. Calculate the derivatives in equation (2.15) using the
product rule and general power rule and show there results
 103
d 2 d dy
3x2 + x (y ) + y 2 (x) + 3y 2 =0
dx dx dx
dy dy
3x2 + x 2y + y 2 (1) + 3y 2 =0
dx dx
dy
(3x2 + y 2 ) + (2xy + 3y 2 ) =0
dx
Solving this last equation for the derivative term gives
dy (3x2 + y 2 )
=
dx 2xy + 3y 2

Make note that once the derivative is solved for, then the form for representing the
derivative can be changed by using some algebra along with the given original implicit
form y 3 = x3 xy 2 . For example, one can write
dy (3x2 + y 2 ) 3x2 y y 3 3x2y (xy 2 x3 ) 3xy x2 y 2
= = = =
dx 2xy + 3y 2 2xy 2 + 3y 3 2xy 2 + 3(xy 2 x3 ) 3x2 + y 2

An alternative method to solve the above problem is to use differentials and find
the differential of each term to obtain

3x2 dx + x 2y dy + dx y 2 + 3y 2 dy = 0

Divide each term by dx and combine like terms to obtain

dy
(2xy + 3y 2 ) = (3x2 + y 2 )
dx
dy
and solving for dx obtain the same result as above.

Example 2-11.
Find the equation of the tangent line to the circle x2 + 2x + y2 6y 15 = 0 which
passes through the point (2, 7).
Solution
The given equation is an implicit equation defining the circle. By completing
the square on the x and y terms one can convert this equation to the form

(x2 + 2x + 1) + (y 2 6y + 9) = 15 + 10 = (x + 1)2 + (y 3)2 = 25

which represents a circle centered at the point (1, 3) with radius 5.

104
This circle is illustrated in the figure 2-5. One
can verify that the point (2, 7) is on the circle by
substituting the values x = 2 and y = 7 into the
given equation to show that these values do indeed
satisfy the equation. Use implicit differentiation
and show
dy dy dy
Figure 2-5. 2x + 2 + 2y 6 =0 or (2y 6) = (2x + 2)
dx dx dx
Circle centered at (1, 3) dy (x + 1)
with radius 5. and so the derivative is given by =
dx y 3
This derivative represents the slope of the circle at a point (x0 , y0 ) on the circle
which is the same as the slope of the tangent line to the point (x0 , y0 ) on the circle.
Therefore, the slope of the tangent line to the circle at the point (2, 7) is obtained by
dy
evaluating the derivative at this point. The notation used to denote a derivative
dx
dy
being evaluated at a point (x0 , y0 ) is . For example, one can say the slope of
dx (x0 ,y0 )
the tangent line to the circle at the point (2, 7) is given by
dy (x + 1) (2 + 1) 3
mt = = = =
dx (2,7) (y 3) (2,7) (7 3) 4

The equation of the tangent line to the circle which passed through the point (2, 7)
is obtained from the point-slope formula y y0 = mt (x x0 ) for the equation of a line.
One finds the equation of the tangent line which passes through the point (2, 7) on
the circle is given by
y 7 = (3/4)(x 2)

Example 2-12.
(a) Consider two lines
1 and
2 which intersect
to form supplementary angles and as illus-
trated in the figure 2-6. Let equal the coun-
terclockwise angle from line
1 to line
2 . One
could define either angle or as the angle
of intersection between the two lines. To avoid
confusion as to which angle to use, define the Figure 2-6.
point of intersection of the two lines as a point Intersection of two lines.
of rotation.
 105
One can think of line
1 as being rotated about this point to coincide with the
line
2 or line
2 as being rotated to coincide with line
1 . The smaller angle of
rotation, either counterclockwise or clockwise, is defined as the angle of intersection
between the two lines.
Assume the lines
1 and
2 have slopes m1 = tan 1 and m2 = tan 2 which are well
defined. We know the exterior angle of a triangle must equal the sum of the two
opposite interior angles so one can write = 2 1 and consequently,
tan 2 tan 1 m2 m1
tan = tan(2 1) = = (2.16)
1 + tan 1 tan 2 1 + m1 m2

Here denotes the counterclockwise angle from line
1 to line
2 . If the lines are
perpendicular, then they are said to intersect orthogonally. In this case the for-
mula given by equation (2.16) becomes meaningless because when the lines intersect
orthogonally then the slopes satisfy m1 m2 = 1.
(b) If two curves C1 and C2 intersect at a point
P , the angle of intersection of the two curves is
defined as the angle of intersection of the tan-
gent lines to the curves C1 and C2 at the in-
tersection point P . Two curves are said to in-
Figure 2-7.
tersect orthogonally when their intersection is
such that the tangent lines at the point of in- Intersection of two curves.
tersection form right angles.
(c) Find the angle of intersection between the circles
x2 + 2x + y 2 4y = 0 and x2 4x + y 2 6y + 8 = 0
Solution First find the points where the two cir-
cles intersect. Eliminating the terms x2 and y2
by subtracting the equations of the circle shows
that the two circles must intersect at points
which lie on the line y = 4 3x. Substitute
this value for y into either of the equations for
the circle and eliminate y to obtain a quadratic
equation in x and show the points of intersec-
Figure 2-8. tion are (0, 4) and (1, 1). As a check, show that
these values satisfy both the given equations.
Intersection of two circles.
106
To find the slopes of the tangent lines at these two points of intersection, use
implicit differentiation to differentiate the given equations for the circles. These
differentiations produce the following equations.
x2 + 2x + y 2 4y =0 x2 4x + y 2 6y + 8 =0
d 2 d d 2 d
(x + 2x + y 2 4y) = (0) (x 4x + y 2 6y + 8) = (0)
dx dx dx dx
dy dy dy dy
2x + 2 + 2y 4 =0 2x 4 + 2y 6 =0
dx dx dx dx
dy (2x + 2) dy (2x 4)
= =
dx (2y 4) dx (2y 6)

The slopes of the tangent lines at the point (0, 4) are given by
dy (2x + 2) 1
For the first circle m1 = = =
dx (0,4) (2y 4) (0,4) 2
dy (2x 4)
and for the second circle m2 = = =2
dx (0,4) (2y 6) (0,4)

This gives the equations of the tangent lines to the point (0, 4) as y 4 = (1/2)x and
y 4 = 2x. Note that the product of the slopes gives m1 m2 = 1 indicating the curves
intersect orthogonally.
Similarly, the slopes of the tangent lines at the point (1, 1) are given by
dy (2x + 2)
m1 = = =2
dx (1,1) 2y 4) (1,1)
dy (2x 4) 1
and m2 = = =
dx (1,1) (2y 6) 1,1) 2

This gives the equations of the tangent lines to the point (1, 1) as y 1 = 2(x 1) and
y 1 = (1/2)(x 1). The product of the slopes gives m1 m2 = 1 indicating the curves
intersect orthogonally. The situation is illustrated in the figure 2-8.

Importance of Tangent Line and Derivative Function f
(x)

Given a curve C defined by the set of points { (x, y) | y = f (x), a x b } where
y = f (x) is a differentiable function for a < x < b. The following is a short list of
things that can be said about the curve C , the function f (x) defining the curve and
the derivative function f
(x) associated with f (x).
1. If P is a point on the curve C , having coordinates (x0 , f (x0 )), the slope of the curve
at the point P is given by f
(x0 ), where f
(x) is the derivative function associated
with the function rule y = f (x) defining the ordinate of the curve.
 107
2. The tangent line to the curve C at the point P is given by the point-slope formula

y f (x0 ) = f
(x0 )(x x0 ) the point (x0 , f (x0 )) is a fixed point on the curve.

3. The curve C is called a smooth curve over the interval a x b if it has a

continuously turning tangent line as the point P moves from (a, f (a) to (b, f (b)).
4. If the continuously turning tangent line suddenly changes at a point (x1 , f (x1 )),
then the derivative function f
(x) is said to have a jump discontinuity at the point
x = x1 . See figure 2-9(a).

Figure 2-9. Analysis of the derivative function f
(x).

5. If as x increases, the tangent line to the curve continuously changes from a positive
slope to a zero slope followed by a negative slope, the curve C is said to have a
local maximum or relative maximum at the point where the slope is zero. If this
local maximum occurs at the point (x1 , f (x1 )),then f (x1 ) f (x) for all points x
near x1 . Similarly, if as x increases the tangent line continuously changes from a
negative slope to a zero slope followed by a positive slope, the curve C is said to
have a local minimum or relative minimum at the point where the slope is zero. If
the local minimum occurs at the point (x2 , f (x2 )), then f (x2 ) f (x) for all points
x near x2 . See figure 2-9(b)
6. If f
(x) > 0 for all values of x as x moves from a to b, the continuously turning
tangent line always has a positive slope which indicates that the function y = f (x)
is an increasing function of x over the interval (a, b). Functions with this property
are called monotone increasing functions. See figure 2-9(c).
7. If f
(x) < 0 for all values of x as x moves from a to b, the continuously turning
tangent line always has a negative slope which indicates that the function y = f (x)
is a decreasing function of x over the interval (a, b). Functions with this property
are called monotone decreasing functions. See figure 2-9(d)
108
Rolles Theorem4
If y = f (x) is a function satisfying (i) it is continuous
for all x [a, b] (ii) it is differentiable for all x (a, b) and
(iii) f (a) = f (b), then there exists a number c (a, b) such
that f
(c) = 0.

This result is known as Rolles theorem. If y = f (x) is a

constant, the theorem is true so assume y = f (x) is different
from a constant. If the slope f
(x) is always positive or
always negative for a x b, then f (x) would be either
continuously increasing or continuously decreasing between
the endpoints x = a and x = b and so it would be impossible
for y = f (x) to have the same value at both endpoints.
This implies that in order for f (a) = f (b) the derivative function f
(x) must change
sign as x moves from a to b. If the derivative function changes sign it must pass
through zero and so one can say there exists at least one number x = c where f
(c) = 0.

The Mean-Value Theorem

If y = f (x) is a continuous function for x [a, b] and is differentiable so that
f
(x) exists for x (a, b), then there exists at least one number x = c (a, b) such that
the slope mt of the tangent line at (c, f (c)) is the same as the slope ms of the secant
line passing through the points (a, f (a)) and (b, f (b)) or
f (b) f (a)
mt = f
(c) = = ms a<c<b
ba

This result is known as the mean-value theorem and its implications are illustrated
in the figure 2-10.
Proof
A sketch showing the secant line and tangent line having the same slope is given
in the figure 2-10. In this figure note the secant line passing through the points
(a, f (a)) and (b, f (b)) and verify that the equation of this secant line is given by the
point-slope formula

f (b) f (a)
y f (a) = (x a)
ba
4
Michel Rolle (1652-1719) A French mathematician. His name is pronounced Roll .
 109
Also construct the vertical line x = , where a < < b. This line intersects the curve
y = f (x) at the point P with coordinates (, f ()) and it intersects the secant line at
 
f (b)f (a)
point Q with coordinates (, f (a) + ba
( a)). Denote the distance from Q to
P as h() and verify that
 
f (b) f (a)
h() = f () f (a) ( a) (2.17)
ba

Figure 2-10. Construction of secant line to curve y = f (x)

Note that h() varies with and satisfies h(a) = h(b) = 0. The function h() satisfies
all the conditions of Rolles theorem so one can say there exists at least one point
x = c where h
(c) = 0. Differentiate the equation (2.17) with respect to and show

dh f (b) f (a)
h
() = = f
() (2.18)
d ba

If there is a value = c such that h
(c) = 0, then equation (2.18) reduces to

f (b) f (a)
mt = f
(c) = = ms (2.19)
ba
which shows that there must exist a point x = c such that the slope of the tangent
line at the point (c, f (c)) is the same as the slope of the secant line through the points
(a, f (a)) and (b, f (b)) as illustrated in the figure 2-10.
The mean-value theorem can be expressed in a slightly different form if in equa-
tion (2.19) one makes the substitution b a = h, so that b = a + h. This produces the
form
f (a + h) = f (a) + hf
(c) where a < c < a + h (2.20)
110
Let denote a real number between 0 and 1 and express the number c in the form
c = a + h, then another form for the mean-value theorem is

f (a + h) = f (a) + hf
(a + h), 0<<1 (2.21)

Figure 2-11. Another form for the mean value theorem.

A physical interpretation of the mean-value theorem, which will prove to be

useful in later sections, is obtained from an examination of the figure 2-11.
In this figure let QR =  h where  0 as h 0, then one can write
QS =QR + RS
or y = f (a + h) f (a) = h + f
(a) h (2.22)
f (a + h) f (a)
or by the mean-value theorem =f
(a) +  = f
(a + h)
h
where  0 as h 0. This result was used earlier in equation (2.14).
d
In summary, if G(x) = G
(x) = g(x), then one form for the mean-value theorem
dx
is

G = G(a + h) G(a) = g(a)h + h or G = G(a + h) G(a) = G
(a)h + h (2.23)

where  0 as h 0.
 111
Cauchys Generalized Mean-Value Theorem
Let f (x) and g(x) denote two functions which are continuous on the interval [a, b].
Assume the derivatives f
(x) and g
(x) exist and do not vanish simultaneously for all
x [a, b] and that g(b)
= g(a). Construct the function

y(x) = f (x)[g(b) g(a)] g(x)[f (b) f (a)] (2.24)

and note that y(a) = y(b) = f (a)g(b) f (b)g(a) and so all the conditions exist such that
Rolles theorem can be applied to this function. The derivative of the function given
by equation (2.24) is

y
(x) = f
(x)[g(b) g(a)] g
(x)[f (b) f (a)]

and Rolles theorem states that there must exist a value x = c satisfying a < c < b
such that
y
(c) = f
(c)[g(b) g(a)] g
(c)[f (b) f (a)] = 0 (2.25)

By hypothesis the quantity g(b) g(a)
= 0 and g
(c)
= 0, for if g
(c) = 0, then equation
(2.25) would require that f
(c) = 0, which contradicts our assumption that the deriva-
tives f
(x) and g
(x) cannot be zero simultaneously. Rearranging terms in equation
(2.25) gives Cauchys generalized mean-value theorem that f (x) and g(x) must satisfy
f (b) f (a) f
(c)
=
, a<c<b (2.26)
g(b) g(a) g (c)

Note the special case g(x) = x reduces equation (2.26) to the form of equation (2.19).
Derivative of the Logarithm Function
Assume b > 0 is constant and y = y(x) = logb x. Use the definition of a derivative
and write
dy y(x + x) y(x)
= y
(x) = lim
dx x0 x
dy log b (x + x) logb (x)
= y
(x) = lim and use the properties of logarithms to write
dx x0 x
112
 
dy
1 x + x
= y (x) = lim logb
dx x0 x x
 
dy 1 x
 x
= y
(x) = lim logb 1 +
dx x0 x x x
 x/x
dy 1 x
= y
(x) = lim logb 1 + (2.27)
dx x0 x x
x
In equation (2.27) make the substitution h = and make note of the fact that
x
h 0 as x 0 to obtain
dy 1 1/h
= y
(x) = lim logb (1 + h) (2.28)
dx x h0
Recall from chapter 1 that lim (1+h)1/h = e and use this result to simplify the equation
h0
(2.28) to the form
dy d 1
= y
(x) = logb x = (logb e) (2.29)
dx dx x
Observe that in the special case b = e one can use the result loge e = ln e = 1 to
simplify the equation (2.29) to the following result.
dy d 1
If y = ln x, x > 0, then = ln x = , x
= 0 (2.30)
dx dx x
If y = logb u, where u = u(x) > 0, the chain rule for differentiation can be employed
to obtain the results
d d du d 1 du
logb u = logb u or logb u = (logb e) (2.31)
dx du dx dx u dx
and in the special case y = ln u, u = u(x) > 0, then
d d du d 1 du
ln u = ln u or ln u = (2.32)
dx du dx dx u dx

The more general situation is that for

dy d 1
y = ln |x|, then = ln |x| = , x
= 0 (2.33)
dx dx x
and if
dy d 1 du
y = ln |u|, then = ln |u| = , u(x)
= 0 (2.34)
dx dx u(x) dx

In differential notation one can write

du
d ln |u| = (2.35)
u
 113
Example 2-13. Find the derivatives of the following functions

(a) y = ln | cos x|, (b) y = log10 |x|, (c) y = logb |u(x)|

Solution
dy d 1 d sin x
(a) = ln | cos x| = cos x = = tan x
dx dx cos x dx cos x
dy d 1
(b) = log10 |x| = (log10 e) , x
= 0
dx dx x
dy d 1 du
(c) = logb |u(x)| = (logb e) , u(x)
= 0
dx dx u(x) dx

Derivative of the Exponential Function

Let y = y(x) = bx , with b > 0 denote a general exponential function. Knowing
how to differentiate the logarithm function can be used to find a derivative formula
for the exponential function. Recall that

y = bx if and only if x = logb y (2.36)

Make use of the chain rule for differentiation and differentiate both sides of the
equation x = logb y with respect to x to obtain
d d
x = logb y
dx dx
d d dy
x = logb y (2.37)
dx dy dx
1 dy
1 = logb e
y dx

This result can be expressed in the alternative form

dy 1 d 1
= y or (bx ) = bx (2.38)
dx logb e dx logb e

1
Using the identity5 logb e = the equation (2.38) can be expressed in the alternative
ln b
form
d x
(b ) = (ln b) bx (2.39)
dx
In the special case b = e there results logb e = ln e = 1, so that the equations (2.38)
and (2.39) simplify to the result
d x
e = ex (2.40)
dx
5 logb x
Use the change of base relation logb a = log a x
in the special case a = e and x = b.
114
Note the exponential function y = ex is the only function equal to its own derivative.
Often times the exponential function y = ex is expressed using the notation y = exp(x).
This is usually done whenever the exponent x is replaced by some expression difficult
to typeset as an exponent. Also note that the functions y = ex and y = ln x are inverse
functions having the property that

eln x = x for x > 0 and ln(ex ) = x for all values of x

If u = u(x), then a generalization of the above results is obtained using the chain
rule for differentiation. These generalizations are
d d du d du
(bu ) = (bu ) or (bu ) = (ln b) bu
dx du dx dx dx (2.41)
d d du d du
and (eu ) = (eu ) or (eu ) = eu
dx du dx dx dx

Because the exponential function y = eu is easy to differentiate, many differentiation

problems are converted to this form. For example, writing y = bx = ex ln b , then
dy d d x ln b
d
= (bx ) = e = ex ln b [x ln b]
dx dx dx dx

which simplifies to the result given by equation (2.39).

In differential notation, one can write
d eu =eu du
d au =au ln a du (2.42)
 u 
a
d =au du, 0 < a < 1 or a > 1
ln a

d n
Example 2-14. The differentiation formula x = nxn1 was derived for n an
dx
d
integer. Show that for x > 0 and r any real number one finds that xr = rxr1
dx
Solution Use the exponential function and write y = xr as y = er ln x , then

dy d r ln x d r
= e = er ln x (r ln x) = xr = rxr1
dx dx dx x
 115
Example 2-15. If y = | sin x|, find dy
dx
Solution Use the exponential function and write y = | sin x| = eln | sin x| , then

dy d ln | sin x| d
= e =eln | sin x| ln | sin x|
dx dx dx
if

1 d | sin x| cos x sin x > 0
=| sin x| sin x = cos x =
sin x dx sin x cos x if sin x < 0

dy
Example 2-16. If y = xcos x with x > 0, find dx
Solution Write y = xcos x as y = e(cos x) ln x , then

dy d (cos x) ln x d
= e =e(cos x) ln x [(cos x) ln x]
dx dx  dx 
cos x 1
=x cos x + ln x ( sin x)
x
 cos x 
=xcos x (ln x)(sin x)
x

Example 2-17. The general power rule for differentiation is expressed

d du
u(x)r = ru(x)r1 (2.43)
dx dx

where r can be any real number. This is sometimes written as

d d r ln u(x)
u(x)r = e
dx dx
d
=er ln u(x) [r ln u(x)]
dx
(2.44)
1 du
=u(x)r r
u(x) dx
du
=ru(x)r1
dx

which is valid whenever u(x)
= 0 with ln | u(x) | and u(x)r well defined.
116
Example 2-18. The exponential function can be used to differentiate the
general power function y = y(x) = u(x)v(x) , where u = u(x) > 0 and u(x)v(x) is well
defined. One can write y = u(x)v(x) = ev(x) ln u(x) and by differentiation obtain
dy d
= ev(x) ln u(x)
dx dx
d
=ev(x) ln u(x) [v(x) ln u(x)]
 dx 
v(x) 1 du dv
=u(x) v(x) + ln u(x)
u(x) dx dx

Derivative and Continuity

If a function y = f (x) is such that both the function f (x) and its derivative f
(x)
are continuous functions for all values of x over some interval [a, b], then the function
y = f (x) is called a smooth function and its graph is called a smooth curve. A smooth
function is characterized by an unbroken curve with a continuously turning tangent .

Example 2-19.
(a) The function y = f (x) = x2 4 has the deriva-
dy
tive dx = f
(x) = 2x which is everywhere contin-
uous and so the graph is called a smooth curve.

(b) The function

 
1
y = f (x) = 2 + exp ln | 2x 3 | = 2 + (|2x 3|)1/3
3

has the derivative

dy 1 2
= e 3 ln|2x3|
dx 3(2x 3)

which has a discontinuity in its derivative at the point x = 3/2 and so the curve is
not a smooth curve.
Maxima and Minima
Examine the curve y = f (x) illustrated in the figure 2-12 which is defined and
continuous for all values of x satisfying a x b. Start at the point x = a and move
along the x-axis to the point b examining the heights of the curve y = f (x) as you
move left to right.
 117

Figure 2-12. Curve y = f (x) with horizontal line indicating critical points.

A local maximum or relative maximum value for f (x) is said to occur at those
points where in moving from left to right the height of the curve increases, then stops
and begins to decrease. A local minimum or relative minimum value of f (x) is said
to occur at those points where in moving from left to right the height of the curve
decreases, then stops and begins to increase. In figure 2-12 the points x1 , x3 , x5, x8 are
where the function f (x) has local maximum values. The points x2 , x4 , x6 are where
f (x) has local minimum values. The end points where x = a and x = b are always
tested separately for the existence of a local maximum or minimum value.
Definition: (Absolute maximum) A function is said to have an absolute max-
imum M or global maximum M at a point (x0 , f (x0 )) if f (x0 ) f (x) for all x D,
where D is the domain of definition of the function and M = f (x0 ).
Definition: (Absolute minimum) A function is said to have an absolute min-
imum m or global minimum m at a point (x0 , f (x0 )) if f (x0 ) f (x) for all x D,
where D is the domain of definition of the function and m = f (x0 ).
For x D one can write m f (x) M where m and M are referred to as extreme
values of the function y = f (x). In the figure 2-12 the point where x = x5 gives
M = f (x5 ) and the point where x = x2 gives m = f (x2 ).Note that for functions defined
on a closed interval, the end points x = a and x = b must be tested separately for a
maximum or minimum value.
118
Definition: (Relative maximum) A function is said to have a relative maximum
or local maximum at a point (x0 , f (x0 ) if f (x0 ) f (x) for all x in some open interval
containing the point x0 .
Definition: (Relative minimum) A function is said to have a relative minimum
or local minimum at a point (x0 , f (x0 )) if f (x0 ) f (x) for all values of x in some
open interval containing the point x0 .

Concavity of Curve
If the graph of a function y = f (x) is such that f (x) lies above all of its tangents
on some interval, then the curve y = f (x) is called concave upward on the interval. In
this case one will have throughout the arc of the curve f

(x) > 0 which indicates that
as x moves from left to right, then f
(x) is increasing. If the graph of the function
y = f (x) is such that f (x) always lies below all of its tangents on some interval, then
the curve y = f (x) is said to be concave downward on the interval. In this case one
will have throughout the arc of the curve f

(x) < 0, which indicates that as x moves
from left to right, then f
(x) is decreasing. Related to the second derivative are
points known as points of inflection.
Definition: (Point of inflection)
Assume y = f (x) is a continuous function which
has a first derivative f
(x) and a second derivative
f

(x) defined in the domain of definition of the
function. A point (x0 , f (x0 )) is called an inflection
point if the concavity of the curve changes at that point. The second derivative
f

(x0 ) may or may not equal zero at an inflection point. One can state that a point
(x0 , f (x0 )) is an inflection point associated with the curve y = f (x) if there exists a
small neighborhood of the point x0 such that
(i) for x < x0 , one finds f

(x) > 0 and for x > x0 , one finds f

(x) < 0
or (ii) for x < x0 , one finds f

(x) < 0 and for x > x0 , one find f

(x) > 0

No local minimum value or maximum value occurs at an inflection point.

A horizontal inflection point is characterized by a tangent line parallel with the

x-axis with f
(a) = 0 and f

(a) = 0 as illustrated by point A. A vertical inflection
point is characterized by a tangent line parallel with the y -axis with f
(b) =
as illustrated by the point B . The point C illustrates a point of inflection where
 119
f
(c)
= 0
and f

(c) = 0 and the concavity changes from concave up to concave down
as x increases across the inflection point.
If the curve is continuous with continuous
derivatives, those points where the concavity
changes, from upward to downward or from
downward to upward, are inflection points with
a zero second derivative.

Sections of the curve which are concave upward will hold water, while those sections
that are concave downward will not hold water.
Comments on Local Maxima and Minima
Examine the figure 2-12 and make note of the following.
(1) The words extrema (plural) or extremum (singular) are often used when referring
to the maximum and minimum values associated with a given function y = f (x).
(2) At a local maximum or local minimum value the tangent line to the curve is
parallel to one of the coordinate axes.
(3) A local maximum or local minimum value is associated with those points x where
f
(x) = 0. The roots of the equation f
(x) = 0 are called critical points. Critical
points must then be tested to see if they correspond to a local maximum, local
minimum or neither, such as the point x7 in figure 2-12.
(4) Continuous curves which have abrupt changes in their derivative at a single
point are said to have cusps at these points. For example, the points where
x = x1 and x = x2 in the figure 2-12 are called cusps. At these cusps one finds
that either f
(x) = or f
(x) has a jump discontinuity. These points must be
tested separately to determine if they correspond to local maximum or minimum
values for y = f (x).
(5) The end points of the interval of definition x = a and x = b must be tested
separately to determine if a local maximum or minimum value exists.
(6) The conditions f
(x) = 0 or f
(x) = at a point x0 are not sufficient conditions for
an extremum value for the function y = f (x) as these conditions may produce an
inflection point or an asymptotic line and so additional tests for local maximum
and minimum values are needed.
(7) If the function y = f (x) is continuous, then between two equal values for the
function, where f
(x) is not zero everywhere, at least one maximum or one min-
imum value must exist. One can also say that between two maximum values
120
there is at least one minimum value or between two minimum values there is at
least one maximum value.
(8) In the neighborhood of a local maximum value, as x increases the function in-
creases, then stops changing and starts to decrease. Similarly, in the neighbor-
hood of a local minimum value, as x increases the function decreases, then stops
changing and starts to increase. In terms of a particle moving along the curve,
one can say that the particle change becomes stationary at a local maximum or
minimum value of the function. The terminology of finding stationary values of
a function is often used when referring to maximum and minimum problems.

First Derivative Test

The first derivative test for extreme values of a function tests the slope of the
curve at near points on either side of a critical point. That is, to test a given func-
tion y = f (x) for maximum and minimum values, one first calculates the derivative
function f
(x) and then solves the equation f
(x) = 0 to find the critical points. If x0
is a root of the equation f
(x) = 0, then f
(x0 ) = 0 and then one must examine how
f
(x) changes as x moves from left to right across the point x0 .
Slope Changes in Neighborhood of Critical Point
If the slope f
(x) changes from
(i) + to 0 to , then a local maximum occurs at the critical point.
(ii) to 0 to +, then a local minimum occurs at the critical point.
(iii) + to 0 to +, then a point of inflection is said to exist at the critical point.
(iv) to 0 to , then a point of inflection is said to exist at the critical point.
Given a curve y = f (x) for x [a, b], the values of f (x) at the end points where
x = a and x = b must be tested separately to determine if they represents relative or
absolute extreme values for the function. Also points where the slope of the curve
changes abruptly, such as the point where x = x1 in figure 2-12, must also be tested
separately for local extreme values of the function.
Second Derivative Test
The second derivative test for extreme values of a function y = f (x) assumes that
the second derivative f

(x) is continuous in the neighborhood of a critical point.
One can then say in the neighborhood of a local minimum value the curve will be
concave upward and in the neighborhood of a local maximum value the concavity
 121
of the curve will be downward. This gives the following second derivative test for
local maximum and minimum values.
(i) If f
(x0 ) = 0 and f

(x0 ) > 0, then f has relative minimum value at x0 .
(ii) If f
(x0 ) = 0 and f

(x0 ) < 0, then f has a relative maximum value at x0 .

If f

(x0 ) = 0, then the second derivative test fails and one must use the first derivative
test. The second derivative test is often used because it is convenient. The second
derivative test is not as general as the first derivative test. If the second derivative
test fails, then resort back to the more general first derivative test.

Example 2-20.
Find the maximum and minimum values of the function y = f (x) = x3 3x
Solution
dy d d
The derivative of the given function is = x3 3 x = 3x2 3 = f
(x). Setting
dx dx dx
f
(x) = 0 one finds

3(x2 1) = 3(x 1)(x + 1) = 0 with roots x = 1 and x = 1

being the critical points.

First derivative test
Selecting the points x = 3/2, x = 1 and x = 1/2 one finds that
15 9
f
(3/2) = 3x2 3 = , f
(1) = 0, f
(1/2) = 3x2 3 =
x=3/2 4 x=1/2 4

and so the slope of the curve changes from + to 0 to indicating a local maximum
value for the function.
Selecting the points x = 1/2, x = 1 and x = 3/2 one finds
9 15
f
(1/2) = 3x2 3 = , f
(1) = 0, f
(3/2) = 3x2 3 =
x=1/2 4 3/2 4

and so the slope of the curve changes from to 0 to + indicating a local minimum
value for the function.
Second derivative test
d2 y
The second derivative of the given function is 2 = f

(x) = 6x. The first and
dx
second derivatives evaluated at the critical points gives
122
(i) at x = 1 one finds f
(1) = 0 and f

(1) = 6(1) = 6 < 0 indicating the curve is
concave downward. Therefore, the critical point x = 1 corresponds to a local
maximum.
(ii) at x = 1 one finds f
(1) = 0 and f

(1) = 6(1) = 6 > 0 indicating the curve is concave
upward. Therefore, the critical point x = 1 corresponds to a local minimum
value.
Sketching the curve
The local minimum value at x = 1 is f (1) = (1)3 3(1) = 2 and the local
maximum value at x = 1 is f (1) = (1)3 3(1) = 2. Consequently, the curve
passes through the points (1, 2) being concave upward and it passes through the
point (1, 2) being concave downward.
Select random points in the neighborhood of
these points for additional information about the
curve. Select the points where x = 2, x = 0
and x = 2 and show the points (2, 2), (0, 0) and
(2, 2) lie on the curve. Plotting these points and
connecting them with a smooth curve gives the
following sketch.

Example 2-21. Snells Law

Refraction is the process where a light or sound wave changes direction when
passing from one isotropic6 medium to another. Examine the figure 2-13 illustrating
a ray of light moving from point P in air to a point Q in water. The point P is
a height h above the air-water interface and the point Q is at a depth d below the
air-water interface. The normal to the air-water interface is used to define angles of
incidence and refraction.
In figure 2-13 the symbol i denotes the angle of incidence and r denotes the
angle of refraction of the ray of light moving from point P in air to a point Q in
water and
is the x-distance between the points P and Q. Fermats law7 states that
light will travel from point P in air to point Q in water along a path P OQ which
6
Having the same physical properties in all directions.
7
Pierre de Fermat (1601-1665) A French lawyer famous for his developing many results in number theory and
calculus.
 123
minimizes the time travel. If c1 denotes the speed of light in air and c2 denotes the
speed of light in water one can find from tables the approximate values

c1 2.99 (10)10 cm/sec c2 2.26 (10)10 cm/sec

Find the relation between the angles i and r such that Fermats law is satisfied.

Figure 2-13. Light ray moving from point P in air to point Q in water.

Solution
Use the formula Distance = (V elocity) (T ime) to obtain the following values.
The time of travel for light in air to move from point P to O is

PO x2 + h2
Tair = =
c1 c1

The time of travel for light in water to move from point O to Q is


OQ (
x)2 + d2
Twater = =
c2 c2

The total time to travel from point P to Q is therefore


x2 + h2 (
x)2 + d2
T = T (x) = Tair + Twater = +
c1 c2
124
Calculating the derivative one finds
dT 1 x 1 (
x)
=

dx c1 x2 + h2 c2 (
x)2 + d2

dT
If the time T has an extreme value, then dx
= 0 and x is required to satisfy the
equation
1 x 1 (
x)
=

c1 x2 + h2 c2 (
x)2 + d2

and from this equation one can theoretically solve for the value of x which makes
T = T (x) have a critical value. This result can be expressed in a slightly different
form. Examine the geometry in the figure 2-13 and verify that
x (
x)
sin i = and sin r =

x + h2
2 (
x)2 + d2

so the condition for an extreme value can be written in the form

sin i sin r
= (2.45)
c1 c2

This result is known as Snells law.8 Show that the second derivative simplifies to
d2 T 1 h2 1 d2
= + >0
dx2 c1 (h2 + x2 )3/2 c2 ((
x)2 + d2 )3/2

By the second derivative test the critical point corresponds to a minimum value for
T = T (x)

x2 x + 1
Example 2-22. Consider the function y = f (x) = 2 and ask the question
x +x+1
Is this function defined for all values of x? If the denominator is not zero, then one

can answer yes to this question. If x2 +x+1 = 0, then x = 12 14 = 12 (1i 3) which
is a complex number and so for real values of x the denominator is never zero. One
can then say the domain of definition for the function is D = R. To determine the
range for the function, rewrite the function in the form x2 (1 y) x(1 + y) + (1 y) = 0

8
This law was discovered by Willebrord Snell (1591-1526) a Dutch astronomer. Let c denote the speed of light
in vacuum and cm the speed of light in medium m. The ratio nm = c/cm is called the absolute index of refraction
and the more general form of Snells law is n1 sin i = n2 sin r .
 125
In order that x be a real quantity it is necessary for
(1 + y)2 >4(1 y)2
1 + 2y + y 2 >4(1 2y + y 2 )
3y 2 + 10y 3 >0

(3y 1)(y 3) >0

1
This requires that 3
<y<3 which determines the range for the function.

x2 x + 1
Figure 2-14. Sketch of y =
x2 + x + 1

As x increases without bound one can write

x2 x + 1 1 1/x + 1/x2
lim = lim =1
x x2 + x + 1 x 1 + 1/x + 1/x2

so that y = 1 is an asymptotic line.

Differentiating the given function one finds
dy (x2 + x + 1)(2x 1) (x2 x + 1)(2x + 1) 2(x2 1)
= f
(x) = =
dx (x2 + x + 1)2 (x2 + x + 1)2

The slope of the curve f
(x) is zero when x = 1 or x = 1. These are the critical
points to be tested. One finds that at x = 1 the height of the curve is y = f (1) = 1/3
and when x = 1, the height of the curve is y = f (1) = 3. A sketch of the function
is given in the figure 2-14. By the first derivative test for x < 1, f
(x) > 0 and for
x > 1, f
(x) < 0 so that x = 1 corresponds to an absolute maximum value. It is
similarly demonstrated that the point x = 1 corresponds to an absolute minimum
value.
126
Example 2-23.
Find the largest rectangle that can be in-
scribed in a given triangle, where the base
of the rectangle lies on the base of the trian-
gle. Let b denote the base of the triangle and
let h denote the height of the triangle.

Solution
Let x denote the base of the rectangle and y the height of the rectangle, then the
area of the rectangle to be maximized is given by A = xy. This expresses the area as
a function of two variables. If y can be related to x, then the area can be expressed
in terms of a single variable and the area can be differentiated. In this way one can
apply the previous max-min methods for analyzing this problem. To begin, observe
that the triangles ABC and ADE are similar triangles so one can write
hy x hy h
= or x = b or y =h x
h b h b

This gives a relationship between the values of x and y. Note that as y varies from
y = 0 to the value y = h, the area A = xy will vary from 0 to a maximum value and
then back to 0. The area of the rectangle can now be expressed as either a function
of x or as a function of y. For example, if A = xy, then one can write either


h h
A =x h x = hx x2 , 0xb
b b


hy b
or A =b y = (hy y 2 ), 0yh
h h
These representation for the area can be differentiated to determine maximum and
minimum values for the area A. Differentiating with respect to x one finds
dA h
=h2 x
dx b

and a critical value occurs when dA dx

= 0 or h 2 hb x = 0, which requires x = b/2 with
y = h/2. Alternatively, if one differentiates with respect to y one finds

dA b
= (h 2y)
dy h

and a critical value occurs when dA

dy
= 0, or when y = h/2 which then gives x = b/2.
In both cases one find the maximum area as A = hb/4. Note that there is only
 127
one critical point as y varies from 0 to h, since the area is zero at the end points
where y = 0 and y = h, the Rolles theorem implies there must be a maximum value
somewhere between. That is, if the area is a continuous function of y and A increases
as y increases from 0, then the only way for A to return to zero is for it to reach a
maximum value, stop and then return to zero.

Logarithmic Differentiation
Whenever one is confronted with functions which are represented by complicated
products and quotients such as

x2 3 + x2
y = f (x) =
(x + 4)1/3

or functions of the form y = f (x) = u(x)v(x) , where u = u(x) and v = v(x) are complicated
functions, then it is recommended that you take logarithms before starting the
differentiation process. For example, to differentiate the function y = f (x) = u(x)v(x) ,
first take logarithms to obtain
 
ln y = ln u(x)v(x) which simplifies to ln y = v(x) ln u(x)

The right-hand side of the resulting equation is a product function which can then
be differentiated. Differentiating both sides of the resulting equation, one finds
d d d d
ln y = [v(x) ln u(x)] = v(x) ln u(x) + ln u(x) v(x)
dx dx dx dx
1 dy 1 du(x) dv(x)
=v(x) + ln u(x)
y dx u(x) dx dx

Solve this equation for the derivative term to obtain

 
dy 1 du(x) dv(x)
= y v(x) + ln u(x) (2.46)
dx u(x) dx dx

where y can be replaced by u(x)v(x) .

128
Differentiation of Inverse Functions
Assume that y = f (x) is a single-valued function of x in an interval (a, b) and the
derivative function dy
dx
= f
(x) exists and is different from zero in this interval. If the
inverse function x = f 1 (y) exists, then it has the derivative
dx 1
=
dy dy
dx

Proof
By hypothesis the function y = f (x) is differentiable and the derivative is nonzero
in an interval (a, b) and so one can use implicit differentiation and differentiate both
d d
sides of y = f (x) with respect to y to obtain y = f (x) which by the chain rule
dy dy
becomes
d d dx dx
y= f (x) or 1 = f
(x)
dy dx dy dy
Consequently, if f
(x)
= 0, then one can write
dx 1 1
=
=
dy f (x) dy
dx

An alternative way to view this result is as follows. If y = f (x), then one can
interchange x and y and write

x = f (y) and solving for y obtain y = f 1 (x)

Observe that by employing the chain rule there results

dx d d 1 d 1 dx
= f
(y) and y= f (x) = f (x)
dy dy dy dx dy

This last equation reduces to

d 1 dx 1 1 1 d 1
1= f (x) or =
=
1 = f (x) = f 1
(x)
dx dy dx f (y) f (f (x)) dx
dy

which gives the result

d 1 1
f (x) = f 1
(x) =
1 (2.47)
dx f (f (x))

provided that the denominator is different from zero.

 129
x+1 dy 1
Example 2-24. If y = f (x) = , show that = f
(x) = 2 . Solving for x
x dx x
1 dx d 1 1
one finds x = f 1 (y) = with derivative = f (y) = . Note that
y1 dy dy (y 1)2

dx d 1 1 1 1 1
= f (y) = 2
= dy = 1 = x2 =
dy dy (y 1) dx x2
(y 1)2

Approached from a different point of view one finds that by interchanging x and
y+1
y in the given function gives x = f (y) = and solving for y gives the inverse
y
1 1
function y = f (x) = . This function has the derivative
x1
d 1 d 1
f (x) = f 1
(x) = (x 1)1 =
dx dx (x 1)2

Using the equation (2.47) one can write

d 1 1 1
2 1
f (x) = f 1
(x) =
1 = f 1 (x) =

=
dx f (f (x)) 1 (x 1)2
2
(f 1 (x))

Example 2-25. If y = f (x) = ex , then interchanging x and y gives x = f (y) = ey

and solving for y one obtains y = f 1 (x) = ln x. Here the functions ex and ln x are
inverse functions of one another. Differentiate y = f 1 (x) = ln x to obtain
dy d 1 d 1
= f (x) = f 1
(x) = ln x =
dx dx dx x

Here f
(x) = ex and f
(f 1 (x)) = eln x = x and

d 1 d 1 1 1
f (x) = ln x = = f 1
(x) =
1 =
dx dx x f (f (x)) x

Differentiation of Parametric Equations

If x = x(t) and y = y(t) are a given set of parametric equations which define y as
a function of x by eliminating the parameter t and the functions x(t) and y(t) are
continuous and differentiable, then by the chain rule one can write
dy
dy dy dx dy y
(t)
= or = dt =
provided x
(t)
= 0.
dt dx dt dx dx x (t)
dt
130
Differentiation of the Trigonometric Functions
To derive the derivatives associated with the trigonometric functions one can
utilize the trigonometric identities
sin(A + B) sin(A B) = 2 cos A sin B
cos(A + B) cos(A B) = 2 sin A sin B

sin A cos B cos A sin B = sin(A B)

sin
as well as the limit relation lim =1 previously derived in the example 1-6.
0

Example 2-26. Find the derivative of y = sin x and then generalize this result
to differentiate y = sin u(x) where u = u(x) is an arbitrary function of x.
Solution
Using the definition of a derivative, if y = sin x, then
dy y(x + h) y(x) sin(x + h) sin x
= lim = lim
dx h0 h h0 h
h h h
dy 2 sin( 2 ) cos(x + 2 ) sin( 2 ) h
= lim = lim h
lim cos(x + ) = cos x
dx h0 h h0
2
h0 2

Therefore, the derivative of the sine function is the cosine function and one can write
d
sin x = cos x (2.48)
dx

Using the chain rule for differentiation this result can be generalized. If y = sin u,
dy dy du d d du
then = or sin u = sin u or
dx du dx dx du dx
d du
sin u = cos u (2.49)
dx dx

Example 2-27. Some examples involving differentiation of the sine function

are the following.
d d
sin(x2 ) = cos(x2 ) 2x sin(e3x ) = cos(e3x ) 3e3x
dx dx
d d   x
sin(x + ) = cos(x + ) sin( x2 + 1) = cos( x2 + 1)
dx dx 2
x +1
 131
Example 2-28. Find the derivative of y = cos x and then generalize this result
to differentiate y = cos u(x) where u = u(x) is an arbitrary function of x.
Solution
Using the definition of a derivative, if y = cos x, then
dy y(x + h) y(x) cos(x + h) cos(x)
= lim = lim
dx h0 h h0 h
h h
dy 2 sin( 2 ) sin(x + 2 ) sin( h2 ) h
= lim = lim h
lim sin(x + )
dx h0 h h0
2
h0 2
dy
= sin x
dx

One finds that the derivative of the cosine function is the negative of the sine function
giving
d
cos x = sin x (2.50)
dx
This result can be generalized using the chain rule for differentiation to obtain the
d d du
result cos u = cos u or
dx du dx
d du
cos u = sin u (2.51)
dx dx

Example 2-29. Some examples involving the derivative of the cosine function
are the following.

d d
cos(x2 ) = sin(x2 ) 2x cos(ex ) = sin(ex) ex
dx dx
d d   3x2
cos(x + ) = sin(x + ) cos( x3 + 1) = sin( x3 + 1)
dx dx 2 x3 + 1

Example 2-30. Find the derivative of y = tan x and then generalize this result
to differentiate y = tan u(x) where u = u(x) is an arbitrary function of x.
Solution
Until you get to a point where you memorize all the rules for differentiating a
function and learn how to combine all these results you are restricted to using the
definition of a derivative.
132
If y = tan x, then
dy y(x + h) y(x) tan(x + h) tan x
= lim = lim
dx h0 h h0 h
 
dy 1 sin(x + h) sin x sin(x + h) cos x cos(x + h) sin x
= lim = lim
dx h0 h cos(x + h) cos x h0 h cos x cos(x + h)
dy sin h 1
= lim lim
dx h0 h h0 cos x cos(x + h)
dy 1
= 2 = sec2 x
dx cos x

If you know the derivatives of sin x and cos x you can derive the derivative of the
tan x by using the quotient rule for differentiation and write
  d d
d d sin x cos x dx
sin x sin x dx
cos x
tan x = =
dx dx cos x cos2 x
d cos2 x + sin2 x 1
tan x = 2
= = sec2 x
dx cos x cos2 x

One finds
d
tan x = sec2 x (2.52)
dx
d d du
The chain rule can be utilized to show tan u = tan u or
dx du dx
d du
tan u = sec2 u (2.53)
dx dx

Example 2-31. Some examples involving the derivative of the tangent func-
tion are the following.
d d
tan(x2 ) = sec2 (x2 ) 2x tan(ex ) = sec2 (ex ) ex
dx dx
d d   1 + 2x
tan(x + ) = sec2 (x + ) tan( x2 + x) = sec2 ( x2 + x)
dx dx 2 x2 + x
 133
Example 2-32. Find the derivative of y = cot x and then generalize this result
to differentiate y = cot u(x) where u = u(x) is an arbitrary function of x.
Solution
cos x
Use the trigonometric identity y = cot x = and write
sin x
dy y(x + h) y(x) cot(x + h) cot x
= lim = lim
dx h0 h h0 h
cos(x+h) cos x
sin(x+h)
sin x cos(x + h) sin x cos x sin(x + h)
= lim = lim
h0 h h0 h sin x sin(x + h)
sin h 1
= lim lim
h0 h h0 sin x sin(x + h)
d 1
cot x = 2 = csc2 x
dx sin x
so that
d
cot x = csc2 x (2.54)
dx
Using the chain rule for differentiation one finds
d du
cot u(x) = csc2 u(x) (2.55)
dx dx

Example 2-33. Find the derivative of y = sec x and then generalize this result
to differentiate y = sec u(x) where u = u(x) is an arbitrary function of x.
Solution
1
Use the trigonometric identity y = sec x = and write
cos x
1
dy sec(x + h) sec x cos(x+h)
cos1 x
= lim = lim
dx h0 h h0 h
h
cos x cos(x + h) sin( 2 ) sin(x + h2 )
= lim = lim h
lim
h0 h cos x cos(x + h) h0 h0 cos x cos(x + h)
2
sin x 1 sin x
= 2 = = sec x tan x
cos x cos x cos x
so that
d
sec x = sec x tan x (2.56)
dx
Using the chain rule for differentiation one finds
d du
sec u(x) = sec u(x) tan u(x) (2.57)
dx dx
134
Example 2-34. Find the derivative of y = csc x and then generalize this result
to differentiate y = csc u(x) where u = u(x) is an arbitrary function of x.
Solution
1
Use the trigonometric identity y = csc x = and write
sin x
1
dy csc(x + h) csc x sin(x+h)
sin1 x
= lim = lim
dx h0 h h0 h
h
sin x sin(x + h) sin( 2 ) cos(x + h2 )
= lim = lim h
lim
h0 h sin x sin(x + h) h0 h0 sin x sin(x + h)
2
cos x 1 cos x
= 2
= = csc x cot x
sin x sin x sin x

so that
d
csc x = csc x cot x (2.58)
dx
Using the chain rule, show that
d du
csc u(x) = csc u(x) cot u(x) (2.59)
dx dx

Example 2-35. Some curves are easily expressed in terms of a parameter. For
example, examine the figure 2-15 which illustrates a circle with radius a which rolls
without slipping along the x-axis. On this circle there is attached a fixed arm of
length 0P = r, which rotates with the circle. At the end of the arm is a point P
which sweeps out a curve as the circle rolls without slipping. This arm initially lies
on the y-axis and the coordinates of the point P in this initial position is (0, (r a)).
As the circle rolls along the x-axis without slipping, the point P has coordinates
(x, y). From the geometry of the problem the coordinates of point P in terms of the
parameter are given by

x = a x0 = a r sin y = a + y0 = a r cos (2.60)

The term a in the parametric equations (2.60) represents arc length as the circle
rolls and the terms r sin and r cos represent projections of the arm onto the x and
y axes respectively.
 135

Figure 2-15. Circle rolling without slipping.

Figure 2-16. Cycloid curves for r = 23 a, r = a and r = a

2

The curve that the point P sweeps out as the circle rolls without slipping has different
names depending upon whether r > a, r = a or r < a, where a is the radius of the
circle. These curves are called
a prolate cycloid if r > a
a cycloid if r = a

a curtate cycloid if r < a

These curves are illustrated in the figure 2-16.

136
To construct a tangent line to some point (x0 , y0 ) on one of the cycloids, one must
be able to find the slope of the curve at this point. Using chain rule differentiation
dy dy dx
one finds = or
d dx d
dy
dy dx dy
= d , where = a r cos and = r sin
dx dx d d
d

The point (x0 , y0 ) on the cycloid corresponds to some value 0 of the parameter. The
slope of the tangent line at this point is given by
dy r sin
mt = =
dx a r cos =0

and the equation of the tangent line at this point is y y0 = mt (x x0 ).

Simple Harmonic Motion

If the motion of a particle or center of mass of a body can be described by either
of the equations

y = y(t) = A cos(t 0 ) or y = y(t) = A sin(t 0 ) (2.61)

where A, and 0 are constants, then the particle or body is said to undergo a
simple harmonic motion. This motion is periodic with least period T = 2/||. The
amplitude of the motion is |A| and the quantity 0 is called a phase constant or phase
angle.
Note 1: By changing the phase constant, one of the equations (2.61) can be trans-
formed into the other. For example,

A sin(t 0 ) = A cos [(t 0 ) /2] = A cos(t 0 ), 0 = 0 + /2

and similarly

A cos(t 0 ) = A sin [(t 0 ) /2] = A sin(t 0 ), 0 = 0 + /2

Note 2: Particles having the equation of motion

y = y(t) = sin t + cos t

 137
where , and are constants, can be written in the form of either equation in

(2.61) by multiplying both the numerator and denominator by 2 + 2 to obtain
 

y = y(t) = 2 + 2  sin t +  cos t (2.62)
2 + 2 2 + 2

, cos 0 = 2+ 2 reduces equation


The substitutions A = 2 + 2 , sin 0 =
2 + 2
(2.62) to the form

y = y(t) = A (sin 0 sin t + cos 0 cos t) = A cos(t 0 )

Example 2-36.
Consider a particle P moving around a circle
of radius a with constant angular velocity . The
points P1 and P2 are the projections of P onto the
x and y axes. The distance of these points from
the origin are described by the x and y-positions
of the particle and are given by x = x(t) = a cos t
and y = y(t) = a sin t. Here both P1 and P2 exhibit a simple harmonic motion about
the origin as the particle P moves counterclockwise about the circle. This simple
harmonic motion has a time period 2/ and amplitude a.
dx dy
The derivatives = x
(t) = a sin t and = y
(t) = a cos t represent the
dt dt
velocities of the points P1 and P2 . These velocities can be used to determine the
velocity of the particle P on the circle. Velocity is the change in distance with
respect to time. If s = a is the distance traveled by the particle along the circle,
ds d
then v = = a = a is the velocity of the particle. This same result can be
dt dt
obtained from the following analysis. The quantity dx = a sin t dt represents a
small change of P in x-direction and the quantity dy = a cos t dt represents a small
change of P in the y-direction. One can define an element of arc length squared
given by ds2 = dx2 + dy 2 . This result can be represented in the form
 2  2
 ds dx dy
ds = (dx)2 + (dy)2 = = + dt (2.63)
dt dt dt

and when the derivatives x
(t) and y
(t) are substituted into equation (2.63) there
ds
results v = = a. The second derivatives of x = x(t) and y = y(t) are found to be
dt
d2 x d2 y
= x = a 2 cos t and = y = a 2 sin t (2.64)
dt2 dt2
138
which can be written in the form

x = 2 x and y = 2 y (2.65)

This shows that one of the characteristics of simple harmonic motion is that the
magnitude of the acceleration of either the point P1 or P2 is always proportional
to the displacement from the origin and the direction of the acceleration is always
opposite to that of the displacement.

LHopitals Rule
One form of LHopitals9 rule, used to evaluate the indeterminate form 00 , is the
following. If f (x) and g(x) are both differentiable functions and satisfy the properties

lim f (x) = 0, lim g(x) = 0, lim g
(x)
= 0,

xx0 xx0 xx0

then one can write

f (x) f
(x)
lim = lim (2.66)
xx0 g(x) xx0 g
(x)
provided the limit on the right-hand side exists.
The proof of the above statement is obtained by using the definition of the
derivative and properties of the limiting process. One can write
f (x) f (x0 )
lim
f
(x) f
(x0 ) xx0 x x0
lim =
=
xx0 g
(x) g (x0 ) g(x) g(x0 )
lim
xx0 x x0
f (x) f (x0 ) f (x)
= lim = lim
xx0 g(x) g(x0 ) xx0 g(x)

because f (x0 ) = g(x0 ) = 0 by hypothesis.

The LHopitals rule can also be used to evaluate the indeterminate form
. If
lim f (x) = and lim g(x) = , then one can write
x x

f (x) f
(x)
lim = lim (2.67)
x g(x) x g
(x)

To show this is true make the substitution x = 1/t so that as x , then t 0 and
write
f (x) f (1/t)
lim = lim , t>0 (2.68)
x g(x) t0 g(1/t)

9
Guillaume Francois Antoine Marquis LHopital (1661-1704) French mathematician who wrote the first calculus
book. LHopitals name is sometimes translated as LHospital with the s silent.
 139
and then apply LHopitals rule to the right-hand side of equation (2.68) to obtain
f (x) f
(1/t)(1/t2 ) f
(1/t) f
(x)
lim = lim
= lim = lim
x g(x) t0 g (1/t)(1/t2 ) t0 g
(1/t) x g
(x)

Still another form of LHopitals rule is that if x0 is a finite real number and

lim f (x) = and lim g(x) =

xx0 xx0

then one can write

f (x) f
(x)
lim = lim (2.69)
xx0 g(x) xx0 g
(x)
This result can be established using the result from equation (2.67). Make the
substitution x = x0 + 1/t so that as x x0 , then t and write
f (x) f (x0 + 1/t) F (t)
lim = lim = lim (2.70)
xx0 g(x) t g(x0 + 1/t) t G(t)

Applying LHopitals rule from equation (2.67) one finds

f (x) F (t) F
(t) f
(x0 + 1/t)(1/t2 ) f
(x)
lim = lim = lim
= lim
= lim
xx0 g(x) t G(t) t G (t) t g (x0 + 1/t)(1/t2 ) xx0 g
(x)

Note that sometimes LHopitals rule must be applied multiple times. That is,
f
(x)
if lim
is an indeterminate form, then apply LHopitals rule again and write
xx0 g (x)

f (x) f
(x) f

(x)

lim = lim
= lim


xx0 g(x) xx0 g (x) xx0 g (x)

1 cos x
Example 2-37. Find lim .
x0 x2
Solution Use LHopitals rule multiple times and write

1 cos x ( sin x) cos x 1

lim 2
= lim = lim =
x0 x x0 2x x0 2 2

Make note of the fact that the functions that are used in equations (2.66), (2.67),
and (2.69) can themselves be derivatives.
One final note about LHopitals rule. There may occur limits10 where a re-
peated application of LHopitals rule puts you into an infinite loop and in such
cases alternative methods for determining the limits must be employed.

10 x2 +1
For example, LHopitals rule applied to limx x produces an infinite loop.
140
Example 2-38.
sin x
(a) Evaluate the limit lim
x0 x
sin x cos x
Solution Using the LHopitals rule one finds lim = lim =1
x0 x x0 1
ln x
(b) Evaluate the limit x
lim
x
ln x 1/x
Solution By LHopitals rule x
lim = lim =0
x x 1
ln(sin x)
(c) Evaluate the limit lim
x0 ln(tan x)
1
ln(sin x) cos x
Solution By LHopitals rule lim = lim sin x = lim cos2 x = 1
x0 ln(tan x) x0 1 x0
sec2 x
tan x

n
1
Example 2-39. Use LHopitals rule to show lim 1 + =e

n
x n
1 1
Solution Write 1 + = en ln(1+ n ) , then by LHopitals rule
n
ln(1 + x1 )
one can show that limx 1 =1
x

Differentiation of Inverse Trigonometric Functions

Examine the inverse trigonometric functions illustrated in the figures 1-18 and
1-19 presented in chapter 1. Observe that these functions are multi-valued functions
and because of this their derivatives depend upon which branch of the function
you are dealing with. In the example 1-5 a branch was assigned to each inverse
trigonometric function. You are not restricted to use these branches all the time. In
using mathematics to solve applied problems it is customary to select the branch of
the inverse trigonometric function which is applicable to the kind of problem you are
solving. In this section derivations for the derivatives of the inverse trigonometric
functions will be given for all possible branches that you might want to deal with.
By definition y = sin1 u is equivalent to sin y = u and the branch where 2 y 2
has been selected. Hence, to find the derivative of y = sin1 u one can differentiate
instead the equivalent relationship sin y = u. Differentiating with respect to x one
finds
d du dy du dy 1 du
sin y = or cos y = or =
dx dx dx dx dx cos y dx
Consequently,
dy d 1 du
= sin1 u =
dx dx cos y dx
 141
Observe that u = sin y is related to cos y, since sin2 y + cos2 y = 1 so that one can write
 
cos y = 1 sin2 y = 1 u2 (2.71)

The sign selected for the square root function depends upon where y is located. If
y = sin1 u is restricted to the first and fourth quadrant, where 2 y 2 , then cos y
is positive and so the plus sign is selected for the square root. However, if y = sin1 u
is restricted to the second or third quadrant, where 2 < sin1 u < 32 , then the function
cos y is negative and so the minus sign is selected for the square root function. This
gives the following differentiation formula for the function y = sin1 u = arcsinu
1 du


, |u| < 1, 2 < sin1 u <
2
d d 1

1 u2 dx
arcsinu = sin u = (2.72)
dx dx 1 du 1 3

, |u| < 1, 2
< sin u< 2
2
1 u dx

To differentiate the function y = arccos u = cos1 u write the equivalent statement

cos y = u and then differentiate both sides of this equivalent equation with respect to
x to obtain
d dy du dy 1 du
cos y = sin y = or =
dx dx dx dx sin y dx
Here u = cos y and sin y are related by
 
sin y = 1 cos2 u = 1 u2

where the sign assigned to the square root function depends upon where y lies. If
y = cos1 u lies in the first or second quadrant, then sin y is positive and so the plus
sign is selected. If y = cos1 u is the third or fourth quadrant, then sin y is negative
and so the minus sign is selected. One can then show that
1 du


, |u| < 1, 0 < cos1 u <
d d 1
2
1 u dx
arccos u = cos u = (2.73)
dx dx 1 du 1

, |u| < 1, < cos u < 2
2
1 u dx

In a similar fashion, if y = arctanu = tan1 u, then write tan y = u and differentiate

both sides of this equation with respect to x and show
d dy du dy 1 du
tan y = sec2 y = or =
dx dx dx dx sec2 y dx
142
Use the identity 1 + tan2 y = sec2 y = 1 + u2 and show
d 1 du
tan1 u = (2.74)
dx 1+ u2 dx

This result holds independent of which quadrant the angle y = tan1 u lies in.
In a similar fashion one can derive the derivative formulas for the inverse func-
tions cot1 u, sec1 u and csc1 u. One finds
d 1 du
cot1 u = (2.75)
dx 1 + u2 dx

a result which holds independent of which quadrant the angle y = cot1 u lies in.
The derivatives for the inverse secant and cosecant functions are found to be
1 du


, 0 < sec1 u < or < sec1 u <
2
d 2
u u 1 dx 2
sec1 u = (2.76)
dx 1 du 3
< sec1 u < 1


, or 2
< sec u < 2
2
u u 1 dx 2
1 du


, < csc1 u < or 3
2
< csc1 u < 2
d
u u2 1 dx 2
csc1 u = (2.77)
dx 1 du
, 0 < csc1 u < 1 3


or < csc u< 2
2
u u 1 dx 2

Hyperbolic Functions and their Derivatives

The hyperbolic functions were introduced around the year 1760 by the mathe-
maticians Vincenzo Riccati11 and Johan Heinrich Lambert.12 These functions were
previously defined in terms of the exponential functions ex and ex . Hyperbolic
functions occur in many areas of physics, engineering and related sciences. Recall
that these functions are defined
ex ex e2x 1
sinh x = = hyperbolic sine function
2 2ex
e + ex
x 2x
e +1
cosh x = = hyperbolic cosine function (2.78)
2 2ex
sinh x e ex
x
e2x 1
tanh x = = = hyperbolic tangent function
cosh x ex + ex e2x + 1
11
Vincenzo Riccati (1707-1775) An Italian mathematician.
12
Johan Heinrich Lambert (1728-1777) A French mathematician.
 143
Analogous to the definition of the trigonometric functions the coth x, sech x and csch x
are defined.
cosh x ex + ex e2x + 1
coth x = = x = hyperbolic cotangent function
sinh x e ex e2x 1
1 2 2ex
sech x = = x = hyperbolic secant function (2.79)
cosh x e + ex e2x + 1
1 2 2ex
csch x = = x = hyperbolic cosecant function
sinh x e ex e2x 1

Figure 2-17. Circular functions and hyperbolic functions.

The set of points C = { (x, y) | x = cos(t), y = sin(t), 0 < t < 2 } defines a circle of
unit radius centered at the origin as illustrated in the figure 2-17. The parameter t
has the physical significance of representing an angle of rotation. This representation
for the circle gives rise to the terminology of calling trigonometric functions circular
functions. In a similar fashion, the set of points

H = { (x, y) | x = cosh (t), y = sinh (t), t R }

defines the right-half of the equilateral hyperbola defined by x2 y2 = 1 as illustrated

in the figure 2-17. If the point (x, y) = ( cosh (t), sinh (t)) on the hyperbola is connected
to the origin by a straight line, then an area between the line, the x-axis and the
hyperbola is formed having an area a. The physical significance of the parameter t in
this representation is that t = 2a. This representation gives rise to the terminology of
144
calling the functions in equations (2.78) and (2.79) the hyperbolic functions. Graphs
of the hyperbolic functions, defined by the equations (2.79), are illustrated in the
figures 2-18 and 2-19.

Figure 2-18. Hyperbolic functions sinh (t), cosh (t), tanh (t).

An examination of the hyperbolic functions illustrated in the figures 2-18 and

2-19 show that
sinh (x) = sinh (x) csch (x) = csch (x)
cosh (x) = cosh (x) sech (x) = sech (x) (2.80)
tanh (x) = tanh (x) coth (x) = coth (x)

which shows that the functions cosh (x) and sech (x) are even function of x symmet-
ric about the y-axis and the functions sinh (x), tanh (x), csch (x) and coth (x) are odd
functions of x being symmetric about the origin.

Approximations
For large values of |x|, with x > 0 For large values of |x|, with x < 0
1 x 1 x
cosh x sinh x e cosh x sinh x e
2 2
tanh x coth x 1 tanh x coth x 1

sech x csch x 2ex 0 sech x csch x 2ex 0

 145

Figure 2-19. Hyperbolic functions csch (t), sech (t), coth (t)

Hyperbolic Identities
One can readily show that the hyperbolic functions satisfy many properties
similar to the trigonometric identities. For example, one can use algebra to verify
that
cosh x + sinh x = ex and cosh x sinh x = ex (2.81)

and

cosh 2 x sinh 2x = 1, 1 sech 2 x = tanh 2 x, 1 + csch 2 x = coth 2 x (2.82)

Algebra can also be used to prove the addition formula

sinh (x + y) = sinh x cosh y + cosh x sinhy
cosh (x + y) = cosh x cosh y + sinh x sinhy (2.83)
tanh x + tanh y
tanh (x + y) =
1 + tanh x tanh y

Example 2-39. Show that sinh (x + y) = sinh x cosh y + cosh x sinh y.

Solution
Use the law of exponents and show that
ex+y exy ex ey exey
sinh (x + y) = =
2 2
146
Now use the results from the equations (2.81) to show
1
sinh (x + y) = [( cosh x + sinhx)( cosh y + sinh y) ( cosh x sinh x)( cosh y sinh y)]
2
which when expanded simplifies to the desired result.

Replacing y by y in the equations (2.83) produces the difference expansions

sinh (x y) = sinh x cosh y cosh x sinhy
cosh (x y) = cosh x cosh y sinh x sinhy (2.84)
tanh x tanh y
tanh (x y) =
1 tanh x tanh y

Substituting y = x in the equations (2.83) produces the results

sinh (2x) =2 sinh x cosh x
cosh (2x) = cosh 2 x + sinh 2 y = 2 cosh 2 x 1 = 1 + 2 sinh 2 x (2.85)
2 tanh x
tanh (2x) =
1 + tanh 2 x

It is left for the exercises to verify the additional relations

   
x+y xy
sinh x + sinh y =2 sinh cosh
2 2
   
x+y xy
cosh x + cosh y =2 cosh cosh (2.86)
2 2
sinh (x + y)
tanh x + tanh y =
cosh x cosh y

   
x+y xy
sinh x sinhy =2 cosh sinh
2 2
   
x+y xy
cosh x cosh y =2 sinh sinh (2.87)
2 2
sinh(x y)
tanh x tanh y =
cosh x cosh y

  
x 1
sinh = ( cosh x 1)
2 2
  
x 1
cosh = ( cosh x + 1) (2.88)
2 2
 
x cosh x 1 sinh x
tanh = =
2 sinh x cosh x + 1
 147
Eulers Formula
Sometime around the year 1790 the mathematician Leonhard Euler13 discovered
the following relation
eix = cos x + i sin x (2.89)

where i is an imaginary unit with the property i2 = 1. This formula is known as

Eulers formula and is one of the most important formulas in all of mathematics.
The Euler formula can be employed to make a connection between the trigonometric
functions and the hyperbolic functions.
In order to prove the Euler formula given by equation (2.89) the following result
is needed.
If a function f (x) has a derivative f
(x) which is everywhere zero within an
interval, then the function f (x) must be a constant for all values of x within the
interval.
The above result can be proven using the mean-value theorem considered earlier.


If f (c) = 0 for all values c in an interval and x1
= x2 are arbitrary points within the
interval, then the mean-value theorem requires that
f (x2 ) f (x1 )
f
(c) = =0
x2 x1

This result implies that f (x1 ) = f (x2 ) for all values x1
= x2 in the interval and hence
f (x) must be a constant throughout the interval.
To prove the Euler formula examine the function

F (x) = (cos x i sin x)eix , i2 = 1 (2.90)

where i is an imaginary unit, which is treated as a constant. Differentiate this

product and show
d
F (x) = F
(x) =(cos x i sin x)eix i + ( sin x i cos x)eix
dx (2.91)
F
(x) = [i cos x + sin x sin x i cos x] eix = 0

Since F
(x) = 0 for all values of x, then one can conclude that F (x) must equal a
constant for all values of x. Substituting the value x = 0 into equation (2.90) gives

F (0) = (cos 0 i sin 0)ei0 = 1

13
Leonhard Euler (1707-1783) A famous Swiss mathematician.
148
so that the constant value is 1 and consequently

1 = (cos x i sin x)eix (2.92)

Multiply both sides of equation (2.92) by (cos x + i sin x) and show

cos x + i sin x =(cos x + i sin x)(cos x i sin x)eix

cos x + i sin x =(cos2 x + sin2 x)eix
cos x + i sin x =eix

which is the Euler formula given by equation (2.89).

Note that if

eix = cos x + i sin x (2.93)

then eix = cos x i sin x (2.94)

since cos(x) = cos x and sin(x) = sin x. Adding and subtracting the above equa-
tions produces the results
eix eix eix + eix
sin x = and cos x = (2.95)
2i 2

Examine the equations (2.95) and then examine the definitions of the hyperbolic
sine and hyperbolic cosine functions to obtain the immediate result that

i sin x = sinh (ix) and cos x = cosh (ix) (2.96)

which states that complex values of the hyperbolic sine and cosine functions give
relations involving the trigonometric functions sine and cosine. Replacing x by ix in
the equations (2.96) produces the results

sinh x = i sin(ix) and cosh x = cos(ix) (2.97)

The results from equations (2.96) and (2.97) together with the definition of the
hyperbolic functions gives the additional relations

tanh x = i tan(ix) and tanh (ix) = i tan x (2.98)

Properties of these complex functions are examined in more detail in a course

dealing with complex variables and their applications.
 149
Example 2-40. Using the above definitions one can show that an alternative
form of de Moivres theorem is

( cosh x sinh x)n = cosh nx sinh nx

Derivatives of the Hyperbolic Functions

For u = u(x) an arbitrary function of x, the derivatives of the exponential func-
tions
d du d du
eu = eu and eu = eu (2.99)
dx dx dx dx
can be used to calculate the derivatives of the hyperbolic functions, since they are
all defined in terms of exponential functions.

Example 2-41. Find the derivatives of the functions sinh u and cosh u where
u = u(x) is a function of x.
Solution
Use the definitions of the hyperbolic sine and cosine functions and write
d eu eu eu + eu du
 
d du
sinh u = = = cosh u
dx dx 2 2 dx dx
d eu + eu eu eu du
 
d du
cosh u = = = sinh u
dx dx 2 2 dx dx

Following the above example, the derivatives of all the hyperbolic functions can
be calculated. One can verify that the following results are obtained.
d du d du
sinh u = cosh u csch u = csch u coth u
dx dx dx dx
d du d du
cosh u = sinh u sech u = sech u tanh u (2.100)
dx dx dx dx
d du d du
tanh u = sech 2 u coth u = csch 2 u
dx dx dx dx

Inverse Hyperbolic Functions and their Derivatives

One can define the inverse hyperbolic functions14 in a manner analogous to how
the inverse trigonometric functions were defined. For example,
14
Sometimes the inverse hyperbolic functions are represented using the notations, arcsinh, arccosh, arctanh,
arccoth, arcsech, arccsch
150
y = sinh 1 x if and only if sinh y = x
y = cosh 1 x if and only if cosh y = x
y = tanh 1 x if and only if tanh y = x
(2.101)
y = coth 1 x if and only if coth y = x
y = sech 1 x if and only if sech y = x
y = csch 1 x if and only if csch y = x

Figure 2-20. Inverse Hyperbolic functions sinh 1 (t), cosh 1 (t), tanh 1 (t).

Graphs of the inverse hyperbolic functions can be obtained from the graphs of
the hyperbolic functions by interchanging x and y on the graphs and axes and then
re-orienting the graph. The sketches given in the figures 2-20 and 2-21 illustrate the
inverse hyperbolic functions.
Examine the figures 2-20 and 2-21 and note the functions cosh 1 t and sech 1 t are
multi-valued functions. The other inverse functions are single-valued. The branches
where cosh 1 t and sech 1 t are positive are selected as the principal branches. If you
want the negative values of these functions, then use the functions cosh 1 t and
sech 1 t.
 151

Figure 2-21. Inverse Hyperbolic functions csch 1 (t), sech 1 (t), coth 1 (t)

The hyperbolic functions are defined in terms of exponential functions. The

inverse hyperbolic functions can be expressed in terms of logarithm functions, which
is the inverse function associated with the exponential functions.

Example 2-42. Express y = sinh 1 x in terms of logarithms.

Solution
If y = sinh 1 x, then sinh y = x or
ey ey
sinh y = =x or ey ey = 2x
2
This last equation can be converted to a quadratic equation in the unknown ey . This
is accomplished by a multiplication of the equation ey ey = 2x by ey to obtain

(ey )2 2x(ey ) 1 = 0

Solving this quadratic equation for the unknown ey one finds


ey = x x2 + 1 (2.102)

Observe that ey is always positive for real values of y and x2 + 1 > x. Hence, in
order that ey remain positive, one must select the positive square root in equation
(2.102). Then solving the equation (2.102) for y one obtains the result

y = sinh 1 x = ln(x + x2 + 1), < x <
152
Example 2-43. Express y = tanh 1 x in terms of logarithms.
Solution
If y = tanh 1 x, then tanh y = x, which implies 1 < x < 1. Use the definition of
the hyperbolic tangent to write
ey ey
tanh y = =x or ey ey = xey + xey
ey + ey

Multiply the last equation through by ey and then solve for e2y to obtain

2y y 1+x
(1 x)e = (1 + x) giving e =
1x

Solving for y gives  

1 1+x
y = ln , 1 < x < 1
2 1x

Example 2-44. Find the derivative of y = cosh 1 u where u = u(x) is a function

of x.
SolutionOne can write cosh y = u and then differentiate both side with respect to x
and obtain
dy du
sinh y =
dx dx
dy dy 1 du 1 du 1 du
and then solve for to obtain = =  =
dx dx sinh y dx cosh y 1 dx
2 u 1 dx
2

where one uses the + sign in the principal value region where y > 0.

The previous examples demonstrate how one can establish the representations
 
sinh 1 x = ln x + x2 + 1 , < x <

cosh 1 x = ln(x + x2 1), x 1
 
1 1+x
tanh 1 x = ln , 1 < x < 1
2 1x
 
1
1 x+1
coth x = ln , x > 1 or x < 1 (2.103)
2 x1

 
1 1 1
sech x = ln + 1 , 0<x<1
x x2

 
1 1
csch 1 x = ln + + 1 , x
= 0
x x2
 153
Relations between Inverse Hyperbolic Functions
1
In the previous equations (2.103) replace x by x
and show
 
1 1
sinh = csch 1 x
x
 
1 1
cosh = sech 1 x (2.104)
x
 
1 1
tanh = coth 1 x
x

Example 2-45.

(a) Examine the logarithm of the product (x + x2 + 1)(x + x2 + 1) = 1 and observe
 
that ln(x + x2 + 1) = ln(x + x2 + 1). This result can be used to show

sinh 1 (x) = ln(x + x2 + 1) = sinh 1 x
 
1 1 1+x
(b) If tanh x = ln , then
2 1x
   
1 1 1x 1 1+x
tanh (x) = ln = ln = tanh 1 x
2 1+x 2 1x

(c) If y = sech 1 x, with y > 0 and 0 < x < 1, then x = sech y. By definition
1
sech y = , so that one can write
cosh y
 
1 1 1
= cosh y or y = cosh = sech 1 x
x x

are equivalent statements.

Using the techniques illustrated in the previous example one can verify the
following identities

sinh 1 (x) = sinh 1 x

tanh 1 (x) = tanh 1 x
(2.105)
coth 1 (x) = coth 1 x
csch 1 (x) = csch 1 x
154
Derivatives of the Inverse Hyperbolic Functions
To obtain the derivatives of the inverse hyperbolic functions one can differentiate
the functions given by the equations (2.103). For example, if
 
y = sinh 1 x = ln x + x2 + 1
dy d d  
then = sinh 1 x = ln x + x2 + 1
dx dx dx
1 d 
= x + (x2 + 1)1/2
x + x2 + 1 dx
 
1 x
= 1+
x + x2 + 1 x2 + 1

x + x2 + 1
=

x + x2 + 1 x2 + 1
d 1
sinh 1 x =
dx 2
x +1

One can use the chain rule for differentiation to generalize this result and obtain
d 1 du
sinh 1 u =
dx u2 + 1 dx

where u = u(x) is a function of x. In a similar fashion all the inverse hyperbolic

functions can be differentiated and one can verify that
d 1 du
sinh 1 u = , < u <
dx u2 + 1 dx
d 1 du
cosh 1 u = , u>1 (2.106)
dx u2 1 dx
d 1 du
tanh 1 u = 2
, 1 < u < 1
dx 1 u dx

Example 2-46. Find the derivative of y = sech 1 x with y > 0.

Solution

1+ 1 x 2
If y = sech 1 x = ln
x


dy d 1 d 1 + 1 x2
then = sech x = ln , 0<x<1
dx dx dx x
2
x 1 1 x2
1 2
= 1x
1+ 1x 2 x2
x
1
= , 0<x<1
x 1 x2
 155
The chain rule for differentiation can be used to generalize this result to
d 1 du
sech 1 u = , 0<u<1
dx u 1 u dx
2

If the lower half of the hyperbolic secant curve is used, then the sign of the above
result must be changed.

Differentiation of the logarithmic functions which define the inverse hyperbolic

functions, one obtains the results
d 1 du
coth 1 u = 2 , u2 > 1
dx u 1 dx
du

1 , sech 1 u > 0, 0 < u < 1
d
u 1+u 2 dx
sech 1 u =
dx du
 1 sech 1 u < 0, 0 < u < 1

, (2.107)
u 1 + u2 dx
du

1 , u>0
d 1

u 1 + u2 dx
csch u =
dx du
 1

, u<0
u 1 + u2 dx

Some additional relations involving the inverse hyperbolic functions are the fol-
lowing.
x
sinh 1 x = tanh 1 sinh 1 x = i sin1 (ix)
x2 +1

sinh 1 x = cosh 1 x2 + 1 cosh 1 x = i cos1 x
x
tanh 1 x = sinh 1 , |x| < 1 tanh 1 x = i tan1 (ix)
1 x2

Example 2-47. As an exercise study Mercator pro-

jections and conformal mappings and show projections of
point P using line from 0 to A gives y2 latitude which dis-
torts map shape and distances and projections of point P
using the line C to B also distorts shapes and distances.
Show the correct conformal projection is y between y1 and
y2 such that dy
d
= sec and

y = ln[sec + tan ] = tanh 1 [sin ]

156
Table of Derivatives
df
Function f (x) Derivative
dx
dy
y = xm dx
= mxm1
dy
y = ax dx
= ax ln a
dy
y = ex dx
= ex
dy
y = sin x dx
= cos x
dy
y = cos x dx
= sin x
dy
y = tan x dx
= sec2 x
dy
y = cot x dx
= csc2 x
dy
y = sec x dx
= sec x tan x
dy
y = csc x dx
= csc x cot x

y = sin1 x dy
dx
= 1
1x2
dy
y = cos1 x dx
= 1
1x2
dy 1
y = tan1 x dx
= 1+x2
dy 1
y = cot1 x dx
= 1+x2
dy
y = sec1 x dx
= 12
x x 1
1 dy 1
y = csc x dx
=
x x2 1
dy
y = sinhx dx
= cosh x
dy
y = cosh x dx
= sinh x
dy
y = tanh x dx
= sech 2 x
dy
y = coth x dx
= csch 2 x
dy
y = sech x dx
= sech x tanh x
dy
y = csch x dx
= csch x coth x
dy 1
y = sinh 1 x = ln(x + 1 + x2 ) dx
=
1+x2
dy 1
y = cosh 1 x = ln(x + x2 1) dx
=
x2 1

dy
y = tanh 1 x = 12 ln 1+x
1x dx
= 1
1x2

dy
y = coth 1 x = 12 ln x+1
x1 dx
= 1
x2 1
dy
y = sech 1 x = cosh 1 ( x1 ) dx
= 1
x 1x2
dy
y = csc1 x = sinh 1 ( x1 ) dx
= 12
x x +1
 157
Table of Differentials

d(c u) = c du
d(u + v) = du + dv d (uv ) = v uv1 du + un (ln u) dv

d(u + v + w) = du + dv + dw d (uu ) = uu (1 + ln u) du

d(u v) = u dv + v du d (eu ) = eu du

d(u v w) = u v dw + u dv w + du v w d (bu ) = bu (ln b) du

  1
u v du u dv d(ln u) = du
d = u
v v2
1
n
d (u ) = nu n1
du d(logb u) = (logb e) du
u

d sin u = cos u du d sin1 u = (1 u2 )1/2 du

d cos u = sin u du d cos1 u = (1 u2 )1/2 du

d tan u = sec2 u du d tan1 u = (1 + u2 )1 du

d cot u = csc2 u du d cot1 u = (1 + u2 )1 du

1
d sec u = sec u tan u du d sec1 u = (u2 1)1/2 du
u
d csc u = csc u cot u du 1
d csc1 u = (u2 1)1/2 du
u
All angles in first quadrant.

d sinh 1 u = (u2 + 1)1/2 du

d sinh u = cosh u du
d cosh 1 u = (u2 1)1/2 du
d cosh u = sinh u du
d tanh 1 u = (1 u2 )1 du
d tanh u = sech 2 u du
d coth 1 u = (u2 1)1 du
d coth u = csch 2 u du
1
d sech u = sech u tanh u du d sech 1 u = (1 u2 )1/2 du
u
1
d csch u = csch u coth u du d csch 1 u = (u2 + 1)1/2 du
u

d(xy) =x dy + y dx
y x dy y dx
x y dx x dy d( ) =
d( ) = x x2
y y2 y x dy y dx
2
x +y 2 d[tan1 ( )] =
d( ) =x dx + y dy x x2 + y 2
2
158
Partial Derivatives
If u = u(x1 , x2 , x3 , . . . , xn) is a function of several independent variables, the dif-
ferentiation with respect to one of the variables is done by treating all the other

variables as constants. The notations , , ,..., are used to denote
x1 x2 x3 xn

these differentiations. The partial derivative symbol indicates all variables dif-
xi
ferent from xi are being held constant. For example, if u = u(x, y) is a function of
two real variables x and y, then the partial derivatives of u with respect to x and y
are defined
u u(x + x, y) u(x, y) u u(x, y + y) u(x, y)
= lim , = lim ,
x x0 x y y0 y

provided these limits exist. The partial derivative operator is used to indicate
x
a differentiation with respect to x holding all other variables constant during the
differentiation processes. So treat the partial derivative operator just like an ordinary
derivative, except all other variables are held constant during the differentiation with

respect to x. Similarly, the partial differential operator is just like an ordinary
y
derivative with respect to y while holding all other variables constant during the
differentiation with respect to y.
u u
Example 2-48. If u = u(x, y) = x3 y2 sin y + cos x, then find and
x y
Solution
Treating y as a constant one finds
u
= (x3 y 2 sin y + cos x)
x x
(x3 y 2 ) (sin y) (cos x)
= +
x x x
3
u (x ) (cos x)
=y 2 0+ If y constant, then sin y is constant.
x x x
=y 2 (3x2 ) sin x
In a similar fashion, if x is held constant, then
u
= (x3 y 2 sin y + cos x)
y y
u (x3 y 2 ) (sin y) (cos x)
= +
y y y y
2
(y ) (sin y)
=x3 + 0 If x is constant, cos x is constant.
y y
u
=x3 (2y) cos y
y
 159
Higher partial derivatives are defined as a derivative of a lower ordered derivative.
For example, The second partial derivatives of u = u(x, y) are defined
2u 2u
   
u u
= , =
x2 x x y 2 y y

2u 2u
   
u u
The second derivatives = , = are called
x y x y y x y x
mixed partial derivatives.If both the function u = u(x, y) and
its first ordered partial
derivatives are continuous functions, then the mixed partial derivatives are equal to
one another, in which case it doesnt matter as to the order of the differentiation
2u 2u
and consequently = .
x y y x
Total Differential
If u = u(x, y) is a continuous function of two variables, then as x and y change,
the change in u is written

u = u(x + x, y + y) u(x, y)

Add and subtract the term u(x, y + y) to the change in u and write

u = [u(x + x, y + y] u(x, y + y)] + [u(x, y + y) u(x, y)]

which can also be expressed in the form

   
u(x + x, y + y) u(x, y + y) u(x, y + y) u(x, y)
u = x + y (2.108)
x y

Now use the mean-value theorem on the terms in brackets to show

 
u(x + x, y + y) u(x, y + y) u
= + 1
x x
  (2.109)
u(x, y + y) u(x, y) u
= + 2
y y

where 1 approaches zero as x 0 and 2 approaches zero as y 0. One can then

express the change in u as
u u
u = x + y + 1 x + 2 y (2.110)
x y

Define the differentials dx = x and dy = y and write the change in u as

u u
u = dx + dy + 1 dx + 2 dy (2.111)
x y
160
Define the total differential of u as
u u
du = dx + dy (2.112)
x y

and note the total differential du differs from u by an infinitesimal of higher order
than dx or dy because 1 and 2 approach zero as x 0 and y 0. The total
differential du, given by equation (2.112) is sometimes called the principal part in
the change in u.
Notation
Partial derivatives are sometimes expressed using a subscript notation. Some
examples of this notation are the following.
3u
=uxxx
2u x3
=uxx
u x2 3u
=ux =uxxy
x 2u x2 y
=uxy
u x y 3u
=uy =uxyy
y 2u xy 2
=uyy
y 2 3u
=uyyy
y 3

In general, if f = f (x, y) is a function of x and y and m = i + j is an integer, then

mf
is the representation of a mixed partial derivative of f .
xi y j
Differential Operator
If u = u(x, y) is a function of two variables, then the differential of u is defined
u u
du = dx + dy (2.113)
x y

and if the variables x = x(t) and y = y(t) are functions of t, then u becomes a function
of t with derivative
du u dx u dy
= + (2.114)
dt x dt y dt
This is obtained by dividing both sides of equation (2.113) by dt. One can think of
equation (2.114) as defining the differential operator
d[ ] [ ] dx [ ] dy
= + (2.115)
dt x dt y dt

where the quantity to be substituted inside the brackets can be any function of x
and y where both x and y are functions of another variable t.
 161
By definition a second derivative is the derivative of a first derivative and so one
can write
d2 u
     
d du d u dx u dy d u dx d u dy
2
= = + = + (2.116)
dt dt dt dt x dt y dt dt x dt dt y dt

since the derivative of a sum is the sum of derivatives. The quantities inside the
parentheses represents a product of functions which can be differentiated using the
product rule for differentiation. Applying the product rule one obtains
d2 u u d dx dx d u
   
u d dy dy d u
= + + +
dt2 x dt dt dt dt x y dt dt dt dt y
2 2
  2
  (2.117)
d u u d x dx d u u d y dy d u
= + + +
dt x dt2 dt dt x y dt2 dt dt y

The equation (2.115) tells us how to differentiate the terms inside the brackets. Here
u u
both and are some functions of x and y and so using the equation (2.115) one
x y
finds
d2 u u d2 x
     
dx u dx u dy
= + +
dt2 x dt2 dt x x dt y x dt
u d2 y
     
dy u dx u dy
+ + +
y dt2 dt x y dt y y dt
(2.118)
d2 u u d2 x
 2 2

dx u dx u dy
= + +
dt2 x dt2 dt x2 dt x y dt
u d2 y dy u dx 2 u dy
 2 
+ + + 2
y dt2 dt y x dt y dt
Functions of more than two variables are treated in a similar fashion.
Maxima and Minima for Functions of Two Variables
Given that f = f (x, y) and its partial derivatives fx and fy are all continuous
and well defined in some domain D of the x, y-plane. For R > 0, the set of points
N = { (x, y) | (x x0 )2 + (y y0 )2 R2 } is called a neighborhood of the fixed point
(x0 , y0 ), where it is assumed that the point (x0 , y0 ) and the neighborhood N are in the
domain D. The function f = f (x, y) can be thought of as defining a surface S over
the domain D with the set of points S = { (x, y, f ) | x, y D, and f = f (x, y) } defining
the surface. The function f is said to have
a relative or local minimum value at (x0 , y0 ) if f (x, y) f (x0 , y0 ) for (x, y) N.
a relative or local maximum value at (x0 , y0 ) if f (x, y) f (x0 , y0 ) for (x, y) N.
Determining relative maximum and minimum values of a function of two variables
can be examined by reducing the problem to a one dimensional problem. Note that
162
the planes x = x0 = a constant and y = y0 = a constant cut the surface f = f (x, y) in
one-dimensional curves. One can examine these one-dimensional curves for local
maximum and minimum values. For example, consider the curves defined by

Cx = { (x, y0 , f ) | f = f (x, y0 ) } Cy = { (x0 , y, f ) | f = f (x0 , y)}

These curves have tangent lines with the slope of the tangent line to the curve Cx
given by f
x
= fx (x, y0 ) and the slope of the tangent line to the curve Cy given
y=y0
f
by y
= fy (x0 , y). At a local maximum or minimum value these slopes must be
x=x0
zero. Consequently, one can say that a necessary condition for the point (x0 , y0 ) to
corresponds to a local maximum or minimum value for f is that the conditions
f f
= fx (x0 , y0 ) = 0 and = fy (x0 , y0 ) = 0 simultaneously.
x (x0 ,y0 ) y (x0 ,y0 )

These are necessary conditions for an extreme value but they are not sufficient
conditions. The problem of determining a sufficient condition for an extreme value
will be considered in a later chapter and it will be shown that

If the function f = f (x, y) and its derivatives fx , fy , fxx, fxy , fyy exist and are
continuous at the point (x0 , y0 ), then for f = f (x, y) to have an extreme value at
the point (x0 , y0 ) the conditions
f f
= fx (x0 , y0 ) = 0 and = fy (x0 , y0 ) = 0
x (x0 ,y0 ) y (x0 ,y0 )

together with the condition fxx (x0 , y0 )fyy (x0 , y0 ) [fxy (x0 , y0 )]2 > 0 that must be
satisfied. One can then say
f (x0 , y0 ) is a relative maximum value if fxx (x0 , y0 ) < 0
f (x0 , y0 ) is a relative minimum value if fxx (x0 , y0 ) > 0

Implicit Differentiation
If F (x, y, . . . , z) is a continuous function of nvariables with continuous partial
derivatives, then the total differential of F is given by
F F F
dF = dx + dy + + dz (2.119)
x y z
 163
In two dimensions, if F (x, y) = 0 is an implicit function defining y as a function
of x, then by taking the total differential one obtains
F F
dF = dx + dy = 0
x y

dy
and solving for the derivative is calculated as
dx
dy F F Fx
= / = ,
dx x y Fy

provided that Fy = Fy
= 0.
In three dimensions, if F (x, y, z) = 0, is an implicit function of three variables
which defines z as a function of x and y, then one can write the total differential as
F F F
dF = dx + dy + dx = 0 (2.120)
x y z

Solving for dz in the equation (2.120) one finds

       
F F F F Fx Fy
dz = / dx + / dy = dx + dy (2.121)
x z y z Fz Fz

Note that if the implicit function F (x, y, z) = 0 defines z as a function of independent

variables x and y, then one can write z = z(x, y) and the total differential of z would
be given by
z z
dz = dx + dy (2.122)
x y
Comparing the equations (2.122) and (2.121), there results the relations
z F F Fx z F F Fy
= / = and = / = (2.123)
x x z Fz y y z Fz

provided that F
z

= 0.
Given an implicit equation F (x, y, z) = 0, one could assume one of the following.
(a) x and y are independent variables.
(b) x and z are independent variables.
(c) y and z are independent variables.
The derivatives in these various cases all give results similar to equations (2.123)
and (2.122) derived above.
164
Exercises

2-1.
(a) Sketch the curve y = x2
(b) Find the equation of the tangent line to this curve which passes through the
point (2, 4).
(c) Find the equation of the tangent line to this curve which passes through the
point (2, 4)
(d) Find the equation of the tangent line to this curve which passes through the
point (0, 0)

2-2. Find the derivatives of the following functions.

1 
5 2
(a) y = 3x + 2x x + 4 (d) y = (g) y = x2 + 1
ax2 + bx + c
1 (h) y = (2x + 1)3 (3x2 x)2
(b) y = 4x3 x2 + x + 1 (e) y = 2
3x x + 1  2
x1
(c) y = ax2 + bx + c 1 (i) y =
(f ) y = 2 x+1
x +x

2-3. Find the derivatives of the given functions.

1 1 (g) y = 3 + 4t + 5t2
(a) y = (d) y = (j) y = x ln x
x3 x 2x2 + x
(h) y =
(b) y = 3 x
3
(e) y = t2 x+1 (k) y = ex ln x
1 1
(c) y = x3/2 (f ) y = + t (i) y = x + (l) y = n a + x
t 2t

2-4. Find the derivatives of the given functions.

3 4
(a) y = + 2 (d) y = (x2 + 4)3 (g) y = e3x (j) y = tan1 ( tanh 3x)
x x
x+1 1
x (e) y = (h) y = ln(3x) (k) y = sin1
(b) y = x 1 + x2
1+x
(i) y = esin x
 1
(c) y = x(1 + x) (f ) y = x x3 x2 (l) y = tan (ln x)

2-5. Find the derivatives of the following functions.

(a) A = r 2 (d) y = sin(3) (g) y = sin2 (3) (j) y = sin(a + bxn )
4
(b) V = r 3 (e) y = cos(3) (h) y = cos2 (3) (k) y = cos1 (ax bx2 )
3
(c) S = 4r 2 (f ) y = tan(3) (i) y = tan2 (3) (l) y = xx + x1/x
 165

2-6. Find the derivatives of the following functions.


(a) y = 3x
2 1 x2
(d) y = e3x
3
(g) y = 1 + sin t (j) y =
1 x  1 + x2
(b) y = tan1 (e) y = sin(ex ) (h) y = t2 3t

a a (k) y = sec(ln( a + bx + cx2 ))

(c) y = e3x (f ) y = cos(tan x) (i) y = e t x
(l) y = ex

dy

2-7. Find the derivative dx
= y
(x) if y = y(x) is defined by the equation

(a) x2 + y 2 = 24 (d) x3 + y 3 = 6x (g) x2 y 2 = 1 (j) xy = (x + y)3

 
x2 y2 y
(b) x3 y 2 = 1 (e) x3 + y 3 = 6y (h) + =1 (k) y = x ln
a2 b2 1+x
(c) xy + x3 y 3 = 1 (f ) x3 + y 3 = 6xy (i) x2 = 4py (l) y = xy

dy

2-8. Find the derivative dx
associated with the given functions.

(a) y = sin1 (3x) (d) y = sin1 (x2 ) (g) y = (1 + x) sin1 x (j) y = (cos 3x)x
 x
1
(b) y = cos1 (3x) (e) y = cos1 (x2 ) (h) y = (1 + x) cos1 x (k) y = 1 +
x
(c) y = tan1 (3x) (f ) y = tan1 (x2 ) (i) y = (1 + x) tan1 x (l) xy = x2 y 3 + x + 3

2-9. Differentiate the given functions.

1
(a) y = ln(3x) (d) y = (3 + x)x , x > 0 (g) y = ln(ax + b) (j) y =
2
x +x
(b) y = ln(ax) (e) (3 + x)x
2
(h) y = eax+b 1
(k) y = x + 2
x
 
x 3
(c) y = ln
x+1 (f ) (3 + x)x (i) y = xx
(l) y = x sin3 (4x)

2-10. Find the derivative of the given functions.

(a) y = sin1 (ax2 ) (d) y = sinh (3x) (g) y = sinh 1 (3x) a + bx + c2

(j) y =
x+1
(b) y = cos1 (ax2 ) (e) y = cosh (3x) (h) y = cosh 1 (3x) (k) y = ln( cosh 3x)
(c) y = tan1 (ax2 ) (f ) y = tanh (3x) (i) y = tanh 1 (3x) (l) y = sin1 x3

2-11. Show the derivative of a function f (x) at a fixed point x0 can be written
f (x) f (x0 )
lim = f
(x0 ) Hint: Make a substitution.
xx0 x x0

2-12. Use the quotient rule to differentiate the functions

cos x 1 1
(a) y = cot x = , (b) y = sec x = , (c) y = csc x =
sin x cos x sin x
166

2-13. Find the first and second derivatives of the following functions.
1 3t x
(a) y= e (c) y = te3t (e) y=
t (x a)(x b)
1 1
(b) y= 1 x2 (d) y = x 4 + 3 sin x (f ) y= 2
x x + a2
dy d2 y

2-14. Find the first derivative dx
and second derivative dx2
associated with the
given parametric curve.
(a) x = a cos t, y = b sin t (d) x = at, y = bt2
(b) x = 4 cos t, y = 4 sin t (e) x = a cosh t, y = b sinh t
(c) x = 3t2 , y = 2t (f ) x = sin(3t + 4), y = cos(5t + 2)

2-15. Differentiate the given functions.

(a) y = x sin(4x2 ) (d) y = cos(4x2 ) (g) y = ln(3x + 4) sin(x2 )
(b) y = x2 e3x (e) y = x2 ln(3x), x > 0 (h) y = ln(x2 + x) cos(x2 )

(c) y = xex (f ) y = x2 tan(3x) (i) y = tan x sec x

2-16. Differentiate the given functions.

sin x x
(a) y = (d) y =
x x+1 (g) y = x2 ln(x2 )
cos x x
(b) y = (e) y = (h) y = sin(x2 ) ln(x3 )
(1 + x2 )3/2
x
x+1 (1 + x2 )3/2 (i) y = sin(x2 ) cos(x2 )
(c) y = (f ) y =
x x

2-17. Find the first and second derivatives of the given functions.
1 4
(a) y = x + (d) y = x2 + 2x + 3 + (g) y = x sin x
x x
2
(b) y = sin (3x) (e) y = cos2 (3x) (h) y = x2 cos x
x
(c) y = (f ) y = tan(2x) (i) y = x cos(x2 )
x+1

2-18. Find the tangent line to the given curve at the specified point.
1
(a) y = x + , at (1, 2) (b) y = sin x, at (/4, 1/ 2) (c) y = x3 , at (2, 8)
x

2-19. Show that Rolles theorem can be applied to the given functions. Find all
values x = c such that Rolles theorem is satisfied.
1
(a) f (x) = 2 + sin(2x), x [0, 1] (b) f (x) = x + , x [1/2, 2]
x
 167

2-20. Sketch the given curves and where appropriate describe the domain of the
function, symmetry properties, x and yintercepts, asymptotes, relative maximum
and minimum points, points of inflection and how the concavity changes.
1
(a) y =x+ (c) y = x4 6x2 (e) y = sin x + cos x
x
x
(b) y= (d) y = (x 1)2 (x 4) (f ) y = x4 + 12x3 + 1
x+1
f (x + h) f (x)

2-21. If f
(x) = lim
h0 h


f (x + 2h) 2f (x + h) + f (x)
(a) Show that f (x) = lim
h0 h2
(b) Substitute f (x) = x3 into the result from part (a) and show both sides of the
equation give the same result.
(c) Substitute f (x) = cos x into the result from part (a) and show both sides of the
equation give the same result. Hint: Use LHopitals rule with respect to the
variable h.

2-22. Find the local maximum and minimum values associated with the given
curves.
(a) y = x2 4x + 3 (g) y = sin(2x), all x
(d) y = 5 48x + x3
2
x x+1
(b) y = (e) y = sin x, all x (h) y = cos(2x), all x
x2 + 1
2 3
(c) y = 2 (f ) y = cos x, all x (i) y = , x [16, 16]
x +4 5 4 cos x

2-23. Find the absolute maximum and absolute minimum value of the given
functions over the domain D.
2
(a) y = f (x) = x2 + , D = { x | x [1/2, 2] }
x
x
(b) y = f (x) = , D = { x | x [1, 2] }
x+1
(c) y = f (x) = sin x + cos x, D = { x | x [0, 2] }
x
(d) y = f (x) = , D = { x | x [2, 2] }
1 + x2

2-24. Show that the given functions satisfy the conditions of the mean-value
theorem. Find all values x = c such that the mean-value theorem is satisfied.
(a) f (x) = 4 + (x 2)2 , x [2, 6] (b) f (x) = 4 x2 , x [0, 2]

2-25. A wire of length
= 4 + is to be cut into two parts. One part is bent into
the shape of a square and the other part is bent into the shape of a circle. Determine
how to cut the wire so that the area of the square plus the area of the circle has a
minimum value?
168


2-26. A wire of length
= 9 + 4 3 is to be cut into two parts. One part is bent
into the shape of a square and the other part is bent into the shape of an equilateral
triangle. Show how the wire is to be cut if the area of the square plus the area of
the triangle is to have a minimum value?


2-27. A wire of length
= 9 + 3 is to be cut into two parts. One part is bent into
the shape of an equilateral triangle and the other part is bent into the shape of a
circle. Show how the wire is to be cut if the area of the triangle plus the area of the
circle is to have a minimum value?

2-28. Find the critical values and determine if the critical values correspond to a
maximum value, minimum value or neither.
(a) y = (x 1)(x 2)2 (c) y = f (x), where f
(x) = x(x 1)2 (x 3)3
x2 7x + 10
(b) y =
x 10
(d) y = f (x), where f
(x) = x2 (x 1)2 (x 3)

2-29. Evaluate the given limits.

xn 1
(a) lim 1+x1 ax 1
x1 x 1 (d) lim (g) lim
x0 x x0 bx 1
 x
1 ln x sin x
(b) lim 1 + (e) lim 2 (h) lim
x x x1 x 1 x1 x 1
2
ax + bx ex ex x sin x
(c) lim 2 (f ) lim (i) lim
x0 bx + ax x0 sin x x0 x2

2-30. Determine where the graph of the given functions are (a) increasing and
(b) decreasing. Sketch the graph.

(a) y = 8 10x + x2 (b) y = 3x2 5x + 2 (c) y = (x2 1)2

2-31. Verify the Leibnitz differentiation rule for the nth derivative of a product
of two functions, for the cases n = 1, n = 2, n = 3 and n = 4.
n  
n n
 n
Dniu Di v

D [u(x)v(x)] = D [uv] =
i
i=0
       
n n n n n1
n n2
2 n
D [uv] = (D u) v + D u Dv + D u D v ++ uDn v
0 1 2 n
 
n n!
where = are the binomial coefficients. The general case can be
m m!(n m)!
proven using mathematical induction.
 169

2-32. Use Eulers formula and show
(a) ei (+2n) = ei where n is an integer.
(b) The polar form of the complex number x + i y = rei
(c) Show de Moivres theorem can be expressed (rei )n = rn ei n

u u 2 u 2 u 2u

2-33. Find the partial derivatives , , , ,
x y x2 x y y 2

(a) u = x2 y + xy 3

(c) u = x2 + y 2 (e) u = xyexy
(b) u = (x2 + y 2 )3

(d) u = x2 y 2 (f ) u = xey + yex

2-34.

(a) Show cosh 1 x = ln(x

+  x2 1),  x 1
1 1
(b) Show sech 1 x = ln + 2
1 , 0< x< 1
x x
1 1
(c) Show sinh (x) = sinh x
d 1
(d) Show cosh 1 x = , x>1
dx 2 x 1

d d2 dn dn f (x)

2-35. Define the operators D = , D2 = 2 , . . . , Dn = n , with Dn f (x) =
dx dx dx dxn
representing the nth derivative of f (x). Find a formula for the indicated derivatives.

(a) Dn (ex ) (d) Dm (xn ), m<n (g) Dn (sin3 x) (j) Dn (cos(ax + b))
1 (k) Dn (ln(x + a))
(b) Dn (ax ) (e) Dn (sin x) (h) Dn ( 2 )
x a2 1
(c) Dn (ln x), x>0 (f ) Dn (cos x) (i) Dn (sin(ax + b)) (l) Dn ( )
x+a

dy d2 y

2-36. Find the first and second derivatives and if x2 y + y2 x = 1
dx dx2

2 2 2

2-37. Find the partial derivatives , , , ,
x y x2 x y y 2
b
(a) = x3 + yx2 3axy (d) = ax + + cxy
y
2 2
(b) = x + y + xy (e) = ax + by + cxy + dx2 y
x2 y2
(c) = sin(xy) (f ) = 3 +4
y x

x2 y2

2-38. Sketch the ellipse + = 1 and then find the tangent lines to this ellipse
4 9
at the following points (a) (0, 3) (b) ( 3, 3/2) (c) (2, 0) (d) ( 3, 3/2) (e) (0, 3)
170

2-39. The area of a circle is given by A = r2 . If the radius r = r(t) changes with
time, then how does the area change with time?

2-40. If s describes the displacement of a particle from some fixed point, as mea-
sured along a straight line, and s = s(t) is a function of time t, then the velocity v
of the particle is given by the change in the displacement with respect to time or
ds
v = v(t) = . The acceleration a of the particle is defined as the rate of change
dt
dv d2 s
of the velocity with respect to time and so one can write a = a(t) = = 2. If
dt dt
4 2
t 11t
s = s(t) = 2t3 + 6t, find the velocity and acceleration as a function of time.
4 2
Find where s increases and decreases.

2-41.
Let z = z(x, y) denote a function representing a surface
in three-dimensional space. Let P denote a point on this
surface with coordinates (x0 , y0 , z(x0 , y0 )).
z z(x0 + x, y0 ) z(x0 , y0 )
(a) If = lim and
x (x0 ,y0 ) x0 x
z z(x0 , y0 + y) z(x0 , y0 )
= lim
y (x0 ,y0 ) y0 y
z z
are the partial derivatives and evaluated
at the point P , then what is the
x y
geometric interpretation of these partial derivatives?
(b) Let the planes x = x0 = a constant, and y = y0 = a constant, intersect the surface
z = z(x, y) in curves C1 and C2 as illustrated. Find the equations of the tangent
lines AA and BB to the curves C1 and C2 at their point of intersection P .
d u du

2-42. Derive the absolute value rule | u |= , where u = u(x) is a function
dx | u | dx
of x and test this rule using the function u = u(x) = x. Hint: |u| = u2

2-43. Determine the sign of the slope to the left and right of the given critical
point.
 171

2-44.
(a) Semi-log graph paper has two perpendicular axes with a logarithmic scale on
one axis and an ordinary scale on the other axis. Show curves of the form
y = x , > 0, > 0 are straight lines when plotted on semi-log graph paper.
(b) Log-Log graph paper has two perpendicular axes with a logarithmic scale on
both axes. Show curves of the form y = x , > 0 are straight lines when
plotted on log-log graph paper.

2-45.
Let s denote the distance between a fixed point
(x0 , y0 ) and an arbitrary point (x, y) lying on the line
ax + by + c = 0
(a) Show that the quantity s2 is a minimum when
the line through the points (x0 , y0 ) and (x, y) is per-
pendicular to the line ax + by + c = 0.
(b) Show the minimum distance d from the point
|ax0 + by0 + c|
(x0 , y0 ) to the line ax + by + c = 0 is given by d =
a2 + b2

2-46.
If r = f () is the polar equation of a curve, then this
curve can be represented in cartesian coordinates as a set
of parametric equations

x = r cos = f () cos y = r sin = f () sin

Show the slope of the curve can be represented

dy f
() sin + f () cos
=

dx f () cos f () sin
172

2-47. The volume of a sphere is given by V = 43 r3 . If the radius r = r(t) changes
with time, then how does the volume change with time?

2-48. A pool is constructed 15 meters long, 8 meters wide and 4 meters deep.
When completed, water is pumped into the pool at the rate of 2 cubic meters per
minute.
(i) At what rate is the water level rising?
(ii) How long does it take to fill the pool?

2-49.
A box having a lid is to be constructed from a
square piece of cardboard having sides of length
.
The box is to be constructed by cutting squares with
sides x from two corners and then cutting rectangles
with sides of length x and y from the opposite corners
as illustrated in the figure. The sides are folded up
and the lid folded over with the sides to be taped.
Find the dimensions of the box having the largest volume.

2-50.
Determine the right circular cone of maximum volume
that can be inscribed inside a given sphere having a radius R.
The situation is illustrated in the figure where
AC = r = base radius of cone. 0B = R = radius of sphere.
AB = h = altitude of cone. 0C = R = radius of sphere.

x

2-51. Sketch the function y = + sin x and determine where the maximum and
2
minimum values are.

2-52. A cylindrical can is to be made such that it

has a fixed volume V . Show that the can using the
least amount of material has its height equal to its
diameter.
 173

2-53. Consider a rectangle inscribed inside a circle of radius R. Find the dimen-
sions of the rectangle with maximum perimeter.

2-54. Find the first and second derivatives if

(a) y = cos(x3 + x) x
(c) y = (e) y = tan(3x)
1 x2 + 1
(b) y =x2 + 3 (d) y = sin(x2 + x) (f ) y =e3x + cosh (2x)
x
dy d2 y

2-55. Find the first and second derivatives , if
dx dx2
(a) x = sin(2t), y = cos(2t) (c) y =t2 sin(2t)e3t (e) y = sin u, u = x2 + x
(b) x =t2 , y = t3 (d) y =u3 , u = x2 + x (f ) y = sinh (3x2 ) + cosh (3x)

2-56. Sketch the curve y = x29+3

(a) Find regions where the slope is positive.
(b) Find regions where the slope is negative.
(c) Find where the slope is zero.
(d) Find regions where the curve is concave up.
(e) Find regions where the curve is concave down.

2-57. Find the first and second derivatives of the given functions.

(a) y = sin1 (3x) (c) y = tan1 ( x) (e) y = sec1 (3x2 )
x
(b) y = cos1 (1 x2 ) (d) y = cot1 (x2 + x) (f ) y = csc1 ( )
3

2-58. Find the derivatives of the given functions.

(a) y = ln(x + 1 + x2 ) (c) y = x2 + x ln |x + | (e) y = sin(3x) cos(2x)
1
(b) y = sin2 (ex ) (d) y = cos x esin x (f ) y = tan x
x

2-59. Use derivative information to sketch the curve over the domain specified.
(a) y = 1 + 3x2 x3 for 1 x 4
(b) y = 1 + (x 1)3 (x 5) for 1 y 6

2-60. Determine the following limits

xm m sin mx
(a) lim (b) lim (1 + ax)1/x (c) lim
x xn n x0 x0 sin nx

2-61. Given the function F (x, y, z) = ax2 +by2 +cxyzd = 0, where a, b, c, d are nonzero
constants. Assume x and y are independent variables and find the following partial
z z
derivatives. (a) (b)
x y
174

2-62. Given the function F (x, y, z) = ax2 +by2 +cxyzd = 0, where a, b, c, d are nonzero
constants. Assume x and z are independent variables and find the following partial
y y
derivatives. (a) (b)
x z

2-63. Given the function F (x, y, z) = ax2 +by2 +cxyzd = 0, where a, b, c, d are nonzero
constants. Assume y and z are independent variables and find the following partial
x x
derivatives. (a) (b)
y z

2-64.
The Heaviside15 step function H() is defined

0, <0
H() =
1, >0
The figure illustrates the step function H(x x0 )
(i) Sketch the functions y1 (x) = H(x x0 ) and y2 (x) = H(x (x0 + ))
(ii) Sketch the function y(x) = y1 (x) y2 (x)
y (x) y2 (x)
(ii) Define the Dirac Delta function (x x0 ) = lim 1 and give a physical
0 
interpretation as to the meaning of this function.

2-65.

The crank arm 0P , of length r (cm), revolves with constant angular velocity
(radians/sec). The connecting rod P Q, of length
(cm), moves the point Q back
and forth driving a piston. Show that point Q has the velocity (cm/sec), given by
dS r 2 sin t cos t
= V = r sin t  Hint: Use law of cosines
dt
2 r 2 sin2 t


2-66. Find the global maximum of the function y(x) = x
x, 0 < x < 10 illustrated.

15
Oliver Heaviside (1850-1925) An English engineer and mathematician.
 175
Chapter 3
Integral Calculus
The integral calculus is closely related to the differential calculus presented in
the previous chapter. One of the fundamental uses for the integral calculus is the
construction of methods for finding areas, arc lengths, surface areas and volumes
associated with plane curves and solid figures. Many of the applications of the dif-
ferential and integral calculus are also to be found in selected areas of engineering,
physics, business, chemistry and the health sciences. These application areas re-
quire additional background material and so investigation into these applied areas
are presented in a later chapter after certain fundamental concepts are developed.
Various concepts related to the integral calculus requires some preliminary back-
ground material concerning summations.
Summations


The mathematical symbol (Greek letter sigma) is used to denote a summation
of terms. If f = f (x) is a function whose domain contains all the integers and m is
an integer, then the notation
m


f (j) = f (1) + f (2) + f (3) + + f (m) (3.1)
j=1

is used to denote the summation of the terms f (j) as j varies from 1 to m. Here j = 1
is called the starting index for the sum and the m above the sigma sign is used to
denote the ending index for the sum. The quantity j is called the dummy summation
index because the letter j does not occur in the answer and j can be replaced by
some other index if one desires to do so.
The following are some examples illustrating how the summation notation is
employed.
(a) If m, n are integers satisfying 1 < m < n, then a summation from 1 to n of the
f (j) terms can be broken up and written as a sum of m terms followed by a
summation of (n m) terms by writing
n

m

n


f (j) = f (j) + f (j) (3.2)
j=1 j=1 j=m+1

(b) The summation index can be shifted to represent summations in different forms.
n


For example, the representation S = (j 1)2 can be modified by making the
j=1
176
substitution k = j 1 and noting that when j = 1, then k = 0 and when j = n,
n1


then k = n 1, so that the sum S can also be expressed S = k2 .
k=0
n

nm


As another example, the sum f (j) can be expressed in the form f (m+k).
j=m+1 k=1
This is called shifting the summation index. This result is obtained by making the
substitution j = m + k and then finding the summation range for the index k. For
example, when j = m + 1, then k = 1 and when j = n, then k = n m giving the above
result.
(c) If c1 , c2 are constants and f (x) and g(x) are functions, then one can write
n

n

n


(c1 f (k) + c2 g(k)) = c1 f (k) + c2 g(k) (3.3)
k=1 k=1 k=1

where the constant terms can be placed in front of the summation signs.
(d) If f (x) = c = constant, for all values of x, then
n


f (j) =f (m) + f (m + 1) + f (m + 2) + + f (n)
j=m

= c + c + c + + c (3.4)
  
(nm+1) values of c

=c (n m + 1)

n


The special sum 1 = 1+1+1++1 = n
  
occurs quite often.
j=1 n ones


n


(e) The notation f (j) is used to denote the limiting process lim
n
f (j) if this
j=1 j=1
limit exists. These summations are sometimes referred to as infinite sums.
(f) Summations can be combined. For example,
m2

m


f (j) + f (m 1) + f (m) = f (j) (3.5)
j=1 j=1

Special Sums
n1

n1


If f (x) = a + xd, then the sum S = f (j) = (a + jd) or
j=0 j=0

n1


S= f (j) = a + (a + d) + (a + 2d) + (a + 3d) + + (a + (n 1)d) (3.6)
j=0
 177
is known as an arithmetic series with a called the first term, d called the common
difference between successive terms,
= a + (n 1)d is the last term and n is the
number of terms. Reverse the order of the terms in equation (3.6) and write

S = (a + (n 1)d) + (a + (n 2)d) + + (a + d) + a (3.7)

and then add the equations (3.6) and (3.7) on a term by term basis to show

2S = n[a + a + (n 1)d] = n(a +
) (3.8)

Solving equation (3.8) for S one finds the sum of an arithmetic series is given by
n1

n a+

S= (a + jd) = (a +
) = n (3.9)
j=0
2 2

which says the sum of an arithmetic series is given by the number of terms multiplied
by the average of the first and last terms of the sum.
n1

n1


If f (x) = arx , with a and r nonzero constants, the sum S = f (j) = ar j or
j=0 j=0

n1

n1


S= f (j) = a rj = a + ar + ar2 + ar3 + + arn1 (3.10)
j=0 j=0

is known as a geometric series, where a is the first term of the sum, r is the common
ratio of successive terms and n is the number of terms in the summation. Multiply
equation (3.10) by r to obtain

rS = ar + ar2 + ar3 + + arn1 + arn (3.11)

and then subtract equation (3.11) from equation (3.10) to show

a arn
(1 r)S = a arn or S= (3.12)
1r

If |r| < 1, then rn 0 as n and in this special case one can write
n1

a arn a
S = lim arj = lim = , |r| < 1 (3.13)
n
j=0
n 1r 1r
178
Archimedes of Syracuse, (287-212 BCE), used infinite summation processes to find
the areas under plane curves and to find the volume of solids. In addition to the
arithmetic and geometric series, Archimedes knew the following special sums
n


1 =1 + 1 +1 + + 1 = n
  
j=1 n terms
n

1
j =1 + 2 + 3 + + n = n(n + 1)
2
j=1
n
(3.14)

1
2 2 2 2 2
j =1 + 2 + 3 + + n = (2n3 + 3n2 + n)
6
j=1
n

1 4
j 3 =13 + 23 + 33 + + n3 = (n + 2n3 + n2 )
j=1
4

Modern day mathematicians now know how to generalize these results to obtain
sums of the form n

Sn = j p = 1p + 2p + 3p + + np (3.15)
j=1

where p is any positive integer. They have found that the sum Sn of the series given
by equation (3.15) must be a polynomial of degree p + 1 of the form

Sn = a0 np+1 + a1 np + a2 np1 + + ap n (3.16)

where n is the number of terms in the series and a0 , a1 , a2 , . . . , ap are constants to be

determined. A more general representation for the sum (3.15) can be found on page
349.
n


Example 3-1. Sum the series j 4 = 14 + 24 + 34 + 44 + + n4
j=1
SolutionLet Sn denote the sum of the series and make use of the fact that Sn
must be a polynomial of degree 5 having the form

Sn = a0 n5 + a1 n4 + a2 n3 + a3 n2 + a4 n (3.17)

where a0 , a1 , a2 , a3 , a4 are constants to be determined and n is the number of terms to

be summed. The sums S1 , S2 , S3 , S4 , S5 give the five conditions
 179
1


S1 = j 4 = 14 = 1
j=1
2


S2 = j 4 = 14 + 24 = 1 + 16 = 17
j=1
3


S3 = j 4 = 14 + 24 + 34 = 1 + 16 + 81 = 98
j=1
4


S4 = j 4 = 14 + 24 + 34 + 44 = 1 + 16 + 81 + 256 = 354
j=1
5


S5 = j 4 = 14 + 24 + 34 + 44 + 54 = 1 + 16 + 81 + 256 + 625 = 979
j=1

to determine the constants a0 , a1 , a2 , a3 , a4. That is, if Sn has the form given by
equation (3.17), then

S1 = a0 + a1 + a2 + a3 + a4 =1
S2 = a0 (2)5 + a1 (2)4 + a2 (2)3 + a3 (2)2 + a4 (2) =17
S3 = a0 (3)5 + a1 (3)4 + a2 (3)3 + a3 (3)2 + a4 (3) =98 (3.18)

S4 = a0 (4)5 + a1 (4)4 + a2 (4)3 + a3 (4)2 + a4 (4) =354

S5 = a0 (5)5 + a1 (5)4 + a2 (5)3 + a3 (5)2 + a4 (5) =979

The equations (3.18) represent 5-equations in 5-unknowns which can be solved using
algebra. After a lot of work one finds the solutions
1 6 1 15 1 10 1
a0 = = , a1 = = , a2 = = , a3 = 0, a4 =
5 30 2 30 3 30 30
n

1
This gives the result Sn = j 4 = 14 + 24 + 34 + + n4 = (6n5 + 15n4 + 10n3 n)
j=1
30

Integration
The mathematical process which represents the inverse of differentiation is known
d
as integration. In the differential calculus the differential operator , which per-
dx
formed differentiation, was employed as a shorthand notation
 for the limiting process
required for differentiation. Define the integral symbol ( ) dx as an operator that
performs the inverse of differentiation which is called integration.
180

Figure 3-1. Differential and integral operators.

Examine the operator boxes illustrated in the figure 3-1 where one box represents
a differential operator and the other box represents an integral operator. If f (x) is
d
an input to the differential operator box, then the output is denoted f (x) = f
(x).
dx
Suppose it is required to undo what has just been done. To reverse the differentiation
process, insert the derivative function into the integral operator box. The output
from the integral operator box is called an indefinite integral and is written

f
(x) dx = f (x) + C (3.19)

and the equation (3.19) is sometimes read as The indefinite integral of f
(x) dx is
equal to f (x)+C . Here f
(x) is called the integrand, f (x) is called a particular integral
and f (x) + C is called the general integral of the indefinite integral of f
(x) dx and C
is called the constant of integration. Recall that two functions f (x) and f (x) + C , C
constant, both have the same derivative f
(x), this is because the derivative of a sum
is the sum of the derivatives and the derivative of a constant is zero. It is customary
when performing an indefinite integral to always add a constant of integration in
order to get the more general result.
Examine the notation for the inputs and outputs associated with the operator
d
boxes illustrated in the figure 3-1. One can state that if G(x) = g(x) then by
dx
definition one can express the indefinite integral in any of the forms
  
dG(x)
dx = G(x) + C or g(x) dx = G(x) + C, or dG(x) = G(x) + C (3.20)
dx

dG(x)
because G(x) + C is the more general function which has the derivative g(x) = .
dx
 
181
The symbol is called an integral sign and is sometimes replaced by the words,
The function whose differential is . The symbol x used in the indefinite integral
given by equation (3.20) is called a dummy variable of integration. It can be replaced
by some other symbol. For example,

d
if G() = g() then g() d = G() + C (3.21)
d

where C is called a constant of integration.

Example 3-2.
The following integrals occur quite often and should be memorized.
 
d
If x = 1, then 1 dx = x + C or dx = x + C
dx
 
d 2
If x = 2x, then 2x dx = x2 + C or d(x2 ) = x2 + C
dx
 
d 3
If x = 3x2 , then 3x2 dx = x3 + C or d(x3 ) = x3 + C
dx
 
d n
If x = nxn1 , then nxn1 n
dx = x + C or d(xn ) = xn + C
dx
 m+1     m+1 
d u um+1 u um+1
If = um , then um du = +C or d = +C
du m + 1 m+1 m+1 m+1
 
d
If sin t = cos t, then cos t dt = sin t + C or d(sin t) = sin t + C
dt
 
d
If cos t = sin t, then sin t dt = cos t + C or d(cos t) = cos t + C
dt

Properties of the Integral Operator


d
If f (x) dx = F (x) + C, then F (x) = f (x)
dx
That is, to check that the integration performed is accurate, observe that one must
have the derivative of the particular integral F (x) always equal to the integrand
function f (x).
If f (x) dx = F (x) + C, then
 
f (x) dx = f (x) dx = [F (x) + C] = F (x) + K

for all constants . Here K = C is just some new constant of integration. This
property is read, The integral of a constant times a function equals the constant
times the integral of the function.
182  
If f (x) dx = F (x) + C and g(x) dx = G(x) + C, then

  
[f (x) + g(x)] dx = f (x) dx + g(x) dx = F (x) + G(x) + C

This property states that the integral of a sum is the sum of the integrals. The
constants C in each of the above integrals are not the same constants. The symbol
C represents an arbitrary constant and all C
s are not the same. That is, the sum of
arbitrary constants is still an arbitrary constant. For example, examine the state-
ment that
 the integral of a sum is the sum of the integrals. If for i = 1, 2, . . ., m you
know fi (x) dx = Fi (x) + Ci , where each Ci is an arbitrary constant, then one could
add a constant of integration to each integral and write
   
[f1 (x) + f2 (x) + + fm (x)] dx = f1 (x) dx + f2 (x) dx + + fm (x) dx

=[F1 (x) + C1 ] + [F2 (x) + C2 ] + + [Fm (x) + Cm ]

=F1 (x) + F2 (x) + + Fm (x) + C

All the arbitrary constants of integration can be combined to form just one arbitrary
constant of integration.
Notation
There are different notations for representing an integral. For example, if
d
F (x) = f (x), then dF (x) = f (x) dx and dF (x) = f (x) dx = F (x) + C or
dx
  
d
f (x) dx = F (x) dx = dF (x) = F (x) + C (3.22)
dx

Examine equation (3.22) and observe dF (x) = F (x) + C . One can think of the

differential operator d and the integral operator as being inverse operators of each

other where the product of operators d produces unity. These operators are

commutative so that d also produces unity. For example,

 
d f (x) dx = d[F (x) + C] = d F (x) + d C = f (x) dx

Some additional examples of such integrals are the following.

 183
   
dv du
If d(uv) = u dv + v du, then u +v dx = d(uv) = uv + C
dx dx

du dv
 
u v du u dv

v u
  
u u
If d = , then dx dx dx = d = +C
2 2
v v v v v
 
dw dw
In general, if dw = dx, then dx = dw = w + C
dx dx

Integration of derivatives
 
d dy d2 y d
If = 2 , or (f
(x)) = f

(x), then multiplying both sides of this equa-
dx dx dx dx
tion by dx and integrating both sides of the equation gives
    2  
d dy d y d

dx = dx (f (x)) dx = f

(x) dx
dx dx dx2 dx
    2 or  
dy d y
d = dx d (f (x)) = f

(x) dx


dx dx2

Since dw = w + C , one finds
 2     
d y dy dy
dx = d = + C or f (x) dx = d(f
(x)) = f
(x) + C



(3.23)
dx2 dx dx
In a similar fashion one can demonstrate that in general
 
dn+1 y dn y
dx = +C or f (n+1) (x) dx = f (n) (x) + C (3.24)
dxn+1 dxn
for n = 1, 2, 3, . . ..
Polynomials 
xn+1
Use the result xn dx = + C obtained from example 3-2 to evaluate the
n+1
integral of a polynomial function
pn (x) = a0xn + a1 xn1 + + an2 x2 + an1 x + an

where a0 , a1 , . . . , an1, an are constants. Also use the result that the integral of a
sum is the sum of the integrals and the integral of a constant times a function is
that constant times the integral of a function. One can then integrate the given
polynomial function to obtain
 

pn (x) dx = a0 xn + a1 xn1 + + an2 x2 + an1 x + an dx
    
n n1 2
=a0 x dx + a1 x dx + + an2 x dx + an1 x dx + an dx

xn+1 xn x3 x2
=a0 + a1 + + an2 + an1 + an x + C
n+1 n 3 2
184
Example 3-3. Recall that if functions are scaled, then the chain rule for
differentiation is used to find the derivative of the scaled function. If you know
d d
F (x) = f (x), then you know F (u) = f (u), no matter what u is, so long as it
dx du
is different from zero and well behaved. Say for example that you are required to
differentiate the function y = F (ax), where a is a constant different from zero, then
you would use the chain rule for differentiation. Make the substitution u = ax and
write y = F (u), then
dy dy du d du
= = F (u) = f (u) a = f (ax) a
dx du dx du dx

In a similar fashion integrals must be adjusted

 when a scaling occurs. If you
know f (x) dx = F (x) + C , then you know f (u) du = F (u) + C , no matter what u
is, so long as it is different from zero and well behaved. Consequently, to evaluate
the integral f (ax) dx you would make the substitution u = ax with du = a dx and
then multiply and divide the given integral by the required scale factor and write
the integral in the form
  
1 1 1 1
f (ax) dx = f (ax) a dx = f (u) du = F (u) + C = F (ax) + C
a a a a
 
2 x3 u3
As another example, if you know then you know u2 du = + C
x dx = + C,
3 3
since x is a dummy variable of integration
 and can be replaced by some other symbol.
To find the integral given by I = (3x+7)2 dx you would make a substitution u = 3x+7
with du = 3 dx and then perform the necessary scaling to write
 
1 2 1 1 u3 1 (3x + 7)3 1
I= (3x + 7) 3 dx = u2 du = +C = + C = (3x + 7)3 + C
3 3 3 3 3 3 9

 
Example 3-4. If cos u du = sin u + C , then to find cos(ax) dx one can scale
the integral by letting u = ax with du = a dx to obtain

1 1 1
cos u du = sin u + C = sin(ax) + C
a a a

General Considerations
2  
 
d y dy
If you plot the functions 2 , , y, y(x) dx, y(x) dx dx you will find that
dx dx
differentiation is a roughening process and integration is a smoothing process.
 185
If you are given a function, say y = y(x) = x3 e5x , then you can use the rules for
differentiation of a product of two functions to obtain
dy
= y
(x) = x3 (5e5x ) + (3x2 )e5x = (5x3 + 3x2 ) e5x
dx

One topic in integral calculus develops ways that enable one to reverse the steps
used in differentiation and work backwards to obtain the original function which was
differentiated plus a constant of integration representing the more general function
yg = yg (x) = x3 e5x + C . In the study of integral calculus one develops integration
methods whereby the integral

(5x3 + 3x2 ) e5x dx = x3 e5x + C

can be obtained. This result can also be expressed in the form


d 3 5x
(x e ) dx = x3 e5x + C
dx

and illustrates the basic relation between differentiation and integration, that if you
dF (x)
know a derivative = f (x), then you can immediately write down the integral
dx
  
dF (x)

dx = F (x) dx = f (x) dx = F (x) + C (3.25)
dx

Many integrals can be simplified by making a change of variable  within the

integral. For example, if it is required to evaluate an integral I = f (x) dx, then
sometimes one can find a change  of variables x = g(u) with dx = g
(u) du which
changes the integration to I = f (g(u))g
(u) du which may or may not be an easier
integral to evaluate. In the sections that follow we will investigate various methods
which will aid in evaluating difficult integrals.
Another thing to look for in performing integrations, is that an integration
might produce two results which appear to be different. For example, student A
might perform an integration and get the result

f (x) dx = F (x) + C

and student B might perform the same integration and get the result

f (x) dx = G(x) + C
186
If both students results are correct, then (i) the constants of integration C need
not be the same constants and (ii) there must exist some relationship between the
functions F (x) and G(x) because they have the same derivative of f (x).
In the differential calculus, if one finds two functions F (x) and G(x) having deriva-
tives F
(x) and G
(x) which are equal and satisfy F
(x) = G
(x), over an interval (a, b),
then one

can say that
the functions F (x) and G(x) differ by a constant and one can



write F (x) dx = G (x) dx or F (x) = G(x) + c.

Example 3-5. Consider the functions F (x) = cos2 x and G(x) = sin2 x, these
dF
functions have the derivatives F
(x) = = 2 cos x sin x and G
(x) = 2 sin x cos x
dx
which are equal. Consequently one can state that
F (x) = G(x) + c or cos2 x = sin2 x + c (3.26)

for all values of x. Substituting x = 0 into equation (3.26) one finds 1 = c and
consequently comes up with the trigonometric identity cos2 x + sin2 x = 1.
This result can also be illustrated using integration. Consider the evaluation of
the integral

2 sin x cos x dx

Student A makes the substitution u = sin x with du = cos x dx and obtains the solution
u2



2 sin x cos x dx = 2 u du = 2 + C1 = sin2 x + C1
2
Student B makes the substitution v = cos x with dv = sin x and obtains the solution
v2



2 sin x cos x dx = 2 v dv = 2 + C2 = cos2 x + C2
2
The two integrals appear to be different, but because of the trigonometric identity
cos2 x + sin2 x = 1, the results are really the same as one result is expressed in an
alternative form of the other and the results differ by some constant.

Table of Integrals
If you know a differentiation formula, then you immediately obtain an integration
formula. That is, if
d


F (u) = f (u), then f (u) du = F (u) + C (3.27)
du
Going back and examining all the derivatives that have been calculated one can
reverse the process and create a table of derivatives and integrals such as the Tables
I and II on the following pages.
 Table I Derivatives and Integrals 187
Function f (u) Derivative Integral

dy up+1
y = up = pup1 up du = + C, p
= 1
du p+1

dy 1 du
y = ln u = = ln | u | +C
du u u

dy au
y = au = au ln a au du = +C
du ln a

dy
y = eu = eu eu du = eu + C
du

dy
y = sin u = cos u cos u du = sin u + C
du

dy
y = cos u = sin u sin u du = cos u + C
du

dy
y = tan u = sec2 u sec2 u du = tan u + C
du

dy
y = cot u = csc2 u csc2 u du = cot u + C
du

dy
y = sec u = sec u tan u sec u tan u du = sec u + C
du

dy
y = csc u = csc u cot u csc u cot u du = csc u + C
du

dy 1 du
y = sin1 u = = sin1 u + C
du 1 u2 1 u2

dy 1 du
y = cos1 u = = cos1 u + C
du 1 u2 1 u2

dy 1 du
y = tan1 u = = tan1 u + C
du 1 + u2 1 + u2

1 dy 1 du
y = cot u = = cot1 u + C
du 1 + u2 1 + u2

dy 1 du
y = sec1 u = = sec1 u + C
du u u2 1 2
u u 1

dy 1 du
y = csc1 u = = csc1 u + C
du u u2 1 u u2 1

dy
y = sinhu = cosh u cosh u du = sinh u + C
du

dy
y = cosh u = sinh u sinh u du = cosh u + C
du

dy
y = tanh u = sech 2 u sech 2 u du = tanh u + C
du
188
Table II Derivatives and Integrals

Function f (u) Derivative Integral


dy 2
y = coth u = csch u csch 2 u du = coth u + C
du 
dy
y = sech u = sech u tanh u sech u tanh u du = sech u + C
du

dy
y = csch u = csch u coth u csch u coth u du = csch u + C
du

y = sinh 1 u dy 1 du
 = = sinh 1 u + C
y = ln(u + 1 + u2 ) du 1 + u2 1 + u2


y = cosh 1 u dy 1 du
 = = cosh 1 u + C
du 2 u2 1
y = ln(u + u2 1) u 1

1
y = tanh u dy 1

du
= tanh 1 u + C

1 1+u =
y = ln du 1 u2 1 u2
2 1u
1
y = coth u dy 1

du
= coth 1 u + C

1 u+1 = 2
y = ln du u 1 u2 1
2 u1

y = sech 1 u dy 1 du
1 = = sech 1 u + C
y = cosh 1 du u 1 u2 u 1u 2
u

y = csch 1 u dy 1 du
1 = = csch 1 u + C
y = sinh 1 du u u2 + 1 u u2 + 1
u

Example 3-6. In the above tables of derivative and integrals the symbol u
is a dummy variable of integration. If u = u(x) is a function of x, then to use an
integration formula from the above table there may be occasions where it is necessary
to scale the integral to be evaluated in order that it agree exactly with the form given
in the above tables. 
(a) To evaluate the integral Ia = (5x2 + 7)5 x dx one can make the substitution
u = 5x2 + 7 and make sure that the correct differential du = 10x dx is used in the
integral formula. This may or may not require that scaling by a constant be
performed. Observe that the given integral needs a constant factor of 10 to have
the correct du to go along with the u specified. Consequently, one can multiply
and divide by 10 in order to change the form of the given integral. This gives
 189
 
1 2 5 1 1 u6 1
Ia = (5x + 7) (10x dx) = u5 du = +C = (5x2 + 7)6 + C
10 10 10 6 60

2
(b) In a similar fashion the integral Ib = e3x x dx is evaluated. If one makes the
substitution u = 3x2 , then du = 6x dx is the required form necessary to use the
above table. This again requires that some type of scaling be performed. One
can write  
1 2 1 1 u 1 2
Ib = e3x (6x dx) = eu du = e + C = e3x + C
6 6 6 6

(c) To evaluate the integral Ic = sin(x4 ) x3 dx make the substitution u = x4 with
du = 4x3 dx and then scale the given integral by writing
 
1 4 1 3 1 1
Ic = sin(x )(4x dx) = sin u du = cos u + C = cos(x4 ) + C
4 4 4 4

(d) To evaluate the integral Id = sin x dx let u = x with du = dx and scale the
integral by writing
 
1 1 1
Id = sin x dx = sin u du = cos x + C

(e) Each of the above integrals has been scaled and placed into the form

f (g(x))g
(x) dx

where is some scaling constant. These type of integrals occur quite frequently
and when you recognize them it is customary to make the substitution u = g(x)
with du = g
(x) dx and simplify the integral to the form

f (u) du

Always perform scaling if necessary to get the correct form for du.

Trigonometric Substitutions
The integration tables given above can be expanded by developing other types
of integrals. The appendix C gives an extended table of integrals representing just
a sampling of the thousands of integrals that have been constructed since calculus
was created.
Always examine the integrand of an integral and try to learn some of the alge-
braic and trigonometric forms that can be converted to integrals of a simpler type.
190
All of the trigonometric identities that you have learned are available and can be
thought of as possible aids for evaluating integrals where the integrand involves
trigonometric functions.
One type of integrand to look for is the powers of the trigonometric functions.
Recall the de Moivre1 theorem that states

(cos x + i sin x)n = cos nx + i sin nx (3.28)

where i is an imaginary unit satisfying i2 = 1. Let cos x + i sin x = y and then

multiply both sides of this equation by cos x i sin x to obtain

(cos x i sin x)(cos x + i sin x) = y (cos x i sin x)

and then expand the left-hand side to show that if

1
cos x + i sin x = y then cos x i sin x = (3.29)
y

An addition and subtraction of the equations (3.29) produces the relations

1 1
2 cos x = y + and 2i sin x = y (3.30)
y y

Apply de Moivres theorem to the quantities y and 1/y from equation (3.29) to show
1
cos nx + i sin nx = y n and cos nx i sin nx = (3.31)
yn

Adding and subtracting the equations (3.31) gives the relations

1 1
2 cos nx = y n + and 2i sin nx = y n (3.32)
yn yn

where n is an integer. The above relations can now be employed to calculate trigono-
metric identities for powers of sin x and cos x. Recall the powers of the imaginary unit
i are represented i2 = 1, i3 = i, i4 = i2 = 1, i5 = i, etc, so that the mth power
of either 2i sin x or2 cos x can be calculated
 by employing the binomial expansion to
m m
1 1
expand the terms y or y + and then using the relations from equations
y y
(3.32) to simplify the results.
1
Abraham de Moivre (1667-1754) a French mathematician.
 191
Using the results from the equations (3.30) one can verify the following algebraic
operations     2
1 1
22 i2 sin2 x = y = y2 + 2 = 2 cos 2x 2
y y2 (3.33)
1
or sin2 x = (1 cos 2x)
2
In a similar fashion show that
 3
3 3 3 1 3 1
2 i sin x = y = y 3 3y + 3
y y y
   
1 1 (3.34)
8(i) sin3 x = y 3 3 3 y = 2i sin 3x 3(2i sin x)
y y
3 1
or sin3 x = sin x sin 3x
4 4
To calculate the fourth power of sin x write
 4
4 4 4 1 4 1
2 i sin x = y = y 4 4y 3 + 6 2 + 4
y y y
   
1 1 (3.35)
16 sin4 x = y 4 + 4 4 y 2 + 2 + 6 = 2 cos 4x 4(2 cos 2x) + 6
y y
3 1 1
or sin4 x = cos 2x + cos 4x
8 2 8
In summary, the use of de Moivres theorem together with some algebra produced
the trigonometric identities
1
sin2 x = (1 cos 2x)
2
3 3 1
sin x = sin x sin 3x (3.36)
4 4
4 3 1 1
sin x = cos 2x + cos 4x
8 2 8

In a similar fashion one can use the results from equation (3.30) and establish
the following identities
 
2 2
1 2 1
2 cos x = y + = cos2 x = (1 + cos 2x)
y 2
 
3 3
1 3 3 1
2 cos x = y + = cos3 x = cos x + cos 3x (3.37)
y 4 4
 
4 4 1 4 3 1 1
2 cos x = y + = cos4 x = + cos 2x + cos 4x
y 8 2 8
Verifying the above results is left as an exercise. The calculation of representations
for higher powers of sin x and cos x are obtained using an expansion similar to the
above examples.
192
 
Example 3-7. Evaluate the integrals 2
sin x dx and cos2 x dx
SolutionUsing the trigonometric identities for sin2 x and cos2 x from equations (3.36)
and (3.37) one can write

   
2 1 2 1
sin x dx = (1 cos 2x) dx cos x dx = (1 + cos 2x) dx
2 2
     
2 1 1 2 1 1
sin x dx = dx cos 2x 2dx cos x dx = dx + cos 2x 2dx
2 4 2 4
 
1 1 1 1
sin2 x dx = x sin 2x + C cos2 x dx = x + sin 2x + C
2 4 2 4

where C represents an arbitrary constant of integration.

 
Example 3-8. Evaluate the integrals sin3 x dx and cos3 x dx
SolutionUsing the trigonometric identities for sin3 x and cos3 x from equations (3.36)
and (3.37) one can write

     
3 3 1 3 3 1
sin x dx = sin x sin 3x dx cos x dx = ( cos x + cos 3x) dx
4 4 4 4
     
3 1 3 1
3
sin x dx = sin x dx sin 3x 3dx cos3 x dx = cos x + cos 3x 3dx
4 12 4 12
 
3 1 3 1
sin3 x dx = cos x + cos 3x + C cos3 x dx = sin x + sin 3x + C
4 12 4 12

Example 3-9. Using the substitutions for sin4 x and cos4 x from the equation
(3.36) and (3.37) one can verify the integrals

3 1 1
sin4 x dx = x sin 2x + sin 4x + C
8 4 32

3 1 1
cos4 x dx = x + sin 2x + sin 4x + C
8 4 32

Trigonometric substitution is just one of many methods which can be applied

to aid in the evaluation of an integral where the integrand contains trigonometric
functions.
 193
Example 3-10. Alternative methods for the integration of odd powers of sin x
and cos x involve using the trigonometric identity sin2 x+cos2 x = 1 as illustrated below.
(a) Integral of odd power of sin x
  
sin2n+1 x dx = (sin2 x)n sin x dx = (1 cos2 x)n sin x dx

Make the substitution = cos x with d = sin x dx and express the above integral
in the form  
sin2n+1 x dx = (1 2 )n d, = cos x

The quantity (1 2 )n can be expanded by the binomial expansion. This creates

m+1
a sum of integrals, each of the form m d = m+1 where m is some constant
integer.
(b) Integral of odd power of cos x
  
cos2n+1 x dx = (cos2 x)n cos x dx = (1 sin2 x)n cos x dx

Make the substitution = sin x with d = cos x dx and express the above integral
in the form
  
2n+1 2 n
cos x dx = (cos x) cos x dx = (1 2 )n d, = sin x

Expand the quantity (1 2 )n using the binomial theorem and then like the
previous example integrate each term of the expansion. Note that each term is
m+1
again an integral of the form m d = m+1 .

Products of Sines and Cosines

To evaluate integrals which are products of the sine and cosine functions such
as   
sin mx sin nx dx, sin mx cos nx dx, cos mx cos nx dx

one can use the addition and subtraction formulas from trigonometry
sin(A + B) = sin A cos B + cos A sin B
sin(A B) = sin A cos B cos A sin B
(3.38)
cos(A + B) = cos A cos B sin A sin B

cos(A B) = cos A cos B + sin A sin B

194
to obtain the product relations
1
sin mx sin nx = [cos(m n)x cos(m + n)x]
2
1
sin mx cos nx = [sin(m n)x + sin(m + n)x] (3.39)
2
1
cos mx cos nx = [cos(m n)x + cos(m + n)x]
2

which, with proper scaling, reduce the above integrals to forms involving simple
integration of sine and cosine functions.

Example 3-11. Evaluate the integral I = sin 5x sin 3x dx
Solution Using the above trigonometric substitution one can write
  
1 1 1
I= [cos 2x cos 8x] dx = cos 2x 2dx cos 8x 8dx
2 4 16

to obtain, after proper scaling of the integrals,


1 1
I= sin 5x sin 3x dx = sin 2x sin 8x + C
4 16

Special Trigonometric Integrals

Examining the previous tables of derivatives and integrals one finds that integrals
of the trigonometric functions tan x, cot x, sec x and csc x are missing. Let us examine
the integration of each
 of these functions.
Integrals of the form tan u du
dw
To evaluate this integral express it in the form w
as this is a form which can
be found in the previous tables. Note that
  
sin u d(cos u)
tan u du = du = = ln | cos u| + C
cos u cos u

An alternative approach is to write

  
sec u tan u d(sec u)
tan u du = du = = ln | sec u| + C
sec u sec u

Therefore one can write


tan u du = ln | cos u| + C = ln | sec u| + C
 195
The reason that there are two equivalent representations for the integral of the tan-
gent function is because of a trigonometric identity and properties of the logarithm
function. Note that cos u sec u = 1 and taking logarithms gives
ln(cos u sec u) = ln 1

ln | cos u| + ln | sec u| = 0
or ln | sec u| = ln | cos u|

Integrals of the form cot u du
The integral of the cotangent function is treated much the same way as the
integral of the tangent function. One can write
  
cos u d(sin u)
cot u du = du = = ln | sin u| + C
sin u sin u
One can then show that
 
csc u cot u
cot u du = ln | sin u| + C = du = ln | csc u| + C
csc u
From this result canyou determine a relationship between ln | sin u| and ln | csc u| ?
Integrals of the form sec u du
dw
The integral of the secant function can be expressed in the form w by writing
  
sec u + tan u sec u tan u + sec2 u
sec u du = sec u du = du
sec u + tan u sec u + tan u
so that  
d(sec u + tan u)
sec u du = = ln | sec u + tan u| + C
sec u + tan u

Integrals of the form csc u du
In a similar fashion one can verify that
 
d(csc u + cot u)
csc u du = = ln | csc u + cot u| + C
csc u + cot u

Method of Partial Fractions

P (x)
The method of partial fractions is used to integrate rational functions f (x) =
Q(x)
where P (x) and Q(x) are polynomial functions and the degree of P (x) is less than the
R(x)
degree of Q(x). If a rational function is such that the degree of R(x) is greater
Q(x)
than the degree of Q(x), then one must use long division and write the rational
function in the form
R(x) P (x)
= a0 xn + a1 xn1 + + an1x + an +
Q(x) Q(x)
where now P (x) is a remainder term with the degree of P (x) less than the degree of
Q(x) and our object is to integrate each term of the above representation.
196
x3
Example 3-12. The function y = is a rational function with degree
(x 1)(x 2)
of numerator greater than the degree of the denominator. One can use long division
and write
x3 7x 6
y= 2
= x+3+
x 3x + 2 (x 1)(x 2)

Recall from the study of algebra that when one sums fractions it is customary to
get a common denominator and then sum the numerators. In developing integration
techniques for rational functions the algebra mentioned above is reversed. It has
been found that to integrate a rational function f (x) = PQ(x) (x)
, where the degree of
P (x) is less than the degree of Q(x), it is easier to first factor the numerator and
denominator terms and then split the fraction into the sum of fractions with simpler
denominators. The function f (x) is then said to have been converted into its simplest
fractional component form and these resulting fractions are called the partial fractions
associated with the given rational function. The following cases are considered.

Case 1 The denominator Q(x) has only first degree factors, none of which are
repeated. For example, Q(x) has the form

Q(x) = (x x0 )(x x1 )(x x2 ) (x xn )

where x0
= x1
= x2
=
= xn . One can then write

P (x) A0 A1 A2 An
f (x) = = + + + +
Q(x) x x0 x x1 x x2 x xn
where A0, A1, A2, . . . , An are constants to be determined.

11x 43
Example 3-13. Evaluate the integral I = 2
dx
x 6x + 5
Solution
11x 43
Here the integrand f (x) = 2 is a rational function with
the degree of the
x 6x + 5
numerator less than the degree of the denominator. Observe that the denominator
has linear factors and so one can write
11x 43 11x 43 A1 A2
f (x) = = = + (3.40)
x2 6x + 5 (x 1)(x 5) x1 x5

where A1 , A2 are constants to be determined. Multiply both sides of equation (3.40)

by the factor (x 1) and show
11x 43 A2 (x 1)
= A1 + (3.41)
x5 x5
 197
Evaluate equation (3.41) using the value x = 1 to show A1 = 8. Next multiply equation
(3.40) on both sides by the other factor (x 5) and show
11x 43 A1 (x 5)
= + A2 (3.42)
x1 x1

Evaluate the equation (3.42) using the value x = 5 to show A2 = 3. One can then
write  
  
11x 43 8 3 dx dx
I= dx = + dx = 8 +3
x2 6x + 5 x1 x5 x1 x5

du
Both integrals on the right-hand side of this equation are of the form and
u
consequently one finds
I = 8 ln |x 1| + 3 ln |x 5| + C

where C is a constant of integration. Observe that C is an arbitrary constant and so

one can replace C by ln K , to make the algebra easier, where K > 0 is also an arbitrary
constant. This is done so that all the terms in the solution will be logarithm terms
and therefore can be combined. This results in the solution being expressed in the
form
 
I = ln K(x 1)8 (x 5)3 

Case 2 The denominator Q(x) has only first degree factors, but some of these
factors may be repeated factors. For example, the denominator Q(x) might have
a form such as
Q(x) = (x x0 )k (x x1 )
(x xn )m

where k,
, . . . , m are integers. Here the denominator has repeated factors of

orders k,
, , m. In this case one can write the rational function in the form

P (x) A1 A2 Ak
f (x) = = + + +
Q(x) x x0 (x x0 )2 (x x0 )k
B1 B2 B

+ + + +
x x1 (x x1 )2 (x x1 )

+
C1 C2 Cm
+ + 2
+ +
x xn (x xn ) (x xn )m

where A1, . . . , Ak , B1, . . . , B
, . . . , C1 , . . . , Cm are constants to be determined.

198

8x4 132x3 + 673x2 1183x + 560
Example 3-14. Evaluate the integral I = dx
(x 2)3 (x 8)(x 9)
Solution
Using the method of partial fractions the integrand can be expressed in the
simpler form
8x4 132x3 + 673x2 1183x + 560 A1 A2 A3 B1 C1
3
= + 2
+ 3
+ + (3.43)
(x 2) (x 8)(x 9) x 2 (x 2) (x 2) x8 x9

where A1 , A2 , A3 , B1, C1 are constants to be determined. The constants B1 and C1 are

found as in the previous example. One can verify that
8x4 132x3 + 673x2 1183x + 560
C1 = =2
(x 2)3 (x 8) x=9
8x4 132x3 + 673x2 1183x + 560
and B1 = =3
(x 2)3 (x 9) x=8

One can then write

8x4 132x3 + 673x2 1183x + 560 2 3 A1 A2 A3
= + +
(x 2)3 (x 8)(x 9) x9 x8 x 2 (x 2)2 (x 2)3

which simplifies to
3x2 8x + 3 A1 A2 A3
3
= + 2
+ (3.44)
(x 2) x 2 (x 2) (x 2)3

Multiply both sides of equation (3.44) by (x 2)3 to obtain

3x2 8x + 3 = A1 (x 2)2 + A2 (x 2) + A3 (3.45)

Differentiate equation (3.45) and show

6x 8 = 2A1 (x 2) + A2 (3.46)

Differentiate equation (3.46) and show

6 = 2A1 (3.47)

giving A1 = 3. Evaluate equations (3.45) and (3.46) at x = 2 to show A3 = 1 and

A2 = 4. The given integral can now be represented in the form
    
dx dx dx dx dx
I=3 +4 +3 +2
x2 (x 2)2 (x 2)3 x8 x9
 199
where each term can be integrated to obtain
4 1 1
I = 3 ln |x 2| + + 3 ln |x 8| + 2 ln |x 9| + C
x 2 2 (x 2)2

or
  1 1 4
I = ln (x 2)3 (x 8)3 (x 9)2  + 2
+C
2 (x 2) x2

Case 3 The denominator Q(x) has one or more quadratic factors of the form
ax2 + bx + c none of which are repeated. In this case, for each quadratic factor
there corresponds a partial fraction of the form

A0 x + B 0
ax2 + bx + c

where A0 and B0 are constants to be determined.


11x2 + 18x + 43
Example 3-15. Evaluate the integral I = dx
(x 1)(x2 + 2x + 5)
Solution
Use partial fractions and express the integrand in the form
11x2 + 18x + 43 A Bx + C
2
= + 2
(x 1)(x + 2x + 5) x 1 x + 2x + 5

where A, B, C are constants to be determined. As in the previous example, the

constant A is given by
11x2 + 18x + 43
A= =9
(x2 + 2x + 5) x=1

One can then write

11x2 + 18x + 43 9 Bx + C
2
= 2
(x 1)(x + 2x + 5) x 1 x + 2x + 5

Simplify the left-hand side of this last equation and show

2x + 2 Bx + C
= 2
x2 + 2x + 5 x + 2x + 5

giving B = 2 and C = 2. Here partial fractions were use to convert the given integral
to the form  
8 2x + 2
I= dx + dx
x1 x2 + 2x + 5
200
which can be easily integrated to obtain I = 9 ln |x 1| + ln |x2 + 2x + 5| + ln K. This
 
result can be further simplified and one finds I = ln K(x 1)9 (x2 + 2x + 5) where K
is an arbitrary constant.

Case 4 The denominator Q(x) has one or more quadratic factors, some of which
are repeated quadratic factors. In this case, for each repeated quadratic factor
(ax2 + bx + c)k there corresponds a sum of partial fractions of the form

A1 x + B 1 A2 x + B 2 Ak x + B k
2
+ 2 2
+ +
ax + bx + c (ax + bx + c) (ax2 + bx + c)k

where A1, B1, . . . , Ak , Bk are constants to be determined.

Before giving an example of this last property let us investigate the use of partial
fractions to evaluate special integrals which arise during the application of case 4
above. These special integrals will then be summarized in a table for later reference.

dx
Integrals of the form , where x < .
2 x2
Use partial fractions and write
   
dx 1 1 1
= + dx
2 x2 2 +x x
1
= [ln | + x| ln | x|] + C
2
    
dx 1  + x 1 1 x
so that 2 2
= ln 
 
 + C = tanh + C, x<
x 2 x
See for example the previous result of equation (2.103) with x replaced by x/ .
dx
Integrals of the form , where x > .
x2 2
Use partial fractions and show
   
dx 1 1 1
= dx
x2 2 2 x x+
1
= [ln |x | ln |x + |] + C
2

This can be simplified to one of the forms

      
dx 1 x  1 1
x 1 1

2 2
= ln 
 
 +C = coth +C = tanh +C, x>
x 2 x+ x

Here the previous results from equation (2.103) have been used to produce the
alternative form above.
 201

dx
Integrals of the form
x2 + 2

Make the substitution x = tan u with dx = sec2 u du to obtain

  
dx sec2 u du 1 1
= = du = u+C
x2 + 2 2 (tan2 u + 1)
 
1 x
where u = tan . Therefore one can write

  
dx 1 1 x
= tan +C
x2 + 2

or by constructing a right triangle representing the substitution, one can write the
equivalent forms
 
dx 1 1 1 1 x2 + 2
= cos  +C = sec +C
x2 + 2 x2 + 2


dx
Integrals of the form
(x2 + 2 )2
Make the trigonometric substitution x = tan with dx = sec2 d and show
   
dx sec2 d 1 sec2 1
= = 3 d = 3 cos2 d
(x + 2 )2
2 4 (tan2 + 1)2 sec4


1 1 1 1
= 3 (1 + cos 2) d = + sin 2 = 3 [ + sin cos ]
2 2 3 2 2

Using back substitution representing in terms of x one finds


  
dx 1 1 x x
= tan + +C
(x2 + 2 )2 2 3 x2 + 2

where C is a general constant of integration added to make the result more general.
dx
Integrals of the form
ax2 + bx + c
dx
Integrals having the form I = , where Q(x) = ax2 + bx + c is a quadratic
Q(x)
factor, can be evaluated if one first performs a completing the square operation on
the quadratic term. One finds that either
2  
2
b 4ac b
ax2 + bx + c =a x + + where 4ac b2 > 0
2a 4a2
 2 
2 b b2 4ac
or ax + bx + c =a x + where b2 4ac > 0
2a 4a2
202
4ac b2
Case 1 If 4ac b2 > 0, make the substitution 2 = so that
4a2
 
dx dx
=  
ax2 + bx + c a x+ b 2
+ 2
2a

b
and then make the additional substitution X = x + 2a
with dX = dx. One then
obtains    
dx 1 dX 1 1 X
= = tan1 +C
ax2 + bx + c a X 2 + 2 a
Back substitution and simplifying gives the result
  
dx 2 1 2ax + b
= tan + C, 4ac b2 > 0
ax2 + bx + c 4ac b2 4ac b2

b2 4ac
Case 2 If b2 4ac > 0, make the substitution 2 = and write
4a2
 
dx dx
2
=  
ax + bx + c b 2
a x+ 2a
2

and then make the additional substitution X = x+ 2ab with dX = dx. This produces
the simplified form
   
dx 1 dX 1 1  X 
2
= = ln  +C
ax + bx + c a X 2 2 a X +

Back substitution and simplifying then gives the final result

 2ax + b b2 4ac 
  
dx 1
 
2
= ln   + C, b2 4ac > 0
ax + bx + c 2 2
b 4ac  2ax + b + b 4ac 

Sums and Differences of Squares

Use the Pythagorean theorem and the defi-
nitions of the trigonometric functions as follows.
In the right triangle illustrated one finds

x2 + y 2 = r2 (3.48)

Divide each term of equation (3.48) by r2 and write

x2 y2
+ =1 = cos2 + sin2 = 1 (3.49)
r2 r2
 203
Divide each term of equation (3.48) by x2 and write
y2 r2
1+ = = 1 + tan2 = sec2 (3.50)
x2 x2
Divide each term of equation (3.48) by y2 and write
x2 r2
+1= = cot2 + 1 = csc2 (3.51)
y2 y2

The above identities are known as the Pythagorean identities and can be used
when one recognizes sums and differences of squared quantities in the integrand of
an integral. Sometimes
 the integrand is simplified by using one of these identities.
dx
Integrals of the form 
2 x2
Make the substitution x = sin with dx = cos d to obtain
   
dx cos cos x
 =  d =  d = d = + C = sin1 + C
2
x 2 2 2
sin 2 2
1 sin

This gives the general result


du u
 = sin1 +C (3.52)
2 u2


dx
Integrals of the form 
x2 + 2
Let x = tan u with dx = sec2 u du
and then form a right triangle with one angle
u and appropriate sides of x and . One can then show
  
dx sec2 u du
 = = sec u du
x2 + 2 tan2 u + 1
= ln | sec u + tan u| + C1

x x2 + 2 
= ln | + | + C1 = ln |x + x2 + 2 | + C

where C = C1 ln is just some new constant. In general one can write


du 
 = ln |u + u2 + 2 | + C
u2 + 2


dx
Integrals of the form 
x2 2
Let x = sec u with and form a right triangle with one angle u
dx = sec u tan u du
and appropriate sides x and . One can then show
204
  
dx sec u tan u du
 = = sec u du = ln | sec u + tan u| + C1
x2 2 sec2 u 1

x x2 2 
= ln | + | + C1 = ln |x + x2 2 | + C

where C = C1 ln is some new constant of integration. In general, one can write

du 
 = ln |u + u2 2 | + C
u2 2


du
Example 3-16. Evaluate the integral I =
u4 + 18x2 + 81 
2 du
Solution Recognize the denominator is the square of (u + 9) and write I =
 (u2 + 9)2
dx
This is an integral of the form previously investigated, so that one can
+ 2 )2 (x2
write
  
1 x x
I = 3 tan1 + 2 +C
2 x + 2
where = 3.

Example 3-17. The Pythagorean identities can be employed when one rec-
ognizes the integrand has sums or differences of squared quantities. Sometimes it
is necessary to complete the square on quadratic terms in order to obtain a sum or
difference of squared terms. For example, to evaluate the integral

dx
I=
36x2 + 48x + 41
one can write
  
1 dx 1 dx 1 dx
I= =  2 2  2 2 =
36 x2 + 48
36
x+ 41
36
36 x2 + 4
x + + 41
36 (x + 23 )2 + 25
36
3 3 36 3

One can now make the substitution u = x + 2/3 with du = dx to obtain


1 du
I= , where = 5/6
36 u2 + 2

One then finds

   
1 1 u 1 1 x + 2/3 1 6x + 4
I= tan1 + C = tan1 +C = tan1 +C
36 36 5/6 5/6 30 5
 205

4x5 15x4 + 13x3 4x2 + 13x + 89
Example 3-18. Evaluate the integral I = dx
(x 3)2 (x2 + 1)2
SolutionThe denominator of the integrand has a repeated linear factor and a re-
peated quadratic factor and so by the properties of partial fractions one can write
4x5 15x4 + 13x3 4x2 + 13x + 89 A B Cx + D Ex + F
= + + 2 + 2 (3.53)
(x 3)2 (x2 + 1)2 x 3 (x 3)2 x +1 (x + 1)2

where A, B, C, D, E, F are constants to be determined. Multiply both sides of equa-

tion(3.53) by (x 3)2 and show


4x5 15x4 + 13x3 4x2 + 13x + 89 2 Cx + D Ex + F
= A(x 3) + B + (x 3) + 2 (3.54)
(x2 + 1)2 x2 + 1 (x + 1)2

Evaluate equation (3.54) using the value x = 3 to find

4x5 15x4 + 13x3 4x2 + 13x + 89
B= =2
(x2 + 1)2 x=3

The equation (3.53) can therefore be written as

4x5 15x4 + 13x3 4x2 + 13x + 89 2 A Cx + D Ex + F
= + 2 + 2 (3.55)
(x 3)2 (x2 + 1)2 (x 3)2 x3 x +1 (x + 1)2

The left-hand side of equation (3.55) simplifies to the form

4x4 5x3 3x2 14x 29 A Cx + D Ex + F
2 2
= + 2 + 2 (3.56)
(x 3)(x + 1) x3 x +1 (x + 1)2

Multiply both sides of equation (3.56) by the factor (x 3) and show


4x4 5x3 3x2 14x 29 Cx + D Ex + F
= A + (x 3) + 2 (3.57)
(x2 + 1)2 x2 + 1 (x + 1)2

Evaluating equation (3.57) using the value x = 3 gives the result

4x4 5x3 3x2 14x 29
A= =1
(x2 + 1)2 x=3

Consequently, the equation (3.56) can be written in the form

4x4 5x3 3x2 14x 29 1 Cx + D Ex + F
2 2
= 2 + 2 (3.58)
(x 3)(x + 1) x3 x +1 (x + 1)2

The left-hand side of equation (3.58) simplifies to give the result

3x3 + 4x2 + 8x + 10 Cx + D Ex + F
2 2
= 2 + 2 (3.59)
(x + 1) x +1 (x + 1)2
206
Multiply equation (3.59) on both sides by the factor (x2 + 1)2 and then expand the
right-hand side of the equation to obtain

3x3 + 4x2 + 8x + 10 =(Cx + D)(x2 + 1) + (Ex + F )

(3.60)
or 3x3 + 4x2 + 8x + 10 =Cx3 + Dx2 + (E + C)x + (F + D)

Comparing the left and right-hand sides of equation (3.60) one finds

C = 3, D = 4, E + C = 8, F + D = 10

From the last two equations one finds E = 5 and F = 6. All this algebra reduces the
integrand of the given integral to a summation of simpler terms where each term
can be easily integrated using a table of integrals if necessary.2 One finds
  
1 2 3x + 4 5x + 6
I= + 2
+ 2 + 2 dx (3.61)
x 3 (x 3) x + 1 (x + 1)2

The first integral in equation (3.61) is


dx
= ln |x 3| (3.62)
x3

and the second integral in equation (3.61) is


dx 2
2 2
= (3.63)
(x 3) (x 3)

The third integral in equation (3.61) needs to be scaled to get part of it in the form
du
which can then be integrated. One can write
u
   
(x + 4/3) 3 (2x + 8/3) 3 2x dx dx
3 dx = dx = +4
x2 + 1 2 x2 + 1 2 x2 + 1 x2 + 1

Evaluating these integrals gives the result


3x + 4 3
dx = ln(x2 + 1) + 4 tan1 x (3.64)
x2 + 1 2

The last integral in equation (3.61) can be scaled and written

    
5x + 6 (x + 6/5) 5 2x + 12/5 5 2x dx dx
dx = 5 dx = dx = +6
(x2 + 1)2 2
(x + 1) 2 2 2
(x + 1) 2 2 (x2 + 1)2 (x2 + 1)2
2
A table of integrals is given in the appendix C.
 207
and then integrated to obtain

   
5x + 6 5 1 1 x 6x 5
2 2
dx = 2
+ 3 tan x + 2 = + 3 tan1 x (3.65)
(x + 1) 2 x +1 x +1 2(x2 + 1)

Combining the above results gives

2 3 6x 5
I = ln |x 3| + ln(x2 + 1) + 4 tan1 x + + 3 tan1 x + C (3.66)
x3 2 2(x2 + 1)

where C is a constant of integration. To check that what has been done is correct one
dI
should note that the final result should satisfy = f (x), where f (x) is the integrand
dx
of the original integral. This check is left as an exercise.

Note that in the case the denominator has a single linear factor (x a), then one
f (x) A (x)
can write = + where A is a constant which can be determined
(x a)g(x) xa (x)
f (x) (x)
from the relation = A + (x a) evaluated at x = a.
g(x) (x)

Example 3-19. Find the partial fraction expansion for representing a function
having the form
ax6 + bx5 + cx4 + dx3 + ex2 + f x + g
f (x) =
(x 1)(x 2)(x 3)3 (x2 + x + 1)(x2 + 3x + 1)4

where a, b, c, d, e, f, g are known constants.

Solution Here the denominator has the unrepeated linear factors (x 1) and (x 2).
The linear factor (x 3) is repeated three times. The quadratic factor (x2 + x + 1) is
unrepeated and the quadratic factor (x2 + 3x + 1) is repeated four times. Using the
properties of partial fractions, represented by the previous cases 1 through 4, the
form for the partial fraction representation of the given function is
A0 B0 C0 D0 E0 F0 x + G0
f (x) = + + + 2
+ 3
+ 2
x 1 x 2 x 3 (x 3) (x 3) x +x+1
A1 x + B1 A2 x + B2 A3 x + B3 A4 x + B4
+ 2 + + 2 + 2
x + 3x + 1 (x2 + 3x + 1)2 (x + 3x + 1)3 (x + 3x + 1)4

where A0 , B0, . . . , A4 , B4 are constants to be determined.

The following table III is a summary of previous results.

208 Table III Summary of Integrals

1 1
sin2 u du = u sin 2u + C
2 4

1 1
cos2 u du = u + sin 2u + C
2 4

3 1
sin3 u du = cos u + cos 3u + C
4 12

3 1
cos3 u du = sin u + sin 3u + C
4 12

3 1 1
sin4 u du = sin 2u + sin 4u + C
8 4 32

3 1 1
cos4 u du = u + sin 2u + sin 4u + C
8 4 32

tan u du = ln | cos u | +C

cot u du = ln | sin u | +C

sec u du = ln | sec u + tan u | +C

csc u du = ln | csc u + cot u | +C

du 1 +u
= ln +C
2 u 2 2 u

du 1 u
= ln +C
u2 2 2 u+
  
du 1 1 u
= tan +C
u2 + 2

  
du 1 1 u u
= 3 tan + 2 +C
(u2 + 2 )2 2 u + 2
  
du 2 1 2au + b
= tan + C, 4ac b2 > 0
au2 + bu + c 4ac b2 4ac b2

du 1 2au + b b2 4ac
= ln + C, b2 4ac > 0
au2 + bu + c b2 4ac 2au + b + b2 4ac
  
du 1 u
 = sin +C
2
u 2

du 
 = ln u + u2 + 2 + C
2 + u2

du 
 = ln u + u2 2 + C
u2 2
 209
Integration by parts
If d(U V ) = U dV + V dU then one can write U dV = d(U V ) V dU and so by inte-
grating both sides of this last equation one obtains the result
  
U dV = d(U V ) V dU
  (3.67)
U dV = U V V dU

The equation (3.67) is known as the integration by parts formula. Another form for
the integration by parts formula is
 


U (x)V (x) dx = U (x)V (x) V (x)U
(x) dx (3.68)

When using integration by parts try to select U = U (x) such that V (x)U
(x) dx is easy
to integrate. If this is not possible, then alternative methods of integration have
to be investigated. Integration by parts is a powerful method for evaluating many
types of integrals. Sometimes it is necessary to apply the method of integration by
parts multiple times before a result is obtained.

Example 3-20. Evaluate the integral I = arctanx dx
Solution
For the given example, let U = arctan x and dV = dx then one can calculate
 
dx
dU = d ( arctanx) = and dV = dx or V = x (3.69)
1 + x2

Substituting the results from the equations (3.69) into the integration by parts
formula (3.67) one finds
 
x dx
arctanx dx = x arctanx
1 + x2

dU
In order to evaluate the last integral, use the integration formula = ln U + C
U
and recognize that if U = 1 + x2 , then it is necessary that dU = 2x dx and so a scaling
must be performed on the last integral. Perform the necessary scaling and express
the integration by parts formula in the form
 
1 2x dx
arctanx dx =x arctanx
2 1 + x2
1
=x arctanx ln(1 + x2 ) + C
2
where C is a constant of integration.
210
Note when using integration by parts and you perform an integration to find V ,
it is not necessary to add the constant of integration for if V is replaced by V + C in
equation (3.67) one would obtain




U dV = U (V + C) (V + C) dU = U V + C U V dU C U

and the constant would disappear. You can always add a general constant of integra-
tion after performing the last integral. This is usually done to make the final result
more general.

The integration by parts formula can be written in different ways. Using the
rule for differentiation of a product, write



d dV dU dV dU
(U V ) = U +V and consequently U V = U dx + V dx
dx dx dx dx dx

or
dV dU



U dx = U V V dx (3.70)
dx dx
which is the form for integration by parts previously

presented. In equation (3.70)
dV
make the substitution = W (x) with V (x) = W (x) dx, then equation (3.70) takes
dx
on the form




dU


U (x)W (x) dx = U (x) W (x) dx W (x) dx dx (3.71)
dx

and interchanging the functions U (x) and W (x) gives the alternative result




dW


U (x)W (x) dx = W (x) U (x) dx U (x) dx dx (3.72)
dx

The above two integration by parts formulas tells us that to integrate a product
of two functions one can select either of the equations (3.71) or (3.72) to aid in the
evaluation of the integral. One usually selects from the above two formulas that
formula which produces an easy to obtain result, if this is at all possible.


Example 3-21. Evaluate the integral x2 sin nx dx
Solution The integration by parts formula may be repeated many times to evaluate
an integral. For the given integral one can employ integration by parts, with U = x2
and dV = sin nx dx to obtain
 211
     
cos nx cos nx
x2 sin nx dx = x2 2x dx
n n
Use scaling and apply integration by parts on the last integral, with U = 2x and
dV = cosnnx dx, to obtain
       
cos nx sin nx sin nx
2x dx = 2x 2 dx
n n2 n2

The last integral can be scaled and integrated. One finds that the last integral
becomes   
sin nx 2
2 dx = cos nx
n2 n3
Back substitution gives the results
    
cos nx sin nx 2
x2 sin nx dx = x2 2x + cos nx + C
n n2 n3

where C is a general constant of integration which can be added at the end of any
indefinite integral.

Reduction Formula  
The use of the integration by parts formula U dV = U V V dU to evaluate
an integral gives a representation of a first integral in terms of a second integral.
Sometimes, when an integration by parts is performed on the second integral, one
finds that it can be reduced to a form of the first integral. When this happens one
can usually obtain a general formula, known as a reduction formula, for evaluating
the first and sometimes the second integral.

Example 3-22. Evaluate the integral Im = sinm x dx where m is a positive
integer. 
Solution Write the integral as Im = sinm1 x sin x dx and use integration by parts
with
U = sinm1 x dV = sin x dx
dU =(m 1) sinm2 x cos x dx V = cos x
to obtain 
Im = sinm1 x cos x + (m 1) sinm2 x cos2 x dx

m1
Im = sin x cos x + (m 1) sinm2 x (1 sin2 x) dx

Im = sinm1 x cos x + (m 1) [Im2 Im ]

and using algebra one can solve for Im to obtain
212
1 (m 1)
Im = sinm1 x cos x + Im2
m m
or  
m 1 m1 (m 1)
sin x dx = sin x cos x + sinm2 x dx (3.73)
m m
This result is known as a reduction formula where the integral of a higher power of
sin x is expressed in terms of an integral involving a lower power of sin x.
Substitute m = 2 into the reduction formula (3.73) and show
 
2 1 1 x 1
sin x dx = sin x cos x + dx = sin x cos x + C (3.74)
2 2 2 2
where a general constant of integration has been added to the final result.
Substitute m = 3 into the reduction formula (3.73) gives
 
3 1 2 2 1 2
sin x dx = sin x cos x + sin x dx = sin2 x cos x cos x + C (3.75)
3 3 3 3
where C is some general constant of integration that has been added to obtain a more
general result. It is left as an exercise to show that the results given by equations
(3.74) and (3.75) are alternative forms of the results obtained in the examples 3-5
and 3-6.


Example 3-23. Using integration by parts on the integral Jm = cosm x dx one
can verify the reduction formula
1 (m 1)
Jm = cosm1 x sin x + Jm2
m m
or  
m 1 (m 1)
cos x dx = cosm1 x sin x + cosm2 x dx
m m

Example 3-24. For m and n integers and held constant during the integration
process, evaluate the integrals
 
m
Sm = x sin nx dx and Cm = xm cos nx dx

Solution Use integration by parts on the Sm integral with U = xm and dV = sin nx dx.
cos nx
One finds dU = mxm1 dx and V = so that
n

cos nx m cos nx m
Sm = xm + xm1 cos nx dx or Sm = xm + Cm1 (3.76)
n n n n
 213
An integration by parts applied to the Cm integral with U = xm and dV = cos nx dx
1
produces dU = mxm1 dx and V = sin nx. The Cm integral then can be represented
n


sin nx m sin nx m
Cm = xm xm1 sin nx dx or Cm = xm Sm1 (3.77)
n n n n

In the equations (3.76) and (3.77) replace m by m1 everywhere and use the resulting
equations to show
 
m cos nx
m m1 sin nx m 1
Sm =x + x Sm2
n n n n
 
m sin nx m m1 cos nx m1
Cm =x x + Cm2
n n n n

which simplify to the reduction formulas

cos nx sin nx m(m 1)
Sm = xm + mxm1 Sm2 (3.78)
n n2 n2
sin nx cos nx m(m 1)
Cm =xm + mxm1 2
Cm2 (3.79)
n n n2

These reduction formula can be used as follows. First show that

cos nx sin nx
S0 = sin nx dx = and C0 = cos nx dx =
n n

and then use integration by parts to show

S1 = x sin nx dx, C1 = x cos nx dx
cos nx sin nx sin nx cos nx
S1 = x + , C1 =x +
n n2 n n2

where the general constants of integration have been omitted. Knowing S0 , S1 , C0 , C1

the reduction equations (3.78) and (3.79) can be used to calculate

S2 , C 2 , S 3 , C 3 , S 4 , C 4 , . . .

The Definite Integral

Consider the problem of finding the area bounded by a given curve y = f (x), the
lines x = a and x = b and the x-axis. The area to be determined is illustrated in the
figure 3-2.
214

Figure 3-2. Area under curve and partitioning the interval [a, b] into n-parts.

The curve y = f (x) is assumed to be such that y > 0 and continuous for all x [a, b].
To find an approximation to the area desired, construct a series of rectangles as
follows.
ba
(1) Divide the interval [a, b] into n-parts by defining a step size x = and then
n
define the points
x0 =a

x1 =a + x = x0 + x
x2 =a + 2x = x1 + x
.. ..
. . (3.80)
xi =a + ix = xi1 + x
.. ..
. .
(b a)
xn =a + nx = a + n = b = xn1 + x
n
This is called partitioning the interval (a, b) into n-parts.
(2) Select arbitrary points ti , within each x interval, such that xi1 ti xi for all
values of i ranging from i = 1 to i = n. Then for all values of i ranging from 1 to
n construct rectangles of height f (ti ) with the bottom corners of the rectangle
touching the x-axis at the points xi1 and xi as illustrated in the figure 3-2.
(3) The area of the ith rectangle is denoted Ai =(height)(base), where the height of
the rectangle is f (ti ) and its base is xi = xi xi1. The sum of all the rectangles
n

n


is given by Snt = Ai = f (ti ) xi which is called the Riemann3 sum for the
i=1 i=1

3
Georg Friedrich Bernhard Riemann (1826-1866) A German mathematician.
 215
function y = f (x). The resulting sum is determined by the partition constructed.
This Riemannian sum represents an approximation to the area under the curve.
This approximation gets better as each xi gets smaller.
Define the limit of the Riemann sum
n

n

 b
lim S t = lim Ai = lim f (ti )xi = f (x) dx, xi1 ti xi (3.81)
n n n x0
a
i=1 n i=1

where the quantity on the right-hand side of equation (3.81) is called the definite
integral from a to b of f (x) dx and the quantity on the left-hand side of equation (3.81)
is the limit of the sum of rectangles as x tends toward zero. The notation for the
definite integral from a to b of f (x) dx has the physical interpretation illustrated in
the figure 3-3.

Figure 3-3. Mnemonic device for determining area under curve.

The quantity f (x) dx = dA is to represent an element of area which is a rectangle

positioned
 a distance x from the origin, having height f (x) and base dx. The integral
sign is an elongated S to remind you that rectangles are being summed and
the lower limit a and upper limit b on the integral sign is to remind you that the
summation of rectangles is a limiting process taking place between the limits x = a
and x = b.
Observe that if F (x) is a differentiable function which is continuous over the
interval [a, b] and F (x) is selected as a particular integral of f (x), then one can write
dF (x)
= F
(x) = f (x), or one can write the indefinite integral
dx
  
dF (x)
f (x) dx = dx = dF (x) = F (x) + C
dx
216
It will now be demonstrated that any function F (x)which is a particular integral
b
of f (x) can be used to evaluate the definite integral f (x) dx. To accomplish this
a
graph the function y = F (x), satisfying F
(x) = f (x), between the values x = a and
x = b and then partition the interval [a, b] into n-parts in the same way as for the
original function y = f (x). Apply the mean-value theorem for derivatives to the points
(xi1 , F (xi1 )) and (xi , F (xi )) associated with the ith x interval of the curve y = F (x).
In using the mean-value theorem make special note that the function F (x) is related
to f (x) by way of differentiation so that F
(x) = f (x) is the slope of the curve y = F (x)
at the point x. One can then calculate the slope of the secant line through the points
(xi1 , F (xi1 )) and (xi , F (xi )) as

F F (xi ) F (xi1 )
ms = =
x xi xi1

The mean-value theorem says there must exist a point ci satisfying xi1 ci xi ,
such that the slope of the tangent line to the curve y = F (x), at the point x = ci ,
is the same as the slope of the secant line. By our choice of F (x), the slope of the
tangent line to F (x) at the point x = ci is given by f (ci ) since F
(x) = f (x). Therefore,
one can write
F (xi ) F (xi1 )
= f (ci ) or F (xi ) F (xi1 ) = f (ci )xi , xi = xi xi1 (3.82)
xi xi1

This mean-value relationship can be applied to each x interval for all values of i
ranging from 1 to n.
Make note of the fact the points ti , i = 1, . . . , n, used to evaluate the definite
integral in equation (3.81) were not specified. They were arbitrary points satisfying
xi1 ti xi for each value of the index i ranging from 1 to n. Note the values ci ,
i = 1, . . . , n which satisfy the mean-value equation (3.82) are special values. Suppose
one selects for equation (3.81) the values ti = ci as i ranges from 1 to n. In this
special case the summation given by equation (3.81) becomes
n


Snc = f (ci )xi = [F (x1 ) F (x0 )] + [F (x2 ) F (x1 )] + + [F (xn ) F (xn1 )]
i=1
n
(3.83)


f (ci )xi =F (xn ) F (x0 ) = F (b) F (a)
i=1

Observe that the summation of terms on the right-hand side of equation (3.83) is a
telescoping sum and can be written as
 217
[F (x1 ) F (x0 )] + [F (x2 ) F (x1 )]
+ [F (x3 ) F (x2 )] + [F (x4 ) F (x3 )]
..
.
+ [F (xn1 ) F (xn2 )] + [F (xn ) F (xn1 )] = F (xn ) F (x0 ) = F (b) F (a)

where as i ranges from 1 to n 1, for each term F (xi ) there is a F (xi ) and so these
terms always add to zero and what is left is just the last term minus the first term.
This result still holds as x 0 and so one can state that the area bounded by the
curve y = f (x), the x-axis, the lines x = a and x = b is given by the definite integral
 b b
f (x) dx = F (x) = F (b) F (a) (3.84)
a a

where F (x) is any particular integral of f (x).

Observe that the constant of integration associated with the indefinite integral
can be omitted when dealing with definite integrals. If this constant were used,
then equation (3.84) would become [F (b) + C] [F (a) + C] = F (b) F (a) which is
the same result as given in equation (3.84). Also note that the name definite
integral indicates that the integral has a definite value of F (b) F (a), which does
not contain the symbol x or the constant C . The symbol x is called a dummy variable
of integration in the definite integral and can be replaced by some other symbol.
The above result is a special case of the fundamental theorem of integral calculus.

Fundamental theorem of integral calculus

Let f (x) > 0 denote a continuous function over the interval a x b. Partition
the interval (a, b) into n subintervals [x0 , x1 ], [x1, x2 ], . . . , [xi1, xi ], . . . , [xn1, xn ] which
may or may not be of equal length. Select an arbitrary point ti [xi1, xi ] for
i = 1, 2, . . . , n and construct the rectangles of height f (ti ) and base xi = xi xi1 with
Ai = f (ti )xi the area of the ith rectangle. A special case of the above situation is
illustrated in the figure 3-2.
 
Let An = ni=1 Ai = ni=1 f (ti )xi denote the Riemannian sum which equals the
sum of the areas of these rectangles and let F (x) denote any function which is
dF (x)
an integral of f (x) with the property F (x) = f (x) dx or = f (x). Then the
dx
fundamental theorem of integral calculus can be expressed
n

 b b
lim An = n
lim f (tk )x = f (x) dx = F (x) = F (b) F (a)
n a
x0
k=1 a
218
Since F
(x) = f (x), the

fundamental theorem of integral calculus is sometimes ex-
b x=b
pressed in the form F
(x) dx = F (x) = F (b) F (a)
a x=a
Note that the limiting summation of rectangles that represents the definite in-
tegral of f (x) from x = a to x = b has the same value no matter how the points ti are
selected inside the interval [xi1 , xi], for i = 1, 2, . . ., n. This is because of the continuity
of f (x) over the interval [a, b]. Recall that f (x) is a continuous function if for every 1
there exists a positive number 1 such that

|f (ti ) f (ci )| < 1 whenever |ti ci | < 1 (3.85)

Suppose it is required that the condition given by (3.85) be satisfied for each value

i = 1, 2, . . . , n. If 1 = , with  as small as desired, and n is selected large enough
(b a)
ba
such that x = < 1 , then one can compare the two summations
n
n

n


Snt = f (ti )xi and Snc = f (ci )xi
i=1 i=1

One finds the absolute value of the difference of these sums satisfies
 

n 
n
|Snt Snc | = 
 
[f (ti )xi f (ci )xi ] |f (ti ) f (ci )| xi
 
i=1 i=1

ba
For n large enough such that each xi = < 1 for all values of the index i and
n
   ba
for all values
|f (ti ) f (ci )| < 1 of the index i, then Snt Snc  n1 x = n =
ba n
This states that the difference between the two sums Snt and Snc can be made as small
as desired for n large enough and in the limit these sums are the same. A similar
type of argument can be made for an arbitrary, unequally spaced, partitioning of
the interval [a, b].
Properties of the Definite Integral
dF (x)
1. If F
(x) = = f (x), then the definite
dx
integral
 x
A(x) = F (x) F (a) = f (t) dt
a

represents the area under the curve y = f (t) be-

tween the limits t = a and t = x. Whenever the
upper or lower limit of integration involves the symbol x, then the dummy variable
of integration in the definite integral is usually replaced by a different symbol.
 219
2. Differentiate the above result to show
 x 
dA(x) dF (x) d
= = f (t) dt = f (x) (3.86)
dx dx dx a

This shows that to differentiate a definite integral with respect to x, where the
upper limit of integration is x and the lower limit of integration is a constant,
one obtains the integrand evaluated at the upper limit x.
3. If the direction of integration is changed, then the sign of the integral changes
 a  b
f (x) dx = f (x) dx (3.87)
b a
4. The interval of integration [a, b] can be broken up into smaller subintervals, say,
[a, 1 ], [1, 2 ], [2, b] and the integral written
 b  1  2  b
f (x) dx = f (x) dx + f (x) dx + f (x) dx (3.88)
a a 1 2

5. Assume the curve y = f (x) crosses the x-axis at some point x = c between the
lines x = a and x = b, such that f (x) is positive for a x c and f (x) is negative
for c x b, then
 c
f (x) dx represents a positive area
a

and the integral

 b
f (x) dx represents a negative area.
c

 b
The definite integral f (x) dx represents the summation of the signed areas
a  b
above and below the x-axis. The integral |f (x)| dx represents a summation of
a
positive areas.
n


6. The summation f (ti ) represents the sum of the heights associated with the
i=1
n
1

rectangles constructed in the figure 3-2, and the sum y = f (ti ) represents
n
i=1
the average height of these rectangles. Using the equation (3.81) show that in
ba
the limit as x 0 and using xi = , this average height can be represented
n
n n  b
1
1
1
y = lim f (ti ) = lim f (ti )xi = f (x) dx
x0
n
n i=1 x0 b a
n i=1
b a a
220
This states that the average value for the height of the curve y = f (x) between
the limits x = a and x = b is given by
 b
1
Average height of curve = y = f (x) dx (3.89)
ba a

7. The integral of a constant times a function equals the constant times the integral
of the function or  b  b
cf (x) dx = c f (x) dx
a a

8. Change of variables in a definite integral

 b
Given an integral of the form I = f (g(x)) g
(x) dx it is customary to make the
a
substitution u = g(x) with du = g
(x) dx. When this is done in a definite integral, then
the limits of integration must also be changed. Thus, when x = a, then u = g(a) and
when x = b, then u = g(b) so that if f (u) is well defined on the interval [g(a), g(b)], then
the given integral can be reduced to the form
 b  g(b)


I= f (g(x)) g (x) dx = f (u) du
a g(a)

9. Area between curves

Let y = f (x) and y = g(x) denote two curves which are continuous on the interval
[a, b] and assume that f (x) g(x) for all x [a, b], then one can state that

 b  b  b  b
f (x) dx g(x) dx or g(x) dx f (x) dx
a a a a

Sketch the curves y = f (x), y = g(x), the lines x = a and x = b and sketch in a
rectangular element of area dA representative of all the rectangular elements being
summed to find the area bounded by the curves and the lines x = a and x = b. Note
that the element of area is given by

dA = (y of upper curve y of lower curve) dx

 221
and these elements of area must be summed between the lines x = a and x = b. The
situation is illustrated in the figure 3-4(a). The area between the curves is obtained
by a summation of the rectangular elements to obtain
 b  b
Area = dA = (g(x) f (x)) dx
a a

There may be times when the given curves dictate that a horizontal element of area
between the curves be used to calculate the area between the curves.

Figure 3-4. Area between curves.

For example, if x = F (y) and x = G(y) are two curves where a vertical element of area
is not appropriate, then try using a horizontal element of area with the element of
area dA given by

dA = (x of right-hand curve x of left-hand curve) dy

and then sum these elements of area between the lines y = c and y = d to obtain
 d  d
Area = dA = (G(y) F (y)) dy
c c

The situation is illustrated in the figure 3-4(b).

If the curves are intersecting curves between the limits of integration and sit-
uations arise where the upper curve switches and becomes a lower curve, then the
integral representing the area must be broken up into integrals over sections otherwise
one obtains a summation of signed areas.
222
 /6
Example 3-25. Consider the definite integral I = sin 2x cos4 2x dx
0
Make the change of variable u = cos 2x with du = 2 sin 2x dx. The new limits of
integration are found by substituting x = 0 and x = /6 into the equation u = cos 2x
1
to obtain ua = cos 2x =1 and ub = cos 2x = . Here du = 2 sin 2x dx
x=0 x=/6 2
and so the given integral must be scaled. The scaled integral can then be written
 /6
1
I= (cos 2x)4 (2 sin 2x dx)
2 0

where now the substitutions for u, du and new limits on integrations can be performed
to obtain
 1/2  1  
1 4 1 41 u5 1 1 1 31
I= u du = u du = = 1 =
2 1 2 1/2 2 5 1/2 10 32 320

Example 3-26. Find the area between the curves y = sin x and y = cos x for
0 x .
Solution
Sketch the given curves over the domain specified and show the curves intersect
where x = /4. The given integral can then be broken up into two parts and one can
write

 /4 /4
A1 = [cos x sin x] dx = sin x + cos x = 21
0 0

A2 = [sin x cos x] dx = cos x sin x =1+ 2
/4 /4

The total area is then A1 + A2 = 2 2
A summation of the signed areas is given by

[cos x sin x] dx = ( 2 1) (1 + 2) = 2 = A1 A2
0

Example 3-27. Find the area of the triangle bounded by the x-axes, the line
y = bh1 x and the line y = h bh2 (x b1 ), where b = b1 + b2 .
 223
Solution
Get into the habit of
(i) Sketching the curve y = f (x) to be integrated.
(ii) Sketching in an element of area dA = f (x) dx or dA = y dx
(iii) Labeling the height and base of the rectangular element of area.
(iv) Sketching the lines x = a and x = b for the limits of integration.

Sketching the above lines one obtains the figure 3-5 illustrated below.

h h
Figure 3-5. Triangle defined by x-axis, y = b1 x and y = h b2 (x b1 )

The big triangle is built up of two smaller right triangles and an element of area
has been constructed inside each of the smaller right triangles. The area of the left
smaller right triangle is given by
 b1  b1  b1
h h h x2 x=b1 h b21 1
A1 = y1 dx = x dx = x dx = = = hb1
0 0 b1 b1 0 b1 2 x=0 b1 2 2
where the element of rectangular area y1 dx is summed from x = 0 to x = b1 . This
result says the area of a right triangle is one-half the base times the height. The
area of the other right triangle is given by
 b  b
  b  b
h h
A2 = y2 dx = h (x b1 ) dx = h dx (x b1 ) dx
b1 b1 b2 b1 b1 b2
 b  b
h x=b h (x b1 )2 x=b
=h dx (x b1 ) dx = h x
b1 b2 b1 x=b1 b2 2 x=b1
h (b b1 )2 h b22 1
=h(b b1 ) = hb2 = hb2
b2 2 b2 2 2
224
where the rectangular element of area y2 dx was summed from x = b1 to x = b = b1 + b2 .
Adding the areas A1 and A2 gives the total area A where
1 1 1 1
A = A1 + A2 = hb1 + hb2 = h(b1 + b2 ) = hb
2 2 2 2

That is, the area of a general triangle is one-half the base times the height.

Example 3-28.
The curve

{ (x, y) | x = r cos , y = r sin , 0 }

is the set of points (x, y) defined by the para-

metric equations x = r cos and y = r sin as
varies from 0 to . Sketch this curve and
show it represents the upper half of a circle
with radius r centered at the origin as illustrated in the figure. An element of rect-
angular area dA = y dx is constructed at a general point (x, 0), where the height of the
rectangle is y and the base of the rectangle
r
is dx. The area between the semi-circle
and the x-axis is given by A = y dx which says the elements of area are to be
r
summed between the limits x = r and x = r. Make the substitutions y = r sin and
dx = r sin d and note that when x = r , then = and when x = r , then = 0.
This gives the integral for the area as
 r  0 
2
A= y dx = r sin (r sin d) = r sin2 d
r 0

Make the trigonometric substitution sin2 = 21 (1 cos 2) and then perform the inte-
grations, after appropriate scaling, by using the previous table of integrals to show

  
r2 r2 1
A= (1 cos 2) d = d cos 2(2d)
2 0 2 0 2 0
r2 1 r 2
A = sin 2 =
2 0 2 0 2

This shows the area of the semi-circle is r2 /2 and so the area of the full circle is r2 .
As an alternative, one can construct an element of area in the shape of a rectangle
which is parallel to the x-axis as illustrated in the figure below. Due to symmetry
 225
this element of area is represented dA = 2x dy and summing these elements of area
in the y-direction from 0 to r gives the total area as
 r  r
A= dA = 2x dy
0 0

Substituting in the values x = r cos and dy = r cos d and noting that y = 0, cor-
responds to = 0 and the value y = r, corresponds to = /2, one obtains the
representation
 /2  /2
A= 2(r cos )(r cos d) = 2r 2 cos2 d
0 0

1
Using the trigonometric identity cos2 = (1 + cos 2) the above integral for the area
2
becomes
   
/2 /2 /2
2 1 1
A = 2r (1+cos 2) d = r 2 d + cos 2 (2 d)
0 2 0 2 0

where the integral of cos 2 has been appropri-

ately scaled. Performing the integrations one
finds

/2 1 /2
A =r 2 + sin 2
0 2 0
2
r
A=
2
which is the same as our previous answer.

Solids of Revolution
Examine the shaded areas in each of the figures 3-6(a),(b),(c) and (d). These
areas are going to be rotated about some axis to create a solid of revolution. The
solid of revolution created depends upon what line is selected for the axis of rotation.
Figure 3-6(a)
Examine rotation of element of area about lines x = 0, y = 0, x = x0 and y = y0 .
Consider a general curve y1 = y1 (x) for a x b such as the curve illustrated in
the figure 3-6(a). To find the area bounded by the curve, the xaxis, and the lines
x = a and x = b one would construct an element of area dA = y1 (x) dx and then sum
these elements from a to b to obtain the area
 b
A= y1 (x) dx (3.90)
a
226
If this area is rotated about the xaxis a solid of revolution is created. To find
the volume of this solid the element of area is rotated about the xaxis to create a
volume element in the shape of a disk with thickness dx. The radius of the disk is
y1 (x) and the volume element is given by

dV = y12 (x) dx

Figure 3-6.
Area element to be rotated about an axis to create volume element.

A summation of these volume elements from a to b gives the volume of the solid as
 b
V = y12 (x) dx (3.91)
a

If the shaded area of figure 3-6(a) is rotated about the yaxis one can create a
cylindrical shell volume element with inner radius x, outer radius x + dx and height
y1 (x). The cylindrical shell volume element is given by
 227
dV =(Volume of outer cylinder Volume of inner cylinder)(height)
   
dV = (x + dx)2 x2 y1 (x) = 2x dx + (dx)2 y1 (x)

The term (dx)2 y1 (x) is an infinitesimal of second order and can be neglected so that
the volume of the cylindrical shell element is given by

dV = 2x y1 (x) dx (3.92)

A summation of these cylindrical shell volume elements gives the total volume
 b
V = 2 x y1 (x) dx (3.93)
a

If the element of area dA = y1 dx is rotated about the line x = x0 one obtains the
volume element in the shape of a cylindrical shell with the volume element given by

dV = 2(x0 x)y1 (x) dx

and the volume of the solid of revolution is obtained from the integral
 b
V = 2 (x0 x)y1 (x) dx
a

which represents a summation of these volume elements to generate the volume of

revolution.
If the element of area is rotated about the line y = y0 one obtains a volume
element in the shape of a washer with the volume element represented
 
dV = Area outer circle Area of inner circle (Thickness)
 
dV = y02 (y0 y1 (x))2 dx
228
The total volume is then given by a summation of these volume elements
 b
V = [y02 (y0 y1 (x))2] dx
a

Figure 3-6(b)
Examine rotation of element of area about lines x = 0, y = 0, x = x0 and y = y0 .
Examine the figure 3-6(b) and show that to determined the area bounded by the
curve x1 = x1 (y), the yaxis and the lines y = , y = is obtained by a summation of
the area element dA = x1 (y) dy from to . The total area is given by

A= x1 (y) dy

If the area is rotated about a line, then a solid of revolution is created. To find
the volume of the solid one can rotate the element of area about the line to create
an element of volume which can then be summed. Consider the element of area
illustrated as being rotated about the axes (i) the xaxis, (ii) the yaxis, (iii) the
line x = x0 and (iv) the line y = y0 to obtain respectively elements of volumes in the
shapes of (i)a cylindrical shell element, (ii) a disk element, (iii) a washer element and
(iv) another cylindrical shell element. Show these volume elements are given by
 
(i) dV = 2y x1 (y) dy (iii) dV = x20 (x0 x1 (y))2 dy
(ii) dV = x21 (y) dy (iv) dV = 2(y0 y) x1 (y) dy
Figure 3-6(c)
Examine rotation of element of area about lines x = 0, y = 0, x = x0 and y = y0 .
Examine the figure 3-6(c) and show the area bounded by the curves y1 = y1 (x),
y2 = y2 (x) and the lines x = a, x = b, is obtained by a summation of the area element
dA = [y1 (x) y2 (x)] dx from a to b. This summation gives the total area as
 b
A= (y1 (x) y2 (x)) dx
a

If this area is rotated about a line, then a solid of revolution is created. To find the
volume of the solid one can rotate the element of area about the same axis to create
an element of volume which can then be summed. Consider the element of area
being rotated about the axes (i) the xaxis, (ii) the yaxis, (iii) the line x = x0 and
(iv) the line y = y0 to obtain respectively elements of volumes in the shapes of (i) a
washer element, (ii) a cylindrical shell element, (iii) another cylindrical shell element
and (iv) another washer element. Show these volume elements can be represented
as follows.
 229
 
(i) dV = y22 (x) y12 (x) dx (iii) dV = 2(x0 x) [y1 (x) y2 (x)] dx
 
(ii) dV = 2x [y1 (x) y2 (x)] dx (iv) dV = (y0 y2 (x))2 (y0 y1 (x))2 dx

Figure 3-6(d)
Examine rotation of element of area about lines x = 0, y = 0, x = x0 and y = y0 .
Examine the figure 3-6(d) and show the area bounded by the curves x1 = x1 (y),
x2 = x2 (y) and the lines y = , y = , is obtained by a summation of the area element
dA = [x1 (y) x2 (y)] dy from to . This summation gives the total area

A= [x1 (y) x2 (y)] dy

If this area is rotated about a line, then a solid of revolution is created. The volume
associated with this solid is determined by a summation of an appropriate volume
elements. These volume elements can be determined by rotating the element of area
about the same line from which the solid was created.
Consider the element of area rotated about the lines (i) the xaxis, (ii) the
yaxis, (iii) the line x = x0 and (iv) the line y = y0 to obtain respectively elements
of volumes in the shapes of (i) a cylindrical shell element, (ii) a washer element,
(iii) another washer element and (iv) another cylindrical shell element. Show these
volume elements can be represented as follows.
 
(i) dV = 2y [x1 (y) x2 (y)] dy (iii) dV = (x0 x2 (y))2 (x0 x1 (y))2 dy
 
(ii) dV = x21 (y) x22 (y) dy (iv) dV = 2(y0 y) (x1 (y) x2 (y)) dy

Example 3-29. Take the semi-circle

{ (x, y) | x = r cos , y = r sin , 0 }

as defined in the previous example and rotate it about the x-axis to form a sphere.
The figure 3-7 will aid in visualizing this experiment.
Note that the vertical element of area when rotated becomes an element of
volume dV in the shape of a disk with radius y and thickness dx. The volume of this
disk is given by dV = y2 dx and a summation of these volume elements from x = r
to x = r gives  r  r
V = dV = y 2 dx
r r
230
Making the same substitutions as in the previous example one finds
 0 
2 3
V = (r sin ) (r sin ) d = r sin3 d
0

Figure 3-7.
Semi-circle rotated about x-axis creating the volume element in shape of a disk.

Now use the trigonometric identity

1
sin3 = (3 sin sin 3)
4
and express the volume integral in the form of an integration of trigonometric func-
tions. After appropriate scaling, make use of the integration table to show that

  
r 3 r 3 1
V = (3 sin sin 3) d = 3 sin d sin 3(3d)
4 0 4 0 3 0



r 3 1 r 3 1
V = 3( cos ) ( cos 3) = 3(1 1)] + (1 1)
4 0 3 0 4 3
4 3
V = r
3
This shows the volume of the sphere of radius r is 4/3 times times the radius cubed.
If the horizontal element of area illustrated in the example 3-24, is rotated about
the x-axis a cylindrical shell element results, like the one illustrated by equation
(3.92), but with x and y interchanged. The inner radius of the cylinder is y and the
outer radius is y + dy and the length of the cylinder is 2x. The element of volume is
given by
dV =(length)[(outer radius)2 (inner radius)2 ]
dV =(2x)[(y + dy)2 y 2 ] = (2x)[y 2 + 2y dy + (dy)2 y 2 ]
dV =2(2x)y dy + (2x)(dy)2
 231
This is an example of an equation where (dy)2 is a higher ordered infinitesimal which
can be neglected. Neglecting this higher ordered infinitesimal gives

dV = 2(2x)y dy

Summation on dy from 0 to r gives the total volume

 r
V = 4 xy dy
0

Substituting x = r cos , y = r sin and dy = r cos d and changing the limits of inte-
gration to ranging from 0 to /2, one finds
 /2
V =4 (r cos )(r sin )(r cos d)
0
 /2
3
V =4r cos2 sin d
0

1
This last integral is recognized as being of the form u2 du = u3 where u = cos and
3
du = sin d. Perform the necessary scaling and then integrate to obtain
  /2
3 1 4 3
V = 4r (cos )3 = r
3 0 3

for the volume of the sphere.

Sometimes one can place axes associated with a solid such that plane sections
at x and x + dx create a known cross sectional area which can be represented by a
function A = A(x) and consequently the plane sections produce a slab shaped volume

element given by dV = A(x) dx. The resulting volume

between the planes x0 and x1 can then expressed as a
summation of these sandwich slices
 x1
V = A(x) dx (3.94)
x0

One can also check cross sections at y and y + dy to see if

there results a known area A = A(y). If this is the case,
then the volume element associated with these plane
232
slices are written dV = A(y) dy and the total volume between the planes y0 and y1 is
expressed as the summation  y1
V = A(y) dy (3.95)
y0

Integration by Parts
Integration by parts associated with a definite integral has the form
 b  b  b
u(x)v
(x) dx = d(u(x)v(x)) dx v(x)u
(x) dx
a a a
 b x=b
 b
u(x)v
(x) dx =u(x)v(x) v(x)u
(x) dx (3.96)
a x=a a
 b  b
u(x)v
(x) dx =u(b)v(b) u(a)v(a) v(x)u
(x) dx
a a

 T
Example 3-30. To integrate I = test dt let u = t with du = dt and dv = est dt
0
with v = 1s est , then the integration by parts formula gives
 T  T
st t st T 1 st T sT 1 T
I= te dt = e e dt = e 2 est
0 s 0 0 s s s 0
T sT 1 sT
= e 2 [e 1]
s s

 /2
Example 3-31. Evaluate the integral J = x sin x dx
/2
Solution Let U = x and dV = sin x dx giving dU = dx and V = cos x, so that integration
by parts produces the result
/2
 /2 /2

J = x cos x cos x dx = cos + cos + sin x =2
/2 /2 2 2 2 2 /2

 b
Example 3-32. Evaluate the integral I = x2 b x dx, where a, b are constants
a
satisfying a < b.
Solution Use integration by parts with

u =x2 dv =(b x)1/2 dx

2
du =2x dx v = (b x)3/2
3
 233
to obtain
x=b
 b b
 b
2 4
I =uv v du = x2 (b x)3/2 + x(b x)3/2 dx
x=a a 3 a 3 a
 b
2 4
I = a2 (b a)3/2 + x(b x)3/2 dx
3 3 a
One can now apply integration by parts again on the last integral using
u =x dv =(b x)3/2 dx
2
du =dx v = (b x)5/2
5
to obtain   
2 2 3/2 4 2 5/2
b 2 b 5/2
I = a (b a) + x(b x) + (b x) dx
3 3 5 a 5 a

b

2 2 3/2 4 2 5/2 2 2 7/2
I = a (b a) + a(b a) + (b x)
3 3 5 5 7 a
2 2 3/2 8 5/2 16 7/2
I = a (b a) + a(b a) + (b a)
3 15 105
2
I= (b a)3/2 (15a2 + 12ab + 8b2 )
105

Physical Interpretation
When using definite integrals the integration by parts formula has the following
physical interpretation. Consider the section of a curve C between points P and Q
on the curve which can be defined by

C = { (x, y) | x = x(t), y = y(t), t0 t t1 } (3.97)

Here the section of the curve C is defined by a set of parametric equations x = x(t)
and y = y(t) for t0 t t1 with the point P having the coordinates (x0 , y0 ) where
x0 = x(t0 ) and y0 = y(t0 ). Similarly, the point Q has the coordinates (x1 , y1 ) where
x1 = x(t1 ) and y1 = y(t1 ). A general curve illustrating the situation is sketched in the
figure 3-8.
Examine the element of area dA1 = y dx and sum these elements of area from x0
to x1 to obtain  x1  t1
dx
A1 = y dx = y(t) dt = Area x0 P Qx1 (3.98)
x0 t0 dt
Similarly, if one sums the element of area dA2 = x dy from y0 to y1 there results
 y1  t1
dy
A2 = x dy = x(t) dt = Area y0 P Qy1 (3.99)
y0 t0 dt
234

Figure 3-8. Physical interpretation for integration by parts.

Examine the figure 3-8 and verify the areas of the following rectangles

A3 = area rectangle 0x1 Qy1 = x1 y1 , A4 = area rectangle 0x0 P y0 = x0 y0 (3.100)

The integration by parts formula can then be expressed

 t1 t1
 t1
dx dy
y(t) dt =x(t)y(t) x(t) dt
t0 dt t0 t0 dt
 t1
(3.101)
dy
=(x1 y1 x0 y0 ) x(t) dt
t0 dt

In terms of areas this result can be written

A1 = A3 A4 A2 or A2 = A3 A4 A1

and is interpreted as saying that the areas A1 and A2 are related and if one these
areas is known, then the other area can also be evaluated.
Improper Integrals
Integrals of the form
   b
I1 = f (x) dx, I2 = f (x) dx, I3 = f (x) dx (3.102)
a
 235
are called improper integrals and are defined by the limiting processes
 b  b  b
I1 = lim f (x) dx, I2 = lim f (x) dx, I3 = lim f (x) dx (3.103)
b a
a
a a a
b

if these limits exist. In general an integral of the form

 b
I= f (x) dx (3.104)
a

is called an improper integral if

(i) The lower limit a is allowed to increase or decrease without bound.
(ii) The upper limit b is allowed to increase or decrease without bound.
(iii) The lower limit a decreases without bound and the upper limit b increases with-
out bound.
(iv) The integrand f (x) is not defined at some point c between the end points a and
b, then the integral is called an improper integral and one must write
 b   b
I= f (x) dx = lim f (x) dx + lim f (x) dx (3.105)
c c+
a <c
a >c

if these limits exist.

Improper integrals occur in a variety of forms in advanced mathematics courses
involving integral transforms. For example,
The Laplace transform of a function is written as the improper integral

L{F (t)} = L{F (t); t s} = F (t)est dt = f (s) (3.106)
0

and represents a transformation of a function F (t) into a function f (s), if the improper
integral exists. Other transforms frequently encountered are
The Fourier exponential transform is written as the improper integral

1
Fe {f (x); x } = f ()ei d = Fe () (3.107)
2

The Fourier sine transform is written as the improper integral


2
Fs {f (x); x } = f (x) sin x dx = Fs () (3.108)
0

The Fourier cosine transform is written as the improper integral


2
Fc {f (x); x } = f (x) cos x dx = Fc () (3.109)
0
236
if the above integrals exist. Numerous other transforms similar to those mentioned
above can be found in many advanced mathematics, physics and engineering texts.
Integrals used to define Functions
Definite integrals are frequently used to define special functions. For example,
the natural logarithm function can be defined
 x
1
ln x = dt
1 t
 x (3.110)
d d 1 1
ln x = dt =
dx dx 1 t x

The natural logarithm of x is represented as the area bounded by the curve 1/t, the
lines t = 1, t = x and the t-axis.
Properties of the natural logarithm function can be obtained from the defining
integral. For example, one finds
(i) ln 1 = 0
 ab  a  ab
1 1 1
(ii) ln(a b) = dt = dt + dt In the last integral make the substitution
1 t 1 t a t
t = a u with dt = a du, so that when t = a, u = 1 and when t = ab, u = b and obtain
 a  b
1 1
ln(a b) = dt + du
1 t 1 u
giving
 
ln(a b) = ln a + ln b
 1/b
1 1 u du
(iii) ln = dt Make the substitution t = b
with dt = b
with new limits on u
b 1 t
from b to 1 and show
   1 du  b
1 b = du
ln = u
b b 1 u
b
 
1
giving ln = ln b
b  
a 1 1
(iv) Using the result from (iii) it follows that ln = ln a = ln a + ln giving the
  b b b
a
result ln = ln a ln b
b
 ar
1
(v) ln (ar ) = dt Make the substitution t = ur with dt = rur1 du with new limits
1 t
of integration u ranging from 1 to a to show
 ar  a
r 1 1
ln (a ) = dt = r du
1 t 1 u

showing that ln (ar ) = r ln a

 237
Other functions defined by integrals
There are many special functions which are defined as a definite integral or im-
proper integral. For example, three functions defined by integrals which occur quite
frequently are the following.
The Gamma function is defined

(z) = tz1 et dt (3.111)
0

and integration by parts shows that

(z + 1) = z(z) (3.112)

Use limits to show (1) = 1 and then use equation

(3.112) to show

(1) = 1, (2) = 1!, (3) = 2(2) = 2!, . . .

and when z = n is an integer the Gamma function reduces to the factorial function
(n) = (n 1)! = (n 1)(n 2)(n 3) 3 2 1 or (n + 1) = n!
The values (0), (1), (2), . . . are not defined.
The error function is defined
 x
2 2
erf (x) = et dt (3.113)
0

The complementary error function is defined


2 2
erfc (x) = 1 erf (x) = et dt (3.114)
x

The error function4 erf (x) occurs in the study of the normal probability dis-
2
tribution and represents the area under the curve 2 et from 0 to x, while the
complementary error function is the area under the same curve from x to .
The above is just a very small sampling of the many special functions which are
defined by integrals.
4
Note alternative forms for the definition of the error function are due to scaling.
238
Arc Length
Let y = f (x) denote a continuous curve for x [a, b] and consider the problem
of assigning a length to the curve y = f (x) between the points (a, f (a)) = P0 and
ba
(b, f (b)) = Pn . Partition the interval [a, b] into n-parts by defining x = and
n
labeling the points

a = x0 , x1 = x0 + x, x2 = x1 + x, . . . , xi = xi1 + x, . . . , xn = xn1 + x = b

as illustrated in the figure 3-9.

Label the points (xi , f (xi )) = Pi for i = 0, 1, 2, . . ., n and construct the line segments
Pi1 Pi for i = 1, 2, . . ., n. These line segments connect the points

(x0 , f (x0 )), (x1 , f (x1 )), (x2 , f (x2 )), . . . , (xi, f (xi )), . . ., (xn1, f (xn1)), (xn , f (xn ))

in succession and form a polygonal line connecting the points (a, f (a)) and (b, f (b)).

Figure 3-9.
Approximation of arc length by summation of straight line segments.

The sum of these line segments can be represented


n n
2


f (xi ) f (xi1 )
sn = (xi xi1 )2 + [f (xi ) f (xi1 )]2 = 1+ xi (3.115)
i=1 i=1
xi

where xi = xi xi1. This sum is an approximation to the length of the curve y = f (x)
between the points (a, f (a)) and (b, f (b)). This arc length approximation gets better
as xi gets smaller or as n gets larger. If in the limit as n , the above sum exists
 239
so that one can write s = n lim sn , then the curve y = f (x) is said to be rectifiable.
The limiting value s is defined to be the arc length of the curve y = f (x) between the
f (xi ) f (xi1 )
end points (a, f (a)) and (b, f (b)). Here lim = f
(xi ) and the infinite sum
x0 xi
becomes a definite integral and so one can express the limiting value of the above
sum as 
 b  b  2
 dy
s= 1 + [f
(x)]2 dx = 1+ dx (3.116)
a a dx
If x [a, b], then define the arc length s = s(x) of the curve y = f (x) between the points
(a, f (a)) and (x, f (x)) as  x 
s = s(x) = 1 + [f
(t)]2 dt (3.117)
a

and define the differential of arc length ds = s
(x) dx. The element of arc length ds
can be determined from any of the following forms
  2  2
 dy dx
ds = dx2 + dy 2 = 1+ dx = + 1 dy
dx dy
 (3.118)
 2  2
dx dy 
ds = + dt = [x
(t)]2 + [y
(t)]2 dt
dt dt

Example 3-33. Find the circumference of the circle defined by x = r cos ,

y = r sin for 0 2 , where r is a constant.
Solution
The element of arc length squared can be written
ds2 =dx2 + dy 2 or

ds = [x
()]2 + [y
()]2 ] d

dx dy
Substituting in the derivatives = x
() = r sin and = y
() = r cos , the element
d d
of arc length is
 
ds = [r sin ]2 + [r cos ]2 d = r sin2 + cos2 d = r d

and the total arc length is a summation of these elements

 2 2
s=r d = r = 2 r
0 0
240
Example 3-34. Find the length of the line segment connecting the points
(x1 , y1 ) and (x2 , y2 ) where x1
= x2 .
Solution
y y
The slope of the line through these points is m = 2 1 and the equation of the
x2 x1
line through these points is y y1 = m(x x1 ). The arc length is given by
  2
 x2  x2 x2
dy   
s= 1+ dx = 1 + m2 dx = ( 1 + m2 ) x = ( 1 + m2 )(x2 x1 )
x1 dx x1 x1

This simplifies to the well known result s = (x2 x1 )2 + (y2 y1 )2

Area Polar Coordinates

The equation of a curve in polar coordinates is given by r = f (). To find the
area bounded by the curve r = f (), the rays = and = , divide the angle
into n-parts by defining =
n and then defining the rays
0 = , 1 = 0 + , . . . , i = i1 + , . . . , n = n1 + =

The area between the rays = i1 , = i and the curve r = f (), illustrated in the
figure 3-10, is approximated by a circular sector with area element
1 2 1
dAi = r i = f 2 (i ) i (3.119)
2 i 2
where i = i i1 and ri = f (i ). A summation of these elements of area between
the rays = and = gives the approximate area
n n n


1
1
dAi = ri2 i = f 2 (i ) i (3.120)
2 2
i=1 i=1 i=1

Figure 3-10.
Approximation of area by summation of circular sectors.
 241
This approximation gets better as i gets smaller. Using the fundamental
theorem of integral calculus, it can be shown that in the limit as n , the equation
1
(3.119) defines the element of area dA = r2 d. A summation of these elements of
2
area gives   
1 2 1
Polar Area = dA = r d = f 2 () d (3.121)
2 2

Example 3-35. Find the area bounded by the polar curve

r = 2r0 cos for 0 .
Solution
One finds that the polar curve r = 2r0 cos , for 0 , is a circle of radius r0
which has its center at the point (r0 , 0) in polar coordinates. Using the area formula
given by equation (3.121) one obtains
  

1 2 sin 2
Area = (2r0 cos ) d = 2r02 2
cos d = r02 (cos 2 + 1) d = r02 + = r02
2 0 0 0 2 0

Make note of the fact that polar curves sometimes sweep out a repetitive curve. For
example, in the polar equation r = 2r0 cos , if varied from 0 to 2 , then the polar
distance r would sweep over the circle twice. Consequently, if one performed the
integration  2
1
(2r0 cos )2 d
2 0

one would obtain twice the area or 2r02 . Therefore, one should always check polar
curves to see if some portions of the curve are being repeated as the independent
variable varies.

Arc Length in Polar Coordinates

In rectangular coordinates, ds2 = dx2 + dy 2 represents the element of arc length
squared. If one changes to polar coordinates using the transformation equations

x = x(r, ) = r cos and y = y(r, ) = r sin (3.122)

then the total differentials dx and dy are given by

x x y y
dx = dr + d dy = dr + d
r and r
dx = cos dr r sin d dy = sin dr + r cos d
242
and represent the total differentials in terms of the variables r and . Squaring and
adding these differentials one finds

ds2 = dx2 + dy 2 = dr 2 + r 2 d2 (3.123)

as the representation for the arc length squared in polar coordinates. Other forms
for representing the element of arc length in polar coordinates are
 
 2  2
dr d 
ds = + r 2 d = 1 + r2 dr = [r
(t)]2 + r 2 [t][
(t)]2 dt (3.124)
d dr

Example 3-36. Find the circumference of the circle r = 2r0 cos given in the
previous example.
 Here an element of arc length is given by the polar coordinate representation
Solution
 2
dr dr
ds = + r 2 d, where = 2r0 sin . Integration of the element of arc length
d d
from 0 to gives
  
2 2 2 2
s= 4r0 sin + 4r0 cos d = 2r0 d = 2r0 = 2r0
0 0 0

Surface of Revolution

When a curve is revolved about the x or y-

axis a surface of revolution results. The problem
of determining the surface area of the resulting
surface of revolution is approached using the fol-
lowing arguments. First consider a right circular cone where the top has been cut
off. The resulting figure is called the frustum of a right circular cone. The top sur-
face is the shape of a circle with radius r1 and the bottom surface is a circle with
radius r2 > r1 . The side surface has a slant height of length
as illustrated in the
accompanying figure. The surface area associated with the side of this figure is given
by5
Side surface area = (r1 + r2 )
(3.125)
5
How this result is derived can be found in example 3-38
 243

Figure 3-11.
Arc length ds rotated about x-axis to form frustum of right circular cone.
Consider next the surface of revolution obtained when a curve y = f (x) is rotated
about the x-axis as illustrated in the figure 3-11. Let ds denote an element of arc
length in cartesian coordinates connecting the points (x, y) and (x + dx, y + dy) on the
curve and observe that when this element is rotated about the x-axis a frustum of
a right circular cone results. The radius of one circle is y and radius of the other
circle is y + dy . The element of surface area dS is the side surface area of the frustum
and given by equation (3.125) and so one can write

dS = [y + (y + dy)] ds (3.126)

The product dy ds is an infinitesimal of the second order and can be neglected so

that the element of surface area can be written in the form

dS = 2y ds (3.127)

By the fundamental theorem of integral calculus a summation of these surface el-

ements gives the total surface area of the surface of revolution. This total surface
area can be expressed in different forms depending upon the representation of the
arc length ds (See equations (3.118).) If y = y(x), then one can write

 b  b  2
dy
S= 2y ds = 2 y 1+ dx (3.128)
a a dx

In a similar fashion, a curve x = F (y), for c y d, rotated about the y-axis

would created a surface of revolution with surface area given by

 d  2
dx
S = 2 x 1+ dy (3.129)
c dy
244
In general, if a curve is rotated about a line, one can express the element of surface
area associated with a surface of revolution as dS = 2 ds where ds is an element
of arc length on the curve expressed in an appropriate form and represents the
distance from the arc length element ds to the axis of revolution.

Example 3-37. Consider the upper half of the circle x2 + y2 = r2 rotated about
dy dy
the x-axis to form a sphere. Here 2x + 2y = 0 or = x/y so that the surface area
dx dx
of the sphere is given by
   
r r r
x2  r
S = 2 y 1 + 2 dx = 2 x2 + y2 dx = 2r dx = 2rx = 4r 2
r y r r r

Example 3-38. h
The line y = (x b), 0 x b is rotated about
b
the y-axis to form a cone. An element of arc length
ds on the line is rotated about the y -axis to form an
element of surface area dS given
 by dS = 2x ds Using  2
dy
ds2 = dx2 + dy 2
in the form ds = 1 + dx dx, the total
surface area of a cone with height h and base radius b
is given by

 b  2
dy
S = 2 x 1+ dx
0 dx

Perform the integration and show the total surface area of the cone is given by

S = b
where
2 = b2 + h2 , (3.130)

with
=
1 +
2 representing the slant height of the cone.

As an exercise, use the above results to obtain the surface area of a frustum
associated with the given right circular cone as follows. If a is the base radius of
cone associated with slant height
1 and b is the base radius associated with cone
of the slant height
1 +
2 , then show, Area of frustum = b(
2 +
1 ) a
1 . Then use
similar triangles and simplify this result and show, Area of frustum =
2 (b+a) which
agrees with equation (3.125).
 245
Mean Value Theorems for Integrals
There are several mean value theorems associated with definite integrals which
can be found under the following names.
(i) The first mean value theorem for integrals. If f (x) is a continuous function for
a x b, then there is a point x = 1 [a, b] such that
 b
f (x) dx = (b a) f (1 ) = (b a)f (a + 1 (b a)), 0 < 1 < 1 (3.131)
a

(ii) The generalized first mean value theorem for integrals.

If both f (x) and g(x) are continuous functions for a x b and the function
g(x) does not change sign for x [a, b], then there exists a point x = 2 [a, b] such
that
 b  b  b
f (x)g(x) dx = f (2 ) g(x) dx = f (a + 2 (b a)) g(x) dx, 0 < 2 < 1 (3.132)
a a a

(iii) Bonnets second mean value theorem for integrals.

If both f (x) and g(x) are continuous functions for a x b and the function g(x)
is a positive monotonic6 decreasing function, then there exists a point x = 3 [a, b]
such that  b  3
f (x)g(x) dx = g(a) f (x) dx (3.133)
a a

where 3 = a + 3 (b a), for 0 < 3 < 1. Alternatively, if g(x) is a positive monotonic

increasing function, then there exists a point x = 4 [a, b] such that
 b  b
f (x)g(x) dx = g(b) f (x) dx (3.134)
a 4

where 4 = a + 4 (b a), for 0 < 4 < 1.

(iv) The generalized second mean value theorem for integrals.
If both f (x) and g(x) are continuous functions for a x b and the function
g(x) is a monotone increasing or monotone decreasing over the interval [a, b], then
there exists a point x = 5 [a, b] such that
 b  5  b
f (x)g(x) dx = g(a) f (x) dx + g(b) f (x) dx (3.135)
a a 5

where 5 = a + 5 (b a), for 0 < 5 < 1.

6
If f (x) is defined on an interval [a, b] and if f is such that whenever a < x1 < x2 < b, there results
f (x1 ) f (x2 ), then f is called a monotone increasing function over the interval [a, b]. If the inequality above is
reversed so that f (x1 ) f (x2 ), then f is called a monotone decreasing function over the interval [a, b].
246
Proof of Mean Value Theorems
If f (x) > 0 is a continuous function over the interval a x b, define the functions
 x  x
G(x) = g(t) dt, H(x) = f (t)g(t) dt, P (x) =G(x)H(b) G(b)H(x)
a a
G
(x) =g(x), H
(x) =f (x)g(x), P
(x) =G
(x)H(b) G(b)H
(x)

and observe that P (a) = P (b) = 0 because G(a) = H(a) = 0 and the way P (x) is defined.
Consequently, it is possible to apply Rolles7 theorem which states that there must
exist a value x = , for a < < b, such that P
() = 0. This requires
 b  b
g() f (t)g(t) dt f ()g() g(t) dt = 0
a a

which simplifies to give the generalized first mean value theorem for integrals
 b  b
f (x)g(x) dx = f () g(x) dx
a a

Note the special case g(x) = 1 produces the first mean value theorem for integrals.
To prove Bonnets second mean value theorem, assume f (x) > 0 is a continuous
function for a x b and consider the cases where g(x) is monotone decreasing and
monotone increasing over the interval [a, b].
Case 1: Assume that g(x) is positive and monotone decreasing over the inter-

val [a, b]. Define the function (x) = g(a) ax f (x) dx which is continuous over the

interval [a, b] and demonstrate (a) = 0 ab f (x)g(x) dx (b). An application
of the intermediate value theorem shows there exists a value x = such that
b
() = g(a) a f (x) dx = a f (x)g(x) dx.
Case 2: Assume that g(x) is positive and monotone increasing over the interval
b
[a, b]. Define the function (x) = g(b) x f (x) dx which is continuous over the in-

terval [a, b] and demonstrate (b) = 0 ab f (x)g(x) dx (a). An application of
the intermediate value theorem shows that there exists a value x = such that
b b
() = g(b) f (x) dx = a f (x)g(x) dx.
To prove the 
generalized second mean value theorem 
for integrals consider the
x b
function F (x) = f (u) du and then evaluate the integral f (x)g(x) dx using integra-
a a
tion by parts to show
 b b
 b  b


f (x)g(x) dx = g(x)F (x) F (x)g (x) dx = g(b)F (b) F (x)g
(x) dx (3.136)
a a a a

7
Michel Rolle (1652-1719) a French mathematician.
 247
The last integral in equation (3.136) can be evaluated as follows. The assumption
that g(x) is a monotonic function implies that the derivative g
(x) is of a constant
sign for x [a, b] so that by the generalized first mean value theorem for integrals the
equation (3.136) can be expressed in the form
 b  b
f (x)g(x) dx =g(b)F (b) F () g
(x) dx = g(b)F (b) F ()[g(b) g(a)]
a a
  b
(3.137)
=g(a)F () + g(b)[F (b) F ()] = g(a) f (x) dx + g(b) f (x) dx
a

Differentiation of Integrals
The general Leibnitz formula for the differentiation of a general integral, where
both the lower and upper limits of integration are given by functions (t) and (t),
is given by the relation
 (t)  (t)
d f (t, ) d d
f (t, ) d = d + f (t, (t)) f (t, (t)) (3.138)
dt (t) (t) t dt dt

To derive the Leibnitz differentiation formula consider the following simpler exam-
ples.
 x
d
Example 3-39. Show that f (t) dt = f (x) where a is a constant.
 dx a
x
Solution Let F (x) = f (t) dt and use the definition of a derivative to obtain
a

x+x x  x+x
dF (x) F (x + x) F (x) a
f (t) dt a
f (t) dt 1
= lim = lim = lim f (t) dt
dx x0 x x0 x x0 x x

Apply the mean value theorem for integrals and show the above reduces to
dF (x) 1
= F
(x) = lim f (x + x) x = f (x), where 0 < < 1
dx x0 x

Consequently, one can write

 x  
d d d
f (t) dt = f (x), f (t) dt = f (), f (t) dt = f ()
dx a d 0 d 0
 a  x
d d
Note that the above results imply that f (t) dt = f (t) dt = f (x).
dx x dx a
248
 (x)
d d d
Example 3-40. Show that f (t) dt = f ((x)) f ((x))
dx (x) dx dx
Solution Write the given integral in the form
      
(x) 0 (x) (x) (x)
d d d
f (t) dt = f (t) dt + f (t) dt = f (t) dt f (t) dt
dx (x) dx (x) 0 dx 0 0

and use chain rule differentiation employing the results from the previous example
to show    
(x) (x) (x)
d d
f (t) dt = f (t) dt f (t) dt
dx (x) dx 0 0
 
 

d d d d
= f (t) dt f (t) dt
d 0 dx d 0 dx
 (x)
d d d
f (t) dt =f ((x)) f ((x))
dx (x) dx dx

Example 3-41.
Consider the function I = I(x, g(x), h(x)) defined by the integral
 h(x)
I = I(x, g, h) = f (x, t) dt (3.139)
g(x)

where the integrand is a function of both the variables x and t and the limits of
integration g and h are also functions of x. The integration is with respect to the
variable t and it is assumed that the integrand f is both continuous and differentiable
with respect to x. The differentiation of a function defined by an integral containing
a parameter x is given by the Leibnitz rule
 h(x)
dI f (x, t) dh dg
= I
(x) = dt + f (x, h(x)) f (x, g(x)) . (3.140)
dx g(x) x dx dx

The above result follows from the definition of a derivative together with the use of
chain rule differentiation.

Consider first the special case of the integral I1 (x) = gh f (x, t) dt where g and h
are constants. Calculate the difference
 h
I1 (x + x) I1 (x) = [f (x + x, t) f (x, t)] dt
g

and then employ the mean value theorem with respect to the x-variable and write
 249
f (x + x, t)
f (x + x, t) f (x, t) = x, 0<<1
x
to write
 h
f (x + x, t)
I1 (x + x) I1 (x) = x dt where 0 < < 1.
g x
Dividing both sides by x and letting x 0 gives the derivative
 h  h
dI1 I1 (x + x) I1 (x) f (x + x, t) f (x, t)
= I1
(x) = lim = lim dt = dt
dx x0 x x0 g x g x
In the special case that both the upper and lower limits of integration are functions
of x, one can employ chain rule differentiation for functions of more than one variable
and express the derivative of I = I(x, g, h) as
dI I I dg I dh
= + + . (3.141)
dx x g dx h dx
where  h(x)
I f (x, t) I I
= dt, = f (x, h(x)), = f (x, g(x))
x g(x) x h g
The equation (3.141) then simplifies to the result given by equation (3.140).

Double Integrals
Integrals of the form
 b d  d  b
I1 = f (x, y) dy dx or I2 = f (x, y) dx dy
a c c a
are called double integrals of the function z = f (x, y) over the rectangular region R
defined by
R = { (x, y) | a x b, c y d }

These double integrals are evaluated from the inside out and can be given the fol-
lowing physical interpretation. The function z = f (x, y), for (x, y) R, can be thought
of as a smooth surface over the rectangle as illustrated in the figure 3-12.

Figure 3-12.
Planes x = a constant and y = a constant intersecting surface z = f (x, y).
250

Figure 3-13.
Elements of volume in the shape of slabs.

In the left figure, the plane, y = a constant, intersects the surface z = f (x, y) in the
curve
z = f (x, y) y= a constant

and the integral  b

f (x, y) dx y=a constant
a

represents the area under this curve. If this area is multiplied by dy , one obtains an
element of volume dV in the shape of a slab with thickness dy as illustrated in the
figure 3-13. This element of volume dV can be expressed as an area times a thickness
to obtain  
b
dV = f (x, y) dx dy
a

If the element of volume is summed from y = c to y = d, there results the total volume
  
d b
V = f (x, y) dx dy
c a

Here the inner integral is integrated first while holding y constant and then the result
is integrated with respect to y from c to d to perform a summation representing
the volume under the surface z = f (x, y). Double integrals and multiple integrals in
general are sometimes referred to as iterated or repeated integrals. When confronted
with multiple integrals always perform the inner integral first and the outside integral
last.
 251
In a similar fashion, the plane, x = a constant, intersects the surface z = f (x, y) in
the curve
z = f (x, y) x= a constant
so that the integral  d
f (x, y) dy x= a constant
c

represents the plane area under the curve. If the resulting area is multiplied by dx,
one obtains a volume element dV in the shape of a slab times a thickness dx. This
slab is given by    d
dV = f (x, y) dy dx
c

If these volume elements are summed from x = a to x = b, then the resulting volume
under the surface is given by
  
b d
V = f (x, y) dy dx
a c

Another way to interpret the previous double integrals is to first partition the
interval [a, b] into n parts by defining x = ba
n
and then partition the interval [c, d]
dc
into m parts by defining y = m . One can then define the points
ba
a =x0 , . . . , xi = a + ix, . . . xn = a + nx = a + n =b
n
dc
c =y0 , . . . , yj = c + jy, . . . ym = c + my = c + m =d
m
where i and j are integers satisfying 0 i n and 0 j m. One can then move to a
point (xi , yj ) located within the rectangle R and construct a parallelepiped of height
f (xi , yj ) and base with sides xi = xi+1 xi and yj = yj+1 yj as illustrated in the
figure 3-14.

Figure 3-14. Element of volume dV = f (x, y) dy dx

252
A summation over the rectangle of these parallelepiped volume elements gives an
approximation to the volume bounded by the surface z = f (x, y), and the planes
x = a, x = b, y = c, y = d and z = 0. This approximation gets better and better as
xi 0 and yj 0. One finds that

n

m     b  

d b d
lim f (xi , yj ) xi yj = f (x, y) dx dy = f (x, y) dx dy
xi 0
yj 0 i=1 j=1 c a a c

where the inner integrals produce a slab, either in the x or y directions and the outer
integrals then represents a summation of these slabs giving the volume under the
surface. Note that if the surface z = f (x, y) oscillates above and below the plane z = 0,
then the result of the double integral gives a summation of the signedvolumes.
The orientation of the surface might be such that it is represented in the form
x = g(y, z), in which case the height of the surface is the distance x above the plane
x = 0. If the surface is represented y = h(x, z), then the height of the surface is
the distance y above the plane y = 0. Hence volume  
integrals can be represented
b1 d1
as double integrals having one of the forms V1 = f (x, y) dy dx if z = f (x, y)
a1 c1
 b2  d2
describes the surface, or V2 = g(y, z) dz dy if x = g(y, z) describes the surface or
a2 c2
 b3  d3
V3 = h(x, z) dz dx if y = h(x, z) describes the surface.
a3 c3
Summations over nonrectangular regions
If the smooth surface z = f (x, y) is defined over some nonrectangular region R
where the region R can be defined
(i) by a lower curve y = g1 (x) and upper curve y = g2 (x) between the limits
axb
(ii) or by a left curve x = h1 (y) and a right curve x = h2 (y) between the limits
cyd
as illustrated in the figure 3-15, then the volume is still obtained by a limiting
summation of the parallelepipeds constructed with base area dx dy and height f (x, y).
 253

Figure 3-15. Summation of dx dy over region R

The volume under the surface is similar to the case where the region R is a
rectangle, but instead the summations of the parallelepipeds are from one curve to
another curve. One can write either of the volume summations
   
x=b y=g2 (x)
f (x, y) dy dx = f (x, y) dy dx
R x=a y=g1 (x)
   
y=d x=h2 (y)
f (x, y) dx dy = f (x, y) dx dy
R y=c x=h1 (y)

The first inner integral sums in the vertical direction to create a slab and the outer
integral sums these slabs from a to b. The second inner integral sums in the horizontal
direction to form a slab and the outer integral sums these slabs from c to d.

Example 3-42.
If the surface (x x0 )2 + (y y0 )2 = r2 is a
cylinder in three-dimensions. If this cylinder is
cut by the planes z = 0 and z = h, a finite cylinder
is formed by the bounding surfaces. It is known
that the total volume V of this finite cylinder is
the area of the base times the height or V = r2 h.
Derive this result using double integrals.
Solution Here the region R is a circle of radius r centered at the point (x0 , y0 ) and

bounded by the upper semi-circle y = y0 + r2 (x x0 )2 and the lower semi-circle

y = y0 r 2 (x x0 )2 . The parallelepiped element of volume is located at position
(x, y) where the base area dx dy is constructed. The height of the parallelepiped to
be summed is h and so the parallelepiped element of volume is given by dV = h dy dx
where the element dy is written first because the inner integral is to be summed in
254
the ydirection from the lower semi-circle to the upper semi-circle. Summing these
volume elements first in the y-direction and then summing in the xdirection gives
  
x=x0 +r y=y0 + r 2 (xx0 )2
V = h dy dx (3.142)
x=x0 r y=y0 r 2 (xx0)2

Here the inner integral produces a slab and then these slab elements are summed
from x0 r to x0 + r. Perform the inner integration to obtain
 
x0 +r y0 + r 2 (xx0 )2 x0 +r 
V =h y dx = 2h r 2 (x x0 )2 dx (3.143)
x0 r y0 r 2 (xx 0 )2 x0 r

Here the integrand involves a difference of squares and suggests that one make the
trigonometric substitution x x0 = r cos with dx = r sin d to obtain
 0  

2 2 2 2 1
V =2h r 1 cos (r sin d) = 2hr sin d = 2hr (1 cos 2) d
0 0 2

  

2 1 1
V =hr d cos 2 2d = hr sin 2 = r 2 h
2
0 2 0 2 0

It is left as an exercise to perform the inner summation in the xdirection first, fol-
lowed by a summation in the ydirection and show that the same result is obtained.

Example 3-43. Evaluate the iterated integral of the function f (x, y) = xy2 over
the region between the parabola x = 2y2 and the line x = 2y.
Solution
Sketch the region over which the integration
is to be performed and then move to a general
point (x, y) within the region and construct an
element of area dx dy and determine which di-
rection is best for the inner integral. For
this problem the given curves intersect where
x = 2y 2 = 2y giving the points of intersection
(0, 0) and (2, 1). One can then construct the fig-
ure illustrated. Let us examine the integrals
   
 
x=2 y= x/2 y=1 x=2y
2
V = xy dy dx and V = xy 2 dx dy (3.144)
x=0 y=x/2 y=0 x=2y 2

where the top inner integral is a summation in the ydirection and the bottom inner
integral represents a summation in the xdirection. Now select the iterated integral
 255
which you think is easiest to integrate. For this problem, both integrals are about
the same degree of difficulty. For the first double integral in equation (3.144) one
finds  
 x=2 
y= y= x/2
x/2 x=2
y3
V = xy 2 dy dx = x dx
x=0 y=x/2 x=0 3 y=x/2
 3 
x=2  x 3
x x
V = dx
x=0 2 3 2
 2 
7/2 2
1 5/2 1 4 1 x 1 x5 4
V = x x dx = =
0 6 2 24 6 2 7/2 24 5 0 35

For the second double integral in equation (3.144) the inner integral is evaluated
first followed by an integration with respect to the outer integral to obtain
 y=1
 x=2y   1 2 x=2y  1
2 2x y2  
V = xy dx dy = y dy = (2y)2 (2y 2 )2 dy
y=0 x=2y 2 0 2 x=2y 2 0 2
 1  5 7
 y=1
y y 4
V = 2y 4 2y 6 dy = 2 2 =
0 5 7 y=0 35

Polar Coordinates

An element of area in polar coordinates can

be constructed by sketching the arbitrary rays
and +d together with the circular arcs of radius
r and r +dr as illustrated in the figure given. The
element of area dA can then be expressed as the
area of the sector with radius r + dr minus the
area of the sector with radius r. This can be
represented
1 1 1 2 1
dA = (r + dr)2d r 2 d = r + 2r dr + dr 2 r 2 d = r dr d + dr 2 d
2 2 2 2

The last term is 12 dr2 d is an infinitesimal of higher order and so this term can
be neglected. One then finds the element of area in polar coordinates is given by
dA = r dr d.
To find an area associated with a region bounded by given curves one can use
cartesian coordinates and write dA = dx dy as an element of area and perform sum-
mations in the xdirection and then the ydirection and express the total area as
256

A= with appropriate limits on the integrals. Alternatively, one can represent
dx dy
the element of area in polar coordinates as dA = r dr d and perform summations in

the rdirection and then the direction to express the total area as A = r dr d
with appropriate limits on the integrals.

Example 3-44. Find the area bounded by the lemniscate r2 = 2a2 cos 2.
Solution
Construct an element of area dA = r dr d in-
side the lemniscate and then make use of sym-
metry by calculating only the area in the first
quadrant. One can then represent the area in
the first quadrant by the double integral

 =/4  r= 2a2 cos 2  /4 2a2 cos 2
1 2
A= r dr d = r d
=0 r=0 0 2 0
 /4
a2 /4 a2
A =a2 cos 2 d = sin 2 =
0 2 0 2
 
a2
The total area under the lemniscate is therefore Atotal = 4 = 2a2 .
2

Cylindrical Coordinates

The coordinate transformation from carte-

sian coordinates (x, y, z) to cylindrical coordi-
nates (r, , z) is given by

x = cos = x2 + y 2
y = sin or = tan1 (y/x)
z =z z =z

where is a radial distance from the zaxis,

is an angular displacement measured from the
xaxis and is called the azimuth or azimuthal
angle and z is the height above the plane z = 0.
The element of volume in cylindrical coordinates
is given by dV = dddz
 257
Example 3-45. Find the volume of a cylinder of radius R and height H .
Solution Use the volume element in cylindrical coordinates and express the volume
as the triple integral    z=H =2 =R
V = dddz
z=0 =0 =0

Evaluating this integral from the inside-outward gives

 z=H  =2  z=H  =2  z=H
2 =R R2 R2 =2
V = d dz = d dz = dz
z=0 =0 2 =0 2 z=0 =0 2 z=0 =0
 z=H z=H
2
V =R dz = R2 z = R2 H
z=0 z=0

Spherical Coordinates
The coordinate transformation from cartesian coordi-
nates (x, y, z) to spherical coordinates (r, , ) is given by

x = r sin cos , y = r sin cos , z = r cos

for 0 and 0 2 , where r is called the radial

distance from the origin, is called the inclination from
the zaxis and is called the azimuth or azimuthal angle.
The element of volume in spherical coordinates is given
by
dV = (r sin d)(r d) dr = r 2 sin dr d d

Example 3-46. Find the volume of a sphere having a radius R.

Solution Using the element of volume for spherical coordinates, the volume of a
sphere is given by the triple integral
 =2  =  r=R
V = r 2 dr sin d d
=0 =0 r=0

Evaluating the triple integral from the inside-outward gives

 =2  =
r3 r=R
V = sin d d
=0 =0 r=0 3
 =2  =  =2
R3 R3 =
V = sin d d = ( cos ) d
3 =0 =0 3 =0 =0
 =2 =2
2 2 3 4 3
V = R3 d = R = R
3 =0 3 =0 3
258

Using Table of Integrals

The appendix C contains a table of integrals. In order to use these tables one
must sometimes make appropriate substitutions in order to convert an integral into
the proper form as given in the tables. For example, if it is required to evaluate the
integral 
e3x dx
I=
a + bex
where a and b are constants and you look for this integral in the table, you will not
find it. This is because the integral is listed under a different form. If you make the
substitution u = ex with du = ex dx The above integral can be written in the form

u2 du
I= where u = ex
a + bu
which is a form that you can locate in the tables. Onefinds under the listing for
x2 dx
integrals containing X = a + bx, the listing number 78 for . Here the variable x
X
in the table of integrals is just a dummy variable of integration and you can replace
x by u and write

u2 du 1  
I= = 3 (a + bu)2 4a(a + bu) + 2a2 ln(a + bu) + C
a + bu 2b
Use back substitution to express the integral in terms of ex .
The Bliss Theorem
In the previous pages there were times when summations were replaced by inte-
grals because of the fundamental theorem of integral calculus. The replacement of
summations by integrals can be also be justified by using the Bliss theorem. Gilbert
Ames Bliss (1876-1961) was an American mathematician who studied mathematics
his whole life. One of the results he discovered is known as Bliss theorem8. This
theorem relates summations and integrations and has the following geometric in-
terpretation. Consider an interval (a, b) which is partitioned into n-subintervals by
defining points
a = x0 < x1 < x2 < . . . < xi1 < xi < . . . < xn = b

and intervals xi = xi xi1 for i = 1, . . . , n. The spacing for the points xi need not
be uniform. Define

= maximum [x1 , x2 , . . . , xi , . . . , xn ]
8
Bliss,G.A., A substitute for Duhamels Theorem, Annals of Mathematics, Vol. 16 (1914-15), Pp 45-49.
 259
as the largest subinterval associated with the selected partition. Bliss showed that if
f (x) and g(x) are single-valued and continuous functions defined in the interval (a, b),
then for each subinterval xi one can select arbitrary points i and i inside or at
the ends of the subinterval (xi1 , xi) such that for each value i = 1, . . . , n one can write

xi1 < i < xi and xi1 < i < xi

One can then form the sum n

f (i )g(i)xi
i=1

and one can also form the integral

 b
f (x)g(x) dx.
a

Bliss theorem states that as n and 0, then

n

 b
lim f (i )g(i)xi = f (x)g(x) dx (3.145)
0
n i=1 a

Note that this important theorem allows one to replace summations over an
interval by integrations over the interval and the fundamental theorem of integral
calculus is a special case of this theorem. The above result is used quite often in
developing methods for finding answers to physical problems where discrete sum-
mations become continuous integrals as the number of summations increase.

Example 3-47.

A right circular cone is obtained by revolving the line y = hr x,

0 x h, about the x-axis. Divide the interval 0 x h into
n-parts each of length x = nh and then form circular disks at position xi = ix = ih n
r ir
having radius yi = h xi = n , for i = 1, 2, 3, . . ., n. Find the total volume V of the disks
in the limit as n .
n
Solution 1 Use the result from page 178, i=1 i2 = 16 (2n3 + 3n2 + n) and write
n n n


r2 h r2h
2 2n3 + 3n2 + n
V = lim yi2 x = lim 2 i2 = lim 3 i = r 2 h lim 3
= r2h
n
i=1
n
i=1
n n n n i=1 n 6n 3

Solution 2 Write  h  h
2 r2 2
V = y dx = 2
x dx = r 2 h
0 0 h 3
260
Exercises

3-1. Evaluate the given integrals.
  
cos t
(a) (2x + 1)3 dx (c) dt (e) (sin x)(cos x) dx
sin2 t
  

(b) sin 4x dx (d) ( 4 sin 2t) cos 2t dt (f ) (ax2 + bx + c) dx

3-2. Evaluate the given integrals.

  
x2
(a) (2 + sin 2t)4 cos 2t dt (c) x4 dx (e) sin(3x + 1) dx
  
(3 + 2x)
(b) dx (d) e3x dx (f ) x cos(3x2 + 1) dx
4 + 3x + x2

3-3. Evaluate the given integrals.

  
2 cos(3x2 )
(a) sec (3x + 4) dx (c) sec(3x + 4) tan(3x + 4) dx (e) x dx
sin(3x2 )
  
2 2
(b) csc (3x + 4) dx (d) (2x + 1) csc(x2 + x) cot(x2 + x) dx (f ) x ex dx

3-4. Evaluate the given integrals.

  
x dx x dx
(a) (c) (e) x cosh (x2 ) dx
1 x4 1 x2 
 
x dx
(b) (d) x sinh (x ) dx 2
(f ) x sech 2 (x2 ) dx
1 + x2

3-5. Evaluate the given integrals.

  
2 (3x + 1) dx
(a) csch (3x + 1) coth (3x + 1) dx (c) csch (3x + 1) dx (e) 
(3x + 1)2 1
  
(3x + 1) dx (3x + 1) dx
(b) sech (3x + 1) tanh (3x + 1) dx (d)  (f )
1 (3x + 1)2 1 (3x + 1)2

3-6. Evaluate the given integrals.

      
1 1 c 
(a) + + 5 dx (c) a+b t+ dt (e) u a + bu2 du
x2 x t
   

(b) x x2 + 25 dx (d) z 2z dz (f ) a + bu du

3-7. Evaluate the definite integrals and give a physical interpretation of what the
integral represents.
 3   B
H
(a) x2 dx (b) sin x dx (c) x dx
1 0 0 B
 261

3-8. If necessary use trigonometric substitution to evaluate the given integrals.
   
dx dx
(a)  (c) (e) a2 u2 du
(3x + 1) (3x + 1)2 + 1 (3x + 1)2 1
 
dx  2
dx a t2
(b)  (d)  (f ) dt
(3x + 1) 1 (3x + 1)2 9 (2x + 1)2 t2

3-9. For C1 , C2 constants, explain why the following integrals are equivalent.
 
du du
(a) = sin1 u + C1 , = cos1 u + C2
1 u2 1 u2
 
du du
(b) = tan1 u + C1 , = cot1 u + C2
1 + u2 1 + u2
 
du du
(c) = sec1 u + C1 , = csc1 u + C2
u u2 1 u u2 1
 
(d) tan u du = ln | cos u | +C1 , tan u du = ln | sec u | +C2
 
(e) cot u du = ln | sin u | +C1 , cot u du = ln | csc u | +C2

3-10. Evaluate the given integrals.

  
(a) sin2 (3x + 1) dx (c) sin3 (3x + 1) dx (e) sin4 (3x + 1) dx
  
2 3
(b) cos (3x + 1) dx (d) cos (3x + 1) dx (f ) cos4 (3x + 1) dx

3-11. Use partial fractions to evaluate the given integrals.

 
3x2 12x + 11 x dx
(a) dx (d)
(x 1)(x 2)(x 3) x2+ 4x 5
 
4x2 8x + 3 2x + 5x7 + 8x6 + 5x5 5x3 2x2 3x 1
8
(b) dx (e) dx
(x 1)2 (x 2) (x 1)(x2 + x + 1)2
 
3x2 + 3x + 2 x3 + x2 + x
(c) dx (f ) 2 dx
(x + 1)(x2 + 1) (x2 1)(x + 2)

3-12. Find the function y = y(x) passing through the given point (x0 , y0 ) whose
dy
derivative satisfies = f (x), if
dx
(a) f (x) = x, (x0 , y0 ) = (1, 3) (d) f (x) = tan2 (3x), (x0 , y0 ) = (0, 1)
(b) f (x) = x + 1, (x0 , y0 ) = (1, 2) (e) f (x) = sin2 (3x), (x0 , y0 ) = (0, 1)
(c) f (x) = sin 3x, (x0 , y0 ) = (0, 1) (f ) f (x) = cos2 (3x), (x0 , y0 ) = (0, 1)
dy
Note: An equation which contains a derivative, like = f (x), is called a differential
dx
equation. The above problem can be restated as, Solve the first order differential
dy
equation = f (x) subject to the initial condition y(x0 ) = y0 .
dx
262

3-13. If necessary use partial fractions to evaluate the given integrals.
  
x2 x 1 dx dx
(a) dx (c) 2 (e)
(x 2)(x 1)2 x(x + x + 1) x(x 1)2
  
x2 dx x2 dx dx
(b) (d) (f )
x2 + 1 (x + 1)2 x(a x)

3-14. Use partial fractions to evaluate the given integrals.

  
dx dx dx
(a) (c) (e)
(x a)(x b) (x a)2 (x b)2 (x a)(x2 + b2 )
  
dx dx dx
(b) (d) (f )
(x a)2 (x b) (x a)(x2 b2 ) (x a )(x2 + b2 )
2 2

3-15. Use integration by parts to evaluate the given integrals.

 1  /2

1
(a) sin x dx (c) x cos x dx (e) x sin 3x dx
0 0 0
  e  16
/2
(b) x sin x dx (d) x ln x dx (f ) x x 1 dx
0 1 1

3-16. Use integration by parts to evaluate the given integrals.

  
(a) x ex dx (c) x2 sin x dx (e) ex cos x dx
  
(b) x2 ex dx (d) ex sin x dx (f ) sec3 x dx

3-17. Use integration by parts to evaluate the given integrals.

   1
(a) ex
sin x dx (c) x ln(x + 1) dx (e) x2 ex dx
0
   1
x
(b) ln x dx (d) e cos x dx (f ) x3 ex dx
0

3-18. Use the fundamental theorem of integral calculus and express the given sums
as a definite integral.


1 1 2 n
(a) lim f( ) + f( ) + + f( )
n n n n n


1 2 n
(b) lim sin( ) + sin( ) + + sin( )
n n n n n
   
1 1 2 n
(c) lim + ++
n n n n n

3-19. Sketch the curves x = y2 6 and x = 4y 1. Find the area enclosed by these
curves.
 263

3-20. It has been found that to integrate rational functions of the sine and cosine
sin x x
functions, the change of variables u = = tan sometimes simplifies the
1 + cos x 2
integration problem.
x
(a) Show that if u = tan , then one obtains
2

2du 2u 1 u2
(i) dx = , (ii) sin x = , (iii) cos x =
1 + u2 1 + u2 1 + u2

dx
(b) Evaluate the integral
 1 + cos x
dx
(c) Evaluate the integral
sin x cos x

3-21.
The intersection of the curves y = x2 4, y = x2 + 4, y = 4 and y = 4 are
illustrated in the figure.
(a) Find the area with vertices ABH
(b) Find the area with vertices BDF H
(c) Find the area with vertices HABF GH
(d) Find the area with vertices ABDEF HA

d  1  6x 2 + 4x2

3-22. If tan (3x2 ) + ln(x4 + x2 ) = + , then find
dx 1 + 9x4 x + x3
  
6x 2 + 4x2
+ dx
1 + 9x4 x + x3

3-23. Evaluate the given integrals.

  
dx x x
(a) (c) e sin e dx (e) (ln x)2 dx
16 + 6x x2
     
1
(b) x2 + dx (d) x a2 x2 dx (f ) xex dx
1 x2

3-24. Evaluate the given integrals.

  
dt dx dx
(a) (c) (e)
t2 + 4t 3 (x + 9)2
2 (3x)2 9
  
dx dx dx
(b) (d) (f ) 
9 x2 (3x)2 + 9 9 (3x)2
264

3-25. Evaluate the given integrals.
  
dx 2
ax + b
(a) (c) sin (3) d (e) dx
x(1 + x) x2 c2
  
ex dx au + b ax + b
(b) (d) du (f )  dx
1 + 3ex u2 + c2 b (x + a)2
2

3-26. Evaluate the given integrals.

  
dx dx
(a) (c) (e) tan1 x dx
e + ex
x 1+ x
  
dx
(b) (d) eln x
dx (f ) sin1 x dx
x3 + 2 x

3-27. Evaluate the given integrals.

  
dx 2 x
(a) (c) (x + 1) e dx (e) sec1 x dx
a + bx
  
x dx 2 3/2
(b) (d) (1 x ) dx (f ) x sin1 x dx
(x 1)2

dy

3-28. Find the derivative if
dx
 (x)  0  (x)
(a) y(x) = f (t) dt (b) y(x) = f (t) dt, (c) y(x) = f (t) dt
0 (x) (x)

3-29. Sketch the curve defined by the set of points

C = { (x, y) | x = t sin t, y = 1 cos t, 0 t 2 }

If the curve C is rotated about the xaxis a surface of revolution isgenerated.

 2  2
dx dy
(i) Show the element of surface area can be expressed dS = 2y(t) + dt
dt dt
(ii) Find the surface area produced by the rotation.
(iii) Find the volume inside the surface.

3-30. Solve for the value of if

  +1   6
2 64 3 1
(a) x dx = (b) x dx = (c) x dx = x, dx
0 3 2 0 8
 
1 n1

3-31. Verify the reduction formula sin x dx = sinn1 x cos x +
n
sinn2 x dx
n n
 265

3-32. 
1 n
(a) Let In = xn ex dx and derive the reduction formula In = xn ex In1

(b) Evaluate the integral x ex dx

3-33.
(a) Find the arc length of the curve y = 23 x3/2
between the values x = 0 and x = 3.
(b) Find the arc length of the curve y = 41 x3 + 3x
1

between the values x = 1 and x = 4.

(c) Find the arc length of the curve defined by
the parametric equations
t3 t2
x= and y =
3 2

between the values t = 0 and t = 1.

3-34. 
n
(a) Let Jn = (ln | x |) dx and derive the reduction formula Jn = x (ln | x |)n nJn1

(b) Evaluate the integral ln | x | dx

3-35. 
n
(a) Let In,m = xm (ln x) dx and derive the reduction formula
1 n
Im,n = xm+1 (ln x)n Im,n1
 m+1 m+1
(b) Evaluate the integral x ln x dx

3-36.

(a) Consider the area bounded by the xaxis, the curve y = 2x + 1 and the lines
x = 0 and x = 3. This area is revolved about the xaxis.
(i) Find the surface area of the solid generated.
(ii) Find the volume bounded by the surface generated.

3-37. Consider the area bounded by the x-axis, the lines x = 1 and x = 8 and the
curve y = x1/3 . This area is revolved about the yaxis.
(i) Find the surface area of the solid generated.
(ii) Find the volume enclosed by the surface.
266

3-38. The

average value of a function y = y(x) over the interval [a, b] is given by
b
1
y = y(x) dx and the weighted average of the function y = y(x) is given by
ba a
 b
w(x)y(x) dx
yw = a b where w = w(x) is called the weight function.
a
w(x) dx
(a) Find the average value of y = y(x) = sin x over the interval [0, ].
(b) Find the weighted average of y = y(x) = sin x over the interval [0, ] with respect
to the weight function w = w(x) = x.
(c) Find the weighted average of y = y(x) = sin x over the interval [0, ] with respect
to the weight function w = w(x) = cos2 x
(d) Where does the weight function place the most emphasis in calculating the
weighted average in parts (b) and (c)?

3-39. Find the area under the given curves from x0 to x1 .


(a) y = x x2 + 1, x0 = 0, x1 = 2 (d) y = sin x, x0 = 0, x1 =
x
(b) y= , x0 = 0, x1 = 2 (e) y = cos x, x0 = 0, x1 = /2
2
x +1
(c) y = ln x, x0 = 1, x1 = e (f ) y = tan x, x0 = 0, x1 = /4

3-40. Consider the triangular area bounded by the xaxis, the line y = x, 0 x 1
and the line y = 2 x, 1 x 2. This area is revolved about the line x = 6 to form a
solid of revolution.
(i) Find the surface area of the solid generated.
(ii) Find the volume of the solid generated.

3-41. The line y = r = a constant, for 0 x h is rotated about the xaxis to form
a right circular cylinder of base radius r and height h.
(a) Use calculus and find the volume of the cylinder.
(b) Use calculus and find the lateral surface area of the cylinder.

3-42. Sketch the region of integration for the double integral

 1  4x
I= f (x, y) dy dx.
0 x

3-43. The line y = hr x for 0 x h is rotated about the xaxis to form a right
circular cone of base r and height h.
(a) Use calculus and find the volume of the cone.
(b) Use calculus and find the lateral surface area of the cone.
 267

3-44.
The radius r of a circle is divided into nparts by
defining a distance x = r/n and then constructing the
points
x0 = 0, x1 = r, x2 = 2r, . . . , xi = i r, . . . , xn = n r = r

The large circle can then be thought of as being com-

posed as a series of concentric circles. The figure on
the right shows the concentric circles constructed with
radii xi and xi+1 .
(a) Use calculus to find the circumference of a circle with radius r.
(b) Use calculus to sum the areas between concentric circles and find the area of a
circle. Hint: Show dA = 2x dx
(c) Use polar coordinates and double integrals to find the area of a circle.

3-45.
(a) Find common area of intersection associated with the
circles x2 + y2 = r02 and (x r0 )2 + y2 = r02
(b) Find the volume of the solid of revolution if this area
is rotated about the xaxis.

Hint: Make use of symmetry.

3-46. Sketch the region R over which the integration is to be performed

 1  y   3 cos  4  x2
(a) f (x, y) dx dy (b) f (r, ) r dr d (c) f (x, y) dy dx
0 y2 0 0 1 1

3-47. Sketch the region of integration, change the order of integration and evaluate
the integral.
     b
3 2 4 x d
axb
(a) 12xy dx dy (b) 3xy dy dx (c) xy dx dy,
1 y1 0 x/2 a c cyd

3-48. Integrate the function f (x, y) = 32 xy over the region R bounded by the curves
y=x and y2 = 4x
(a) Sketch the region of integration.
(b) Integrate with respect to x first and y second.
(c) Integrate with respect to y first and x second.
268

3-49. Evaluate the double integral and sketch the region of integration.
 1  x
I= 2(x + y) dy dx.
0 0

3-50. For

f (x) = x and b > a > 0, find the number c such that the mean value
b
theorem f (x) dx = f (c)(b a) is satisfied. Illustrate with a sketch the geometrical
a
interpretation of your result.

3-51. Make appropriate substitutions and show

 
dx 1 x dx 1 x+a
= tan1 + C = tan1 +C
2 + x2 b2 + (x + a)2 b b
 
dx 1 x dx 1 x+a
2 2
= tanh 1 + C, x< 2 2
= tanh 1 + C, x+a <b
x b (x + a) b b

3-52.  
nT T
(a) If f (x) = f (x + T ) for all values of x, show that f (x) dx = n f (x) dx
0 0
T a
(b) If f (x) = f (T x) for all values of x, show that f (x) dx = f (x) dx
a 0

3-53.
(a) Use integration by parts to show
   
1 ax b 1 ax b
eax sin bx dx = e sin bx eax cos bx dx and eax cos bx dx = e cos bx+ eax sin bx dx
a a a a

(b) Show that 

a sin bx b cos bx
eax sin bx dx = eax +C
a2 + b2

b sin bx + a cos bx
eax cos bx dx = eax +C
a2 + b2

3-54. Make an appropriate substitution to evaluate the given integrals

  
e4x + e3x
(a) (e2x + 3)m e2x dx (b) dx (c) (ex + 1)2 ex dx
ex + ex

3-55. Evaluate the given integrals

  
x+a (x + a)(x + b) (x + a)(x + b)(x + c)
(a) dx (b) dx (c) dx
x3 x3 x3

3-56. Evaluate the given integrals

  
x4 + 1
(a) ln(1 + x) dx (b) dx (c) (a + bx)(x + c)m dx
x1
 269

3-57. Use a limiting process to evaluate the given integrals
  x 
st t
(a) e dt, s > 0 (b) e dt (c) test dt, s > 0
0 0

3-58. Consider a function Jn (x) defined by an infinite series of terms and having
the representation


xn 1 x2 x4 (1)mx2m
Jn (x) = n + + + 2m +
2 n! 22 1!(n + 1)! 24 2!(n + 2)! 2 m!(n + m)!

where n is a fixed integer and m represents the mth term of the series. Here m takes
on the values m = 0, 1, 2, . . .. Show that
 x
J1 (x) dx = 1 J0 (x)
0

The function Jn (x) is called the Bessel function of the first kind of order n.


3-59. Determine a general integration formula for In = xn ex dx and then evaluate

the integral I4 = x4 ex dx

 a  a

3-60. Show that if g(x) is a continuous function, then g(x) dx = g(a x) dx
0 0

 2T  b

3-61. Let h(x) = h(2T x) for all values of x. Show that h(x) dx = h(x) dx
b 0


3-62. If for m, n positive integers one has Im,n = cosm x sin nx dx, then derive the
reduction formula
(m + n)Im,n = cosm x cos nx + m Im1,n1

3-63. If f (x) = f (x) for all values of xa, then f (x) is acalled an even function. Show
that if f (x) is an even function, then f (x) dx = 2 f (x) dx
a 0

3-64. If g(x) = g(x) for all valuesofa x, then g(x) is called an odd function. Show
that if g(x) is an odd function, then g(x) dx = 0
a

 2T  T

3-65. If f (2T x) = f (x) for all values of x, then show f (x) dx = 2 f (x) dx
0 0
270

3-66. Let A = A(y) denote the cross-sectional area of a pond at height y measured
from the bottom of the pond. If the maximum depth of the pond is h, then set up
an integral to represent volume of water in the pond.

3-67. Determine if the given improper integral exists. If the integral exists, then
evaluate the integral. Assume > 0 in parts (e) and (f).
 1  1  
dx dx dx if p < 1
(a) (c) (e)
1 x2 x 0 ( x)p if p 1
 1
0  
dx dx dx if p > 1
(b) (d) 2 +1 (f )
0 1 x2 x 0 ( + x)p if p 1

3-68. A particle moves around the circle x2 + y2 = r02 with constant angular velocity
of cm/s. Find the amplitude and period of the simple harmonic motion described
by (a) the projection of the particles position on the x-axis. (b) the projection of
the particles position on the y-axis.

3-69. Find the angle of intersection associated with the curves r = sin and r = cos
which occurs in the region r > 0 and 0 < < 2

3-70. Make use of symmetry when appropriate and sketch a graph of the following
curves.
a2 x x2
(a) y= (b) y= (c) y=
a2 x2 x + a2
2 x2 a2
b  b 

3-71. Let A1 = 0 sin xand
2b
dx A2 = 0 sin xb
dx
(a) Sketch the representation of the area A1 and evaluate the integral to find the
area A1 .
(b) Sketch the representation of the area A2 and evaluate the integral to find the
area A2 .
(c) Which area is larger?

3-72. A plane cuts a sphere of radius r forming a spherical cap of height h. Show

the volume of the spherical cap is V = h2 (3r h)
3

3-73. A solid sphere x2 + y2 + z 2 = r2 is placed in a drill press and a cylindrical hole
is drilled through the center of the sphere. Find the volume of the resulting solid if
the diameter of the drill is r, where 0 < < 1/4
   
dx x ax x

3-74. Show that  = 2 sin1 = 2 cos1 = 2 tan1
x(a x) a a ax
 271
Chapter 4
Sequences, Summations and Products
There are many different types of functions that arise in the application of
mathematics to real world problems. In chapter two many of the basic functions used
in mathematics were investigated and derivatives of these functions were calculated.
In chapter three integration was investigated and it was demonstrated that definite
integrals can be used to define and represent functions. Many of the functions
previously introduced can be represented in a variety of ways. An infinite series is
just one of the ways that can be used to represent functions. Some of the functions
previously introduced are easy to represent as a series while others functions are very
difficult to represent. In this chapter we investigate selected methods for representing
functions. We begin by examining summation methods and multiplication methods
to represent functions because these methods are easy to understand. In order to
investigate summation and product methods to represent functions, one must know
about sequences.
Sequences
A sequence is defined as a one-to-one correspondence between the set of positive
integers n = 1, 2, 3, . . . and a set of real or complex quantities u1 , u2 , u3 , . . ., which are
given or defined in some specific way. Such a sequence of terms is often expressed
as {un }, n = 1, 2, 3, . . . or alternatively by {un }
n=1 or for short just by {un }. The set of

real or complex quantities {un }, for n = 1, 2, 3, . . . is called an infinite sequence. The

indexing or numbering for the terms in the sequence can be selected to begin with
any convenient number. For example, there may be times when it is convenient to
examine sequences such as

{un }, n = 0, 1, 2, 3, . . . or {un }, n = , + 1, + 2, . . .

where is some convenient starting index.

A sequence of real numbers can be represented as a function or mapping from the
set of integers to the set of real numbers f : N R and is sometimes represented as
f (1), f (2), . . . or f1 , f2 , . . .. Similarly, a sequence of complex numbers can be represented
as a function or mapping from the set of integers to the set of complex numbers
f : N C.
272
Example 4-1. The following are some examples of sequences.



1 1 1 1 1 1
(a) , un = , 1, , , , . . . , , . . .
n n=1 n 2 3 4 n



1 1 1 1 1
(b) , un = , 1, 1, , , . . . , , . . .
n! n=0 n! 2 6 n!



4n 4n 20 4n
(c) , un = , 12, 8, , 6, . . ., ,...
n 2 n=3 n2 3 n2
 
 n  n 1 3 3 n
(d) (1)n sin , un = (1)n sin , , , 1, , . . . , (1)n sin ,...
6 n=1 6 2 3 2 6

Limit of a Sequence
The limit of a sequence, if it exists, is the problem of determining the value
of the sequence {un} as the index n increases without bound. A sequence such as
un = 3 + n4 for n = 1, 2, 3, . . . has the values

u1 = 7, u2 = 5, u3 = 13/3, u4 = 4, u5 = 19/5, u6 = 11/3, . . .

and as n increases without bound one can write

 
4
lim un = lim 3+ =3
n n n
This limit statement means that for n sufficiently large the values un are as close as
desired to the value 3.
Some sequences {un } do not have a limit as the index n increases without bound.
For example, the sequence un = (1)n for n = 1, 2, 3, . . . oscillates between the values
+1 and 1 and does not have a limit. Another example of a sequence which does
not have a limit is the sequence {vn } where vn = 3n for n = 1, 2, 3, . . .. Here the values
vn increase without bound as n increases.
Convergence of a sequence
A sequence {un }, n = 1, 2, 3, . . ., where un can be a real or complex number, is
said to converge to a number
or have a limit
, if to each small positive number
 > 0 there exists an integer N such that

|un
| <  for every integer n > N (4.1)

If the sequence converges, then the limit is written

lim un =
(4.2)
n

If the sequence does not converge it is said to diverge.

 273
Divergence of a sequence
A sequence {un } is said to diverge in the positive direction if for every number
M > 0, there exists a number N > 0, such that un > M for all integers n > N .
This is sometimes expressed limn un = . Similarly, a sequence {vn } is said to
diverge in the negative direction if there exists numbers M > 0 and N > 0 such that
vn < M for all integers n > N . This is sometimes expressed limn vn = .

Figure 4-1. Interpretation of convergence of a sequence {un }.

When the sequence converges, the elements un of the sequence tend to concen-
trate themselves around the point  for large values of the index n. The terms un do
not have to approach  at any specified rate nor do they have to approach  from a
particular direction. However, there may be times where un approaches  only from
the left and there may be other times when un approaches  only from the right.
These are just special cases associated with the more general definition of a limit
given above.
There are two geometric interpretations associated with the above limit state-
ment. The first whenever un = xn + i yn is a set of complex numbers and the second
geometric interpretation arises whenever un = xn represents a sequence of real num-
bers. These geometric interpretations are illustrated in the figure 4-1. In the case
274
the sequence {un } is a sequence of complex numbers, the quantity |z | <  repre-
sents an open disk centered at the point  = 1 + i 2 and the statement lim un = 
n
can be interpreted to mean that there exists an integer N , such that for all integers
m > N , the terms um are trapped inside the circular disk of radius  centered at . In
the case where the terms un = xn , n = 1, 2, 3, . . . are real quantities, the interpretation
of convergence is that for all integers m > N , the terms um are trapped inside the
interval ( ,  + ). These regions of entrapment can be made arbitrarily small by
making the quantity  > 0 small. In either case, there results an infinite number of
terms inside the disk or interval illustrated in the figure 4-1. A sequence which is
not convergent is called divergent or non-convergent.
Relation between Sequences and Functions
There is a definite relation associated with limits of sequences and limits of
functions. For example, if {un } is a sequence and f = f (x) is a continuous function
defined for all x 1 with the property that un = f (n) for all integers n 1, then if 
exists, the following limit statements are equivalent

lim un = , and lim f (x) = 

n x

One can make use of this property to find the limits of certain sequences.
ln n
Example 4-2. Evaluate the limit n
lim un where un =
n
Solution
ln x
Let f (x) = and use LHopitals rule to show
x

ln x 1/x 2
lim f (x) = lim = lim = lim = 0
x x x x 1/2 x x x

The limit properties for functions also apply to sequences. For example, if the
sequences {un } and {vn } are convergent sequences and n
lim vn
= 0, then one can write
lim (un vn ) = lim un lim vn
n n n

lim (un vn ) = lim un lim vn

n n n

un lim un
n
lim ( )=
n vn lim vn
n

If k is a constant and n
lim un = U exists, then the limit lim k un = k U also exists.
n
 275
Establish Bounds for Sequences
A real sequence {un }, n = 1, 2, 3, . . . is said to be bounded if there exists numbers
m and M such that m un M for all integers n. The number M is called an upper
bound and the number m is called a lower bound for the sequence.
The previous squeeze theorem1 from chapter 1 can be employed if there exists
three sequences {fj }, {gj }, {hj }, j = 1, 2, 3, . . . where the sequences {fj } and {hj } have
the same limit so that
lim fj =  and lim hj = 
j j

If one can verify the inequalities fj gj hj , for all values of the index j , then the
terms gj are sandwiched in between the values for fj and hj for all values j and
consequently one must have lim gj = .
j

n!
Example 4-3. Evaluate the limit lim un where un = , where n! is n-factorial.
n nn
Solution
An examination of un for n = 1, 2, 3, . . ., m, . . . shows that
21 321 m (m 1) 3 2 1
u1 = 1, u2 = , u3 = , um = ,
22 33 mm

Observe that the um term is positive and can be written as

 
m (m 1) 3 2 1 m (m 1) 3 2 1
um = =
m
 m
 m m m m mm m
m times
where the term within the parentheses is less than 1. Consequently, one can say um
is bounded with
1
0 < um
m
Here m1 0 as m increases without bound and so by the sandwich principle one can
say that um 0 as m increases without bound.

Every convergent sequence {un } is bounded. If n lim un = , then there exists

neighborhoods N
given by
(i) A circular neighborhood about  given by |z | < . This occurs when the values
un are complex numbers and there are an infinite number of values un from the
sequence inside the circle and a finite number of points un outside the circle.
1
Also known as the pinching lemma or the sandwich lemma
276
(ii) An interval neighborhood given by   < un <  + . Here the values un are real
numbers where there is an infinite number of points un inside the interval and
only a finite number of points outside the interval.
These neighborhoods N
represent bounded sets. By increasing the radius of the
bounded sets N
to encompass the finite number of points outside N
, a new set can
be constructed which will still be bounded and contain all the terms of the sequence.
If the real sequence {xj }, j = 1, 2, 3, . . . is a convergent sequence, then
(i) The sequence must be bounded and
(ii) The limit of the sequence is unique.
Make note that it is sometimes convenient to replace the small quantity  used
in the definition of convergence by some other small quantity such as 2 , 2
, or


M
, (M > 0 constant). The small quantity used is usually selected so that when
many applications of the inequality (4.1) are made, the final results add up to some
convenient number.
To show the sequence must be bounded, use the definition of convergence of a
sequence with a fixed value for . If the limit of the sequence is , then there exists
an integer N such that for integers n satisfying n > N
|xn | = |xn  + | |xn | + || <  + || = M1 for n > N
That is, if the sequence has a limit, then there exists a value N such that each term of
the sequence xN+1 , xN+2 , . . . are less then M1 in absolute value. Let us now examine the
terms x1 , x2, . . . , xN . Let M2 = max {|x1 |, |x2|, . . . , |xN |}. By selecting M = max{M1 , M2 },
it follows that |xj | < M for all values of j and so the sequence is bounded. If the
sequence is not bounded, then it diverges.
To show the limit of the sequence is unique, assume that there are two limits,
say  and 
. Now use the method of reductio ad absurdum to show this assumption
is false. The assumption of convergence of the sequence insures that for every  > 0,
there exists a value N1 such that for all integers n > N1 there results |xn | < 2
. If

is also a limit, then there must exist an integer N2 such that for all n > N2 there
results |xn 
| < 2
. Here the selection /2 has been used as the small quantity in the
definition of convergence rather than . The reason for this change is to make the
final answer come out in terms of the quantity . Define, N = max{N1 , N2}, then for
all integers n > N it follows that
 
| 
| = | xn + xn 
| | xn | + |xn 
| <+ =
2 2
and since  can be made arbitrarily small, then  must equal 
. Hence our original
assumption was false.
 277
Additional Terminology Associated with Sequences
1. A real sequence u1 , u2, u3, . . . is said to be monotone increasing
if u1 u2 u3 . . .. Alternatively, define a monotone increasing sequence {un }
as one where un+1 un for all values of n greater than some fixed integer N .
2. A real sequence u1 , u2 , u3 , . . . is said to be monotone decreasing
if u1 u2 u3 . . .. An alternative definition of a monotone decreasing sequence
{un } is that un+1 un for all values of n greater than some fixed integer N .
3. Consider an infinite sequence u1 , . . . , uj1 , . . . , uj2 , . . . , uj3 , . . . , un1, un, un+1 , . . . where
j1 < j2 < j3 < < jn < represent a selected subset from the real numbers N .
The sequence of numbers uj1 , uj2 , uj3 , . . . or {ujk } k=1 is called a subsequence of the

given sequence.
4. A sequence is called oscillating, either finite oscillatory or infinitely oscillatory,
depending upon whether the terms are bounded or unbounded. For example, the
sequence {cos n 3
}, for n = 1, 2, 3, . . ., is said to oscillate finitely, because the terms
remain bounded. In contrast, consider the sequence of terms {(1)nn2 } for the
values n = 1, 2, 3, . . . This sequence is said to oscillate infinitely, because the terms
become unbounded. In either case the sequence is called a nonconvergent se-
quence. A finite oscillatory sequence {xn }, n = 1, 2, 3, . . . is a sequence of real num-
bers which bounce around between finite limits and does not converge. An ex-
ample of a finite oscillatory sequence are the numbers {1, 0, 1, 1, 0, 1, 1, 0, 1, . . .}
with the pattern repeating forever. Oscillatory sequences occur in certain ap-
plied mathematics problems quite frequently.
5. A number L is called a limit point of the sequence {un }, n = 1, 2, 3, . . ., if for every
 > 0, | un L |<  for infinitely may values of n. A sequence may have more than
one limit point. For example, the sequence 1, 2, 3, 1, 2, 3, 1, 2, 3, . . . with the pattern
1, 2, 3 repeating forever, has the limit points 1, 2, 3. Note the following special
cases. (i) A finite set cannot have a limit point. (ii) An infinite set may or may
not have a limit point.
6. A real sequence {un }, n = 1, 2, 3, . . ., is called a null sequence if for every small
quantity  > 0 there exists an integer N such that |un | <  for all values of n > N .
7. Every bounded, monotonic sequence converges. That is, if the sequence is in-
creasing and bounded above, it must converge. Similarly, if the sequence is
decreasing and bounded below, it must converge. Another way of examining
these situations is as follows.
278
If {un } is an increasing sequence and the un are bounded above, then the set
SL = { un | n 1 } has a least upper bound L. Similarly, if {vn } is a decreasing
sequence and the vn are bounded below, the set of values SG = { vn | n 1 } has
greatest lower bound G. One can then show

lim un = L
n
and lim vn = G
n

since the sequences are monotonic.

8. Cauchys convergence criteria
The existence of a limit for the sequence n lim un , can be established by using
the Cauchy condition for convergence. A sequence {uj }, j = 1, 2, 3, . . ., is said to be
Cauchy convergent or to satisfy the Cauchy convergence criteria if for every small
quantity  > 0, there exists an integer N > 0, such that if one selects any two terms
from the sequence, say um and un , where n > N and m > n > N , then |um un | < .
The Cauchy criteria is another way of saying that for large values of m and n the
terms of the sequence will always stay close together. Also observe that in using the
Cauchy criteria it is not necessary to know the exact limit in order to demonstrate
convergence.
There are alternative ways for expressing the Cauchy convergence condition.
Some of these representations are the following.
(i) A sequence of points u1 = (x1 , y1 ), u2 = (x2 , y2 ), . . . , um = (xm, ym ), . . . , un = (xn , yn)
tends to a limit point if and only if, for every  > 0 there exists an integer N = N ()

such that (xm xn )2 + (ym yn )2 <  whenever m > N and n > N . That is, the
distance between the points um and un can be made as small as desired if m and
n are selected large enough.
(ii) A necessary condition for the existence of the limit n lim un , is for every  > 0,
there exists an integer N such that for all integers n > N and for every integer
k > 0, the condition |un+k un | <  is satisfied.

The Cauchy condition follows from the following argument. If the limit of the
sequence {un } is , then for every  > 0 one can find an integer N such that for
integers n and m both greater than N one will have
 
|un | < and |um | <
2 2

and consequently one can add and subtract  to obtain

 279
 
|un um | = |(un ) + ( um )| |un | + |um | < + =
2 2
In more advanced mathematics courses one can show that the Cauchy conver-
gence condition is both a necessary and sufficient condition for the existence of a
limit for the sequence {un }.

Example 4-4. If {un } is a monotone increasing sequence with

u1 u2 u3 un

and for each value of the index n one can show un K where K is some constant,
then one can state that the sequence {un } is a convergent sequence and is such that
lim un K .
n
To prove the above statement let SL = { x | x = un } and select for the set
SL a least upper bound and call it L so that one can state un L for all values
n = 1, 2, 3, . . .. Note that if L is the least upper bound, then every number K > L is
also an upper bound to the set SL . If  > 0 is a small positive number, then one can
state that L  is not an upper bound of SL , but L +  is an upper bound of the set
SL . Let N denote an integer such that L  < uN . Such an integer N exists since the
infinite set {un } is monotone increasing. Once N is found, one can write that for all
integers n > N one must have L  < uN < un < L + , which can also be written as
the statement

for all integers n > N the inequality |un L| < 

is satisfied. But this is the meaning of the limit statement lim un = L. Consequently,
n
one can state that the sequence is convergent and if each un K , then lim un K .
n

Stolz -Cesaro Theorem

Let {an }
n=1 and {bn }n=1 denote sequences of real numbers such that the terms

bn are strictly increasing and unbounded. The Stolz2 -Cesaro3 theorem states that if
the limit
an+1 an
lim =
n bn+1 bn

2
Otto Stolz (1842-1905) an Austrian mathematician.
3
Ernesto Cesaro (1859-1906) an Italian mathematician.
280
exits, then the limit
an
lim =
n bn

will also exists with the limit . This result is sometimes referred to as the LHopitals
rule for sequences.

A proof of the Stolz -Cesaro theorem is along the following lines. If the limit
an+1 an
lim =  exists, then for every  > 0 there must exists an integer N such
n bn+1 bn
that for all n > N there results the inequality
 
 an+1 an  an+1 an
  <  or   < <+
b
n+1 bn bn+1 bn

By hypothesis, {bn } is strictly increasing so that bn+1 bn
= 0 and consequently one

can write
( )(bn+1 bn ) < an+1 an < ( + )(bn+1 bn ) (4.3)

Let K denote a large number satisfying K > N and then sum each term in equation
(4.3) from N to K and show
K
 K
 K

( ) (bn+1 bn ) < (an+1 an ) < ( + ) (bn+1 bn )
n=N n=N n=N

which simplifies to

( )(bK+1 bN ) < aK+1 aN < ( + )(bK+1 bN )

By dividing each term by bK+1 one obtains

   
bN aK+1 aN bN
( ) 1 < < ( + ) 1
bK+1 bK+1 bK+1 bK+1

For N fixed and large enough values of K , the above inequality reduces to
aK+1
( ) < < ( + ) (4.4)
bK+1

because the other sequences are null sequences. The final equation (4.4) implies that
an
lim = .
n bn
 281
Examples of Sequences
The following table gives some examples of sequences.

Table 4.1 Example of Sequences

Example Sequence Comments
1 1, 0, 1, 0, 1, 0, 1, 0, . . . Divergent sequence with limit points 0 and 1
A bounded oscillating sequence
1 3 1 4
2 , , , , . . . , n1 , n1
4 4 5 5 n
,... Divergent sequence with limit points 0 and 1.
3 1, 2, 3, 4, . . . Divergent sequence.
4 0, 21 , 23 , 34 , 45 , 56 , 67 , . . . Convergent sequence with limit 1.
5 0, 32 , 45 , 67 , 89 . . . Convergent subsequence of previous sequence.
6 1, 21 , 13 , 14 , . . . A null sequence.
If | r |< 1, sequence converges.
7 {r n }
n=0 If r = 1, constant sequence with limit 1.
If r > 1 or r 1, sequence diverges.

Infinite Series
Consider the infinite series


un = u1 + u2 + u3 + + um + (4.5)
n=1

where the terms u1 , u2, u3 , . . . are called the first, second, third,. . . terms of the series.
The set of terms {un } usually represents a set of real numbers, complex numbers or
functions. In the discussions that follow it is assumed that the terms of the series
are represented by one of the following cases.
(i) um are real numbers um = m
(ii) um are complex numbers um = m + i m
(iii) um are functions of a real variable um = um (x)
(iv) um are functions of a complex variable um = um (z) for z = x + i y
for m = 1, 2, 3, . . .. Examine the cases (i) and (ii) above, where the terms of the infinite
series are constants.
282
The infinite series given by equation (4.5) is sometimes represented in the forms
 
un or un , where N denotes the set of integers {1, 2, 3, . . .}. This is done as a
nN
shorthand representation of the series and is a way of referring to the formal series
after it has been properly defined and no confusion arises as to its meaning. In
equation (4.5) the index n is called a dummy summation index. This summation
index can be changed to any other symbol and it is sometimes shifted by making
a change of variable. For example, by making the substitution n = k m, where m
is some constant, the summation index is shifted so that when n = 1, k takes on
the value m + 1 and the series given by equation (4.5) can be represented in the


alternative form ukm . Because k is a dummy index it is possible to replace k
k=m+1
by the original index n to obtain the equivalent representation


unm = u1 + u2 + u3 + , m is a constant integer. (4.6)
n=m+1

The indexing for an infinite series can begin with any convenient indexing. For
example, it is sometimes more advantages to consider series of the form



un = u0 + u1 + u2 + , or un = u + u+1 + u+2 + (4.7)
n=0 n=

where is some convenient starting index.

Sequence of Partial Sums


Given an infinite series uj = u1 + u2 + u3 + , form the sequence of terms
j=1
U1 , U2 , U3 , . . . defined by

U1 = u1 , U2 = u1 + u2 , Un = u1 + u2 + u3 + + un
m

where the finite sum Um = uj = u1 + u2 + + um represents the summation of
j=1
the first m terms from the infinite series. The sequence of terms {Um }, m = 1, 2, 3, . . .
is called the sequence of partial sums associated with the infinite series given by
equation (4.5). The notation of capital letters with subscripts or Greek letters with
subscripts is used to denote partial sums. For example, if the given infinite series

is j=1 aj , then the sequence of partial sums is denoted by the sequence {An } where
n
An = a1 + a2 + + an = j=1 aj for n = 1, 2, . . . or alternatively use the notation
n
n = j=1 aj .
 283
Convergence and Divergence of a Series



The infinite series uj is said to converge to a limit U , or is said to have a
j=1
sum U , whenever the sequence of partial sums {Un }, has a finite limit, in which
case one can write lim Un = U . If the sequence of partial sums {Un } becomes
n
unbounded, is oscillatory or the limit lim Un does not exist, then the infinite
n
series is said to diverge.

Example 4-5. Divergent finite oscillatory and infinite oscillatory

Consider the series (1)n = 1 1 + 1 1 + 1 1 + which has the partial sums
n=0
These partial sums oscillate between the finite values
U1 = 1, U2 = 0, U3 = 1, U4 = 0, .
of 0 and 1 and so the sequence of partial sums is called a finite oscillatory sequence.
In this case the series is said to diverge.



In contrast consider the series (1)n 2n = 1 2 + 4 8 + 16 32 + which has the
n=0
partial sums U1 = 1, U2 = 1, U3 = 3, U4 = 5, U5 = 11, . Here the sequence of partial
sums become infinite oscillatory and so the series is said to diverge.

Example 4-6. Harmonic series

1
Consider the harmonic series H = which has the nth partial sum4
m=1
m

1 1 1 1
Hn = 1 + + + ++
2 3 4 n

Nicole Oresme (1323-1382), a French mathematician and scholar who studied infinite
series, examined this series. His analysis considers the above finite sum, using the
value n = 2m, where he demonstrated that the terms within the partial sum Hn can
be grouped together and expressed in the form
 
1 1 1 1 1 1 1 1 1 1
Hn = 1 + + ( + ) + ( + + + ) + + m1
+ m1 ++ m
2 3 4 5 6 7 8 2 +1 2 +2 2
4
The nth partial sum of the harmonic series occurs in numerous areas of mathematics, statistics and probability
theory and no simple formula has been found to represent the sum Hn . The sums Hn are also known as harmonic
d
numbers, with H0 = 0. A complicated formula for Hn is given by Hn = + dz Log [(z)] where is the
Euler-Mascheroni constant and (z) is the Gamma function.
284
Observe that using a term by term comparison of the above finite sums one can state
Hn > hn , where n = 2m . The finite sum hn , becomes unbounded and so the harmonic
n
1
series diverges5 . Express the nth partial sum of harmonic series as Hn = and
m=1
m
express the harmonic series using the partial sums H2 , H4 , H8 , H16 , . . . by writing
       
1 1 1 1 1 1 1 1 1 1 1
H =1 + + + + + + + + ++ + ++ +
2 3 4 5 6 7 8 9 16 17 32
H = H2 + (H4 H2 ) + (H8 H4 ) + (H16 H8 ) + + (H2n Hn ) +

where
1 1 1 1 1 1 1
H2n Hn = + ++ > + ++ =
n+1 n+2 2n 2n 2n  2n 2
n terms
Hence, the 2nth partial sum can be written
H2n =H2 + (H4 H2 ) + (H8 H4 ) + + (H2n H2n1 )
3 1 1 1
H2n > + + + + n terms
2 2 2 2
3 1 1
H2n > + (n 1) = 1 + n
2 2 2

and 12 n increases without bound with increasing n and by comparison H2n also in-
creases without bound as n increases.
2
Note that the harmonic mean of two numbers n1 and n2 is defined as n = 1 1 .
n1
+ n2
The harmonic series gets its name from the fact that every term of the series, after
the first term, is the harmonic mean of its neighboring terms. For example, examining
1 1 1
the three consecutive sums + + from the harmonic series, one can show
m1 m m+1
1 1
that the harmonic mean of n1 = m1 and n2 = m+1 is given by n = m1 .

Example 4-7. Convergent Series

Triangular numbers, illustrated in the figure 4-2, were known to the Greeks.
The first few triangular numbers are 1, 3, 6, 10, . . . and one can verify that the nth
n(n + 1)
triangular number is given by the formula .
2

5
The harmonic series is a very slowly diverging series. For example, it would take a summation of over
1.509(10)43 terms before the sum reached 100.
 285

Figure 4-2. Triangular numbers.

 2 2 2 2 2
Consider the infinite sum = + + ++ + which
n=1
n(n + 1) 1 2 2 3 3 4 m(m + 1)
represents the sum of the reciprocal of the triangular numbers. The mth partial sum
of this series is given by
2 2 2 2
Um = + + ++
12 23 34 m(m + 1)

2 2 2
Observe that by partial fractions one can write = so that
m(m + 1) m m+1

2 2 2 2 2
Um = (2 1) + (1 ) + ( ) + + ( )
3 3 4 m m+1

This is called a telescoping series because of the way the terms add up. The resulting
sum is
2 2
Um = 2 with limit lim Um = lim (2 )=2
m+1 m m m+1
Consequently, the infinite series converges with sum equal to 2.

n

In general, one should examine the nth partial sum Un = uj of a given series
j=1
such as (4.5) to determine if the limit of the sequence of partial sums lim Un is
n
infinite, becomes finite oscillatory or infinite oscillatory or the limit does not exist,

then the series un is said to be a divergent series. Whenever the limit lim Un
n
exists with a value U , then U is called the sum of the series and the series is called
convergent. The convergence of the series can be represented in one of the forms

 N

un = lim Un = U = lim uj
n N
n=0 j=0
286
Comparison of Two Series
Consider two infinite series which differ only in their starting values



um = u1 + u2 + + u + u+1 + and un = u + u+1 + (4.8)
m=1 n=

where > 1 is an integer. Observe that it follows from the above definition for
convergence that if one of the series in equation (4.8) converges, then the other
series must also converge. Similarly, if one of the series from equation (4.8) diverges,
then the other series must also diverge. In dealing with an infinite series there are
many times where it is convenient to chop off or truncate the series after a finite
number of terms counted from the beginning of the series. One can then deal with
the remaining part. This is because the portion chopped off is a finite number of
terms representing some constant being added to the series. Consequently, it is
possible to add or remove a finite number of terms to or from the beginning of an
infinite series without affecting the convergence or divergence of the series.
Test For Divergence


If the infinite series un converges to a sum U , then a necessary condition for
n=1
convergence is that the nth term of the series approach zero as n increases without
bound. This necessary condition is expressed n lim un = 0. This requirement follows
from the following arguments.
n
 n1

If Un = ui is the nth partial sum, and Un1 = ui is the (n 1)st partial
i=1 i=1
sum, then for convergence, both of these partial sums must approach a limit U as n
increases without bound and so

lim Un = U and lim Un1 = U (4.9)

n n

By subtracting the nth and (n 1)st partial sums one obtains Un Un1 = un and
consequently
lim un = lim (Un Un1 ) = U U = 0 (4.10)
n n

The condition limn un = 0 is a necessary condition for convergence of the infinite



series ui . If this condition is not satisfied, then one can say the infinite series ui
i=1 i=1
diverges. The nth term of a series approaching zero as n increases without bound is
a necessary condition for any series to converge, but it doesnt guarantee convergence
of the series. If the nth term of the series approaches zero, then additional testing
 287
must be done to determine if the series converges or diverges. Take for example the
harmonic series 1 1 1 1 1 1
H= = 1+ + + + ++ +
k 2 3 4 5 n
k=1
1
considered earlier, one can observe that the nth term n
approaches zero, but further
investigation shows that the series diverges
Cauchy Convergence
The Cauchy convergence condition associated with an infinite series examines
the sequence of partial sums {Uj } for j = 1, 2, 3, . . . and requires for convergence that
for every given small number  > 0, there exists an integer N such that for any two
integers m and n satisfying n > m > N , one can show that |Un Um | < . Observe that
the sequence of partial sums Un and Um are written
n
 m

Un = uj and Um = uj , n>m
j=1 j=1

so that for Cauchy convergence of the sequence of partial sums it is required that
|Un Um | = |um+1 + um+2 + + un | be less than the given small quantity  > 0. This
test holds because if n
lim Un =  and limm Um = , then by definition of a limit one
can select integer values for n and m so large that one can write
 
| Un  |< and | Um  |<
2 2
where  > 0 is any small positive quantity and m and n are sufficiently large, say
both m and n are greater than N . It follows then that
 
| Un Um |=| (Un ) (Um ) || |Un  | + | Um  |< + =
2 2
The Cauchy test is an important test for convergence because it allows one to test
for convergence without actually finding the limit of the sequence.

Example 4-8. Geometric series

Consider the geometric series a + a r + a r2 + a r3 + + a rn + where a and r are
nonzero constants. The sequence of partial sums is given by
A1 =a

A2 =a + a r
.. ..
. .
An =a + a r + a r 2 + a r 3 + + a r n1
288
Recall that by multiplying An by r and subtracting the result from An one obtains
a a rn
(1 r)An = a a r n or An = (4.11)
1r 1r
The convergence or divergence of the sequence {An } depends upon the sequence {rn}.
Skipping the trivial case where r = 0, the sequence {An } converges if the sequence
{r n } converges. Consider the sequence {r n }, for n = 0, 1, 2, 3, . . . in the following cases
|r| < 1, r > 1, r = 1, r < 1 and r = 1.
1
(i) If |r| < 1, write |r| = 1+ where > 0. Using the binomial expansion show
1 1 1
n
= (1 + )n > 1 + n for n > 2, or r n = n
< and consequently for a
r (1 + ) 1 + n
given  > 0, with 0 <  < 1, write
1 1 1
|r n | = |r|n = < < for all n > N >
(1 + )n 1 + n 
1
Here 1+n 0 as n . This is an example of the sandwich theorem and
demonstrates limn rn = 0.
(ii) If r > 1, then rn for n = 1, 2, 3, . . . increases without bound. Consequently, for any
ln M
given positive number M , rn > M for all integers n > and so the sequence
ln r
{r n } diverges.
(iii) If r = 1, then rn = 1 for all integers n and the sequence for {An } diverges.
(iv) If r < 1 the sequence {rn } diverges since it becomes infinitely oscillatory with
r 2n + and r 2n+1 .
(v) The special case r = 1 also gives a finite oscillating sequence since r2n = 1 and
r 2n+1 = 1 for n = 1, 2, 3, . . . and so in this case the sequence {r n } diverges.
In summary, the geometric series has the finite sum given by equation (4.11)
and the infinite sum
a
A = lim An = a + a r + a r 2 + a r 3 + + a r n + = , for |r| < 1
n 1r
otherwise, the geometric series diverges.

The Integral Test for Convergence

Assume f (x) is a given function satisfying the following properties.
(i) f (x) is defined and continuous for x M for some positive integer M .
(ii) f (x) has the property that f (n) = un for all integers n M .
(iii) f (x) decreases as x increases which implies un = f (n) is a monotonic decreasing
function.
(iv) f (x) satisfies lim f (x) = 0.
x
 289

One can then say that the infinite series n=1 un

(a) converges if the integral M f (x) dx converges.

(b) diverges if the integral M f (x) dx diverges.

  T
where f (x) dx = lim f (x) dx is an improper integral.
M T M

The integral test for convergence of an infinite series compares the area under the
curve y = f (x) with overestimates and underestimates for this area. The following is

a proof of the integral test in the case M = 1. Given the infinite series n=1 un , with

un > 0 for all values of n, one tries to find a continuous function y = f (x) for 1 x <
which decreases as x increases and is such that f (n) = un for all values of n. It is
then possible to compare the summation of the infinite series with the area under
the curve y = f (x) using rectangles. This comparison is suggested by examining the
rectangles sketched in the figure 4-3.

Figure 4-3.
Overestimates and underestimates for area under curve y = f (x).

Assume there exists a function f (x) > 0 which decreases as x increases with the
 n+1
property f (n) = un and that limx f (x) = 0. The integral f (x) dx represents
n
the area under the curve y = f (x) bounded by the x-axis and the lines x = n and
x = n + 1. Using the mean value theorem for integrals the value of this integral is
f () for n < < n + 1.
290
The assumption f (x) decreases as x increases implies the inequality
 n+1
un = f (n) f (x) dx = f () f (n + 1) = un+1 for all values of n. (4.12)
n

The inequality (4.12) can now be applied to each interval (n, n + 1) to calculate over-
estimates and underestimates for the area under the curve y = f (x), for x satisfying
n x n + 1 and for n = 1, 2, 3, . . .. A summation of the inequalities given by equa-
tion (4.12), for n = 1, 2, . . . , N 1, gives a summation of areas under the curve and
produces the inequality
 N
u2 + u3 + + uN f (x) dx u1 + u2 + u3 + + uN1 (4.13)
1

A graphic representation of this inequality is given in the following figure.

 N
f (x) dx
1

Figure 4-4. Pictorial representation of the inequality (4.13)

The inequality (4.13) gives an underestimate and overestimate for the area under
the curve y = f (x) of figure 4-3. Consider the following cases.  N
Case 1: If in the limit as N increases without bound the integral lim f (x) dx exists,
N 1
say with value S , then the left-hand side of equation (4.13) indicates the se-
quence of partial sums is monotonic increasing and bounded above by S and
consequently the infinite series must converge.  N
Case 2: If in the limit as N increases without bound the integral lim f (x) dx is un-
N 1
bounded or the integral does not exist, then the right-hand side of the inequality
(4.13) indicates that the infinite series diverges.
 291
Example 4-9. The p-series

1 1 1 1 1
Consider the p-series which is defined H= p = p + p + p + p +
n=1
n 1 2 3 4
(Case I) In the case p = 1, the above series is called the harmonic series or the p-series
of order 1. Sketch the curve y = f (x) = 1/x and construct rectangular overestimates
for the nth partial sum. One can then verify that the nth partial sum of the harmonic
series satisfies
n 
1 1 1 1 n+1
1
1+ + ++ = dx = ln(n + 1)
2 3 n i 1 x
i=1

In the limit as n increases without bound the logarithm function diverges and so the
harmonic series diverges using the integral test.
(Case II) In the case p 0, the p-series diverges because it fails the test of the nth
term approaching zero as n increases without bound.
1
(Case III) In the case p is positive and p
= 1 use the function f (x) = p and show
x
  T   
1 1 1 1
dx = lim dx = lim 1 (4.14)
1 xp T 1 xp T p 1 T p1

1
If p > 1 the right-hand limit from equation (4.14) has the value p1 and so the
integral exists and consequently the p-series converges. If 0 p < 1 the limit on the
right-hand side of equation (4.14) diverges and so by the integral test the p-series
also diverges.
In conclusion, the p-series

 1 1 1 1 1
H= p
= p
+ p
+ p
+ +
n=1
n 1 2 3 4p
converges for p > 1 and diverges for p 1

Example 4-10. Estimation of error


Let un denote a given infinite series which converges by the integral test.
n=1
If U denotes the true sum of this series and Un denotes its nth partial sum, then
|U Un | = Rn is the remainder term that was omitted and represents the true error
in using Un as an approximate value for the sum of the series.
Let En denote an estimate for the error associated with the truncation of an
infinite series after the nth term where the nth partial sum Un is being used as an
292
estimate for the true value U of the series. The function f (x) used in the integral
test for convergence is a decreasing function of x and so

|Rn | f (x) dx
n

If one sets En = n f (x) dx, then the magnitude of the remainder |Rn| En or the
true error is less than the estimated error En.
 N
Here un = f (n) so that the partial sum UN = N n=1 un = k=1 f (k) can be used as

an approximation to the infinite sum. A better approximation for the sum is given
 
by UN = N k=1 f (k) + N f (x) dx since the integral representing the tail end of the area

under the curve in figure 4-3 is a good approximation to the sums neglected.

Example 4-11.
1 1 1
Sum 100 terms of the infinite series S = 2 + 2 + + 2 + and estimate the
1 2 m
error associated with this sum.
Solution: Let Sn denote the nth partial sum

n
1 1 1 1
Sn = 2
= 2 + 2 ++ 2
m=1
m 1 2 n

Use a computer and verify that S100 = 1.63498 so that an estimate for the error
between the finite sum and the infinite sum is given by

1
E100 = dx = 0.01
100 x2

Therefore, one can state that the difference between the true sum and approximate
sum is |S S100 | < 0.01 or 1.62498 < S < 1.64498. The exact value for S is known to be
2 /6 = 1.64493... and this exact value can be compared with our estimate. Observe
 1
that the value S100 + 100 x2 dx gives a better estimate for the sum.
It is important that you make note of the fact that the integral test does not give
the sum of the series. For example,
 
1 2 1
S= = and dx = 1
m=1
m2 6 1 x2
 293
Alternating Series Test
An alternating series has the form


(1)j+1uj = u1 u2 + u3 u4 + u5 u6 + , uj > 0 (4.15)
j=1

where each term of the series is positive, but the sign in front of each term alter-
nates between plus and minus. An alternating series converges if the following two
conditions are satisfied.
(i) For a large enough integer N, the terms un of the series are decreasing in absolute
value so that |un+1 | |un|, for all values of n > N.
(ii) The nth term approaches zero as n increases without bound so that one can
write n
lim un = 0 or lim |un | = 0.
n

To prove the above statement one can examine the sequence of partial sums UN ,
starting with N = 1, and make use of the fact that un+1 un to obtain the situation
illustrated in the figure 4-5.

partial sums

Figure 4-5.
Sequence of partial sums for alternating series.

The even sequence of partial sums U2 , U4 , U6 , . . . forms a bounded monotone in-

creasing sequence. The odd sequence of partial sums U1 , U3 , U5 , . . . forms a bounded
monotone decreasing sequence. Both the even {U2n } and odd {U2n+1 } sequence of
partial sums converge and consequently

lim U2n = U
n
and lim U2n+1 = V
n
294
Note also that for n large, the term u2n+1 of the series must approach zero and
consequently
lim u2n+1 = lim (U2n+1 U2n ) = lim U2n+1 lim U2n = V U = 0
n n n n

which implies that U = V and the alternating series converges.

Example 4-12. Alternating series


(1)n
Consider the series . For n + 1 > n, then ln(n + 1) > ln n and consequently
n=2
ln n
1 1 1
< indicating the terms un = decrease as n increases or equivalently,
ln(n + 1) ln n ln n
1
the sequence {un } is monotonic decreasing. In addition, lim = 0. These two
n ln n
conditions guarantee the convergence of the given alternating series.

Example 4-13. Estimation of error

Denote by Um the mth partial sum of a convergent alternating series and let U
denote the true sum of the series. For n > m examine the difference between the nth
partial sum and mth partial sum by writing
   
n m
    n 
   
|Un Um | =  (1)i+1ui (1)i+1ui  =  (1)i+1ui 
   
i=1 i=1 i=m+1

Using an appropriate selection of the value n (i.e. being either even or odd and
greater than m) one can write
 
 
n 
 
 (1)i+1ui  = um+1 (um+2 um+3 ) (um+4 um+5 ) (un1 un ) < um+1
 
i=m+1

This inequality is independent of the value of n so that

 
 n
 
 i+1 i+1 
|U Un | =  (1) ui (1) ui  < un+1
 
i=1 i=1

This last inequality is sometimes referred to as the Leibniz condition and implies that
by selecting an error En as the absolute value of the (n + 1) st term of an alternating
series, then one can write |Rn| = |U Un | En. Hence, to obtain the sum of an
alternating series accurate to within some small error  > 0, one must find an integer
value n such that |un+1 | = En+1 < , then it can be stated with confidence that the nth
partial sum Un and the true sum U of the alternating series satisfies the inequality
Un  < U < Un + .
 295
Bracketing Terms of a Convergent Series


Let A = a0 + a1 + a2 + + am + = an denote a convergent series and define
n=0
the sequence of terms {j }
=1 which is a strictly increasing sequence of nonnegative

integers. That is the terms j1 , j2 , . . . are positive integers satisfying

j1 < j2 < < jn < jn+1 <

The series for A can then be bracketed into nonoverlapping groups or partitions as
follows

A = (a0 + + aj1 ) + (aj1 +1 + + aj2 ) + (aj2 +1 + + aj3 ) + + (ajn +1 + + ajn+1 ) +

where there is a finite number of terms in each group. This is equivalent to defining
the infinite series
b0 =a0 + a1 + + aj1

b1 =aj1 +1 + aj1 +2 + + aj2

 .. ..
B= bn where . .
n=0
bn =ajn +1 + ajn +2 + + ajn+1
.. ..
. .

The partial sums Bn = b0 + b1 + + bn of the bracketed series, by definition, must

equal the partial sum Ajn+1 of the original series and consequently the sequence of
partial sum {Bn} is a subsequence of the partial sums {An }. In advanced calculus it
is shown that every subsequence of a convergent sequence converges and from this
result it can be concluded that
If a convergent series A has its terms bracketed into nonoverlapping groups to
form a new series B, then the series B converges to the sum of the original series.
The converse of the above statement is not true. For example, the bracketed
series (1 1) + (1 1) + (1 1) + converges, but the unbracketed series has partial
sums which oscillate and hence is divergent.
296
Example 4-14.
Consider the infinite series

(1)n1 1 1 1 1
A= = 1 + + + (1)m1 +
n=1
n 2 3 4 m

and say this series is bracketed into groups of two terms as follows
     

1 1 1 1 1 1
B= 1 + + + =
2 3 4 5 6 n=1
(2n 1)(2n)

Here both series converge to the same value as the bracketing does not effect the
convergence of a converging series.

Comparison Tests



Consider two infinite series, say un and vn where the terms of the series
n=1 n=1
un and vn are nonnegative. Let M denote a positive integer, then

(i) if un vn for all integers n > M and the infinite series n=1 vn converges, then

the infinite series n=1 un is also convergent.


(ii) if un vn 0 for all integers n > M and if the infinite series vn diverges,
n=1


then the infinite series un must also diverge.
n=1

To prove statement (i) above, let n=1 vn denote a convergent series with sum
V and let
Un = u1 + u2 + + un , and Vn = v1 + v2 + + vn
 
denote the nth partial sums associated with the infinite series un and vn respec-
tively. If un vn and V is the value of the converging series, then the sequence of
partial sums {Un } and {Vn} satisfy Un Vn V and consequently the sequence {Un }
is an increasing bounded sequence which must converge.


If the series vn diverges, then the sequence of partial sums {Vn } increases
n=1
without bound. If un vn for all n, then Un Vn for all n and so Un must also


increase without bound, indicating that the series un must also diverge.
n=1
 297
Ratio Comparison Test



If un is a series to be compared with a known convergent series cn , then
n=1 n=1
if the ratios of the (n + 1)st term to the nth term satisfies
un+1 cn+1
, (4.16)
un cn


then the series un is a convergent series.
n=1



If un is a series to be compared with a known divergent series dn , then if
n=1 n=1
the ratios of the (n + 1)st term to the nth term satisfies
un+1 dn+1
, (4.17)
un dn


then the series un is a divergent series.
n=1
To prove the above statements, make the assumption that the inequalities hold
for all integers n 0. The proofs can then be modified to consider the cases where
the inequalities hold for all integers n N . The proof of the above statements follows
by listing the inequalities (4.16) and (4.17) for the values n = 0, 1, 2, . . ., (m 1). This
produces the listings
u1 c1 u1 d1

u0 c0 u0 d0
u2 c2 u2 d2

u1 c1 u1 d1
u3 c3 u3 d3
(4.18)
u2 c2 u2 d2
.. .. .. ..
. . . .
um cm um dm

um1 cm1 um1 dm1

Multiply the terms on the left-hand sides and right-hand sides of the above listings
and then simplify the result to show
u0 u0
um cm , and um dm (4.19)
c0 d0

A summation of the terms on both sides of the above inequalities produces a com-
parison of the given series with known convergent or divergent series multiplied by
some constant.
298
Example 4-15. Comparison test (set up an inequality)
Two known series used quite frequently for comparison with other series are the

 
1
geometric series ar n and the p-series p
. These series are used in comparison
n=0 n=1
n
tests because inequalities involving powers of known quantities are easy to construct.

 1
For example, to test the series for convergence or divergence, use the
n=1
n2 +1
comparison test with the known p-series. The inequality n2 + 1 > n2 implies that
1 1
2
< 2 and consequently by summing both sides of this inequality one finds
n +1 n

 
1 1
< . It is known that the p-series, with p = 2, converges and so by
n=1
n2 + 1 n=1
n2

comparison the given series converges.

 a
As another example, consider the series where a > 0 and b > 0 are
n=1
3n +b

1
constants and compare this series with the geometric series . Since 3n + b > 3n ,
n=1
3n

a a a
then < n and consequently, converges, so that the given series must
3n + b 3 n=1
3 n

also converge.

Example 4-16. Estimation of error


Let un denote a convergent infinite series, then use the nth partial sum Un to
n=1
estimate the true value U of the series. The true error in using the nth partial sum
as an estimated value for the sum is given by


Rn = |U Un | = un
i=n+1

where Rn is the remainder after the summation of n-terms of the series.


If there exists a known comparison series vn such that for all integers n greater
n=1
than some fixed integer N there results the inequality |un | vn , then it follows that
|un+1 + un+2 + + un+m | |un+1 | + |un+2 | + + |un+m |

vn+1 + vn+2 + + vn+m vi
i=n+1
 299
 n+m 
  

 
which implies that for all n > N , |Rn | = |U Un | = lim  ui  vi . Conse-
m  
i=n=1 i=n+1


quently if one selects En = vi as an error estimate for the series sum, then write
i=n+1
which says the true error must be less than the estimated error En obtained
|Rn | En
by summation of terms greater than vn from the known comparison series.

Absolute Convergence



Consider the two series an and | an | where the second series has terms
n=1 n=1
which are the absolute value of the corresponding terms in the first series. By


definition the series an is called an absolutely convergent series if the series of
n=1


absolute values | an | is a convergent series.
n=1

Example 4-17.
1 1 1
(a) The series S1 = 1 + + is called the alternating harmonic series. By the
2 3 4
alternating series test it is a convergent series. It is not an absolutely convergent
series because the series of absolute values is the harmonic series which is a
known divergent series.
1 1 1 1
(b) The series S2 = 2 2 + 2 2 + is an absolutely convergent series because
1 2 3 4
the corresponding series of absolute values is the p-series or order 2 which is a
known convergent series.

Example 4-18.



If |uj | is an absolutely convergent series, then the series uj must also be
j=1 j=1
a convergent series.
To prove the above statement examine the sequence {U n } where U n denotes the
nth partial sum associated with the series of absolute values
n

U n = |u1 | + |u2 | + + |un | = |uj | for n = 1, 2, 3, . . .
j=1

For convergence of the series of absolute values the Cauchy convergence criteria
requires that there exist an integer N such that for all integers n and m satisfying,
n > m > N , one has
300

|U n U m | = ||um+1 | + |um+2 | + + |un || <  (4.20)

where  > 0 is any small number. Select the value N large enough that one can apply


the Cauchy convergence criteria to the infinite series uj for the same given value
j=1
of  > 0 using the same values of m and n. For Cauchy convergence of the series

 m

uj , it is required that the difference of the mth partial sum Um = uj and nth
j=1 j=1
n

partial sum, Un = uj , for n > m, must satisfy |Un Um| = |um+1 + um+2 + + un| < .
j=1
Using the generalized triangle inequality, the absolute value of a sum is less than or
equal to the sum of the absolute values. That is,

|um+1 + um+2 + + un | ||um+1 | + |um+2 | + + |un || < . (4.21)

But this is the Cauchy condition which is required for convergence of the infinite


series |uj |.
j=1



Another proof is to consider the two series un and |un | with partial sums
n=1 n=1

Un =u1 + u2 + u3 + + un
and U n =|u1 | + |u2 | + |u3 | + |un |

Our hypothesis is that the sequence of partial sums {U n } converges. Our problem
is to show that the sequence of partial sums {Un } also converges. Assume the series

n=1 un has both positive and negative terms so that the partial sum {Un } can be

written as Un = Pn Nn where Pn is the sum of the positive terms and Nn is the

sum of the absolute values of the negative terms within the nth partial sum. This
implies that the partial sum U n can be expressed U n = Pn + Nn. Here the sequence
{U n } is a bounded increasing sequence with some limit limn U n = U which implies
the inequalities
Pn U n U and Nn U n U

so that the limit U is an upper bound for the sequences {Pn } and {Nn}. Both the
sequences {Pn } and {Nn} are monotone increasing and bounded sequences and must
 301
converge to some limiting values. If these limiting values are called P and N , then
one can employ the limit theorem from calculus to write

lim Un = lim (Pn Nn ) = lim Pn lim Nn = P N

n n n n

which shows the infinite series n=1 un is a convergent series.

If n=1 an is an absolutely convergent series, then write
 
N  N

 
lim  an  lim |an |
N   N
n=1 n=1

as this is just a limit associated with the generalized triangle inequality.

Also note the terms within an absolutely convergent series can be rearranged
with the resulting series also being absolutely convergent with the same sum as the
original series.

Slowly Converging or Slowly Diverging Series

In using the comparison test to determine whether a series converges or diverges,
one should be aware that some series converge very slowly while other series diverge
very slowly. To estimate the value of a very slowly convergent series to within some
error bound, it is sometimes necessary to sum an excessive number of terms.
  
If an andbn are two convergent series, one says that the series an con-
 bn
verges at a slower rate than the series bn if the condition lim = 0 is
n an
satisfied.
In a similar fashion
  
If an and bn are two divergent series, one says that the series an diverges
 an
at a slower rate than the series bn if the condition lim = 0 is satisfied.
n bn

Example 4-19.


 1
The series an = will converge at a slower rate than the series
n=2 n=2
n(ln n)2
1

 
1 bn n2 (ln n)2
bn = 2
because lim = lim
1
= lim =0
n=2 n=2
n n an n n n
n(ln n)2

This result follows by using LHopitals rule to evaluate

(ln x)2 2(ln x) x1 2x1
lim = lim = lim =0
x x x 1 x 1
302
Example 4-20.


 1
The series an = diverges at a slower rate than the comparison series
n=2 n=2
n ln n
1

 
1 an 1
bn = , because lim = lim n ln n = lim = 0.
n n bn n 1 n ln n
n=2 n=2
n
As an exercise estimate how many terms of each series needs to be summed for the
 31
result to be greater than 3. Verify that the answer is 8718
n=2 an > 3 and n=2 bn > 3

Ratio Test
The following tests investigate the ratio of certain terms in an infinite series
as the index of the terms increases without bound. The ratio test is sometimes
referred to as the dAlemberts test, after Jean Le Rond dAlembert (1717-1783) a
French mathematician. The ratio test examines the absolute

value of the ratio of

the (n + 1)st term divided by the nth term of the series ui as the index n increases
i=1
without bound.

dAlembert ratio test

If the terms un are different from zero for n = 1, 2, 3, . . . and the limit
 un+1 
lim   = q exists 6 , then
n un  
(i) The series n=1 un is absolutely convergent if q < 1.

(ii) The series n=1 un diverges if q > 1.
(iii) The test fails if q = 1.
The above result can be proven using the geometric series. Assume that for all
values of n greater than some value N it is possible to show that
 
 un+1 
 
 u  r < 1, (4.22)
n

Define the quantities

v0 =uN+1, v1 = uN+2, v2 = uN+3,

and use the inequality given by equation (4.22) to show |vm| r|vm1 |. This is
accomplished by setting n = N +1, N +2, . . . in equation (4.22) to obtain the inequalities
6
If all the terms of the series are nonnegative, then the absolute value sign can be removed
 303
|v1 | r|v0 |

|v2 | r|v1 | r 2 v0
.. .. (4.23)
. .
|vm | r|vm1 | r m v0

The original series can then be split into two parts. The first part N j=1 |uj | is a

finite series leaving the series j=N+1 |uj | representing the second part which can be
compared with a geometric series. The second part satisfies the inequality


 |v0 |
|uj | |vi | |v0 |(1 + r + r 2 + r 3 + ) , |r| < 1 (4.24)
1r
j=N+1 i=0

and so the series is absolutely convergent by the comparison test.

If for all values of n greater than some value N it is possible to show
 
 un+1 
 
 u =q>1 (4.25)
n

then the terms of the series un become a monotonic increasing sequence of positive
numbers and consequently the nth term cannot approach zero as n increases without
bound. Under these conditions the given series is divergent.

Example 4-21. Ratio test


m+1 2 3 4 n+1
Consider the series m
= + 2 + 3 ++ n +
m=1
2 2 2 2 2
n+2  
un+1 2 n+1 1 n+2 1
Using the ratio test one finds lim = lim = lim = <1 and so
n un n n + 1 n 2 n+1 2
2n
the given series is absolutely convergent.
Example 4-22. Ratio test

1
Test the harmonic series for convergence using the ratio test. The ratio test
n=1
n
1
un+1 n+1 n 1
produces the limit n
lim = lim 1 = lim = lim = 1 and so the
un n
n
n n + 1 n 1 + 1/n

ratio test fails and so some other test must be used to investigate convergence or
divergence of the series.
Example 4-23. Ratio test

en
Test the series to determine if the series converges. Using the ratio test
n=1
n!
en+1
un+1 (n+1)! e
one finds lim = lim en = lim =0<1 and so the given series converges.
n un n n n+1
n!
304
Root Test
 
n
If the limit lim |un | = L exists, then the series un is
n
n=1
(i) absolutely convergent if L < 1
(ii) is divergent if L > 1
(iii) The test fails if L = 1


n
Note that if n
|un | < q < 1 for all n > N , the |un | < q < 1 so that the series |ui |
 i=N
converges by comparison with the geometric series. If n |un | > q > 1, the nth term of
the series does not approach zero and so the series diverges.
Certain Limits
Three limits that prove to be very useful are the following.
n
1. If > 0 and is any real number, then n
lim = 0. This limit follows by
(1 + )n
 
x x x
examining the function f (x) = = x ln(1+) = x .
(1 + )x e e ln(1+)
(ln n)
2. If is any real number and > 0 is real, then lim = 0. This limit follows
 n
 n

(ln x) ln x
by examining the function g(x) =
= .
x x/
3. A consequence of the ratio test is that if all the terms of the sequence {un } are
  
 un+1 
such that un
= 0 and the limit limn  un  < 1, then the series un is absolutely
convergent, which implies that limn un = 0 as this is a necessary condition for
convergence. Hence, the ratio test can be used to investigate certain limits which
approach zero.

Example 4-24. Limits


n1000 n1000
If un = , investigate the infinite series un and find the limit n
lim
(1 + )n n=1
(1 + )n
where  > 0 is a constant.
Solution: If one tests the ratio
(n + 1)1000
 1000  1000
un+1 (1 + )n+1 n+1 1 1 1
= 1000
= = 1+
un n n (1 + ) n (1 + )
n
(1 + )
un+1 1
in the limit as n increases without bound, one finds n
lim = < 1 and so
 un (1 + )
the sum un converges. The convergence of the series implies that the nth term
n1000
approaches zero, so that lim = 0, provided  > 0.
n (1 + )n
 305
Power Series
A series of the form


cn xn = c0 + c1 x + c2 x2 + c3 x3 + + cm xm +
n=1

is called a power series centered at the origin. Here x is a variable and the terms
c0 , c1 , . . . , cm , . . . are constants called the coefficients of the power series. If x is assigned
a constant value, the power series then becomes a series of constant terms and it
can be tested for convergence or divergence. One finds that in general power series
converge for some values of x and diverge for other values of x. If the power series
converges for |x| < R, then R is called the radius of convergence of the power series.
A series having the form


cn (x x0 )n = c0 + c1 (x x0 ) + c2 (x x0 )2 + c3 (x x0 )3 + + cm (x x0 )m +
n=1

is called a power series in (x x0 ) centered at the point x0 with coefficients cm for

m = 0, 1, 2, . . ..

Example 4-25. Power series

Consider the power series


an (x x0 )n = a0 + a1 (x x0 ) + a2 (x x0 )2 +
n=0

where the coefficients an , for n = 0, 1, 2, . . . are constants and independent of the

value selected for x. This series converges for certain values of x and diverges for
other values of x. For convergence, the ratio test is required to satisfy the limiting
condition
     
 un  n   
lim   = lim  an (x x0 )  = lim  an (x x0 )  = |x x0 | < 1
n u  n  a n1  n  an1 
n1 n1 (x x0 ) R
 
 an1 
lim 
where R = n  . Consequently, the power series is absolutely convergent when-
an 
ever |xx0| < R and is divergent whenever x satisfies |xx0 | > R. For real variables, the
power series converges for x satisfying R < x x0 < R and the interval (x0 R, x0 + R)
is called the interval of convergence for the power series and R is called the radius
of convergence of the power series. For complex variables the region of convergence
is the interior of the circle |z z0 | < R in the complex plane.
306
A power series


f (x) = an (x x0 )n = a0 + a1 (x x0 ) + a2 (x x0 )2 + + am (x x0 )m + (4.26)
n=0

will satisfy one of the following conditions.

 an1 
lim 
(i) If n  = 0, then the infinite series (4.26) converges only for x = x0 and

an
diverges for all other values of x.
 an1 
lim 
(ii) If n  = , then the infinite series (4.26) converges absolutely for all

an
values of x. 
a 
(iii) If lim  n1  = R exists and is nonzero, then the infinite series (4.26) converges
n an
for | x x0 |< Rand diverges for | x x0 |> R.
In general, if a power series


f (x) = ai (x x0 )i = a0 + a1 (x x0 ) + a2 (x x0 )2 + + an (x x0 )n +
i=0

exists and converges for | x x0 |< R, then one can say

(i) The function f (x) is a continuous function on the interval |x x0 | < R.
(ii) The function f (x) has the derivative

f
(x) = a1 + 2a2 (x x0 ) + 3a3 (x x0 )2 + + nan (x x0 )n1 +

and this series for the derivative converges over the same interval |x x0 | < R.
(iii) The function f (x) has the integral given by

a1 a2 an
f (x) dx = a0 (x x0 ) + (x x0 )2 + (x x0 )3 + + (x x0 )n+1 +
2 3 n+1

plus some arbitrary constant of integration can be added to this result. The
series for the integral also converges over the interval |x x0 | < R.
Operations with Power Series


Two power series given by f (x) = fn (x x0 )n with radius of convergence Rf
n=0
n

and g(x) = gn (x x0 )n with radius of convergence Rg can be added
n=0



a(x) = f (x) + g(x) = fn (x x0 )n + gn (x x0 )n
n=0 n=0
 307
or they can be subtracted



b(x) = f (x) g(x) = fn (x x0 )n gn (x x0 )n
n=0 n=0

by adding or subtracting like powers of (x x0 ). The resulting series has a radius of

converge R = smaller of { Rf , Rg }.
The product of the power series for f (x) and g(x) can be expressed



 
f (x)g(x) = fn (x x0 )n gn (x x0 )n
n=0 n=0


= fn gk (x x0 )n+k
n=0 k=0


 n

= fj gnj (x x0 )n
n=0 j=0


= cn (x x0 )n
n=0

n

where cn = fj gnj is the Cauchy product, sometimes referred to as the convolution
j=0
of the sequences fn and gn .
The two power series for f (x) and g(x) can be divided and written


fn (x x0 )n
f (x) 
= n=0
= hn (x x0 )n
g(x) 
gn (x x0 )n n=0

n=0

where the coefficients hn are related to the coefficients fn and gn by comparison of

the coefficients of like powers of (x x0 ) on both sides of the expression




  
fn (x x0 )n = gn (x x0 )n hn (x x0 )n
n=0 n=0 n=0

Maclaurin Series
Let f (x) denote a function which has derivatives of all orders and assume the
function and each of its derivatives has a value at x = 0. Also assume that the
function f (x) can be represented within some interval of convergence by an infinite
series of the form

f (x) = c0 + c1 x + c2 x2 + c3 x3 + + cn xn +
308
where c0 , c1, c2 , . . . , cn, . . . are constants to be determined. If the above equation is to
be an identity, then it must be true for all values of x. Substituting x = 0 into the
equation gives f (0) = c0 . The series can be differentiated on a term by term basis as
many times as desired. For example, one can write
f
(x) =c1 + 2c2 x + 3c3 x2 + 4c4 x3 +
f

(x) =2!c2 + 3!c3 x + 4 3c4 x2 + 5 4x3 +

f


(x) =3!c3 + 4!c4 x + 5 4 3x2 +

.. ..
. .
f (n) (x) =n!cn + (n + 1)!cn+1x +
..
.

Substituting x = 0 into each of the above derivative equations gives the results

f

(0) f


(0) f (n) (0)

c1 = f
(0), c2 = , c3 = , , cn = ,
2! 3! n!

This shows that f (x) can be represented in the form

f

(0) 2 f


(0) 3 f (n) (x) n

f (x) = f (0) + f
(0)x + x + x ++ x + (4.27)
2! 3! n!

or
 f (m) (0) m
f (x) = x (4.28)
m=0
m!

where f (0) (0) = f (0) and 0! = 1 by definition. The series (4.27) is known as a Maclau-
rin7 series expansion of the function f (x) in powers of x. This type of series is useful
in determining values of f (x) in the neighborhood of the point x = 0 since if |x| is less
than 1, then the successive powers xn get very small for large values of n.
In the special case f (x) = g(x + h) one finds, for h constant, the derivatives
f (x) = g
(x + h), f

(x) = g

(x + h), etc. Evaluating these derivatives at x = 0 gives

f (0) = g(h), f
(0) = g
(h), f

(0) = g

(h), etc., so that the equation (4.27) takes on the
form
x2 x3 xn
g(x + h) = g(h) + g
(h)x + g

(h) + g


(h) + + g (n) (h) + (4.29)
2! 3! n!
which is called Taylors form for Maclaurins results.
7
Colin Maclaurin (1698-1746) a Scottish mathematician.
 309
Example 4-26. Some well known Maclaurin series expansions are the following.

x3 x5 x7 x9 x11
sin x =x + + + |x| <
3! 5! 7! 9! 11!
x2 x4 x6 x8 x10
cos x =1 + + + |x| <
2! 4! 6! 8! 10!
x3 x5 x7 x9 x11
sinh x =x + + + + + + |x| <
3! 5! 7! 9! 11!
x2 x4 x6 x8 x10
cosh x =1 + + + + + + |x| <
2! 4! 6! 8! 10!
x2 x3 x4 x5 x6
ex = Exp(x) =1 + x + + + + + + |x| <
2! 3! 4! 5! 6!
(x ln a)2 (x ln a)3
ax = ex ln a =1 + x ln a + + + <x<
2! 3!
x2 x3 x4 x5 x6
ln(1 + x) =x + + + 1x<1
2 3 4 5 6
1
=1 + x + x2 + x3 + x4 + |x| < 1
1x
x2 x3
(1 + x) =1 + x + ( 1) + ( 1)( 2) + |x| < 1
2! 3!
3 5 7
1 x 1 3 x 1 3 5 x
sin1 x =x + + + + 1<x<1
2 3 24 5 246 7 
1 1 1 x3 1 3 x5 1 3 5 x7
cos x = sin x = x + + + + 1<x<1
2 2 2 3 24 5 246 7

Note that many functions do not have a Maclaurin series expansion. This occurs
whenever the function f (x) or one of its derivatives cannot be evaluated at x = 0.
For example, the functions ln x, x3/2 , cot x are examples of functions which do not
have a Maclaurin series expansion.

Example 4-27. The following are series expansions of selected special func-
tions occurring in advanced mathematics, engineering, mathematical physics and
the sciences.
J (x) The Bessel function of the first kind of order
x


(1)k x2k
J (x) =
2 22k k! ( + k + 1)
k=0

where (x) is the gamma function.

Y (x) The Bessel function of the second kind of order
J (x) cos() J (x)
Y (x) =
sin()
310

Figure 4-6. The Bessel functions J0 (x), Y0 (x), J1 (x), Y1 (x)

x sin t
The sine integral Si(x) = 0 t
dt

 (1)n x2n+1
Si(x) =
n=0
(2n + 1)(2n + 1)!

The cosine integral

x cos t1
Ci(x) = + ln x + 0 t
dt, | arg x| <


(1)nx2n
Ci(x) = + ln x +
n=1
2n(2n)!
 
where = limn 1 + 12 + 13 + + n1 ln n =
0.57721 . . . is called the Euler-Mascheroni con-
stant.

2
x 2
The error function erf(x) =
0
et dt

2  (1)nx2n+1
erf(x) =
n=0 n! (2n + 1)
 311
The hypergeometric function F (, ; ; x)

 k k xk
F (, ; ; x) =
k k!
k=0

where a 0 = 1 and a k = a(a + 1)(a + 2) (a + k 1) for k a nonnegative integer,

is called the factorial rising function8 or upper factorial. In expanded form the
hypergeometric function is written
x ( + 1)( + 1) x2
F (, ; ; x) =1 + + +
1! ( + 1) 2!
( + 1) ( + n 1)( + 1) ( + n 1) xn
+ +
( + 1) ( + n 1) n!

The hypergeometric function is related to many other functions. Some example

relationships are the following.
ln(1 x) 1
F (1, 1 ; 2 ; x) = F (, ; ; x) =
x (1 x)
1 3 tan1 x x
F ( , 1 ; ; x) = lim F (, ; ; ) =ex
2 2 x
1 1 1
F (, ; ; sin2 x) = cos(2x) F ( , 1 ; ; x2) =1 x2
2 2 2

Taylor and Maclaurin Series

Brook Taylor (1685-1731) an English mathematician and Colin Maclaurin (1698-
1746) a Scottish mathematician both studied the representation of functions f (x) in
terms of a series expansion in powers of the independent variable x. If f (x) is a real
or complex function which is infinitely differentiable in the neighborhood of a fixed
point x0 , then f (x) can be represented in the form

f (x) = Pn (x, x0 ) + Rn (x, x0) (4.30)

where Pn (x, x0 ) is called a nth degree Taylor polynomial centered at x0 and Rn (x, x0)
is called a remainder term. The Taylor polynomial of degree n has the form
f
(x0 ) f

(x0 ) f (n) (x0 )
Pn (x, x0 ) = f (x0 ) + (x x0 ) + (x x0 )2 + + (x x0 )n (4.31)
1! 2! n!
8 k
There is a factorial falling function or lower factorial defined by a = a(a 1)(a 2) (a (k 1)) for
k a nonnegative integer. There are alternative notations to represent the factorial rising and falling functions. Some
texts use the notation x(n) for the rising factorial function and the notation (x)n for the falling factorial function.
(x + n) (x + 1)
In terms of gamma functions one can write x(n) = x n = n
and (x)n = x = .
(x) (x n + 1)
312
and the remainder term is represented
 x
1
Rn (x, x0) = (x t)n f (n+1) (t) dt (4.32)
n! x0

If the point x0 = 0, then the series is called a Maclaurin series.

One method of deriving the above series expansion is to assume that a given
function f (x) has derivatives of all orders and these derivatives all have a finite value
at some point x0 . Also assume the function f (x) can be represented by a convergent
infinite series of the form

f (x) = c0 + c1 (x x0 ) + c2 (x x0 )2 + + cn (x x0 )n + (4.33)

where c0 , c1, . . . are constants to be determined. Substituting x = x0 into this equation

gives f (x0 ) = c0 . The assumed series representation can be differentiated on a term
by term basis as many times as desired. The first n-derivatives are
f
(x) =c1 + 2c2 (x x0 ) + 3c3 (x x0 )2 + 4c4 (x x0 )3 +
f

(x) =2!c2 + 3!c3 (x x0 ) + 4 3c4 (x x0 )2 +

f


(x) =3!c3 + 4!c4 (x x0 ) +

.. ..
. .
f (n) (x) =n!cn + (n + 1)!cn+1(x x0 ) +

Substituting the value x = x0 into each of the above derivatives produces the results
f

(x0 ) f


(x0 ) f (n) (x0 )
c1 = f
(x0 ), c2 = , c3 = , , cn = ,
2! 3! n!

This shows that f (x) can be represented in the form

f

(x0 ) f (n) (x0 )
f (x) = f (x0 ) + f
(x0 )(x x0 ) + (x x0 )2 + + (x x0 )n + (4.34)
2! n!

or
 f (m) (x0 )
f (x) = (x x0 )m (4.35)
m=0
m!

which is known as a Taylor series expansion of f (x) about the point x0 . Note that
when x0 = 0 the Taylor series expansion reduces to the Maclaurin series expansion.
The validity of the infinite series expansions given by the Maclaurin and Taylor
series is related to the convergence properties of the resulting infinite series. In
general, the Taylor series given by equation (4.33) will satisfy one of the following
 313
conditions (i) The infinite series converges for all values of x (ii) the series converges
only when x = x0 or (iii) The infinite series converges for x satisfying | x x0 |< R and
diverges for | xx0 |> R, where R > 0 is a real number called the radius of convergence
of the power series. Note that in the case where there is a radius of convergence R
and x is an endpoint of the interval (x0 R, x0 + R), then the infinite series may or
may not converge. Usually the ratio test, and the root test are used to determine
the radius of convergence of the infinite series. The endpoints of the interval of
convergence must be tested separately to determine convergence or divergence of
the series.
Using the mean value theorem for integrals the remainder term can be reduced
to one of the forms
(x x0 )n+1
Rn (x, x0) =f (n+1) (1 ) , (4.36)
(n + 1)!
f (n+1) (2 )(x 2 )n (x x0 )
or Rn (x, x0) = (4.37)
n!

where 1 , 2 are constants satisfying x0 < 1 < x and x0 < 2 < x. The equation (4.36)
is known as the Lagrange form of the remainder term and equation (4.37) is known
as the Cauchy form for the remainder term.
Another method to derived the above results involves integration by parts. Con-
sider the integral  x
f (x) f (x0 ) = f
(t) dt (4.38)
x0

where x0 and x are held constant. An integration of the right-hand side is performed
using integration by parts with U = f
(t), dU = f

(t) dt and dV = dt and V = t x. Here
x is treated as a constant of integration so that
 x x
 x



f (t) dt =f (t)(t x) (t x) f

(t) dt
x0 x0 x0
 x  x (4.39)
f
(t) dt =f
(x0 )(x x0 ) + (x t) f

(t) dt
x0 x0

Now evaluate the integral on the right-hand side of equation (4.39) using integration
by parts to show
 x  x


(x x0 )2


(x t)2



f (t) dt = f (x0 )(x x0 ) + f (x0 ) + f (t) dt (4.40)
x0 2! x0 2!
314
Continue to use integration by parts n-times to obtain
 x
(x x0 )2 (x x0 )n
f
(t) dt = f
(x0 )(x x0 ) + f

(x0 ) + + f (n) (t) + Rn (x, x0 ) (4.41)
x0 2! n!

where the remainder term is given by

 x
1
Rn (x, x0 ) = (x t)n f (n+1) (t) dt (4.42)
n! x0

Example 4-28. Some additional Series Expansions

x3 2x5 (1)n122n (22n 1)B2n x2n1

tan x =x + + ++ + |x| <
3 15 (2n)! 2
3 5 n1 2n 2n1
1 x x 2x (1) 2 B2n x
cot x = 0 < |x| <
x 3 45 945 (2n)!
x2 5x4 61x6 (1)n E2n x2n
sec x =1 + + + ++ + |x| <
2 24 720 (2n)! 2
3 5 n1 2n1 2n1
1 x 7x 31x (1) 2(2 1)B2n x
csc x = + + ++ + 0 < |x| <
x 6 360 15, 120 (2n)!
x3 2x5 22n (22n 1)B2n x2n1
tanh x =x + ++ + |x| <
3 15 (2n)! 2

Where Bn are the Bernoulli numbers and En are the Euler numbers. These numbers
are defined9 from the expansions
x B1 x B2 x2 B4 x4 B6 x6 B2n x2n
=1 + + + + ++ +
ex 1 1! 2! 4! 6! (2n)!
x 1 1 x2 1 x4 1 x6
=1 x + +
ex 1 2 6 2! 30 4! 42 6!
2ex E1 x E2 x2 E3 x3
=E0 + + + +
e2x + 1 1! 2! 3!
2ex x2 x4 x6 x8
=1 + 5 61 + 1385
e2x + 1 2! 4! 6! 8!

and produce the numbers

1 1 1
B0 =1, B1 = 1/2, B2 = , B4 = , B6 = , , B2n+1 = 0 for n > 1
6 30 42
E0 =1, E2 = 1, E4 = 5, E6 = 61, , E2n+1 = 0 for n = 0, 1, 2, . . .

9
There are alternative definitions of the Bernoulli and Euler numbers which differ by subscripting notation,
signs and scale factors.
 315
Taylor Series for Functions of Two Variables
Using the above results it is possible to derive a Taylor series expansion asso-
ciated with a function of two variables f = f (x, y). Assume the function f (x, y) is
defined in a region about a fixed point (x0 , y0 ), where the points (x0 , y0 ) and (x, y) can
be connected by a straight line. Such regions are called connected regions. Further,
let f (x, y) possess nth-order partial derivatives which also exist in the region which
surrounds the fixed point (x0 , y0 ). The Taylors series expansion of f (x, y) about the
point (x0 , y0 ) is given by
f (x0 , y0 ) f (x0 , y0 )
f (x0 + h, y0 + k) =f (x0 , y0 ) + h+ k
x y
  (4.43)
1 2 f (x0 , y0 ) 2 2 f (x0 , y0 ) 2 f (x0 , y0 ) 2
+ h + 2 hk + k +
2! x2 xy y 2
where h = x x0 and k = y y0 . This expansion can be represented in a simpler form
by defining the differential operator

D=h +k , h and k are constants.
x y
The Taylor series can then be represented in the form
n
1 j
f (x0 + h, y0 + k) = D f (x, y) + Rn+1 , (4.44)
j=0
j!

where all the derivatives are evaluated at the point (x0 , y0 ). The remainder term can
be expressed as
1
Rn+1 = D(n+1) f (x, y), to be evaluated at (x, y) = (, ) (4.45)
(n + 1)!
where the point (, ), lies somewhere on the straight line connecting the points
(x0 + h, y0 + k) and (x0 , y0 ).
The equation (4.43) or (4.44) is derived by introducing a new independent vari-
able t which is the parameter for the straight line defined by the equations
dx dy
x = x0 + ht, y = y0 + kt, with = h, and =k
dt dt
where h and k are constants and 0 t 1. Consider the function of the single variable
t defined by
F (t) = f (x, y) = f (x0 + ht, y0 + kt)

which is a composite function of the single variable t. The composite function can
be expanded in a Maclaurin series about t = 0 to obtain
t2 tn t(n+1)
F (t) = F (0) + F
(0)t + F

(0) + + F (n) (0) + F (n+1) () , 0 < < t. (4.46)
2! n! (n + 1)!
Evaluation of equation (4.46) at t = 1 gives f (x0 + h, y0 + k).
316
The first n derivatives of the function F (t) are calculated using chain rule differ-
entiation. The first derivative is
f (x, y) dx f (x, y) dy
F 0 (t) = +
x dt y dt
(4.47)
f (x, y) f (x, y)
= h+ k.
x y

By differentiating this expression, the second derivative can be determined as

2 f (x, y) 2 f (x, y) dx
 
00
F (t) = h+ k
x2 y x dt
 2 2

f (x, y) f (x, y) dy
+ h+ k
x y y 2 dt

or
2 f (x, y) 2 2 f (x, y) 2 f (x, y) 2
F 00 (t) = h + 2 hk + k . (4.48)
x2 x y y 2
Continuing in this manner, higher derivatives of F (t) can be calculated. For
example, the third derivative is
3f 2 3f 3 f 2 dx
 
000
F (t) = h + 2 2 hk + k
x 3 x y xy 2 dt
 3 3 3

f 2 f f 2 dy
+ 2
h +2 2
hk + k
x y xy y 3 dt

or
3f 3 3f 2 3f 2 3f 3
F 000 (t) = h + 3 h k + 3 hk + k , (4.49)
x 3 x 2 y xy 2 y 3

Using the operator D = h +k a pattern to these derivatives can be constructed
x y
 
0
F (t) = Df (x, y) = h +k f (x, y)
x y
 2
00 2
F (t) = D f (x, y) = h +k f (x, y)
x y
 3

F 000 (t) = D3 f (x, y) = h +k f (x, y)
x y
.. ..
. .
 n

F (n) (t) = Dn f (x, y) = h +k f (x, y).
x y
 317
 n
n
Here the operator D = h +k can be expanded just like the binomial ex-
x y
pansion and
  n
 
(n) n f n n n1 nf n n2 2 n f
F (t) = D f (x, y) = h + h k + h k
xn 1 xn1y 2 xn2y 2
  (4.50)
n nf n
n f
++ hkn1 + k ,
n1 xy n1 y n
 
n n!
where = are the binomial coefficients.
m m!(n m)!
In order to calculate the Maclaurin series about t = 0, each of the derivatives
must be evaluated at the value t = 0 which corresponds to the point (x0 , y0 ) on the
line. Substituting these derivatives into the Maclaurin series produces the result
given by equation (4.43), where all derivatives are understood to be evaluated at
the point (x0 , y0 ).
In order for the Taylor series to exist, all the partial derivatives of f through
the nth order must exist at the point (x0 , y0 ). In this case, write f C n over the
connected region containing the points (x0 , y0 ) and (x, y). The notation f C n is
read, f belongs to the class of functions which have all partial derivatives through
the nth order, and further, these partial derivatives are continuous functions in the
connected region surrounding the point (x0 , y0 ).

In a similar fashion it is possible to derive the Taylor series expansion of a

function f = f (x, y, z) of three variables. Assume the Taylor series expansion is to be
about the point (x0 , y0 , z0 ), then show the Taylor series expansion has the form
n
 1 j
f (x0 + h, y0 + k, z0 + ) = D f (x, y, z) + Rn+1 (4.51)
j!
j=0

where    
f f f
Df = h +k + f= h +k + (4.52)
x y x x y x
is a differential operator and h = x x0 , k = y y0 and  = z z0 . After expanding the
derivative operator Dj f for j = 0, 1, 2, . . ., each of the derivatives are to be evaluated
at the point (x0 , y0 , z0 ). The term Rn+1 is the remainder term given by
1
Rn+1 = D(n+1) f (x, y, z) (4.53)
(n + 1)! (x,y,z)=(,,)
318
where the point (, , ) is some unknown point on the line connecting the points
(x0 , y0 , z0 ) and (x0 + h, y0 + k, z0 + ).
Functions of n-variables f = f (x1 , x2 , . . . , xn ) have their Taylor series expansions
derived in a manner similar to the above by employing a differential operator of the
form  

D= h1 + h2 + + hh (4.54)
x1 x2 xn
where h1 = x1 x10 , h2 = x2 x20 , . . . , hn = xn xn0 .

Example 4-29. Hypergeometric series

The hypergeometric series defines a power series in x in terms of three parameters
a, b, c as
ab x a(a + 1)b(b + 1) x2 a n b n xn
F (a, b ; c ; x) = 1 + + ++ +
c 1! c(c + 1) 2! c n n!
where
n1

a n = a(a + 1)(a + 2)(a + 3) (a + n 1) = (a + i)
i=0

is called the rising factorial function and the symbol is used to denote a product
of terms as the index i ranges from 0 to n 1. Apply the ratio test and examine the
ratio of successive terms and show
a n+1 b n+1 xn+1
un+1 c n+1 (n + 1)! (a + n)(b + n) x
= n n n
=
un a b x (c + n) (n + 1)
c n n!

Divide the numerator and denominator by n2 and show

  
un+1 1 + na 1 + nb
=   x
un 1 + nc 1 + n1

and in the limit as n increases without bound one obtains the limit x for the ratio of
successive terms. Hence, in order for the series to converge it is required that |x| < 1.

Alternative Derivation of the Taylor Series

The above results for the representation of f (x) as a power series can also be
derived by considering the definite integral
 x
f (x) f (x0 ) = f
(t) dt (4.55)
x0
 319
where x0 and x are held constant. An integration of the right-hand side is performed
using integration by parts with U = f
(t), dU = f

(t) dt and dV = dt and V = t x. Here
x is treated as a constant of integration so that
 x x
 x



f (t) dt =f (t)(t x) (t x) f

(t) dt
x x0 x
 x0  x0 (4.56)





f (t) dt =f (x0 )(x x0 ) + (x t) f (t) dt
x0 x0

Now evaluate the integral on the right-hand side of equation (4.56) using integration
by parts to show
 x  x
(x x0 )2 (x t)2



f
(t) dt = f
(x0 )(x x0 ) + f

(x0 ) + f (t) dt (4.57)
x0 2! x0 2!
Continue to use integration by parts n-times to obtain
 x
(x x0 )2 (x x0 )n
f
(t) dt = f
(x0 )(x x0 ) + f

(x0 ) + + f (n) (t) + Rn (x, x0 ) (4.58)
x0 2! n!
or
(x x0 )2 (x x0 )n
f (x) = f (x0 ) + f
(x0 )(x x0 ) + f

(x0 ) + + f (n) (t) + Rn (x, x0 ) (4.59)
2! n!
where Rn (x, x0) is called the remainder term and is given by
 x
1
Rn (x, x0 ) = (x t)n f (n+1) (t) dt (4.60)
n! x0

Remainder Term for Taylor Series

Use the generalized mean value theorem for integrals
 x  x
F (t)G(t) dt = F () G(t) dt, x0 < < x (4.61)
x0 x0

to evaluate the integral used in the representation of the remainder term as given
n
by equation (4.60). Let F (t) = f (n+1) (t) and G(t) = (xt)
n! in equation (4.61) and show
 x  x
1 (x t)n (x x0 )n+1
Rn (x, x0 ) = (x t)n f (n+1) (t) dt = f (n+1) (1 ) dt = f (n+1) (1 )
n! x0 x0 n! (n + 1)!
where x0 < 1 < x. This is the Lagrange form of the remainder term associated with
(n+1)
(t)(xt)n
a Taylor series expansion. Alternatively, substitute F (t) = f n!
and G(t) = 1
into the equation (4.61) to obtain
 x  x
1 n (n+1) f (n+1) (2 )(x 2 )n
Rn (x, x0 ) = (x t) f (t) dt = 1 dt
n! x0 n! x0
(4.62)
(n+1)
f (2 )(x 2 )n
Rn (x, x0 ) = (x x0 )
n!
where x0 < 2 < x. This is the Cauchy form for the remainder term associated with
a Taylor series expansion.
320
Schomilch and Roche Remainder Term
Still another form for the remainder term associated with the Taylor series ex-
pansion is obtained from the following arguments. Let f (x), f
(x), . . . , f (n+1) (x) all be
defined and continuous on the interval [x0 , x0 + h] and construct the function
n
(x0 + h x)m (m)
F (x) = f (x) + f (x) + (x0 + h x)p+1 A (4.63)
m=1
m!

where A and p are nonzero constants. Select the constant A such that
F (x0 + h) =f (x0 + h)
n
hm (m) (4.64)
and F (x0 ) =f (x0 ) + f (x) + hp+1 A = f (x0 + h),
m=1
m!

then F (x) satisfies all the conditions of Rolles theorem so there must exist a point
x = = x0 + h, 0 < < 1, such that F
() = 0. Differentiate the equation (4.63) and
show
n 
 


(x0 + h x)m (m+1) m(x0 + h x)m1 (m)
F (x) = f (x)+ f (x) f (x) (p+1)(x0 +hx)p A
m=1
m! m!

which can be simplified as follows.

n n
(x0 + h x)m (m+1) m(x0 + h x)m1 (m)
F
(x) = f (x) f (x) (p + 1)(x0 + h x)p A
m=1
m! m=2
m!
n n1
 (x0 + h x)i

(x0 + h x)m (m+1)
F (x) = f (x) f (i+1) (x) (p + 1)(x0 + h x)p A
m=1
m! i=1
i!
(x0 + h x)n (n+1)
F
(x) = f (x) (p + 1)(x0 + h x)p A
n!
(4.65)
At x = = x0 + h it follows that
(h h)n (n+1)
F
() = f () (p + 1)(h h)p A = 0
n!

a condition which requires that

[h(1 )]np (n+1)
A= f () (4.66)
(p + 1) n!

Substituting A from equation (4.66) into the equation (4.64) produces the result
n
hm (m) hp+1 [h(1 )]np (n+1)
f (x0 + h) = f (x0 ) + f (x0 ) + f ()
m=1
m! (p + 1) n!
 321
Let x = x0 + h and write the above equation in the form
n
(x x0 )m (m)
f (x) = f (x0 ) + f (x0 ) + Rn (x, x0 ) (4.67)
m=1
m!

where Rn(x, x0 ) is the Schlomilch10 and Roche11 form of the remainder term given by
(x x0 )p+1 (x )np (n+1)
Rn (x, x0 ) = f () (4.68)
(p + 1) n!

where = x0 + h, for 0 < < 1 and p is a constant satisfying 0 p n. Note that

in the special case p = 0 there results the Cauchy form for the remainder and when
p = n there results the Lagrange form for the remainder.
No one can sum an infinite number of terms on the computer. In order to use a
Taylor series expansion to represent a function for computational purposes, one must
chop of the infinite series or truncated it after n-terms. Knowing and controlling
the error associated with the part of the infinite series that is thrown away is very
important in applications and use of Taylor series when summation with a computer
is used. If the remainder term is known, then it is possible to work backwards by first
specifying an error tolerance and then determining the number of terms n required
to achieve this error tolerance for the values of x being used in the application of
the Taylor series.

Example 4-30. (Alternative derivation of LHopitals rule)

Assume the functions f (x) and g(x) have the Taylor series representations
f

(x0 )
f (x) =f (x0 ) + f
(x0 )(x x0 ) + (x x0 )2 +
2!
g

(x0 )
g(x) =g(x0 ) + g
(x0 )(x x0 ) + (x x0 )2 +
2!

which are convergent series in some neighborhood of the point x0 . One can then
f (x)
express the limit xx
lim in the form
0 g(x)

f

(x0 )
f (x0 ) + f
(x0 )(x x0 ) + (x x0 )2 +
lim 2! (4.69)
xx0 g

(x0 )
g(x0 ) + g
(x0 )(x x0 ) + (x x0 )2 +
2!

The LHopitals rule can be derived by considering the following cases.

10
Oscar Xaver Schlomilch (1823-1901) a German mathematician.
11
Edouard Albert Roche (1820-1883) a French mathematician.
322
Case 1 If f (x0 ) = 0 and g(x0 ) = 0 but g
(x0 )
= 0, then equation (4.69) reduces to
f

(x0 ) f


(x0 )
f
(x0 ) + (x x0 ) + (x x0 )2 + f
(x0 ) f
(x)
lim 2! 3! = = lim (4.70)
xx0
g

(x0 ) g


(x0 ) g
(x0 ) xx0 g
(x)
g (x0 ) + (x x0 ) + (x x0 )2 +
2! 3!
and so LHopitals rule takes on the form
f (x) f
(x)
lim = lim

xx0 g(x) xx0 g (x)
Case 2 If f (x0 ) = 0, f
(x0 ) = 0, g(x0) = 0 and g
(x0 ) = 0, but g

(x0 )
= 0, then equation (4.69)
reduces to the form
f

(x0 ) f


(x0 )
+ (x x0 ) + f

(x0 ) f

(x)
lim

2! 3! = = lim (4.71)
xx0 g (x0 ) g


(x0 ) g

(x0 ) xx0 g

(x)
+ (x x0 ) +
2! 3!
and so in this case the LHopitals rule takes on the form
f (x) f
(x) f

(x)
lim = lim
= lim


xx0 g(x) xx0 g (x) xx0 g (x)
The above examples illustrate that one can reapply LHopitals rule whenever
the ratio of derivatives gives an indeterminate form.

Indeterminate forms 0 , , 00 , 0 , 1
f (x) 0 f (x)
If the limit lim
xx0 g(x)
= or xx
lim = , then the limits are said to have
0 0 g(x)
indeterminate forms and are calculated using the LHopitals rule
f (x) f
(x)
lim = lim

xx0 g(x) xx0 g (x)
if the limit exists.
Other indeterminate forms are
lim f (x)g(x) = 0
xx0
or lim f (x)g(x) = 0
xx0

lim [f (x) g(x)] =

xx0

lim f (x)g(x) = 00
xx0

lim f (x)g(x) = 0
xx0

lim f (x)g(x) = 1
xx0

The general procedure used to investigate these other indeterminate forms is

to use some algebraic or trigonometric transformation that reduces these other in-
determinate forms to the basic forms 00 or
so that LHopitals rule can then be
applied. The following examples illustrate some of these techniques.
 323
Example 4-31. Evaluate the limit lim [csc x ln(1 + x)]
x0
Solution If x = 0 is substituted into the functions given one obtains 0 which is
an indeterminate form. Products of functions can be written in alternative forms
using algebra. For example,
f (x) g(x)
f (x) g(x) =
1
or f (x) g(x) =
1
g(x) f (x)
ln(1 + x)
If the given limit is expressed in the form lim , then one can use LHopitals
x0 sin x
rule to investigate the limit. One finds
1
ln(1 + x)
lim = lim 1 + x = 1
x0 sin x x0 cos x

Example 4-32. Evaluate the limit lim (sec x tan x)

x/2
SolutionInvestigating this limit one finds that it depends upon how x approaches
/2. One finds lim (sec x tan x) = lim sec x lim tan x = ( ) which is an
x/2 x/2 x/2
indeterminate form. Using appropriate trigonometric identities the given limit can
be expressed in an alternative form where LHopitals rule applies. One can write
 
1 sin x 1 sin x cos x 0
lim (sec x tan x) = lim = lim = lim = =0
x/2 x/2 cos x cos x x/2 cos x x/2 sin x 1

Example 4-33. Evaluate the limit lim |x|x

x0
Solution This limit gives the indeterminate form 00 . One can use the identity
|x| = eln |x| and write
lim |x|x = lim exln |x| = elimx0 xln |x|
x0 x0

Here limx0 x ln |x| gives the indeterminate form 0 (). Writing

ln |x|
lim x ln |x| = lim 1
x0 x0
x

one can apply LHopitals rule and show

1
ln |x| x
lim x ln |x| = lim 1
= lim = lim (x) = 0
x0 x0 x0 1 x0
x x2

Consequently, one can write lim |x|x = lim exln |x| = e0 = 1

x0 x0
324
2
Example 4-34. Evaluate the limit lim (x + 9)1/x
x
Solution This limit gives the indeterminate form 0 . One can use the identity
(x + 9) = eln(x+9) and write

2 1 ln(x+9)
lim (x + 9)1/x = lim e x2 ln(x+9) = elimx x2
x x

Apply LHopitals rule to this last limit and show

1
ln(x + 9) 1
lim 2
= lim x+9 = lim =0
x x x 2x x 2x(x + 9)

Therefore
2 ln(x+9)
lim (x + 9)1/x = elimx x2 = e0 = 1
x

Example 4-35. Evaluate the limit lim (2 3x )1/x

x0
Solution This limit gives the indeterminate form 1 . Write

1 x ln(23x )
lim (2 3x )1/x = lim e x ln(23 )
= elimx0 x
x0 x0

Recall that
d x d x ln 3
3 = e = ex ln 3 ln 3 = 3x ln 3
dx dx
and consequently when LHopitals rule is applied to the above limit, one finds
1
ln(2 3x ) x
(0 3x ln 3)
lim = lim 2 3 = ln 3
x0 x x0 1

Therefore,
ln(23x ) 1
lim (2 3x )1/x = elimx0 x = e ln 3 = 31 =
x0 3

Modification of a Series
Let un 0 for all n and let {vn } denote a bounded sequence satisfying |vn | < K,

where K is a constant. If the infinite series un is a convergent series, then the
n=1


series un vn will also be a convergent series.
n=1
 325
This follows from an analysis of the Cauchy condition for convergence. Select an
integer value N so large that for all integer values n > m > N the Cauchy convergence
condition satisfies
n
 
|Un Um | = |um+1 + um+2 + + un | |ui | <
K
i=m+1

 n 
   n

 
then write  ui vi  |ui vi | and since the terms vi are bounded, it follows that
 
i=m+1 i=m+1
|ui vi | = ui |vi | ui K so that the Cauchy condition for convergence becomes
 
n m
  n
 n
 

 
 ui vi ui vi  |ui vi | K |ui | < K =
  K
i=1 i=1 i=m+1 i=m+1


so that the infinite series ui vi is convergent.
i=1
Conditional Convergence

An infinite series n=1 un is called a conditionally convergent series or semi-
convergent series, if the given series is convergent but the series of absolute values
is not convergent.
As an example, consider the alternating series given by
1 1 1 1
1 + +
2 3 4 5

This alternating series converges, however it is not absolutely convergent because

the series of absolute values turns into the harmonic series which diverges. The
series is therefore said to be conditionally convergent.
In dealing with an absolutely convergent series, the rearrangement of terms does
not affect the sum of the series. However, in dealing with a conditionally convergent
series, the value of the sum can be changed by using some special rearrangement of
terms and some rearrangements of terms can even make the series diverge. Condi-
tionally convergent series are very sensitive to any changes made to the summation
process.

Algebraic Operations with Series

Examine the series operations of addition, subtraction and multiplying a series
by a constant term together with the operation of multiplying two infinite series as
these are operations that occur quite frequently when dealing with infinite series.
326
Addition and Subtraction
Two convergent series can be added or subtracted if one is careful to maintain
 
parenthesis. That is, given two series A = n=0 an and B = n=0 bn then these series

can be added or subtracted on a term by term basis to obtain




S= an + bn = (an + bn ) =(a0 + b0 ) + (a1 + b1 ) + (a2 + b2 ) + (a3 + b3 ) +
n=0 n=0 n=0
  
D= an bn = (an bn ) =(a0 b0 ) + (a1 b1 ) + (a2 b2 ) + (a3 b3 ) +
n=0 n=0 n=0

The use of parentheses is important because the bn terms may be negative and
in such cases the removal of parenthesis is not allowed. That is, the addition or
subtraction of two infinite series is on a term by term basis with parenthesis being
 
used to group terms. The partial sums are given by An = nm=0 am and Bn = nm=0 bm
so that the sum S and difference D can be expressed
n
 n

lim Sn = lim (am + bm ) lim Dn = lim (an bn )
n n n n
m=0 m=0
S = lim An + lim Bn D = lim An lim Bn
n n n n

S= A + B D= A B

Multiplication by a Constant

A series n=0 an can be multiplied by a nonzero constant c to obtain the series



c an = (c an ) = c a0 + c a1 + c a2 + c a3 +
n=0 n=0

The multiplication of each term by a nonzero constant does not affect the conver-
gence or divergence of the series.
Cauchy Product


If the infinite series an = a0 + a1 + a2 + a3 + and the infinite series
n=0


bn = b0 + b1 + b2 + b3 + are multiplied, then the product series can be written
n=0
 327
a0 b0 + a0 b1 + a0 b2 + a0 b3 + + a0 bn +
+ a1 b0 + a1 b1 + a1 b2 + a1 b3 + + a1 bn +

+ a2 b0 + a2 b1 + a2 b2 + a2 b3 + + a2 bn +
+ a3 b0 + a3 b1 + a3 b2 + a3 b3 + + a3 bn + (4.72)
+

+ an b0 + an b1 + an b2 + an b3 + + an bn +
+
and this result can be grouped into a summation in any convenient way. The Cauchy
method of grouping is to use a summation of terms on a diagonal starting in the
upper left corner of the sum given by (4.72) and drawing diagonal lines from column
n to row n and then summing the results. This gives the elements {cn } from the
double array defined as the diagonal elements
c0 =a0 b0
c1 =a1 b0 + a0 b1
c2 =a2 b0 + a1 b1 + a0 b2

c3 =a3 b0 + a2 b1 + a1 b2 + a0 b3
.. ..
. .
n

cn =an b0 + an1 b1 + an2 b2 + + a1 bn1 + a0 bn = ai bni
i=0

and consequently the product series, called the Cauchy product, can be represented



 n
    
an bn = cn = ai bni (4.73)
n=0 n=0 n=0 n=0 i=0

The Cauchy product is often used in multiplying power series because the result
is also a power series. The Cauchy product is just one of several different definitions
which can be used for the representation of a multiplication of two infinite series.
If the summation of the series begins with the index 1, instead of 0, then



 n
    
an bn = cn = ai bn+1i (4.74)
n=1 n=1 n=1 n=1 i=1
328
Bernoulli Numbers
The sequence of numbers {Bn} defined by the coefficients of the Maclaurin series
expansion

x
xn
= B n , |x| < 2
ex 1 n=0 n!

are called Bernoulli12 numbers. Multiply by ex 1 and use the Maclaurin series for
the exponential function along with the Cauchy product to show

 

n

xn
xn
xn
xn
xn
x= Bn Bn = Bj Bn
n=0
n! n=0
n! n=0
n! n=0 j=0
j!(n j)! n=0 n!

Now equate coefficients of like powers of x above and show B0 = 1 and

n n  

n!
n
Bn = Bj = Bj for n 2 (4.75)
j!(n j)! j
j=0 j=0

Verify that B0 = 1, B1 = 12 , B2 = 16 , B3 = 0, B4 = 301 , . . . Alternatively, expand the

x
function x in a Maclaurin series about x = 0 and show
e 1

x 1 1 x2 1 x4 1 x6 1 x8 5 x10 691 x12 7 x14 3617 x16 43867 x18

= 1 x+ + + + +
ex 1 2 6 2! 30 4! 42 6! 30 8! 66 10! 2730 12! 6 14! 510 16! 798 18!

These expansions produce the Bernoulli numbers

n 0 1 2 4 6 8 10 12 14 16
Bn 1 21 1
6
1
30 1
42
1
30 5
66
691
2730 7
6
3617
510

Here the odd Bernoulli numbers are given by B2n+1 = 0 for n 1.

Euler Numbers
The sequence of numbers {En } defined by the coefficients of the Maclaurin series
expansion

2ex
xn
f (x) = = E n , |x| <
e2x + 1 n=0 n! 2

are called Euler13 numbers. The function f (x) is an even function of x which implies
that the odd Euler numbers satisfy E2n+1 = 0 for all n 0.
12
Named after Jakob Bernoulli (1654-1705) a Swiss mathematician. Due to scaling, indexing and sign conven-
tions, there are alternative definitions for the Bernoulli numbers, sometimes denoted Bn (see table of integrals).
13
Named after Leonhard Euler (1701-1783) a Swiss mathematician. Due to scaling, indexing and sign con-
ventions, there are alternative definitions for the Euler numbers, sometimes denoted En (see table of integrals).
 329
Consequently,
 
2ex 2 xn x2m
f (x) = 2x = x = sech x = E n = E 2m
e +1 e + ex n=0
n! m=0
(2m)!
A multiplication by e2x + 1 and a Maclaurin series expansion of e2x together with an
application of the Cauchy product formula demonstrates that
  


xn  2n xn  xn 

xn   2nk n!
n
xn

 xn
2 = En + En = Ek + En
n! n! n! n! k!(n k)! n! n!
n=0 n=0 n=0 n=0 n=0 k=0 n=0

By equating like powers of x show E0 = 1 and

n
 n!2nk
En = 2 Ek , for n 1
k!(n k)!
k=0
2ex
Alternatively, expand the function f (x) = e2x +1 in a Maclaurin series and show
2 4 6 8 10
x x x x x x12 x14
f (x) = 1 + 5 61 + 1385 50521 + 2702765 199360981 +
2! 4! 6! 8! 10! 12! 14!
These expansions produce the Euler numbers.
n 0 2 4 6 8 10 12 14
En 1 -1 5 -61 1385 -50521 2702765 -199360981
Here the odd Euler numbers are zero and E2n+1 = 0 for n = 0, 1, 2, . . ..

Example 4-36. Additional Series Expansions

One can verify the following infinite series expansions.
x3 2x5 22n (22n 1)Bn x2n1
tan x =x + + ++ + |x| <
3 15 (2n)! 2
3 5 2n 2n1
1 x x 2x 2 Bn x
cot x = ++ + 0 < |x| <
x 3 45 945 (2n)!
x2 5x4 En x2n
sec x =1 + + ++ + |x| <
2 24 (2n)! 2
3 5 2n1 2n1
1 x 7x 31x 2(2 Bn x
csc x = + + ++ + 0 < |x| <
x 6 360 15, 120 (2n)!
x3 2x5 22n (22n 1)Bn x2n1
tanh x =x + (1)n1 + |x| <
3 15 (2n)! 2
3 5 n1 2n 2n 2n1
1 x x 2x (1) 2 (2 1)Bn x
coth x = + ++ + 0 < |x| <
x 3 45 945 (2n)!
x2 5x4 61x6 (1)n En x2n
sech x =1 + ++ + |x| <
2 24 720 (2n)! 2
3 5 n 2n1 2n1
1 x 7x 31x (1) 2(2 1)Bn x
csch x = + ++ + 0 < |x| <
x 6 360 15, 120 (2n)!
where Bn are the Bernoulli numbers and En are the Euler numbers.
330
Functions Defined by Series
If {fn (x)} , n = 0, 1, 2, . . ., denotes an infinite sequence of functions defined over an
interval [a, b], then other functions can be constructed from these functions.
Many functions F (x) are defined by an infinite series having the form


F (x) = cj fj (x) (4.76)
j=0

where c0 , c1, c2 , . . . are constants. To study the convergence or divergence of such

series one should consider the sequence of finite sums {Fn (x)} where
n

Fn (x) = cj fj (x) = c0 f0 (x) + c1 f1 (x) + + cn fn (x) (4.77)
j=0

for n = 0, 1, 2, . . .. The sequence {Fn (x)} is called the sequence of partial sums asso-
ciated with the infinite series (4.76). The infinite series is said to converge if the
sequence of partial sums converges. If the sequence of partial sums diverges, then
the infinite series (4.76) is said to diverge.
The sequence is said to converge uniformly on an interval a x b to a function
F (x), if for every  > 0 there exists an integer N such that

|Fn (x) F (x)| < , for all n > N and for all x [a, b] (4.78)

Example 4-37.
(a) From the sequence of functions {sin nx} one can define the Fourier sine series
expansions


F (x) = bn sin nx (4.79)
n=1

where the bn coefficients are constants.

(b) From the sequence of functions {cos nx} one can define the Fourier cosine series
expansions


G(x) = a0 + an cos nx (4.80)
n=1

where the an coefficients are constants. The study of Fourier series expansions
has many applications in advanced mathematics courses.
 331
Generating Functions
Any function g(x, t) which has a power series expansion in the variable t having
the form


g(x, t) = n (x) tn = 0 (x) + 1 (x) t + 2 (x) t2 + + m (x) tm + (4.81)
n=0

is called a generating function which defines the set of functions {n (x)} for the
values n = 0, 1, 2, . . .. In the above definition scaling of the terms sometimes occurs.
For example, the starting index n = 0 can be changed to some other value and
tn
sometimes tn is replaced by . Some examples of generating functions are the
n!
following.


1
(i) g(x, t) = = xn tn
1 xt n=0

1 t cos
(ii) g(x, t) = = (cos n) tn
1 2t cos + t2 n=0

t sin
(iii) g(x, t) = = (sin n) tn
1 2t cos + t2 n=1

1
(iv) g(x, t) = = (enx ) tn
1 tex n=0

(v) g(x, t) = (1 2xt + t2 )1/2 = Pn (x) tn Legendre polynomials {Pn (x)}
n=0
 

1 xt
(vi) g(x, t) = (1 t) exp = Ln (x) tn Laguerre polynomials {Ln(x)}
1t n=0
There are many other special functions which can be defined by special gener-
ating functions.

Functions Defined by Products

Given a sequence {fn } of numbers or functions, one can define the finite product
n

fi = f1 f2 f3 fn and then the infinite product is written
i=1


 n

fi = f1 f2 f3 where fi = lim
n
fi = lim f1 f2 fn
n
i=1 i=1 i=1

n

if this limit exists. Let Sn denote the finite product Sn = fi and take the logarithm
i=1
of both sides to obtain n n
 
ln Sn = ln fi = ln fi (4.82)
i=1 i=1
332



One can then say that the infinite product fi is convergent or divergent depending
i=1


upon whether the infinite sum ln fi is convergent or divergent.
i=1

Example 4-38. Some examples of infinite products

(a) The French mathematician and astronomer Francois Viete (1540-1603) discov-


 
2 2 2+ 2 2+ 2+ 2
ered the representation =
2 2
2
(b) The English mathematician John Wallis (1616-1703) discovered the representa-
2 2 4 4 6 6 8 8
tion =
2 1 3 3 5 5 7 7 9
(c) The German mathematician Karl Theodor Wilhelm Weierstrass (1815-1897) rep-
resented the Gamma function as the infinite product

1
z  z 
= zez 1+ exp
(z) n=1
n n

where = 0.577215665 . . . is known as the Euler-Mascheroni constant.

(d) Euler represented the function sin as the infinite product
2 2 2

sin = 1 2 1 2 2 1 2 2
2 3

 1
(e) One definition of the Riemann zeta function is (z) = z
. Another form
n=1
n


1
is (z) = where {pn } denotes the sequence of prime numbers. The
n=1
1 pz
n
Riemann zeta function has many uses in number theory.

Continued Fractions
Continued fractionsoccasionally arise in the representation of various kinds of
mathematically quantities. A continued fraction has the form
b1
f = a0 + (4.83)
b2
a1 +
b3
a2 +
b4
a3 +
b5
a4 +
a5 +

where the coefficients a0 , a1 , . . . , b1 , b2, . . . can be real or complex quantities. They can
be constants or functions of x.
 333
In general, when using the continued fraction representation14 given by equation
(4.83) the coefficients a0 , {ai } and {bi }, i = 1, 2, 3, . . . can be constants or functions of
x and these coefficients can be finite in number or infinite in number. The pattern
of numerator over denominator can go on forever or the ratios can terminate after
a finite number of terms. A finite continued fraction has the form

b1
fn = a0 + (4.84)
b2
a1 +
b3
a2 +
b4
a3 +
bn
a4 + +
an
bn
which terminates with the ratio .
an
Terminology
(i) The numbers b1 , b2, b3 , . . . are called the partial numerators.
(ii) The numbers a1 , a2 , a3, . . . are called the partial denominators.
(iii) If the partial numerators bi , for i = 1, 2, 3, . . . are all equal to 1 and all the ai
coefficients have integer values, then the continued fraction is called a simple or
regular continued fraction. A simple continued fraction is sometimes represented
using the shorthand list notation f = [a0 ; a1 , a2 , a3 , . . .] where the ai , i = 0, 1, 2, . . .
are called the quotients of the regular continued fraction.
(iv) The continued fraction is called generalized if the terms ai and bi for i = 1, 2, 3, . . .
do not have any restrictions as to their form.
(v) The ratio of terms notation as illustrated by the equations (4.83) and (4.84) is
awkward and takes up too much space in typesetting and is often abbreviated
to the shorthand Pringsheim15 notation
b1 | b2 | bn |
fn = a0 + + ++ (4.85)
| a1 | a2 | an

bn
for a finite continued fraction terminating with the an
term and in the form

b1 | b2 | bn |
f = a0 + + ++ + (4.86)
| a1 | a2 | an
14
Take note that the starting index is zero. Some notations use a different starting index which can lead to
confusion at times.
15
Alfred Israel Pringsheim (1850-1941) a German mathematician.
334
for an infinite continued fraction. Historically, the shorthand notation originally
used for representing an infinite continued fraction was of the form
b1 b2 b3
f = a0 + ... (4.87)
a1 + a2 + a3 +

where the three dots indicates that the ratios continue on forever.
(vi) If the continued fraction is truncated after the nth term, the quantity fn is called
the nth convergent.
(vii) The continued fraction is called convergent if the sequence of partial convergents
{fn } converges, otherwise it is called a divergent continued fraction.

Evaluation of Continued Fractions

Consider a regular continued fraction which has been truncated after the nth
ratio
1| 1| 1| 1 | 1 |
fn = a0 + + + ++ + (4.88)
| a1 | a2 | a3 | an1 | an
To evaluate this continued fraction start at the bottom and calculate backwards
through the continued fraction. For example calculate the sequence of rational
numbers
1 1 1 1 1
r1 = an1 + , r2 = an2 + , , rn2 = a2 + , rn1 = a1 + , rn = fn = a0 +
an r1 rn3 rn2 rn1

For example, consider the continued fraction

1| 1| 1| 1| 1|
f5 = 1 + + + + +
|2 |3 |4 |5 |6

and start at the bottom and calculate the ratios

1 31 6 130 31 421
r1 =5 + = , r2 = 4 + = , r3 = 3 + = ,
6 6 31 31 130 130
130 972 421 1393
r4 =2 + = , r5 = 1 + =
421 421 972 972

Continued fractions have a long history of being used to approximate numbers

and functions. In 1655 John Wallis discovered an iterative scheme for calculating
the partial convergents of a continued fraction in the forward direction. His iterative
scheme can be written as follows. Define

A1 = 1, A0 = a0 , B1 = 0, B0 = 1 (4.89)
 335
and for j = 1, 2, 3, 4, . . . define the recursion relations

Aj = aj Aj1 + bj Aj2 , Bj = aj Bj1 + bj Bj2 (4.90)

or the matrix equivalent

    
Aj Aj1 Aj2 aj
= (4.91)
Bj Bj1 Bj2 bj
An
then the ratio fn = is the nth partial convergent and represents the continued
Bn
b
fraction after having been truncated after the n term. A proof of
an
the above assertion is a proof by mathematical induction. For j = 1 one obtains
A1 =a1 A0 + b1 A1 = a1 a0 + b1
B1 =a1 B0 + b1 B1 = a1
so that the ratio f1 is given by
A1 a1 a0 + b1 b1
f1 = = = a0 +
B1 a1 a1
Similarly, for j = 2 one finds
A2 =a2 A1 + B2 A0 = a2 (a1 a0 + b1 ) + b2 a0
B2 =a2 B1 + b2 B0 = a2 a1 + b2
so that the second partial convergent is written
A2 a0 a1 a2 + a0 b2 + a2 b1 a2 b1 b1
f2 = = = a0 + = a0 +
B2 a1 a2 + b2 a1 a2 + b2 b2
a1 +
a2
Hence, the recursion relations hold for j = 1 and j = 2. Assume the recursion relations
holds for j = n such that
An an An1 + bn An2
fn = = (4.92)
Bn an Bn1 + bn Bn2
Observe that the partial convergent for fn+1 is obtained from the partial convergent
b
for fn by replacing an by an + n+1 . Making this substitution in equation (4.92) one
an+1
obtains

bn+1 bn+1
an + an+1 An1 + bn An2 an An1 + bn An2 + A
an+1 n1

= bn+1
bn+1 an Bn1 + bn Bn2 +
an + an+1
Bn1 + bn Bn2 an+1 Bn1
bn+1
An + A
an+1 n1 an+1 An + bn+1 An1
= bn+1
= = fn+1
Bn + B an+1 Bn + bn+1 Bn1
an+1 n1

and so the truth of the nth proposition implies the truth of the (n + 1)st proposition.
336
Convergent Continued Fraction
An
Examine the sequence of partial convergents fn = associated with a given
Bn
An
continued fraction. If the limit lim fn = lim = f exists, then the continued
n n Bn
fraction is called convergent. Otherwise, it is called a divergent continued fraction.
Regular Continued Fractions
Regular continued fractions of the form
1| 1| 1|
f = a0 + + ++ + (4.93)
| a1 | a2 | an

are the easiest to work with and are sometimes represented using the list notation

f = [a0 ; a1 , a2 , a3 , . . . , an , . . .] (4.94)

Example 4-39. (Continued fraction)

The representation of a number x as a regular continued fraction of the form
of equation (4.93) is accomplished using the following algorithm to calculate the
partial denominators ai for i = 1, 2, 3, . . ., n.
1
a0 =[x], x1 =
x a0
1
a1 =[x1 ], x2 =
x1 a1 (4.95)
.. ..
. .
1
an =[xn ], xn+1 =
xn an
where [x] is the greatest integer in x function. For example, to represent the number
x = = 3.1415926535897932385 . . . as a continued fraction one finds
1
a0 =[x] = 3, x1 = = 7.0625133059310457698 . . .
3
1
a1 =[x1 ] = 7, x2 = = 15.9965944066857199 . . .
x1 7
1
a2 =[x2 ] = 15, x3 = = 1.0034172310133726 . . .
x2 a2
1
a3 =[x3 ] = 1, x4 = = 292.63459101440 . . .
x3 a3

and so one representation of as a continued fraction has the list form given by
f = = [3 ; 7, 15, 1, 292, . . .] which gives the following rational number approximations
for .
22 333 355 103993
f1 = 3, f2 = , f3 = , f4 = , f5 = ,
7 106 113 33102
 337
Continue the above algorithm and show

f = = [3 ; 7, 15, 1, 292, 1, 1, 1, 2, 1, 3, 1, 14, . . .]

A generalized continued fraction expansion for can be obtained from the arctanx
function evaluated at x = 1 to obtain the representation
1| 1| 4| 9| 16 | 25 | 36 |
= + + + + + + +
4 |1 |3 |5 |7 |9 | 11 | 13

where all the partial numerators after the first term are squares and the partial
denominators are all odd numbers.
Other examples of mathematical constants represented by regular continued
fractions are
e =[2 ; 1, 2, 1, 1, 4, 1, 1, 6, 1, 1, 8, 1, 1, 10, 1, 1, 12, 1, . . .]
=[0 ; 1, 1, 2, 1, 2, 1, , 4, 3, 13, 5, 1, 1, 8, 1, 2, 4, 40, 1, . . .]

Eulers Theorem for Continued Fractions

Euler showed that the infinite series U = u1 + u2 + u3 + can be represented by
the continued fraction
u1 | u2 | u1 u3 | u2 u4 | u3 u5 | un un+2 |
U= (4.96)
|1 | u1 + u2 | u2 + u3 | u3 + u4 | u4 + u5 | un+1 + un+2

The convergence or divergence of this continued fraction is then closely related to

the convergence or divergence of the infinite series which it represents.
Gauss Representation for the Hypergeometric Function
Carl Fredrich Gauss (1777-1855) a famous German mathematician showed that
the hypergeometric function could be represented by the continued fraction
(ac)b (bc1)(a+1) (ac1)(b+1) (bc2)(a+2)
2 F1 (a+ 1, b ; c + 1 ; z) 1| z
c(c+1) | (c+1)(c+2)
z| (c+2)(c+3)
z| (c+3)(c+4)
z|
= + + + + +
2 F1 (a, b ; c ; z) |1 | 1 | 1 | 1 | 1

Representation of Functions
There are many areas of mathematics where functions f (x) are represented in
the form of an infinite generalized continued fraction having the form
b1 (x)
f (x) = a0 (x) + (4.97)
b2 (x)
a1 (x) +
b3 (x)
a2 (x) +
bn (x)
a3 (x) + +
an (x) + rn+1 (x)
338
bn+1(x)
where rn+1 (x) = . This continued fraction is often expressed in the
an+1 (x) + rn+2 (x)
more compact form
b1 (x) | b2 (x) | b3 (x) | bn (x) |
f (x) = a0 (x) + + + ++ + (4.98)
| a1 (x) | a2 (x) | a3 (x) | an (x)

in order to conserve space in typesetting. It is customary to select the functions

a0 (x), ai(x) and bi (x), i = 1, 2, 3, . . . as simple functions such as some linear function of
x or a constant, but this is not a requirement for representing a function. If one
selects the functions bi (x) and ai (x), for i = 1, 2, 3, . . . as polynomials, then whenever
the continued fraction is truncated, the resulting function fn (x) becomes a rational
function of x. The converse of this statement is that if f (x) is a rational function of
x, then it is always possible to construct an equivalent continued fraction.
Observe also that the reciprocal function is given by
1 1 | b1 (x) | b2 (x) | b3 (x) | bn (x) |
= + + + ++ + (4.99)
f (x) | a0 (x) | a1 (x) | a2 (x) | a3 (x) | an (x)

Example 4-40. (Arctangent function)

Assume the function arctan x has the continued fraction expansion
x
arctanx = (4.100)
x2
a1 +
4x2
a2 +
9x2
a3 +
16x2
a4 +
25x2
a5 +
a6 +

where a1 , a2 , . . . are constants to be determined. Note

1 arctanx
= lim =1
a1 x0 x

and the continued fraction (4.100) has the form

x x 1 4
arctanx = = r1 = 1 = x2 x4 +
a1 + r1 arctanx 3 45
x2 x2 4 2 36 4
r1 = = r2 = 3= x x +
a2 + r2 r1 5 175

Continuing this iterative processes one obtains

 339

4x2 4x2 9 2 16 4
r2 = = r3 = 5= x x +
a3 + r3 r2 7 49
9x2 9x2 16 2 400 4
r3 = = r4 = 7= x x +
a4 + r4 r3 9 891
.. ..
. .
(nx)2 (nx)2 (n + 1)2
rn = = rn+1 = [2(n + 1) 1] = x2 +
an+1 + rn+1 rn 2(n + 1) + 1

Observe that lim ri (x) = 0 so it is possible to calculate the coefficients ai , i = 1, 2, 3, . . .

x0
and show
(a1 , a2 , a3, a4 , a5 , . . . , an , . . .) = (1, 3, 5, 7, 9, . . ., 2n + 1, . . .)

Evaluating the arctangent function at x = 1 gives the continued fraction

1| 1| 4| 9| 16 | 25 | 36 |
= + + + + + + +
4 |1 |3 |5 |7 |9 | 11 | 13

given in the previous example 4-29.

Fourier Series
Consider two functions f = f (x) and g = g(x) which are continuous over the
interval a x b. The inner product of f and g with respect to a weight function
r = r(x) > 0 is written (f, g) or (g, f ) and is defined
 b
(f, g) = (g, f ) = r(x)f (x)g(x) dx (4.101)
a

The inner product of a function f with itself is called a norm squared and written
 f 2 . The norm squared is defined
 b
2
 f  = (f, f ) = r(x)f 2 (x) dx (4.102)
a

with norm given by  f = (f, f ). If the inner product of two functions f and g
with respect to a weight function r is zero, then the functions f and g are said to be
orthogonal functions.
340
Example 4-41. The set of functions {1, sin x, cos x} are orthogonal functions over
the interval (0, ) with respect to the weight functions r = r(x) = 1. This is because
the various combinations of inner products satisfy

(1, sin x) = (1) sin x dx = 0
 0
(1, cos x) = (1) cos x dx = 0
0

(sin x, cos x) = sin x cos x dx = 0
0

The given functions have the norm squared values


2
(1, 1) = 1  = (1)2 dx =
0


(sin x, sin x) = sin x 2 = sin2 x dx =
2
 0

(cos x, cos x) = cos x 2 = cos2 x dx =
0 2

A set or sequence of functions {f1 (x), f2 (x), . . ., fn (x), . . . , fm(x), . . .} is said to be

orthogonal over an interval (a, b) with respect to a weight function r(x) > 0 if for all
integer values of n and m, with n
= m, the inner product of fm with fn satisfies
 b
(fm , fn ) = (fn , fm ) = r(x)fm (x)fn (x) dx = 0 m
= n. (4.103)
a

Here the inner product is zero for all combinations of m and n values with m
= n.
If the sequence of functions {fn (x)}, n = 0, 1, 2, . . . is an orthogonal sequence one can
write for integers m and n that the inner product satisfies the relations
 b

0, m
= n
(fm , fn ) = (fn , fm ) = r(x)fn (x)fm (x) dx =
a  fn 2 , m=n
This result can be expressed in the more compact form


2 0 m
= n
(fm , fn ) = ||fn || mn = (4.104)
||fn ||2 m=n
where ||fn ||2 is the norm squared and mn is the Kronecker delta defined to have a
value of unity when m and n are equal and to have a value of zero when m and n are
unequal.

0, m
= n
mn = (4.105)
1, m=n
In the special case where ||fn ||2 = 1, for all values of n, the sequence of functions
{fn (x)} is said to be orthonormal over the interval (a, b).
 341
Example 4-42. If the set of functions {gn(x)} is an orthogonal set of functions
over the interval (a, b) with respect to some given weight function r(x) > 0, then the
set of functions fn (x) = ggn (x)
n
is an orthonormal set. This result follows since
 b
gn (x) gm (x) 1
(fn , fm ) = (fm , fn ) = r(x) dx = (gn , gm)
a  gn   gm   gn   gm 

since the norm squared values are constants. The above inner product representing
(fn , fm ) is zero if m
= n and has the value 1 if m = n.

Example 4-43. Show the set of functions {1, sin nx

L
, cos nx
L
} is an orthogonal
set over the interval L x L with respect to the weight function r = r(x) = 1.
Solution Using the definition of an inner product one can show
 L
nx nx
(1, sin )= sin dx = 0 for n = 1, 2, 3, . . .
L L L
 L
nx nx
(1, cos )= cos dx = 0 for n = 1, 2, 3, . . .
L L L
 L
nx mx nx mx
(sin , sin )= sin sin dx = 0 n
= m (4.106)
L L L L L
 L
nx mx nx mx
(cos , cos )= cos cos dx = 0 n
= m
L L L L L
 L
nx mx nx mx
(cos , sin )= cos sin dx = 0 for all n, m values.
L L L L L

The given set of functions have the norm-squared values

 L
2
(1, 1) = ||1|| = dx = 2L
L
 L
nx nx nx 2 nx
(sin , sin ) =|| sin || = sin2 dx = L for all values of n (4.107)
L L L L L
 L
nx nx nx 2 nx
(cos , cos ) =|| cos || = cos2 dx = L for all values of n
L L L L L

A Fourier16 trigonometric series representation of a function f (x) is expressing

f (x) in a series having the form


nx  nx
f (x) = a0 + an cos + bn sin where L x L (4.108)
n=1
L n=1
L

16
Jean Baptgiste Joseph Fourier (1768-1830) A French mathematician.
342
with a0 and an , bn for n = 1, 2, 3, . . . are constants called the Fourier coefficients. If the
Fourier coefficients are properly defined, then f (x) is said be represented in the form
of a trigonometric Fourier series expansion over the interval (L, L). The interval
(L, L) is called the full Fourier interval associated with the series expansion.
One can make use of the orthogonality properties of the set {1, sin nx L
, cos nx
L
} to
obtain formulas for determining the Fourier coefficients of the Fourier trigonometric
expansion. For example, if one integrates both sides of equation (4.108) from L to
L one finds
 L  L
  L   L
nx nx
f (x) dx = a0 dx + an cos dx + bn sin dx
L L n=1 L L n=1 L L

and this result can be expressed in terms of inner products as

 
2 nx nx
(1, f (x)) = a0  1  + an (1, cos )+ bn (1, cos )
n=1
L n=1
L

By the above orthogonality properties one finds

L 
f (x) dx L
(1, f (x)) L 1
a0 = = L = f (x) dx (4.109)
 1 2 dx 2L L
L

If one multiplies both sides of equation (4.108) by sin mx

L
and then integrates
both sides of the resulting equation from L to L, the result can be expressed in
terms of inner products as
 
mx a0 mx nx mx nx mx
(f (x), sin ) = (1, sin )+ an (cos , sin )+ bn (sin , sin )
L 2 L n=1
L L n=1
L L

and by the orthogonality of these functions one finds the above equation reduces to
mx mx 2
(f (x), sin ) = bm  sin 
L L

because the only nonzero inner product occurs when the summation index n takes on
the value m. This shows that the coefficients bm, for m = 1, 2, 3, . . . can be determined
from the relations

(f (x), sin mx
L
) 1 L
mx
bm = mx 2 = f (x) sin dx for m = 1, 2, 3, . . . (4.110)
 sin L  L L L
 343
Similarly, if one multiplies both sides of equation (4.108) by cos mx
L and then
integrates both sides of the resulting equation from L to L, one can make use of
inner products and orthogonality properties to show
(f (x), cos mx
L )
L


1 mx
am = mx 2 = f (x) cos dx for m = 1, 2, 3, . . . (4.111)

cos L
L L L
In summary, the equations (4.109), (4.110), (4.111) demonstrate that the Fourier
coefficients can be determined from an appropriate inner product divided by a norm
squared
(1, f ) (cos( nx
L
), f ) (sin( nx
L
), f )
a0 = , an = nx , bn = nx (4.112)

1
2
cos( L )
2
sin( L )
2
Note that the set of functions {1, sin nxL
, cos nx
L
} are periodic functions with period
2L and consequently the Fourier trigonometric series will produce a periodic function
for all values of x. The notation f(x) is introduced to define the periodic extension
of f (x) outside the full Fourier interval (L, L). One can write

 nx  nx
f (x) = a0 + an cos + bn sin where L x L
n=1
L n=1
L
or
 nx  nx
f(x) = a0 + an cos + bn sin where <x<
n=1
L n=1
L
The above definitions are introduced due to the fact that f (x) = f(x) because the
original function f (x) need only be defined over the full Fourier interval and f (x) is
not necessarily a periodic function, whereas the function f(x) is periodic and satisfies
f(x + 2L) = f(x) for all values of x.

Example 4-44. (Fourier Series.)

Represent the exponential function as a Fourier series

 nx nx 
ex = a0 + an cos + bn sin over the interval (L, L)
n=1
L L
Solution
One must find the Fourier coefficients a0 , an , bn, n = 1, 2, 3, . . . associated with the
exponential function ex . The Fourier coefficients are calculated from the relations
(4.109), (4.110) and (4.111). One finds
L
(ex , 1)


1 1
a0 = = ex dx = sinh L
||1||2 2L L L
(e , cos nx
x
L )
L
2L(1)n sinh L


1 nx
an = nx 2 = ex cos dx =
|| cos L || L L L L2 + n2 2
(ex , sin nx
L )
L
2n(1)n sinh L


1 nx
bn = nx 2 = ex sin dx =

sin L
L L L L2 + n2 2
344
which gives the Fourier trigonometric series representation of ex as
 
sinh L  2L(1)n sinh L nx 2n(1)n sinh L nx
ex = + 2 2 2
cos 2 2 2
sin (4.113)
L n=1
L +n L L +n L

Figure 4-7.
Fourier trigonometric representation of the function ex compared with ex

The figure 4-7 illustrates a graphical representation of two curves. The first curve
plotted illustrates the given function f (x) = ex for all values of x while the second
curve plotted illustrates f(x) = ex , the Fourier trigonometric series representation.
Note that because the set of functions {1, sin nx
L
, cos nx
L
} are periodic of period 2L the
Fourier series given by equation (4.113) only represents ex on the interval (L, L).
The Fourier series does not represent ex for all values of x. The interval (L, L) is
called the full Fourier interval. Outside the full Fourier interval the Fourier series
gives the periodic extension of the values of f (x) inside the full Fourier interval.

Properties of the Fourier trigonometric series

Conditions for the existence of a Fourier series are: (i) f (x) must be single-valued
and piecewise continuous over the interval (L, L). (ii) The function f (x) is bounded
with a finite number of maxima and minima and a finite number of discontinuities
over the interval (L, L). (iii) The integrals defining the Fourier coefficients must
exist.
 345
The Fourier series, when it exists, represents f (x) on the interval (L, L) which is
called the full Fourier interval. The Fourier series evaluated at points x outside the
full Fourier interval gives the periodic extension of f (x) defined over the full Fourier
interval.
In order for a function f (x) to have a Fourier series representation one must
be able to calculate the Fourier coefficients a0 , an, bn given by the equations (4.109),
(4.110)and (4.111). Consequently, some functions will not have a Fourier series.
For example, the functions x1 , x12 are examples of functions which do not have a
Fourier trigonometric series representation over the interval (L, L). Note that these
functions are unbounded over the interval.
If the functions f (x) and f
(x) are piecewise continuous over the interval (L, L)
then the Fourier series representation for f (x) (a) Converges to f (x) at points where
f (x) is continuous. (b) Converges to the periodic extension of f (x) if x is outside the
full Fourier interval (L, L). (c) At points x0 where there is a finite jump discontinu-
ity, the Fourier trigonometric series converges to 12 f (x+


0 ) + f (x0 ) which represents

the average of the left and right-hand limits associated with the jump discontinuity.
The function SN (x) = a0 + N nx nx
 
is called the N th partial

n=1 an cos L + bn sin L

sum associated with the Fourier series and represents a truncation of the series
after N terms of both the sine and cosine terms are summed. One usually plots
the approximating function SN (x) when representing the Fourier series f(x) graph-
ically. Whenever the function f (x) being approximated has a point where a jump
discontinuity occurs, then the approximating function SN (x) has oscillations in the
neighborhood of the jump discontinuity as well as an overshoot of the jump in
the function. These effects are known as the Gibbs17 phenomenon. The Gibbs
phenomenon always occurs whenever one attempts to use a series of continuous
functions to represent a discontinuous function. The Gibbs phenomenon is illus-
trated in the figure 4-7. These effects are not eliminated by increasing the value of
N in the partial sum.
Fourier Series of Odd Functions
If f (x) = f (x) for all values of x, then f (x) is called an odd function of x and
f (x) is symmetric about the origin. In this special case the Fourier series of f (x)
reduces to the Fourier sine series
17
Josiah Willard Gibbs (1839-1903) An American mathematician.
346

 nx
f(x) = bn sin (4.114)
n=1
L
where  L
2 nx
bn = f (x) sin dx (4.115)
L 0 L
Fourier Series of Even Functions
If f (x) = f (x) for all values of x, then f (x) is called an even function of x and
f (x) is symmetric about the yaxis. In this special case the Fourier series of f (x)
reduces to a Fourier cosine series

 nx
f(x) = a0 + an cos (4.116)
n=1
L

where
 L  L
1 2 nx
a0 = f (x) dx, an = f (x) cos dx for n = 1, 2, 3, . . . (4.117)
L 0 L 0 L
Options
If you are only interested in the function f (x) defined on the interval 0 x L,
then you can represent this function in three different ways. (1) You can extend
f (x) to the full Fourier interval by making it into an odd function. This extension
produces a Fourier sine series. (2) You can extend f (x) to the full Fourier interval by
making into an even function. This extension produces a Fourier cosine series. (3)
You can extend f (x) is some arbitrary fashion so f (x) is neither even nor odd, then
one obtains the full Fourier trigonometric series for the Fourier expansion of f (x).

Figure 4-8.
Function f (x) extended as (a) an odd function (b) an even function (c) neither
 347
Example 4-45. Given the function f (x) = x for 0 < x < L. Extend this function
to the full Fourier interval (L, L) and express f (x) as (i) a Fourier sine series (ii) a
Fourier cosine series (c) a Fourier trigonometric series.

Solution
(a) If f (x) is extended as an odd function, then f (x) = x for L < x < L so that the
Fourier trigonometric series

 
nx nx
f(x) = a0 + an cos( )+ bn sin( ) (4.118)
n=1
L n=1
L

reduces to a Fourier sine series since

 L
(1, f ) 1
a0 = = x dx = 0
 1 2 L L
 L
(cos( nx
L
), f ) 1 nx
an = nx = x cos( ) dx = 0
 cos( L ) 2 L L L

(sin( nx ), f ) 1 L
nx L
bn = L
= x sin( ) dx = 2(1)n
 sin( nx
L
) 2 L L L n

This gives the Fourier sine series

 L nx
f1 (x) = 2(1)n sin( )
n=1
n L

A graph of f1 (x) over the interval (3L, 3L) is illustrated in the following figure.

Note that f1 (x) is periodic and has jump discontinuities at the points 3L, L, L
and 3L where the Gibbs phenomena is readily observed.
348
(b) If f (x) is extended to the

full Fourier interval as an even function, then it can be
x, 0<x<L
represented as f (x) = and the Fourier trigonometric series
x, L < x < 0
(4.118) reduces to a Fourier cosine series since
 
L
(1, f ) 1 L
a0 = 2
= 2 x dx =
1 2L 0 2
 
(cos( nx ), f ) 1 L
nx 2L
an = L
nx = 2 x cos( ) dx = 2 2 (1 + (1)n)
 cos( L ) 2 L 0 L n
(sin( nx
L ), f )
bn = nx =0
 sin( L ) 2

This gives the Fourier cosine series

L  2L nx
f2 (x) = + (1 + (1)n ) cos( )
2 n=1 n2 2 L

A graph of f2 (x) over the interval (3L, 3L) is illustrated in the following figure.

x, 0<x<L
(c) If f (x) is defined f (x) = , then f (x) is neither an odd nor even
0, L < x < 0
function and so there results a Fourier trigonometric series with coefficients
 L
(1, f ) 1 L
a0 = 2
= x dx =
1 2L 0 4

(cos( nx ), f ) 1 L
nx L
an = L
nx = x cos( ) dx = 2 2 (1 + (1)n )
 cos( L ) 2 L 0 L n

(sin( nx
L
), f ) 1 L
nx (1)n
bn = nx = x sin( ) dx =
 sin( L ) 2 L 0 L n

This gives the Fourier series

L  L n nx  L nx
f3 (x) = + (1 + (1) ) cos( ) (1)n sin( )
4 n=1 n2 2 L n=1
n L
 349
A graph of f3 (x) over the interval (3L, 3L) is given in the following figure.

Note the Gibbs phenomena results because of the jump h

discontinuity
i
in the
periodic extension of the function. Also note that f3 (x) = 12 f1 (x) + f2 (x) .

Some Spectacular Results

1. Summation of positive powers
For m a positive integer
(B + n + 1)m+1 B m+1
1m + 2m + 3m + + nm =
m+1
where the right-hand side of the above equation is evaluated as follows.
(a) Expand (B + n + 1)m+1 in a binomial series.
(b) In the binomial expansion replace B k by the Bernoulli number Bk
2. Summation of negative powers
For m a positive integer with B2m a Bernoulli number and (m) the Riemann
zeta function, then

X 1 1 1 1 (1)m (2)2m B2m
= 1 + + + + = = (2m)
n=1
n2m 22m 32m 42m 2(2m)!

Euler showed that

2 4 6
(2) = , (4) = , (6) =
6 90 945
One just doesnt sit down and come up with wonderful formulas like the ones
above. It takes a lot of work to make a discovery. If you dont have a lot of informa-
tion about a subject, then you dont know what questions to ask about the subject.
Therefore, one can say that the more information you have about different subjects,
a better understanding of interrelationships between subjects can be developed. Re-
sults like the above come about by a person getting deeply involved in the subject
matter and investigating simple ideas which in turn lead to complicated results.
350
Exercises

4-1. Examine the given sequence {un } and determine if it converges or diverges. If
the sequence converges, then find its limit.
3n 1 n2
(a) un = (c) un = 1 + ( )n (e) un =
4n 3 2 en
 n
n2 ln n 1
(b) un = 2 (d) un = (f ) un = 1 +
2n 1 n n

4-2. Examine the given sequence {vn} and determine if it converges or diverges. If
the sequence converges, then find its limit.
n
(a) vn = (c) vn = 1 + (1)n (e) vn = sin(n/2)
1 2n
2n2 + 3n + 4 1 + (1)n 2n
(b) vn = (1)n 2 (d) vn = (f ) vn = n
n +n+1 n 3

4-3. Find the sum of the given series

50
 10 100


(a) (3 + 5n) (c) (2 + 7n) (e) (0.02)n1
n=1 n=1 n=1
20  m1
 3 5
10
 10
 3 2
(b) ( + j) n1
(d) 4(3) (f ) (2 + 2)
2 2 2+ 2
j=1 n=1 m=1

4-4. Find the sum of the given geometric series.

1 1 1 1
S1 = + + + +
4 8 16 32
S2 =0.6 + 0.06 + 0.006 + 0.0006 +

S3 =(2 3) + (7 4 3) + (26 15 3) + (97 56 3) +
1 1
S4 = 6 2 + 6 2+
3 3

4-5. Find the sum of the given series.
 i
   i   
 1 
1
i
1  1 1
S1 = S3 = + S5 =
3 3 4 n=1
n + x n + x+1
i=1 i=1
 i     i   
 1 
1
i
1 N
1
S2 = S4 = S6 = ln 1 +
4 3 4 n
i=1 i=1 n=1

4-6. Determine values of x for which the given series converges.

 

xn nxn 1
(a) (b) (c)
n=1
2n2 n n=1
2n n=1
nx2n
 351

4-7. Probability Theory
Assume that a random variable X can take on any of the values {1, 2, 3, . . ., k, . . .},
where k is an integer. If pk is the probability that X takes on the value k, then the
probabilities p1 , p2 , . . . , pk, . . . must be selected such that


(i) Each pk 0 and (ii) pk = 1 7(a)
k=1

In statistics the quantity E(X) is called the expected value of X and is defined


E(X) = k pk 7(b)
k=1

provided the series converges.

Show that the given probabilities satisfy each of the conditions 7(a) and then
calculate the expected value given by equation 7(b).
 k  k
1 1 2 3
(i) pk = k (ii) pk = (iii) pk = 3
2 2 3 4

4-8. For constants a, b and r with a > 0 and b > 0 the given series are known to
converge. Find their sums and required condition for convergence.
 N

1
S1 = i/2
S3 = ar k
i=1
a k=0

 2i N
a +1   k 
S2 = S4 = ar + bkr k1
bi
i=1 k=0

4-9. Use partial fractions and convert the given series to telescoping series and
find their sums.
1 1 1 1
(a) + + ++ +
13 35 57 (2n 1)(2n + 1)
1 1 1 1
(b) + + ++ +
12 23 34 (n 2)(n 1)
1 2 3 n
(c) 2
+ 2 + 2 ++ +
3 15 35 (4n 1)2
2

4-10. Examine the N th partial sum associated with the given infinite series and
determine if the series converge. If the given series converges, find its sum.



1 n
(a) (c)
n=1
n(n + 1)(n + 2) n=1
(n + 1)!



1 n
(b) (d)
n=1
n(n + 1) n=1
(n + 1)(n + 2)(n + 3)

Hint: Use partial fractions.

352

4-11. Use the integral test to determine convergence or divergence of the given
series.
 1  ln n  3n + 4
(a) (c) (e)
n=1
n n=1
n2 n=1
n2

 
ln n 1 n
(b) (d) (f )
n=1
n n=1
n ln n n=1
en

4-12. Assume that f (x) is a given function satisfying the following properties.
(i) The function f (x) is a continuous function such that f (x) > 0 for all values of x.
lim nP f (n) exists and the limit is different from zero.
(ii) For p > 0 the limit n


Show that f (n) converges if p > 1 and diverges for 0 < p 1.
n=1
Hint: See modification of a series.

4-13. Use the comparison test to determine convergence or divergence of the given
series.


 
1 1 cos n
(a) (c) (e)
n=1
n(n + 3)(n + 6) n=1
3+2 n n=1
n2 + 1




1 1 1
(b) (d) (f )
n=1
3 + 2n n=1
2
3n + 2n + 1 n=1
n2 ln n

4-14.
(a) Verify that the given series converge.
(b) Find the sum of the first four terms of each series and give an estimate for the
error between the exact solution and your calculated value.
(c) Find the sum of the first eight terms of each series and give an estimate for the
error between the exact solution and your calculated value.

 

1 n+1 1 1 1
(i) (ii) (1) (iii) (iv) (1)n+1
n=1
n3 n=1
n3 n=1
n4 n=1
n4

4-15. Show that the given series converge and determine which series converges
at the slower rate.
 
1 n
(i) A= (ii) B=
n=1
n 3n n=1
5n

4-16. Show that the given series diverge and determine which series diverges at
the slower rate.
1  1
(i) A= (ii) B=
n=1
n n=1
ln n
 353

4-17. Newtons root finding method To deter-
mine where a given curve y = f (x) crosses the
x-axis one can select an initial guess x0 and if
f (x0 )
= 0 one can then calculate f
(x0 ). From
the values f (x0 ) and f
(x0 ) one can construct the
tangent line to the curve y = f (x) at the point
(x0 , f (x0 )). This tangent line given by y f (x0 ) = f
(x0 )(x x0 ).
(a) Show the tangent line intersects the x-axis at the point x1 = x0 f (x0 )/f
(x0 )
(b) Form the sequence {xn } where xn = xn1 f (xn1)/f
(xn1 ) for n = 1, 2, 3, . . .
(c) Give a geometric interpretation to what this sequence is doing. Hint: What has
been done once can be done again.
(d) If y = f (x) = x2 3x + 1 and x0 = 1, find using a calculator x1 , x2 , x3 and x4
(e) If y = f (x) = x2 3x + 1 and x0 = 2, find using a calculator x1 , x2 , x3 and x4
(f) Sketch the curve y = f (x) = x2 3x + 1 and find the roots of the equation f (x) = 0.
(g) What happens if the initial guess x0 is bad? Say x0 = 3/2 for the above example.

xn

4-18. Let fn (x) =
n(n + 1)



(a) Show that fn (9/10) converges. (b) Show that fn (10/9) diverges.
n=1 n=1

 1

4-19. Given the infinite series , with p > 0.
n=2
n [ln n]p
(a) Show the series converges for p > 1.
(b) Show the series diverges for p 1.
Hint: Let f0 (t) = t, f1 (t) = ln f0 (t), f2 (t) = ln f1 (t), . . . , fn+1(t) = ln fn (t) and show
 
dt fm+1 (t), p=1
p = 1 p1
f0 (t)f1 (t)f2 (t) fm1 (t) [fm (t)] (p1)
[fm (t)] , p
= 1

 
1 1
and then examine p
=
n=2
n [ln n] f (n)[f1 (n)]p
n=2 0

 
1

4-20. Show that if the series un converges, then the series diverges.
n=1
u
n=1 n

4-21. If Un = 13 n 3 = 13 n(n + 1)(n + 2), show Uk = Uk+1 Uk = (k + 1)(k + 2) and find

n1
 n1

the sum of the series Sn = 1 2 + 2 3 + 3 4 + + n (n + 1) = (k + 1)(k + 2) = Uk
k=0 k0
354

4-22. Reversion of series
(a) Given the series

(x 1)2 (x 1)3
y = y(x) = 1 (x 1) + + 22(a)
2! 3!

and it is required that you solve for x 1 in terms of y to obtain a series of the
form
(x 1) = A1 (y 1) + A2 (y 1)2 + A3 (y 1)3 + A4 (y 1)4 + 22(b)

where A1 , A2 , A3 , A4 . . . are constants to be determined. Substitute equation 22(b)

into equation 22(a) and expand all terms. Equate like powers of (y 1) and
obtain a system of equations from which the constants A1 , A2 , A3 , A4 , . . . can be
determined.
(b) If the original series in part(a) is y = e(x1) and the series obtained by reversion
is the function (x 1) = ln y, expanded in a series about y = 1, then examine
the approximation for x 1 by truncation of your series after the A4 term. Let
E = E(y) denote the error in using this truncated series to solve for x 1. Plot a
graph of the error E = E(y) for 1 y < 2.

4-23. Examine the given alternating series to determine if they converge.

  
1 (1)n1 (1)n+1
(a) (c) (e)
n=1
(3)n1 n=1
n! n=1
n

  
1 (1)n (1)n+1 n
(b) (1)n (d) (f )
n=2
ln n n=1
n n=1
n2 + 1

4-24. Use the root test to determine if the given series converge.

  
1 n2 1
(a) (c) (e)
n=2
[ln n]n n=1
2n n=1
nn
   n   n
nn n 1
(b) (d) (f )
n=1
24n n=1
2
n +1 n=1
n+1

4-25. Test the following series to determine convergence or divergence.


 
(1)n n3
(a) (c) sinn (/3) (e)
n=1
n! n=1 n=1
3n

 10  
n n 3n
(b) n
(d) n
(f )
n=1
(1.001) n=1
2 n=1
n!
 355

4-26. Determine whether the given series converge or diverge.
1 2 3 4 n
(a) + + + ++ +
2 3 4 5 n+1
3 5 7 2n + 1
(b) + + ++ +
19 35 51 16n + 1
(c) sin(1) + 2 sin(1/2) + 3 sin(1/3) + + n sin(1/n) +

1
(d) n cos( )
n=1
n

2
(e)
n=1
n

4-27. Determine the convergence or divergence of the given series.




n 1 1
(a) (c) (e)
n=1
n+1 n=1
n3/2 n=1
2n + 4n

  
(n!)2 1 2n
(b) (d) (f )
n=1
(2n)! n=1
n3 +n n=1
n2

4-28. Determine if the given series is (a) conditionally convergent, (b) absolutely
convergent or (c) divergent.
  
(1)n (1)n+1 (1)n+1 ln n
(a) (c) (e)
n=2
n n=1
n n=1
n2
  
cos(n) (1)n+1 (1)n+1 nn
(b) (d) (f )
n=1
n2 n=1
6n + 3 n=1
n!

4-29. Determine if the given series is (a) conditionally convergent, (b) absolutely
convergent or (c) divergent.
 

(1)n (1)n n2
(a) (c) (e) (1)n+1
n=0
3n + 2 n=2
2 ln n n=1
3n


 
(1)n+1 n n
(b)  (d) (1)n+1 (f ) (1) n+1

n=1 (n + 1)(n + 2) n=1

n2 + 1 n=1
n2 +1

4-30. Find the interval where the power series converges absolutely.
 

x2n n (x 1)n xn
(a) (c) (e) (1)n1
n=1
n 2n n=1
3n n=1
3n

 n
 n 
(x 2) (3x) (3x + 2)n
(b) (1)n+1 (d) (f ) (1)n1
n=1
n n=1
ln(n + 1) n=1
4n
x x x

4-31. Let y = f (x) = | | + | | + | | + and show that
1 1 1
dy 1| 2x | x| x|
= f
(x) = + + + +
dx | 1 | 1 | 1 | 1
x
Hint: Show that y =
1+y
356

4-32. Explain the difference between (a) the limit of a sequence and (b) the limit
point of a sequence.

4-33. Examine the binomial series for the expansion of (a+b)n when n is an integer.
n(n 1) n2 2 n(n 1)(n 2) n3 3
(a + b)n =an + nan1 b + a b + a b + + bn
    2!   3!    
n n n 0 n n1 1 n n1 2 n 1 n1 n 0 n
(a + b) = a b + a b + a b ++ a b + a b 33(a)
0 1 2 n1 n
  
n n nj j
(a + b) = a b
j
j=0
 
n!
n mn
where = m! (nm)! are the binomial coefficients.
m 0, m>n
n  
 n  
n n nj j  n j nj
(a) Show that (a + b) = a b = a b
j j
j=0 j=0
(b) Newton generalized the binomial expansion to
 

r r
(a + b) = ark bk
k
k=0 33(b)
r(r 1) r2 2 r(r 1)(r 2) r3 3
(a + b)r =ar + rar1 b + a b + a b +
2! 3!
where r represents an arbitrary real number.
(i) Show that when r is a nonnegative integer, the equation 33(b) reduces to
equation 33(a).
(ii) (Difficult problem) Write equation 33(b) in the form ar (1 + x)r where x = b/a.
Examine the series expansion for f (x) = (1 + x)r . Then use the Lagrange and
Cauchy forms of the remainder Rn to show the equation 33(b) converges if
|a| > |b| and diverges if |a| |b|, where x = b/a.

1 1 1

4-34. Let y = g(x) = x + | | + | | + | | + and show that
x x x
dy 1| x| 1| 1| 1
= g
(x) = + ++ Hint: Show that y = x +
dx |2 |x |x |x y
sin x | cos x | sin x | cos x |

4-35. Let y = h(x) = | +
| 1
+
| 1
+
| 1
+ and show that
1
dy (1 + y) cos x + y sin x
= h
(x) =
dx 1 + 2y + cos x sin x
sin x
Hint: Show that y =
1 + cos
1+y
x
 357

4-36. The continued fraction function
Pn (x) 1 | 1| 1 | 1|
yn = yn (x) = = 0 + + ++ +
Qn (x) | 1 | 2 | n |x

where 0 , 1 , 2 , . . . , n are constants, represents a rational function of x.

dy 1
(a) Show that 1 = 2
dx [Q1 (x)]
dy2 1
(b) Show that =
dx [Q2 (x)]2
dy3 1
(c) Show that =
dx [Q3 (x)]2
dyn 1
(d) Can you show that in general = (1)n
dx [Qn (x)]2
  2    2    2 


4-37. Euler used the product formula sin = 1 1 1
2 n

2 1
to represent sine of . Differentiate this relation and show cot = 1+2 2 n2 2
n=1

 

4-38. Assume that un and vn are two infinite series of positive terms and that
there exists an integer N such that for all n > N the inequality un Kvn for some
positive constant K .
 
(i) If the series vn converges, prove the series un converges.
 
(ii) If the series un diverges, prove the series vn diverges.

 1

4-39. Show that the alternating p-series (1)n+1 converges if p > 0.
n=1
np

1

4-40. Consider the geometric series = 1 + z + z 2 + z 3 + + z n + where
1z
z = rei , |z| < 1 and i2 = 1. Show that by equating real and imaginary parts
1 r cos
=1 + r cos + r 2 cos 2 + + r n cos n +
1 2r cos + r 2
r sin
=r sin + r 2 sin 2 + + r n sin n +
1 2r cos + r 2
Hint: Use Euler identity ei = cos + i sin

4-41.
(a) Show {sin nx}, n = 1, 2, 3, . . . is an orthogonal sequence over the interval (0, ) with
respect to the weight function r = 1.
(b) Scale the above sequence to construct an orthonormal sequence over the given
interval.
358

4-42. Calculate the inner products and norm squared values associated with the
given sequence of functions {fn (x)} using the given interval (a, b) and weight function
r(x), for n = 1, 2, 3, . . ..
nx
(a) {fn } = {sin }, (0, L), r = 1
L
nx
(b) {f0 , fn } = {1, cos }, (0, L), r = 1
L
nx nx
(c) {f0 , f2n , f2n1} = {1, cos , sin }, (L, L), r = 1
L L
(d) {f0 , f1 , f2 } = {1, 1 x, x2 4x + 2}, (0, ), r = ex

4-43. Even and Odd Functions  L

(a) If Go (x) = Go (x) for all values of x, show that Go (x) dx = 0
L
 L  L
(b) If Ge (x) = Ge (x) for all values of x, show that Ge (x) dx = 2 Ge (x) dx
L 0
(c) Let Fo (x) denote an odd function of x and Fe (x) denote an even function of x.
Similarly, let Go (x) denote an odd function of x and Ge (x) denote an even function
of x, show that
(i) H(x) = Fo (x)Go (x) is an even function of x
(ii) H(x) = Fo (x)Ge (x) is an odd function of x
(iii) H(x) = Fe (x)Ge (x) is an even function of x
(d) Determine which functions are even or odd.
nx nx
(i) Fe (x) cos( ) (iii) Fo (x) cos( )
L L
nx nx
(ii) Fe (x) sin( ) (iv) Fo (x) sin( )
L L

(e) Use the above properties to simplify the Fourier series representation of f (x) over
the interval (L, L), as given by equation (4.108), if
(i) The function f (x) is an even function.
(ii) The function f (x) is an odd function.

4-44. (Newtons method for nonlinear system)

To solve the system of simultaneous nonlinear equations

f (x, y) = 0, g(x, y) = 0

in the two unknowns x and y, one can use Newtons method which is described as
follows.
 359
Start with an initial guess of the solution and call it x0 and y0 . Now expand f
and g in Taylor series expansions about the point (x0 , y0 ). These expansions can be
written
f (x0 , y0 ) f (x0 , y0 )
f (x0 + h, y0 + k) = f (x0 , y0 ) + h+ k
x y
1  
+ fxx h2 + 2fxy hk + fyy k2 +
2!
g(x0, y0 ) g(x0 , y0 )
g(x0 + h, y0 + k) = g(x0, y0 ) + h+ k
x y
1  
+ gxxh2 + 2gxy hk + gyy k2 + .
2!
Usually the initial guess (x0 , y0 ) is such that f (x0 , y0 ) and g(x0 , y0 ) are not zero. It is
desired to find values h and k such that the equations
f (x0 + h, y0 + k) = 0 and g(x0 + h, y0 + k) = 0
are satisfied simultaneously. Now assume that the values h and k to be selected are
small corrections to x0 and y0 so that second-order terms h2 , hk, k2 , and higher order
product terms are small and can consequently be neglected in the above Taylor
series expansion. These assumptions produce the linear system of equations
f (x0 , y0 ) f (x0 , y0 )
f (x0 , y0 ) + h+ k=0
x y
g(x0, y0 ) g(x0 , y0 )
g(x0, y0 ) + h+ k=0
x y
which can then be solved to determine the correction terms h and k.
(a) Show by letting h = x1 x0 and k = y1 y0 that an improved estimate for the
solution to the simultaneous equations f (x, y) = 0 and g(x, y) = 0, is given by

x1 = x0 + h = x0 +


y1 = y0 + k = y0 + ,

where    f (x0 ,y0 ) 
 f (x , y ) f (x0 ,y0 )  
 0 0  f (x0 , y0 ) 
= y
 and =  g(xx0 ,y0 )
 g(x0, y0 ) g(x0 ,y0 )
y

x
g(x0, y0 ) 
 f (x ,y ) 
 0 0 f (x0 ,y0 ) 
 
and is the determinant of the coefficients given by =  g(xx0 ,y0 ) y
g(x0 ,y0 ) .
 
x y
(b) Illustrate Newtons method by solving the nonlinear system of equations
f (x, y) = 2x2 3y + 1 = 0 g(x, y) = 8x + 11 3y 2 = 0
Hint: Nonlinear equations may have multiple solutions, a unique solution, or
no solutions at all. Sometimes a graph is helpful in estimating a solution if one
exists.
360

4-45.
Verify the Fourier series representation for the functions illustrated. In each
graph assume the maximum amplitude of each function is +1 and the minimum
amplitude of each function is either zero or -1 depending upon the graph.

 
4 1 (2n + 1)x
(a) f (x) = sin
n=0 2n + 1 a
1

11 nx
(b) f (x) = sin
2 n=1 n a
 
1 4  1 (2n + 1)x
(c) f (x) = 2 cos
2 n=0 (2n + 1)2 a

2  (1)n nx
(d) f (x) = sin
n=1 n a

8  1 n nx
(e) f (x) = 2 2
sin( ) sin( )
n=1 n 2 a

4-46. Show the Fourier trigonometric series


nx  nx
f (x) = a0 + an cos + bn sin where L x L
n=1
L n=1
L

 nx
can also be expressed in the form f (x) = a0 + cn sin( + n ) by finding the values
n=1
L
cn and n .

a0  nx

nx

4-47. Let f (x) + an cos + bn sin denote the Fourier series repre-
2 n=1
L L
sentation of f (x) over the full Fourier interval (L, L).
(a) Use the Euler formulas
nx nx nx nx
einx/L = cos + i sin and einx/L = cos i sin
L L L L
 361
and show
nx ei nx/L + ei nx/L nx ei nx/L ei nx/L
cos = and sin =
L 2 L 2i

(b) Define C0 = a20 , Cn = 12 (an ibn ), Cn = 12 (an + ibn) and show the Fourier series can
be represent in the complex form

  L
inx/L 1
f (x) Cn e where Cn = f (x)einx/L dx
n=
2L L

4-48. You are sick and your doctor prescribes medication XY Z to be taken -times
a day based upon the concentration of XY Z . Find . First get over the shock of
being asked such a question. To solve the problem you must make some assumptions
such as the following.
(i) At time = 0 you take medication XY Z and this produces a concentration C0 of
XY Z in your blood stream.
(ii) The concentration C0 decays exponentially with time so that after a time
the concentration of XY Z in your blood is C0 ek , where k is called the decay
constant.
(iii) At times , 2, 3, . . ., n you take the medication XY Z and consequently you build
up a certain residual concentration of XY Z in your blood stream given by

C0 ek + C0 e2k + C0 e3k + + C0 ekn

(a) If you continue the prescribed dosage forever, then the residual concentration
would be
  m
C= C0 emk = C0 ek ek
m=1 m=0

Sum this series and determine the residual concentration C .

(b) If Cs denotes the maximum safe concentration of XY Z that the human body can
stand, then show that must be selected to satisfy
 
ek
C0 + C = C0 1 + Cs
1 ek

(c) Assume k, C0 , Cs are known values and solve for .

362

4-49.
1
(a) Verify the well known result = 1 + 2 3 + 4 5 + 6 7 +
1+
and memorize this result.
(b) Assume y = f (x) = tan1 x = a0 + a1 x + a2 x2 + a3 x3 + a4 x4 + where a0 , a1 , a2 , . . . are
constants to be determined. Show
dy 1
= f
(x) = = a1 + 2a2 x + 3a3 x2 + 4a4 x3 +
dx 1 + x2
1
(c) Use the well known result from part (a) to expand f
(x) = and compare the
1 + x2
expansion in part (b) with the expansion in part (c) to determine the coefficients
a1 , a2 , a3 , . . .
(d) Pick a particular value of x whereby you can determine a0 and then give the
series expansion for tan1 x

4-50. Apply the method outlined in the previous problem to determine the series
expansion for sin1 x.

4-51. Show that

x2 x3
(a) a = 1 + x ln a + (ln a)2 + (ln a)3 +
x
2! 3!
h h2 h3 5h4
(b) x + h = x + 3 3/2 + 4 5/2 6 7/2 +
2 x 2 x 2 x 2 x

4-52.
(a) Assume the series expansion y =ax = a0 + a1 x + a2 x2 + a3 x3 + a4 x4 + 52-(a)
dy
and show =ax ln a = a1 + 2a2 x + 3a3 x2 + 4a4 x3 + 52-(b)
dx
(b) Substitute equation 52-(a) into equation 52-(b) and compare coefficients to show
x2 x3 x4 x5
ax = 1 + x ln a + (ln a)2 + (ln a)3 + (ln a)4 + (ln a)5 +
2! 3! 4! 5!

dy cos x

4-53. If y = sin x + sin x + sin x + sin x + show that =
dx 2y 1

1 1 6

4-54. Show that ex sin x = 1 + x2 + x4 + x +
3 120
1 1 11 4 1 5 61 6

4-55. Show that ex cos x = 1 + x + x2 x3 x x + x +
2 3 24 5 720
5 19 6

4-56. Show that ex tan x = 1 + x2 + x4 + x +
6 30
 363
Chapter 5
Applications of Calculus
Selected problems from various areas of physics, chemistry, engineering and the
sciences are presented to illustrate applications of the differential and integral calcu-
lus. Many of these selected topics require knowledge of basic background material,
such as terminology and fundamentals, associated with the area of application. Con-
sequently, much of this chapter gives a presentation of selected basic material from
areas of engineering, physics, chemistry and the sciences which is required knowledge
for the understanding of many scientific applications of the differential and integral
calculus.
Related Rates
The rate of change of a quantity Q = Q(t) with respect to time t is denoted by the
dQ
derivative . Problems which involve rates of change of two or more time dependent
dt
variables are referred to as related rate problems . The general procedure for
solving related rate problems is something like the following.
1. If necessary, define the variables of the problem and make note of the units of
measurement being used. For example, one could write [Q] = cubic centimeters
which is read1 The dimension of Q is cubic centimeters.
2. Find how the variables of the problem are related for all values of time t being
considered.
3. Determine if the variables of the problem, or their derivatives, have known values
at some particular instant of time.
4. Find the rate of change relation between the variables by differentiating the
relation or relations found in step 2 above.
5. Evaluate the results in step 4 at the particular instant of time specified.

Example 5-1. Consider a large inverted right circular cone with altitude H
and base radius R where water runs into the cone at the rate of 3 cubic feet per
second. How fast is the water level rising when the water level, as measured from
the vertex of the cone, is 4 feet? Here the base radius R and height H of the cone
are considered as fixed constants.

1
Notation introduced by J.B.J. Fourier, theorie analytique de la chaleur, Paris 1822.
364
Solution Let r = r(t), [r] = feet, denote the radius of the water level at time t and let
h = h(t), [h] = feet, denote the height of the water level at time t, [t] = minutes. One
can then express the volume V of water in the cone at time t as
2
V = V (t) = r h, [V ] = cubic feet (5.1)
3

Using similar triangles one finds that there is

a relation between the variables r and h given by
r R R
= or h r= (5.2)
h H H
dV
The given problem states that = 3, where
dt
dV 3 dh
[ ] = ft /min and it is required to find , when
dt dt
h = 4 feet. Differentiating equation (5.1) with
respect to time t gives


dV 2 dh dr
= r + 2r h (5.3)
dt 3 dt dt
and differentiating equation (5.2) with respect to t gives
dr R dh
= (5.4)
dt H dt
since R and H are constants. Substituting the results from the equations (5.2) and
(5.4) into the equation (5.3) gives
R2 2

     
dV dh R R dh
= h +2 h h
dt 3 H2 dt H H dt
which simplifies to
dV R2 dh
= 2 h2 (5.5)
dt H dt
Now one can evaluate the equation (5.5) when h = 4 to obtain
R2 3 H2


dh dh dh
3 = 2 (4)2 or = , = ft/min (5.6)
H dt dt 16 R2 dt
Alternatively, one could have substituted the equation (5.2) into the equation (5.1)
to obtain
R2 3
V = h (5.7)
3 H2
and then differentiate the equation (5.7) with respect to time t to obtain
R2
 
dV 2 dh
= 3h
dt 3 H2 dt
and evaluating this last equation when h = 4 gives the same result as equation (5.6).
 365
Example 5-2. Two roads intersect at point 0 at an angle of 60 degrees. Assume
car A moves away from 0 on one road at a speed of 50 miles per hour and a second
car B moves away from point 0 at 60 miles per hour on the other road. Let a denote
the distance from point 0 for car A and let b denote the distance from point 0 for
car B. How fast is the distance between the cars changing when a = 1 mile and b = 2
miles.
Solution

Let r denote the distance between the cars A and B

and use the law of cosines to show

r 2 = a2 + b2 2ab cos

Differentiate this relation with respect to time t and show


dr da db db da
2r = 2a + 2b 2 a + b cos , is a constant
dt dt dt dt dt

dr
and then solve for the rate of change to obtain
dt


da db db da
a +b a + b cos
dr dt dt dt dt
=
dt a2 + b2 2ab cos

Substitute into this equation the values

a =1 mi b =2 mi

da db =
=50 mi/hr =60 mi/hr 3
dt dt
dr
and show = 30 3 miles per hour.
dt

Example 5-3. Boyles2 law resulted from a study of an ideal compressed gas
at a constant temperature. Boyle discovered the relation P V = C = constant, where P
represents pressure, [P ] = Pascal, abbreviated Pa, and V represents volume, [V ] = cm3
and C is a constant. If at some instant the pressure is P0 and the volume of the gas
has the value V0 and the pressure is increasing at the rate r0 , [r0 ] = Pa/min, then at
what rate is the volume decreasing at this instant?
2
Robert Boyle (1627-1691) an Irish born chemist/mathematician.
366
Solution Here Boyles law is P V = P0 V0 = constant, where the pressure and volume
are changing with respect to time. Differentiating this relation with respect to time
t gives the relation
dV dP d
P + V = (P0 V0 ) = 0 (5.8)
dt dt dt
dP
Evaluating the equation (5.8) at the instant where = r0 , P = P0 and V = V0 , one
dt
finds
dV dV V0
P0 + r0 V0 = 0 or = r0
dt dt P0
The minus sign indicates that the volume is decreasing and the volume rate of change
dV
has dimension, [ ] = cm3 /min.
dt
Note that Boyles law is a special case of the more general gas law given by
PV
= C = Constant relating pressure P , volume V and temperature T all having
T
appropriate units of measurements.

Newtons Laws
Isaac Newton used his new mathematical knowledge of calculus to formulate
basic principles of physics in studying the motion of objects and particles. The
following are known as Newtons laws of motion.
(i) Newtons First Law
A body at rest tends to stay at rest or a body in a uniform straight line
motion tends to stay in motion unless acted upon by an external force.
(ii) Newtons Second Law
The time rate of change of momentum3 of a body is proportional to the
resultant force that acts upon it.
(iii) Newtons Third Law
For every action there is an equal and opposite reaction.

In the following discussions the symbols F, x, v, a, m, p, t are used to denote force,

distance, velocity, acceleration, mass, momentum and time. Time t is measured
in units of seconds, abbreviated (s). The symbol F is to denote force, measured
in units of Newtons abbreviated4 (N). The quantity x denotes distance, measured
dx
in meters, abbreviated (m). The velocity is denoted v = and represents the
dt
3
Momentum is defined as mass times velocity.
4
If a unit of measurement is named after a person, then the unit is capitalized, otherwise it is lower case.
 367
change in distance with respect to time. The velocity is measured in units of meters
d2 x dv
per second, abbreviated (m/s). The second derivative of distance 2 = = a or
dt dt
derivative of the velocity with respect to time t, is called the acceleration, which is
measured in units (m/s2 ). The symbol m denotes the mass5 of a body, measured in
units called kilograms, abbreviated (kg) and the momentum p = mv is defined as the
mass times the velocity and is measured in units (kg m/s).
The first law concerns the inertia of a body. A body at rest, unless acted upon
by an external force, will remain at rest. In terms of the above symbols one can
write the first law as F = 0 or v = 0. That is, if the body is at rest, then it has no
forces acting on it and if the body is in a state of uniform motion, then there is no
change in the velocity. An external force is required to change the state of rest or
uniform motion.
The momentum p of a body is defined as the mass times the velocity and written
d
p = mv . Consequently, the second law can be expressed F = (mv), where is a
dt
proportionality sign. The units of measurement for force, mass, distance, velocity
and time are selected to make the proportionality constant unity so that one can
write Newtons second law as
d dv dm
F = (mv) = m + v (5.9)
dt dt dt

If the mass is constant and does not change with time, then the second law can be
expressed
dv d2 x
F =m = m 2 = ma (5.10)
dt dt
The units of measurement used for the representation of Newtons laws are either the
meter-kilogram-second system (MKS), the centimeter-gram-second system (CGS) or
the foot-pound-second system (FPS) where


M KS

FPS

CGS








F in N F in lb F in dynes











m in kg






m in slugs







m in gm





2
2

in cm/s2


a in m/s
a in ft/s
a

5
Note the subtle distinction between the notation used to denote mass (m) and meters (m).
368
1 N = 105 dynes = 0.2248 lbs-force

1 Kg = 6.852 (10)2 slugs = 1000 gm

9.807 m/s2 = 32.17 ft/s2 = 980.7 cm/s2
Here F denotes a summation of the forces acting in the direction of motion.
Note that if the sum of the forces or resultant force is zero, then the object is said
to be in translational equilibrium. If the velocity of a body is constant, but its mass
is changing, then the equation of motion (5.9) becomes
dm
F =v (5.11)
dt

In terms of symbols, the third law can be expressed by examining two bodies,
call them body A and body B. If body A exerts a force FAB on body B, then body
B exerts a force FBA on body A and the third law requires that FAB = FBA, that is
the forces are equal and opposite.

Newtons Law of Gravitation

Newtons law of gravitation states that the cen-

ters of mass associated with two solids m1 and m2
experience an inverse square law force F of attrac-
tion given by
Gm1 m2
F = (5.12)
r2

where r is the distance between the centers of mass and G = 6.673 1011 m3 /kg s2 is
a proportionality constant called the gravitational constant.
If m1 = me is the mass of the Earth and m2 = m is the mass of an object at
a height h above the surface of the Earth, then the force of gravity between these
masses is given by

Gme m Gme
Fg = =m (5.13)
(re + h)2 (re + h)2
where re denotes the radius of the Earth6 . Write the quantity in brackets as

  2
Gme Gme h Gme
2
= 2 1+ 2 (5.14)
(re + h) re re re
6
The radius of the Earth is approximately 6400 km4000 mi and the mass of the Earth is approximately
6.035 (10)24 kg .
 369
since h is much less than the radius of the Earth re . The equation (5.14) can be used
to define the following terms.
The acceleration of gravity g is defined
Gme
g= (5.15)
re2
and the weight W of an object of mass m due to gravity is defined

W = Fg = m g (5.16)

That is, the weight of an object is the force (force of gravity), by which an
object of mass m is pulled vertically downward toward the center of the Earth.
The dimensions of g and W are given by [g] = m/s2 , and [W ] = kg m/s2 = N . The
acceleration of gravity varies slightly over the surface of the Earth because the
radius of the Earth is not constant everywhere. If re is assumed to be constant,
then the acceleration of gravity is found to have the following values in the MKS,
FPS and CGS system of units

g = 9.807 m/s2 , g = 32.17 ft/s2 , g = 980.7 cm/s2 (5.17)

Example 5-4. Approximating Value of Escape Velocity

A rocket launched straight upward from the surface of the Earth will fall back
down if it doesnt achieve the correct velocity. Let r = r(t) denote the distance of
the rocket measured from the center of the Earth and let m = m(t) denote the mass
of the rocket which changes with time. The forces acting on the rocket as it moves
upward are the thrust from the engines, the pull of gravity and resistance due to air
friction called a drag force. By Newtons second law one can write
 
d d dr
(mv) = m = Ftotal = Fthrust Fgravity Fdrag (5.18)
dt dt dt
dr
where v = v(t) = is the velocity of the rocket. This is an equation, called a
dt
differential equation, which describes the motion of the rocket. When you learn
more about aerodynamics you will learn how to represent the thrust force and drag
forces on the rocket and then you can solve the resulting differential equation.
Instead, let us solve a much simpler problem created by making assumptions
which will reduce the equation (5.18) to a form which is tractable7
7
When confronted with a very difficult problem to solve, one can always make assumptions to simplify the
problem to a form which can be solved. Many times an analysis of the simplified solution produces an incite into
how to go about solving the more difficult original problem.
370
(i) Neglect the thrust force and drag force and consider only the gravitational force.
(ii) Assume the mass of the rocket remains constant.
(iii) Assume that at time t = 0, the initial velocity of the rocket is v0 and the position
of the rocket is given by r(0) = re , where re is the radius of the Earth.
These assumptions greatly simplify the differential equation (5.18) to the form
dv Gme m dv dv dr dv
m = where = = v (5.19)
dt r2 dt dr dt dr
and one obtains after simplification the differential equation
dv Gme
v = 2 (5.20)
dr r
One can separate the variables and express equation (5.20) in the form
dr
v dv = Gme (5.21)
r2
An integration of both sides of this separated equation gives the result
v2
 
dr Gme
v dv = Gme = = +C (5.22)
r2 2 r
where C is a constant of integration. The constant C is selected such that the initial
conditions are satisfied. This requires
v02 Gme v02 Gme
= +C = C= (5.23)
2 re 2 re
Substitute this value for C into the equation (5.22) and simplify the result to obtain
 
2 1 1
v = v02 + 2Gme (5.24)
r re
2Gme
In equation (5.24) the term is always positive so that if it is required that
r
2Gme
v02 0
re
then one can say that the velocity of the rocket will always be positive. This condi-
tion can be written
  1/2
2Gme Gme 
v0 = 2 2 re = 2g re 11200 m/s 7 miles per second
re re
Gme
where g = is the acceleration of gravity. This value is a good approximation of
re2
the velocity necessary to overcome the gravitational forces pulling the rocket back
to Earth. This velocity is called the escape velocity.
 371
Work
Let W denote the work done by a constant force F , which has moved an object
in a straight line a distance x in the direction of the force. Here W is defined8 as the
scalar quantity

Work = Force times distance W =Fx (5.25)

If the force F = F (x) varies continuously as the distance x changes, then if the
object is moved in a straight line an increment dx, the increment of work done dW
is expressed
dW = F (x) dx

and the total work done in moving an object from x1 to x2 in a straight line is given
by the integral  x 2

W = F (x) dx (5.26)
x1

The equation (5.26) tells us that the work done is nothing more than the area under

the curve F = F (x) between the values x1 and x2 .

Example 5-5.

If the constant force F acts at an angle to

the direction of motion, then the component of
force in the direction of motion is F cos and the
work done in moving an object is the component
of force in direction of the displacement times the displacement or W = (F cos ) s
If the force F = F (s) varies as a function of displacement s, then the increment
of work done in moving an object the incremented distance ds is

dW = (F (s) cos ) ds

and the total work done moving an object from s1 to s2 is

 s2
W = F (s) cos ds
s1

8
If there are many discrete forces acting on a body at different times, then one can define the work as the
average force times the displacement.
372
Recall that force is measured in units called Newtons, where 1 N = 1 kg m/s2 . Dis-
placement is measured in meters (m) so that work is force times distance and is
measured in units of Newton-meters or (N m) and one can write [W ] = N m, which
is read, The dimension of work is Newton-meter. By definition 1 N m = 1 Joule,
where Joule is abbreviated (J).
Energy
In the language of science the term energy is a scalar measure of a physical
systems ability to do work. There are many different kinds of energy. A few selected
types of energy you might have heard of are chemical energy, kinetic energy, various
kinds of potential energy, internal energy, elastic energy due to stretching or twisting,
heat energy, light energy and nuclear energy.
Kinetic Energy Ek
The energy associated with a body in motion is called kinetic energy and is
1
denoted by Ek . The kinetic energy is defined Ek = mv 2 , where m is the mass of
2
the body, [m] = kg and v is the velocity of the body, [v] = m/s. Kinetic energy is a
positive scalar quantity measured in the same units as work. One can verify that
[Ek ] = kg m2 /s2 = kg m/s2 m = N m = J


Example 5-6. The work done by a constant force moving an object in a F

straight line through a distance s during a time t is given by the integral
 s
W = F ds (5.27)
0

ds
Let s denote distance traveled during a time t with =v denoting the velocity
dt
dv d2 s
and a = = 2 denoting the acceleration. Using Newtons second law of motion
dt dt
one can write
dv d2 s ds d2 s dv
F = ma = m =m 2, where =v and 2
= =a (5.28)
dt dt dt dt dt
Substituting the equation (5.28) into the equation (5.27) gives
s s t
d2 s d2 s ds
    
dv
W = m ds = m 2 ds = m dt (5.29)
0 dt 0 dt 0 dt2 dt
Observe the equation (5.29) is written as an integration with respect to time by
ds dv d2 s
using the relations v = and = 2 . If the object has an initial velocity v0 at
dt dt dt
time t = 0, then the integration (5.29) can be expressed in the form
 t  t   t
dv 1 2 1 1 1
W = m v dt = m d v = mv 2 = mv 2 mv02 (5.30)
0 dt 0 2 2 0 2 2
 373
The equation (5.30) is a representation of the work-energy relation

The work done by forces acting on a body

equals the change in kinetic energy of the body

Potential Energy Ep
The energy associated with a body as a result of its position with respect to
some reference line is called the potential energy and is defined Ep = m g h, where m
is the mass of the body, [m] = kg, g is the acceleration of gravity, [g] = m/s2 and h
is the height of the body above the reference line, [h] = m. The potential energy is
sometimes called the gravitational potential energy. The potential energy is measured
in units of kg m/s2 m = N m = J and has the same units of measurement as work.
The work done against gravity in lifting a weight from a height h1 to a height h2 is
given by
 h2  h2 h2
W = Fg dx = mg dx = mg x = (mg h2 mg h1 ) = Ep
h1 h1 h1

where Fg = W is the weight acting downward. One can say the work done equals
the change in potential energy.

Example 5-7. Consider a ball of mass m which is thrown vertically

upward with an initial velocity v0 . Neglect air resistance
so that the only force acting on the ball is the force due
to gravity and construct a coordinate system with the
origin placed at the point where the ball is released.
Here the upward direction is taken as positive and by
Newtons second law one can write
dv
m = mg (5.31)
dt

since the weight of the ball is mg and this force is acting downward. Separate the
variables in equation (5.31) and then integrate to obtain
 v  t
m dv = mg dt or mv mv0 = mgt (5.32)
v0 0
374
If y denotes the distance of the ball above the reference axis, then the velocity of
dy
the ball is given by v = . The equation (5.32) can now be expressed in the form
dt
dy
m = mv0 mgt (5.33)
dt
dy
since the velocity v = represents the change in the height of the ball as a function
dt
of time. Multiply equation (5.33) by dt and integrate to obtain
 y  t
1
m dy = [mv0 mgt] dt or my = mv0 t mgt2 (5.34)
0 0 2

Solve equation (5.32) for the variable t and substitute for t in equation (5.34) and
then simplify to show
1 1
mv 2 + mgy = mv02 (5.35)
2 2
which can be interpreted as stating that the sum of the kinetic energy plus the
potential energy of the ball always has a constant value. Note that when the ball
reaches its maximum height, where y = h, the velocity of the ball is zero, and at this
time the equation (5.35) shows that the initial kinetic energy of the ball equals the
potential energy of the ball at its maximum height.

There are many more types of energy and all these energy types obey the law
of conservation of energy which states that there is no change in the total energy in
the Universe. Another way of saying this is to state that energy can be transformed,
but it cannot be created or destroyed.
First Moments and Center of Gravity
Consider a force F acting perpendicular to a
plane containing a line 0 0. The first moment
of a force F , also called a torque, is defined

Moment =(Force)(Lever arm distance)

(5.36)
M =F

where the lever arm  is understood to represent

the shortest perpendicular distance from the line
0 0 to the line of action of the force F . The moment is a measure of the ability of
the force to produce a rotation about the line 0 0. In general a quantity times a
distance to a point, or times a distance to a line, or times a distance to a plane, is
called a moment of that quantity with respect to a point, line or plane.
 375
Centroid and Center of Mass
In the figure 5-1 the x-axis is considered as a see-saw with weights W1 and
W2 placed at the positions (x1 , 0) and (x2 , 0) respectively. Consider the problem of
determining where one would place a fulcrum so that the see-saw would balance.
Let (x, 0) denote the point where the fulcrum is placed and let 2 = x2 x denote
a lever arm associated with the weight W2 and let 1 = x x1 denote the lever arm
associated with the weight W1 . The see-saw will balance if x is selected such that
the sum of the moments9 about the fulcrum equals zero. This requires that
M1 = 1 W1 = 2 W2 = M2 or (x x1 ) W1 = (x2 x) W2 (5.37)

Figure 5-1. Balancing of weights using moments.

Another way to express the balancing of the see-saw is to examine the distances
x x1 and x x2 . One distance is positive and the other is negative and the product
(x x1 )W1 gives a positive moment and the product (x x2 )W2 gives a negative
moment. One can then say that the moments produced by the weights balance if x
is selected such that the sum of the moments is zero or
2

W1 x1 + W2 x2
(x xi )Wi = 0 or x = (5.38)
W1 + W2
i=1

The point (x, 0) is then called the center of gravity or centroid of the system.

9
By placing the fingers of the right-hand in the direction of the force and letting the fingers move in the direction
of rotation produced by the force, then the thumb points in a positive or negative direction. If the z -axis comes
out of the page toward you, then this is the positive direction assigned to the moment. The moment M1 = 1 W1
is then said to be a positive moment and the moment M2 = 2 W2 is called a negative moment. The sum of the
moments equal to zero is then written 2 W2 + 1 W1 = 0.
376
If there are n-weights W1 , W2 , . . . , Wn placed at the positions (x1 , 0), (x2, 0), . . ., (xn, 0)
respectively, then the centroid of the system is defined as that point (x, 0) where the
sum of the moments produces zero or
n n

W1 x1 + W2 x2 + + Wn xn Wi xi
(x xi )Wi = 0 or x = = i=1
n (5.39)
i=1
W1 + W2 + + Wn i=1 Wi

n
If W = i=1 Wi is the total sum of the weights, then equation (5.39) can be written
as
W x = W1 x1 + W2 x2 + + Wn xn (5.40)

and this equation has the following interpretation. Imagine a three-dimensional

right-handed xyz Cartesian system of axes with the z -axis in figure 5-1 coming out
of the page. Each weight Wi then produces a moment Mi = Wi xi about the z -axis.
That is, for each value i = 1, 2, . . ., n, the distance xi denotes the lever arm and Wi
denotes the force. The total sum of these moments gives the right-hand side of
equation (5.40). The left-hand side of equation (5.40) is then interpreted as stating
that if the sum of the weights W was placed at the position (x, 0), it would create
a moment M = W x equivalent to summing each individual moment produced by all
the weights. The position x is then called the center of gravity or centroid of the
system. Another interpretation given to equation (5.39) is that the numerator is a
weighted sum of the x-values and the denominator is the sum of the weights so that
x is then a weighted average of the x-values.
One can generalize the definition of a first moment by defining a first moment
associated with just about any quantity. For example, one can define first moments
such as
Mf =(f orce)(lever arm) Mv =(volume)(lever arm)
.. ..
Mm =(mass)(lever arm) . . (5.41)
Ma =(area)(lever arm) Mq =(quantity)(lever arm)
where the lever arm is understood to represent the shortest perpendicular distance
from some reference point, line or plane to the quantity.

Example 5-8. Let m1 , m2 , . . . , mn denote n point masses located respectively at

the points (x1 , y1 ), (x2 , y2 ), . . . , (xn, yn). Find the center of mass (x, y) of this system of
masses.
 377
Solution
Use the x and y-axes as the lines about which
one can take first moments associated with the
given point masses. If m = ni=1 mi is the total


sum of all the point masses, then if this mass

were placed at the point (x, y) it would produce
first moments about these axes given by

Mx = m y and My = m x

These moments must be equivalent to the sum of the first moments produced by
each individual mass so that one can write
n

n

My = m x = mi xi and Mx = m y = mi yi
i=1 i=1

The center of mass of the system then has the coordinates (x, y) where
n n
mi xi My mi yi Mx
x = i=1
n = and y = i=1
n =
i=1 mi m i=1 mi m

Here the center of mass of the system of masses has coordinates (x, y) where x is a
weighted sum of the xi values and y is a weighted sum of the yi values for positions
ranging from i = 1, 2, . . ., n

Centroid of an Area
Moments can be used to find the centroid of an area bounded by the curve
y = f (x) > 0, the x-axis and the lines x = a and x = b. Partition the interval [a, b] into
n equal parts with

ba
a = x0 , x1 = x0 + x, x2 = x0 + 2x, . . . , xn = x0 + nx = b where x =
n

Consider the center of the rectangular element of area illustrated in the figure 5-2
which has the coordinates (i , yi ), where i = xi1 + x
2
and yi = 12 f (i ). The center of
this element of area has a first moment about the y-axis given by

My = (lever arm)(area) = (i )[f (i ) xi]

and it also has a first moment about the x-axis given by

 
1
Mx = (lever arm)(area) = f (i ) [f (i ) xi ]
2
378
A summation of the first moments associated with each rectangle produces a sum
from 1 to n giving the total moments
n n


1
My = i f (i )xi Mx = [f (i )]2 xi
2
i=1 i=1

Figure 5-2. Moments for Centroid of an area.

Neglecting infinitesimals of higher order and using the fundamental theorem of in-
tegral calculus one finds that in the limit as xi 0, the above sums become the
definite integrals
 b  b
1
My = xf (x) dx Mx = [f (x)]2 dx (5.42)
a a 2
The total area under the curve y = f (x) is given by the definite integral
 b
A= f (x) dx
a

and if this total area were concentrated and placed at the point (x, y) it would produce
moments about the x and y-axes given by Mx = A y and My = A x. The centroid is
that point (x, y) where
 b  b
1
Mx = A y = [f (x)]2 dx and My = A x = xf (x) dx (5.43)
a 2 a

M M
from which one can solve for x and y to obtain y = x and x = y .
A A
In a similar fashion one can use the fundamental theorem of integral calculus
to show the lever arms associated with the first moments of the center point of an
element of area can be expressed in terms of the x and y coordinates associated
with the element of area. One can then verify the following lever arm equations
associated with the elements of area illustrated.
 379
1.) For the center point of the element of area
dA = y dx
lever arm to y-axis is x
lever arm to x-axis is y/2
2.) For the center point of the element of area
dA = (y2 y1 ) dx
lever arm to y-axis is x
lever arm to x-axis is 12 (y1 + y2 )

3.) For the center point of the element of area

dA = x dy
lever arm to y-axis is x/2
lever arm to x-axis is y
4.) For the center point of the element of area
dA = (x2 x1 ) dy
lever arm to y-axis is 12 (x1 + x2 )
lever arm to x-axis is y

Note that in determining the above lever arm distances the infinitesimals of
higher order have been neglected.
For example, associated with the last figure there is an element of area given by
dA = (x2 x1 ) dy = [g(y) f (y)] dy and the total area is given by
 d
A= [g(y) f (y)] dy
c

This element of area has a moment about the x-axis given by

dMx = (lever arm)(area) = y dA = y[g(y) f (y)] dy

and a moment about the y-axis given by

1 1 2
g (y) f 2 (y) dy

dMy = (lever arm)(area) = [g(y) + f (y)][g(y) f (y)] dy =
2 2
Summing these moments one finds
 d  d
Mx = y[g(y) f (y)] dy and My = [g 2 (y) f 2 (y)] dy
c c

with the centroid (x, y) found from the relations

My Mx
x = and y =
A A
380Symmetry
If an object has an axis of symmetry, then
the centroid of the object must lie on this line.
For example, if an area has a line of symmetry,
then when the area is rotated 180 about this
line the area has its same shape. Examine the
rectangle when rotated about a line through its center and find out the rectangle is
unchanged. One can say the centroid for the rectangle is at its geometric center.

Example 5-9. Use the equations (5.43) and find the centroid of a rectangle of
height h and base b.
Solution
Here y = f (x) = h is a constant and so one
can write
 b  b
1 2
My = xf (x) dx = xh dx = hb
0 0 2
 b  b
1 1 2 1
Mx = [f (x)]2 dx = h dx = bh2
0 2 0 2 2
The total area of the rectangle is A = bh and so the centroid (x, y) is determined by
the equations
My b Mx h
x = = and y = =
A 2 A 2

Example 5-10. Find the centroid of the area bounded by the x-axis, the y-
axis and the ellipse defined by the parametric equations x = a cos , y = b sin , for
0 /2 and a > b > 0 constants.
Solution

The area to be inves-

tigated is the upper quad-
rant of an ellipse. Move out
a distance x from the origin
and construct an element of area dA = y dx. and substitute y = b sin , x = a cos with
dx = a sin d and show the total area is
 a  0  /2
A= y dx = b sin (a sin ) d = ab sin2 d
0 /2 0
 /2
1
A =ab [1 cos 2] d = ab
0 2 4
 381
The element of area dA has a moment about the y-axis given by
 a  0  /2
1 2
My = xy dx = (a cos )(b sin )(a sin ) d = a2 b sin2 cos d = a b
0 /2 0 3
The element of area has a moment about the x-axis given by
 a  0
1 1 2 2
Mx = y(y dx) = b sin (a sin ) d
0 2 /2 2

ab2 /2 3 ab2 /2 3
 

1 1
Mx = sin d = sin sin 3 d = ab2
2 0 2 0 4 4 3
My 4a My 4 b
The centroid (x, y) is given by x = = and y = =
A 3 A 3

Example 5-11. Find the centroid of the triangle with vertices (0, 0), (b, 0), (c, h)

The equation of the line 1 with slope h/c is

given by y = (h/c)x. The equation of the line 2
with slope h/(c b) is given by y = [h/(c b)] (x b)
Construct a horizontal element of area
dA = (x2 x1 ) dy

and show



(c b) c b
dA = b + y y dy = b y dy
h h h
and after summing these elements of area one finds
 h

b 1
A= b y dy = bh
0 h 2
This element of area has a moment about the x-axis given by
 h

b 1
Mx = y b y dy = bh2
0 h 6
and moment about the y-axis given by
 2 
h
c2 2

1 (c b) 1
My = b+ y 2 y dy = hb(b + c)
0 2 h h 6

The centroid for the given triangle is (x, y) where

My 1 Mx h
x = = (b + c) and y = =
A 3 A 3
382
Table 5-1 Centroids of Some Simple Shapes
Shape Area x y

1 1 1
Triangle bh (1 + )b h
2 3 3

b h
Rectangle bh
2 2

Quadrant 4r 4r
of circle r2
4 3 3

Quadrant 4a 4b
of ellipse ab
4 3 3

2 sin
Wedge r2 r 0
3
 383
Centroids of composite shapes
If an area is composed of some combination of simple shapes such as triangles,
rectangles, circles or some other shapes where the centroids of each shape have
known centroids, then the resultant moment about an axis is the algebraic sum of
the moments of the component shapes and the centroid of the composite shape is
given by x = MAy and y = MAx , where A is the total area of the composite shape.
Whenever the centroids of all the individual shapes which make up the total shape
are known, then integration is not required.

Example 5-12. If the composite shape is composed of n known shapes having

area A1 with centroid (x1 , y1 )
area A2 with centroid (x2 , y2 )
..
.
area An with centroid (xn , yn )
then the total area of the composite shape is

A = A1 + A2 + A3 + + An

The total moment produced about the yaxis from each area is

My = A1 x1 + A2 x2 + A3 x3 + + An xn

The total moment produced about the xaxis from each area is

Mx = A1 y1 + A2 y2 + A3 y3 + + An yn

The centroid of the composite shape is therefore

My Mx
x = and y =
A A

Centroid for Solid of Revolution

Consider the area bounded by the curve y = f (x) > 0 and the lines x = a, x = b > a
and the x-axis which is revolved about the x-axis to form a solid of revolution. One
can construct an element of volume dV for this solid using the disk generated when
dA is rotated about the x-axis. The volume element associated with this disk is
 b
dV = y 2 dx with V = y 2 dx
a
384
where the integral represents a summation of the volume elements. The moment of
the disk about the plane perpendicular to the axis of rotation which passes through
the origin is  b
dM = x dV = x y 2 dx and M= xy 2 dx
a

where x is the lever arm distance from the plane to the volume element and M is a
summation of these moments. The above integral is called the first moment of the
solid of revolution with respect to the plane through the origin and perpendicular
to the axis of rotation. The centroid x is then defined as
 b
xy 2 dx
M
x = = a b (5.44)
V
y 2 dx
a

and by symmetry the position of x is on the axis of rotation.

Centroid for Curve

Let y = f (x) define a smooth continuous

curve for a x b. At the position (x, y) on
the curve construct the element of arc length

ds = dx2 + dy 2 .
The first moments about the x and y-axes associated with the curve

C = { (x, y) | y = f (x), a x b }

are defined as a summation of the first moments associated with the element of arc
length ds. One can define

dMx = (lever arm)(element of arc length) = y ds

dMy = (lever arm)(element of arc length) = x ds

A summation of these first moments gives

 b  b
Mx = y ds and My = x ds
a a

If the given curve C has an arc length s given by


 b  2
dy
s= 1+ dx,
a dx
 385
then the centroid of the curve C is defined as the point (x, y) where
My Mx
x = and y = (5.45)
s s
That is, if the arc length s could be concentrated and placed at a point (x, y) called
the centroid, then the centroid is selected so that the first moment x s is the same
as that produced by the summation of individual moments about the y-axis and
the first moment y s is the same as that produced by the summation of individual
moments about the x-axis.

Example 5-13. Consider the arc of a circle which lies in the first quadrant.
This curve is defined C = { (x, y) | x = r cos , y = r sin , 0 /2 } where r is the
radius of the circle. Find the centroid associated with this curve.
Solution

The element of arc length squared is given

by ds2 = dx2 + dy 2 so that one can write

 /2  /2  2  2
dx dy
s= ds = + d
0 0 d d
 /2  /2 /2 1
s= ds = r d = r = r
0 0 0 2

Use the equations (5.45) and calculate the first moments of the curve about the
x and y -axes to obtain
 /2  /2 /2
Mx = y ds = r sin r d = r 2 ( cos ) = r2
0 0 0
 /2  /2 /2
My = x ds = r cos r d = r 2 (sin ) = r2
0 0 0

My 2 Mx 2
The centroid (x, y) is then x = = r and y = = r
s s

Higher Order Moments

The various first moments defined by the relation (5.41) can be further general-
ized to second moments by replacing the lever arm by the lever arm squared. Second
moments about a line  are denoted I and one can write
I =(f orce)(lever arm)2 I =(volume)(lever arm)2
.. ..
I =(mass)(lever arm)2 . . (5.46)
I =(area)(lever arm)2 I =(quantity)(lever arm)2
386
where the lever arm is understood to represent a perpendicular distance from some
reference line
. Second moments are referred to as moments of inertia. In the
study of rotational motion of rigid bodies it is found that the moment of inertia is
a measure of how mass distribution affects changes to the angular motion of a body
as it rotates about an axis.
Third order moments would involve the lever arm cubed and nth order moments
would involve the lever arm raised to the nth power. Third order and higher order
moments arise in the study of statistics, mechanics and physics.

Example 5-14.
Given a region R one can construct at a general
point (x, y) R an element of area dA. This element
of area has a second moment of inertia about the
yaxis given by
dIyy = x2 dA

and summing these second moments of inertia over

the region R gives the total second moment about
the y-axis as



Iyy = x2 dA
R

In a similar fashion, the second moment of inertia of the element dA about the xaxis
is given by
dIxx = y 2 dA

and a summation of these second moments over the region R gives the total second
moment about the x-axis as


Ixx = y 2 dA
R

If the moment axis is perpendicular to the plane in which the region R lies, say
a line through the origin and perpendicular to the x and y axes, then the second
moment with respect to this line is called a polar moment of inertia and is written









2 2 2 2
J00 = r dA = (x + y ) dA = x dA + y 2 dA = Iyy + Ixx
R R R R

which shows the polar moment of inertia about the line through the origin is the
sum of the moments of inertia about the x and y axes.
 387
An examination of the Figure 5-3 shows that if y is the line x = x0 parallel to
the yaxis, then an element of area dA has a second moment with respect to the
line y given by  
Iy y = |x x0 |2 dA = (x x0 )2 dA (5.47)
R R

If x is the line y = y0 which is parallel to the xaxis, then the element of area dA
has the second moment with respect to line x given by
 
Ix x = |y y0 |2 dA = (y y0 )2 dA (5.48)
R R

Figure 5-3.
Second moments with respect to lines parallel to the x and y axes.

Expanding equation (5.47) one finds


Iy y = (x2 2xx0 + x20 ) dA
 R  
2 2 (5.49)
= x dA 2x0 x dA + x0 dA
R R R
Iy y =Iyy 2x0 My + x20 A

Similarly, if one expands the equation (5.48) one finds that

Ix x = Ixx 2y0 Mx + y02 A (5.50)

The results given by the equations (5.49) and (5.50) are known as the basic equations
for representing the parallel axes theorem from mechanics. This theorem states that
if you know the area of a region and the first and second moments of the region
about one of the coordinate axes, then you can find the second moment about any
axis parallel to the coordinate axes by using one of the above results.
388
Example 5-15. Let d = dx dy dz denote an element of volume and d = dm
denote an element of mass, where is the density of the solid. The second moments
of mass with respect to the x, y and z axes are given by




Ixx = (y 2 + z 2 ) d




Iyy = (x2 + z 2 ) d




Izz = (x2 + y 2 ) d

where integrations are over the volume defining the solid.

Example 5-16.

Consider a particle P with constant mass m

rotating in a circle of radius r about the origin
having a tangential force F producing the mo-
tion. Let denote the angular displacement of
d
the particle, = the angular velocity of the
dt
d d2
particle, = 2 = ,
the angular acceleration
dt dt
of the particle. Newtons second law of motion
dv
can then be expressed F = m .
dt
ds d
If s = r is the distance traveled by the particle, then is the =v =r = r
dt dt
change in distance with respect to time or speed10 of the particle, so that Newtons
second law can be expressed
 
dv d d d2
F =m =m r = mr 2 (5.51)
dt dt dt dt
Note that if F is the tangential force acting on the particle, then M = F r is the torque
or first moment of the force about the origin. Consequently, multiplying equation
d2
(5.51) on both sides by r one finds F r = mr2 2 . Here F r = M is the first moment or
dt
torque about the origin, mr2 = I is the mass times the lever arm squared or moment
of inertia of the mass about the origin. The Newtons law for rotational motion can
therefore be expressed in the form
d2
M = I or M =I (5.52)
dt2
10
Speed is a scalar quantity representing the magnitude of velocity which is a vector quantity.
 389
Example 5-17. Find the centroid and moments of inertia about the x and y
axes associated with the semi-circle x2 + y2 = R2 for x > 0.

Solution The area inside the semi-circle is A = 2 R2 . At

a general point (x, y) within the semi-circle construct an
element of area dA = dx dy , then the first moment about the
yaxis is given by

dMy = x dA = x dx dy

and a summation over all elements dA within the semi-circle gives the first moment

 x=R  y= R2 x2  R 
My = x dy dx My = 2x R2 x2 dx
x=0 y= R2 x2 0

Make the substitution u = R2 x2 with du = 2x dx and show the integration

2 My 2R3 /3 4R
produces the result My = R3 , so that x = = 2
= . By symmetry, the
3 A R /2 3
centroid must lie on the line y = 0 so that the centroid of the semi-circle lies at the
4R
point (x, y) = ( , 0).
3
The second moments of the area element about the x and yaxes gives

dIxx = y 2 dA = y 2 dy dx and dIyy = x2 dA = x2 dx dy

A summation over the area of the semi-circle gives the integrals

 x=R  y= R2 x2  R
2 4
Ixx = y 2 dy dx = (R2 x2 )3/2 dx = R
x=0 y= R2 x2 3 0 8
 x=R  y=R2 x2  R
4
Iyy = x2 dy dx = 2x2 (R2 x2 )1/2 dx = R
x=0 y= R2 x2 0 8

where the substitution x = R sin can be used to aid in evaluating the above integrals.
Let Icc denote the moment of inertia about theline x = x0 = 4R 3
through the
8
centroid. The parallel axis theorem shows that Icc = R4 .
8 9
390
Example 5-18. Find the centroid and moments of inertia about the x and
yaxes associated with the circular sector bounded by the rays = 0 and = 0
and the circle r = R.
Solution The area inside the circular sector is given by
A = 0 R2 . Move to a general point (x, y) within the sector
and construct an element of area dA = r drd. The x and y
lever arms are given by x = r cos and y = r sin . The first
moment of this area element about the yaxis is given by

dMy = x dA = (r cos )r drd

and a summation of these elements over the area of the sector gives
 r=R  0
2 3
My = cos d r 2 dr = R sin 0
r=0 0 3
My 2 sin 0
The x value for this area is given by x = = R . By symmetry, the
A 3 0
centroid lies on the ray = 0 where (x, y) = ( 23 R sin00 , 0).
The elements for the second moments about the x and yaxes are given by

dIxx = y 2 dA = (r 2 sin2 ) r drd and dIyy = x2 dA = (r 2 cos2 ) r drd

and summing these elements over the area of the sector gives the moments of inertia
R  0
R4

Ixx = r 3 sin2 d dr = (20 sin 20 )
0 0 8
r 0
R4
 
Iyy = r 3 cos2 d dr = (20 + sin 20 )
0 0 8

Moment of Inertia of an Area

The moment of inertia of a general plane area

with respect to an axis can be calculated as fol-
lows. Construct elements of area dA all in the
shape of a rectangle within the plane area and
which are parallel to the axis not passing through
the area. By definition, the element of moment
of inertia associated with an element of area dA
 391
Table 5-2 Moments of Inertia of Some Simple Shapes
Shape Area Ixx Iyy

1 1 1 3
Triangle bh bh3 b h(1 + + 2 )
2 12 12

1 3 1 3
Rectangle bh bh b h
3 12

Quadrant 1 1
of circle r2 r4 r4
4 16 16

Quadrant 1 1
of ellipse ab ab3 a3 b
4 16 16

r4 r4
Wedge r2 (20 sin 20 ) (20 + sin 20 )
8 8
392
is the lever arm squared times the element of area or dI = 2 dA, where represents
the distance from the axis to an element of area dA. The total moment of inertia is
then a summation over all rectangular elements. If and are values denoting the
extreme distances of the plane area from the axis, then the moment of inertia of the
plane area is determined by evaluating the integral
 =
I = 2 dA (5.53)
=

Here it is assumed that the element of area dA can be expressed in terms of the
distance .
Moment of Inertia of a Solid

In the study of mechanics one fre-

quently encounters the necessity to cal-
culate the moment of inertia of a solid
which is generated by revolving a plane
area about an axis. If a rectangular ele-
ment of area is construct within the plane
area and is rotated about an axis  not
passing through the area, then a cylindrical shell shaped volume element dV is
generated. Multiplying this volume element by the density of the solid creates an
element of mass dm = dV . The element of the moment of inertia is then given by

dI = 2 dm = 2 dV (5.54)

where is the distance of dm from the axis of rotation. If the extreme distances
of the plane area from the axis of rotation have the values a and b, then the total
moment of inertia about the rotation axis is
 =b  =b
2
I = dm = 2 dV (5.55)
=a =a

If the solid is homogeneous, then the density is a constant so that the moment of
inertia can be expressed  =b
I = 2 dV (5.56)
=a
 393
Moment of Inertia of Composite Shapes
To calculate the moment of inertia of a composite area about a selected axis
(i) Calculate the moment of inertia of each component about the selected axis.
(ii) Next one need only sum the moments of inertia calculated in step (i) to calculate
the moment of inertia of the given composite area.
That is, the moment of inertia of a composite area about an axis is equal to the
sum of the moments of the component areas with respect to the same axis. Note that
if a component of the shape is removed, then this places a hole in the composite
shape and in this case the moment of inertia of the component removed is then
subtracted from the total sum.
Pressure
The average density of a substance is defined as its mass m divided by its
volume V or = m V
, where [] = kg/m3 , [m] = kg, [V ] = m3 . The relative density of a
substance is defined as the ratio of density of substance divided by the density of
water. Pressure is a scalar quantity defined as the average force per unit of area and
its unit of measurement is the Pascal, abbreviated (Pa), where 1 Pa = 1 N m2 .
Liquid Pressure
Integration can be used to determine the forces acting on submerged objects.
F
Pressure at a point is p = lim , and represents a derivative of the force with
A0 A
respect to area. An area submerged in water experiences only a pressure normal to
its surface and there are no forces parallel to the area. This is known as Pascals
law. Knowing the pressure at a point, one can use integration to calculate the total
force acting on a submerged object. The pressure p representing force per unit of
area must be known when constructing water-towers, dams, locks, reservoirs, ships,
submarines, under-water vessels as the total force acting on a submerged object
must be known for certain design considerations.
Consider two points P1 and P2 beneath a fluid having a constant density . If
h = |P1 P2 | is the distance between the points and is the constant density of the
liquid, then the change in pressure between the points P1 and P2 is given by

p = g h (5.57)

where h, [h] = m, is measured positive in the downward direction, , [] = kg/m3 ,

is the density and g , [g] = m/s2 is the acceleration of gravity so that [p] = N/m2 .
Defining w = g , [w] = N/m3 as the weight of the liquid per unit volume, one can
394
write p = w h. Note the pressure increases with depth and depends only on the
quantities vertical distance, density and acceleration of gravity. The pressure is then
the same at all points of a submerged area lying on a horizontal line of constant
depth within the liquid.
Consider a vertical plane area submerged in
a liquid. Using the fact that the pressure is
the same at all point lying on a horizontal
line at constant depth, one can construct an
element of area dA = (h) dh on the submerged
area as illustrated, where  = (h) is the hor-
izontal dimension of the element of area.
The element of force exerted by the liquid on one side of the submerged element
of area is given by
dF = p dA = w h dA = w h (h) dh

and so the total force acting on the submerged area is given by the summation of
forces  h2
F = w h (h) dh (5.58)
h1

The representation for the total force given by equation (5.58) assumes that the
element of area can be expressed in terms of the depth h. If one selects a different
way of representing the position of the submerged object, say by constructing an
x, y -axes somewhere, then the above quantities have to be modified accordingly.
Gas Pressure
The equation (5.57) is valid for the change in gas pressure between two points P1
and P2 for small volumes. However, for small volumes the gas pressure is very small
and p remains small unless h is very large. One usually makes use of the fact that
the gas pressure is essentially constant at all points within a volume of reasonable
size. When dealing with volumes of a very large size, like the Earths atmosphere,
the equation (5.57) is no longer valid. Instead, one usually uses the fact that (i)
the pressure decreases as the height h above the Earth increases and (ii) the density
of the air varies widely over the surface of the Earth. Under these conditions one
uses the approximate relation that the change in pressure with respect to height is
proportional to g and one writes
dp
= g (5.59)
dh
 395
The negative sign indicating that the pressure decreases with height. Note that the
pressure has a wide range of values over the Earths surface, varying with tempera-
ture, humidity, molar mass of dry air and sea level pressure. The average sea level
pressure being 101.325 kPa or 760 mmHg. One can find various empirical formulas
for variations of the density determined by analyzing weather data.
Chemical Kinetics
In chemistry a chemical reaction describing how hydrogen (H2 ) and oxygen (O2 )
combine to form water is given by
k
2H2 + O2
2H2 O

This reaction is a special case of a more general chemical reaction having the form
kf

m B +m B +m B +
n1 A1 + n2 A2 + n3 A3 +  (5.60)
1 1 2 2 3 3
kr
where A1 , A2 , A3 , . . . represent molecules of the reacting substances, called reactants
and B1 , B2, B3 , . . . represent molecules formed during the reaction, called product el-
ements of the reaction. The coefficients n1 , n2 , n3 , . . . and m1 , m2 , m3 , . . . are either
positive integers, zero or they have a fractional value. These values indicate the
proportion of molecules involved in the reaction or proportions involved when the
reactants combine. These coefficients are referred to as stoichiometric coefficients.
The constants kf and kr are positive constants called the forward and reverse re-
action rate coefficients. If kr = 0, then the reaction goes in only one direction.
The stoichiometric representation of a reaction gives only the net result of a re-
action and does not go into details about how the reaction is taking place. Other
schemes for representing a reaction are used for more complicated reactions. One
part of chemistry is the development of mathematical models which better describe
the mechanisms of how elements and compounds react and involves the study of re-
action dynamics of chemicals. This sometimes requires research involving extensive
experimental and theoretical background work in order to completely understand all
the bonding and subreactions which occur simultaneously during a given reaction.
Simple chemical reactions can be described using our basic knowledge of calculus.
Rates of Reactions
A reaction is called a simple reaction if there are no intermediate reactions or
processes taking place behind the scenes. For example, a simple reaction such as
k
n1 A1 + n2 A2
m1 B1 + m2 B2 (5.61)
396
states that n1 molecules of A1 and n2 molecules of A2 combined to form m1 molecules
of B1 and m2 molecules of B2 . Let [A1 ], [A2], [B1], [B2] denote respectively the concen-
trations of the molecules A1 , A2 , B1, B2 with the concentration measured in units of
moles/liter. The stoichiometric reaction (5.61) states that as the concentrations of
A1 and A2 decrease, then the concentrations of B1 and B2 increase. It is assumed
that A1 and A2 decrease at the same rate so that
1 d[A1 ] 1 d[A2 ]
= (5.62)
n1 dt n2 dt

and the concentrations of B1 and B2 increase at the same rate so that

1 d[B1] 1 d[B2 ]
= (5.63)
m1 dt m2 dt

Equating the equations (5.62) and (5.63) gives

1 d[A1 ] 1 d[A2 ] 1 d[B1] 1 d[B2]
= = = (5.64)
n1 dt n2 dt m1 dt m2 dt

Here a standard rate of reaction is achieved by taking the rate of change of each
substance and dividing by its stoichiometric coefficient. Also note that the minus
signs are used to denote a decrease in concentration and a plus sign is used to denote
an increase in concentration.
The Law of Mass Action
There are numerous and sometimes complicated rate laws for describing the
chemical kinetics of a reaction. These complicated rate laws are avoided in presenting
this introduction to chemical kinetics. For simple chemical reactions at a constant
temperature which have the form of equation (5.60), let x = x(t) denote the number
of molecules per liter which have reacted after a time t. Many of these simple
equations obey the law of mass action which states that the rate of change of x = x(t)
with respect to time t can be represented
dx
= k[A1 ]n1 [A2 ]n2 [A3 ]n3 (5.65)
dt

Each of the superscripts n1 , n2 , n3 , . . . have known values and k is a rate or velocity

coefficient having units of 1/time. The order of the chemical reaction is represented
by the sum n1 + n2 + n3 + of these exponents. If the sum is one, the reaction is
called a first-order reaction or unimolecular reaction. If the sum is two, the reaction
is called a second-order reaction or bimolecular. If the sum is three, the reaction is
 397
called a third-order reaction or trimolecular, etc. Note that trimolecular and higher
order reactions are rare.
An example of a unimolecular reaction is a substance disintegrating and this
type of reaction can be represented
d[A]
= k[A], [A] = [A](t) (5.66)
dt

and this equation is a way of stating that the rate of change of a decaying substance
is proportional to the amount present. The proportionality constant k being called
d[A]
the rate coefficient. Separate the variables in equation (5.66) and write = k dt
[A]
and then integrate both sides from 0 to t, assuming that at time t = 0, [A](0) = [A]0
is the initial concentration. One can then write
 [A]  t
d[A]
= k dt
[A]0 [A] 0

and then these equations can be integrated to obtain

ln [A] ln [A]0 = kt = [A] = [A](t) = [A]0 ekt (5.67)

where the reaction rate k has dimensions of 1/time. Use equation (5.67) and plot
[A](t) versus t one finds the result is a straight line on semi-log paper. A second-order
reaction or bimolecular reaction has the form
kf

B
A1 + A2  (5.68)
1
kr

and represents a reversible bimolecular reaction. Here kf is the forward rate constant
and kr is the reverse rate constant. One can alternatively write the forward and
reverse reactions as two separate equations. If kr = 0, then there is no reverse
reaction. Apply the law of mass action to the stoichiometric reaction (5.68) gives
the differential equations
d[A1 ]
=kr [B1 ] kf [A1 ][A2 ]
dt
d[A2 ]
=kr [B1 ] kf [A1 ][A2 ] (5.69)
dt
d[B1]
= kr [B1] + kf [A1 ][A2 ]
dt
Note that the rate coefficients kr and kf can have very large differences in magnitudes
thus driving the reaction more in one direction than the other and for the reaction
398
(5.68) the rate coefficients kf and kr do not have the same units of measurements.
To show this one should perform a dimensional analysis on each of the terms in the
equations (5.69). Each group of terms in equation (5.69) must have the same units of
measurements so that by examining the dimensions of each term in a group one can
show the reaction rates kf and kr do not have the same dimensions. For example, if
the concentrations are measured in units of mol/liter, then the terms on the left-hand
side of the equations (5.69) all have units of mol/liter per second and consequently
each group of terms on the right-hand side of the equations (5.69) must also have this
same unit of measurement. This requires that kr have units of 1/second and kf have
units of mol 1 . If different units of measurement are used one must perform
second
liter
a similar type of analysis of the dimensions associated with each group of terms.
The requirement that each group of terms have the same dimensions is known as
requiring that the equations be homogeneous in their dimensions. If an equation is
not dimensionally homogeneous, then you know it is wrong.
In the equations (5.69) let kr = 0 to obtain
d[A1 ]
= kf [A1 ][A2 ]
dt
d[A2 ]
= kf [A1 ][A2 ] (5.70)
dt
d[B1 ]
= kf [A1 ][A2 ]
dt

Let y = y(t) = [B1] denote the concentration of B1 as a function of time t. If at time

t = 0 the concentrations of A1 and A2 are denoted by [A1 ]0 and [A2 ]0 , then after a
time t one can express the amount of A1 and A2 by using the equations

[A1 ] = [A1 ](t) = [A1 ]0 y and [A2 ] = [A2 ](t) = [A2 ]0 y

Substituting these values into the last of the equations (5.70) gives the result
dy
= kf ([A1 ]0 y)([A2 ]0 y) (5.71)
dt

Let 1 = [A1 ]0 and 2 = [A2 ]0 and assume 1
= 2 so that equation (5.71) can be
expressed in the form
dy
= kf dt (5.72)
(1 y)(2 y)
 399
where the variables have been separated. Use partial fractions and write
1 A B
= +
(1 y)(2 y) 1 y 2 y

1 1
and show A = 2 1
and B = 2 1
= A. The equation (5.72) can then be expressed
the following form
A dy A dy
= kf dt, 1
= 2 (5.73)
1 y 2 y
which is easily integrated to obtain

A ln |1 y| + A ln |2 y| = kf t + C

where C is a constant of integration. Solving for y gives

 
 2 y 
A ln 
  =kf t + C
1 y 
 
 2 y 
ln 
  =kf (2 1 )t + C , where C = C(2 1 ) (5.74)
1 y 
2 y
=Kekf (2 1 )t , where K = eC is some new constant.
1 y

Evaluate equation (5.74) at time t = 0 with y(0) = 0 to show K = 12 . Substituting

this value into the last equation in (5.74) and using algebra to solve for y one finds
that 
1 ekf (2 2 )t
y = y(t) = 1 2 (5.75)
1 2 ekf (2 1 )t
In the special case 1 = 2 , the equation (5.72) takes on the form
dy
= kf dt (5.76)
(1 y)2

As an exercise integrate both sides of this equation and show

21 kf t
y = y(t) = (5.77)
1 + 1 kf t

Differential Equations
Equations which contain derivatives which are of the form
dn y dn1y d2 y dy
L(y) = a0 (x) n
+ a1 (x) n1
+ + an2 (x) 2
+ an1 (x) + an (x)y = 0 (5.78)
dx dx dx dx
400
where a0 , a1 , . . . , an are constants or functions of x, are called linear nth order homo-
geneous differential equations and linear differential equations of the form

dn y dn1y d2 y dy
L(y) = a0 (x) n
+ a1 (x) n1
+ + an2 (x) 2
+ an1 (x) + an (x)y = F (x) (5.79)
dx dx dx dx

are called linear nth order nonhomogeneous differential equations. The symbol L is
a shorthand notation to denote the linear differential operator
dn ( ) dn1( ) d2 ( ) d( )
L( ) = a0 (x) n
+ a1 (x) n1
+ + an2 (x) 2
+ an1(x) + an (x)( ) (5.80)
dx dx dx dx

An operator L is called a linear differential operator when it satisfies the conditions

L(y1 + y2 ) = L(y1 ) + L(y2 )
(5.81)
L(y) = L(y)

where is a constant. The first condition is satisfied because a derivative of sum is

the sum of the derivatives and the second condition is satisfied because the derivative
of a constant times a function is the constant times the derivative of the function.
Differential equations with ordinary derivatives, not having the forms of equations
(5.78) or (5.79) , are called nonlinear differential equations. In general, linear differ-
ential equations are easier to solve than nonlinear differential equations.
A solution of the differential equation (5.78) is any continuous function y = y(x)
which can be differentiated n-times and one can show that when the function y = y(x)
and its derivatives are substituted into the equation (5.78) then an identity results.
The function y = y(x) is then said to have satisfied the conditions specified by the
differential equation. The differential equation (5.78) has a nth derivative term and
consequently it would require n-integrations to obtain the solution. The general
solution will therefore contain n arbitrary constants. Sometimes it is possible to
integrate the differential equation and determine the solution by integration meth-
ods. Sometimes the given differential equation has a special form where short cut
methods have been developed for obtaining a solution.
The general procedure to solve the linear nth order nonhomogeneous equation
(5.79) is to first solve the homogeneous differential equation (5.78) by finding n
linearly independent solutions {y1 (x), y2 (x), . . . , yn (x)}, called a fundamental set of so-
lutions, where each function yi (x) satisfies L(yi (x)) = 0 for i = 1, 2, . . ., n. The general
solution of the linear homogeneous differential equation (5.78) is then any linear
 401
combination of the functions in the fundamental set. The general solution of the
linear homogeneous equation can be expressed

yc = c1 y1 (x) + c2 y2 (x) + + cn yn (x) (5.82)

where c1 , c2 , . . . , cn are constants and yc is called a complementary solution. After

determining the complementary solution one then tries to find any function yp which
satisfies L(yp ) = F (x). The function yp is then called a particular solution of the
nonhomogeneous linear differential equation (5.79). The general solution to the linear
nonhomogeneous differential equation (5.79) is written

y = yc + yp (5.83)

which represents a sum of the complementary and particular solutions.

Spring-mass System
Consider a vertical spring which is suspended from a support as illustrated in
figure 5-4(a). Consider what happens when a weight W is attached to a linear spring,
and the spring stretches some distance s0 , and the weight remains at rest in an
equilibrium position as illustrated in the figure 5-4(b). The weight is in equilibrium
because the downward force W is offset by the upward spring restoring force and
these forces must be equal and in opposite directions. If the weight is displaced
from this equilibrium position and then released, it undergoes a vibratory motion
with respect to a set of reference axes constructed at the equilibrium position as
illustrated in figure 5-4(c).
In order to model the above problem, the following assumptions are made:
(a) No motion exists in the horizontal direction.
(b) A downward displacement is considered as positive.
(c) The spring is a linear spring and obeys Hookes11 law which states that the
restoring force of the spring is proportional to the spring displacement.

Using Hookes law the spring force holding the weight in equilibrium can be
calculated. In figure 5-4(b), there is no motion because the weight W acting down
is offset by the spring force acting upward. Let fs denote the spring force illustrated
in figure 5-4(d). Using Hookes law, the spring force fs is proportional to the dis-
placement s and is written fs = Ks, where K is the proportionality constant called
11
Robert Hooke (16351703) English physicist.
402
the spring constant. The graph of fs versus displacement s is therefore a straight
line with slope K .

Figure 5-4. Spring-mass system.

Summation of the forces in equilibrium is represented in figure 5-4(b) which illus-

trates the spring force equal to the weight acting down or fs = Ks0 = W . This
determines the spring constant K as
W
K= . (5.84)
s0

In figure 5-4(c), the spring force acting on the weight is given by fs = K(s0 + y) where
y is the displacement from the equilibrium position. The vibratory motion can be
describe by using Newtons second law that the sum of the forces acting on the mass
must equal the mass times acceleration. The motion of the weight is thus modeled
by summing the forces in the y direction and writing Newtons second law as
d2 y d2 y
m = W fs = W Ks0 Ky = Ky, or m + Ky = 0 (5.85)
dt2 dt2

or
d2 y
+ 2 y = 0, 2 = K/m (5.86)
dt2
 403

Here the substitution 2 = K/m or = K/m has been made to simply the
representation of the differential equation describing the motion of the spring-mass
system. The quantity is called the natural frequency of the undamped system.

Simple Harmonic Motion

Consider the spring illustrated which is stretched
a distance y. Hookes law states the for a linear
spring, the restoring force Fs is proportional to the
displacement y and one can write Fs = Ky , where K
is the proportionality constant called the spring constant. The negative sign indicat-
ing that the restoring force is in the opposite direction of the spring displacement.

The elastic potential energy of the spring is defined as follows. The work done
in stretching a spring a distance y is given by
W =(average f orce)(displacement)
 
1 1
W = K y (y) = K y 2
2 2

In stretching the spring using a force Ky, the spring exerts an opposite force Ky
which does negative work. This negative work is called the elastic potential energy
of the spring and it is denoted by

1
Ep = Ky 2
2
dy
Multiply equation (5.85) by dt to obtain
dt

dy d2 y dy
m 2
dt + K y dt = 0 (5.87)
dt dt dt

Note that the integration of each term in equation (5.87) is of the form u du for
an appropriate value of u. One can verify that an integration of equation (5.87)
produces the result that the kinetic energy plus spring potential energy is a constant
and represented   2
1 dy 1
m + Ky 2 = E (5.88)
2 dt 2
404
where E is a constant of integration. Observe that the terms in equation (5.88)
represent   2
1 1 dy
Ek = mv 2 = m = Kinetic energy of system
2 2 dt
1
Ep = K y 2 = Spring potential energy
2
E =Total energy of the system

Equation (5.88) can be integrated to obtain the displacement y = y(t) as a function

of time t. Write equation (5.88) as
 2
dy 2E K 2 2E K
y = 2 A2 y 2 , where A2 = and 2 =

= (5.89)
dt m m K m

Take the square root of both sides to obtain the differential equation
dy 
= A2 y 2
dt

This is a differential equation where the variables can be separated to obtain

dy
 = dt
A2 y 2

and then integrated by making the substitution

A cos d
y = A sin , dy = A cos d to obtain = d = dt
A cos

where another integration produces

y
= t + 0 or = sin1 = t + 0 or y = y(t) = A sin(t + 0 ) (5.90)
A

where 0 is a constant of integration. Any periodic motion y = y(t) represented by

either of the equations

y = A sin(t + 0 ) or y = A cos(t + 0 ) (5.91)

is said to be a simple harmonic motion with amplitude A. The period P associated

with the oscillation is the time taken to complete one oscillation. The period of
the motion described by equation (5.90) or (5.91) is P = 2
. The frequency of the
1
motion is f = P = 2 and represents the number of oscillation performed in one
second, where 1 unit cycle per second is called a Hertz. The angle 0 in equation
(5.91) is called the phase angle or phase shift associated with the oscillations.
 405
Note that equations (5.91) differ only by a phase constant, since one can write

A sin(t + 0 ) =A cos(t + 0 /2) = A cos(t + 0 ), 0 = 0 /2

A cos(t + 0 ) =A sin(t + 0 + /2) = A sin(t + 0 ), 0 = 0 + /2

In general, given an equation of the form

y = y(t) = A1 cos t + A2 sin t


one can multiply both the numerator and denominator by A21 + A22 to obtain
 
2  A1 A2

2
y = y(t) = A1 + A2 cos t +  2 sin t
A21 + A22 A1 + A22

so that the oscillatory motion can be expressed in the following form.

y =y(t) = A [sin 0 cos t + cos 0 sin t]

y =y(t) = A sin(t + 0 )

d2 y
Simple harmonic motion can be characterized by observing that the acceleration
dt2
satisfies the conditions
d2 y
(i) It is always proportional to its distance from a fixed point 2 = 2 y .
dx
(ii) It is always directed toward the fixed point.

Damping Forces
Observe the sign of the spring force in equation (5.85). If y > 0, the restoring force
is in the negative direction. If y < 0, the restoring force is in the positive direction.
The directions of the forces are important because forces are vector quantities and
must have both a magnitude and a direction. The direction of the forces is one check
that the problem is correctly modeled.
If additional forces are added to the spring mass system, such as damping forces
and external forces, then equation (5.85) must be modified to include these addi-
tional forces. In figure 5-5, assume a damper and an external force are attached to
the spring as illustrated.
406

Figure 5-5. Spring mass system with additional forces.

If there is a damping force FD which opposes the motion of the mass and the
magnitude of the damping force is proportional to the velocity12, then this can be
represented by
dy
FD = (5.92)
dt
where > 0 is the proportionality constant called the damping coefficient. The sign
dy
of the damping force is determined by the sign of the derivative . Note that if y
dt
is increasing and dy
dt
> 0 , the damping force is in the negative direction, whereas, if y
is decreasing and dy
dt
< 0, the damping force acts in the positive direction. In figure
5-5, the quantity F (t) denotes an external force applied to drive the mass.
The use of Newtons second law of motion, together with the summation of
forces, one can construct a mathematical model describing the motion of the spring
mass system with damping and external force. The illustration in the figure 5-5 can
be used as an aid to understanding the following equation
d2 y dy
m 2
= Ky + F (t) (5.93)
dt dt
d d2
or my + y + Ky = F (t) = , = (5.94)
dt dt2

where the right-hand side of equation (5.93) represents a summation of the forces
acting on the spring-mass system. Each term in equation (5.94) represents a force
12
Various assumptions can be made to model other types of damping.
 407
term and for the equation to be dimensionally homogeneous, every term must have
dimensions of force. The quantity my is called the inertial force, y is the damping
force, Ky is the spring force and F (t) is an external force. Here m, and K are all
positive constants with dimensions [m] = [W ] lbs lbs lbs
[g] = ft/sec2 , [] = ft/sec , [K] = ft with
y and t having the dimensions [y] = ft and [t] = seconds. It is left as an exercise to
verify that the equation (5.94) is dimensionally homogeneous.13
To solve the differential equation (5.94) one first solves the homogeneous equa-
tion
d2 y dy
m + + Ky = 0 (5.95)
dt2 dt
by finding a set of two independent solutions {y1 (t), y2 (t)} called a fundamental set
of solutions to the homogeneous differential equation. The general solution to the
homogeneous differential equation is then any linear combination of the solutions
from the fundamental set. The general solution to equation (5.95) can be expressed

yc = c1 y1 (t) + c2 y2 (t) where c1 , c2 are arbitrary constants. (5.96)

This general solution is called the complementary solution and is usually denoted
using the notation yc . Any solution of the nonhomogeneous differential equation
(5.94) is denoted using the notation yp and is called a particular solution. The general
solution to the differential equation (5.94) can then be expressed as y = yc + yp .
If the homogeneous differential equation has constant coefficients, one can assume
an exponential solution y = exp(t) = et , constant, to obtain the fundamental set
of solutions.

d2 y dy
Example 5-19. Solve the differential equation 2 + 3 + 2y = 2e3t
dt dt
Solution Assume an exponential solution y = et to the homogeneous differential
equation
d2 y dy
2
+ 3 + 2y = 0 (5.97)
dt dt
2
and substitute y = et , dy
dt
= et , ddt2y = 2 et into the homogeneous differential equa-
tion (5.97) to obtain the algebraic equation

2 + 3 + 2 = ( + 2)( + 1) = 0 (5.98)
13
The word homogeneous is used quite frequently in the study of differential equations and its meaning depends
upon the context in which it is used.
408
called the characteristic equation. The roots of this equation = 2 and = 1
are called the characteristic roots. Substituting these characteristic roots into the
assumed solution produces the fundamental set {e2t , et }. The complementary solu-
tion is then a linear combination of the functions in the fundamental set. This gives
the complementary solution
yc = c1 e2t + c2 et (5.99)

where c1 , c2 are arbitrary constants.

The right-hand side of the given nonhomogeneous equation is an exponen-
tial function and since it is known that derivatives of exponential functions give
exponential functions, one can assume that a particular solution yp must have
the form yp = Ae3t where A is a constant to be determined. The processes of
examining the derivatives of the right-hand side of the nonhomogeneous equa-
tion and forming a linear combination of the basic terms associated with the
right-hand side function and all its derivatives, is called the method of undeter-
mined coefficients for obtaining a particular solution. Substituting the functions
dy d2 y
yp = Ae3t , dtp = 3Ae3t , dt2p = 9Ae3t into the given nonhomogeneous differential
equation one finds
9Ae3t 9Ae3t + 2Ae3t = 2e3t (5.100)

Simplify equation (5.100) and solving for the constant A one finds A = 1, so that
yp = e3t is a particular solution. The general solution is then given by

y = yc + yp = c1 e2t + c2 et + e3t (5.101)

Example 5-20. (Representation of solution)

Obtain a general solution to the linear homogeneous differential equation
d2 y
+ 2 y = 0, is a constant (5.159)
dt2
which represents simple harmonic motion.
Solution When dealing with homogeneous linear differential equations with constant
coefficients one should assume an exponential solution y = et having derivatives
dy d2 y
= et and 2 = 2 et . Substitute the assumed exponential solution into the
dt dt
given differential equation to obtain the characteristic equation

2 et + 2 et = 0 = 2 + 2 = ( i )( + i ) = 0 (5.104)
 409
The characteristic roots are the complex numbers = i and = i , where i is an
imaginary unit satisfying i2 = 1. These characteristic roots are substituted back
into the assumed exponential solution to produce the fundamental set of solutions

{ ei t , ei t } (5.104)

Observe that any linear combination of the solutions from the fundamental set is also
a solution of the differential equation (5.159), so that one can express the general
solution to the homogeneous differential equation as14

y = c1 ei t + c2 ei t (5.105)

where c1 , c2 are arbitrary constants. Use Eulers identity ei = cos + i sin and
consider the following special cases of equation (5.105).
(i) If c1 = 12 and c2 = 12 , the general solution becomes the real solution
1 i t 1 i t
y1 = y1 (t) = e + e = cos t
2 2
1 1
(ii) If c1 = 2i
and c2 = 2i
, the general solution becomes the real solution
1 i t 1
y2 = y2 (t) = e ei t = sin t
2i 2i

The functions cos t and sin t are real linearly independent solutions to the
differential equation (5.159) and consequently one can state that the set of solutions

{ cos t, sin t } (5.106)

is a fundamental set of solutions to the equation (5.159), and

y = k1 cos t + k2 sin t (5.107)

is a general solution15 to the given differential equation (5.159), where k1 and k2 are
arbitrary constants.

Multiply and divide equation (5.107) by k12 + k22 to obtain
 
k1 k2

y= k12 + k22  cos t +  sin t (5.108)
k12 + k22 k12 + k22

14
Electrical engineers prefer to use this form for the solution.
15
Mechanical engineers prefer this form for the solution.
410

The substitutions
k1 k2

A= k12 + k22 , sin 0 =  , cos 0 = 
k12 + k22 k12 + k22

allow one to express the general solution in the form

y = A sin(t + 0 )

The substitutions
k2 k1

A= k12 + k22 , sin 0 =  , cos 0 = 
k12 + k22 k12 + k22

allows one to express the general solution in the form

y = A cos(t 0 )

This example illustrates that one has many options available in representing the
form for the general solution to a linear homogeneous differential equation with
constant coefficients. The resulting form is closely associated with the selection of
the two independent functions which make up the fundamental set of solutions.

Example 5-21. (Representation of solution)

Solve the linear homogeneous differential equation
d2 y
2 y = 0, is a constant (5.109)
dt2

Solution The given differential equation is a linear homogeneous second order dif-
ferential equation with constant coefficients and so one can assume an exponential
dy d2 y
solution of the form y = y(t) = e t which has the derivatives = e t and 2 = 2 e t .
dt dt
Substitute the assumed solution and its derivatives into the above differential equa-
tion to obtain the characteristic equation

2 e t 2 e t = 0 = 2 2 = ( )( + ) = 0 (5.110)

giving the characteristic roots = and = from which one can construct the
fundamental set of solutions
{ e t , e t } (5.111)
 411
A general solution is then any linear combination of the functions in the fundamental
set and so can be represented in the form

y = y(t) = c1 e t + c2 e t (5.112)

where c1 , c2 are arbitrary constants. A special case of equation (5.112) occurs when
c1 = 12 and c2 = 12 and one finds the special solution
1 t 1 t
y1 = y1 (t) = e + e = cosh t (5.113)
2 2
1
The special case where c1 = 2
and c2 = 12 produces the solution
1 t 1 t
y2 = y2 (t) = e e = sinh t (5.114)
2 2
The functions cosh t and sinh t are linearly independent solutions to the differential
equation (5.109) and therefore one can construct the fundamental set of solutions

{ cosh t, sinh t } (5.115)

and from this fundamental set one can construct the general solution in the form

y = y(t) = k1 cosh t + k2 sinh t (5.116)

where k1 and k2 are arbitrary constants.

If one selects the constants k1 and k2 such that

k1 = A cosh t0 and k2 = A sinh t0 t0 is a constant

then the general solution can be expressed in the form

y = y(t) = A ( cosh t cosh t0 sinh t sinh t0 ) = A cosh (t t0 )

Alternatively, one can select the constants

k1 = A sinh t0 and k2 = A cosh t0

and express the general solution in the alternative form

y = y(t) = A ( sinh t cosh t0 cosh t sinh t0 ) = A sinh (t t0 )

This is another example, where the form selected for the fundamental set of
solutions can lead to representing the general solution to the differential equation
in a variety of forms. In selecting a particular form for representing the solution
one should select a form where the representation of the solution and any required
auxiliary conditions are easily handled.
412
Mechanical Resonance
In equation (5.94), let F (t) = F0 cos t, with is a constant, and then construct
the general solution to equation (5.94) for this special case. To solve

L(y) = my + y + Ky = F0 cos t (5.117)

it is customary to first solve the homogeneous equation

L(y) = my + y + Ky = 0. (5.118)

This is an ordinary differential equation with constant coefficients and this type of
equation can be solved by assuming an exponential solution y = exp(t) = et . Sub-
stituting this assumed value for y into the differential equation (5.118) one obtains
an equation for determining the constant(s) . This resulting equation is called the
characteristic equation associated with the homogeneous differential equation and
the roots of this equation are called the characteristic roots. One finds the charac-
teristic equation
m 2 + + K = 0

with characteristic roots

  2
2 4mK K
= = . (5.119)
2m 2m 2m m

(i) If the characteristic roots are denoted by 1 , 2 and these roots are distinct,
then the set of solutions {e1 t , e2 t } is called a fundamental set of solutions to the
homogeneous differential equation and the general solution is denoted by the
linear combination

y = c1 e1 t + c2 e2 t , c1 , c2 are arbitrary constants (5.120)

(ii) If the characteristic roots are equal and 1 = 2 , then one member of the funda-
mental set is e1 t . It has been found that each time a characteristic root repeats
itself, then one must multiply the first solution by t. This rule gives the sec-
ond member of the fundamental set as te1 t The fundamental set is then given
by {e1 t , te1 t } and produces the general solution as a linear combination of the
solutions in the fundamental set. The general solution can be written

y = c1 e1 t + c2 te1 t , c1 , c2 are arbitrary constants. (5.121)

 413
(iii) If the characteristic roots are imaginary and of the form 1 = +i and 2 = i ,
then one can use the Euler formula eix = cos x + i sin x to express the general
solution in either of the forms
y =c1 e(+i )t + c2 e(i )t , complex form for the solution
(5.122)
y =et (C1 cos t + C2 sin t) , real form for the solution

where c1 , c2, C1 , C2 represent arbitrary constants.

The general solution to the homogeneous differential equation (5.118) is called
the complementary solution. In equation (5.119), the discriminant (/2m)2 K/m > 0,
determines the type of motion that results. The following cases are considered.
CASE I (Homogeneous Equation and Overdamping)
If (/2m)2 K/m > 0, equation (5.119) has two distinct roots 1 and 2 where both
1 and 2 are negative, then the corresponding complementary solution of equation
(5.118) has transient terms e1 t and e2 t and the general solution is of the form

yc = c1 e1 t + c2 e2 t , 1 < 0, 2 < 0 (5.123)

where c1 , c2 are arbitrary constants. This type of solution illustrates that if the
damping constant is too large, then no oscillatory motion can exist. In such a
situation, the system is said to be overdamped.
CASE II (Homogeneous equation and underdamping)
For the condition (/2m)2 K/m < 0, let 02 = K/m (/2m)2 and obtain from the
characteristic equation (5.119) the two complex characteristic roots

1 = /2m + i0 and 2 = /2m i0 .

These characteristic roots produce a complementary solution of the form

yc = et/2m (c1 sin 0 t + c2 cos 0 t)

or 
yc = c21 + c22 et/2m cos(0 t ) (5.124)

with c1 , c2 arbitrary constants. Here 0 represents the damped natural frequency of

the system. If is small, 0 is approximately the natural frequency of the undamped

system given by = K/m.
414
The solution equation (5.124) denotes a
damped oscillatory solution which can be vi-
sualized by plotting the curves

y1 = c21 + c22 et/2m and y2 = y1

as envelopes of the oscillation cos(0 t ) as

is illustrated in figure 5-6. The term /2m
is called the damping constant or damping
factor. Figure 5-6. Damped oscillations.

CASE III (Homogeneous equation and critical damping) If (/2m)2 K/m = 0, equation
(5.119) has the repeated roots 1 = 2 = /2m which produces the solution

yc = (c1 + c2 t)et/2m . (5.125)

By reducing the damping constant one gets to a point where oscillations begin
to occur. The motion is then said to be critically damped. The critical value for
the damping constant in this case is denoted by c and is determined by setting

the discriminant equal to zero to obtain c = 2m where = K/m is the natural
frequency of the undamped system.
Particular Solution
Associated with the complementary solution from one of the cases I, II, or III,
is the particular solution of the nonhomogeneous equation (5.117). The particular
solution can be determined by the method of undetermined coefficients. Examine the
function(s) on the right-hand side of the differential equation and all the derivatives
associated with these function(s). Select the basic terms which keep occurring in the
function and all of its derivatives and form a linear combination of these basic terms.
For the equation (5.117) the basic terms which occur by continued differentiation
of the right-hand side are the functions {cos t, sin t} multiplied by some constant.
One can then assume that the particular solution is of the form

yp = A cos t + B sin t (5.126)

with A and B unknown constants to be determined. Substituting this assumed par-

ticular solution into the differential equation (5.117) produces the equation

[(K m2 )A + B] cos t + [A + (K m2 )B] sin t = F0 cos t (5.127)

 415
Now equate the coefficients of like terms to obtain the system of equations
(K m2 )A + B = F0
(5.128)
A + (K m2 )B = 0,

which are equations used to determine the constants A and B . Solving equations
(5.128) gives
(K m2 )F0 F0
A= and B = (5.129)

where
= (K m2 )2 + 2 2 .

The particular solution16 can then be expressed as

(K m2 )F0 F0 F0
yp = cos t + sin t = cos(t ) (5.130)

where is a phase angle defined by tan = /(K m2 ) for m2
= K. The general

solution to equation (5.119) can then be written y = yc + yp . In the general solution,
the complementary solutions are transient solutions, and the particular solution rep-
resents the steady state oscillations. The amplitude of the steady state oscillations
is given by
F F0 F0
Amp = 0 =  2 2 2 2
=  (5.131)
(K m ) + m ( ) + 42 2 (/c )2
2 2 2

where is the natural frequency of the undamped system and c = 2m is the critical
value of the damping. For = 0 (no damping), the denominator in equation (5.131)
becomes m|2 2| and approaches zero as tends toward . Thus, with no damping,
as the angular frequency of the forcing term approaches the natural frequency
of the system, the denominator in equation (5.119) approaches zero, which in turn
causes the amplitude of the oscillations to increase without bound. This is known
as the phenomenon of resonance. For
= 0, there can still be a resonance-type
behavior whereby the amplitude of the oscillations become large for some specific
value of the forcing frequency .
Define the resonance frequency as the value of which produces the maximum
amplitude of the oscillation, if an oscillation exists.
16 
A B

Recall that A cos t + B sin t = A2 + B 2 cos t + sin t = A2 + B 2 cos(t )
A2 +B2 A2 +B2
416

Figure 5-7. Amplitude versus frequency for a forced system.

This amplitude, given by equation (5.131), has a maximum value when the
denominator is a minimum. Let

H = ( 2 2 )2 + 42 2 (/c )2

denote this denominator. The quantity H has a minimum value with respect to
when the derivative of H with respect to is zero. Calculating this derivative gives
dH
= 2( 2 2 )(2) + 8 2 (/c )2 = 0
d

when
2 = 2 1 2(/c )2 .

(5.132)

The phenomenon of resonance is illustrated graphically in figure 5-7 by plotting the

amplitude, as given by equation (5.131), versus for various values of the ratio /c .

In practical problems, it is important to be able to design vibratory structures to

avoid resonance. For example, printing presses vibrating at the correct frequency can
act as forcing functions to cause large vibrations and even collapse of the supporting
floor. High winds can act as forcing functions to cause resonance oscillations of
structures. Flutter of aircraft wings, if not controlled properly, can result in dynamic
instability of an aircraft.
 417
Resonance can also be a desired phenomenon such as in tuning an electrical
circuit for a maximum response of a voltage of a specified frequency. In the study
of electrical circuits where frequency is a variable, it is desirable to have frequency
response characteristics for the circuit in a graphical form similar to figure 5-7.
Train yourself to look for curves which have shapes similar to those of the curves
in figure 5-7. Chances are some kind of resonance phenomenon is taking place.
Curves similar to the curves of figure 5-7 usually result from vibration models used
to study a wide variety of subjects and are the design basis of a large number of
measuring devices. The following is a brief list of subject areas where it is possible to
find additional applications of the basic equations of vibratory phenomena and res-
onance. Examine topics listed under mechanical vibrations, earthquake modeling,
atomic vibrations, atomic cross sections, vibrations of atoms in crystals, scattering
of atoms, particles, and waves from crystal surfaces, sound waves, string instru-
ments, tidal motions, lasers, electron spin resonance, nuclear magnetic resonance,
and behavior of viscoelastic materials.
Torsional Vibrations
Torsional vibrations are similar in form to the spring mass system and the dif-
ferential equation of the motion can be obtained from the example 5-15 presented
earlier. From this example the relation

d2
Torques = M = I = I 2 (5.133)
dt
is employed from equation (5.52) where denotes the angular displacement, I is the
d2
moment of inertia of the body, and = = 2 is the angular acceleration. Consider
dt
a disk attached to a fixed rod as in figure 5-8.
If the disk is rotated through an angle ,
there is a restoring moment M produced by
the rod. Hookes law states that the restor-
ing moment is proportional to the angular
displacement and

M = KT (5.134)

where KT is called the spring constant of the

Figure 5-8. Torsional vibrations.
shaft and is the proportionality constant as-
sociated with the angular displacement .
418
From the relation in equation (5.133) there results

d2 d2
M = KT = I or I + KT = 0 (5.135)
dt2 dt2

as the equation of motion describing the angular displacement .

By adding a linear damper and external force to equation (5.135), a more general
equation results
d2 d
I 2
+ + KT = F (t). (5.136)
dt dt
From strength of materials, the constant KT is given by the relation KT = GJ/L,
where G is called the shearing modulus of the rod material, J is the polar moment
of inertia of the rod cross section, and L is the length of the shaft.
The simple pendulum
For the pendulum illustrated the forces about 0 are the weight
of the mass and the radial force along the string. The radial
force passes through the origin and so does not produce a mo-
ment about the origin. The moment of inertia of the mass m
about 0 is given by I = m2 and the torque about 0 is given by
T = (mg)( sin )and consequently the equation of motion can
d2
be expressed T = mg sin = m2 2
dt
Simplification reduces the equation of motion to the form
d2 g
+ 2 sin = 0, 2 =
dt2 

For small oscillations one can make the approximation sin and write the equation
for the oscillating pendulum in the form
d2
+ 2 = 0
dt2

which is the equation of a simple harmonic oscillator.

Electrical Circuits
The basic elements needed to study electrical circuits are as follows:

(a) Resistance R is denoted by the symbol

The dimension of resistance is ohms17 and written [R] = ohms.
17
George Simon Ohm (17871854), German physicist.
 419

(b) Inductance L is denoted by the symbol

The dimension of inductance is henries18 and written [L] = henries.

(c) Capacitance C denoted by the symbol

The dimension of capacitance is farads19 and written [C] = farads.

(d) Electromotive force (emf) E or V denoted by the symbols or

The dimensions of electromotive force is volts20 and written [E] = [V ] = volts.

(e) Current I is a function of time, denoted I = I(t), with dimensions of amperes21

and written [I] = amperes.
(f) Charge Q on the capacitance is a function of time and written Q = Q(t),
with dimensions [Q] = coulombs.
The basic laws associated with electrical circuits are as follows: The current is
the time rate of change of charge. This can be represented with the above notation
as
dQ dQ
I= with [I] = amperes, [ ] = coulombs/second (5.137)
dt dt

Figure 5-9. Figure 5-10.

Voltage drop VR across a resistor. Voltage drop VL across inductor.

The voltage drop VR across a resistance, see figure 5-9, is proportional to the
current through the resistance. This is known as Ohms law. The proportionality
18
Joseph Henry (17971878), American physicist.
19
Michael Faraday (1791 1867) English physicist.
20
Alessandro Volta (17451827) Italian scientist.
21
Andre Marie Ampere (17751836) French physicist.
420
constant is called the resistance R. In symbols this can be represented as VR = RI
where
[VR ] = volts = [R][I] = (ohm) (ampere) (5.138)

The voltage drop VL across an inductance, see figure 5-10, is proportional to the
time rate of change of current through the inductance. The proportionality constant
is called the inductance L. In symbols this can be represented as
dI dI
VL = L where [VL] = volts = [L] [ ] = (henry)(ampere/second) (5.139)
dt dt

The voltage drop VC across a capacitance, see

figure 5-11, is proportional to the charge Q of
the capacitance. The proportionality constant
is denoted 1/C. In symbols this is represented
as
Q
VC = (5.140)
C
Figure 5-11.

1
Voltage drop VC across capacitor. where [VC ] = volts = [Q] = coulombs/farad
C
The Kirchoff laws for an electric circuit are.
Kirchhoff s22 first law :
The sum of the voltage drops around a closed circuit must equal zero.
Kirchhoff s second law:
The amount of current into a junction must equal the current leaving the junc-
tion.
The place in an electrical circuit where two or more circuit elements are joined
together is called a junction. A closed circuit or loop occurs whenever a path con-
structed through connected elements within a circuit closes upon itself. Voltage
drops are selected as positive, whereas voltage rises are selected as negative.

Example 5-22. For the RC-circuit illustrated in figure 5-12, set up the differ-
ential equation describing the rate of change of the charge Q on the capacitor. Make
the assumption that Q(0) = 0.
22
Gustav Robert Kirchhoff (19241887) German physicist.
 421

Figure 5-12. An RC-series circuit.

Solution For a path around the circuit illustrated in figure 5-12, the Kirchhoffs
voltage law would be written
VR + VC E = 0.
dQ
Let I = I(t) = dt denote the current in the circuit at any time t. By Kirchhoffs first
law:      
V oltage drop V oltage drop Applied
across R
+ across C
= emf
VR + VC = E
Q
RI + C
= E.
This gives the differential equation
dQ 1
L(Q) = R + Q=E (5.141)
dt C

where R, C and E are constants. The solution of the homogeneous differential equa-
tion
dQ 1
R + Q=0
dt C
can be determined by separating the variables and integrating to obtain
 
dQ 1 dQ 1 t
= dt and = dt = ln Q = +
Q RC Q RC RC

where is a constant of integration. Solving for Q one finds is Qc = c1 exp(t/RC)

where c1 = e is just some new constant. Since the right-hand side of the nonho-
mogeneous differential equation is a constant, one can assume a particular solution
of the form Qp = c2 E where c2 is a constant to be determined. Substituting this
assumed solution into the nonhomogeneous differential equation and solving for c2
one finds c2 = C and so the particular solution can be written Qp = CE. The general
422
solution of equation (5.141) is represented by the sum Q = Qc + Qp and the solution
satisfying Q(0) = 0 is given by

Q = Q(t) = EC(1 et/RC ) (5.142)

The relation (5.142) is employed to determine the current I and voltages VC and VR
as
dQ E
I = I(t) = = et/RC
dt R
Q (5.143)
VC = = E(1 et/RC )
C
VR = RI = E et/RC

In equations (5.142) and (5.143) the term exp(t/RC) is called a transient term
and the constant = RC is called the time constant for the circuit. In general,
terms of the form exp (t/) are transient terms, and such terms are short lived and
quickly or slowly decay, depending upon the magnitude of the time constant = .
The following table illustrates values of exp(t/) for t equal to various values of the
time constant.
Time t exp(t/)
0.3679
2 0.1353
3 0.0498
4 0.0183
5 0.0067

The values in the above table gives us valuable information concerning equations
such as (5.142) and (5.143). The table shows that decaying exponential terms are es-
sentially zero after five time constants. This is because the values of the exponential
terms are less than 1 percent of their initial values.
Solutions to circuit problems are usually divided into two parts, called transient
terms and steady state terms. Transient terms eventually decay and disappear and
do not contribute to the solution after about 5 time constants. The steady state
terms are the part of the solution which remains after the transient terms become
negligible.
 423
Example 5-23. For the parallel circuit illustrated in figure 5-13, apply Kirch-
hoffs first law to each of the three closed circuits.

Figure 5-13. A parallel circuit.

Note that each closed circuit has the same voltage drop. This produces the
following equations.
E = RI1
dI2
E=L (5.144)
dt

1
E= I3 dt.
C
Kirchhoffs second law applied to the given circuit tells us

I = I1 + I2 + I3 . (5.145)

If the impressed current I is given, the above four equations can be reduced to one
ordinary differential equation from which the impressed voltage E can be found.
Write equation (5.145) in the form

E 1 dE
I= + E dt + C .
R L dt

called a differentialintegral equation. By differentiation of this equation there re-

sults an ordinary linear second-order differential equation
dI 1 dE E d2 E
= + +C 2 ,
dt R dt L dt

where E is the dependent variable to be determined.

Conversely, if E is given and I is unknown, then equations (5.144) give us I1 , I2 ,
and I3 , and equation (5.145) can be used to determine the current I.
424
Thermodynamics
Experiments on a fixed mass of gas has established the following gas laws which
relate the pressure P , volume V and absolute temperature T .
Boyles Law If T is held constant, then pV = constant.
V
Charless Law If P is held constant, then = constant.
T
P
Gay-Lussac Law If V is held constant, then = constant.
T
These laws are summarized using the gas equation
P1 V1 P2 V2
=
T1 T2

where pressure P can be measured in units [N/m2 ], volume V can be measured in

units [m3 ] and absolute temperature T is measured in units [K].
The ideal gas absolute temperature is defined using Boyles law which states
P V T which produces the equation of state for an ideal gas, which can be expressed
in the form
P V =nRT


N 3 J
[ 2 ][m ] =[mol] [K]
m mol K
where n is the number of moles of gas and R = 8.314472 molJ K is the ideal gas constant


or universal molar gas constant. Note that real gases may or may not obey the ideal
gas law. For gases which are imperfect, there are many other proposed equations of
state. Some of these proposed equations are valid over selected ranges and conditions
and can be found under such names as Van der Waals equation, Berthelot equation,
Dieterici equation, Beattie-Bridgeman equation, Virial equation.
The zeroth law of thermodynamics states that if two bodies are in thermal equi-
librium with a third body, then the two bodies must be in thermal equilibrium with
each other. The zeroth law is used to develop the concept of temperature. Here
thermodynamic equilibrium infers that the system is (i) in chemical equilibrium and
(ii) there are no pressure or temperature gradients which would cause the system
to change with time. The first law of thermodynamics is an energy conservation
principle which can be expressed dQ = dU + dW where dQ is the heat supplied to
a gas, dU is the change in internal energy of the gas and dW is the external work
done. The second law of thermodynamics examines processes that can happen in
an isolated system and states that the only processes which can occur are those for
which the entropy either increases or remains constant. Here entropy S is related
 425
to the ability or inability of a systems energy to do work. The change in entropy is
defined as dS = dQ/T where dQ is the heat absorbed in an isothermal and reversible
process and T denotes the absolute temperature.
Recall that the ability of gases to change when subjected to pressure and tem-
perature variations can be described by the equation of state of an ideal gas

P V = nRT, (5.146)

where P is the pressure [N/m2 ], V is the volume [m3 ], n is the amount of gas [moles], R
is the universal gas constant [J/mol K], and T is the temperature [K]. For an ideal
gas, the gas constant R can also be expressed in terms of the specific heat at constant
pressure Cp, [J/mol K] and the specific heat at constant volume Cv , [J/mol K] by
Mayers equation R = Cp Cv . Equation (5.146) is illustrated in the pressure-volume
diagram of figure 5-14.
The curves where T is a constant are called isothermal curves and are the hy-
perbolas labeled (b) and (c) illustrated in figure 5-14. These curves correspond to
the temperature values T1 and T2 . When a gas undergoes changes of state it can do
so by an isobaric process (P is a constant) illustrated by line (a) in figure 5-14, an
isovolumetric process (V is a constant) illustrated by the line (e) in figure 5-14, an
isothermal process (T is a constant) illustrated by the hyperbolas with T = T1 and
T = T2 in figure 5-14, or an adiabatic process (no heat is transferred) represented by
the curve (d) in figure 5-14.

Figure 5-14. Pressure-Volume diagram.

426
The first law of thermodynamics states that when a gas undergoes a change,
the equation dU = dQ + dW must be satisfied, where dU is the change in internal
energy, dQ is the change in heat supplied to the gas, and dW is the work done. An
adiabatic process is one in which dQ = 0. For an adiabatic process, the first law
of thermodynamics requires dU = dW . The work done dW is related to the volume
change by the relation dW = P dV, and the change in internal energy is related to
the temperature change by the relation dU = CpdT . For an adiabatic process
dP dV
+ =0 (5.147)
P V

Integrate the equations (5.147) and show the adiabatic curve (d) in figure 5-14 can
be described by any of the equations
1
T V 1 = Constant, TP = Constant, or P V = Constant,

where = Cp/Cv is the ratio of the specific heat at constant pressure to the specific
heat at constant volume. Also note that during an adiabatic process dQ = 0 so that
the work done by the system undergoing a change in volume is given by the integral
of dW which is represented by the shaded area in the figure 5-14. This shaded area
is represented by the integral
 v2
work done = P dV
v1

Radioactive Decay
The periodic table of the chemical elements lists all 118 known chemical elements
using the notation A, where A represents a shorthand notation used to signify the
name of an element, is the atomic mass number or total number of protons and
neutrons in the nucleus of the element and is the atomic number or number of
protons in the nucleus of the element. Isotopes of an element all have the same
number of protons in the nucleus, but a different number of neutrons. For example,
carbon is denoted 12 13 14
6 C and the elements 6 C, 6 C are isotopes of carbon. Many of the

isotopes experience a process known as nuclear decay or radioactive decay, where an

isotope will emit some particles in a continuous way and lose some of its mass over
time.
Let A denote the quantity of a radioactive substance, measured in grams, with
the derivative dA
dt
denoting the rate of disintegration or amount of mass lost as a
 427
function of time t. In general, the amount of mass lost during radioactive decay is
proportional to the amount present and so can be represented by the mathematical
statement
dA
= kA (5.148)
dt
where k is a proportionality constant and the minus sign indicates mass being lost.
The proportionality constant k is referred to as the decay constant.
If A = A0 at time t = 0 one can separate the variables in equation (5.148) and
write
dA
= kdt (5.149)
A
Integrate both sides of equation (5.149) and show
 A  t
dA
= k dt (5.150)
A0 A 0

Here the limits of integration indicate that at time t = 0, A = A0 and at time t, then
A = A(t). After integrating equation (5.150) one obtains

A t
 
A
ln A = kt = ln = kt = A = A0 ekt (5.151)
A0 0 A0

If, by experiment, it is found that p percent of A0 disappears in T years, then

p

1 100 A0 is the amount remaining after T years and so this information can be
used to determine the decay constant k. At time T one has the equation
 p 
1 A0 = A0 ekT (5.152)
100
 p 
which implies ln 1 = kT and so one can solve for the decay constant k and
100
find
1  p 
k= ln 1 (5.153)
T 100
The half-life of a radioactive material is the time it takes for 50-percent of the
material to disappear. Consequently, if A = 21 A0 at time t = , the equation (5.151)
requires that  
1 1 1
A0 = A0 ek = ln = k = = ln(2) (5.154)
2 2 k
The table below gives the half-life of some selected elements from the periodic table.
428

Element Isotope Half-Life

Using the results from equa-
105
tion (5.154) one can express the Silver (Ag) Ag 41.29 days
111
radioactive decay curve given by Ag 7.45 days
107
equation (5.151) in the form Ag 44 seconds
197
 t/ Gold (Au) Au 7.4 seconds
A 1
= e(ln 2) t/ = (5.155) Iodine (I) 125
I 59 days
A0 2
123
I 13 hours
The figure 5-15 is a sketch of
131
A I 8 days
y = versus x = t/ . Examine
A0 137
this figure and note the values of Cesium (Ce) Ce 30 years
238
A for the values t = , 2, 3, 4, . . .. Uranium (U) U 4.46 (10)9 years
232
One can then construct the table Thorium (Th) Th 14.05 (10)9 years
of values illustrated.
The figure 5-15 illustrates that after a time of one half-life, then one-half of the
material is gone. After another time span of one half-life, half of the remaining
material is gone.

Figure 5-15. Radioactive decay curve A = A0 e(ln 2) t/

 429
The reduction in the amount of material by one-half is
t/ A/A0
illustrated by the scaling of the A and t axis of the radioac-
0 1
tive decay curve given by equation (5.155). If one uses the
1 1/2 axes A/A0 and t/ , then for each span of one half-life there
2 1/4 is a decrease in the amount of material by one-half.
3 1/8 Different radioactive substances are used for scientific
4 1/16 research in many disciplines. For example, archaeology uses
.. .. 14
C for dating of ancient artifacts.
. .

Carbon-12 is a stable element and its isotope carbon-14 is radioactive with a

half-life of 5730 years. All living organisms contain both 12 C and 14 C in known
ratios. However, after an organism dies, the carbon-12 amount remains the same
but the carbon-14 begins to decay. By measuring the proportions of 12 C and 14 C in
dead organisms one can estimate the elapsed time since death.
Geologist use 238 U, 206 P b, 232 T h, and 208 P b to determine the age of rocks. They
measure the relative amounts of these radioactive substances and compare ratios of
these amounts with rocks from an earlier age.
Radioactive substances are used for tracers and imaging in chemistry, biology
and medicine.

Economics
Suppose that it cost C = C() dollars to produce number of units of a certain
product. The function C(x) is called the cost function for production of x items. Let
r = r(x) denote the price received from the sale of 1 unit of the item and let P = P (x)
denote the profit from the sale of x items. This profit can be represented

P = P (x) =(number of items sold)(selling price of 1 unit) cost of production

P = P (x) =x r(x) C(x)

where the function x r(x) is called the revenue function.

As a first approximation for the representations of r(x) and C(x) one can assume
that they are linear functions of x and one can write

r = x and C(x) = a + bx

These assumptions have the following interpretations for , , a, b all constants.

430
(i) The minus sign in the representation for r indicates that an increase in the
selling price will cause a decrease in sales. This can be seen by plotting the
curve x = 1 r versus r which is a straight line with slope 1/ . This line tells
one that as r increases (price increases), then the number of sales x decreases.
(ii) The constant has to be large enough such that x remains positive as you
dont want to give away the product.
(iii) In the cost function, the constant a represents the overhead for the maintenance
of the production facilities and the variable term bx represent the additional cost
of production associated with producing x units. The units of measurements for
each term must be in dollars so [a] = $ and [b] = $/unit and [x] = number of units.
Hence, one can interpret b as the cost to produce 1 unit.

Using the above assumptions the profit from the sale of x items is given by

P = P (x) = x( x) (a + bx)

and if a profit is to be made from the sale of just one item, then it is required that
> + a + b. The derivative of the profit with respect to x is

dP
= x() + ( x) b
dx
dP b
The profit is a maximum when = 0 or x = is a critical point to be investi-
dx 2
d2 P
gated. The second derivative gives 2 = 2 < 0 indicating that the critical point
dx
produces a maximum value. These results are interpreted
b
(i) x = items should be produced for a maximum profit.
2
+b
(ii) The sale price for each item should be r = dollars per unit.
2
In economics the term R(x) = x r(x) is called the revenue function and its deriva-
tive dR
dx
is called the marginal revenue. The term P (x) is called the profit function
and its derivative dPdx
is called the marginal profit. The term C(x) is called the cost
function and its derivative dC dx
is called the marginal cost function.
By collecting data from production costs and sales over a period of time one can
construct better approximations for the price function and cost function and other
models similar to the above can be constructed and analyzed.
 431
Population Models
Mathematical modeling is used to study the growth and/or decay of a pop-
ulation. The population under study can be human populations subjected to a
spreading disease, insect populations which can affect crops, bacteria growth or cell
growth in the study of the spread of a disease or cancer cell growth. Predator-prey
models are used to study the advance and decline of populations based upon food
supplies. The effect of a certain type of medicine on the spread of bacteria or virus
growth is still another example of population changes which can be studied using
mathematics.
One begins by making some assumptions and starting with a simple model which
is easy to solve. By adding perturbations to the simple model it can be made more
complex and applicable to the type of problem one is trying to model. This type of
modeling has produced many extremely accurate results and the models predictive
capability has given much incite into the study of population growth or decay.
For example, an over simplified population growth model for say predicting
census changes is to let N denote the current population number and then assume
that the rate of change of a population is proportional to the number present. The
resulting model is represented
dN
= N
dt
Here > 0 is a proportionality constant. The conditions that at time t = 0 the
population is N0 can be used as an initial condition that the model must satisfy.
This model is simple and easy to solve. The variables can be separated and the
result integrated giving
 N  t N t
dN
= dt = ln N = t = N = N0 et
N0 N 0 N0 0

This result states that there is an exponential increase in the population with time.
One immediate method to modify the model is to investigate what happens if is
allowed to change with time. If = (t) then the above integrations become
 N  t N
 t t
dN (t) dt
= (t) dt = ln N = (t) dt = N = N0 e 0
N0 N 0 N0 0

Whenever the exponent of e gets too cumbersome it is sometimes convenient to

represent the solution in the form

 t 
N = N0 exp (t) dt
0
432
To try and make the census population model more accurate one can make
assumptions that include the rate of births and rate of deaths associated with the
current population. If one makes the assumptions that the birth rate is proportional
to N , say N and the death rate is proportional to N 2 , say N 2 , where and are
positive constants. The population model then has the form
dN
= N N 2 = ( N ) N (5.156)
dt

which states the rate of change of the population with time is determine by the birth
rate minus the death rate. Analyze this differential equation to see if it makes sense
by
(i) Determining conditions for when dN dt
> 0 which would indicate the population is
increasing.
(ii) Determining conditions for when dN dt
< 0 which would indicate the population is
decreasing.
(iii) Determining conditions for when dN dt
= 0 which would indicate no change in the
population.
Setting the equation (5.156) equal to zero, implies that N = N (t) is a constant,
since dN
dt
= 0. One finds the constant solutions N = N (t) = 0 and N = N (t) = / , are
constant solutions for all values of time t. These solutions are called steady-state
solutions and they do not change with time.
In order for dN
dt
> 0, one must require that N > 0 and ( N ) > 0 or / > N . In
order for dNdt
< 0, one must require that either N < 0 and ( N ) > 0 or N > 0 and
( N ) < 0 as these conditions would indicate the population was decreasing.
One can add additional assumptions such as (i) N is never zero and (ii) either
N0 < N < / for t > 0 producing an increasing population or (iii) N0 > N > /
producing a decreasing population. In either of the cases where dN dt
is different from
zero, one can separate the variables in equation (5.156) and write
dN
= dt
( N )N

An integration of this equation gives

 N  t
dN
= dt
N0 ( N )N 0
 433
To integrate the left-hand side of the above equation use partial fractions and show
the above integral reduces to
 N
  t
/ 1/
+ dN = dt
N0 N N 0

Scaling the integral properly, one can integrate this equation to obtain

 N    
 N  t  N   N0 
ln  = t = ln   ln 
 N0  = t

N  N0 0 N 

Solving for N gives the solution

N0 et N0
N = N (t) = t
= (5.157)
b N0 + N0 e N0 + ( N0 ) et

The equation (5.156) is called the logistic equation. The solution of this equation is
given by equation (5.157) which gives the limiting value t
lim N (t) = / . A graphical
representation of the logistic equation solutions are given in the figure 5-16.

Figure 5-16. Solutions to the logistic equation.

There are many more population models which are much more complicated than
the simple ones considered in this introduction.

Approximations
If the Greek letter epsilon  is positive and very small, then this is expressed by
writing 0 < || << 1. For very small  one can truncate certain Taylor series expan-
sions to obtain the following formulas to approximate f (x0 + ). These approximate
expansions are denoted using the symbol to represent approximation.
434
3
(1 + )n 1 + n  sin   a 1 +  ln a
3!
1
1 2 e 1 + 
1+ cos  1
1  2! 
1 3 ln(x + ) ln x +
1+ 2 tan   + x
3
It is left as an exercise to verify the above approximations.

Partial Differential Equations

Examples of partial differential equations can be found in just about all of the
scientific disciplines. For example, partial differential equations are employed to
describe such things as fluid motion, quantum mechanical interactions, diffusion
processes, wave motion and electric and magnetic phenomena. The following are
some examples of partial differential equations. The derivation of these well known
partial differential equations are presented in more advanced courses.
The vibrating string
Let u = u(x, t) denote the displacement of a
string at position x and time t, where the string
is stretched between the points (0, 0) and (L, 0).
The assumption that T , the tension in the string,
is much greater than the weight of the string,
produces the equation describing the vibrations of the string. The partial differential
equation describing the vibrations of the string is given by
2u 2
2 u Tg
= , u = u(x, t), 2 =
t2 x2

where T is the string tension, g the acceleration of gravity and is the weight per
unit length of string. This equation is called the one-dimensional wave equation and
is subject to boundary conditions u(0, t) = 0 and u(L, t) = 0 and initial conditions

u(x, 0) =f (x) = the initial shape of the string

u(x, 0)
and =g(x) = the initial velocity of the string
t
The above quantities have the following dimensions [x] = cm, [t] = s, [u] = cm,
[] = gm/cm, [T ] = dynes/cm and [g] = cm/sec2 .
The above one-dimensional wave equation is a special case of a more general
three-dimensional wave equation.
 435
One-dimensional heat flow
The partial differential equation describing the one-dimensional heat flow in a
rod along the x-axis is given by
2T T
k 2
= c , T = T (x, t)
x t

where T = T (x, t) is the temperature at position x and time t in an insulated rod of

length L, where k is called the thermal conductivity of the solid, is the volume
density of the solid and c is the specific heat of the solid. The above quantities have
the following dimensions [x] = cm, [t] = s, [T ] = C, [c] = cal/gm C , [] = gm/cm3 and
[k] = cal/sec cm2 C/cm.
The one-dimensional heat equation is a special case of the more general three-
dimensional diffusion equation.
Easy to Solve Partial Differential Equations
Partial differential equations of the form
u
= f (x, y)
x

can be integrated partially with respect to x to obtain

  
u
dx = f (x, y) dx = u(x, y) = f (x, y) dx + (y)
x

Here, y is held constant during the integration process and so the constant of inte-
gration can be any arbitrary function of y, represented here by (y).
Similarly, the partial differential equation
u
= g(x, y)
y

can be integrated partially with respect to y to obtain

  
u
dy = g(x, y) dy = u(x, y) = g(x, y) dy + (x)
y

Here x is held constant during the integration process and so any arbitrary func-
tion of x is considered as a constant of integration. This constant of integration is
represented by (x).
436
Partial differential equations of the form
2u
= h(x, y)
x y

can be integrated with respect to y to obtain


u
= h(x, y) dy + (x)
x

where (x) is an arbitrary function of x representing the constant of integration

during a partial integration with respect to y. One can then integrate with respect
to x and obtain the solution in the form
  
u(x, y) = h(x, y) dy dx + (x) dx + (y)

where (y) is the constant of integration associated with a partial integration with

respect to x. Note if (x) is arbitrary, then (x) dx is just some new arbitrary
function of x.
If you use partial differentiation to differentiate each of the above solutions,
holding the appropriate variables constant, you wind up with the integrand that
you started with. These partial differentiations are left as an exercise.

Example 5-24. Determine by integration the solution u = u(x, y) of the given

partial differential equations. Remember, that when dealing with functions of more
than one variable, you are going to be holding one of the variables constant during
a partial differentiation or partial integration.
u
(i) The solution to the partial differential equation x = 0 is
u = u(x, y) = (y) where (y) is an arbitrary function of y .
(ii) The solution to the partial differential equation uy = 0 is u = u(x, y) = (x) where
(x) is an arbitrary function of x.
u
(iii) The solution to the partial differential equation x = x + y is given by
2
u = u(x, y) = x2 + xy + (y) where (y) is an arbitrary function of y .
(iv) The solution to the partial differential equation u y
= x + y is given by
y2
u = u(x, y) = xy + 2 + (x) where (x) is an arbitrary function of x.
Note that if a variable is held constant during a partial integration, then an
arbitrary function of that variable can be considered as a constant of integration.
 437
Example 5-25. Show that for f, g arbitrary functions which are continuous
and differential, then the function u = u(x, t) = f (x ct) + g(x + ct) is a solution to the
one-dimensional wave equation

2u 2
2 u
= c , u = u(x, t), c is a constant
t2 x2

Solution Use the chain rule for differentiation and show

u
= f  (x ct)(c) + g  (x + ct)(c)
t
2u
= f  (x ct)(c2 ) + g  (x + ct)(c2 )
t2
u
= f  (x ct) + g  (x + ct)
x
2u
= f  (x ct) + g  (x + ct)
x2

Substitute the derivatives in the one-dimensional wave equation and obtain the
identity
c2 f  + c2 g  = c2 f  + c2 g 

Maximum and Minimum for Functions of Two Variables

Finding the maximum and minimum values associated with a function z = f (x, y),
which is defined and continuous over a domain D, is similar to what has been done for
functions of one variable. A function z = f (x, y) is said to have a relative maximum
value at a point (x0 , y0 ) D if f (x, y) f (x0 , y0 ) is satisfied for all points (x, y) in some
-neighborhood of the point (x0 , y0 ). Here a -neighborhood of the point (x0 , y0 ) is
defined at the set of points

N = { (x, y) | (x x0 )2 + (y y0 )2 2 } (5.158)

A function z = f (x, y) is said to have a relative minimum at a point (x0 , y0 ) if the

condition f (x, y) f (x0 , y0 ) is satisfied for all (x, y) in some -neighborhood of the
point (x0 , y0 ). If (x0 , y0 ) is a critical point to be tested for a relative maximum or
minimum point, then one can reduce the test to a study of one-dimensional problems.
Construct the three-dimensional curves

z = f (x0 , y) and z = f (x, y0 )

438
This is equivalent to letting the plane x = x0 cut the surface z = f (x, y) in the curve
z = f (x0 , y) and then letting the plane y = y0 cut the surface z = f (x, y) to produce
the curve z = f (x, y0 ). If the curve z = f (x0 , y) has a relative maximum or minimum
z
value, then y = 0. If the curve z = f (x, y0 ) has a relative maximum or minimum
z
value, then x = 0. Hence, a necessary condition that the point (x0 , y0 ) have relative
maximum or minimum value is for
z z
=0 and =0 simultaneously (5.159)
x y

Note that if one of the functions f (x0 , y) or f (x, y0 ) has a maximum at (x0 , y0 )
and the other function has a minimum at the point (x0 , y0 ), then the point (x0 , y0 )
is called a saddle point. A surface with saddle point is illustrated in the following
figure.

Figure 5-17. Surface in the shape of a saddle.

The study of maximum and minimum values are investigated in more detail in
the next volume.
 439
Exercises

5-1. Consider a spherical balloon at the instant when the radius of the balloon is
r0 [cm]. If air is entering the balloon at the rate of [cm3 /s], then at what rate is
the radius of the balloon changing at this instant?

5-2. Air expands adiabatically (no heat loss or gain) according to the gas law
pv 1.4 = constant, where p is the pressure [dyne/cm2 ] and v is the volume [cm3 ].
(a) If the volume is increasing at a rate [cm3 /s], then find the corresponding rate
of change in pressure.
(b) If the pressure is decreasing at a rate [dyne/cm2 s] then find the corresponding
rate of change in the volume.

5-3. A women who is 5.5 feet tall walks away from a street lamp, where the lamp
is 10 feet above the ground. She walks at a rate of 4 ft/s
(a) At what rate is her shadow changing when she is 4 feet from the lamp post?
(b) Is the length of shadow increasing or decreasing as she walks away from the
lamp?
(c) At what instant is the shadow 5.5 feet long?

5-4. For a thin lens in air, let x denote the distance of the object from the lens
and let y denote the distance of the image from the lens. The distances x and y are
1 1 1
related by the thin lens formula + = where f is a constant representing the
x y f
focal length of the lens.

(a) Show the thin lens formula can be written

in the Newtonian form S1 S2 = f 2 where
S1 = x f > 0 and S2 = y f > 0.
dx
(b) If x changes at a rate = r0 , then find a
dt
formula for the rate of change of y.

5-5. The sides of an equilateral triangle increase at the rate of r0 cm/hr. Find a
formula for the rate of change of the area of an equilateral triangle when the length
of a side is x0 cm.
440

5-6. A meteorologist, at a secret location, collects data and comes up with an
atmospheric pressure formula p = p0 e0h where [p] = lbs/ft2 , h has dimensions of feet
and represents the altitude above sea-level. In the atmospheric pressure formula the
quantities p0 and 0 are known constants.
(a) Find the dimensions of the constants p0 and 0 .
(b) If the meteorologist gets into a balloon which rises at a rate of 10 ft/s, then find
a formula representing the rate of change in the pressure when the altitude is h0
feet.

5-7. Given the parabola y y0 = (x x0 )2 where x0 , y0

are known constants.
(a) At the point (, ) on the curve a tangent line is
constructed. Find the equation of the tangent
line.
(b) The tangent line makes an angle with the x-
axis as illustrated. If changes at the rate of 1/2
cm/s, then at what rate does change?

5-8. Charles23 law, sometimes referred to as the law of volumes, states that at a
constant pressure the volume V of a gas and gas temperature T satisfy the relation
V
= C = constant, where V is the volume of the gas in cubic centimeters and T
T
is the absolute temperature in degrees Kelvin. If at a certain instant when V has
the volume V0 and T has the temperature T0 , it is know that the volume of gas is
dV
changing at the rate = r0 , then find how the temperature is changing.
dt

5-9. The Gay-Lussac24 law states that if the mass and volume of an ideal gas
are held constant, then the pressure of the gas varies directly with the gas absolute
temperature. If P denotes pressure measured in Pascals and T is the absolute temper-
P
ature in degrees Kelvin, then the Gay-Lussac law can be expressed = C = constant.
T
If at some instant when P has the value P0 and T has the value T0 , it is known that
dT
the temperature is change at the rate = r0 , then find how the pressure is changing
dt
at this instant.
23
Jacques Charles (1746-1823) French physicist and physical chemist as well as a balloonist.
24
Joseph Louis Gay-Lussac (1778-1850) A French chemist who studied the expansion of gases.
 441

5-10. A rock is thrown off a cliff so that after a time t its height above the ground
is h = h(t) = 200 16t2 .
(a) Find a formula representing the velocity of the rock.
(b) Find a formula representing the acceleration of the rock.
(c) What is the rocks velocity when it hits the ground?

5-11.

A spherical water tank has a radius of r = 12 feet. As-

sume h = h(t) is the depth of the water in the spherical
tank. The empty space above the water level inside
the tank defines a spherical cap whose volume is given
by Vcap = 3 (2r h)2 (r + h)
(a) Show the volume of water in the spherical tank is given by V = V (h) = 3 (3rh2 h3 )
(b) If water is entering the tank at 10 gallons per minute, then how fast is the water
level rising when h = 10 feet? Hint: Use 1 gallon =0.1336 cubic feet.

5-12. A spherical water tank has radius of r = 12 feet. Assume that h = h(t) is the
depth of the water in the spherical water tank and R = R(t) is the radius of the top
surface of the water. Find a relationship between dhdt
and dR
dt
.

5-13. A ball is shot from an air gun inclined at an angle with the horizontal.

The height of the ball as a function of time is given by y = y(t) = 16t2 + 50 3 t and
the horizontal distance traveled is given by x = x(t) = 50 t.
Answer the following questions.
(a) Find the maximum height of the ball.
(b) Find the time when the maximum height is achieved.
(c) Find the time when the ball hits the ground.
(d) Find the x position where the maximum height is achieved.
(e) Eliminate time t from x = x(t) and y = y(t) to obtain y as a function of x.

5-14. Empirical data obtained by shooting bullets into maple wood blocks pro-
duces the formula

v = v(x) = K 1 2x, 0 < x < 1/2, (K is a constant )
for the speed [ft/s] of the bullet after it has penetrated the wood a distance x feet.
Find the rate at which the speed of the bullet is decreasing after it enters the wood.
442

5-15. Use the results from table 5-1 to find the centroid of the given composite
shapes.

5-16. Find the centroid of the area bounded by the curves

(a) The parabola y = x2 and the line y = y0 > 0
(b) The parabola y = x2 and the lines x = x0 > 0 and y = 0
(c) The parabola y = x2 and the lines x = x0 > 0, x = x1 > x0 > 0 and y = 0

5-17. Find the centroid of the area bounded by the curves

(a) The parabola x = 6y y2 and the line x = 0
(b) The parabola x = 6y y2 and the line x = 1
(c) The parabola x = 6y y2 and the line 2y + x = 0

5-18. Find the centroid of the solid produced by rotation of the given area about
the axis specified.

Area defined by bounding curves Axis of rotation

(a) by hx = 0, x = b, y=0 x-axis
(b) by hx = 0, x = b, y=0 y -axis
(c) by hx = 0, x = b, y=0 The line y = 2
(d) y 2 = 4x, x = 0, y = y0 > 0 y -axis
(e) y 2 = 4x, x = 0, y = y0 > 0 x-axis
(f) y 2 = 4x, x = 0, y = y0 > 0 The line x = 1
b, h, y0 are all positive constants.

5-19. The curve y = 4 x2 , 0 x 2, is revolved about the y-axis to form a solid

of revolution. Find the centroid of this solid.
(a) Use disk shaped volume elements.
(b) Us cylindrical shell shaped volume elements.
 443

5-20. Newtons law of cooling states that the temperature T of a body cools with
time t at a rate proportional to the temperature difference T Tenv between the body
temperature T and the temperature of the environment Tenv . Newtons law of cooling
dT
can therefore be expressed by the differential equation = k(T Tenv ), where
dt
k > 0 is a proportionality constant and the negative sign indicates the temperature
is decreasing. (a) Use integration techniques to obtain the general solution to this
differential equation. (b) If the body initially has a temperature T = 100 C and is
cooling in an environment at 0 C , find T = T (t). (c) If the body cools to 80 C in
20 minutes, find the proportionality constant k. (d) Find the time it takes for the
body to cool from 90 C to 70 C . (e) Give units of measurement for each term in the
solution to part (a).

5-21. It has been found that under certain conditions, the number density N
3
(#/cm ) of a certain bacteria increases at a rate proportional to the amount N
present. If at time t = 0, N = N0 is the initial number of bacteria per cubic centimeter
and if after 5 hours, the value of N has been found to increase to 3N0 , then find the
equation representing N = N (t) as a function of time t. Give units of measurement
for all terms in your equation.

5-22. Pappuss Theorem

Pappuss25 theorem states that if a region

R is rotated about a line which does not pass
through the region, then the volume of the solid
of revolution equals the area of the region R mul-
tiplied by the distance traveled by the centroid
of the region R. Assume the region R illustrated
is rotated about the
x1
y -axis.
(a) Show the volume of the solid formed is V = 2 xh(x) dx
x0
 x1
1
(b) Show the x-position of the centroid is x = xh(x) dx, where A is the area of
A x0
the region R.
(c) Prove Pappuss theorem.
(d) Prove Pappuss theorem if the region R is rotated about the x-axis.
25
Pappus of Alexandria (290-350) A Greek geometer.
444

5-23. Use the Pappus theorem from the previous example
(a) to find the volume of a right circular cone obtained by revolving the line y = hr x,
0 x h about the x-axis
(b) to find the volume of a sphere obtained by revolving the semi-circle x2 + y2 = r2
r x r , y > 0 about the x-axis.

5-24. Solve the given differential equations using integration techniques.

dy dy dy
(a) =0 (d) = ex (g) = y2
dx dx dx
dy dy dy
(b) =1 (e) =y (h) = sin x
dx dx dx
dy d2 y dy dy
(c) =x (f ) 2
= (i) = cos x
dx dx dx dx

5-25. Solve each of the given differential equations by separating the variables
and applying integration techniques.
dy 1+x dy 1 + x2 dy 1 + x3
(a) = (b) = (c) =
dx 1+y dx 1 + y2 dx 1 + y3

5-26.
d2 y
(a) Solve the differential equation 2 + 2 y = cos t, where and are constants.
dt
(b) For what value does resonance occur?

5-27.
Use a plane to cut the regular pyramid with height h and
square base having sides of length b and form an element of
volume which can then be summed to determine the volume
of the pyramid. (a) Find the volume of this pyramid. (b) Find
the volume of a frustum of this pyramid.


5-28. A piece of cardboard having length  = 21 (15 + 33) and width w = 12 (15 33)
is to be made into a box by cutting squares of length x from each corner and then
turning up the sides followed by reinforcing the sides with tape. Find the box that
can be constructed which has the maximum volume.

5-29. Find the centroid of the region bounded by the following curves.

(a) y = x2 and y = 2x (b) y = x2 and y = mx, m > 0 (c) y 2 = x and y = 2(x1)

 445

5-30.

The face of a dam is a rectangular plate of

width w and length . The plate is inclined at
an angle so that the longer side is at the water
3
level. Let = 62.4 lbs/ft denote the water density and construct an x, y-axis lying
on the plate with origin at one corner as illustrated in the sketch.
(i) Show the distance y along the plate, measured from the water surface, corre-
sponds to a water depth h = y sin
(ii) Show a rectangular element of area dA on the plate is given by dA =  dy
(iii) At depth h, the pressure p acting on the element of area dA is given by p = h.
(iv) Show the element of force dF acting on the element of area is dF = p dA = p  dy
(v) Find the total force acting normal to the face of the dam.

5-31. Assume a body falls from rest from a height of 100 meters in air and the
body experiences a drag force proportional to its velocity.
(i) Show that Newtons law of motion is represented
dv
m = mg kv
dt
where k is a proportionality constant.
(ii) Separate the variables and then integrate to determine the velocity as a function
of time.
(iii) When does the body hit the ground? What is its velocity when is hits the
ground.
(iv) Give units of measurement for all terms in the equations you used to obtain your
answer.

5-32. For each of the given differential equations assume an exponential solution
ex and find
(a) The characteristic equation (c) A fundamental set of solutions
(b) The characteristic roots (d) The general solution
dy d2 y dy
(a) y = 0 (d) 2
+3 + 2y = 0
dx dx dx
d2 y d2 y dy
(b) 2
+ 2 y = 0 (e) 2
+ 6y = 0
dx dx dx
d2 y d3 y d2 y dy
(c) 2 y = 0 (f ) + 6 + 11 + 6y = 0
dx2 dx 3 dx 2 dx
446

5-33.

For the RLC circuit illustrated Kirchhoffs

law produces the differential equation
di q dq
L + Ri + = E, i=
dt C dt
dq
(a) Set L = 0 and i = and then solve the initial-value problem that q = 0 when t = 0
dt  
and show q = q(t) = CE 1 et/RC and then find the current i in the circuit.
(b) Set E = 0, R = 0 to find the discharge of a condenser through an inductance L.
Assume the initial-values q = q0 and i = 0 at time t = 0
2
Hint: One method is to let i = dq dt
and write ddt2q = dt
di di dq
= dq dt
di
= i dq and then
separate variables.
(c) Set C = 0, E = 0 to find the decay of current in the circuit containing a resistance
and inductance. Assume the initial-value i = i0 at time t = 0 and show the
current decays according to the law i = i0 e(R/L)t . Find the current at the times
t = L/R, t = 2L/R, t = 3L/R, t = 4/R and t = 5L/R

5-34. A Paradox
1
The curve y = for 1 x T is revolved about the x-axis to form a surface.
x
(a) Find the volume V = V (T ) bounded by the surface and the planes x = 1 and
x = T . (b) Find the surface area S = S(T ) and show S(T ) > 2 ln T . (c) Show that in
the limit as T that V (T ) is finite, but S(T ) becomes infinite.
The above results shows that you can take paint and fill up the infinite volume,
but you cant paint the surface of this volume. Question: If you fill up the volume
with paint and then pour it out, does this count as painting the outside surface?

5-35. Assume y1 = y1 (x) is a solution of the differential equation

d2 y dy
L(y) = + P (x) + Q(x)y = 0
dx2 dx
What condition must u = u(x) satisfy, if y2 (x) = u(x)y1 (x) is also a solution to the
same differential equation?
d2 y dy
(a) Determine the general solution to the differential equation 2 + 2 + y = 0
dx dx
if it is known that y1 = y1 (x) = ex is a solution of this equation.
(b) Determine the general solution to the differential equation
d2 y dy
x2 2
2x + 2y = 0 if it is know that y1 = y1 (x) = x is a solution of this
dx dx
equation.
 447

5-36. Use partial integration to solve the given partial differential equations for
the most general representation of the unknown function u = u(x, y).

u 2u u
(a) = 6x2 + y (d) y = x+y (g) = x2 + y
x x y y
u 2u u
(b) x +u = y+x (e) = x+y (h) = xy
x y 2 x
u u 2 u 2u
(c) y +u =x+y (f ) + =1+y (i) = xy
y x x2 x y

Hint for (f): After one integration, multiply through by ex .

5-37. Assume f (x + iy) = u(x, y) + i v(x, y) is such that u = u(x, y) and v = v(x, y) are
real continuous functions with partial derivatives of the first and second order which
u v u v
satisfy the Cauchy-Riemann conditions = and =
x y y x
2 2 2 2
u u v v
(a) Show that + 2 =0 (b) Show that + 2 =0
x2 y x2 y

5-38. Find the largest rectangle that can be inscribed inside a circle of radius r.

5-39. Find the length of each of the given curves.

(a) y = x2 from x = 0 to x = 1 (c) y = cosh x from x = 0 to x = 1

(b) y 2 = 4x from x = 0 to x = 1 (d) y = ex from x = 0 to x = 1

5-40. A cylindrical can is to be constructed to hold V0 cubic meters of material.

The cost of construction for the sides of the cylinder is c0 dollars per square meter
and the cost of constructing the top and bottom is 3c0 dollars per square meter.
This is because the side of the cylinder can be considered as a rectangle with height
h and width 2r , where h is the height of the cylinder and r is the radius of the
cylinder. The top and bottom of the cylinder are circles and consequently more
manufacturing techniques and waste of material occurs in their construction. Find
the dimensions of the cylinder requiring minimum cost of construction. Hint: The
volume of the cylinder is V = r2 h = V0 .

5-41. Evaluate the following limits

u(x + h, y) u(x, y) u(x, y + k) u(x, y)
(a) lim (b) lim
h0 h k0 k

assuming these limits exist.

448

5-42.
(a) (Resistors in Series) Show that Ohms
law requires the voltage drop in moving around
the series circuit requires V = iR1 + iR2 + + iRn
and so one can replace the sum of the resistors
by an equivalent resistance Req given by
Req = R1 + R2 + + Rn .
(b) (Resistors in Parallel) Use Kirchhoffs law
and show the current i in the parallel circuit
must satisfy i = i1 + i2 + + in and that V = i1 R1 ,
V = i2 R 2 , . . . , V = in R n .
(c) Show an equivalent resistance Req must satisfy V = Req i and from this result show
1 1 1 1
= + ++
Req R1 R2 Rn

5-43. Assume an open container with vertical sides where the bottom of the
container has the same shape as the top of the container. If water evaporates from
this open container at a rate which is directly proportional to the exposed surface
area, use calculus to show that the depth of water in the container changes at a
constant rate and it doesnt matter what shape the top and bottom have as long as
they are the same.


5-44. Evaluate the integral I = tan4 x dx for 0 < x <
2
dz
(a) Use the substitution z = tan x and show dx = 1+z2
so that the integral becomes

z4 z 4 + z 2 (z 2 + 1) + 1
    
1
I= dz = dz = z2 1 + dz
1 + z2 z2 + 1 1 + z2

(b) Integrate the result from part (a) and then use back substitution to express the
integral I in terms of x.


5-45. Show that integrals of the type I = f (sin x, cos x) dx where f (u, v) is a rational
x
function of u, v , can be simplified by making the substitution z = tan
2
2z 1 z2 2dz 2 x 1
(a) Show sin x = , cos x = , dx = Hint: Show cos =
1+z 2 1+z 2 1+z 2 2 1 + tan2 x2
 2
1 + cos x
(b) Evaluate the integral I = dx
cos4 x
 449

5-46. The Trapezoidal Rule
Given a curve y = f (x), a x b, one can
partition the interval [a, b] into n-parts by defin-
ing a step size h = x = (ba)n
and then labeling
the points
a = x0 , x1 = x0 +h, . . . , xn1 = x0 +(n1)h, xn = x0 +nh

Next construct trapezoids, the ith trapezoid has

the vertices xi1 , yi1, xi , yi as illustrated in the
figure, where yi1 = f (xi1 ) and yi = f (xi ).
(a) Show the area of the ith trapezoid is given by
1 h
Ai = (yi1 + yi )h = [f (xi1) + f (xi )]
2 2
(b) Show the area of all the trapezoids is
n n
 n1



1 h

A = total area = Ai = (yi1 + yi )h = (y0 + yn ) + 2 yi
i=1 i=1
2 2 i=1

(c) Computer Problem

For the functions given over the interval
 b
specified compare the area under
the curve using the definite integral f (x) dx, with the trapezoidal rule for
a
approximating the area. Fill in the following table for values of n = 10, 50 and
100 associated with each function.

Area under curve y = f (x) for a x b

Function Interval Trapezoidal Integration
n b
n= f (x) axb i=1 Ai a
f (x) dx
x 0x2
x2 0x2
x3 0x2
sin x 0x
3
cos x 2
x 2

(d) If the theoretical error of approximation using the trapezoidal rule is

 (b a)3



E =  f (), where a < < b, compare your errors with the maximum
12n2
theoretical error.
450

5-47. For a > 0, verify the improper integrals
 
ax b 1
(a) e sin bx dx = 2 (b) eax cos bx dx =
0 a + b2 0 a2 + b2

5-48. If the Gamma function

 is defined by the improper integral
(x) = et tx1 dx
0
(a) Use integration by parts to show (x + 1) = x(x)
(b) Show for n an integer (n + 1) = n!

5-49. Consider two particles starting at the origin at the same time and moving
along the x-axis such that their positions at any time t are given by

s1 (t) = 2t2 + t and s2 (t) = 11t 3t2

(a) At what time will the particles have the same position and what will be their
velocities at this position?
(b) Find the particles positions when they have the same speed? What is this same
speed?
(c) Describe the motion of each particle.

5-50. Find the maximum and minimum values for the given functions

(a) f (x, y) = x2 + y 2 2x 4y 20 (b) g(x, y) = 4x2 + 9y 2 16x 54y + 61

5-51. Given a point (x0 , y0 )
= (0, 0) lying in the first quadrant. Pick a point x1 > x0
on the x-axis and draw a line from (x1 , 0) through (x0 , y0 ) which intersects the y-axis.
Find the shortest line from the x-axis, through the point (x0 , y0 ) which intersects the
y -axis Hint: If  is the length of the line segment, then minimize 2 .

5-52. Find the maximum and minimum distances from the origin to points on
the circle (x 6)2 + (y 8)2 = 25

5-53. A particle undergoes simple harmonic motion on the x-axis according to

the law x = x0 + a cos t + b sin t, where x0 , a, b, are given constant values.
(a) Find the center and amplitude of the motion.
(b) Find the period and frequency associated with the motion.
(c) Find the maximum acceleration.
 451

5-54.
A right circular conical water tank, as illus-
trated has a top radius R and height H . As-
sume water is in the tank at a depth h0 . Let
= 62.5 lbs/f t3 denote the density of water.
(a) Show weight of disk at height h produces
force dF = r2 dh
(b) Show element of work done in lifting disk a
distance H h is

dW = ( r 2 dh)(H h)

(c) Show the work done in pumping the water out over the top of the tank is
 2  h0
R
W = h2 (H h) dh
H 0

and then evaluate this integral.

 v2

5-55. For p pressure and v volume, the integral W = p dv occurs in the study
v1
of thermodynamics and represents work done by a gas.
(a) Evaluate this integral for an isothermal expansion where pv = c = constant.
(b) Evaluate this integral for an adiabatic expansion where pv = c = constant, where
= 1.41 is also a constant.

5-56.
A botanical gardens is planning the construction of flower
beds to display their hosta plants. The flower beds are to be
rectangular and constructed inside a rectangular area having
a known perimeter P . There is to be a walk surrounding each
flower bed having dimensions of s-feet on each side and e-feet
on each end. Design studies are to begin where the exact values
of P ,s and e are to be supplied for each flower bed. For a given
value of P, s and e find the dimensions of the flower bed if the
area of the flower bed is to be a maximum.
452
APPENDIX A
Units of Measurement
The following units, abbreviations and prefixes are from the
Systeme International dUnites (designated SI in all Languages.)

Prefixes.
Abbreviations
Prefix Multiplication factor Symbol
exa 1018 W
peta 1015 P
tera 1012 T
giga 109 G
mega 106 M
kilo 103 K
hecto 102 h
deka 10 da
deci 101 d
centi 102 c
milli 103 m
micro 106
nano 109 n
pico 1012 p
femto 1015 f
atto 1018 a

Basic Units.
Basic units of measurement
Unit Name Symbol
Length meter m
Mass kilogram kg
Time second s
Electric current ampere A

Temperature degree Kelvin K
Luminous intensity candela cd

Supplementary units
Unit Name Symbol
Plane angle radian rad
Solid angle steradian sr

Appendix A
 453
DERIVED UNITS
Name Units Symbol
Area square meter m2
Volume cubic meter m3
Frequency hertz Hz (s1 )
Density kilogram per cubic meter kg/m3
Velocity meter per second m/s
Angular velocity radian per second rad/s
Acceleration meter per second squared m/s2
Angular acceleration radian per second squared rad/s2
Force newton N (kg m/s2 )
Pressure newton per square meter N/m2
Kinematic viscosity square meter per second m2 /s
Dynamic viscosity newton second per square meter N s/m2
Work, energy, quantity of heat joule J (N m)
Power watt W (J/s)
Electric charge coulomb C (A s)
Voltage, Potential difference volt V (W/A)
Electromotive force volt V (W/A)
Electric force field volt per meter V/m
Electric resistance ohm (V/A)
Electric capacitance farad F (A s/V)
Magnetic flux weber Wb (V s)
Inductance henry H (V s/A)
Magnetic flux density tesla T (Wb/m2 )
Magnetic field strength ampere per meter A/m
Magnetomotive force ampere A

Physical Constants:
4 arctan 1 = = 3.14159 26535 89793 23846 2643 . . .
n
limn 1 + n1 = e = 2.71828 18284 59045 23536 0287 . . .

Eulers constant = 0.57721 56649 01532 86060 6512 . . .

= limn 1 + 12 + 13 + + n1 log n


Eulers constant
Speed of light in vacuum = 2.997925(10)8 m s1
Electron charge = 1.60210(10)19 C
Avogadros constant = 6.0221415(10)23 mol1
Planks constant = 6.6256(10)34 J s
Universal gas constant = 8.3143 J K 1 mol1 = 8314.3 J Kg 1 K 1
Boltzmann constant = 1.38054(10)23 J K 1
StefanBoltzmann constant = 5.6697(10)8 W m2 K 4
Gravitational constant = 6.67(10)11 N m2 kg 2

Appendix A
454
APPENDIX B
Background Material
Geometry

Rectangle
Area = (base)(height) = bh
Perimeter = 2b + 2h

Right Triangle
1 1
Area = (base)(height) = bh
2 2
Perimeter = b + h + r
where r2 = b2 + h2 is the Pythagorean theorem

Triangle with sides a, b, c and angles A, B, C

1 1 1
Area = (base)(height) = bh = b(a sin C)
2 2 2
Perimeter = a + b + c
a b c
Law of Sines = =
sin A sin B sin C
2 2 2
Law of Cosines c = a + b 2ab cos C

Trapezoid
1
Area = (b1 + b2 )h
2
Perimeter = b1 + b2 + c1 + c2
h h
c1 = c2 =
sin 1 sin 2

Circle
Area = 2
Perimeter = 2
Equation x2 + y2 = 2
 455
Sector of Circle
1
Area = r2 , in radians
2
s = arclength = r , in radians
Perimeter = 2r + s

Rectangular Parallelepiped

V = Volume = abh
S = Surface area = 2(ab + ah + bh)

Parallelepiped
Composed of 6 parallelograms
V = Volume = (Area of base)(height)

A = Area of base = bc sin

height = h = a cos

Sphere of radius
4
V = Volume = 3
3
S = Surface area = 4 2

Frustum of right circular cone

V = Volume = (a2 + ab + b2 )h
3
Lateral surface area = (a + b)
456
Chord Theorem for circle
a2 = x(2R x)

Right Circular Cylinder

V = Volume = (Area of base)(height) = (r2)h
Lateral surface area = 2r h
Total surface area = 2r h + 2(r2 )

Right Circular Cone

1
V = Volume = r 2 h
3

Lateral surface area = r  = r h2 + r2
height = h, base radius r

Algebra

Products and Factors

(x + a)(x + b) = x2 + (a + b)x + ab
(x + a)2 = x2 + 2ax + a2
(x b)2 = x2 2bx + b2
(x + a)(x + b)(x + c) = x3 + (a + b + c)x2 + (ac + bc + ab)x + abc
x2 y 2 = (x y)(x + y)
x3 y 3 = (x y)(x2 + xy + y 2 )
x3 + y 3 = (x + y)(x2 xy + y 2 )
x4 y 4 = (x y)(x + y)(x2 + y 2 )

2 b b2 4ac
If ax + bx + c = 0, then x =
2a
 457
Binomial Expansion
For n = 1, 2, 3, . . . an integer, then
n(n 1) n2 2 n(n 1)(n 2) n3 3
(x + y)n = xn + nxn1 y + x y + x y + + yn
2! 3!

where n! is read n factorial and is defined

n! = n(n 1)(n 2) 3 2 1 and 0! = 1 by definition.

Binomial Coefficients
The binomial coefficients can also be defined by the expression
 
n n!
= where n! = n(n 1)(n 2) 3 2 1
k k!(n k)!

where for n = 1, 2, 3, . . . is an integer. The binomial expansion has the alternative

representation
         
n n n n n1 n n2 2 n n3 3 n n
(x + y) = x + x y+ x y + x y + y
0 1 2 3 n

Laws of Exponents
Let s and t denote real numbers and let m and n denote positive integers.
For nonzero values of x and y

0
x = 1, x
= 0 s t
(x ) =x st x1/n = n x

xs xt =xs+t (xy)s =xs y s xm/n = n xm
 1/n
xs 1 x x1/n n
x
=xst xs = s = 1/n =
xt x y y n y

Laws of Logarithms
If x = by and b
= 0, then one can write y = logb x, where y is called the logarithm
of x to the base b. For P > 0 and Q > 0, logarithms satisfy the following properties
logb (P Q) = logb P + logb Q
P
logb = logb P logb Q
Q
logb QP =P logb Q
458
Trigonometry
Pythagorean identities
Using the Pythagorean theorem x2 + y2 = r2 associated with a right triangle with
sides x, y and hypotenuse r, there results the following trigonometric identities,
known as the Pythagorean identities.
 x 2  y 2  y 2  r 2  2  2
x r
+ =1, 1+ = , +1 = ,
r r x x y y
cos2 + sin2 =1, 1 + tan2 = sec2 , cot2 + 1 = csc2 ,

Angle Addition and Difference Formulas

sin(A + B) = sin A cos B + cos A sin B, sin(A B) = sin A cos B cos A sin B
cos(A + B) = cos A cos B sin A sin B, cos(A B) = cos A cos B + sin A sin B
tan A + tan B tan A tan B
tan(A + B) = , tan(A B) =
1 tan A tan B 1 + tan A tan B

Double angle formulas

2 tan A
sin 2A =2 sin A cos A =
1 + tan2 A
1 tan2 A
cos 2A = cos2 A sin2 A = 1 2 sin2 A = 2 cos2 A 1 =
1 + tan2 A
2 tan A 2 cot A
tan 2A = 2 =
1 tan A cot2 A 1

Half angle formulas


A 1 cos A
sin =
2 2

A 1 + cos A
cos =
2 2

A 1 cos A sin A 1 cos A
tan = = =
2 1 + cos A 1 + cos A sin A

The sign depends upon the quadrant A/2 lies in.

Multiple angle formulas

sin 3A =3 sin A 4 sin3 A, sin 4A =4 sin A cos A 8 sin3 A cos A

cos 3A =4 cos3 A 3 cos A, cos 4A =8 cos4 A 8 cos2 A + 1
3 tan A tan3 A 4 tan A 4 tan3 A
tan 3A = , tan 4A =
1 3 tan2 A 1 6 tan2 A + tan4 A
 459
Multiple angle formulas

sin 5A =5 sin A 20 sin3 A + 16 sin5 A

cos 5A =16 cos5 A 20 cos3 A + 5 cos A
tan5 A 10 tan3 A + 5 tan A
tan 5A =
1 10 tan2 A + 5 tan4 A
sin 6A =6 cos5 A sin A 20 cos3 A sin3 A + 6 cos A sin5 A
cos 6A = cos6 A 15 cos4 A sin2 A + 15 cos2 A sin4 A sin6 A
6 tan A 20 tan3 A + 6 tan5 A
tan 6A =
1 15 tan2 A + 15 tan4 A tan6 A
Summation and difference formula

A+B AB AB A+B
sin A + sin B =2 sin( ) cos( ), sin A sin B =2 sin( ) cos( )
2 2 2 2
A+B AB AB A+B
cos A + cos B =2 cos( ) cos( ), cos A cos B = 2 sin( ) sin( )
2 2 2 2
sin(A + B) sin(A B)
tan A + tan B = , tan A tan B =
cos A cos B cos A cos B

Product formula

1 1
sin A sin B = cos(A B) cos(A + B)
2 2
1 1
cos A cos B = cos(A B) + cos(A + B)
2 2
1 1
sin A cos B = sin(A B) + sin(A + B)
2 2

Additional relations

sin A sin B AB
= tan( )
cos A + cos B 2
sin(A + B) sin(A B) = sin2 A sin2 B, sin A sin B AB
= cot( )
sin(A + B) sin(A B) = cos2 A cos2 B, cos A cos B 2
A+B
cos(A + B) cos(A B) = cos2 A sin2 B, sin A + sin B tan( 2 )
=
sin A sin B AB
tan( )
2
460
Powers of trigonometric functions
1 1 1 1
sin2 A = cos 2A, cos2 A = + cos 2A
2 2 2 2
3 1 3 1
sin3 A = sin A sin 3A, cos3 A = cos A + cos 3A
4 4 4 4
4 3 1 1 4 3 1 1
sin A = cos 2A + cos 4A, cos A = + cos 2A + cos 4A
8 2 8 8 2 8
Inverse Trigonometric Functions

1
sin1 x = cos1 x sin1 = csc1 x
2 x
1
cos1 x = sin1 x cos1 = sec1 x
2 x
1
tan x = cot1 x
1
tan1 = cot1 x
2 x
Symmetry properties of trigonometric functions

sin = sin() = cos(/2 ) = cos(/2 + ) = + sin( ) = sin( + )

cos = + cos() = sin(/2 ) = + sin(/2 + ) = cos( ) = cos( + )
tan = tan() = cot(/2 ) = cot(/2 + ) = tan( ) = + tan( + )

cot = cot() = tan(/2 ) = tan(/2 + ) = cot( ) = + cot( + )

sec = + sec() = csc(/2 ) = + csc(/2 + ) = sec( ) = sec( + )
csc = csc() = sec(/2 ) = + sec(/2 + ) = + csc( ) = csc( + )

Transformations
The following transformations are sometimes useful in simplifying expressions.
u
1. If tan = A, then
2
2A 1 A2 2A
sin u = , cos u = , tan u =
1 + A2 1+A 2 1 A2

y
2. The transformation sin v = y, requires cos v = 1 y 2 , and tan v =

1 y2
Law of sines

a b c
= =
sin A sin B sin C
Law of cosines
a2 =b2 + c2 2bc cos A
b2 =c2 + a2 2ac cos B

c2 =a2 + b2 2ab cos C

 461
Special Numbers
Rational Numbers
All those numbers having the form p/q, where p and q are integers and q is
understood to be different from zero, are called rational numbers.
Irrational Numbers
Those numbers that cannot be written as the ratio of two numbers are called
irrational numbers.
The Number
The Greek letter (pronounced pi) is an irrational number and can be defined
as the limiting sum1 of the infinite series
(1)n
 
1 1 1 1 1
=4 1 + + ++ +
3 5 7 9 11 2n + 1

Using a computer one can verify that the numerical value of to 50 decimal places
is given by

= 3.1415926535897932384626433832795028841971693993751 . . .

The number has the physical significance of representing the circumference C of

a circle divided by its diameter D. The symbol for the ratio C/D was introduced
by William Jones (1675-1749), a Welsh mathematician. It became a standard nota-
tion for representing C/D after Euler also started using the symbol for this ratio
sometime around 1737.
The Number e
The limiting sum
1 1 1
1+ + ++ +
2! 3! n!
is an irrational number which by agreement is called the number e. Using a computer
this number, to 50 decimal places, has the numerical value

e = 2.71828182845904523536028747135266249775724709369996 . . .

The number e is referred to as the base of the natural logarithm and the function
f (x) = ex is called the exponential function.
1
Limits are very important in the study of calculus.
462
Greek Alphabet

Letter Name Letter Name

A alpha N nu
B beta xi
gamma O o omicron
delta pi
E  epsilon P rho
Z zeta sigma
H eta T tau
theta upsilon
I iota phi
K kappa X chi
lambda psi
M mu omega
Notation
By convention letters from the beginning of an alphabet, such as a, b, c, . . . or the
Greek letters , , , . . . are often used to denote quantities which have a constant
value. Subscripted quantities such as x0 , x1 , x2 , . . . or y0 , y1 , y2 , . . . can also be used to
represent constant quantities. A variable is a quantity which is allowed to change
its value. The letters u, v, w, x, y, z or the Greek letters , , are most often used to
denote variable quantities.
Inequalities
The mathematical symbols = (equals),
= (not equal), < (less than), << (much
less than), (less than or equal), > (greater than), >> (much greater than)
(greater than or equal), and | | (absolute value) occur frequently in mathematics to
compare real numbers a, b, c, . . .. The law of trichotomy states that if a and b are real
numbers, then exactly one of the following must be true. Either a equals b, a is less
than b or a is greater than b. These statements are expressed using the mathematical
notations2
a = b, a < b, a>b
2
In mathematical notation, the statement b > a, read b is greater than a, can also be represented a < b or
a is less than bdepending upon your way of looking at things.
 463
Inequalities can be defined in terms of addition or subtraction. For example, one
can define
a<b if and only if a b < 0
a>b if and only if a b > 0, or alternatively
a>b if and only if there exists a positive number x such that b + x = a.

In dealing with inequalities be sure to observe the following properties associated

with real numbers a, b, c, . . .
1. A constant can be added to both sides of an inequality without changing the
inequality sign.

If a < b, then a + c < b + c for all numbers c

2. Both sides of an inequality can be multiplied or divided by a positive constant

without changing the inequality sign.

If a < b and c > 0, then ac < bc or a/c < b/c

3. If both sides of an inequality are multiplied or divided by a negative quantity,

then the inequality sign changes.

If b > a and c < 0, then bc < ac or b/c < a/c

4. The transitivity law

If a < b, and b < c, then a < c
If a = b and b = c, then a = c
If a > b, and b > c, then a > c
5. If a > 0 and b > 0, then ab > 0
6. If a < 0 and b < 0, then ab > 0 or 0 < ab

7. If a > 0 and b > 0 with a < b, then a < b
A negative times a negative is a positive
To prove that a real negative number multiplied by another real negative number
gives a positive number start by assuming a and b are real numbers satisfying a < 0
and b < 0, then one can write

a + a < a or 0 < a and b + b < b or 0 < b

464
since equals can be added to both sides of an inequality without changing the in-
equality sign. Using the fact that both sides of an inequality can be multiplied by a
positive number without changing the inequality sign, one can write
0 < (a)(b) or (a)(b) > 0

Another way to show a negative times a negative

is a positive is as follows. Think of a number line
with the number 0 separating the positive num-
bers and negative numbers. By agreement, if a
number on this number line is multiplied by -1,
then the number is to be rotated counterclockwise 180 degrees. If the positive num-
ber x is multiplied by -1, then it is rotated counterclockwise 180 degrees to produce
the number x. If the number x is multiplied by -1, then it is to be rotated 180
degrees counterclockwise to produce the positive number x. If a > 0 and b > 0, then
the product a(b) scales the number b to produce the negative number ab. If the
number ab is multiplied by 1, which is equivalent to the product (a)(b), one
obtains by rotation the number +ab.
Absolute Value
The absolute value of a number x is defined
if x 0

x,
|x| =
x, if x < 0
The symbol is often used to represent equivalence of two equations. For example,
if a and b are real numbers the statements
|x a| b b x a b ab x a+b

are all equivalent statements involving restrictions on the real number x.

An important inequality known as the triangle inequality is written
|x + y| |x| + |y| (1.1)

where x and y are real numbers. To prove this inequality observe that |x| satisfies
|x| x |x| and also |y| y |y|, so that by adding these results one obtains

(|x| + |y|) x + y |x| + |y| or |x + y| |x| + |y| (1.2)

Related to the inequality (1.2) is the reverse triangle inequality

|x y| |x| |y| (1.3)

a proof of which is left as an exercise.

 465
Cramers Rule
The system of two equations in two unknowns
1 x + 1 y =1 
1 1
 
x
or = 1
2 x + 2 y =2 2 2 y 2

has a unique solution if 1 2 2 1 is nonzero. The unique solution is given by

......
......
......
2 1.............
....
........


......
...... ..
..
..........
...... ..
.... .........

2

2

2 2

=
1............................ 1
x=

, y=

where

......
.....
. ..
..
. 1 2
...... 2 1

1 1

1 1

......
...... 2

.
2........................
...... .......
...... ....

2 2

2 2
+1 2

is a single number called the determinant of the coefficients.

The system of three equations in three unknowns
1 x + 1 y + 1 z =1
2 x + 2 y + 2 z =2 has a unique solution if the determinant of the coefficients
3 x + 3 y + 3 x =3

1 1 1

2 2 2
= 1 2 3 + 1 2 3 + 1 2 3 3 2 1 3 2 1 3 2 2

3 3 3

is nonzero. A mnemonic device to aid in calculating the determinant of the co-

efficients is to append the first two columns of the coefficients to the end of the
array and then draw diagonals through the coefficients. Multiply the elements along
an arrow and place a plus sign on the products associated with the down arrows
and.
a .minus
.
sign associated
..
......
..
.......
with ..
.......
the products of the up arrows. This gives the figure
....... ....... ....... .... .... ....
....... ....... ....... ................ ............... ................
....... ....... ....... ........ ......... .........
.......

.......
.......
.......
.......
....... ...
..

....... ...........
. ..
...
.......
....... ....... ....... ........ ... ...

....... ....... ..
...... ......
1 ........... ....... 1 ............ 1
....... 1....................... 1.......................
....... .
.
....... ...... ....... ...... ....... ............
= 2 3 + 1 2 3 + 1 2 3 3 2 1 3 2 1 3 2 2

2 .............2 .........
. ..
....... ...............2

. .
....... ...................2 ....... 2
1
....... ....... .......
..... ................... ........................ .......

...... ......
..
.....3
.... . .... . ..........
.
.......3 ............ 3 ..............

.......
.......
3.............. ..3............ .
...... ...... . ............ ............ ............
...... ...... ...... ............ . . . .
...... ...... ...... ........ ................... ..................
...... ...... ...... . . .

The solution of the three equations, three unknown system of equations is given
by the determinant ratios

1 1 1

1 1 1

1 1 1

2 2 2

2 2 2

2 2 2

3 3 3

3 3 3

3 3 3

x =

, y =

, z =

1 1 1

1 1 1

1 1 1

2 2 2

2 2 2

2 2 2

3 3 3

3 3 3

3 3 3

and is known as Cramers rule for solving a system of equations.

466

Appendix C
Table of Integrals
Indefinite Integrals
General Integration Properties


dF (x)
1. If = f(x) , then f(x) dx = F (x) + C
dx



2. If f(x) dx = F (x) + C , then the substitution x = g(u) gives f(g(u)) g
(u) du = F (g(u)) + C



dx 1 1 x du 1 u+
For example, if = tan + C , then = tan1 +C
x2 + 2 (u + )2 + 2




3. Integration by parts. If v1 (x) = v(x) dx, then u(x)v(x) dx = u(x)v1 (x) u
(x)v1 (x) dx

4. Repeated integration by parts or generalized integration by parts.

  
If v1 (x) = v(x) dx, v2 (x) = v1 (x) dx, . . . , vn(x) = vn1(x) dx, then



u(x)v(x) dx = uv1 u
v2 + u

v3 u


v4 + + (1)n1 un1 vn + (1)n u(n) (x)vn (x) dx

5. If f 1
(x) is the inverse function of f(x) and if f(x) dx is known, then



1
f (x) dx = zf(z) f(z) dz, where z = f 1 (x)

6. Fundamental theorem of calculus.

If the indefinite integral of f(x) is known, say
f(x) dx = F (x) + C , then the definite integral

b
b
b
dA = f(x) dx = F (x)]a = F (b) F (a)
a a

represents the area bounded by the x-axis, the curve

y = f(x) and the lines x = a and x = b.

7. Inequalities.


b b
(i) If f(x) g(x) for all x (a, b), then f(x) dx g(x) dx
ba a
(ii) If |f(x) M | for all x (a, b) and a f(x) dx exists, then



 b  b
 
 f(x) dx f(x) dx M (b a)
 a  a

Appendix C


467


u (x) dx
8. = ln |u(x)| + C
u(x)


(u(x) + )n+1
9. (u(x) + )n u
(x) dx = +C
(n + 1)


u
(x)v(x) v
(x)u(x) u(x)
10. dx = +C
v2 (x) v(x)


u
(x)v(x) u(x)v
(x) u(x)
11. dx = ln | |+C
u(x)v(x) v(x)


u
(x)v(x) u(x)v
(x) u(x)
12. dx = tan1 +C
u2 (x) + v2 (x) v(x)


u
(x)v(x) u(x)v
(x) 1 u(x) v(x)
13. dx = ln | |+C
u2 (x) v2 (x) 2 u(x) + v(x)


u
(x) dx 
14.  = ln |u(x) + u2 (x) + | + C
u2 (x) +


dx dx


,
=
u(x) dx u(x) + u(x) +
15. =

dx


dx
(u(x) + )(u(x) + )
, =

u(x) + (u(x) + )2


u (x) dx 1 u(x)
16. 2
= ln | |+C
u (x) + u(x) u(x) +


u
(x) dx 1 u(x)
17.  = sec1 +C
2
u(x) u (x) 2


u
(x) dx 1 u(x)
18. 2 2 2
= tan1 +C
+ u (x)


u
(x) dx 1 u(x)
19. = ln | |+C
u2 (x) 2
2 2 u(x) +


 
2u du x
20. f(sin x) dx = 2 f 2 2
, u = tan
1+u 1+u 2



du
21. f(sin x) dx = f(u) , u = sin x
1 u2


 
1 u2 du x
22. f(cos x) dx = 2 f 2 2
, u = tan
1+u 1+u 2



du
23. f(cos x) dx = f(u) , u = cos x
1 u2


 du
24. f(sin x, cos x) dx = f(u, 1 u2 ) , u = sin x
1 u2


 
2u 1 u2 du x
25. f(sin x, cos x) dx = 2 f , , u = tan
1 + u2 1 + u2 1 + u2 2


 2 
 2 u
26. f(x, + x) dx = f , u udu, u2 = + x




27. f(x, 2 x2 ) dx = f( sin u, a cos u) cos u du, x = sin u

Appendix C
468
General Integrals






28. c u(x) dx = c u(x) dx 29. [u(x) + v(x)] dx = u(x) dx + v(x) dx





1
30. u(x) u
(x) dx = | u(x) |2 +C 31. [u(x) v(x)] dx = u(x) dx v(x) dx]

2


[u(x)]n+1
32. un (x) u
(x) dx = +C 33. u(x) v (x) dx = u(x) v(x) u
(x) v(x) dx

n+1

u
(x)
34. F
[u(x)] u
(x) dx = F [u(x)] + C 35. dx = ln | u(x) | +C


u(x)
u

36. dx = u + C 37. 1 dx = x + C

2 u

xn+1 1
38. xn dx = +C 39. dx = ln | x | +C

n+1
x
1 1 u
40. eau u
dx = eau + C 41. au u
dx = a +C

a
ln a
42. sin u u
dx = cos u + C 43. cos u u
dx = sin u + C



44. tan u u
dx = ln | sec u | +C 45. cot u u
dx = ln | sin u | +C



46. sec u u
dx = ln | sec u + tan u | +C 47. csc u u
dx = ln | csc u cot u | +C



48. sinh u u
dx = cosh u + C 49. cosh u u
dx = sinh u + C



50. tanh u u
dx = ln cosh u + C 51. coth u u
dx = ln sinh u + C



u
52. sech u u
dx = sin1 (tanh u) + C 53. csch u u
dx = ln tanh +C


2
1 1 u 1
54. sin2 u u
dx = u sin 2u + C 55. cos2 u u
dx = + sin 2u + C

2 4
2 4
2

56. tan u u dx = tan u u + C 57. cot2 u u
dx = cot u u + C



58. sec 2 u u
dx = tan u + C 59. csc2 u u
dx = cot u + C



1 1 1 1
60. sinh2 u u
dx = sinh 2u u + C 61. cosh2 u u
dx = sinh 2u + u + C

4 2
4 2
62. tanh2 u u
dx = u tanh u + C 63. coth2 u u
dx = u coth u + C



64. sech 2 u u
dx = tanh u + C 65. csch 2 u u
dx = coth u + C



66. sec u tan u u
dx = sec u + C 67. csc u cot u u
dx = csc u + C



68. sech u tanh u u
dx = sech u + C 69. csch u coth u u
dx = csch u + C

Integrals containing X = a + bx, a
= 0 and b
= 0

X n+1
70. X n dx = + C , n
= 1
b(n + 1)


X n+2 aX n+1
71. xX n dx = + C, n
= 1, n
= 2
b2 (n + 2) b2 (n + 1)


b a bc
72. X(x + c)n dx = (x + c)n+2 + (x + c)n+1 + C
n+2 n+1

Appendix C
 469

 
1 X n+3 2aX n+2 a2 X n+1
73. x2 X n dx = + +C
b3 n + 3 n+2 n+1



1 am
74. xn1 X m dx = xn X m + xn1 X m1 dx
n+m m+n



Xm 1 X m+1 mn+1 b Xm
75. n+1
dx = n
+ dx
x na x n a xn


dx 1
76. = ln X + C
X b


x dx 1
77. = 2 (X a ln | X |) + C
X b


x2 dx 1

78. = 3 X 2 4aX + 2a2 ln | X | + C

X 2b


dx 1 x
79. = ln | | +C
xX a X


dx a + 2bx 2b X
80. = 2 + 3 ln | | + C
x3 X a xX a x


dx 1
81. = +C
X2 bX


x dx 1  a
82. = ln | X | + +C
X2 b2 X

 
x2 dx 1 a2
83. = 3 X 2a ln | X | +C
X2 b X


dx 1 1 X
84. 2
= 2 ln | | +C
xX aX a x


dx a + 2bx 2b X
85. = + 3 ln | | +C
x2 X 2 2
a xX a x


dx 1
86. 3
= +C
X 2bX 2

 
x dx 1 1 a
87. = + +C
X3 b2 X 2X 2

 
x2 dx 1 2a a2
88. = ln | X | + +C
X3 b3 X 2X 2


dx 1 1 X
89. = + ln | | +C
xX 3 2aX 2 aX x

Appendix C
470


dx b 2b 1 3b X
90. = 3 3 + 4 ln | |
x2 X 3 2
2a X a X a x a x

 
x dx 1 1 a
91. = 2 + + C, n
= 1, 2
Xn b (n 2)X n2 (n 1)X n1

 
x2 dx 1 1 2a a2
92. = + + C, n
= 1, 2, 3
Xn b3 (n 3)X n3 (n 2)X n2 (n 1)X n1


2 3/2
93. X dx = X +C
3b

2
94. x X dx = (3bx 2a)X 3/2 + C
15b2

2
95. x2 X dx = (8a2 12abx + 15b2 x2 )X 3/2 + C
105b3



X dx
96. dx = 2 X + a
x x X



X X b dx
97. dx = +
x2 x 2 x X


dx 2
98. = X +C
X b


x dx 2
99. = 2 (bx 2a) X + C
X 3b

2
x dx 2
100. = 3
(8a2 4abx + 3b2 x2 ) X + C
X 15b

1 X a



a ln | | +C1 , a>0
dx X+ a
101. =
x X
2 1 X

tan + C2 , a<0
a a



dx X b dx
102. =
2
x X ax 2a x X



2 2na
103. xn X dx = xn X 3/2 xn1 X dx
(2n + 3)b (2n + 3)b



X 1 X 3/2 (2n 5)b X
104. dx = dx
xn (n 1)a xn1 2(n 1)a xn1



xm X n an
105. xm1 X n dx = + xm1 X n1 dx + C
m+n m+n



Xn X n+1 nm+1 b Xn
106. m+1
dx = + dx
x ma xm m a xm

Appendix C
 471



Xn Xn X n1
107. dx = +a dx
x n x

Integrals containing X = a + bx and Y = + x, (b
= 0,
= 0, = a b
= 0 )

dx 1 Y
108. = ln | | +C
XY X

 
x dx 1 a
109. = ln | X | ln | Y | + C
XY b


x2 dx x a2 2
110. = = 2 ln |X| + 2 ln |Y | + C
XY b b

 
dx 1 1 Y
111. = + ln | | + C
X2Y X X


x dx a Y
112. = 2 ln | | + C
X2Y bX X

 
x2 dx a2 1 2 a(a 2b)
113. = 2 + 2 ln |Y | + ln |X| + C
X2Y b X b2


X b Y
114. dx = x + 2 ln | | + C
Y X



+ 2bY 2 dx
115. XY dx = XY
4b 8b XY



dx 1 (m + n 2)b dx
116. = + , m
= 1
XnY m (m 1)X n1 Y m1 (m 1) X n Y m1

2 1 X
tan , +C1 < 0



dx
117. =
Y X 1 ln | X | + C2 ,


> 0
X +

2 1 X



tan + C1 , b < 0, bY > 0
dx b bY
118. =
XY
2 X
tanh1
+ C2 , b > 0, bY > 0
b bY



x dx 1 (b + a) dx
119. = XY
XY b 2b XY



Y 1 dx
120. dx = XY
X b 2b XY



X 2 dx
121. dx = X+
Y Y X

Appendix C
472
Integrals containing terms of the form a + bxn
 
1 1 b



tan x + C, ab > 0
dx ab a
122. =
a + bx2 1
 
 a + abx 


ln   + C, ab < 0
2 ab  a abx 


x dx 1 a
123. = ln |x2 + | + C
a + bx2 2b b



x2 dx x a dx
124. =
a + bx2 b b a + bx2



dx x 1 dx
125. = +
(a + bx2 )2 2a(a + bx2 ) 2a a + bx2

 
dx 1  x2 
126. = ln  +C
x(a + bx2 ) 2a  a + bx2 



dx 1 b dx
127. 2 2
=
x (a + bx ) ax a a + bx2



dx 1 x 2n 1 dx
128. = +
(a + bx2 )n+1 2na (a + bx2 )n 2na (a + bx2 )n

    
dx 1 2x  ( + x)2 
129. 3 3 3
= 2
2 3 tan1
+ ln 
2 2 2
 +C
+ x 6 3  x + x 

    
x dx 1 2x  ( + x)2 
130. = 2 3 tan1
ln  2
  +C
3 + 3 x3 6 2 3 x + 2 x2 

If X = a + bxn, then



xm X p apn
131. xm1 X p dx = + xm1 X p1 dx
m + pn m + pn



xm X p+1 m + pn + n
132. xm1 X p dx = + xm1 X p+1 dx
an(p + 1) an(p + 1)



xmn X p+1 (m n) a
133. x m1 P
X dx = xmn1 X p dx
b(m + pn) b(m + pn)



xm X p+1 (m + pn + n) b
134. x m1 p
X dx = xm+n1 X p dx
am am



xmn X p+1 mn
135. xm1 X p dx = xmn1 X p+1 dx
bn(p + 1) bn(p + 1)



xm X p bpn
136. xm1 X p dx = xm+n1 X p1 dx
m m

Appendix C
 473
Integrals containing X = 2ax x2 , a
= 0

 
(x a) a2 xa
137. X dx = X+ sin1 +C
2 2 |a|

 
dx xa
138. = sin1 +C
X |a|

 
xa
139. x X dx = sin1 +C
|a|

 
x dx xa
140. = X + a sin1 +C
X |a|


dx xa
141. = +C
X 3/2 a2 X


x dx x
142. 3/2
= +C
X a X


dx 1 x
143. = ln | |+C
X 2a x 2a


x dx
144. = ln |x 2a| + C

X
dx 1 1 1 x
145. = + ln | |+C
X2 4ax 4a2 (x 2a) 4a2 x 2a


x dx 1 1 x
146. 2
= + 2 ln | |+C

X 2a(x 2a) 4a x 2a

1 (2n + 1)a
147. xn X dx = xn1 X 3/2 + xn1 X dx, n
= 2
n+2 n+2



X dx 1 X 3/2 n3 X
148. n
= n
+ n1
dx, n
= 3/2
x (3 2n)a x (2n 3)a x

Integrals containing X = ax2 + bx + c with = 4ac b2,
= 0, a
= 0

 
1 2ax + b


ln + C1 , <0




2ax + b +
dx 2 2ax + b
149. = tan1 + C2 , >0
X





1
+ C3 , =0

a(x + b/2a)

x dx 1 b 1
150. = ln | X | dx
X 2c 2a X



x2 dx x b 2ac dx
151. = 2 ln |X| +
X a 2a 2a2 X

Appendix C
474



dx 1 x2 b dx
152. = ln | |
xX 2c X 2c X



dx b X 1 2ac dx
153. = 2 ln | 2 | +
x2 X 2c x cx 2c2 X



dx bx + 2c b dx
154. =
X2 X X



x dx bx + 2c b dx
155. 2
=
X X X



x2 dx (2ac )x + bc 2c dx
156. 2
= +
X aX X




dx 1 b dx 1 dx
157. = +
xX 2 2cX 2c X2 c xX




dx 1 3a dx 2b dx
158. =
x2 X 2 cxX c X2 c xX 2
1

ln |2 aX + 2ax + b| + C1 , a>0
a



 
dx 1 2ax + b
1
159. = sinh
a
+ C2 , a >, > 0
X


 
1 sin1 2ax +b


+ C3 , a < 0, < 0

a

x dx 1 b dx
160. = X
X a 2a X

 

x2 dx x 3b 2b2 dx
161. = 2 X+ 2

X 2a 4a 8a X

1 2 cX 2c

ln | + + b| + C1 , c>0


c x x


 
dx 1 bx + 2c
162. = sinh1 + C2 , c > 0, > 0
x X

c x
 

1 bx + 2c

sin1 + C3 , c < 0, < 0
c x



dx X b dx
163. =
2
x X cx 2c x X



1 dx
164. X dx = (2ax + b) X +
4a 8a X



1 3/2 b(2ax + b) b dx
165. x X dx = X X
3a 8a2 16a2 X



6ax 5b 3/2 4b2
166. 2
x X dx = X + X dx
24a2 16a2

Appendix C
 475




X b dx dx
167. dx = X + +c
x 2 X x X




X X dx b dx
168. dx = +a +
x2 x X 2 x X


dx 2(2ax + b)
169. 3/2
= +C
X X


x dx 2(bx + 2c)
170. 3/2
= +C
X X



x2 dx (b2 )x + 2bc 1 dx
171. 3/2
= +
X a X a X




dx 1 1 dx b dx
172. = +
xX 3/2 x X c x X 2c X 3/2




dx ax2 + 2bx + c b2 2ac dx 3b dx
173. = + 2
x2 X 3/2 c2 x X 2c2 X 3/2 2c x X


dx 2(2ax + b)
174. = +C

X X X  
dx 2(2ax + b) 1 8a
175. = + +C
X2 X 3 X X

 

(2ax + b) 3 32 dx
176. X X dx = X X+ +
8a 8a 128a2 X

 

(2ax + b) 5 152 53 dx
177. 2
X X dx = X X + 2
X+ +
8a 16a 128a2 1024a3 X


x dx 2(bx + 2c)
178. = +C

X2 X X

x dx (b2 )x + 2bc 1 dx
179. = +
X X a X a X



X2 X b
180. xX X dx = X X dx
5a 2a

 
181. f(x, ax2 + bx + c) dx Try substitutions (i) ax2 + bx + c = a(x + z)


(ii) ax2 + bx + c = xz+ c and if ax2 +bx+c = a(xx1 )(xx2 ), then (iii) let (xx2 ) = z 2(xx1 )

Integrals containing X = x2 + a2


dx 1 x 1 a 1 x 2 + a2
182. = tan1 + C or cos1 +C or sec 1 +C

X a a a x + a2
2 a a
x dx 1
183. = ln X + C
X 2

Appendix C
476


x2 dx x
184. = x a tan1 + C
X a


x3 dx x2 a2
185. = ln |x2 + a2 | + C
X 2 2


dx 1 x2
186. = 2 ln | | + C
xX 2a X


dx 1 1 x
187. 2
= 2 3 tan1 + C
x X a x a a


dx 1 1 x2
188. = ln | |+C
x3 X 2a2 x2 2a4 X


dx x 1 x
189. 2
= 2 + 3 tan1 + C

X 2a X 2a a
x dx 1
190. = +C
X2 2X


x2 dx x 1 x
191. = + tan1 + C
X2 2X 2a a


x3 dx a2 1
192. 2
= + ln |X| + C
X 2X 2


dx 1 1 x
193. = 2 + 4 ln | | + C
xX 2 2a X 2a X


dx 1 x 3 x
194. = 5 tan1 + C
x2 X 2 a4 X 4
2a X 2a a


dx 1 1 1 x2
195. = 4 6 ln | | + C
x3 X 2 4
2a x 2 2a X a X


dx x 3x 3 x
196. = 2 2 + 4 + 5 tan1 + C
X3 4a X 8a X 8a a



dx x 2n 3 dx
197. = + , n>1
Xn 2(n 1)a2 X n1 (2(n 1)a2 X n1


x dx 1
198. n
= +C
X 2(n 1)X n1



dx 1 1 dx
199. n
= 2 n1
+ 2
xX 2(n 1)a X a xX n1

Integrals containing the square root of X = x2 + a2

1 a2
200. X dx = xX + ln |x + X| + C
2 2

Appendix C
 477

1
201. x X dx = X 3/2 + C
3


1 1 a2
202. x2 X dx = xX 3/2 a2 x X ln |x + X| + C
4 8 8

1 a2
203. x3 X dx = X 5/2 X 3/2 + C
5 3



X a+ X
204. dx = X a ln | |+C
x x



X X
205. dx = + ln |x + X| + C
x2 x


X X 1 a+ X
206. dx = 2 ln | |+C
x3 2x 2a x


dx x
207. = ln |x + X| + C or sinh1 +C
X a


x dx
208. = X +C
X


x2 dx x a2
209. = X ln |x + X| + C
X 2 2


x3 dx 1
210. = X 3/2 a2 X + C
X 3


dx 1 a+ X
211. = ln | |+C
x X a x


dx X
212. = 2 +C
x2 X a x


dx X 1 a+ X
213. = 2 2 + 3 ln | |+C
x3 X 2a x 2a x


1 3/2 3 2 3
214. X 3/2 dx = X + a x X + a4 ln |x + X| + C
4 8 8


1 5/2
215. xX 3/2 dx = X +C
5


1 5/2 1 1 1
216. x2 X 3/2 dx = X a2 xX 3/2 a4 x X a6 ln |x + X| + C
6 24 16 16


1 7/2 1 2 5/2
217. x3 X 3/2 dx = X a X +C
7 5

Appendix C
478


X 3/2 1 3/2 2
3 a+ X
218. dx = X + a X a ln | |+C
x 3 x


X 3/2 X 3/2 3 3 2
219. dx = + x X + a ln |x + X| + C
x2 x 2 2


X 3/2 X 3/2 3 3 a+ x
220. dx = + X a ln | |+C
x3 2x2 2 2 x


dx x
221. 3/2
= +C
X 2
a X


x dx 1
222. = +C
X 3/2 X


x2 dx x
223. = + ln |x + X| + C
X 3/2 X


x3 dx a2
224. = X + +C
X 3/2 X


dx 1 1 a+ X
225. = 3 ln | |+C
xX 3/2 a2 X a x


dx X x
226. = 4 +C
x2 X 3/2 a x a4 X


dx 1 3 3 a+ X
227. = + ln | |+C
x3 X 3/2 2a2 x2 X 2a4 X 2a5 x



228. f(x, X) dx = a f(a tan u, a sec u) sec 2 u du, x = a tan u

Integrals containing X = x2 a2 with x2 > a2

 
dx 1 xa 1 x 1 a
229. = ln + C or coth1 + C or tanh1 + C

X 2a x + a a a a x
x dx 1
230. = ln X + C

X 2
x2 dx a xa
231. = x + ln | |+C
X 2 x+a


x3 dx x2 a2
232. = + ln |X| + C
X 2 2


dx 1 X
233. = 2 ln | 2 | + C
xX 2a x


dx 1 1 xa
234. 2
= 2 + 3 ln | |+C
x X a x 2a x+a

Appendix C
 479


dx 1 1 x2
235. 3
= 2 4 ln | | + C
x X 2a x 2a X


dx x 1 xa
236. = 2 3 ln | |+C
X2 2a X 4a x+a


x dx 1
237. = +C
X2 2X


x2 dx x 1 xa
238. 2
= + ln | |+C
X 2X 4a x+a


x3 dx a2 1
239. 2
= + ln |X| + C
X 2X 2


dx 1 1 x2
240. 2
= 2 + 4 ln | | + C
xX 2a X 2a X


dx 1 x 3 xa
241. = 4 4 5 ln | |+C
x2 X 2 a x 2a X 4a x+a


dx 1 1 1 x2
242. = + ln | |+C
x3 X 2 2a4 x2 2a4 X a6 X



dx x 2n 3 dx
243. n
= 2 n1
, n>1
X 2(n 1)a X 2(n 1)a2 X n1


x dx 1
244. n
= +C
X 2(n 1)X n1



dx 1 1 dx
245. = 2
xX n 2(n 1)a2 X n1 a xX n1

Integrals containing the square root of X = x2 a2 with x2 > a2

1 a2
246. X dx = x X ln |x + X| + C
2 2

1
247. x X dx = X 3/2 + C
3


1 1 a4
248. x2 X dx = xX 3/2 + a2 x X ln |x + X| + C
4 8 8

1 1
249. x3 X dx = X 5/2 + a2 X 3/2 + C
5 3


X x
250. dx = X a sec 1 | | + C
x a

Appendix C
480


X X
251. 2
dx = + ln |x + X| + C
x x


X X 1 x
252. dx = 2 + sec1 | | + C
x3 2x 2a a


dx
253. = ln |x + X| + C
X


x dx
254. = X +C
X


x2 dx 1 a2
255. = x X+ ln |x + X| + C
X 2 2


x3 dx 1
256. = X 3/2 + a2 X + C
X 3


dx 1 x
257. = sec 1 | | + C
x X a a


dx X
258. = 2 +C
2
x X a x


dx X 1 x
259. = 2 2 + 3 sec1 | | + C
x3 X 2a x 2a a


x 3/2 3 2 3
260. X 3/2 dx = X a x X + a4 ln |x + X| + C
4 8 8


1 5/2
261. xX 3/2 dx = X +C
5


1 1 1 a6
262. x2 X 3/2 dx = xX 5/2 + a2 xX 3/2 a4 x X + ln |x + X| + C
6 24 16 16


1 7/2 1 2 5/2
263. x3 X 3/2 dx = X + a X +C
7 5


X 3/2 1 x
264. dx = X 3/2 a2 X + a3 sec1 | | + C
x 3 a


X 3/2 X 3/2 3 3
265. 2
dx = + x X a2 ln |x + X| + C
x x 2 2


X 3/2 X 3/2 3 3 x
266. dx = + X a sec1 | | + C
x3 2x2 2 2 a


dx x
267. 3/2
= +C
X 2
a X

Appendix C
 481


x dx 1
268. 3/2
= +C
X X


x2 dx x a2
269. 3/2
= + C
X X X


x3 dx
270. 3/2
= X + ln |x + X| + C
X


dx 1 1 x
271. = 3 sec1 | | + C
xX 3/2 2
a X a a


dx X x
272. = 4 +C
x2 X 3/2 a x 4
a X


dx 1 3 3 x
273. = 5 sec1 | | + C
x3 X 3/2 2a2 x2 X 2a4 X 2a a

Integrals containing X = a2 x2 with x2 < a2

 
dx 1 a+x 1 x
274. = ln + C or tanh1 + C
X 2a ax a a


x dx 1
275. = ln X + C
X 2


x2 dx a a+x
276. = x + ln | |+C
X 2 ax


x3 dx 1 a2
277. = x2 ln |X| + C
X 2 2


d 1 x2
278. = 2 ln | | + C
xX 2a X


dx 1 1 a+x
279. = 2 + 3 ln | |+C
x2 X a x 2a ax


dx 1 1 x2
280. 3
= 2 2 + 4 ln | | + C
x X 2a x 2a X


dx x 1 a+x
281. = 2 + 3 ln | |+C
X2 2a X 4a ax


x dx 1
282. 2
= +C
X 2X


x2 dx x 1 a+x
283. 2
= ln | |+C
X 2X 4a ax

Appendix C
482


x3 dx a2 1
284. 2
= + ln |X| + C
X 2X 2


dx 1 1 x2
285. 2
= 2 + 4 ln | | + C
xX 2a X 2a X


dx 1 x 3 a+x
286. = 4 + 4 + 5 ln | |+C
x2 X 2 a x 2a X 4a ax


dx 1 1 1 x2
287. = + + ln | |+C
x3 X 2 2a4 x2 2a4 X a6 X



dx x 2n 3 dx
288. n
= 2 n1
+
X 2(n 1)a X 2(n 1)a2 X n1


x dx 1
289. = +C
Xn 2(n 1)X n1



dx 1 1 dx
290. = + 2
xX n 2(n 1)a2 X n1 a xX n1

Integrals containing the square root of X = a2 x2 with x2 < a2

1 a2 x
291. X dx = x X + sin1 + C
2 2 a

1
292. x X dx = X 3/2 + C
3


1 1 1 x
293. x2 X dx = xX 3/2 + a2 x X + a4 sin1 + C
4 8 8 a

1 1
294. x3 X dx = X 5/2 a2 X 3/2 + C
5 3


X a+ X
295. dx = X a ln | |+C
x x


X X x
296. dx = sin1 + C
x2 x a


X X 1 a+ X
297. dx = + ln | |+C
x3 2x2 2a x


dx x
298. = sin1 + C
X a


x dx
299. = X +C
X

Appendix C
 483


x2 dx 1 a2 x
300. = x X+ sin1 + C
X 2 2 a


x3 dx 1
301. = X 3/2 a2 X + C
X 3


dx 1 a+ X
302. = ln | |+C
x X a x


dx X
303. = 2 +C
2
x X a x


dx X 1 a+ X
304. = 2 2 3 ln | |+C
x3 X 2a x 2a x


1 3 3 x
305. X 3/2 dx = xX 3/2 + a2 x X + a4 sin1 + C
4 8 8 a


1
306. xX 3/2 dx = X 5/2 + C
5


1 1 1 a6 x
307. x2 X 3/2 dx = xX 5/2 + a2 xX 3/2 + a4 x X + sin1 + C
6 24 16 16 a


1 7/2 1 2 5/2
308. x3 X 3/2 dx = X a X +C
7 5


X 3/2 1 3/2 2 a+ X
309. 3
dx = X a X a ln | |+C
x 3 x


X 3/2 X 3/2 3 3 x
310. 2
dx = x X a2 sin1 + C
x x 2 2 a


X 3/2 X 3/2 3 3 a+ X
311. dx = X + a ln | |+C
x3 2x2 2 2 x


dx x
312. 3/2
= +C
X 2
a X


x dx 1
313. 3/2
= +C
X X


x2 dx x x
314. = sin1 + C
X 3/2 X a


x3 dx a2
315. 3/2
= X + +C
X X


dx 1 1 a+ X
316. = 3 ln | |+C
xX 3/2 a2 X a x

Appendix C
484


dx X x
317. = 4 + +C
x2 X 3/2 a x a4 X


dx 1 3 3 a+ X
318. = + 5 ln | |+C
x3 X 3/2 2a2 x2 X 2a4 X 2a x

Integrals Containing X = x3 + a3

 
dx 1 (x + a)3 1 2x a
319. = 2 ln | |+ tan1 +C
X 6a X 3a2 3a

 
x dx 1 X 1 2x a
320. = ln | | + tan1 +C
X 6a (x + a)3 3a 3a

2
x dx 1
321. = ln |X| + C
X 2


dx 1 x3
322. = 3 ln | | + C

xX 3a X  
dx 1 1 X 1 2x a
323. = ln | | tan1
+C
x2 X a2 x 6a4 (x + a)3 3a4  3a 


dx x 1 (x + a)3 2 2x a
324. 2
= 3 + 5 ln | |+ tan1 +C
X 3a X 9a X 3 3a5 3a

 
x dx x2 1 X 1 2x a
325. = 3 + ln | |+ tan1
+C

X2 3a X 18a4 (x + a)3 3 3a4 3a
2
x dx 1
326. 2
= +C

X 3X
dx 1 1 x3
327. 2
= 2 + 6 ln | | + C

xX 3a X 3a X

dx 1 x2 4 x dx
328. =

x2 X 2 a6 x 3a6 X 3a6 X 
dx 1 9a5 x 15a2 x 1 2x a
329. = + + 10 3 tan ( ) + 10 ln |x + a| 5 ln |x 2
ax + a 2
| +C
X3 54a3 X 2 X 3a

Integrals containing X = x4 + a4

 
dx 1 X 1 2ax
330. = ln | | tan1
+C
X 4 2a3 (x2 2ax + a2 )2 2 2a3 x 2 a2

 2
x dx 1 x
331. = 2 tan1 +C
X 2a a2  

2
x dx 1 X 1 2ax
332. = ln | | tan 1
+C
X 4 2a (x2 + 2ax + a2 )2 2 2a x 2 a2

3
x dx 1
333. = ln |X| + C

X 4
dx 1 x4
334. = 4 ln | | + C
xX 4a X

 
dx 1 1 (x2 2ax + a2 )2 1 2ax
335. = 4 ln | |+ tan 1
+C
x2 X a x 24a5 X 2 2a5 x 2 a2

 
dx 1 1 x2
336. = 4 2 6 tan1 +C
x3 X 2a x 2a a2

Appendix C
 485
Integrals containing X = x4 a4

x
dx 1 xa 1
337. = 3 ln | | 3 tan1 +C
X 4a x+a 2a a


x dx 1 x 2 a2
338. = 2 ln | 2 |+C
X 4a x + a2


x2 dx 1 xa 1 x
339. = ln | |+ tan1 +C
X 4a x+a 2a a


x3 dx 1
340. = ln |X| + C
X 4


dx 1 X
341. = 4 ln | 4 | + C
xX 4a x

x
dx 1 1 xa 1
342. 2
= 4 + 5 ln | | + 5 tan1 +C
x X a x 4a x+a 2a a


dx 1 1 x 2 a2
343. 3
= 4 2 + 6 ln | 2 |+C
x X 2a x 4a x + a2

Miscelaneous algebraic integrals

dx 1 x+a
344. = tan1 +C
b2 + (x + a)2 b b


dx 1 x+a
345. = tanh1 +C
b2 (x + a)2 b b


dx 1 x+a
346. = coth1 +C
(x + a)2 b2 b b


dx x
347.  1
= 2 sin +C
x(a x) a


dx x
348.  = 2 sinh1 +C
x(a + x) a


dx 1 x
349.  = 2 cosh +C
x(x a) a


dx b+x
350. = 2 tan 1
+ C, a>x
(b + x)(a x) a x


dx xb
351. = 2 tan1 + C, a > x > b
(x b)(a x) ax

1 x+b


dx 2 tanh x+a + C1 , a>b
352. = 
(x + b)(x + a) 2 tanh1 x+a + C , a<b
x+b 2

Appendix C
486

 n
dx 1 a
353. = n sin1 +C
2n
x x a 2n na xn


x+a  x
354. dx = x2 a2 + a cosh1 + C
xa a


a+x x 
355. dx = a sin1 a2 x2 + C
ax a


ax a2  x  (x 2a) 
356. x dx = cos1 + a2 x2 + C, a>x
a+x 2 a 2


a+x a2 x x + 2a  2
357. x dx = sin1 a x2 + C
ax 2 a 2


x+b  b x
358. (x + a) dx = (x + a + b) x2 b2 + (2a + b) cosh1 + C
xb 2 b


dx 
359. = ln |x + a + 2ax + x2 | + C
2ax + x2
1  c 

 x ax2 + c + ln | ax + ax2 + c| + c,
a>0
2 2 a
360. ax2 + c dx =
 
1 x ax2 + c + c sin1 a
x + C, a<0
2 2 a c


1 + ax 1 1
361. dx = sin1 x 1 x2 + C
1 ax a a

 2 2

dx 1 1 (a + c )x + (ab + cd)
362. = tan + C, ad bc
= 0
(ax + b)2 + (cx + d)2 ad bc ad bc

 
dx 1  (a + c)x + (b + d) 
363. = ln   + C, ad bc
= 0
(ax + b)2 (cx + d)2 2(bc ad)  (a c)x + (b d) 

 2 2 2

x dx 1 1 (a + c )x + (ab + cd)
364. = tan + C, ad bc
= 0
(ax2 + b)2 + (cx2 + d)2 2(ad bc) ad bc

 
dx 1 1 x 1 x
365. 2 2 2 2
= 2 tan1 tan1 +C
(x + a )(x + b ) b a2 a a b b

 2 
(x2 + a2 )(x2 + b2 ) 1 (a c2 )(b2 c2 ) 1 x (a2 d2 )(b2 d2 ) 1 x
366. dx = x + tan tan +C
(x2 + c2 )(x2 + d2 ) d 2 c2 c c d d

       
ax2 + b 1 ad bc c 1 af be e
367. 2 2
dx = tan 1
x + tan1
x +C
(cx + d)(ex + f) cd ed fc d ef fc ed f
 2a2 x2 + 2ac + b2 bb2 + 4ac 

 
x dx 1
368. =

ln 

 + C, b2 + 4ac > 0
(ax2 + bx + c)2 + (ax2 bx + c)2 4b b2 + 4ac  2a2 x2 + 2ac + b2 + b b2 + 4ac 

Appendix C
 487

 
dx 1 1 x 1 x
369. 2 2 2 2
= 2 tan1 tan1 +C
(x + a )(x + b ) b a2 a a b b

 
(x2 + 2 )(x2 + 2 ) 1 (2 2 )( 2 2 ) x (2 2 )( 2 2 ) x
370. dx = x + 2 tan1 tan1 +C
(x2 + 2 )(x2 + 2 ) 2

   
x2 + 1 1  
371. 2 2
dx = tan1 x + tan1 x +C
(x + )(x + )   

 
dx 2x + a + b
372.  = cosh1 + C, a
= b
(x + a)(x + b) ab


dx xb
373.  = 2 tan1 +C
(x b)(a x) a x

 2 2

dx 1 1 ( + )x + ( + )
374. = tan +C
(x + )2 + (x + )2

 2 2 2 4 4

x dx 1 1 (a + b )x (a + b )
375.  = sin +C
(a2 + b2 x2 ) (a2 x2 )(x2 b2 ) 2ab (a2 b2 )(a2 + b2 x2 )

 
(x + b) dx 1 x 2 + c2 b 1 a x 2 + c2
376. = sin1
+ cosh +C
(x2 + a2 ) x2 + c2 a 2 c2 x 2 + a2 a a 2 c2 c x 2 + a2

 

px + q p 2 pb dx
377. dx = ln |ax + bx + c| + q
ax2 + bx + c 2a 2a ax2 + bx + c


( a x)2 2 3 1 2 x + a 2 1 2 x a
378. dx = tan tan +C
(a2 + ax + x2 ) x a 3a 3a 3a

 1  1 x
379. (a + x) a2 + x2 dx = (2x2 + 3ax + 2a2 ) a2 + x2 + a2 sinh1 + C
6 2 a


x 2 + a2 1 ax 3
380. dx = tan1 2 +C
x 4 + a2 x 2 + a4 a 3 a x2


x 2 a2 1 x2 ax + a2
381. dx = ln +C
x 4 + a2 x 2 + a4 2a3 x2 + ax + a2

Integrals containing sin ax

1
382. sin ax dx = cos ax + C
a


1 x
383. x sin ax dx = sin ax cos ax + C
a2 a

 
2 2 x2
384. 2
x sin ax dx = 2 x sin ax + cos ax + C
a a3 a

Appendix C
488
3x2 6x x3

   
6
385. x3 sin ax dx = 4 sin ax + cos ax + C
a2 a a a



1 n n(n 1)
386. xn sin ax dx = xn cos ax + 2 xn1 sin ax xn2 sin ax dx
a a a2

a3 x 3 a5 x 5 a7 x 7 (1)n x2n+1 x2n+1

sin ax
387. dx = ax + ++ +
x 3 3! 5 5! 7 7! (2n + 1) (2n + 1)!



sin ax 1 cos ax
388. 2
dx = sin ax + a dx
x a x

a2



sin ax a 1 sin ax
389. dx = cos ax sin ax dx
x3 2x 2x2 2 x



sin ax sin ax a cos ax
390. dx = + dx
xn (n 1)xn1 n1 xn1


dx 1
391. = ln | csc as cot ax| + C
sin ax a

a3 x 3 7a5 x5 2(22n1 1)Bn a2n+1 x2n+1

 
x dx 1
392. = 2 ax + + + + + +C
sin ax a 18 1800 (2n + 1)!
where Bn is the nth Bernoulli number B1 = 1/6, B2 = 1/30, . . . Note scaling and shifting

ax 7a3 x3 2(22n1 1)Bn a2n+1 x2n+1

dx 1
393. = + + ++ ++C
x sin ax ax 6 1080 (2n 1)(2n)!


x sin 2ax
394. sin2 ax dx = +C
2 4a

x2


x sin 2ax cos 2ax
395. x sin2 ax dx = +C
4 4a 8a2


1 1 1
396. x2 sin2 ax dx = cos 2ax + (3 6a2 x2 ) sin 2ax + C
6a 4a2 24a3

cos ax cos2 ax


397. sin3 ax dx = + +C
a 3a


1 1 3 3
398. x sin3 ax dx = x cos 3ax 2
sin 3ax x cos ax + 2 sin ax + C
12a 36a 4a 4a


3 sin 2ax sin 4ax
399. sin4 ax dx = x + +C
8 4a 32a


dx 1
400. = cot ax + C
sin2 ax a

Appendix C
 489


x dx x 1
401. 2 = cot ax + 2 ln | sin ax| + C
sin ax a a


dx cos ax 1 ax
402. 3 = 2 + ln | tan | + C
sin ax 2a sin ax 2a 2



dx cos ax n2 dx
403. = +
sinn ax (n 1)a sinn1 ax n 1 sinn2 ax


dx 1  ax 
404. = tan +C
1 sin ax a 4 2

 
dx 2 1 a tan(ax/2) 1
405. = tan + C, a>1
a sin ax a a2 1 a2 1


x dx x  ax  2  ax 
406. = tan + 2 ln | sin |+C
1 sin ax a 4 2 a 4 2


dx 1  ax 
407. = tan +C
1 + sin ax a 4 2

 
dx 2 1 + a tan(ax/2)
408. = tan1 + C, a>1
a + sin ax a a2 1 a2 1


x dx x  ax  2  ax 
409. = tan + 2 ln | sin |+C
1 + sin ax a 4 2 a 4 2


dx 1
410. 2 = tan1 ( 2 tan x) + C
1 + sin x 2


dx
411. = tan x + C
1 sin2 x


dx 1  ax  1 3
 ax 
412. = tan + tan +C
(1 sin ax)2 2a 4 2 6a 4 2


dx 1  ax  1  ax 
413. 2
= tan tan3 +C
(1 + sin ax) 2a 4 2 6a 4 2
2 1
 ax 
 tan tan + + C, 2 > 2
2

a 2 2






dx
1 tan ax +  2 2
414. = ln
2
+ C,

2 < 2
+ sin ax

2
a 2 ax

tan 2 + + 2 2



1 tan ax + C,

 
=
a 2 4
 
2 + 2


dx 1 1
415. = tan tan ax + C
2 + 2 sin2 ax

a 2 + 2

Appendix C
490

 
1 2 2
tan1 tan ax + C, 2 > 2





dx a 2 2

416. =
2 2 sin2 ax

1 2 2 tan ax +
2 < 2

ln  + C,



2a 2 2 2 2 tan ax




1 n1
417. sinn ax dx = sinn1 ax cos ax + sinn2 ax dx
an n



dx cos ax n2 dx
418. n = n1 + n2
sin ax (n 1)a sin ax n 1 sin ax



1 n
419. xn sin ax dx = xn cos ax + xn1 cos ax dx
a a


+ sin ax  ax 
420. dx = x + tan +C
1 sin ax a 4 2



+ sin ax b a dx
421. dx = x +
a + b sin ax b b a + b sin ax



dx x dx
422.
=
+ sin ax + sin ax

Integrals containing cos ax

1
423. cos ax dx = sin ax + C
a


1 x
424. x cos ax dx = 2
cos ax + sin ax + C
a a

x2

 
2x 2
425. x2 cos ax dx = cos ax + 3 sin ax + C
a2 a a



1 n n(n 1)
426. x cos ax dx = xn sin ax + 2 xn1 cos ax
n
xn2 cos ax dx
a a a2

a2 x 2 a4 x 4 a6 x 6 (1)n a2n x2n

cos ax
427. dx = ln |x| + ++ ++C
x 2 2! 4 4! 6 6! (2n) (2n)!



cos ax dx cos ax a sin ax
428. n
= n1
dx
x (n 1)x n1 xn1


dx 1
429. = ln | sec ax + tan ax| + C
cos ax a

1 a2 x 2 a4 x 4 5a6 x6 En a2n+2 x2n+2

 
x dx
430. = 2 + + + + + +C
cos ax a 2 4 2! 6 4! (2n + 2) (2n)!

a2 x 2 5a4 x4 En a2n x2n

dx
431. = ln |x| + + ++ ++C
x cos ax 4 96 2n(2n)!
where En is the nth Euler number E1 = 1, E2 = 5, E3 = 61, . . . Note scaling and shifting

Appendix C
 491


dx 1 ax
432. = tan +C
1 + cos ax a 2


dx 1 ax
433. = cot +C
1 cos ax a 2


ax
434. 1 cos ax dx = 2 2 cos +C
2


ax
435. 1 + cos ax dx = 2 2 sin +C
2


x sin 2ax
436. cos2 ax dx = + +C
2 4a


x2 1 1
437. x cos2 ax dx = + x sin 2ax + 2 cos 2ax + C
4 4a 8a


sin ax sin3 ax
438. cos3 ax dx = +C
a 3a


3 1 1
439. cos4 ax dx = x+ sin 2ax + sin 4ax + C
8 4a 32a


dx 1
440. = tan ax + C
cos2 ax a


x dx x 1
441. 2
= tan ax + 2 ln | cos ax| + C
cos ax a a


dx 1 sin ax 1 ax 
442. = + ln | tan + |+C
cos3 ax 2a cos2 ax 2a 4 2


dx 1 ax
443. = cot +C
1 cos ax a 2


x dx x ax 2 ax
444. = cot + 2 ln | sin | + C
1 cos ax a 2 a 2


dx 1 ax
445. = tan +C
1 + cos ax a 2


x dx x ax 2 ax
446. = tan + 2 ln | cos |+C
1 + cos ax a 2 a 2


dx 1
447. 2
= tan1 ( 2 cot ax) + C
1 + cos ax 2a


dx 1
448. = cot ax + C
1 cos2 ax a

Appendix C
492


dx 1 ax 1 ax
449. 2
= cot cot3 +C
(1 cos ax) 2a 2 6a 2


dx 1 ax 1 ax
450. 2
= tan + tan2 +C
(1 + cos ax) 2a 2 6a 2
 

2 1 ax


 2
tan tan + C, 2 > 2
dx a 2 + 2
451. =  
+ cos ax 1  + + tan ax 
2 


 ln 
 ax + C, 2 < 2
a 2 2 + tan 2 



dx x dx
452.
=
+ cos ax + cos ax



dx sin ax dx
453. = ,
=
( + cos ax)2 a( 2 2 )( + cos ax) 2 2 + cos ax

 
dx 1 tan ax
454. =  tan1  +C
2 + 2 cos2 ax a 2 + 2 2 + 2
 
1 tan ax




 tan1  + C, 2 > 2
dx a 2 2 2 2
455. = 
 tan ax  2 2 

2 2 cos2 ax 1  


 ln    + C, 2 < 2
2a 2 2  tan ax + 2 2 



dx sec (n2) ax tan ax n 2
456. = + secn2 ax dx + C
cosn ax (n 1)a n1

Integrals containing both sine and cosine functions

1
457. sin ax cos ax dx = sin2 ax + C
2a


dx 1
458. = ln | cot ax| + C
sin ax cos ax a


cos(a b)x cos(a + b)x
459. sin ax cos bx dx = + C, a
= b
2(a b) 2(a + b)


sin(a b)x sin(a + b)x
460. sin ax sin bx dx = +C
2(a b) 2(a + b)


sin(a b)x sin(a + b)x
461. cos ax cos bx dx = + +C
2(a b) 2(a + b)


sinn+1 ax
462. sinn ax cos ax dx = +C
(n + 1)a


cosn+1 ax
463. cosn ax sin ax dx = +C
(n + 1)a

Appendix C
 493


sin ax dx 1
464. = ln | sec ax| + C
cos ax a


cos ax dx 1
465. = ln | sin ax| + C
sin ax a

 
x sin ax dx 1 a3 x 3 a5 x 5 2a7 x7 22n (22n 1)Bn a2n+1 x2n+1
466. = 2 + + ++ +C
cos ax a 3 5 105 (2n + 1)!

 
x cos ax dx 1 a3 x 3 a5 x 5 22n Bn a2n+1 x2n+1
467. = 2 ax +C
sin ax a 9 225 (2n + 1)!


cos ax dx 1 ax a3 x3 22n Bn a2n1 x2n1
468. = + C
x sin ax ax 2 135 (2n 1)(2n)!


sin ax a3 x 3 2a5 x5 22n (22n 1)Bn a2n1 x2n1
469. dx = ax + + ++ + + C
x cos ax 9 75 (2n 1)(2n)!


sin2 ax 1
470. 2
dx = tan ax x + C
cos ax a


cos2 ax 1
471. dx = cot ax x + C
sin2 ax a


x sin2 ax 1 1 1
472. dx = x tan ax + 2 ln | cos ax| x2 + C
cos2 ax a a 2


x cos2 ax 1 1 1
473. 2 dx = x cot ax + 2 ln | sin ax| x2 + C
sin ax a a 2


cos ax 1
474. dx = ln | sin ax| + C
sin ax a


sin3 ax 1 1
475. dx = tan2 ax + ln | cos ax| + C
cos3 ax 2a a


cos3 ax 1 1
476. dx = cot2 ax ln | sin ax| + C
sin3 ax 2a a


x 1
477. sin(ax + b) sin(ax + ) dx = cos(b ) sin(2ax + b + ) + C
2 4a


x 1
478. sin(ax + b) cos(ax + ) dx = sin(b ) cos(2ax + b + ) + C
2 4a


x 1
479. cos(ax + b) cos(ax + ) dx = cos(b ) + sin(2ax + b + ) + C
2 4a
x sin 2ax sin 2bx sin 2(a b)x sin 2(a + b)x


+ + C, b
= a
4 8a 8b 16(a b) 16(a + b)
480. sin2 ax cos2 bx dx =
x sin 4ax

+ C, b=a
8 32a

Appendix C
494


dx 1
481. = ln | tan ax| + C
sin ax cos ax a


dx 1 ax  1
482. 2 = ln | tan + | +C
sin ax cos ax a 4 2 a sin ax


dx 1 ax 1
483. = ln | tan | + +C
sin ax cos2 ax a 2 a cos ax


dx 2 cos 2ax
484. = +C
sin2 ax cos2 ax a


sin2 ax sin ax 1  ax 
485. dx = + ln | tan + |+C
cos ax a a 2 4


cos2 ax cos ax 1 ax
486. dx = + ln | tan | + C
sin ax a a 2

  ax 
dx 1  cos + sin ax 
487. = 1 + (1 + sin ax) ln  2
ax
2 
ax  + C
cos ax (1 + sin ax) 2a(1 + sin ax) cos 2 sin 2


dx 1 ax 1 ax
488. = sec2 + ln | tan | + C
sin ax (1 + cos ax) 4a 2 2a 2



dx 1 ax dx
489. = ln | tan |
sin ax ( + sin ax) a 2 + sin ax

   
dx 1 ax   + sin ax 
490. = 2 ln | tan + | ln  + C,
=
cos ax ( + sin ax 2 a 4 2 cos ax 

 
dx 1 ax + cos ax
491. = 2 ln | tan | + ln | | + C,
=
sin ax ( + cos ax) 2 a 2 a sin ax



dx 1 ax  dx
492. = ln | tan + |
cos ax ( + cos ax) a 4 2 + cos ax
 
2 1 + ( ) tan ax 2 2 > 2 + 2
tan + C,



a R R R = 2 + 2 2
 R + ( ) tan ax 

 


1  2 


ln   + C,
ax 
2 < 2 + 2




a R  + R + ( ) tan 2
dx 
493. = 1  ax 
+ cos ax + sin ax ln  + tan  + C, =

a 2
 
cos ax ax

1 2 + sin 2

 


ln 
ax ax  + C,
 =


a ( + ) cos 2 + ( ) sin 2



1 ax
ln |1 + tan | + C, ==

a  ax  2
dx 1
494. = ln | tan |+C
sin ax cos ax 2a 2 8

Appendix C
 495


sin ax dx x
495. = ln | sin ax cos ax| + C
sin ax cos ax 2


cos ax dx x 1
496. = + ln | sin axx cos ax| + C
sin ax cos ax 2 2a


sin ax dx 1
497. = ln | + sin ax| + C
+ sin ax a


cos ax dx 1
498. = ln | + sin ax| + C
+ sin ax a


sin ax cos ax dx 1
499. = ln |2 cos2 ax + 2 sin2 ax| + C,
=
2 cos2 ax + 2 ax 2a( 2 2 )

 
dx 1
500. = tan 1
tan ax + C
2 sin2 ax + 2 cos2 ax a

 
dx 1  tan ax 
501. = ln  +C
2 sin2 ax 2 cos2 ax 2a  tan ax + 


sinn ax tann+1 ax
502. (n+2
dx = +C
cos ax (n + 1)a


cosn ax cot(n+1) ax
503. dx = +C
sin(n+2) ax (n + 1)a


dx x
504. = 2 2
+ ln | sin ax + cos ax| + C
sin ax
+ cos + a( + 2 )
2
ax


dx x
505. = 2 2
2
ln | sin ax + cos ax| + C
+ cos ax + a( + 2 )
sin ax

n

cos ax cot(n1) ax
506. dx = cot(n2) ax dx
sinn ax (n 1)a



sinn ax tann1 ax sinn2 ax
507. dx = dx
cosn ax (n 1)a cosn2 ax


sin ax 1
508. (n+1)
dx = sec n ax + C
cos ax na


sin x + cos x [( + )x + ( ) ln | sin x + cos x|]
509. dx = +C
sin x + cos x 2 + 2

2 1 ab x
2


tan tan ln |a + b cos x| + C, a>b
+ sin x a b2 a+b 2 b
510. dx =
a + b cos x
2 ba x

tanh1 tan ln |a + b cos x| + C, a<b
2
b a 2 b + a 2 b

Appendix C
496
 
1 1 a


dx
tan tan x + C, a>b
511. = a a2 b 2 a2 b 2
a2 b2 cos2 x 
1 tanh1 a tan x + C,

a b2 a2 b2a2
b>a

 
dx 1 b
512. = 2 tan x tan1 +C
(a cos x + b sin x)2 a + b2 a
 
1 1 a(a cos2 x + 2b cos x + c)


sin + C, b2 > ac, a < 0


a 2
b ac

 



sin x dx 1 1 a(a cos2 x + 2b cos x + c)
513. = sinh
a
+ C, b2 > ac, a > 0
a cos2 x + 2b cos x + c b 2 ac

 
2 x + 2b cos x + c)

1 1 a(a cos
a cosh

+ C, b2 < ac, a > 0
ac b2

2
a(a sin x + 2b sin x + c)
1 sin1



+ C, b2 > ac, a < 0


a b 2 ac





2
cos x dx
1 a(a sin x + 2b sin x + c)
514.  = sinh1 + C, b2 > ac, a > 0
a sin2 x + 2b sin x + c
a b 2 ac





2


1 a(a sin x + 2b sin x + c)


cosh1 + C, b2 < ac, a > 0
a ac b2

Integrals containing tan ax, cot ax, sec ax, csc ax

Write integrals in terms of sin ax and cos ax and see previous listings.
Integrals containing inverse trigonmetric functions


x x 
515. sin1 dx = x sin1 + a2 x2 + C

a a
1 x 1 x

516. cos dx = x cos a2 x 2 + C

a a
1 x 1 x a
517. tan dx = x tan ln |x2 + a2 | + C
a a 2


x x a
518. cot1 dx = x cot1 +ln |x2 + a2 | + C
a a 2
x 


x x sec1 a ln |x + x2 a2 | + C, 0 < sec 1 x
a < /2
519. sec 1 dx = a
a x
x sec1

x
+ a ln |x + x2 a2 + C, /2 < sec1 a <
a
x 


x x csc1 + a ln |x + x2 a2 | + C, 0 < csc 1 x
a < /2
520. csc 1
dx = a
a x
x csc1

x
a ln |x + x2 a2 | + C, /2 < csc 1 a <0
a

 2 
x x a2 x 1 
521. x sin1 dx = sin1 + x a2 x2 + C
a 2 4 a 4

 
x x2 a2 x 1  2
522. x cos1 dx = cos1 x a x2 + C
a 2 4 a 4

Appendix C
 497


x 1 x a
523. x tan1 dx = (x2 + a2 ) tan1 ln |x2 + a2 | + C
a 2 a 2


x 1 x a
524. x cot1 dx = (x2 + a2 ) cot1 + x + C
a 2 a 2

1 x a 2


x x2 sec1
x a2 + C, 0 < sec1 xa < /2
525. x sec 1
dx = 2 a 2
1 x2 sec1 x + a x2 a2 + C,
a 
/2 < sec1 xa <
2 a 2
1 x a

x2 csc1 +
x2 a2 + C, 0 < csc1 xa < /2
1 x 2 a 2
526. x csc dx =
1 x2 csc1 x a x2 a2 + C,
a 
/2 < csc 1 xa < 0
2 a 2


x 1 x 1 
527. x2 sin1 dx = x3 sin1 + (x2 + 2a2 ) a2 x2 + C
a 3 a 9


x 1 x 1 
528. x2 cos1 dx = x3 cos1 (x2 + 2a2 ) a2 x2 + C
a 3 a 9


x 1 x a a3
529. x2 tan1 dx = tan1 x2 + ln |x2 + a2 | + C
a 3 a 6 6


x 1 x a a3
530. x2 cot1 dx = cot1 + x2 ln |a2 + x2 | + C
a 3 a 6 6

1 3 1 x a  2 a3 
x


x

x sec x x a 2 ln |x + x2 a2 | + C, 0 < sec 1 a < /2
531. 2
x sec 1
dx = 3 a 6 6
a 1 3
x a  a3 
x sec1 + x x2 a2 + ln |x + x2 a2 | + c, /2 < sec1 x
a <
3 a 6 6

1 x a  2 a3 


x x3 csc1
+ x x a2 + ln |x + x2 a2 | + C, 0 < csc 1 x
a
< /2
532. 2
x csc 1
dx = 3 a 6 6
a 1 x3 csc1
x a  2 a3 
x
x x a2 ln |x + x2 a2 | + C, /2 < csc 1 a
<0
3 a 6 6

1 x x 1  x 3 13  x 5 135
533. sin1 dx = + + + + + C
x a a 233 a 2455 a 24677



1 x 1 x
534. cos1 dx = ln |x| + sin1 dx
x a 2 x a


1 x x 1  x 3 1  x 5 1  x 7
535. tan1 dx = 2 + 2 2 + + C
x a a 3 a 5 a 7 a



1 x 1 x
536. cot1 dx = ln |x| tan1 dx
x a 2 x a


1 x a 1  x 3 1 3  x 5 1 3 5  x 7
537. sec1 dx = ln |x| + + + + + + C
x a 2 x 233 a 2455 a 24677 a

Appendix C
498

 
1 x a 1  x 3 13  x 5 135  x 7
538. csc1 dx = + + + + +C
x a x 233 a 2455 a 24677 a


1 x 1 x 1 a + a2 x 2
539. sin1 dx = sin1 ln | |+C
x2 a x a a a


1 1 x 1 1 x 1 a + a2 x 2
540. cos dx = cos + ln | |+C
x2 a x a a a


1 x 1 x 1 x 2 + a2
541. 2
tan1 dx = tan1 ln | |+C
x a x a 2a a2



1 1 x 1 1 x 1 1 x
542. 2
cot dx = cot + tan1 dx
x a x a 2a x a

1 x 1 


1 x sec1 +
x2 a2 + C, 0 < sec 1 x
a < /2
543. sec 1
dx = x a ax
x2 1 sec1 x 1 x2 a2 + C,
a 
x
/2 < sec 1 a <
x a ax 
1 1 x 1 x


1 x

csc x2 a2 + C, 0 < csc 1 a
< /2
544. csc 1
dx = x a ax
x2 1 csc1 x + 1 x2 a2 + C,
a 
x
/2 < csc1 a
<0

x a ax
x x
545. sin 1
dx = (a + x) tan 1
ax + C
a+x a


x x
546. cos1 dx = (2a + x) tan1 2ax + C
a+x 2a

Integrals containing the exponential function

1
547. eax dx = eax + C
a

 
x 1
548. xeax dx = eax + C
a a2

 
2 ax x2 2x 2
549. x e dx = 2 + 3 eax + C
a a a



1 n ax n
550. xn eax dx = x e xn1 eax dx
a a


1 ax ax (ax)2 (ax)3
551. e dx = ln |x| + + + ++C
x 1 1! 2 2! 3 3!



1 ax 1 a 1 ax
552. e dx = eax + e dx
xn (n 1)xn1 n1 xn1


eax 1
553. dx = ln | + eax | + C
+ eax a

Appendix C
 499

 
a sin bx b cos bx
554. eax sin bx dx = eax + C
a2 + b 2

 
a cos bx + b sin bx
555. eax cos bx dx = eax + C
a2 + b 2

 

a sin bx nb cos bx n(n 1)b2
556. e ax n
sin bx dx = ax
e n1
sin bx + 2 eax sinn2 bx dx
a2 + n 2 b 2 a + n2 b 2

 

a cos bx + nb sin bx n(n 1)b2
557. e ax n
cos bx dx = e ax
cos n1
bx + 2 eax cosn2 bx dx
a2 + n 2 b 2 a + n2 b 2

Another way to express the above integrals is to define

Cn = eax cosn bx dx and Sn = eax sinn bx dx, then one can write the reduction formulas

a cos bx + nb sin bx ax n(n 1)b2

Cn = 2 2 2
e cosn1 bx + 2 Cn2
a +n b a + n2 b 2
a sin bx nb cos bx ax n1 n(n 1)b2
Sn = e sin bx + Sn2
a2 + n 2 b 2 a2 + n 2 b 2

 
[2ab b(a2 + b2 )x] cos bx + [a(a2 + b2 )x a2 + b2 ] sin bx
558. ax
xe sin bx dx = eax + C
(a2 + b2 )2

 
[a(a2 + b2 )x a2 + b2 ] cos bx + [b(a2 + b2 )x 2ab] sin bx
559. xe ax
cos bx dx = eax + C
(a2 + b2 )2



1 ax 1 1 ax
560. eax ln x dx = e ln x e dx
a a x

 
a sinh bx b cosh bx ax
561. eax sinh bx dx = e + C, a
= b

(a b)(a + b)
1 2ax x
562. eax sinh ax dx = e +C

4a
 2 
a cosh bx b sinh bx ax
563. ax
e cosh bx dx = e + C, a
= b

(a b)(a + b)
1 2ax x
564. eax cosh ax dx = e + +C

4a 2
dx x 1
565. = ln | + eax | + C
+ eax a


dx x 1 1
566. = 2+ ln | + eax | + C
( + eax )2 a( + eax ) a2
 
1 1 ax


tan e + C, > 0


dx a
567. = 
 eax / 

eax + eax
1  

ln    + C, < 0
2a  eax + / 

Appendix C
500

 
a2 + 4b2 a2 cos(2bx) 2ab sin(2bx)
568. eax sin2 bx dx = eax + C
2a(a2 + 4b2

 
a2 + 4b2 + a2 cos(2bx) + 2ab sin(2bx)
569. eax cos2 bx dx = eax + C
2a(a2 + 4b2
Integrals containing the logarithmic function


570. ln x dx = x ln |x| + C

1 2 1
571. x ln x dx = x ln |x| x2 + C
2 4


1 1
572. xn ln x dx = xn+1 + xn+1 ln |x| + C, n
= 1
(n + 1)2 n+1


1 1
573. ln x dx = (ln |x|)2 + C
x 2


dx
574. = ln | ln |x|| + C
x ln x


1 1 1
575. 2
ln x dx = ln |x| + C
x x x


576. (ln |x|)2 dx = x(ln |x|)2 2x ln |x| + 2x + C

1 n 1 n+1
577. (ln |x|) dx = (ln |x|) + C, n
= 1
x n+1



n n
578. (ln |x|) dx = x(ln |x|) n (ln |x|)n1 dx

x
579. ln |x2 + a2 | dx = x ln |x2 + a2 | 2x + 2a tan1 +C
a


x+a
580. ln |x2 a2 | dx = x ln |x2 a2 | 2x + a ln | |+C
xa


2 (ax + b)2 (b a)2 a 1
581. (ax + b) ln(x + ) dx = ln(x + ) 2 (x + )2 (b a)x + C
2a 2 4


582. (ln ax)2 dx = x(ln ax)2 2x ln ax + 2x + C

Integrals containing the hyperbolic function sinh ax

1
583. sinh ax dx = cosh ax + C
a

Appendix C
 501


1 1
584. x sinh ax dx = x cosh ax 2 sinh ax + C
a a

 
x2 2 2x
585. x2 sinh ax dx = + 3 cosh ax sinh ax + C
a a a2



1 n n
586. xn sinh ax dx = x cosh ax xn1 cosh ax dx
a a


1 (ax)3 (ax)5
587. sinh ax dx = ax + + ++C
x 3 3! 4 5!



1 1 1
588. sinh ax dx = sinh ax + a cosh ax dx
x2 x x



1 sinh ax a 1
589. n
sinh ax dx = n1
+ cosh ax dx
x (n 1)x n1 xn1


dx 1 ax
590. = ln | tanh | + C
sinh ax a 2

 
x dx 1 (ax)3 2(22n 1)Bn a2n+1 x2n+1
591. = 2 ax + frac7(ax)5 1800 + + (1)n + +C
sinh ax a 18 (2n + 1)!


1 1
592. sinh2 ax dx = x sinh 2ax x + C
2a 2



n 1 n1
593. sinh ax dx = sinhn1 ax cosh ax sinhn2 ax dx
na n


1 1 1
594. x sinh2 ax dx = x sinh 2ax 2 cosh 2ax x2 + C
4a 8a 4


dx 1
595. = coth ax + C
sinh2 ax a


dx 1 1 ax
596. = csch ax coth ax ln | tanh | + C
sinh3 ax 2a 2a 2


x dx 1 1
597. 2 = x coth ax + 2 ln | sinh ax| + C
sinh ax a a


1 1
598. sinh ax sinh bx dx = sinh(a + b)x sinh(a b)x + C
2(a + b) 2(a b)


1
599. sinh ax sin bx dx = [a cosh ax sin bx b sinh ax cos bx] + C
a2 + b2


1
600. sinh ax cos bx dx = [a cosh ax cos bx + b sinh ax sin bx] + C
a2 + b 2

Appendix C
502
 


dx 1  eax + 2 + 2 
601. = 

ln  

+C
+ sinh ax 2
+ 2  ax
e + + +  2 2

dx cosh ax dx
602. 2
= 2 2
+ 2 2
( + sinh ax) a( + ) + sinh ax + + sinh ax
 
1 1 2 2 tanh ax


 tan + C, 2 > 2


dx a 2 2
603. = 
 + 2 2 tanh ax 

2 + 2 sinh2 ax 1  


 ln    + C, 2 < 2
2a 2 2  2 2
tanh ax 
 


dx 1  + 2 + 2 tanh ax 
604. =  ln

 

+C
2 2 sinh2 ax 2a + 2 2  2 2
+ tanh ax 

Integrals containing the hyperbolic function cosh ax

1
605. cosh ax dx = sinh ax + C

a
1 1
606. x cosh ax dx = x sinh ax 2 cosh ax + C
a a

 
2 x2 2
607. x2 cosh ax dx = x cosh ax + + 3 sinh ax + C
a2 a a



1 n n
608. xn cosh ax dx = x sinh ax xn1 sinh ax dx
a a


1 (ax)2 (ax)4 (ax)6
609. cosh ax dx = ln |x| + + + + + C
x 2 2! 4 4! 6 6!



1 1 1
610. cosh ax dx = cosh ax + a sinh ax dx
x2 x x



1 1 cosh ax a sinh ax
611. cosh ax dx = + dx, n>1
xn n 1 xn1 n1 xn1


dx 2
612. = tan1 eax + C
cosh ax a

 
x dx 1 a2 x 2 a4 x 4 5a6 x6 En a2n+2 x2n+2
613. = 2 + + + (1)n + +C
cosh ax a 2 8 144 (2n + 2) (2n)!


1 1
614. cosh2 ax dx = x + sinh ax cosh ax + C
2 2



1 n1
615. coshn ax dx = coshn1 ax sinh ax + coshn2 ax dx
na n


1 2 1 1
616. x cosh2 ax dx = x + x sinh 2ax 2 cosh 2ax + C
4 4a 8a

Appendix C
 503


dx 1
617. 2 = tanh ax + C
cosh ax a


x dx 1 1
618. 2 = x tanh ax 2 ln | cosh ax| + C
cosh ax a a



dx 1 x sinh ax n2 dx
619. n = n1 + n2
cosh ax (n 1)a cosh ax n 1 cosh ax, n > 1


1 1
620. cosh ax cosh bx dx = sinh(a b)x + sinh(a + b)x + C
2(a b) 2(a + b)


1
621. cosh ax sin bx dx = [a sinh ax sin bx b cosh ax cos bx] + C
a2 + b2


1
622. cosh ax cos bx dx = [a sinh ax cos bx + b cosh ax sin bx] + C
a2 + b2
ax
2 1 e +


2 2 tan  + C, 2 > 2


dx 2
2
623. =  
 eax + 2 2 
+ cosh ax  1  

ln    + C, 2 < 2
a 2 2  eax + + 2 2 


dx 1
624. = tanh ax + C
1 + cosh ax a


x dx x ax 2 ax
625. = tanh 2 ln | cosh |+C
1 + cosh ax a 2 a 2


dx 1 ax
626. = coth +C
1 + cosh ax a 2



dx sinh ax dx
627. =
( + cosh ax)2 a( 2 2 )( + cosh ax) 2 2 + cosh ax
 

1  tanh ax + 2 2 
 



 ln    + C, 2 > 2
dx 2a 2 2  2
tanh ax  2
628. =
2 2 cosh2 ax 1 tanh ax

 tan1  + C, 2 < 2
a 2 2 2 2

 
dx 1 1 tanh ax
629. =  tanh  +C
2 + 2 cosh2 ax a 2 + 2 2 + 2

Integrals containing the hyperbolic functions sinh ax and cosh ax

1
630. sinh ax cosh ax dx = sinh2 ax + C
2a


1 1
631. sinh ax cosh bx dx = cosh(a + b)x + cosh(a b)x + C
2(a + b) 2(a b)

Appendix C
504


1 1
632. sinh2 ax cosh2 ax dx = sinh 4ax x + C
32a 8


1
633. sinhn ax cosh ax dx = sinhn+1 ax + C, n
= 1
(n + 1)a


1
634. coshn ax sinh ax dx = coshn+1 ax + C, n
= 1
(n + 1)a


sinh ax 1
635. dx = ln | cosh ax| + C
cosh ax a


cosh ax 1
636. dx = ln | sinh ax| + C
sinh ax a


dx 1
637. = ln | tanh ax| + C
sinh ax cosh ax a

 
x sinh ax 1 a3 x 3 a5 x 5 22n (22n 1)Bn a2n+1 x2n+1
638. dx = 2 + + (1)n + +C
cosh ax a 3 15 (2n + 1)!

 
x cosh ax 1 a3 x 3 a5 x 5 22n Bn a2n+1 x2n+1
639. dx = 2 ax + + + (1)n1 + +C
sinh ax a 9 225 (2n + 1)!


sinh2 ax 1
640. 2 dx = x tanh ax + C
cosh ax a


cosh2 ax 1
641. 2 dx = x coth ax + C
sinh ax a


x sinh2 ax 1 1 1
642. 2
dx = x2 x tanh ax + 2 ln | cosh ax| + C
cosh ax 2 a a


x cosh2 ax 1 1 1
643. dx = x2 x coth ax + 2 ln | sinh ax| + C
sinh2 ax 2 a a


sinh ax a3 x 3 22n (22n 1)Bn a2n1 x2n1
644. dx = ax + + (1)n1 + + C
x cosh ax 9 (2n 1)(2n)!


cosh ax 1 ax a3 x3 22n Bn a2n1 x2n1
645. dx = + + + (1)n + + C
x sinh ax ax 3 135 (2n 1)(2n)!


sinh3 ax 1 1
646. 3 dx = ln | cosh ax| tanh2 ax + C
cosh ax a 2a


cosh3 ax 1 1
647. 3
dx = ln | sinh ax| coth2 ax + C
sinh ax a 2a


dx 1 1 ax
648. = sech ax + ln tanh | + C
sinh ax cosh2 ax a a 2

Appendix C
 505


dx 1 1
649. 2 = tan1 (sinh ax) csch ax + C
sinh ax cosh ax a a


dx 2
650. 2 2 = coth ax + C
sinh ax cosh ax a


sinh2 ax 1 1
651. dx = sinh ax tan1 (sinh ax) + C
cosh ax a a


cosh2 ax 1 1 ax
652. dx = cosh ax + ln | tanh | + C
sinh ax a a 2

 
dx 1  1 + sinh ax  1
653. = ln   + tan1 eax + C
cosh ax (1 + sinh ax) 2a  cosh ax  a


dx 1 ax 1
654. = ln | tanh | + +C
sinh ax (cosh ax + 1) 2a 2 2a(cosh ax + 1)


dx 1 ax 1
655. = ln | tanh | +C
sinh ax (cosh ax 1) 2a 2 2a(cosh ax 1)


dx x
656. sinh ax
= 2 2
2 2 )
ln | sinh ax + cosh ax| + C
+ a(
cosh ax


dx x
657. cosh ax
= 2 2
+ 2 2 )
ln | sinh ax + cosh ax| + C
+ a(
sinh ax  

1 1 b cosh ax + c sinh ax

2

sec + C, b2 > c2
dx a b c2 b 2 c2
658. =  
b cosh ax + c sinh ax 1 1 b cosh ax + c sinh ax

csch + C, b2 < c2
a c2 b 2 c2 b 2
Integrals containing the hyperbolic functions tanh ax, coth ax, sech ax, csch ax

Express integrals in terms of sinh ax and cosh ax and see previous listings.

Integrals containing inverse hyperbolic functions

x x 
659. sinh1 dx = x sinh1 x2 + a2 + C
a  a 


1 x x cosh (x/a) x2 a2 , cosh1 (x/a) > 0
1
660. cosh dx = 
a x cosh1
(x/a) + x2 a2 , cosh1 (x/a) < 0


x x a
661. tanh1 dx = x tanh1 + ln |a2 x2 | + C
a a 2


x x a
662. coth1 dx = x coth1 + ln |x2 a2 | + C
a a 2
x x


x xsech 1 + a sin1 + C, sech 1 (x/a) > 0
663. sech 1 dx = a a
a xsech 1 x a sin1 x + C, sech 1 (x/a) < 0
a a

Appendix C
506


x x x
664. csch 1 dx = xcsch 1 a sinh1 , + for x > 0 and for x < 0
a a a

 
x x2 a2 x 1 
665. x sinh1 dx = + sinh1 xx x2 + a2 + C
a 2 4 a 4

1 x 1 


x (2x2 a2 ) cosh1 x x2 a2 + C,
cosh1 (x/a) > 0
666. x cosh1 dx = 4 a 4
1 (2x2 a2 ) cosh1 x + 1 x x2 a2 + C,
a 
cosh1 (x/a) < 0
4 a 4


1 x ax 1 2 1 x
667. x tanh dx = + (x a2 ) tanh +C
a 2 2 a


x ax 1 2 x
668. x coth1 dx = + (x a2 ) coth1 + C
a 2 2 a

1 2 1 x 1  2


x sech a a x2 , sech 1 (x/a) > 0
1 x 2 a 2
669. xsech dx =
1 xsech 1 x + 1 a a2 x2 + C,
a 
sech 1 (x/a) < 0
2 a 2


1 x 1 2 1 x a 2
670. xcsch dx = x csch x + a2 + C, + for x > 0 and for x < 0
a 2 a 2


x 1 x 1 
671. x2 sinh1dx = x3 sinh1 + (2a2 x2 ) x2 + a2 + C
a 3 a 9

1 3 1 x 1 2 


x

x cosh (x + 2a 2
) x2 a2 + C, cosh1 (x/a) > 0
672. x2 cosh1 dx = 3 a 9
1 x3 cosh1 x + 1 (x2 + 2a2 ) x2 a2 + C,
a 
cosh1 (x/a) < 0
3 a 9


x a 1 x 1
673. x2 tanh1 dx = x2 + x3 tanh1 + a3 ln |a2 x2 | + C
a 6 3 a 6


x a 1 x 1
674. x2 coth1 dx = x2 + x3 coth1 + a3 ln |x2 a2 | + C
a 6 3 a 6



x 1 x 1 x3 dx
675. x2 sech 1 dx = x3 sech 1
a 3 a 3 x 2 + a2



1 x 1 x a x2 dx
676. 2
x csch dx = x3 csch 1
a 3 a 3 x 2 + a2



n x 1 1 n+1 1 x 1 xn+1 dx
677. x sinh dx = x sinh
a n+1 a n+1 x 2 a2


1 1 x 1 xn+1



x n+1
cosh , cosh1 (x/a) > 0
n 1 x n+1 a n+1 x 2 a2
678. x cosh dx =

a
1 xn+1 cosh1 x + 1
xn+1 dx
, cosh1 (x/a) < 0
n+1 a n+1 x 2 a2


n+1
x 1 x a x dx
679. xn tanh1 dx = xn+1 tanh1
a n+1 a n+1 a2 x 2

Appendix C
 507


n+1
x 1 x a x dx
680. xn coth1 dx = xn+1 coth1
a n+1 a n+1 a2 x 2


1 n+1 1 x a xn dx



x sech + , sech 1 (x/a) > 0
x n + 1 a n + 1 a 2 x2
681. xn sech 1 dx =
1 x a


xn dx
a

xn+1 sech 1 , sech 1 (x/a) < 0
n+1 a n+1 a2 x 2



x 1 x a xn dx
682. xn csch 1 dx = xn+1 csch 1 , + for x > 0, for x < 0
a n+1 a n+1 x 2 + a2

x (x/a)3 1 3(x/a)5 1 3 5(x/a)7

+ + + C, |x| > a


a 233 2445 24677

  2
1 x 1 2x (a/x)2 1 3(a/x)4 1 3 5(a/x)6
683. sinh1 dx = ln | | + + + C, x>a
x a
2 a 222 2444 24666

 2


1 2x (a/x)2 1 3(a/x)4 1 3 5(a/x)6
ln | | + + + + C, x < a
2 a 222 2444 24666

  2 
1 1 x 1 2x (a/x)2 1 3(a/x)4 1 3 5(a/x)6
684. cosh dx = ln | | + + + + +C
x a 2 a 222 2444 24666
+ for cosh1 (x/a) > 0, for cosh1(x/a) < 0


1 x x (x/a)3 (x/a)5
685. tanh1 dx = + + ++C
x a a 32 52


1 x ax 1 2 x
686. coth1 dx = + (x a2 ) coth1 + C
x a 2 2 a

1 a 4a (x/a)2 1 3(x/a)4


1 x

ln | | ln | | + C, sech 1 (x/a) > 0
687. sech 1 dx = 2 x x 222 2444
x a 2 4
1 ln | a | ln | 4a | + (x/a) + 1 3(x/a) + ,

sech 1 (x/a) < 0
2 x x 222 2444
1 x 4a (x/a)2 1 3(x/a)4

ln | | ln | | + + + C, 0<x<a




2 a x 222 2444

1 x 1 x x (x/a)2 1 3(x/a)4
688. csch 1 dx = ln | | ln | | + , a < x < 0
x a
2 a 4a 222 2444

3 5
a + (a/x) 1 3(a/x) + + C,


|x| > a
x 233 2455

Integrals evaluated by reduction formula

1 n1
689. If Sn = sinn x dx, then Sn = sinn1 x cos x + Sn2
n n


1 n1
690. If Cn = cosn x dx, then Cn = sin x cosn1 x + Cn2
n n


sinn ax 1
691. If In = dx, then In = sinn1 ax + In2

cos ax (n 1) a
cosn ax 1
692. If In = dx, then In = cosn1 ax + In2
sin ax (n 1) a

Appendix C
508



m
693. If Sm = x sin nx dx and Cm = xm cos nx dx, then

1 m m 1 m
Sm = x cos nx + Cm1 and Cm = xm sin nx Sm1
n n n n



1
694. If I1 = tan x dx, and In = tann x dx, then In = tann1 x In2 , n = 2, 3, 4, . . .
n1


sinn ax sinn1 ax
695. If In = dx, then In = + In2
cos ax (n 1)a


cosn ax cosn1 ax
696. If In = dx, then In = + In2
sin ax (n 1)a


697. If In,m = sinn x cosm x dx, then

1 n1
In,m = sinn1 x cosm+1 x + In2,m
n+m n+m
1 n+m+2
In,m = sinn+1 x cosm+1 x + In+2,m
n+1 n+1
1 m1
In,m = sinn+1 x cosm1 x + In,m+2
n+m n+m
1 n+m+2
In,m = sinn+1 x cosm+1 x + In,m+2
m+1 m+1
1 n1
In,m = sinn1 x cosm+1 x + In2,m+2
m+1 m+1
1 m1
In,m = sinn+1 x cosm1 x + In+2,m2
n+1 n+1



698. If Sn = eax sinn bx dx and Cn = eax cosn bx dx, then
 
a cos bx + nb sin bx n(n 1) b2
Cn =eax cosn1 bx 2 2 2
+ 2 Cn2
a +n b a + n2 b 2
 
a sin bx nb cos nx n(n 1) b2
Sn =eax sinn1 ax 2 2 2
+ 2 Sn2
a +n b a + n2 b 2


1 n
699. If In = xm (ln x)n dx, then In = xm+1 (ln x)n In1
m+1 m+1

Integrals involving Bessel functions

700. J1 (x) dx = J0 (x) + C

701. xJ1 (x) dx = xJ0 (x) + J0 (x) dx

702. xn J1 (x) dx = xn J0 (x) + n xn1J0 (x) dx

Appendix C
 509



J1 (x)
703. dx = J1 (x) + J0 (x) dx
x


704. x J1 (x) dx = x J (x) + C

705. x J+1 (x) dx = x J (x) + C

J1 (x) 1 J1 (x) 1 J0 (x)
706. dx = + dx
xn n xn1 n xn1


707. xJ0 (x) dx = xJ1 (x) + C

708. x2 J0 (x) dx = x2 J1 (x) + xJ0 (x) J0 (x) dx

709. xn J0 (x) dx = xn J1 (x) + (n 1)xn1 J0 (x) (n 1)2 xn2J0 (x) dx

J0 (x) J1 (x) J0 (x) 1 J0 (x)
710. dx = dx
xn (n 1)2 xn2 (n 1)xn1 (n 1)2 xn2



711. Jn+1 (x) dx = Jn1 (x) dx 2Jn (x)

x
712. xJn (x)Jn (x) dx = [Jn
(x)Jn (x) Jn
(x)Jn (x)] + C
2 2


713. If Im,n = xm Jn (x) dx, m n, then

Im,n = xm Jn1 (x) + (m + n 1) Im1,n1

714. If In,0 = xn J0 (x) dx, then In,0 = xnJ1 (x) + (n 1)xn1J0 (x) (n 1)2 In2,0 Note that



I1,0 = xJ0 (x) dx = xJ1 (x) + C and I0,1 = J1 (x) dx = J0 (x) + C Note also that the integral


I0,0 = J0 (x) dx cannot be given in closed form.

Appendix C
510
Definite integrals
General integration properties

b
dF (x) b
1. If = f(x), then f(x) dx = F (x)|a = F (b) F (a)
dx a
2.




b b
f(x) dx = lim f(x) dx, f(x) dx = lim f(x) dx
b b
0 0 a a

b
b
3. If f(x) has a singular point at x = b, then f(x) dx = lim
0
f(x) dx
a a

b
b
4. If f(x) has a singular point at x = a, then f(x) dx = lim
0
f(x) dx
a a+

b
c
b
5. If f(x) has a singular point at x = c, a < c < b , then f(x) dx = f(x) dx + f(x) dx
a a c+

6.

b
b
cf(x) dx =c f(x) dx, c constant
b
a
a a

a f(x) dx = f(x) dx,
a b
f(x) dx =0,
b
c
b
a

b
b f(x) dx = f(x) dx + f(x) dx
a a c
f(x) dx = f(b x) dx
0 0

7. Mean value theorems

b
f(x) dx =f(c)(b a), acb
a

b
b
f(x)g(x) dx =f(c) g(x) dx, g(x) 0, a c b
a a

b

b
b
f(x)g(x) dx = f(a) g(x) dx f(x)g(x) dx = f(b) g(x) dx
a a a
a<<b a<<b
The last mean value theorem requires that f(x) be monotone increasing and nonnegative
throughout the interval (a, b)
8. Numerical integration
ba
Divide the interval (a, b) into n equal parts by defining a step size h = n .

Two numerical integration schemes are

(a) Trapezoidal rule with global error (ba)
12
h2 f

() for a < < b.

b
h
f(x) dx = [f(x0 ) + 2f(x1 ) + 2f(x2 ) + 2f(xn1 ) + f(xn )]
a 2
(b) Simpsons 1/3 rule with global error (ba)
90
h4 f (iv) () for a < < b.

b
2h
f(x) dx = [f(x0 ) + 4f(x1 ) + 2f(x2 ) + 4f(x3 ) + 2f(x4 ) + + 2f(xn2 ) + 4f(xn1 ) + f(xn )]
a 3

Appendix C
 511

nL
L
9. If f(x) is periodic with period L, then f(x+L) = f(x) for all x and f(x) dx = n f(x) dx,
0 0
for integer values of n.
10.




x x x x
1
dx dx dx f(x) = (x u)n1 f(u) du
0 0 0 (n 1)! 0
  
n integration signs

Integrals containing algebraic terms

1
(m)(n)
11. xm1 (1 x)n1 dx = B(m, n) = , m > 0, n > 0
0 (m + n)

1  2
dx 1 1
12. = ( )
0 1 x4 4 2 4

1
dx
13. 2n n/2
=
0 (1 x ) 2n sin 2n

1
1 dx
14.  =
0 x x(1 x) ( )

1
xp xp dx p
15. = tan , |p| < q
0 xq xq x 2q 2q

1
xp + xp dx p
16. = sec , |p| < q
0 xq + xq x 2q 2q

1
xp1 x1p p
17. 2
dx = cot , 0<p<2
0 1x 2 2

a
dx
18. =
0 a2x 2 2

a  2
19. a2 x2 dx = a

0 4
dx
20. 2 + a2
=
x

0 1 2a
x
21. dx = , 0<<1
1 + x sin

0 1 1
x +x
22. dx = , 0<<1
0 1+x sin


xm dx m
23. = sec
0 1 + x2 2 2


x1
24. dx = cot
0 1 x2 2 2

Appendix C
512


dx
25. n
= cot

0 1x n n
dx 1
26. 2 2 2 2 2
=
0 (a x + c )(x + b ) 2bc c + ab


dx 1
27. =
0 (a2 + x2 )(b2 + x2 ) 2 ab (a + b)


dx 1
28. =
0 (a2 x2 )(x2 + p2 ) 2p a2 + p2


x2 dx p
29. =

0 (a2 x2 )(x2 + p2 ) 2 a2 + p 2
x2 dx
30. 2 2 2 2 2 2
=
0 (x + a )(x + b )(x + c ) 2(a + b)(b + c)(c + a)


x dx
31. =
0 1 + x2 2


x dx
32. =
0 (1 + x)(1 + x2 ) 4

Integrals containing trigonometric terms

1
sin1 x
33. dx = ln 2
0 x 2

/2
tan1 ( ab tan ) d b
34. = ln |1 + |
0 tan 2 a

/2

35. sin2 x dx =
0 4

/2

36. cos2 x dx =
0 4

/2
dx cos1 (b/a)
37. =
a + b cos x a2 b 2

0 /2
(m)(n)
38. sin2m1 x cos2n1 x dx = B(m, n) = , m > 0, n > 0
0 (m + n)

/2
( p+1 q+1
2 )( 2 )
39. sinp x cosq x dx =
0 2( p+q
2 + 1)

/2
dx
40. m =
0 1 + tan x 4


0, p
= q
41. cos p cos q d =
0 , p=q
2

Appendix C
 513


0, p
= q
42. sin p sin q d =
0 , p=q
2


0, p+q even
43. sin p cos q d = 2p
0 2 , p+q odd
p q2


x dx 2
44. =
0 a2 2
cos x 2a a2 1


dx
45. =

0 a + b cos x a2 b 2
sin d 2
46. = tanh1 a
0 1 2a cos + a2 a


sin 2 d 2 2
47. 2
= 2 (1 + a2 ) tanh1 a
0 1 2a cos + a a a



x sin x dx a ln(1 + a),
|a| < 1
48. 2
= 
1

0 1 2a cos x + a
ln 1 + , |a| > 1
a
ap


, a2 < 1
cos p d 1 a2
49. 2
= p
0 1 2a cos + a a ,

a2 > 1
a2 1 p
a


[(p + 1) (p 1)a2 ], a2 < 1
cos p d (1 a2 )3
50. =
0 (1 2a cos + a )
2 2 ap


2 3
[(1 p) + (1 + p)a2 ], a2 > 1
(a p1) 
a 



2 5
(p + 2)(p + 1) + 2(p + 2)(p 2)a2 + (p 2)(p 1)a4 , a2 < 1
cos p d 2(1 a )
51. =
0 (1 2a cos + a )
2 3 ap  


2 5
(1 p)(2 p) + 2(2 p)(2 + p)a2 + (2 + p)(1 + p)a4 , a2 > 1
2(a 1)

2
dx 2a
52. 2
= 2
(a + b sin x) (a b2 )3/2

0 2
dx 2
53. =
0 a + b sin x a b2
2

2
dx 2
54. =
0 a + b cos x a b2
2

2
dx 2a
55. = 2
0 (a + b sin x)2 (a b2 )3/2

2
dx 2a
56. = 2
0 (a + b cos x)2 (a b2 )3/2

L 
mx nx 0, m
= n, m, n integers
57. sin sin dx = L
L L L 2
, m=n

Appendix C
514

L
mx nx
58. cos sin dx = 0 for all integer m, n values
L L L


L
mx nx 0,
m
= n
59. cos cos dx = L2 , m = n
= 0
L L L

L, m=n=0

m
x dx sin m
60. 2
=
0 1 + 2x cos + x sin m sin



sin x /2,
>0
61. dx = 0, =0
0 x

/2, <0



sin x sin x 0,
>>0
62. dx = /2, 0<<
0 x

/4, =>0


sin x sin x 2 , 0 <
63. 2
dx =
0 x , >0
2


sin2 x
64. dx =
0 x2 2


1 cos x
65. 2
dx =

0 x 2
cos x a
66. dx = e

0 x 2 + a2 2a
x sin x
67. 2 2
dx = ea
0 x(x + a ) 2


sin x
68. dx =
0 xp 2(p) sin(p/2)


cos x
69. dx =
0 xp 2(p) cos(p/2)


tan x
70. dx =
0 x 2


sin x
71. 2 2
dx = 2 (1 ea )
0 x(x + a ) 2a


sin2 x
72. 2
dx =
0 x 2


sin3 x 3
73. dx =
0 x3 8


sin4 x
74. dx =
0 x4 3

Appendix C
 515


 
1 b2 b2
75. sin ax2 cos 2bx dx = cos sin
0 2 2a a a

 
1 b2 b2
76. 2
cos ax cos 2bx dx = cos + sin
0 2 2a a a


dx
77. = 3
0 x4 + 2a2 x2 cos 2 + a4 4a cos

 
a2
78. 2
cos x + 2 dx = cos( + 2a)
0 x 2 4

 
a2
79. sin x2 + 2 dx = sin( + 2a)
0 x 2 4


tan bx dx
80. = 2 tanh bp
0 x(p2 + x2 ) 2p


x tan bx dx
81. = tanh bp
0 p2 + x 2 2 2


x cot bx dx
82. 2 2
= coth bp
0 p +x 2


sin ax dx sinh ap
83. 2 2
= , a<b
0 sin bx (p + x ) 2p sinh bp


cos ax dx cosh ap
84. = , a<b
0 cos bx (p2 + x2 ) 2p cosh bp


sin ax dx sinh ap
85. = 2 , a<b
0 cos bx (p2 + x2 ) 2p cosh bp


sin ax x dx sinh ap
86. = , a<b
0 cos bx (x2 + p2 ) 2 cosh bp


cos ax x dx cosh ap
87. 2 2
= , a<b
0 sin bx (p + x ) 2 sinh bp

Integrals containing exponential and logarithmic terms

1
ln x1 2
88. dx =
0 1+x 12

1
ln x1 2
89. dx =
0 (1 x) 6

1
1
3
ln x 4
90. dx =
1x 15

0 1
ln(1 + x) 2
91. dx =
0 x 12

Appendix C
516

1
ln(1 x) 2
92. dx =
0 x 6

1
ln x1 2 a
93. (ax2 + bx + c) dx = (a + b + c) (a + b)
0 1x 6 4

1
ln 1
94. x dx = ln 2
0 1x 2 2

1
1 xp1 1 1 1
95. p
(ln )2n1 dx = (1 2n )(2)2n B2n1
0 (1 x)(1 x ) x 4n p

1  
xm xn 1 + m
96. dx = ln  
0 ln x 1+n

n n!

1 (1) (p + 1)n+1 ,

n an integer
97. p n
x (ln x) dx =
0 (n + 1)
(1)n
, n noninteger
(p + 1)n+1

/4

98. ln(1 + tan x) dx = ln 2
8

0 /2
1
99. ln sin d = ln( )
0 2 2
 a + a2 + b2 

 
100. ln(a + b cos x) dx = ln 
 

0  2 

2 
101. ln(a + b cos x) dx = 2 ln |a + a2 b2 |
0

2 
102. ln(a + b sin x) dx = 2i ln |a + a2 b 2 |
0

1
103. eax dx =
0 a


(n + 1
104. xn eax dx =
0 an+1


2
x2 1 1 1
105. ea dx = = ( )
0 2a 2a 2


2
x2 ( m+1
2 )
106. xn ea dx =
0 2am+1


a
107. eax cos bx dx =
0 a2 + b 2


b
108. eax sin bx dx =
0 a2 + b2

Appendix C
 517


sin bx b
109. eax dx = tan1
0 x a


eax ebx b
110. dx = ln
0 x a


2
x2 b2 /4a2
111. ea cos bx dx = e
0 2a


(ax2 +b/x2 ) 1 2ab
112. e dx = e
0 2 a


2n x2 (2n 1)(2n 3) 5 3 1
113. x e dx =
0 2n+1 n

x2

k b
+x
2
a 2kb/a
114.
e a2dx = 2
e  
0 sin rx dx 2 k sin(ar sin + 2) r cos
115. = 1 e
0 x(x4 + 2a2 x2 cos 2 + a4 ) 2a4 sin 2


cos rx dx sin( + ar sin ) ar cos
116. = 3 e
0 x4 + 2a2 x2 cos 2 + a4 2a sin 2


 
sin rx dx ar 3
117. = 6 3e ar
2ear/2
cos
0 x(x6 + a6 ) 6a 2


 
cos rx dx ar 3 2
118. = 5 ear 2ear/2 cos( + )
0 x 6 + a6 6a 2 3


sin x dx
119. =
0 x(1 x2 )

 
eqx epx 1  p2 + b2 
120. cos bx dx = ln  2
0 x 2 q + b2 


eqx epx p q
121. sin bx dx = tan1 tan1
0 x b b


sin px sin qx p q
122. eax dx = tan1 tan1
0 x a b

 
ax cos px cos qx 1  a2 + a2 
123. e dx = ln  2
0 x 2 a + p2 


2 a a2 /4
124. xex sin ax dx = e
0 4

 
2 a2 2
125. x2 ex cos ax dx = 1 ea /4
0 4 2

 
2 a3 2
126. x3 ex sin ax dx = 3a ea /4
0 8 2

Appendix C
518

 
2 a4 2
127. x4 ex cos ax dx = 3 3a2 + ea /4
0 8 4

 3
ln x
128. dx = 2
0 x1


x sin rx dx
129. = (a cos br + b sin br) ear
(x b)2 + a2 a


sin rx dx  
130. 2 2
= 2 2
a (cos br b sin br) ear
x[(x b) + a ] a(a + b )


cos rx dx
131. 2 2
= ear cos br
(x b) + a a


sin rx dx
132. = ear sin br
(x b)2 + a2 a


2 2
133. ex cos 2nx dx = en

xp1 ln x 2
134. dx = cot p, 0<p<1
0 1+x sin p


135. ex ln x dx =
0

2
136. ex ln x dx = ( + 2 ln 2)
0 4

 
ex + 1 2
137. ln x
dx =
e 1 4

0
x dx 2
138. =
0 ex 1 6


x dx 2
139. =
0 ex + 1 12

Integrals containing hyperbolic terms

1
sinh(m ln x) m
140. dx = tan , |m| < 1
0 sinh(ln x) 2 2


sin ax a
141. dx = tanh( )
0 sinh bx 2b 2b


cos ax a
142. dx = sech ( )
0 cosh bx 2b 2b


x dx 2
143. = 2
0 sinh ax 4a

Appendix C
 519


sinh px p
144. dx = tan( ), |p| < q
0 sinh qx 2q 2q

 

cosh ax cosh bx  cos b 
145. 
dx = ln  2
,
 cos a2 
< b < a <
0 sinh x

sin p


sinh px q
146. cos mx dx = p m , q > 0, p2 < q 2
0 sinh qx 2q cos q
+ cosh q

p m
sinh px sin 2q sinh 2q
147. sin mx dx = p
0 cosh qx q cos q + cosh mq

p m
cosh px cos 2q cosh 2q
148. cos mx dx = p
0 cosh qx q cos q + cosh m
q

Miscellaneous Integrals

x
x
149. 1 [1 ] d = F (, ; + 1; x) See hypergeometric function
0


150. cos(n x sin ) d = Jn (x)
0

a
(m)(n)
151. (a + x)m1 (a x)n1 dx = (2a)m+n1
a (m + n)


f(x) f()
152. If f

(x) is continuous and dx converges, then
1 x


f(ax) f(bx) b
dx = [f(0) f()] ln
0 x a

153. If f(x) = f(x) so that f(x) is an even function, then

 

1
f x dx = f(x) dx
0 x 0

154. Elliptic integral of the first kind

d
 = F (, k), 0<k<1
0 1 k 2 sin2
155. Elliptic integral of the second kind


1 k 2 sin2 d = E(, k)
0

156. Elliptic integral of the third kind

d
 = (, k, n)
0 (1 + n sin ) 1 k 2 sin2
2

Appendix C
520
Appendix D
Solutions to Selected Problems
Chapter 1

1-1.

(a) (b) (c)

A B = {x| 2 < x < 7} A B = {x| 2 < x 7} A B = {x|x < 4}

A B = {x|2 x 4} AB = A B = { x | 5 < x < 3}

1-2.

1-5. (a) A (A B) = (A U ) (A B) and use distributive law to write

(A U ) (A B) = A (U B) = A U = A

1-8.
(a) Sa bounded above
.u.b. = 4, Sa bounded below g.
.b = 4
(b) Sb bounded above
.u.b. = 3, Sb is not bounded below
(c) Sc bounded above
.u.b. = 25, Sc bounded below g.
.b = 0
(d) Sd is not bounded above, Sc bounded below g.
.b. = 27

Solutions Chapter 1
 521

1-9.
(a) y 4 = 2(x 2)
2
(c) y 4 = (x 2)
3

1-10.
(a) y = (3/4)x + 7/4, m = 3/4, b = 7/4
(c) Polar form of 3x + 4y = 7 is r cos( ) = d, where d = 7/5 and tan = 4/3

1-12.
(b) x
= a, x
= b, x
= c
(c) r0
(d) x3 + 1 > 0 = x > 1

1-13.

1-14.

1-15.

Solutions Chapter 1
522

1-16.

1-18.

1-21.

1-22.

(x + h)2 x2
(e) (g) f (g(x)) = g 2 (x) = (3 2x)2
h

1-24. (d) f 1 (x) = x3 4

1-26.

1-27.

Solutions Chapter 1
 523

1-31. (a) e (b) e
4 3 1

1-34. (c) y = x, (d) y = x, (e) y = x, (f) y = x, (g) y = x, (h)
3 4 2
4
y= x
3
1

1-37. (b) y 4 = (x 3)
2


1-38. (c) Multiply numerator and denominator by x+h+ x and simplify.
(d) 2x
1
(f)
4

1-39. (c) Limit does not exist.

(e) Limit does not exist.

All other limits have finite values.

1-41. (a) 4
(c) Multiply numerator and denominator by (1 + cos h)
(e) 1

1

1-42.
2 x

1-43. (a) sin x oscillates between -1 and +1.

(b) sin( x1 ) oscillates between -1 and +1.
(c) Put in form sin

1-44. (d) 3y 4x = 1

Solutions Chapter 1
524
1 1

1-45. (b) Show (1 + p)n > 1 + np and n
<
(1 + p) 1 + np
(d) Show (1 + p)n > 1 + np and 1 + np as n

1-47. (c) Let the point (3/2, 9/4) approach the point (1, 1), then secant line ap-
proaches the tangent line.
(d) y 1 = 2(x 1)

1-48. (d) Complete the square y2 6y + 9 + 12x 3 9 = 0 and then simplify to obtain
(y 3)2 = 12(x 1) From this equation show the focus is at (-2,3), the vertex is at
(1,3), the directrix is the line x = 4 ant the latus rectum is 12.

1-49. (d) Complete the square on the x and y terms to obtain

25(y 2 6y + 9) + 16(x2 4x + 4) = 689 4(25) 4(16) = 400

which simplifies to
(y 3)2 (x 2)2
+ =1
42 52
representing an ellipse. Here the foci are at (5,3) and (-1,3), the directrices are at
x = 31/3 and x = 19/3, the latus rectum is 32/5, the eccentricity is 3/5 and the
center is (2,3).

1-50. (d) Complete the square (x2 2x + 1) 4(y2 8y + 16) = 67 + 1 64 = 4 which

simplifies to
(x 1)2 (y 4)2
=1
22 12

which is a hyperbola centered at (1,4). The foci are at (1 2 5, 4) and (1 + 2 5, 4), the

verticies are at (-1,0) and (3,0, the directrices are at x = 1 5 and x = 1 + 5, the

eccentricity is 2 5 5 and the asymptotic lines are y 4 = 1
2
(x 1) and y 4 = 12 (x 1).

1-51. If yline = yparabola , then x + b = 2 x so that x must satisfy x2 + 2xb + b2 = 4x or
x2 + (2b 4)x + b2 = 0. This is quadratic equation with roots


4 2b (2b 4)2 4b2
x=
2



The discriminant is (2b 4)2 4b2 = 16(1 b) which tells one that
b = 1, one point of intersection
b < 1, two points of intersection
b > 1, no points of intersection
Solutions Chapter 1
 525

1-52. (a) If a, b, c, d, e, f are defined by equations (1.97), then

a + c = a(cos2 + sin2 ) + c(sin2 + cos2 ) = a + c

1-53. Rotate axes /4 radians and show that for x = x 12 y 12 and y = x 12 + y 12 ,

then
x2 y 2
xy = = a2
2 2

1-54. (x + 1)2 = y 2

9

1-57. F = C + 32
5

1-58. (a) line x = 3 (c) circle (x 2)2 + y2 = 22

1-59. (a) parabola opens to right  < < 2  (c) ellipse centered at (2,0)
intersecting x-axis at (-2,0) and (6,0).

1-60. (a) parabola opens upward 2 +  < < 5

2
 (c) ellipse centered at (1, 0)
intersecting x-axis at (3,0) and (-1,0).

1-62. y=3 and x = 2

Solutions Chapter 1
526
Chapter 2

2-1. (b) y 4 = 4(x 2) (c) y 4 = 4(x + 2)

dy b + 2ax

2-2. (d) =
dx (ax + bx + c)2
2
dy
(h) = 3(2x + 1)2 (2)(3x2 x)2 + (2x + 1)3 2(3x2 x)(6x 1)
dx
which can also be expressed in the form
dy
= 2x(3x 1)(1 + 2x)2 (1 + x + 21x2 )
dx
dy 1 dy dy 1 1

2-3. (d) = 3/2 (j) = 1 + ln x (l) = (a + x) n 1
dx 2x dx dx n

dy x 1 dy x(3x2 2x)
3

2-4. (b) = + (f) = + x x2
dx (1 + x)2 2 x(1 + x) dx 2 x3 x2
dy x
(k) = 
dx (1 + x2 )3/2 1 1 1+x2

dy

2-5. (g) = 6 cos(3) sin(3) = 3 sin(6)
d
dy a 2bx
(k) =

dx 1 (ax bx2 )
2

dy x 1/x 1 ln x
(l) = x (1 + ln x) + x 2
dx x2 x

2x

2-6. (j) y
= 
2
(1 + x2 )2 1x
1+x2
(b + 2cx) sec[ln(a + bx + cx2 )] tan[ln(a + bx + cx2 )]
(k) y
=
a + bx + cx2

x 3x2 y + 6xy + 3y 2 yxy1

2-7. (e) y
= 2
(j) y
= (l) y
=
y 2 x 3x2 6xy 3y 2 1 x4 ln x

2x 1+x

2-8. (d) y
= (g) y
= +sin1 x (j) y
= (cos(3x))x (ln | cos 3x| 3x tan(3x))
1 x4 1x 2

 

x x

2-9. (d) y = (3 + x) + ln(3 + x) (i) y
= xx (1 + ln x)
3+x

6x cos(4x) sin2 (4x)
(l) y
=  + sin3 (4x)
sin3 (4x)

2ax 3 3

2-10. (b) y
= (h) y
= =
1 a2 x4 3x 1 1 + 3x 9x2 1
2
3x
(l) y
=
1 x6

Solutions Chapter 2
 527
f (x0 + x) f (x0 )

2-11. lim , let x = x0 + x
x0 x

3x cos x

2-13. (d) y
= + 4 + 3 sin x
2 4 + 3 sin x


9x cos2 x 3 cos x 3x sin x
y = +
4(4 + 3 sin x)3/2 4 + 3 sin x 2 4 + 3 sin x

ab x2
(e) y =
(a x)2 (b x)2
2(a2 b + ab(b 3x) + x2 )
y

=
(a x)3 (x b)3

2-14.
dy
dy dx dy dy
= = = dt
dx dt dt dx dx
   dt
d dy d dy dx d2 y dx
Note that = = 2
dt dx dx dx dt dx dt
so differentiating the above with respect to t gives
 2 d2 y 2
dy d x
dy d2 x d2 y dx d2 y d2 y dt2 dx dt2
+ 2 = 2 = 2 =  dx 2
dx dt2 dx dt dt dx
dt

(b)
dx dy
x = 4 cos t y = 4 sin t with
= 4 sin t = 4 cos t and
dt dt
d2 x d2 y
= 4 cos t = 4 sin t
dt2 dt2
dy 4 cos t x d2 y 4 sin t (x/y)(4 sin t) y 2 + x2
so that = = and = =
dx 4 sin t y dx2 (4 sin t)2 y3

Another method. x2 + y2 = 16 so that 2x + 2yy
= 0 = y
= x/y

x2 + y 2
and 2 + 2yy

+ 2(y
)2 = 0 = y

=
y3

1 + 2x

2-15. (e) y
= x + 2x ln(3x) (h) y
= cos(x2 ) 2x ln(x + x2 ) sin(x2 )
x + x2

1 3x2 3 sin(x2 )

2-16. (e) y
= (h) y
= 2x cos(x2 ) ln(x3 ) +
(1 + x2 )3/2 (1 + x2 )5/2 x

2-17. (e)
y
= 6 cos(3x) sin(3x)

y

= 18 cos2 (3x) + 18 sin2 (3x)

Solutions Chapter 2
528

2-18. (b) y
= cos x so that m = cos 4 = 1
2
and therefore
1 1
(y ) = (x )
2 2 4

2-20. (d)

2-21.
f
(x + h) f
(x)
f

(x) = lim
h0 h
f (x + h) f (x)
but f
(x) = lim
h0 h
 
f (x+2h)f (x+h) f (x+h)f (x)
h h
therefor, f

(x) = lim
h0 h2
f (x + 2h) 2f (x + h) + f (x)
f

(x) =
h2
x2 1

2-22. (b) y
= (x2 +1)2
critical points at x = 1

(e) local maximums at x = (2n + 1)/2

local minimums at x = (2n + 3)/2 where n is an integer.

2-23. (d) Critical points at x = 1 and curve symmetric about origin.

2-26. Let x + y =
with x used for square and y used for triangle, then area of
1
square is As = (x/4)(x/4) = x2 /16 and the area of the triangle is At = sin (y/3)2 .
2 3
The sum of these areas can be expressed

x2 3 2
A= + y
16
36
x2 3
A= + (
x)2
16 36

Show x = 4 3 and y = 9 when A is a minimum.

Solutions Chapter 2
 529

2-28. (c) If f
(x) = x(x 1)2 (x 3)2 , then
f

(x) = 3x(x 1)2 (x 3)2 + 2x(x 1)(x 3)3 + (x 1)2 (x 3)3 = (x 1)(x 3)2 (6x2 13x + 3)
f
(x) = 0
at x = 0, x = 1 and x = 3, these are the critical values.
At x = 0, f

(x) = (1)2(3)3 < 0 = f (x) is local maximum.
At x = 1, f

(1) = 0, second derivative test fails, so use the first derivative test
using the values x = 1/2, x = 1 and x = 3/2.
At x = 1/2, f
(1/2) < 0 negative slope
At x = 1, f
(1) = 0 zero slope
At x = 3/2, f
(3/2) < 0 negative slope

Hence, the point where x = 1 corresponds to a point of inflection.

At x = 3, f

(3) = 0 second derivative test again fails, so use first derivative test
with x = 2.5, x = 3 and x = 3.5. One finds
f
(2.5) < 0 negative slope
f
(3) = 0 zero slope
f
(3.5) > 0 positive slope

so that x = 3 corresponds to a local minimum.

ln a

2-29. (b) e (d) 1/2 (g)
ln b

2-30.

2-33. (c)
2u x2 1
2
= 2 2 3/2
+

u x x (x + y ) x + y2
2
=

x x2 + y 2 2u 2u xy
= = 2
u y x y y x (x + y 2 )3/2
=

y x2 + y 2 2u y 2 1
2
=
3/2
+

y x2 + y 2 x2 + y 2

Solutions Chapter 2
530

2-35.
(a) D ex =ex
D2 ex =2 ex
.. ..
. .
Dn ex =n ex

(b) D(ax ) =D(ex ln a ) = ln a ax

D2 (ax ) =(ln a)2 ax

.. ..
. .
Dn (ax ) =(ln a)n ax
1
(c) D(ln x) =
x
2 1
D (ln x) = 2
x
(1)(2)
D3 (ln x) =
x3
.. ..
. .
(1)(2)(3) (n + 1) (1)n1(n 1)!
Dn (ln x) = n
=
x xn
(d) D(xn ) =nxn1

D2 (xn ) =n(n 1)xn2

D3 (xn ) =n(n 1)(n 2)xn3
.. ..
. .
Dm (xn ) =n(n 1)(n 2) (n (m 1))xnm, m<n
n!
or one can write Dm (xn ) = xnm , m<n
(n m)!

(e) D(sin x) = cos x = sin(x + )
2

D2 (sin x) = sin x = sin(x + 2 )
2
3
D (sin x) = cos x = sin(x + 3 )
2
.. ..
. .

Dn (sin x) = sin(x + n )
2

Solutions Chapter 2
 531

2-35.

(f ) D(cos x) = sin x = cos(x +)
2

D2 (cos x) = cos x = cos(x + 2 )
2
3
D (cos x) = sin x = cos(x + 3 )
2
.. ..
. .

Dn (cos x) = cos(x + n )
2
(g) Express problem so that the results from parts (e) and (f) can be employed.
3 1
Use sin3 x = sin x sin(3x), then
4 4
3 3
D(sin3 x) = D(sin x) sin(3x + )
4 4 2
2
3 3
D2 (sin3 x) = D2 (sin x) sin(3x + 2 )
4 4 2
.. ..
. .
3 3n
Dn (sin3 x) = Dn (sin x) sin(3x + n ) or
4 4 2
n
3 3
Dn (sin3 x) = sin(x + n ) sin(3x + n )
4 2 4 2

2-40. v = t3 6t2 + 11t t and a = 3t2 12t + 11

ds
Hint: Show v = = (t 1)(t 2)(t 3)
dt
d d 2 1 du u du

2-42. |u| = u = (u2 )1/2 2u =
dx dx 2 dx |u| dx

2-44.

(a) ln y = ln + x ln = Y = mx + b where Y = ln y, m = ln , b = ln
(b) ln y = ln + ln x = Y = X + b where Y = ln y, X = ln x, b = ln

Solutions Chapter 2
532

2-45.
 2
(c ax)
Let y = y(x) = s2 =(x x0 )2 + y0
b
dy

=2 ax2 + b2 (x x0 ) + a(c + by0 ) = 0 at the point where

dx
(a2 + b2 )x b2 x0 + ac + aby0 = 0
b2 x0 ac aby0
or x = x =
a2 + b2

(c + ax0 + by0 )2
Show that ymin = y(x ) = sm in2 = and that
a2 + b2

|c + ax0 + by0 |
Smin = d =
a2 + b2
dy
dy dx dy dy d

2-46. Chain rule differentiation requires = or = dx
dx d d dx d

2 2

2-50. Cone with maximum volume has base radius r = 3
R and height h = 43 R.


2-53. A square with side 2R.

2-56.

m mn

2-60. (a)
n

2-65. By the law of cosines
2 = r2 + s2 2rs cos t. Differentiate this relation and
show
ds r 2 cos t sin t
= r sin t
dt s r cos t


From the law of cosines show s r cos t =
2 r2 sin2 t

Solutions Chapter 2
 533
Chapter 3

4 3

3-1. (a) 3
x + 2x2 + x + C or 16 (2x + 1)3 + C1
(b) csc(t) + C
(c) 1
2
cos2 x + C1 or 12 sin2 x + C2

1 5 4 3 2 1 5

3-2. (a) 10 sin (2t) + sin (2t) + 4 sin (2t) + 8 sin (2t) + 8 sin(2t) + C1 or 10 [2 + sin(2t)] + C2
2
4x
(c) 4 ln 2
+C
(e) 13 cos(3x + 1) + C

1 1 1

3-3. (a) 3
tan(3x + 4) + C (c) 3
sec(3x + 4) + C (e) 6
ln | sin(3x2 )| + C

3-4. (a) 1
2
sin1 (x2 ) + C (c) 1 x2 (e) 1
2
sinh (x2 ) + C

1 1 1



3-5. (a) +C (c) coth (3x + 1) + C (e) (3x + 1)2 1
3 sinh (3x+1) 3 3

1


3-6. (a) x + 5x + ln x + C (c) 2c t + at + 23 bt3/2 + C (e) 1
3b (a + bu2 ) a + bu2

26

3-7. (a) 3
Area under parabola
(b) 2 Area under sine curve
(c) 12 BH Area of triangle

3-8. (a) Let u = 3x + 1 = tan with du = 3 dx = sec2d 


1 1 1  1 + u2 1 
then csc d = ln | csc + cot | = ln  + 
3 3 3  u u
1
1
= ln |1 + 1 + u2 | + ln u + C where u = 3x + 1
3 3

3-9. Two functions with the same derivative differ by some constant value.

3-10. Use table III with appropriate scaling.

3-11. (a) ln |x 3| + ln |x 2| + ln |x 1| + C
(c) tan1 x + ln(x + 1) + ln(x2 + 1) + C
3 1 1
(e) x + x2 + x3 + x4 + ln |x 1| + C
2 2 1 + x + x2

Solutions Chapter 3
534

 x y
1 1 5

3-12. (a) dy = x dx dy = x dx = y 3 = (x2 1) or y = x2 +
1  y 1  2 2 2
1 x 1
(c) dy = sin(3x) dx dy = sin(3x) 3dx = y 1 = [1 cos(3x)]
1 3 0 3
4 1
or y = cos(3x)
3 3  y 
1 x 2 1
(e) dy = sin2 (3x) dx dy = sin (3x) 3 dx = y 1 = [6x sin(6x)]
1 3 0 12
1 1
or y = 1 x sin(6x)
2 12
1 1 1

3-13. (b) x tan1 x + C (d) x 2 ln(x + 1) + C (f ) ln x ln(x a) + C
x+1 a a

1 1

3-14. (b) + [ln(x b) ln(x a)] + C
(b a)(x a) (a b)2
ln(x2 b2 ) ln(x a) a  
1 x
(d) + tan +C
2(b2 a 2 ) a2 b2 b(b2 a2 )  b
1 1   1  
1 x 1 x
(f ) 2 tan + tan +C
a + b2 b b a b

1 e2

3-15. (b) 1 (d) + (f ) 100 15
4 4
1 1 1

3-16. (b) (x2 2x+2)ex +C (d) (sin xcos x)ex +C (f ) sec x tan x+ ln | sec x+tan x|+C
2 2 2

eax x x2 1 x2

3-17. (a) (a sin bx b cos bx)+ C (c) ln(x + 1) + ln(x + 1) + C (e) e 2
a2 + b2 2 4 2 2
 1

3-18. (a) f (x) dx
0

 5

3-19. A= [(4y 1) (y 2 6)] dy, = A = 36
1

3-21. (a) Use symmetry to find

 4

32
area ABH = area BCD = ( y + 4 4 y) dy = ( 2 1)
0 3
 
1 x 1

3-23. (d) (a2 x2 )3/2 + C (e) 2x 2x ln x + x(ln x) + C 2
(f ) 2 eax + C
3 a a

1 1

3-24. (a) ln[t + 2 + t2 + 4t 3] + C (d) tan1 x + C (e)
[ln(x 1) ln(x + 1)] + C
9 18


2 2
1 (a + x) b (a + x)

2

3-25. (a) 2 ln(1+ x)+C 2 2
(f ) a b (x + a) +(ba ) tan +C
b2 a2 2ax x2

Solutions Chapter 3
 535
1 


3-26. (c) 2 x 2 ln(1 + x) + C (f ) x x2 + (2x 1) sin1 x + C
2
 

5 1 2 3

3-27. (c) (x2 + 1) ex (d) 1 x2 x x + sin1 x + C
8 4 8

64

3-28. Surface S = and volume V = 5 2
3

3-29. (b) = 1

3-32. (b)arc lengths = 16

28 2

3-35. (i) Surface area S= (ii) volume V = 12
3
  762

3-36. (i) Surface area S = (145)3/2 (10)3/2 (ii) Volume V =
27 7
2 2 4

3-37. (a) (b) (c)
3
1

3-38. (d) 2 (f ) ln 2
2

3-39. The area of three sides need to be calculated. Call these surface areas S1 , S2

and S3 . Show S = S1 + S2 + S3 = 11 2 + 9 2 + 20
 2  r  2  r

3-44. (a) A = r d (b) A = 2 x dx (c) A = r drd
0 0 0 0

3-45. The figure illustrated in the problem 3-45, without the axes, is known as the
symbol the Pythagoreans use to represent their society.
(a) Divide areainto four symmetric parts and showthe area of one of these parts

3 2 3
is A1 = r02 and the total area is A = 4A1 = r02
6 8 3 2
5
(b) Volume is given by V = r3
12 0
1 2 2

3-47. (a) 80 (b) 8 (c) [ a c b2 c2 a2 d2 + b2 d2 ]
4

3-50. c = 12 (a + b)

Solutions Chapter 3
536

3-52. The function f (x) is periodic so that

f (x) = f (x + T ) = f (x + 2T ) = = f (x + (m 1)T ) = f (x + T ) =

Write
 nT  T  2T  mT  T
I= = f (x) dx + f (x) dx + + f (x) dx + + f (x) dx
0 0 T (m1)T (n1)T

n  mT
or I = m=1 (m1)T Let x = u + (m 1)T with dx = du and note that when
f (x) dx
x = (m 1)T , then u = 0 and when x = mT , then u = T , so that
n 
 T n 
 T  T
I= f (u + (m 1)T ) du = f (u) du = n f (u) du
m=1 0 m=1 0 0

3-54.

1 (u + 3)m+1
(a) let u = e2x , du = 2e2x dx then interal has value +C
2 m+1
 4
u + u3
(b) let u = ex , du = ex dx and show integral is du
u2 + 1
u4 + u3 1u
Show
2
= 1 + u + u2 + and integral is
u +1 1 + u2
u2 u3 1
u+ + + tan1 u ln(1 + u2 ) + C
2 3 2
1
(c) let u = ex , du = ex dx with integral (u + 1)3 + C
3

3-55.
a 1
(a) 2
+C
2x x
ab (a + b)
(b) 2 + ln x + C
2x x
abc (bc + ab + ac)
(c) + x + (a + b + c) ln x + C
x x

3-56.
(a) x + ln(1 + x) + x ln(1 + x) + C
x4 + 1 2
(b) Hint : = 1 + x + x2 + x3 +
x1 x1
1

12x + 6x2 + 4x3 + 3x4 + 24 ln(x 1) + C

12
(m + 2)a + b(m + 1) c
(c) (x + c)m+1 + C
m2 + 3m + 2

1 1

3-57. (a) (b) ex (c)
s s2

Solutions Chapter 3
 537

3-58. Integrate term by term and show
x2 x4 x2m
J0 (x) =1 + + (1)m +
22 1!1! 24 2!2! 22m m!m!
x x3 x5 x2m1
J1 (x) = 3 + 5 + + (1)m1 2m1 +
2 2 1!2! 2 2!3! 2 (m 1)!m!

3-59. In = xn ex nIn1 , I4 = (x4 4x3 + 12x2 24x + 24) ex

3-60. Let u = a x, with du = dx, then when x = 0, u = a and when x = a, u = 0 so

that  a  0  a  a
g(a x) dx = g(u)(du) = g(u) du = g(x) dx
0 a 0 0

 2T  T  2T

3-65. Write f (x) dx = f (x) dx + f (x) dx and in the second integral let
0 0 T
x = 2T u with dx = du, so that when x = T, u = T and when x = 2T, u = 0, then use
f (2T x) = f (x) and write
 2T  T  0  T  T  T
f (x) dx = f () d + f (2T u)(du) = f () d + f (u) du = 2 f (x) dx
0 0 T 0 0 0

NoteDefinite integrals have dummy variables of integration.

 h

3-66. V = A(y) dy
0

3-67. 
 1p
dx if p < 1
1p
(e) =
0 ( x)pDoesnt exist if p 1
  1p

dx p1
if p > 1
(f ) p
=
0 ( + x) Doesnt exist if p 1

3-73. Material removed is cylinder with spherical end caps of height h = r r0

where r0 = r 1 2 /4. Diameter of drill is r implies radius of cylinder is r/2.
4
V0 =volume sphere before drilling = r3
3
r 2
V1 =volume of cylinder without spherical caps = ( ) (2r0)
  2
V2 =volume of 2 spherical caps = 2 h2 (3r h)
3
V3 =volume removed by drill = V1 + V2
V4 =Volume remaining after drilling = V0 V3 = V0 (V1 + V2 )

Solutions Chapter 3
538
Chapter 4

4-1. (a) 3/4 (c) 1 (e) 0

4-2. (a) 1/2 (d) 0 (f ) 0

m
 r m+1 1

4-3. S= rn = 1 + r + r2 + + rm , S=
n=0
r1

310 1
(d) 4 = 118, 096
2
1 (.02)100
(e) = 1.02041
1 .02
 10
3 2
2+ 2
1
(f ) (2 + 2) = 6.37237
3 2
2+ 2
1

74 3 r 1

4-4. For S3 use r= and show S3 = = ( 3 1)
2 3 1r 2

4-5. S3 = 0.83333 . . . and S4 = 0.16666 . . .

4-6. (b) Use ratio test and show convergence for |x| < 2



4-7. (ii) E(X) = kpk = 3
k=1

  

4-8. S2 = i=1 r1i + r2i r1 = a2 /b r2 = 1/b
a2 1
Show for a, b real and b > a2 > 1, then series converges to 2
+
ba b1

Note if S = 1 + r + r2 + converges, then

dS  d 1 1
= 1 + 2r + 3r 2 + 4r 3 + = kr k1 = (1 r) =
dr dr (1 r)2
k=0
a b
Consequently, one can show S4 = +
1 r (1 r)2
 
 1  1 1

4-9. (b) =
n=3
(n 2)(n 1) n=3 n 2 n 1
     
1 1 1 1 1 1
S= + + + which converges to S = 1.
1 2 2 3 3 4

Solutions Chapter 4
 539
n 1/2 1 1/2

4-10. (d) un = = + so that the N th partial
(n + 1)(n + 2)(n + 3) n+1 n+2 n+3
sum can be written
N
 5N + N 2
UN = un =
n=1
12(6 + 5N + N 2 )

1
Divide numerator and denominator by N 2 and show lim UN =
N 12
 

4-11. (d) If f (x) = x ln1 x , then MT f (x) dx = MT x ln1 x dx

Let u = ln x with du = x1 dx and show MT f (x) dx = ln[ln x]TM = ln[ln T ] ln[ln M ]
Show this result increases without bound as T so the integral diverges and
series diverges.

1  1

4-12. Let un = , where is the p-series which converges p > 1 and diverges
np n=1
np
p 1. Now select vn = np f (n) and follow results for modification of a series.
1 1

4-13. (d) Since 3n2 + 2n + 1 > n2 , then 2 < 2 and we know the p-series,
3n + 2n + 1 n
with p = 2 converges.
4
 1 1

4-14. (iv) (1)n+1 = 0.945939 with error E < = 0.0016
n=1
n4 54
8
 1 1
(1)n+1 = 0.94694 with error E| < = 0.000152416
n=1
n4 94

4-15. (A) converges slower than (B)

x2n 3xn + 1

4-17. xn = xn1 x0 = 2, . . . , x4 = 2.61803
2xn 3
3 5
Exact roots are x = Convergence to desired root depends upon position
2
of initial guess.

4-18. Converges |x| < 1 and diverges for |x| > 1

 
  un+1  1

4-20. Assume un converges by ratio test so that lim 
n  un 
 = p < 1. Let vn = un
n=1


and examine vn using the ratio test to show series diverges.
n=1

Solutions Chapter 4
540

4-21. This is a telescoping series with
n1

Sn = (Uk+1 Uk ) = (U1 U0 ) + (U2 U1 ) + (U3 U2 ) + + (Un Un1)
k=0
1
so that Sn = Un U0 = n(n + 1)(n + 2)
3
Note that this result can be generalized. If one is given a Uk and calculates the
N
 N

difference Uk = Uk+1 Uk , then one can write Uk = (Uk+1 Uk ) = UN+1 U1
k=1 k=1
m

What would Uk produce for the answer?
k=

4-22.

1 1 1 1
x 1 = ln(y) = (y 1) + (y 1)2 (y 1)3 + (y 1)4 (y 1)5 +
2 3 4 5

1 2 1 3 1 4
Error = E = E(y) = ln(y) (y 1) + (y 1) (y 1) + (y 1)
2 3 4
Over the interval 1 y 2 the error curve is as illustrated

4-23. (b) (c) (e) all converge

1

4-24. (a) limn n
un = limn ln n
=0 hence converges.

3


4-25. (c) sin(/3) = 2 gives geometric series with sum 3 + 2 3

4-26. Use the nth term test 0 and note (e) is a form of the harmonic series.

4-27. (f) divergent

nn n n n n

4-28. (f) = >n divergent
n! n(n 1)(n 2) 1

n n
 n

4-29. (f) n<n and n2 +1
< n2 +1
. The series 2
can be compared with the
n=1
n +1

 1
p-series to show absolute convergence.
n=1
np

Solutions Chapter 4
 541

4-30. Use ratio test

4-33. Write (a + b)r = [a(1 + b/a)]r = ar (1 + b/a)r and let x = b/a and then examine the
series f (x) = (1 + x)r . Here
f
(x) =r(1 + x)r1

f

(x) =r(r 1)(1 + x)r2

.. ..
. .
f (n) (x) =r(r 1)(r 2) (r (n 1)(1 + x)rn

This series has the power series expansion

     
r r r 2 r 3
f (x) = (1 + x) = 1 + x+ x + x + + Rn
1 2 3

Using the Lagrange form of the remainder

 
xn (n) r n
Rn = f (x) = x (1 + x)rn
n! n

where 0 < < 1. For 0 x < 1 and n > r, then one can show (1 + x)rn < 1 and
r  n
n
x 0 as n The Lagrange form for the remainder doesnt aid in the analysis
of the region 1 < x 0 and so one can use the Cauchy form of the remainder to
analyze the remainder in this region.
The Cauchy form of the remainder is
r(r 1)(r 2) (r n + 1) (1 )n1 xn
Rn = 1 + x)nm , 0<<1
(n 1)! (
1x
If |x| < 1, then 1+x
<1 so that
(1 )n1 1
<
1 1m (1 + x)1m
(1 + x)n1 1+x

1
For 1 < x 0 the term (1+x)1m
=K is some constant independent of the index n
and the term
  
r(r 1) (r n + 1) (r 1)! r1
=r =r
(n 1)! (n 1)!(r n)! n1

Therefore,  
 r1  n
|Rn | < K|r| 
  |x| 0 as n
n1 
So the binomial series converges as |Rn| 0 for |x| < 1

Solutions Chapter 4
542
1 x

4-36. y1 = 0 + 1
1 + x
= 0 + 1 x+1 Here Q1 = 1 x + 1 and

dy1 (1 x + 1)(1) x(1 ) 1

= =
dx (1 x + 1)2 [Q1 (x)]2

4-41. For n, m integers and n
= m

 
sin[(m n)x] sin[(m + n)x] 
(sin nx, sin mx) = sin nx sin mx dx = = 0, m
= n
0 2(m n) 2(m + n) 0
 
2 x sin 2nx 2
 sin nx  = (sin nx, sin nx) = sin (nx) dx =  = 2
0 2 4n 0
  
2
or  sin nx = . The set of functions sin nx is therefore an orthonormal set
2
with norm-squared equal to unity.

4-42. (c)
 L
nx nx
(1, cos )= cos dx = 0
L L L
 L
nx nx
(1, sin )= sin dx = 0
L L L
 L
nx mx nx mx
(cos , cos )= cos sin dx = 0, m
= n
L L L L L
 L
nx mx nx mx
(cos , cos )= cos cos dx = 0
L L L L L
 L
nx mx nx mx
(sin , sin )= sin sin dx = 0 m
= n
L L L L L
(1, 1) =  1 2 = 2L
nx nx nx 2
(sin , sin ) =  sin  =L
L L L
nx nx nx 2
(cos , cos ) =  cos  =L
L L L
 L 
1 2 L nx

4-43. (e) (i) If f (x) is even, bn = 0 and a0 = f (x) dx, an = f (x) cos dx
L 0 L 0 L

2 L nx
(ii) If f (x) is odd, a0 = an = 0 and bn = f (x) sin dx
L 0 L

Solutions Chapter 4
 543

4-44. Roots are (2, 3) and (1, 1)

Cs
C0 ek C0 1

4-48. (a) C = (c) k = ln
Cs

1 ek
C0



4-53. y = sin x + y = y 2 = sin x + y Differentiate this relation to obtain given
answer.

4-55. If ex cos x = a0 + a1 x + a2 x2 + a3 x3 + , then its derivative is

ex cos x [cos x x sin x] = a1 + 2a2 x + 3a3 x2 + 4a4 x3 +

Show
3 5 7 6 1 8
cos x x sin x = 1 x2 + x4 x + x +
2 24 720 4480
and substitute the top line equation and third line equation into the second line
equation to obtain
3 5
(a0 + a1 x + a2 x2 + )(1 x2 + x4 + ) = a1 + 2a2 x + 3a3 x2 +
2 24

Expanding the left-hand side of this equation gives

3 3
a0 + a1 x + (a2 a0 )x2 + (a3 a1 )x3 + = a1 + 2a2 x + 3a3 x2 +
2 2

Equate like powers of x to find a relation between the coefficients. Use top line
equation to show at x = 0 that a0 = 1 and from equating like powers of x one finds

a1 = 1, a2 = 1/2, a3 = 1/3, a4 = 11/24, etc

Solutions Chapter 4
544
Chapter 5
4 3 dV dr dr dr

5-1. V = r , = 4r 2 , = 4r02 = =
3 dt dt dt dt 4r02

dp 1.4 dv dp p dv

5-2. pv 1.4 = c, v + p(1.4)v 0.4 =0 or = 1.4
dt dt dt v dt

dp
(a) dt
= 1.4 vp
dv v
(b) dt
= 1.4 p

5-3.

+s s d ds 22
= , if dt
= 4 ft/s, then show = ft/s.
10 5.5 dt 4.5
1 1 1

5-4. (a) Let S1 = x f > 0 and S2 = y f > 0, then + = or
xf +f y f +f f
1 1 1
+ = Simplify this expression and show f 2 = S1 S2
S1 + f S2 + f f
dy y2
(b) Differentiate the lens law and show = 2 r0
dt x

3 2 dA 2 3 dx

5-5. Area of equilateral triangle with side x is A= x so that = x
8 dt 8 dt
Now substitute x = x0 and dx
dt
= r0

5-6. p = p0 e0 h

(a) [p0 ] = lbs/f t2 , [0 ] = 1/f t

dh dp dh
(b) = 10 f t/s = p0 e0 h (0 ) Substitute for dh
dt
and find the pressure
dt dt dt
decreases with height.
dy

5-7. yy0 = (xx0 )2has the derivative = 2(xx0 ). When x = , the slope of the
dx
tangent is m = 2( x0 ) and the equation of the tangent line is y = 2( x0 )(x).
d d d 2 d
Here tan = m = 2(x0 ) and if d
dt
= 1/2 cm/s, then sec2 = 2 or = 2
dt dt dt 1 + tan dt
d 2 1
or = 2 2
dt 1 + 4 ( x0 ) 2

dT r0

5-8. = where c0 = V0 /T0 .
dt c0

dP

5-9. = c0 r 0 where c0 = P0 /T0 .
dt

Solutions Chapter 5
 545

5-10.
h =200 16t2 f t
dh
v= = 32 t f t/s
dt
dv d2 h
a= = 2 = 32 f t/s2
dt dt

When h = 0, then t = 210/4 and v = 80 2 f t/s

5-11.
4 pi dV dh dh
(a) V = r3 (2rh)2 (r+h) = (3rh2 h3 ) and = (3r 2h 3h2 ), r is a constant
3 3 3 dt 3 dt dt
dh
(b) = 30/( 420)
dt

5-12. (h r)2 + R2 = r 2 Differentiate this relation and show

dR dh
R = (h r) r < h < 2r
dt dt
dR dh
R =(r h) 0<h<r
dt dt


5-13. (a) 1125/16 f t (b) 50 3/2 (c) 50 3/16 (d) 625 3/8 (e) parabola

dv K

5-14. =
dx 1 2x

My 60 Mx 44 11

5-15. (a) My = 60 Mx = 44 x = = = 3, y = = =
A 20 A 20 5
My 3 Mx 3

5-16. (b) x = = x0 y = = x20
A 4 A 10
21

5-17. (b) x = y = 3
5
b 3h + 16

5-18. (c) x = ( )
4 h+6

5-19. y = 4/3

1

5-20. (a) T Tenv = T0 ekt , T0 constant (b) T = 100ekt (c) k = ln(4/5) (d)
20
t = 20 ln(9/7)/ln(5/4)

t

5-21. N = N0 e 5 ln 3

Solutions Chapter 5
546

5-22.  x1
(a) dV = 2xh(x) dx V = 2xh(x) dx
 x1 x0
(b) dA = h(x) dx A= h(x) dx
x0  x1 
1 x1
(c) dM = x dA = xh(x) dx M= xh(x) dx x = xh(x) dx
x0 A x0

(d) (Area)(distance traveled by centroid)=A 2 x=(volume)= xx01 2xh(x) dx
1
 x1  x1
A 2 A x0
xh(x) dx = 2 x0
xh(x) dx reduces to an identity.

5-23. 1 2 1
(a) y = r x = h A = hr Volume=V=2 yA = r2 h
3 3 2 3

4 4
(b) y = r, A = r 2 Volume=2 yA = r3
3 2 3
 
dy dy

5-24. (e) = dx, = dx, ln y = x + C or y = y0 ex , y0 = eC
y y
(f)  
d2 y dy dy
dx = dx gives =y+C
dx2 dx dx
Now separate the variables and write
 
dy dy
= dx with = dx giving ln(y + C) = x + C2
y+C y+C

which can also be expressed in the form y + C = y0 ex where y0 = eC2 is a new constant.
 
2 y3 x3

5-25. (b) (1 + y ) dy = (1 + x2 ) dx gives y + =x+ +C
3 3

5-26. Assume solution y = et to the homogeneous equation y

+ 2 y = 0 and show

= i and = i are characteristic roots which lead to the real fundamental set
{cos t, sin t} and complementary solution yc = c1 cos t + c2 sin t.
Examine the right-hand side and its derivative to find basic terms needed for
assumed solution. The terms cos t and sin t (multiplied by some constants) are the
only basic terms and so one can assume a particular solution yp = A cos t + B sin t
where A and B are constants to be determined. Substitute this solution and its
derivatives into the nonhomogeneous equation and equate coefficients of like terms
and show A, B must be selected to satisfy
A2 + 2 A =1
B2 + 2 B =0

Solutions Chapter 5
 547
1 1
giving B = 0 and A = 2 2 . This gives the particular solution yp = 2 2 cos t.
The general solution is therefore
1
y = yc + yp = c1 cos t + c2 sin t + cos t
2 2
Resonance occurs as .

5-27.
h hy hy b
Use similar triangles and write = or x =
b/2 x h 2
 2 2
hy b
This gives element of volume dV = 4x2 dy = 4 dy
h 4
Sum these elements from 0 to h and show volume of pyramid is V = 31 b2 h.

5-28.

The box has volume V = x(w 2x)(
2x) = 4x3 2(w +
)x2 + wlx
dV
Here = 12x2 4(w +
)x + w
= 12x2 60x + 48 = (2x 8)(6x 6)
dx
d2 V 2
The second derivative When x = 1, ddxV2 < 0 hence maximum box
is 2 = 24x 60.
dx
achieved. If x = 4, solution is meaningless as the volume becomes negative.

5-29.
1 1 1
(a) Area = Mx = My =
6 15 12
m3 m5 m4
(b) Area = Mx = My =
6 15 12
2

5-30. F =
sin
2

5-32. (d) y

+ 3y2
y = 0 assume solution y = et get characteristic equation ( + 2)( +
1) = 0 with characteristic roots = 2 and = 1. This gives the fundamental set
{e2x , ex} and general solution y = c1 e2x + c2 ex , with c1 , c2 arbitray constants.
di q di dq q di 1

5-33. (b) L + = 0 or L + = 0 = i = q Separate the variables
dt C dq dt C dq LC
1
and write i di = LC q dq and then integrate to obtain
i2 1 q2 K
= +
2 LC 2 2
Here K/2 is selected as the constant of integration to help simplify the algebra. If
1 2
i = 0, q = q0 at t = 0, then K = LC q0 . This gives

dq 1 dq 1
= q00 q 2 =
= dt
dt LC 2
q0 q 2 LC

Solutions Chapter 5
548
Integrate this equation and show
   
t dq q0 t
q = q0 cos and i= = sin
LC dt LC LC

5-35. Hypothesis, y1 = y1 (x) satisfies the given differential equation so that y

+
P (x)y
+ Q(x)y = 0. If y2 = uy1 = u(x)y1 (x), then by differentiation

y2
=uy1
+ u
y1
y2

=uy1

+ 2u
y1
+ u

y1

We desire to select u = u(x) such that y2 = y2 (x) is also a function which satisfies the
differential equation. If y2

+ P (x)y2
+ Q(x)y2 = 0, the u = u(x) must be selected such
that
uy1

+ 2u
y1
+ P (x)[uy1
+ u
y1 ] + Q(x)[uy1 ] = 0

Rearrange terms and show

y2

+ P (x)y2
+ Q(x)y2 = u [y1

+ P (x)y1
+ Q(x)y1 ] +u
[2y1
+ P (x)y1 + u

y1 = 0
  
zero by hypothesis

Therefore u = u(x) must be selected to satisfy

u
[2y1
+ P (x)y1 ] + u

y1 = 0

dv
Make the substitution u
= v and u

= dx and show

1 P (x) dx
v= e
y12

then another integration gives u.

5-36. (e) Integrate with respect to y and show

u y2
= xy + + f (x), f (x) arbitrary function
y 2

Integrate again with respect to y holding x constant to obtain

xy 2 y3
u= + + yf (x) + g(x) where g(x) is another arbitrary function
2 6

(f) Integrate with respect to x and show

u
+ u = x + xy + f (y) f (y) arbitrary function
x

Solutions Chapter 5
 549
u
Now multiply by ex and show ex + uex = (uex ) = xex + xyex + f (y)ex
x x
Integrate with respect to x and show
uex = ex (x 1) + yex (x 1) + f (y)ex + g(y), g(y) arbitrary
u

5-37. (a) If x = v
y
and u
y
v
= x , then differentiate the first equation with respect
to x and differentiate the second equation with respect to y and show
2u 2v 2u 2v
= and =
x2 y x y 2 x y
2u 2u
then by addition of these equations one obtains + 2 = 0.
x2 y

5-38.

Area= A = (2x)(2y) = 4xy = 4x r2 x2 since x2 + y2 = r2
dA 4x2

= + 4 r 2 x2
dx r 2 x2
Show the derivative is zero when x = r/ 2 and y = r/ 2 and a
square is the maximum inscribed rectangle. To show it is a maxi-

mum examine dA
dx
for x < r/ 2 and dA
dx
for x > r/ 2
 2 
x 2 2 2 dy

5-39. y = e for 0 x 1 use ds = dx + dy and write ds = 1 + dx dx so
 1

that s = 1 + e2x dx To evaluate this integral make the substitution u = ex with
0  e

du
du = ex dx and show s = 1 + u2 Make another substitution w2 = 1 + u2 with
1 u
2w dw = 2u du and show
 1+e2
w dw 1
s= w
2
w 1 w 1 2
2
 1+e2 2
w 1+1
s= dw
2 w2 1
 1+e2 
dw
2
1+e2
s= dw = w]1+e tanh 1 w
2 1 w2 2 2



s = 1 + e2 2 + tanh 1 2 tanh 1 ( 1 + e2 )

5-40. The side area is As = 2rh and the ends (top and bottom) have area Ae = 2r2
where r2 h = V0 is to be satisfied. The cost of construction is
 
C = c0 (2rh) + 3c0 2r 2
V0
Substitute into this equation h = r2
to express C = C(r) as a function of r. Show
dC
=0 3
when r = V 0 /6 and h = 62/3 3 V 0 .
Show curve for C = C(r) is concave up at
dr
this point and hence a minimum is achieved.

Solutions Chapter 5
550
u u

5-41. (a) x (b) y

dV dV

5-43. V = AH so that if dt
= kA and A is a constant, then dt
= A dh
dt
= kA which
dh
implies = k is a constant. Here k is the proportionality constant.
dt
4 1

5-44. I =x tan x + sec2 x tan x + C
3 3

5-45. (b) I = 53 tan x + 13 sec2 x tan x + C

5-46. For n = 50
50
  b
f (x) f (x) dx
i=1 a

x 2 2
x2 2.6672 8/3
x3 4.0016 4
sin x 1.99934 2
cos x -1.99934 2


5-48. (x) = et tx1 dx, Integrate by parts with U = tx1 and dV = et dt to
0
obtain (x) = (x 1)(x 1) Now replace x by x + 1 to obtain (x + 1) = x(x)

(1) = et dt = 1
0
(2) =1(1) = 1

(3) =2(2) = 2 1 = 2!
(4) =3(3) = 3 2! = 3!
.. ..
. .
(n + 1) =n(n) = n (n 1)! = n!

5-49. (a) s1 (t) = s2 (t) when 2t2 + t = 11t 3t2 or t = 0 and t = 2.

ds1 ds2
= v1 = 4t + 1 = v2 = 11 6t
dt dt

At t = 0, v1 = 1 and v2 = 11. At t = 2, v1 = 9 and v2 = 1.

(b) v1 = v2 when t = 1 at positions s1 (1) = 3 and s2 (1) = 8.
(c) v1 = dsdt1 > 0 steadily increases, while v2 = dsdt2 > 0 for t < 11/6, s2 increases and
ds2
dt
< 0 for t > 11/6, then s2 decreases.

Solutions Chapter 5
 551
f f

5-50. (a) = 2x 2, and = 0 when x = 1.
x x
f f
= 2y 4 and = 0 when y = 2. f (1, 2) = 25 is a minimum value, since for
y y
all (x, y) is a neighborhood of (1, 2) we have f (x, y) > f (1, 2).

5-51.
y0
m = slope = and equation of line is
x1 x0
(y0 ) x0 y 0
y y0 = (x x0 ), when x = 0, y1 = y0 + x1 x0
.
(x1 x0 )
Therefore
 2
2 x0 y0

= x21 + y12 = x21 + y0 + , x0 , y0 f ixed
x1 x0

 1/3 2/3
Show x 1
= 0, when x1 = x0 + x0 y0 This is the value of x1 which will produce the
shortest line.

5-52. x = 3, y = 4 and x = 9, y = 12

Solutions Chapter 5
552
Index

abscissa 6
absolute maximum 117 Cauchy product 326
absolute value function 11 Cauchys mean-value theorem 111
acceleration of gravity 369 center of gravity 374
addition 325 center of mass 374
addition of series 326 centroid 375, 380
adiabatic process 425 centroid of area 377
algebraic function 20 centroid of curve 384
algebraic operations 325 centroid of composite shapes 383
alternating series test 293 chain rule differentiation 99
amplitude 403 change of variables 220
amplitude versus frequency 416 characteristic equation 407
analysis of derivative 106 characteristic roots 407, 412
angle of intersection 104, 105 charge 419
angle of intersection for lines 40 Charless law 424
arc length 238 chemical kinetics 395
Archimedes 178 chemical reaction 395
arctangent function 338 circle 18, 59
area between curves 220 circular functions 142
area polar coordinates 240, 256 circular neighborhood 275
Area under a curve 215 circumference of circle 239
arithmetic series 176 closed interval 3
asymptotic lines 55, 68 comparison test 296, 298
axis of symmetry 59, 380 complementary error function 237
complementary solution 414
B composite function 98, 315
concavity 118
belongs to 2 conditional convergence 325
Bernoulli numbers 314, 328 conic sections 57
Bessel functions 309 conic sections polar coordinates 70
bimolecular reaction 397 conjugate hyperbola 69
binomial coefficients 356 conservation of energy 374
binomial series 356 constant of integration 181
binomial theorem 92 contained in 3
Bonnets second mean value theorem 245 continued fraction 332
bounded increasing sequence 300 continuity 116
bounded sequence 324 continuous function 54, 88
bounded set 2 convergence of a sequence 272
bounds for sequence 275 convergence of series 283
Boyles law 365, 424 convergent continued fraction 336
bracketing terms 295 coordinate systems 5
cosine function 24
C critical damping 414
current 419
capacitance 419 curves 16
cartesian coordinates 5 cycles per second 403
Cauchy convergence 278, 287 cycloid 134
Cauchy form for the remainder 313 cylindrical coordinates 256

Index
 553
equations for line 36
D equivalence 5
error function 237, 310
dAlembert ratio test 302 escape velocity 369
damped oscillations 413 estimation of error 291, 294, 298
damping force 405 Euler numbers 328
de Moivres theorem 149, 168, 190 Euler-Mascheroni constant 310
decreasing functions 12 Eulers formula 147
definite integrals 213 Eulers identity 409
derivative 87 evaluation of continued fraction 334
derivative notation 90 even and odd functions 358
derivative of a product 95 even function of x 26
derivative of a quotient 97 existence of the limit 278
derivative of the logarithm 111 exponential function 21, 113
derivative of triple product 96 exterior angle 40
derivatives of inverse hyperbolic functions 149 extrema 119
derivatives of trigonometric functions 131 extreme value 162
determinants and parabola 62 extremum 119
difference between sets 3
differential equations 399 F
differentials 101
differentiation of composite function 98 finite oscillatory 277, 283
differentiation of implicit functions 102 finite oscillatory sequence 277
differentiation of integrals 247 finite sum 282
differentiation operators 90 first derivative test 120
differentiation rules 91 first law of thermodynamics 424
Dirac delta function 174 first mean value theorem for integrals 245
direction of integration 219 first moment 374
directrix 59 focal parameter 59
discontinuous function 54 focus 59
disjoint sets 3 Fourier cosine transform 235
distance between points 8 Fourier exponential transform 235
distance from point to line 171 Fourier series 339
divergence of a sequence 273 Fourier sine transform 235
divergence of series 283 frequency of motion 403
domain of definition 33 full Fourier interval 345
double integrals 249 function 271
dummy summation index 282 function changes sign 108
dummy variable of integration 184 functions 8, 20
functions defined by products 330
E functions defined by series 330
functions of two variables 159
elastic potential energy 403 fundamental theorem of integral calculus 217
electrical circuits 418
electromotive force 419 G
element of volume 226, 249
ellipse 63 Gamma function 237
empty set 1 gas pressure 394
energy 372 Gay-Lussac law 424
epsilon-delta definition of limit 46 general equation of line 38
equality of sets 3 general equation of second degree 71
equation of line 36 generalized mean value theorem for integrals 245
equation of state 425 generalized mean-value theorem 111

Index
554
generalized second mean value theorem 245 intersection 3
generalized triangle inequality 300 intersection of circles 105
geometric interpretations 273 intersection of two curves 105
geometric series 177, 287, 350, 357 interval neighborhood 275
graph compression 29 interval notation 3
graph expansion 29 interval of convergence 305
graph scaling 29 inverse functions 31, 128
graphic compression 29 inverse hyperbolic functions 153
graphs 8 inverse of differentiation 179
graphs of trigonometric functions 24 inverse operator 31
inverse trigonometric functions 34, 140
H isothermal curves 425
iterative scheme 334
half-life 427
harmonic series 283 J
harmonic series of order p 291
Heaviside 174 jump discontinuity 43, 89, 107
higher derivatives 90
higher order moments 385 K
higher partial derivatives 159
Hookes law 401 kinetic energy 372
horizontal inflection point 118 Kirchoffs laws 420
horizontal line test 31 Kronecker delta 340
hyperbola 66
hyperbolic functions 25, 142, 149 L
hyperbolic identities 145
hypergeometric function 311 LHopitals rule 138, 321
hypergeometric series 318 Lagrange form of the remainder 313
Laplace transform 235
I latus rectum 59, 67
law of exponents 145
implicit differentiation 106, 162 law of mass action 396
improper integrals 234 left-hand limits 40
increasing functions 12 left-handed derivative 89
indefinite integral 180 Leibnitz 85
indeterminate forms 43, 322 Leibnitz differentiation rule 168
inductance 419 Leibnitz formula 247
infinite oscillatory 283 Leibnitz rule 248
infinite series 281 length of curve 238
infinitesimals 41 limit 46, 272
inner product 339 limit of a sequence 272
integral notation 182 limit of function 42
integral sign 181 limit point of sequence 277
integral test 288 limit theorem 50
integral used to define functions 236, 248 limiting value 43
integration 179 limits 40, 46, 304
integration by parts 209, 232 linear dependence 13
integration of derivatives 183 linear homogeneous differential equation 410
integration of polynomials 183 linear independence 13
intercept form for line 38 linear spring 403
intercepts 38 lines 36
intermediate value property 54 liquid pressure 393
intersecting lines 40, 104 local maximum 107, 117, 161

Index
 555
local minimum 107, 117, 161 O
logarithm base e 23
logarithmic differentiation 127 odd function of x 26
logarithmic function 21, 111 one-to-one correspondence 271
log-log paper 171 one-to-one function 31, 33
lower bound 2, 275 open interval 3
operator 90
M operator box 90, 180
order of reaction 396
Maclaurin Series 311, 315 ordered pairs 33, 55
mapping 271 ordinate 6
maxima 107, 116, 161 orientation of the surface 252
mean value theorem for integrals 245, 289 orthogonal intersection 40
mean value theorem 108, 245 orthogonal intersection 104
mechanical resonance 412 orthogonal lines 105
method of undetermined coefficients 414 orthogonal sequence 340
minima 107, 116, 161 orthonormal 340
mirror image 33 oscillating sequence 277
modification of series 324 overdamping 413
moment of force 374
moment of inertia of solid 392 P
moment of inertial of area 390
moment of inertia of composite shapes 393 parabola 60
moments of inertia 385 parallel circuit 423
momentum 367 parallelepiped volume elements 252
monotone decreasing 107, 245, 277 parametric equations 129
monotone increasing 107, 245, 277 parametric equation for line 38
multiple-valued functions 14 parametric representation 17, 134
multiplication 325 partial denominators 333
partial derivatives 158, 160
N partial fractions 195, 284
partial numerators 333
natural logarithm 23 partial sums 282
natural logarithm function 236 particular solution 414
necessary condition for convergence 286 period of oscillation 403
negative slope 37 periodic motion 403
neighborhoods 275 perpendicular distance 39
Newton 85 perpendicular lines 39
Newton root finding 353 phase shift 403
Newtons law of gravitation 368 piecewise continuous 345
Newtons laws 366 piecewise continuous functions 11
nonconvergence 277 piston 174
nonrectangular regions 252 plane curves 14
not in 2 plotting programs 14
notation for limits 40, 42 point of inflection 118
notations for derivatives 90 point-slope formula 37, 88
nth term test 286 polar coordinates 5
n-tuples 2 polar coordinates 240, 255
null sequence 277 polar form for line 39
number pairs 5 polar graph 14
polynomial function 20, 95
positive monotonic 245
positive slope 37

Index
556
potential energy 373 Riemann sum 215
power rule 100 right circular cone 172
power rule for differentiation 99 right-hand limits 40
power series 305 right-handed derivative 89
pressure 393 Rolles theorem 107, 246
pressure-volume diagram 425 rotation of axes 30
principal branches 35 rules for differentiation 91
product rule 95
products of sines and cosines 193 S
Proof of Mean Value Theorems 246
proper subsets 3 scale factors 29
properties of definite integrals 218 scaling for integration 183
properties of integrals 181 scaling of axes 28
properties of limits 50 Schlomilch and Roche remainder term 320
p-series 291 secant line 85
Pythagorean identities 203 second derivative test 120
Pythagorean theorem 8 second derivatives 90
second law of thermodynamics 424
Q second moments 385
sectionally continuous 43
quadrants 6 semi-convergent series 325
quotient rule 97 semi-log paper 171
sequence of partial sums 282
R sequence of real numbers 271
sequences and functions 274
radioactive decay 426 series 281, 325
radius of convergence 305 series circuit 421
radius of Earth 369 set complement 4
range of function 33 set operations 3
rate of reaction 395 set theory 1
ratio comparison test 297 sets 1
ratio test 302 shearing modulus 418
rational function 20 shift of index 282
ray from origin 7 shifting of axes 28
rectangular coordinates 6 shorthand representation 282
rectangular graph 14 signed areas 220
reductio ad absurdum 276 simple harmonic motion 136, 403
reduction formula 211 simple pendulum 418
reflection 29 sine function 24
refraction 123 sine integral function 310
regular continued fractions 336 single-valued function 14, 31
related rates 363 slicing method 231
relative maximum 107, 117, 161 slope 37, 85
relative minimum 107, 117, 161 slope changes 120
remainder term 313, 319 slope condition for orthogonality 104
Remainder Term for Taylor Series 319 slope of line 36
representation of functions 337 slope-intercept form for line 38
resistance 419 slowly converging series 301
resonance 412 slowly diverging series 301
resonance frequency 416 smooth curve 107
restoring force 403 smooth function 116
reverse reaction rate 395 Snells law 122
reversion of series 354 solids of revolution 225

Index
 557
special functions 309 total differential 159
special limit 304 transcendental function 21
special sums 176 transformation equations 28
special trigonometric integrals 194 translation of axes 28
spherical coordinates 257 transverse axis 67
spring-mass system 401 triangular numbers 284
squeeze theorem 53, 275 trigonometric functions 24, 129
Stolz -Cesaro theorem 279 trigonometric substitutions 189
subscript notation 160 truncation of series 286
subsequence 277 two point equations of line 36
subsets 3 two-point formula 37
subtraction 325
subtraction of series 326 U
summation notation 175
summation of forces 401 union 3
sums and differences of squares 202 units of measurement 367
surface area 242 universal set 1
surface of revolution 242 upper bound 2, 275
symmetric functions 26 using table of integrals 258
symmetry 26, 31, 380
V

Venn diagram 4
T vertex 59
voltage drop 419
table of centroids 382 volume of sphere 172
table of derivatives 156 volume under a surface 253
table of differentials 157
table of integrals 186, 208 W
table of moments of inertia 390
tangent function 24 weight function 340
tangent line 86 weight of an object 369
Taylor series 311, 315, 318 work 371
Taylor series two variables 315 work done 403
telescoping series 284, 351
terminology for sequences 277 Z
thermodynamics 424
torque 374 zero slope 37
torsional vibrations 417 zeroth law of thermodynamics 424
total derivative 160

Index
 Public Domain Images
Courtesy of Wikimedia.Commons