DALE HOFFMAN

C O N T E M P O R A RY
CALCULUS
Contents

0 Welcome to Calculus 1
0.1 A Preview of Calculus . . . . . . . . . . . . . . . . . . . . 1
0.2 Lines in the Plane . . . . . . . . . . . . . . . . . . . . . . . 7
0.3 Functions and Their Graphs . . . . . . . . . . . . . . . . . 20
0.4 Combinations of Functions . . . . . . . . . . . . . . . . . 30
0.5 Mathematical Language . . . . . . . . . . . . . . . . . . . 45

1 Limits and Continuity 53

1.0 Tangent Lines, Velocities, Growth . . . . . . . . . . . . . . 53
1.1 The Limit of a Function . . . . . . . . . . . . . . . . . . . 62
1.2 Properties of Limits . . . . . . . . . . . . . . . . . . . . . . 73
1.3 Continuous Functions . . . . . . . . . . . . . . . . . . . . 84
1.4 Definition of Limit . . . . . . . . . . . . . . . . . . . . . . 98

2 The Derivative 109

2.0 Introduction to Derivatives . . . . . . . . . . . . . . . . . 109
2.1 The Definition of Derivative . . . . . . . . . . . . . . . . . 117
2.2 Derivatives: Properties and Formulas . . . . . . . . . . . 130
2.3 More Differentiation Patterns . . . . . . . . . . . . . . . . 144
2.4 The Chain Rule . . . . . . . . . . . . . . . . . . . . . . . . 155
2.5 Applications of the Chain Rule . . . . . . . . . . . . . . . 165
2.6 Related Rates . . . . . . . . . . . . . . . . . . . . . . . . . 175
2.7 Newton’s Method . . . . . . . . . . . . . . . . . . . . . . . 187
2.8 Linear Approximation and Differentials . . . . . . . . . . 196
2.9 Implicit and Logarithmic Differentiation . . . . . . . . . 208

3 Derivatives and Graphs 217

3.1 Finding Maximums and Minimums . . . . . . . . . . . . 217
3.2 Mean Value Theorem . . . . . . . . . . . . . . . . . . . . . 229
3.3 The First Derivative and the Shape of f . . . . . . . . . . 238
3.4 The Second Derivative and the Shape of f . . . . . . . . . 248
3.5 Applied Maximum and Minimum Problems . . . . . . . 258
3.6 Asymptotic Behavior of Functions . . . . . . . . . . . . . 272
3.7 L’Hôpital’s Rule . . . . . . . . . . . . . . . . . . . . . . . . 283
A Answers A1

D Derivative Facts A29

H How to Succeed in Calculus A31

T Trigonometry Facts A35

0
Welcome to Calculus
Calculus was first developed more than 300 years ago by Sir Isaac Newton and Gottfried Leibniz to help
them describe and understand the rules governing the motion of planets and moons. Since then, thousands
of other men and women have refined the basic ideas of calculus, developed new techniques to make the
calculations easier, and found ways to apply calculus to a wide variety of problems besides planetary motion.
They have used calculus to help understand physical, biological, economic and social phenomena and to
describe and solve related problems.
The discovery, development and application of calculus is a great intellectual achievement—and now you
have the opportunity to share in that achievement. You should feel exhilarated. You may also be somewhat
concerned (a common reaction among students just beginning to study calculus). You need to be concerned
enough to work to master calculus, yet confident enough to keep going when you (at first) don’t understand
something.
Part of the beauty of calculus is that it relies upon a few very simple ideas. Part of the power of calculus is
that these simple ideas can help us understand, describe and solve problems in a variety of fields. This book
tries to emphasize both the beauty and the power.
In Section 0.1 (Preview) we will look at the main ideas that will continue throughout the book: the problems
of finding tangent lines and computing areas. We will also consider a process that underlies both of those
problems: the limiting process of approximating a solution and then getting better and better approximations
until we finally get an exact solution.
Sections 0.2 (Lines), 0.3 (Functions) and 0.4 (Combinations of Functions) contain review material. These
sections emphasize concepts and skills you will need in order to succeed in calculus. You should have worked
with most of these concepts in previous courses, but the emphasis and use of the material here may be
different than in those earlier classes.
Section 0.5 (Mathematical Language) discusses a few key mathematical phrases. It considers their use and
meaning and some of their equivalent forms. It will be difficult to understand the meaning and subtleties of
calculus if you don’t understand how these phrases are used and what they mean.

0.1 A Preview of Calculus

Calculus can be viewed as an attempt—a historically successful attempt—to solve two fundamental problems.
In this section we begin to examine geometric forms of those two problems and some fairly simple attempts
to solve them. At first, the problems themselves may not appear very interesting or useful—and the methods
for solving them may seem crude—but these simple problems and methods have led to one of the most
beautiful, powerful and useful creations in mathematics: Calculus.
2 contemporary calculus

Finding the Slope of a Tangent Line

Suppose we have the graph of a function y = f ( x ) and we want to find
an equation of a line tangent to the graph at a particular point P on the
graph (see margin). (We will offer a precise definition of “tangent” in
Section 1.0; for now, think of the tangent line as a line that touches the
curve at P and stays close to the graph of y = f ( x ) near P.)
We know that P is on the tangent line, so if its x-coordinate is x = a,
then the y-coordinate of P must be y = f ( a): P = ( a, f ( a)). The only
other information we need to find an equation of the tangent line is its
slope, mtan , but that is where the difficulty arises.
In algebra, we needed two points in order to determine a slope. So
far, we only have the point P. Let’s simply pick a second point, call
it Q, on the graph of y = f ( x ). If the x-coordinate of Q is b, then the
y-coordinate is f (b): Q = (b, f (b)). So the slope of the line through P
and Q is
rise f (b) − f ( a)
m PQ = =
run b−a
If we drew the graph of y = f ( x ) on a wall, put nails at the points P
and Q, and laid a straightedge on the nails, then the straightedge would
have slope m PQ . But the slope m PQ can be very different from the value
we want (the slope mtan of the tangent line). The key idea is that when
the point Q is close to the point P, then the slope m PQ should be close to
the slope we want, mtan . Physically, if we slide the nail at Q along the
f (b)− f ( a)
graph toward the fixed point P, then the slope, m PQ = b− a , of the
straightedge gets closer and closer to the slope, mtan , of the tangent
line. If the value of b is very close to a, then the point Q is very close to
P, and the value of m PQ is very close to the value of mtan .
Rather than defacing walls with graphs and nails, we can instead
f (b)− f ( a)
calculate m PQ = b− a and examine the values of m PQ as b gets
closer and closer to a. We say that mtan is the limiting value of m PQ as
b gets very close to a, and we write:
f (b) − f ( a)
mtan = lim
b→ a b−a
Eventually we will call the slope mtan of the tangent line the deriva-
tive of the function f ( x ) at the point P, and call this part of calculus
differential calculus. Chapters 2 and 3 begin the study of differential
calculus.
The slope of the tangent line to the graph of a function will tell us
important information about the function and will allow us to solve
problems such as:

• The U.S. Postal Service requires that the length plus the girth of a
package not exceed 84 inches. What is the largest volume that can
be mailed in a rectangular box?
 welcome to calculus 3

• An oil tanker was leaking oil and a 4-inch-thick oil slick had formed.
When first measured, the slick had a radius of 200 feet, and the
radius was increasing at a rate of 3 feet per hour. At that time, how
fast was the oil leaking from the tanker?

Derivatives will even help us find solutions to equations such as

x2 = sin( x ) and x9 + 5x5 + x3 + 3 = 0.

Problems

1. Sketch the lines tangent to the curve shown below 3. The graph below shows the temperature of a cup
at x = 1, 2 and 3. Estimate the slope of each of of coffee during a 10-minute period. (Each of
the tangent lines you drew. your answers in (a)–(c) should have the units “de-
grees per minute.”)

(a) What was the average rate of cooling from

minute 0 to minute 10?
(b) What was the average rate of cooling from
minute 7 to minute 8? From minute 8 to
minute 9?
(c) What was the rate of cooling at minute 8? At
minute 2?
(d) When was cold milk added to the coffee?

2. A graph of the weight of a “typical” child from

age 0 to age 24 months appears below. (Your
answers should have the units “kg per month.”)
(a) What was the average weight gain from
month 0 to month 24?
(b) What was the average weight gain from
month 9 to month 12? From month 12 to
month 15?
(c) Approximately how fast was the child gaining
weight at age 12 months? At age 3 months?

4. Describe a method for determining the slope of a

steep hill at a point midway up the hill

(a) using a ruler, a long piece of string, a glass of

water and a loaf of bread.
(b) using a protractor, a piece of string and a
helium-filled balloon.
4 contemporary calculus

Finding the Area of a Shape

Suppose we need to find the area of a leaf as part of a study of how
much energy a plant gets from sunlight. One method for finding the
area would be to trace the shape of the leaf onto a piece of paper
and then divide the region into “easy” shapes such as rectangles and
triangles (whose areas we can easily calculate). We could add all of
these “easy” areas together to approximate the area of the leaf.
A modification of this method would be to trace the shape onto a
piece of graph paper and then count the number of squares completely
inside the edge of the leaf to get a lower estimate of the area, and count
the number of squares that touch the leaf to get an upper estimate of
the area. If we repeat this process with smaller and smaller squares, we
will have to do more counting and adding, but our estimates should be
closer together—and closer to the actual area of the leaf.

(We could also approximate the area of the leaf using a sheet of
paper, scissors and an accurate scale. How?)
We can calculate the area A between the graph of a function y = f ( x )
and the x-axis by using similar methods. We can divide the area into
strips of width w and determine the lower and upper values of y = f ( x )
on each strip. Then we can approximate the area of each rectangle and
add all of the little areas together to get Aw , an approximation of the
exact area. The key idea is that if w is small, then the rectangles are
narrow, and the approximate area Aw should be very close to the actual
area A. If we take narrower and narrower rectangles, the approximate
areas get closer and closer to the actual area:

A = lim Aw
w →0

The process described above is the basis for a technique called

integration, and this part of calculus is called integral calculus. Integral
calculus and integration will begin in Chapter 4.
The process of taking the limit of a sum of “little” quantities will
give us important information about a function and will also allow us
to solve problems such as:
 welcome to calculus 5

• Find the length of the graph of y = sin( x ) over one period (from
x = 0 to x = 2π).

• Find the volume of a torus (“doughnut”) of radius 1 inch that has a

hole of radius 2 inches.

• A car starts at rest and has an acceleration of 5 + 3 sin(t) feet per

second per second in the northerly direction at time t seconds. Where
will the car be, relative to its starting position, after 100 seconds?

Problems
5. Approximate the area of the leaf on the previous page using

(a) the grid on the left.

(b) the grid on the right.

6. A graph showing temperatures during the month of November

appears below.

(a) Approximate the shaded area between the temperature curve and
the 65◦ line from Nov. 15 to Nov. 25.
(b) The area of the “rectangle” is (base)(height) so what are the units
of your answer in part (a)?
(c) Approximate the shaded area between the temperature curve and
the 65◦ line from Nov. 5 to Nov. 30.
(d) Who might use or care about these results?

7. Describe a method for determining the volume of a compact flu-

orescent light bulb using a ruler, a large can, a scale and a jug of
water.
6 contemporary calculus

A Unifying Process: Limits

We used similar processes to “solve” both the tangent line problem
and the area problem. First, we found a way to get an approximate
solution, and then we found a way to improve our approximation.
Finally, we asked what would happen if we continued improving our
approximations “forever”: that is, we “took a limit.”
For the tangent line problem, we let the point Q get closer and closer
and closer to P, the limit as b approached a.
In the area problem, we let the widths of the rectangles get smaller
and smaller, the limit as w approached 0. Limiting processes underlie
derivatives, integrals and several other fundamental topics in calculus,
and we will examine limits and their properties in detail in Chapter 1.

Two Sides of the Same Coin: Differentiation and Integration

Just as the set-up of each of the two basic problems involved a limiting
process, the solutions to the two problems are also related. The process
of differentiation used to solve the tangent line problem and the process
of integration used to solve the area problem turn out to be “opposites”
of each other: each process undoes the effect of the other process. The
Fundamental Theorem of Calculus in Chapter 4 will show how this
“opposite” effect works.

Extensions of the Main Problems

The first five chapters present the two key ideas of calculus, show “easy”
ways to calculate derivatives and integrals, and examine some of their
applications. And there is more.
Through the ensuing chapters, we will examine new functions and
find ways to calculate their derivatives and integrals. We will extend
the approximation ideas to use “easy” functions, such as polynomials,
to approximate the values of “hard” functions, such as sin( x ) and e x .
In later chapters, we will extend the notions of “tangent lines” and
“areas” to 3-dimensional space as “tangent planes” and “volumes.”

Success in calculus will require time and effort on your part, but
such a beautiful and powerful field is worth that time and effort.
 welcome to calculus 7

0.2 Lines in the Plane

The first graphs and functions you encountered in algebra were straight
lines and their equations. These lines were easy to graph, and the
equations were easy to evaluate and to solve. They described a variety
of physical, biological and financial phenomena such as d = rt relating
the distance d traveled to the rate r and time t spent traveling, and
C = 59 ( F − 32) for converting the temperature in degrees Fahrenheit
(F) to degrees Celsius (C).
The first part of calculus — differential calculus — will deal with
ideas, techniques and applications of tangent lines to the graphs of
functions, so it is important that you understand the graphs, properties
and equations of straight lines.

The Real Number Line

The real numbers (consisting of all integers, fractions, rational and
irrational numbers) can be represented as a line, called the real number
line. Once we have selected a starting location, called the origin, a
positive direction (usually up or to the right), and unit of length, then
every number can be located as a point on the number line. If we
move from a point x = a to a point x = b on the line, then we will
have moved an increment of b − a. We denote this increment with the
symbol ∆x (read “delta x” ).

• If b is larger than a, then we will have moved in the positive direction,

The capital Greek letter delta (∆) appears
and ∆x = b − a will be positive. often in calculus to represent the
“change” in something.
• If b is smaller than a, then ∆x = b − a will be negative and we will
have moved in the negative direction. Caution: ∆x does not mean ∆ times x,
but rather the difference between two x-
coordinates.
• Finally, if ∆x = b − a = 0, then a = b and we did not move at all.

We can also use the ∆ notation and absolute values to express the
distance that we have moved. On the number line, the distance from
x = a to x = b is
(
b − a if b ≥ a
dist( a, b) =
a − b if b < a
or: q
dist( a, b) = |b − a| = |∆x | = (∆x )2
The midpoint of the interval from x = a to x = b is the point M
such that dist( a, M) = dist( M, b), or | M − a| = |b − M |. If a < M < b,
a+b
M − a = b − M ⇒ 2M = a + b ⇒ M =
2
It’s not difficult to check that this formula also works when b < M < a.
8 contemporary calculus

Example 1. Find the length and midpoint of the interval from x = −3

to x = 6.

(−3)+6
Solution. dist(−3, 6) = |6 − (−3)| = |9| = 9; M = 2 = 23 . ◀

Solutions to Practice problems are at the Practice 1. Find the length and midpoint of the interval from x = −7
end of each section. to x = −2.

The Cartesian Plane

Two perpendicular number lines, called coordinate axes, determine
a real number plane. The axes intersect at a point called the origin.
Each point P in the plane can be described by an ordered pair ( x, y) of
numbers that specify how far, and in which directions, we must move
from the origin to reach the point P. We can locate the point P = ( x, y)
in the plane by starting at the origin and moving x units horizontally
and then y units vertically. Similarly, we can label each point in the
plane with the ordered pair ( x, y), which directs us how to reach that
point from the origin.
This coordinate system is called the rectangular coordinate system
In this book, a point in the plane will be
labeled either with a name, say P, or with
or the Cartesian coordinate system (after René Descartes), and the
an ordered pair, say ( x, y), or with both: resulting plane the Cartesian plane.
P = ( x, y). The coordinate axes divide the plane into four quadrants, labeled
quadrants I, II, III and IV moving counterclockwise from the upper-right
quadrant.
We will often call the horizontal axis the x-axis and the vertical axis
the y-axis and then refer to the plane as the xy-plane. This choice of x
and y as labels for the axes is a common choice, but we will sometimes
prefer to use different labels — and even different units of measurement
on the two axes.

Increments and Distance Between Points In The Plane

If we move from a point P = ( x1 , y1 ) in the plane to another point
Q = ( x2 , y2 ), then we will need to consider two increments or changes.

• The increment in the x (horizontal) direction is x2 − x1 , denoted by

∆x = x2 − x1 .

• The increment in the y (vertical) direction is y2 − y1 , denoted by

∆y = y2 − y1 .

Computing the distance between the points P = ( x1 , y1 ) and Q =

( x2 , y2 ) involves a simple application of the Pythagorean Theorem:
q q
dist( P, Q) = (∆x )2 + (∆y)2 = ( x2 − x1 )2 + ( y2 − y1 )2
 welcome to calculus 9

The midpoint M of the line segment joining P and Q is:

 
x1 + x2 y1 + y2
M= ,
2 2
where we have just used the one-dimension midpoint formula for each
coordinate.

Example 2. Find an equation describing all the points P = ( x, y)

equidistant from Q = (2, 3) and R = (5, −1).

Solution. The points P = ( x, y) must satisfy dist( P, Q) = dist( P, R)

so: q q
( x − 2)2 + (y − 3)2 = ( x − 5)2 + (y − (−1))2
By squaring each side we get:

( x − 2)2 + ( y − 3)2 = ( x − 5)2 + ( y + 1)2

Expanding we get:

x2 − 4x + 4 + y2 − 6y + 9 = x2 − 10x + 25 + y2 + 2y + 1

and canceling like terms yields:

−4x − 6y + 13 = −10x + 2y + 26

so y = 0.75x − 1.625, the equation of a line. Every point on the line

y = 0.75x − 1.625 is equally distant from both Q and R. ◀

Practice 2. Find an equation describing all points P = ( x, y) equidistant

from Q = (1, −4) and R = (0, −3).

A circle with radius r and center at the point C = ( a, b) consists of

all points P = ( x, y) at a distance of r from the center C: the points P
that satisfy dist( P, C ) = r.

Example 3. Find an equation of a circle with radius r = 4 and center

C = (5, −3).

Solution. A circle consists of the set of points P = ( x, y) at a fixed

distance r from the center point C, so this circle will be the set of points
P = ( x, y) at a distance of 4 units from the point C = (5, −3); P will be
on this circle if dist( P, C ) = 4.
Using the distance formula and rewriting:
q
( x − 5)2 + (y + 3)2 = 4 ⇒ ( x − 5)2 + (y + 3)2 = 16

which we can also express as x2 − 10x + 25 + y2 + 6y + 9 = 16. ◀

Practice 3. Find an equation of a circle with radius r = 5 and center

C = (−2, 6).
10 contemporary calculus

The Slope Between Points in the Plane

In one dimension (on the number line), our only choice was to move
in the positive direction (so the x-values were increasing) or in the
negative direction. In two dimensions (in the plane), we can move
in infinitely many directions, so we need a precise way to describe
direction.
The slope of the line segment joining P = ( x1 , y1 ) to Q = ( x2 , y2 ) is

rise y − y1 ∆y
m = slope from P to Q = = 2 =
run x2 − x1 ∆x

The slope of a line measures how fast we rise or fall as we move

from left to right along the line. It measures the rate of change of the
y-coordinate with respect to changes in the x-coordinate. Most of our
work will occur in two dimensions, and slope will be a very useful
concept that will appear often.
If P and Q have the same x-coordinate, then x1 = x2 ⇒ x = 0. The
∆y
line from P to Q is thus vertical and the slope m = ∆x is undefined
because ∆x = 0.
If P and Q have the same y-coordinate, then y1 = y2 ⇒ ∆y = 0, so
∆y 0
the line is horizontal and the slope is m = ∆x = ∆x = 0 (assuming
∆x ̸= 0).

Practice 4. For P = (−3, 2) and Q = (5, −14), find ∆x, ∆y, and the
slope of the line segment from P to Q.

If the coordinates of P or Q contain variables, then the slope m is

∆y
still given by m = ∆x , but we will need to use algebra to evaluate and
simplify m.

Example 4. Find the slope of the line segment from P = (1, 3) to

Q = (1 + h, 3 + 2h).

Solution. y1 = 3 and y2 = 3 + 2h, so ∆y = (3 + 2h) − (3) = 2h; x1 = 1

and x2 = 1 + h, so ∆x = (1 + h) − (1) = h. The slope is:

∆y 2h
m= = =2
∆x h
In this example, the value of m is constant (2) and does not depend
on the value of h. ◀

Practice 5. Find the slope and midpoint of the line segment from
P = (2, −3) to Q = (2 + h, −3 + 5h).

Example 5. Find the slope between the points P = ( x, x2 + x ) and

Q = ( a, a2 + a) for a ̸= x.
 welcome to calculus 11

Solution. y1 = x2 + x and y2 = a2 + a ⇒ ∆y = ( a2 + a) − ( x2 + x );
x1 = x and x2 = a, so ∆x = a − x and the slope is:

∆y ( a2 + a ) − ( x 2 + x )
m= =
∆x a−x
a2 − x 2 + a − x ( a − x )( a + x ) + ( a − x )
= =
a−x a−x
( a − x ) (( a + x ) + 1)
= = (a + x) + 1
a−x
Here the value of m depends on the values of both a and x. ◀

Practice 6. Find the slope between the points P = ( x, 3x2 + 5x ) and

Q = ( a, 3a2 + 5a) for a ̸= x.

In application problems, it is important to read the information and

the questions very carefully — including the units of measurement of
the variables can help you avoid “silly” answers.

Example 6. In 1970, the population of Houston was 1, 233, 535 and in

1980 it was 1, 595, 138. Find the slope of the line through the points
(1970, 1233535) and (1980, 1595138).
∆y 1595138 − 1233535 361603
Solution. m = = = = 36, 160.3 but
∆x 1980 − 1970 10
36, 160.3 is just a number that may or may not have any meaning to
you. If we include the units of measurement along with the numbers
we will get a more meaningful result:

1595138 people − 1233535 people 361603 people people

= = 36, 160.3
year 1980 − year 1970 10 years year

which says that during the decade from 1970 to 1980 the population of
Houston grew at an average rate of 36, 160 people per year. ◀

If the x-unit is time (in hours) and the y-unit is distance (in kilome-
ters), then
∆y km
m=
∆x hours
so the units for m are km (“kilometers per hour”), a measure of
hour
velocity, the rate of change of distance with respect to time.
If the x-unit is the number of employees at a bicycle factory and the
y-unit is the number of bicycles manufactured, then

∆y bicycles
m=
∆x employees

bicycles
and the units for m are (“bicycles per employee”), a measure
employee
of the rate of production per employee.
12 contemporary calculus

Equations of Lines
Every (non-vertical) line has the property that the slope of the segment
between any two points on the line is the same, and this constant slope
property of straight lines leads to ways of finding equations to represent
non-vertical lines.

Point-Slope Form
In calculus, we will usually know a point on a line and the slope of that
line, so the point-slope form will be the easiest to apply. Other forms
of equations for lines can be derived from the point-slope form.
If L is a non-vertical line through a known point P = ( x1 , y1 ) with a
known slope m, then the equation of the line L is:

Point-Slope: y − y1 = m( x − x1 )

Example 7. Find an equation of the line through (2, −3) with slope 5.

Solution. We can simply use the point-slope formula: m = 5, y1 = −3

and x1 = 2, so y − (−3) = 5( x − 2), which simplifies to y = 5x − 13 ◀

An equation for a vertical line through a point P = ( a, b) is x = a.

All points Q = ( x, y) on the vertical line through the point P have the
same x-coordinate as P.

Two-Point Form
If two points P = ( x1 , y1 ) and Q = ( x2 , y2 ) are on the line L, then we
can calculate the slope between them and use the first point and the
point-slope equation to find an equation for L:

y2 − y1
Two-Point: y − y1 = m( x − x1 ) where m =
x2 − x1

Once we have the slope, m, it does not matter whether we use P or

Q as the point. Either choice will result in the same equation for the
line once we simplify it.

Slope-Intercept Form
It is common practice to rewrite an equation of a line into the form
y = mx + b, the slope-intercept form of the line. The line y = mx + b
has slope m and crosses the y-axis at the point (0, b).
∆y
Practice 7. Use the ∆x definition of slope to calculate the slope of the
line y = mx + b.
 welcome to calculus 13

The point-slope and the two-point forms are usually more useful for
finding an equation of a line, but the slope-intercept form is usually the
most useful form for an answer because it allows us to easily picture
the graph of the line and to quickly calculate y-values given x-values.

Angles Between Lines

The angle of inclination of a line with the x-axis is the smallest angle θ
that the line makes with the positive x-axis as measured from the x-axis
∆y
counterclockwise to the line. Because the slope m = ∆x and because
opposite
tan(θ ) = in a right triangle, m = tan(θ ).
adjacent

The slope of a line is the tangent of its angle of inclination.

Parallel Lines
Two parallel lines L1 and L2 make equal angles with the x-axis, so their
angles of inclination will be equal and hence so will their slopes.
Similarly, if the slopes, m1 and m2 , of two lines are equal, then the
equations of the lines (in slope-intercept form) will always differ by a
constant:

y1 − y2 = (m1 x + b1 ) − (m2 x + b2 ) = (m1 − m2 ) x + (b1 − b2 ) = b1 − b2

which is a constant, so the lines will be parallel.

The two preceding ideas can be combined into a single statement:

Two non-vertical lines L1 and L2 with slopes m1 and m2

are parallel if and only if m1 = m2 .

Practice 8. Find an equation of the line that contains the point (−2, 3)
and is parallel to the line 3x + 5y = 17.

Perpendicular Lines
If two lines are perpendicular, the situation is a bit more complicated.
Assume L1 and L2 are two non-vertical lines that intersect at the
origin (for simplicity), with P = ( x1 , y1 ) and Q = ( x2 , y2 ) points away
from the origin on L1 and L2 , respectively. Then the slopes of L1 and L2
y y
will be m1 = x1 and m2 = x22 . The line connecting P and Q forms the
1
third side of triangle OPQ, which will be a right triangle if and only if
L1 and L2 are perpendicular. In particular, L1 and L2 are perpendicular
if and only if the triangle OPQ satisfies the Pythagorean Theorem:

(dist(O, P))2 + (dist(O, Q))2 = (dist( P, Q))2

14 contemporary calculus

or:

( x1 − 0)2 + ( y1 − 0)2 + ( x2 − 0)2 + ( y2 − 0)2

= ( x1 − x2 )2 + ( y1 − y2 )2

Squaring and simplifying, this reduces to 0 = −2x1 x2 − 2y1 y2 , so:

y2 x y x 1 1
= − 1 ⇒ m 2 = 2 = − 1 = − y1 = −
x2 y1 x2 y1 x
m1
1

We have just proved the following result:

Two non-vertical lines L1 and L2 with slopes m1 and m2

are perpendicular if and only if their slopes are negative
1
reciprocals of each other: m2 = − .
m1

Practice 9. Find an equation of the line that goes through the point
(2, −5) and is perpendicular to the line 3y − 7x = 2.

Example 8. Find the distance (that is, the shortest distance) from the
point (1, 8) to the line L : 3y − x = 3.

Solution. This is a sophisticated problem that requires several steps to

solve: First we need a picture of the problem. We will find an equation
for the line L∗ through the point (1, 8) and perpendicular to L. Then
we will find the point P where L and L∗ intersect. Finally, we will find
the distance from P to (1, 8).
Step 1: L has slope 13 so L∗ has slope m = − 11 = −3, and L∗ has
3
equation y − 8 = −3( x − 1), which simplifies to y = −3x + 11.
Step 2: We can find the point where L intersects L∗ by replacing the
y-value in the equation for L with the y-value from our equation for L∗ :

3(−3x + 11) − x = 3 ⇒ x = 3 ⇒ y = −3x + 11 = −3(3) + 11 = 2

which tells us that L and L∗ intersect at P = (3, 2).

Step 3: Finally, the distance from L to (1, 8) is just the distance from
the point (1, 8) to the point P = (3, 2), which is
q √
(1 − 3)2 + (8 − 2)2 = 40 ≈ 6.325
√
The distance is (exactly) 40, or (approximately) 6.325. ◀

Angle Formed by Intersecting Lines

If two lines that are not perpendicular intersect at a point (and neither
line is vertical), then we can use some geometry and trigonometry to
determine the angles formed by the intersection of those lines.
 welcome to calculus 15

Because θ2 (see figure at right) is an exterior angle of the triangle

ABC, θ2 is equal to the sum of the two opposite interior angles, so
θ2 = θ1 + θ ⇒ θ = θ2 − θ1 . From trigonometry, we then know that:

tan(θ2 ) − tan(θ1 ) m − m1
tan(θ ) = tan(θ2 − θ1 ) = = 2
1 + tan(θ2 ) tan(θ1 ) 1 + m2 m1
 
The range of the arctan function is − π2 , π2 , so θ = arctan 1m+2m−2mm1
 
1
always gives the smaller of the angles. The larger angle is π − θ (or
180◦ − θ ◦ if we measure the angles in degrees).

The smaller angle θ formed by two non-perpendicular  lines with

m2 − m1

slopes m1 and m2 is: θ = arctan
1 + m2 m1

Example 9. Find the point of intersection and the angle between the
lines y = x + 3 and y = 2x + 1.

Solution. Solving the first equation for y and then substituting into
the second equation:

( x + 3) = 2x + 1 ⇒ x=2 ⇒ y = 2+3 = 5

The point of intersection is (2, 5). Because both lines are in slope-
intercept form, it is easy to see that m1 = 1 and m2 = 2:

m2 − m1 2−1
   
θ = arctan = arctan
1 + m2 m1 1+2·1
 
1
= arctan ≈ 0.322 radians = 18.43◦
3

The lines intersect at an angle of (approximately) 18.43◦ . ◀

0.2 Problems

1. Estimate the slope of each line shown below. 2. Estimate the slope of each line shown below.
16 contemporary calculus

3. Compute the slope of the line that passes through: 7. The blocks in a city are all perfect squares. A
(a) (2, 4) and (5, 8) friend gives you directions to a good restaurant:
(b) (−2, 4) and (3, −5) “Go north 3 blocks, turn east and go 5 blocks, turn
(c) (2, 4) and ( x, x2 ) south and go 7 blocks, turn west and go 3 blocks.”
How far away (straight-line distance) is it?
(d) (2, 5) and (2 + h, 1 + (2 + h)2 )
(e) ( x, x2 + 3) and ( a, a2 + 3) 8. At the restaurant (see previous problem), a fellow
diner gives you directions to a hotel: “Go north 5
4. Compute the slope of the line that passes through:
blocks, turn right and go 6 blocks, turn right and
(a) (5, −2) and (3, 8)
go 3 blocks, turn left and go 2 blocks.” How far
(b) (−2, −4) and (5, −3) away is the hotel from the restuarant?
(c) ( x, 3x + 5) and ( a, 3a + 5)
9. The bottom of a 20-foot ladder is 4 feet from the
(d) (4, 5) and (4 + h, 5 − 3h) base of a wall.
(e) (1, 2) and ( x, 1 + x2 )
(a) How far up the wall does the ladder reach?
(f) (2, −3) and (2 + h, 1 − (1 + h)2 )
(b) What is the slope of the ladder?
(g) ( x, x2 ) and ( x + h, x2 + 2xh + h2 )
(c) What angle does it make with the ground?
(h) ( x, x2 ) and ( x − h, x2 − 2xh + h2 )
10. Let P = (1, −2) and Q = (5, 4). Find:
5. A small airplane at an altitude of 5, 000 feet is
flying east at 300 feet per second (a bit over 200 (a) the midpoint R of the line segment PQ.
1
miles per hour), and you are watching it with a (b) the point T that is 3 of the way from P to Q:
small telescope as it passes directly overhead. 1
dist( P, T ) = dist( P, Q)
(a) What is the slope of the telescope 5, 10 and 20 3
seconds after the plane passes overhead? 2
(c) the point S that is 5 of the way from P to Q.
(b) What is the slope of the telescope t seconds
11. If P = (2, 3), Q = (8, 11) and R = ( x, y), where:
after the plane passes overhead?
(c) After the plane passes overhead, is the slope of x = 2a + 8(1 − a), y = 3a + 11(1 − a), 0 ≤ a ≤ 1
the telescope increasing, decreasing or staying
the same? (a) Verify that R is on the line segment PQ.
(b) Verify that dist( P, R) = (1 − a) · dist( P, Q).
12. A rectangular box is 24 inches long, 18 inches
wide and 12 inches high.
(a) Find the length of the longest (straight) stick
that will fit into the box.
(b) What angle (in degrees) does that stick make
with the base of the box?
13. The lines y = x and y = 4 − x intersect at (2, 2).
(a) Show that the lines are perpendicular.
6. You are at the origin, (0, 0), and are watching a (b) Graph the lines together on your calculator
small bug at the point (t, 1 + t2 ) at time t seconds. using the “window” [−10, 10] × [−10, 10].
(a) What is the slope of your line of vision when (c) Why do the lines not appear to be perpendicu-
t = 5, 10 and 15 seconds? lar on the calculator display?
(b) What is the slope of your line of vision at an (d) Find a suitable window so that the lines do
arbitrary time t? appear perpendicular.
 welcome to calculus 17

14. Two lines both go through the point (1, 2), one 20. Find an equation of the circle with center C and
with slope 3 and one with slope − 13 . radius r when
(a) Find equations for the lines. (a) C = (2, 7) and r = 4
(b) Choose a suitable window so that the lines will (b) C = (3, −2) and r = 1
appear perpendicular, and then graph them to- (c) C = (−5, 1) and r = 7
gether on your calculator.. (d) C = (−3, −1) and r = 4
15. Sketch the line with slope m that goes through 21. Explain how to show, without graphing, whether
the point P, then find an equation for the line. a point P = ( x, y) is inside, on, or outside the
(a) m = 3, P = (2, 5) circle with center C = (h, k) and radius r.
(b) m = −2, P = (3, 2) 22. A box with a base of dimensions 2 cm and 8 cm
(c) m = − 12 , P = (1, 4) is definitely big enough to hold two semicircular
rods with radii of 2 cm (see below).
16. Sketch the line with slope m that goes through
the point P, then find an equation for the line. (a) Will these same two rods fit in a box 2 cm high
and 7.6 cm wide?
(a) m = 5, P = (2, 1)
(b) Will they fit in a box 2 cm high and 7.2 cm
(b) m = − 23 , P = (1, 3) wide? (Suggestion: Turn one of the rods over.)
(c) m = π, P = (1, −3)
17. Find an equation for each line.
(a) L1 goes through the point (2, 5) and is parallel
to 3x − 2y = 9.
(b) L2 goes through the point (−1, 2) and is per- 23. Show that an equation of the circle with center
pendicular to 2x = 7 − 3y. C = (h, k) and radius r is ( x − h)2 + (y − k)2 = r2 .
(c) L3 goes through the point (3, −2) and is per- 24. Find an equation of the line tangent to the circle
pendicular to y = 1. x2 + y2 = 25 at the point P when:
18. Find a value for the constant (A, B or D) so that: (a) P = (3, 4)

(a) the line y = 2x + A goes through (3, 10). (b) P = (−4, 3)

(b) the line y = Bx + 2 goes through (3, 10). (c) P = (0, 5)

(c) the line y = Dx + 7 crosses the y-axis at the (d) P = (−5, 0)

point (0, 4). 25. Find an equation of the line tangent to the circle
(d) the line Ay = Bx + 1 goes through the points with center C = (3, 1) at the point P when:
(1, 3) and (5, 13). (a) P = (8, 13)
19. Find the shortest distance between the circles with (b) P = (−10, 1)
centers C1 = (1, 2) and C2 = (7, 10) with radii r1 (c) P = (−9, 6)
and r2 when: (d) P = (3, 14)
(a) r1 = 2 and r2 = 4 26. Find the center C = (h, k ) and the radius r of the
(b) r1 = 2 and r2 = 7 circle that goes through the three points:
(c) r1 = 5 and r2 = 8 (a) (0, 1), (1, 0) and (0, 5)
(d) r1 = 3 and r2 = 15 (b) (1, 4), (2, 2) and (8, 2)
(e) r1 = 12 and r2 = 1 (c) (1, 3), (4, 12) and (8, 4)
 18 contemporary calculus

27. How close does 28. How close does

(a) the line 3x − 2y = 4 come to the point (2, 5)? (a) the line 2x − 5y = 4 come to the point (1, 5)?
(b) the line y = 5 − 2x come to the point (1, −2)? (b) the line y = 3 − 2x come to the point (5, −2)?
(c) the circle with radius 3 and center at (2, 3) (c) the circle with radius 4 and center at (4, 3)
come to the point (8, 3)? come to the point (10, 3)?

29. Follow the steps below (and refer to the figure) to find a formula for
the distance from the origin to the line Ax + By = C.

(a) Show that the line L given by Ax + By = C has slope m = − A

B.
(b) Find the equation of the line L∗ that goes through (0, 0) and is
perpendicular to L.
(c) Show that L and L∗ intersect at the point:
 
AC BC
( x, y) = ,
A2 + B2 A2 + B2

(d) Show that the distance from the origin to the point ( x, y) is:

|C |
√
A2 + B2

30. Show that a formula for the distance from the point ( p, q) to the line
Ax + By = C is:
| Ap + Bq − C |
√
A2 + B2
(The steps will be similar to those in the previous problem, but the
algebra will be more complicated.)
 welcome to calculus 19

0.2 Practice Answers

1. Length = dist(−7, −2) = |(−7) − (−2)| = | − 5| = 5.
(−7) + (−2) −9
The midpoint is at = = −4.5.
2 2
2. dist( P, Q) = dist( P, R) ⇒ ( x − 1)2 + (y + 4)2 = ( x − 0)2 + (y + 3)2 ;
squaring each side and simplifying eventually yields y = x − 4.

3. The point P = ( x, y) is on the circle when it is 5 units from the center

C = (−2, 6), so dist( P, C ) = 5. Then dist (( x, y), (−2, 6)) = 5, so
p
( x + 2)2 + (y − 6)2 = 5 ⇒ ( x + 2)2 + (y − 6)2 = 25.

4. ∆x = 5 − (−3) = 8 and ∆y = −14 − 2 = −16, so:

∆y −16
slope = = = −2
∆x 8

∆y (−3 + 5h) − (−3) 5h

5. slope = = = = 5. The midpoint is at
 ∆x ( 2 + h ) − 2   h 
(2) + (2 + h) (−3 + 5h) + (−3) h 5h
, = 2+ , −3 + .
2 2 2 2

∆y (3a2 + 5a) − (3x2 + 5x ) 3( a2 − x 2 ) + 5( a − x )

6. slope = = = =
∆x a−x a−x
3( a + x )( a − x ) + 5( a − x )
= 3( a + x ) + 5
a−x
7. Let y1 = mx1 + b and y2 = mx2 + b. Then:

∆y (mx2 + b) − (mx1 + b) m ( x2 − x1 )
slope = = = =m
∆x x2 − x1 x2 − x1

8. The line 3x + 5y = 17 has slope − 35 , so the slope of the parallel line

is m = − 35 . Using the form y = − 35 x + b and the point (−2, 3) on
the line, we have 3 = − 35 (−2) + b ⇒ b = 95 ⇒ y = − 35 x + 59 , or
5y + 3x = 9.

9. The line 3y − 7x = 2 has slope 73 , so the slope of the perpendicular

line is m = − 37 . Using the form y = − 37 x + b and the point (2, −5)
on the line, we have −5 = − 37 (2) + b ⇒ b = − 29 3
7 ⇒ y = −7x − 7 ,
29

or 7y + 3x = −29.
20 contemporary calculus

0.3 Functions and Their Graphs

When you prepared for calculus, you learned to manipulate functions

by adding, subtracting, multiplying and dividing them, as well as
evaluating functions of functions (composition). In calculus, we will
continue to work with functions and their applications. We will create
new functions by operating on old ones. We will gather information
from the graphs of functions and from derived functions. We will find
ways to describe the point-by-point behavior of functions as well as
their behavior “close to” certain points and also over entire intervals.
We will find tangent lines to graphs of functions and areas between
graphs of functions. And, of course, we will see how these ideas can be
used in a variety of fields.
This section and the next one review information and procedures
you should already know about functions before we begin calculus.

What is a function?
Let’s begin with a (very) general definition of a function:

A function from a set X to a set Y is a rule for assigning to each

element of the set X a single element of the set Y. A function
assigns a unique (exactly one) output element from the set Y to
each input element from the set X.

The rule that defines a function is often given in the form of an

equation, but it could also be given in words or graphically or by a
table of values. In practice, functions are given in all of these ways, and
we will use all of them in this book.
In the definition of a function, the set X of all inputs is called the
domain of the function. The set Y of all outputs produced from these
inputs is called the range of the function. Two different inputs (elements
in the domain) can be assigned to the same output (an element in the
range) but one input cannot lead to two different outputs.
Most of the time we will work with functions whose domains and
ranges are real numbers, but there are other types of functions all
around us. Final grades for this course provide an example of a function.
For each student, the instructor will assign a final grade based on some
rule for evaluating that student’s performance. The domain of this
function consists of all students registered for the course, and the range
consists of grades (either letters like A, B, C, D, F, or numbers like 4.0,
3.7, 2.0, 1.7, 0.0). Two students can receive the same final grade, but
only one grade can be assigned to each student.
 welcome to calculus 21

Function Machines
Functions are abstract structures, but sometimes it is easier to think of
them in a more concrete way. One such way is to imagine that a function
is a special-purpose computer, a machine that accepts inputs, does
something to those inputs according to a defining rule, and produces
an output. The output is the value of the function associated with the
given input value. If the defining rule for a function f is “multiply the
input by itself,” f (input) = (input)(input), then the figure and table
in the margin show the results of putting the inputs x, 5, 2.5, a, c + 3
and x + h into the machine f .

Practice 1. If we have a function machine g whose rule is “divide 3 by

the input and add 1,” g( x ) = 3x + 1, what outputs do we get from the
inputs x, 5, a, c + 3 and x + h? What happens if the input is 0? input output
x x2
You expect your calculator to behave as a function: each time you a a2
press the same sequence of keys (input) you expect to see the same 2.5 6.25
display (output). In fact, if your calculator did not produce the same x+h ( x + h )2

output each time, you would need a new calculator. On many calculators there is a feature
that does not produce the same output
each time you use it. What is it?
Functions Defined by Equations
If the domain of a function consists of a collection of real numbers
(perhaps all real numbers) and the range is also a collection of real
numbers, then the function is called a numerical function. We can give
the rule for a numerical function in several ways, but usually write it
as a formula. If the rule for a numerical function, f , is “the output is
the input number multiplied by itself,” then we could write the rule as

f ( x ) = x · x = x2

The use of an “x” to represent the input is simply a matter of conve-

nience and custom. We could also represent the same function by:

• f ( a ) = a2

• f (#) = #2 or

• f (input) = (input)2 .

For the function f defined by f ( x ) = x2 − x, we can see that:

• f ( 3 ) = 32 − 3 = 6

• f (0.5) = (0.5)2 − (0.5) = −0.25

• f (−2) = (−2)2 − (−2) = 6

Notice that the two different inputs 3 and −2 both lead to the output
of 6. That is allowable for a function.
22 contemporary calculus

We can also evaluate f if the input contains variables. If we replace

the “x” with something else in the notation “ f ( x ),” then we must
replace the “x” with the same thing everywhere in the formula:

• f ( c ) = c2 − c

• f ( a + 1) = ( a + 1)2 − ( a + 1) = ( a2 + 2a + 1) − ( a + 1) = a2 + a

• f ( x + h) = ( x + h)2 − ( x + h) = ( x2 + 2xh + h2 ) − ( x + h)

and, in general: f (input) = (input)2 − (input)

For more complicated expressions, we can just proceed step-by-step:

( x + h )2 − ( x + h ) − x 2 − x



f ( x + h) − f ( x )
=
h h
( x2 + 2xh + h2 ) − ( x + h) − ( x2 − x )


=
h
2
2xh + h − h h(2x + h − 1)
= = = 2x + h − 1
h h

Practice 2. For the function g defined by g(t) = t2 − 5t, evaluate g(1),

g( x +h)− g( x )
g(−2), g(w + 3), g( x + h), g( x + h) − g( x ) and h .

Functions Defined by Graphs and Tables of Values

The graph of a numerical function f consists of a plot of ordered pairs
( x, y) where x is in the domain of f and y = f ( x ), such as the graph
of f ( x ) = sin( x ) for −4 ≤ x ≤ 7 in the margin. A table of values of
a numerical function consists of a list of (some of) the ordered pairs
( x, y) where y = f ( x ).
A function can be defined by a graph or by a table of values, and
these types of definitions are common in applied fields. The outcome
of an experiment will depend on the input, but the experimenter may
not know the “rule” for predicting the outcome. In that case, the
experimenter usually represents the function of interest as a table of
measured outcome values versus input values, or as a graph of the
outcomes versus the inputs. The table and graph in the margin show
the deflections obtained when weights were loaded at the end of a
wooden stick. The next graph shows the temperature of a hot cup of tea
as a function of the time as it sits in a 68◦ F room. In these experiments,
the “rule” for the function is that f (input) = actual outcome of the
experiment.
Tables have the advantage of presenting the data explicitly, but it is
often difficult to detect patterns simply from lists of numbers.
Graphs usually obscure some of the precision of the data, but more
easily allow us to detect patterns visually — we can actually see what is
happening with the numbers
 welcome to calculus 23

Creating Graphs of Functions

Most people understand and can interpret pictures more quickly than
tables of data or equations, so if we have a function defined by a table
of values or by an equation, it is often useful (and necessary) to create
a picture of the function: a graph.

A Graph from a Table of Values

length (in.) weight (lbs.)
If we have a table of values for a function, perhaps consisting of mea-
13.5 0.4
surements obtained from an experiment, then we can plot the ordered 14.5 0.9
pairs in the xy-plane to get a graph consisting of a collection of points. 15.0 0.7
16.0 0.9
The table in the margin shows the lengths and weights of trout
18.0 1.2
caught (and released) during several days of fishing. The graph plots 18.5 1.6
those values along with a line that comes “close” to the plotted points. 19.5 1.5
20.5 1.7
From the graph, you could estimate that a 17-inch trout should weigh 20.5 2.1
slightly more than one pound.

A Graph from an Equation

Creating the graph of a function given by an equation is similar to
creating one from a table of values: we need to plot enough points
( x, y) where y = f ( x ) so we can be confident of the shape and location
of the graph of the entire function. We can find a point ( x, y) that
satisfies y = f ( x ) by picking a value for x and then calculating the
value for y by evaluating f ( x ). Then we can enter the ( x, y) value in a
table or simply plot the point ( x, y).
If you recognize the form of the equation and know something about
the shape of graphs of that form, you may not have to plot many points.
If you do not recognize the form of the equation, then you will need to
plot more points, maybe 10 or 20 or 234: it depends on how complicated
the graph appears and on how important it is to you (or your boss) to
have an accurate graph. Evaluating y = f ( x ) at many different values
for x and then plotting the points ( x, y) is usually not very difficult, but
it can be very time-consuming. Fortunately, calculators and computers
can often do the evaluation and plotting for you.

Is Every Graph the Graph of a Function?

The definition of “function” requires that each element of the domain
(each input value) be sent by the function to exactly one element of
the range (to exactly one output value), so for each input x-value there
will be exactly one output y-value, y = f ( x ). The points ( x, y1 ) and
( x, y2 ) cannot both be on the graph of f unless y1 = y2 . The graphical
interpretation of this result is called the Vertical Line Test.
24 contemporary calculus

Vertical Line Test for a Function: A graph is the graph of a

function if and only if a vertical line drawn through any point
in the domain intersects the graph at exactly one point.

The figure in the margin shows the graph of a function, followed by

two graphs that are not graphs of functions, along with vertical lines
that intersect those graphs at more than one point. Non-functions are
not “bad,” and are often necessary to describe complicated phenomena.

Reading Graphs Carefully

Calculators and computers can help students, reporters, financial an-
alysts and scientists create graphs quickly and easily. Because of this,
graphs are being used more often than ever to present information and
justify arguments. This text takes a distinctly graphical approach to the
ideas and meaning of calculus. Calculators and computers can help us
create graphs, but we need to be able to read them carefully. The next
examples illustrate some types of information that can be obtained by
carefully reading and understanding graphs.

Example 1. A boat starts from St. Thomas and sails due west with the
velocity shown in the margin.

(a) When is the boat traveling the fastest?

(b) What does a negative velocity away from St. Thomas mean?
(c) When is the boat the farthest away from St. Thomas?

Solution. (a) The greatest speed is 10 mph at t = 3 hours. (b) It means

that the boat is heading back toward St. Thomas. (c) The boat is farthest
from St. Thomas at t = 6 hours. For t < 6 the boat’s velocity is positive,
and the distance from the boat to St. Thomas is increasing. For t > 6
the boat’s velocity is negative, and the distance from the boat to St.
Thomas is decreasing. ◀

Practice 3. You and a friend start out together and hike along the same
trail but walk at different speeds, as shown in the figure.

(a) Who is walking faster at t = 20?

(b) Who is ahead at t = 20?
(c) When are you and your friend farthest apart?
(d) Who is ahead when t = 50?

Example 2. Which has the largest slope: the line through the points A
and P, the line through B and P, or the line through C and P?

Solution. The line through C and P: m PC > m PB > m PA . ◀

 welcome to calculus 25

Practice 4. In the figure, the point Q on the curve is fixed, and the
point P is moving to the right along the curve toward the point Q.
As P moves toward Q, is the indicated value increasing, decreasing,
remaining constant, or doing something else?

(a) x-coordinate of P
(b) x-increment from P to Q
(c) slope from P to Q

Example 3. The graph of y = f ( x ) appears in the margin. Let g( x ) be

the slope of the line tangent to the graph of f ( x ) at the point ( x, f ( x )).

(a) Estimate the values g(1), g(2) and g(3).

(b) For what value(s) of x is g( x ) = 0?
(c) At what value(s) of x is g( x ) largest?
(d) Sketch the graph of y = g( x ).

Solution. (a) The figure in the margin shows the graph of y = f ( x ) x f (x) g( x )

with several tangent lines to the graph of f . From this graph, we can 0 −1.0 0.5
1 0.0 1.0
estimate that g(1) (the slope of the line tangent to the graph of f at 2 2.0 0.0
(1, 0)) is approximately equal to 1. Similarly, g(2) ≈ 0 and g(3) ≈ −1. 3 1.0 −1.0
(b) The slope of the tangent line appears to be horizontal (slope = 0) 4 0.0 −1.0
5 −1.0 0.0
at x = 2 and at x = 5. 6 −0.5 0.5
(c) The tangent line to the graph of f appears to have greatest slope
(be steepest) near x = 1.5.
(d) We can build a table of values of g( x ) and then sketch the graph
of these values. A graph of y = g( x ) appears below. ◀

Practice 5. Water flows into a container (see margin) at a constant rate

of 3 gallons per minute. Starting with an empty container, sketch the
graph of the height of the water in the container as a function of time.
26 contemporary calculus

0.3 Problems

√ x
In Problems 1–4, match the numerical triples to the 7. If f ( x ) = x2 + 3, g( x ) = x − 5 and H ( x ) = x −2 :
graphs. For example, in Problem 1, A: 3, 3, 6 is (a) evaulate f (1), g(1) and H (1).
“over and up” so it matches graph (a). (b) graph f ( x ), g( x ) and H ( x ) for −5 ≤ x ≤ 10.
1. A: 3, 3, 6; B: 12, 6, 6; C: 7, 7, 3 D: 2, 4, 4 (c) evaluate f (3x ), g(3x ) and H (3x ).
(d) evaluate f ( x + h), g( x + h) and H ( x + h).
8. Find the slope of the line through the points P
and Q when:
2. A: 7, 10, 7; B: 17, 17, 25; C: 4, 4, 8 D: 12, 8, 16 (a) P = (1, 3), Q = (2, 7)
(b) P = ( x, x2 + 2), Q = ( x + h, ( x + h)2 + 2)
(c) P = (1, 3), Q = ( x, x2 + 2)
(d) P, Q as in (b) with x = 2, x = 1.1, x = 1.002
3. A: 7, 14, 10; B: 23, 45, 22; C: 8, 12, 8 D: 6, 9, 3
9. Find the slope of the line through the points P
and Q when:
(a) P = (1, 5), Q = (2, 7)
4. A: 6, 3, 9; B: 8, 1, 1; C: 12, 6, 9 D: 3.7, 1.9, 3.6 (b) P = ( x, x2 + 3x − 1),
Q = ( x + h, ( x + h)2 + 3( x + h) − 1)
(c) P, Q as in (b) with x = 1.3, x = 1.1, x = 1.002
10. If f ( x ) = x2 + x and g( x ) = 3x , evaluate and sim-
5. Water is flowing at a steady rate into each of the f ( a + h) − f ( a) g( a + h) − g( a)
plify and when
bottles shown below. Match each bottle shape h h
a = 1, a = 2, a = −1, a = x.
with the graph of the height of the water as a √
function of time 11. If f ( x ) = x2 − 2x and g( x ) = x, evaluate and
f ( a + h) − f ( a) g( a + h) − g( a)
simplify and
h h
when a = 1, a = 2, a = 3, a = x.
12. The temperatures shown below were recorded
during a 12-hour period in Chicago.
(a) At what time was the temperature the highest?
Lowest?
(b) How fast was the temperature rising at 10 a.m.?
At 1 p.m.?
(c) What could have caused the drop in tempera-
ture between 1 p.m. and 3 p.m.?
6. Sketch shapes of bottles that will have the water
height versus time graphs shown below.
 welcome to calculus 27

13. The graph below shows the distance of an air- 15. Refer to the curve shown below.
plane from an airport during a long flight. (a) Sketch the lines tangent to the curve at x = 1,
2, 3, 4 and 5.
(a) How far was the airplane from the airport at 1
p.m.? At 2 p.m.? (b) For what value(s) of x is the value of the func-
tion largest? Smallest?
(b) How fast was the distance changing at 1 p.m.?
(c) For what value(s) of x is the slope of the tan-
(c) How could the distance from the plane to the gent line largest? Smallest?
airport remain unchanged from 1:45 p.m. until
2:30 p.m. without the airplane falling?

16. The figure below shows the height of the wa-

ter (above and below mean sea level) at a Maine
beach.
(a) At which time(s) was the most beach exposed?
14. The graph below shows the height of a diver The least?
above the water level at time t seconds. (b) At which time(s) was the current the strongest?

(a) What was the height of the diving board?

(b) When did the diver hit the water?
(c) How deep did the diver get?
(d) When did the diver return to the surface?

17. Imagine that you are ice skating, from left to

right, along the path shown below. Sketch the
path you will follow if you fall at points A, B and
C.
 28 contemporary calculus

18. Define s( x ) to be the slope of the line through 21. Using the graph of y = f ( x ) below, let g( x ) be
the points (0, 0) and ( x, f ( x )) where f ( x ) is the the slope of the line tangent to the graph of f ( x )
function graphed below. For example, s(3) = at the point ( x, f ( x )). Complete the table, esti-
slope of the line through (0, 0) and (3, f (3)) = 43 . mating values of the slopes as best you can.
(a) Evaluate s(1), s(2) and s(4).
(b) For which integer value of x between 1 and 7
is s( x ) smallest?

19. Let f ( x ) = x + 1 and define s( x ) to be the slope

of the line through the points (0, 0) and ( x, f ( x )),
as shown below. For example, s(2) = slope of the x f (x) g( x )
line through (0, 0) and (2, 3) = 23 . 0 1 1
(a) Evaluate s(1), s(3) and s(4). 1
(b) Find a formula for s( x ). 2
3
4

22. Sketch the graphs of water height versus time for

water pouring into a bottle shaped like:

(a) a milk carton

(b) a spherical glass vase
(c) an oil drum (cylinder) lying on its side

20. Define A( x ) to be the area of the rectangle (d) a giraffe

bounded by the coordinate axes, the line y = 2 (e) you
and a vertical line at x, as shown below. For
23. Design bottles whose graphs of (highest) water
example, A(3) = area of a 2 × 3 rectangle = 6.
height versus time will look like those shown
(a) Evaluate A(1), A(2) and A(5). below.
(b) Find a formula for A( x ).
 welcome to calculus 29

0.3 Practice Answers

1. Create an input-output table using the function rule:

input output
3
x x +1
3
5 5 + 1 = 1.6
3
a a +1
3
c+3 c +3 + 1
3
x+h x+h + 1
3
If x = 0, then g(0) = 0 + 1 is not defined (because of division by 0).

2. g(t) = t2 − 5t
g ( 1 ) = 12 − 5 ( 1 ) = − 4
g(−2) = (−2)2 − 5(−2) = 14
g(w + 3) = (w + 3)2 − 5(w + 3) = w2 + 6w + 9 − 5w − 15 = w2 + w − 6
g( x + h) = ( x + h)2 − 5( x + h) = x2 + 2xh + h2 − 5x − 5h
g( x + h) − g( x ) = ( x2 + 2xh + h2 − 5x − 5h) − ( x2 − 5x ) = 2xh + h2 − 5h
g( x + h) − g( x ) 2xh + h2 − 5h
= = 2x + h − 5
h h
3. (a) Friend (b) Friend (c) At t = 40. Before that, your friend is walking
faster and increasing the distance between you. Then you start to
walk faster than your friend and start to catch up. (d) Friend. You
are walking faster than your friend at t = 50, but you still have not
caught up.

4. (a) The x-coordinate is increasing. (b) The x-increment ∆x is decreas-

ing. (c) The slope of the line through P and Q is decreasing.

5. See the figure:

30 contemporary calculus

0.4 Combinations of Functions

Sometimes a physical or economic situation behaves differently depend-

ing on various circumstances. In these situations, a more complicated
formula may be needed to describe the situation.

Multiline Definitions of Functions: Putting Pieces Together

Sales Tax: Some states have different rates of sales tax depending
on the type of item purchased. As an example, for many years food
purchased at restaurants in Seattle was taxed at a rate of 10%, while
most other items were taxed at a rate of 9.5% and food purchased at
grocery stores had no tax assessed. We can describe this situation by
using a multiline function: a function whose defining rule consists of
several pieces. Which piece of the rule we need to use will depend on
what we buy. In this example, we could define the tax T on an item
that costs x to be:


 0 if x is the cost of a food at a grocery store
T (x) = 0.10x if x is the cost of food at a restaurant

 0.095x if x is the cost of any other item

To find the tax on a $2 can of stew, we would use the first piece of
the rule and find that the tax is $0. To find the tax on a $30 restaurant
bill, we would use the second piece of the rule and find that the tax is
$3.00. The tax on a $150 textbook requires using the third rule: the tax
would be $14.25.
Wind Chill Index: The rate at which a person’s body loses heat
depends on the temperature of the surrounding air and on the speed
of the air. You lose heat more quickly on a windy day than you do on a
day with little or no wind. Scientists have experimentally determined
this rate of heat loss as a function of temperature and wind speed, and
the resulting function is called the Wind Chill Index, WCI. The WCI
is the temperature on a still day (no wind) at which your body would
lose heat at the same rate as on the windy day. For example, the WCI
value for 30◦ F air moving at 15 miles per hour is 9◦ F: your body loses
heat as quickly on a 30◦ F day with a 15 mph wind as it does on a 9◦ F
day with no wind.
If T is the Fahrenheit temperature of the air and v is the speed of the
wind in miles per hour, then the WCI can be expressed as a multiline
function of the wind speed v (and of the temperature T):


 √
T if 0 ≤ v ≤ 4
10.45+6.69 v−0.447v
WCI = 91.4 − 22 (91.5 −
T ) if 4 < v ≤ 45

1.60T − 55 if v > 45

 welcome to calculus 31

The WCI value for a still day (0 ≤ v ≤ 4 mph) is just the air
temperature. The WCI for wind speeds above 45 mph are the same
as the WCI for a wind speed of 45 mph. The WCI for wind speeds
between 4 mph and 45 mph decrease as the wind speeds increase. This
WCI function depends on two variables: the temperature and the wind
speed; but if the temperature is constant, then the resulting formula for
WCI only depends on the wind speed. If the air temperature is 30◦ F
(T = 30), then the formula for the Wind Chill Index is:

30◦ if 0 ≤ v ≤ 4 mph
√


WCI30 = 62.19 − 18.70 v + 1.25v if 4 mph < v ≤ 45 mph
−7◦ if v > 45 mph



The WCI graphs for temperatures of 40◦ F, 30◦ F and 20◦ F appear below:
From UMAP Module 658, “Windchill,”
by William Bosch and L.G. Cobb, 1984.

Practice 1. A Hawaiian condo rents for $380 per night during the tourist
season (from December 15 through April 30), and for $295 per night
otherwise. Define a multiline function that describes these rates.
Example 1. Define f ( x ) by:


2 if x < 0
f (x) = 2x if 0 ≤ x < 2

 1 if 2 < x
Evaluate f (−3), f (0), f (1), f (4) and f (2). Graph y = f ( x ) on the
interval −1 ≤ x ≤ 4.

Solution. To evaluate the function at different values of x, we must

first decide which line of the rule applies. If x = −3 < 0, then we
need to use the first line, so f (−3) = 2. When x = 0 or x = 1, we
need the second line of the function definition, so f (0) = 2(0) = 0 and
f (1) = 2(1) = 2. At x = 4 we need the third line, so f (4) = 1. Finally,
at x = 2, none of the lines apply: the second line requires x < 2 and
the third line requires 2 < x, so f (2) is undefined. The graph of f ( x )
appears in the margin. Note the “hole” above x = 2, which indicates
f (2) is not defined by the rule for f . ◀
32 contemporary calculus

Practice 2. Define g( x ) by:



 x if x < −1

 2 if − 1 ≤ x < 1
g( x ) =



− x if 1 < x ≤ 3
1 if 4 < x


Graph y = g( x ) for −3 ≤ x ≤ 6 and evaluate g(−3), g(−1), g(0), g( 21 ),

g(1), g( π3 ), g(2), g(3), g(4) and g(5).
Practice 3. Write a multiline definition for the function whose graph
appears in the margin.
We can think of a multiline function as a machine that first examines
the input value to decide which line of the function rule to apply:

Composition of Functions: Functions of Functions

Basic functions are often combined with each other to describe more
complicated situations. Here we will consider the composition of
functions — functions of functions.

The composite of two functions f and g, written f ◦ g, is:

f ◦ g( x ) = f ( g( x ))

The domain of the composite function f ◦ g( x ) = f ( g( x )) consists

of those x-values for which g( x ) and f ( g( x )) are both defined: we can
evaluate the composition of two functions at a point x only if each step
in the composition is defined.
If we think of our functions as machines, then composition is simply
a new machine consisting of an arrangement of the original machines.
 welcome to calculus 33

The composition f ◦ g of the function machines f and g shown in the

margin is an arrangement of the machines so that the original input
x goes into machine g, the output from machine g becomes the input
into machine f , and the output from machine f is our final output. The
composition of the function machines f ◦ g( x ) = f ( g( x )) is only valid
if x is an allowable input into g (that is, x is in the domain of g) and if
g( x ) is then an allowable input into f .
The composition g ◦ f involves arranging the machines so the original
input goes into f , and the output from f then becomes the input for g
(see right side of margin figure).
√
Example 2. For f ( x ) = x − 2, g( x ) = x2 and
(
3x if x < 2
h( x ) =
x − 1 if 2 ≤ x

evaluate f ◦ g(3), g ◦ f (6), f ◦ h(2) and h ◦ g(−3). Find the formulas

and domains of f ◦ g( x ) and g ◦ f ( x ).
√ √
Solution. f ◦ g(3) = f ( g(3)) = f (32 ) = f (9) = 9 − 2 = 7 ≈ 2.646;
√ √
g ◦ f (6) = g ( f (6)) = g( 6 − 2) = g( 4) = g(2) = 22 = 4; f ◦ h(2) =
√ √
f (h(2)) = f (2 − 1) = f (1) = 1 − 2 = −1, which is undefined;
h ◦ g(−3) = h ( g(−3)) = h(9) = 9 − 1 = 8; f ◦ g( x ) = f ( g( x )) =
√
f ( x2 ) = x2 − 2, and the domain of f ◦ g consists of those x-values for
√
which x2 − 2 ≥ 0, so the domain of f ◦ g is all x such that x ≥ 2 or
√ √ √
2
x ≤ − 2; g ◦ f ( x ) = g ( f ( x )) = g( x − 2) = x − 2 = x − 2, but
this last equality is true only when x − 2 ≥ 0 ⇒ x ≥ 2, so the domain
of g ◦ f is all x ≥ 2. ◀
x
√
Practice 4. For f ( x ) = x− 3 , g( x ) = 1 + x and
(
2x if x ≤ 1
h( x ) =
5−x if 1 < x

evaluate f ◦ g(3), f ◦ g(8), g ◦ f (4), f ◦ h(1), f ◦ h(3), f ◦ h(2) and

h◦ g(−1). Find formulas for f ◦ g( x ) and g ◦ f ( x ).

Shifting and Stretching Graphs

Some common compositions are fairly straightforward; you should
recognize the effect of these compositions on graphs of the functions.

Example 3. The margin figure shows the graph of y = f ( x ). Graph

(a) 2 + f ( x )
(b) 3 · f ( x )
(c) f ( x − 1)
34 contemporary calculus

Solution. All of the new graphs appear in the margin.

(a) Adding 2 to all of the values of f ( x ) rigidly shifts the graph of

f ( x ) upward 2 units.
(b) Multiplying all of the values of f ( x ) by 3 leaves all of the roots
(zeros) of f fixed: if x is a root of f then f ( x ) = 0 ⇒ 3 · f ( x ) =
3(0) = 0 so x is also a root of 3 · f ( x ). If x is not a root of f ,
then the graph of 3 f ( x ) looks like the graph of f ( x ) stretched
vertically by a factor of 3.
(c) The graph of f ( x − 1) is the graph of f ( x ) rigidly shifted 1 unit
to the right. We could also get these results by examining the
graph of y = f ( x ), creating a table of values for f ( x ) and the
new functions:
x f (x) 2 + f (x) 3 f (x) x−1 f ( x − 1)
−1 −1 1 −3 −2 f (−2) not defined
0 0 2 0 −1 f (0 − 1) = −1
1 1 3 3 0 f (1 − 1) = 0
2 1 3 3 1 f (2 − 1) = 1
3 2 4 6 2 f (3 − 1) = 1
4 0 2 0 3 f (4 − 1) = 2
5 −1 1 −3 4 f (5 − 1) = 0

and then graphing those new functions. ◀

If k is a positive constant, then

• the graph of y = k + f ( x ) will be the graph of y = f ( x ) rigidly

shifted up by k units

• the graph of y = k f ( x ) will have the same roots as the graph of f ( x )

and will be the graph of y = f ( x ) vertically stretched by a factor of k

• the graph of y = f ( x − k) will be the graph of y = f ( x ) rigidly

shifted right by k units

• the graph of y = f ( x + k) will be the graph of y = f ( x ) rigidly

shifted left by k units

Practice 5. The figure in the margin shows the graph of g( x ). Graph:

(a) 1 + g( x )
(b) 2g( x )
(c) g( x − 1)
(d) −3g( x )
 welcome to calculus 35

Iteration of Functions
Certain applications feed the output from a function machine back into
the same machine as the new input. Each time through the machine is
called an iteration of the function.
5
+x x
Example 4. Suppose f ( x ) = and we start with the input x = 4
2
and repeatedly feed the output from f back into f . What happens?

Solution. Creating a table:

iteration input output

5
4 +4
1 4 f (4) = 2 = 2.625000000
5
2.625 +2.625
2 2.625000000 f ( f (4)) = 2 = 2.264880952
3 2.264880952 f ( f ( f (4))) = 2.236251251
4 2.236251251 2.236067985
5 2.236067985 2.236067977
6 2.236067977 2.236067977

Once we have obtained the output 2.236067977, we will just keep getting
the same output (to 9 decimal places). You might recognize this output
√
value as an approximation of 5.
√
This algorithm always finds ± 5. If we start with any positive input,
√
the values will eventually get as close to 5 as we want. Starting with
√
any negative value for the input will eventually get us close to − 5.
We cannot start with x = 0, as 05 is undefined. ◀

Practice 6. What happens if we start with the input value x = 1 and

9
+x
iterate the function f ( x ) = x several times? Do you recognize the
2
resulting number? What do you think will happen to the iterates of
A
+x
g( x ) = x ? (Try several positive values of A.)
2

Two Useful Functions: Absolute Value and Greatest Integer

Two functions (one of which should be familiar to you, the other
perhaps not) possess useful properties that let us describe situations in
which an object abruptly changes direction or jumps from one value to
another value. Their graphs will have corners and breaks, respectively.
The absolute value function evaluated at a number x, y = f ( x ) =
| x |, is the distance between the number x and 0. Some calculators and computer program-
If x is greater than or equal to 0, then | x | is simply x − 0 = x. If x is ming languages represent the absolute
value function by abs( x ) or ABS( x ).
negative, then | x | is 0 − x = − x = −1 · x, which is positive because:

−1 · (negative number) = a positive number

36 contemporary calculus

Definition of | x |: (
x if x ≥ 0
|x| =
−x if x < 0

√
We can also write: | x | = x2 .
The domain of y = f ( x ) = | x | consists of all real numbers. The
range of f ( x ) = | x | consists of all numbers larger than or equal to zero
(all non-negative numbers). The graph of y = f ( x ) = | x | (see margin)
has no holes or breaks, but it does have a sharp corner at x = 0.
The absolute value will be useful for describing phenomena such
as reflected light and bouncing balls that change direction abruptly or
whose graphs have corners. The absolute value function has a number
of properties we will use later.

Properties of | x |: For all real numbers a and b,

• | a| = 0 · | a| = 0 if and only if a = 0

• | ab| = | a| · |b|
This last property is widely known as the • | a + b| ≤ | a| + |b|
triangle inequality.

Taking the absolute value of a function has an interesting effect on

the graph of the function: for any function f ( x ), we have
(
f ( x ) if f ( x ) ≥ 0
| f ( x )| =
− f ( x ) if f ( x ) < 0
In other words, if f ( x ) ≥ 0, then | f ( x )| = f ( x ), so the graph of | f ( x )|
is the same as the graph of f ( x ). If f ( x ) < 0, then | f ( x )| = − f ( x ), so
the graph of | f ( x )| is just the graph of f ( x ) “flipped” about the x-axis,
and it lies above the x-axis. The graph of | f ( x )| will always be on or
above the x-axis.
Example 5. The figure in the margin shows the graph of f ( x ). Graph:
(a) | f ( x )|
(b) |1 + f ( x )|
(c) 1 + | f ( x )|
Solution. The graphs appear below:
 welcome to calculus 37

In (b), we rigidly shift the graph of f up 1 unit before taking the

absolute value. In (c), we take the absolute value before rigidly shifting
the graph up 1 unit. ◀

Practice 7. The figure in the margin shows the graph of g( x ). Graph:

(a) | g( x )|
(b) | g( x − 1)|
(c) g (| x |)
The greatest integer function evaluated at a number x, y = f ( x ) =
⌊ x ⌋, is the largest integer less than or equal to x.
The value of ⌊ x ⌋ is always an integer and ⌊ x ⌋ is always less than
or equal to x. For example, ⌊3.2⌋ = 3, ⌊3.9⌋ = 3 and ⌊3⌋ = 3. If x is
positive, then ⌊ x ⌋ truncates x (drops the fractional part of x). If x is
negative, the situation is different: ⌊−4.2⌋ ̸= −4 because −4 is not less
than or equal to −4.2: ⌊−4.2⌋ = −5, ⌊−4.7⌋ = −5 and ⌊−4⌋ = −4.
Historically, many textbooks have used the square brackets [ ] to
represent the greatest integer function, while calculators and many
programming languages use INT( x ).

Definition of ⌊ x ⌋:
(
x if x is an integer
⌊x⌋ =
largest integer strictly less than x if x is not an integer

The domain of f ( x ) = ⌊ x ⌋ is all real numbers. The range of f ( x ) =

⌊ x ⌋ is only the integers. The graph of y = f ( x ) = ⌊ x ⌋ appears in the
margin. It has a jump break — a “step” — at each integer value of x,
so f ( x ) = ⌊ x ⌋ is called a step function. Between any two consecutive
integers, the graph is horizontal with no breaks or holes.
The greatest integer function is useful for describing phenomena that
change values abruptly, such as postage rates as a function of weight.
As of January 26, 2014, the cost to mail a first-class retail “flat” (such as
a manila envelope) was $0.98 for the first ounce and another $0.21 for
each additional ounce.
The ⌊ x ⌋ function can also be used for functions whose graphs are
“square waves,” such as the on and off of a flashing light.
Example 6. Graph y = ⌊1 + 0.5 sin( x )⌋.

Solution. One way to create this graph is to first graph y = 1 +

0.5 sin( x ), the thin curve in margin figure, and then apply the greatest
integer function to y to get the thicker “square wave” pattern. ◀

Practice 8. Sketch the graph of y = x2 for −2 ≤ x ≤ 2.

 
38 contemporary calculus

A Really “Holey” Function

The graph of ⌊ x ⌋ has a break or jump at each integer value, but how
many breaks can a function have? The next function illustrates just how
broken or “holey” the graph of a function can be.
Define a function h( x ) as:
(
2 if x is a rational number
h( x ) =
1 if x is an irrational number

Then h(3) = 2, h 53 = 2 and h − 25 = 2, because 3, 53 and − 25 are all

√ √
rational numbers. Meanwhile, h(π ) = 1, h( 7) = 1 and h( 2) = 1,
√ √
because π, 7 and 2 are all irrational numbers. These and some
other points are plotted in the margin figure.
In order to analyze the behavior of h( x ) the following fact about
rational and irrational numbers is useful.
Fact: Every interval contains both rational and irrational numbers.

Equivalently: If a and b are real numbers and a < b, then there is

• a rational number R between a and b (a < R < b)

• an irrational number I between a and b (a < I < b).

The above fact tells us that between any two places where y =
h( x ) = 1 (because x is rational) there is a place where y = h( x ) is 2,
because there is an irrational number between any two distinct rational
numbers. Similarly, between any two places where y = h( x ) = 2
(because x is irrational) there is a place where y = h( x ) = 1, because
there is a rational number between any two distinct irrational numbers.
The graph of y = h( x ) is impossible to actually draw, because every
two points on the graph are separated by a hole. This is also an example
of a function that your computer or calculator cannot graph, because in
general it can not determine whether an input value of x is irrational.
Example 7. Sketch the graph of
(
2 if x is a rational number
g( x ) =
x if x is an irrational number

Solution. A sketch of the graph of y = g( x ) appears in the margin.

When x is rational, the graph of y = g( x ) looks like the “holey” hori-
zontal line y = 2. When x is irrational, the graph of y = g( x ) looks like
the “holey” line y = x. ◀

Practice 9. Sketch the graph of

(
sin( x ) if x is a rational number
h( x ) =
x if x is an irrational number
 welcome to calculus 39

0.4 Problems
1. If T is the Celsius temperature of the air and v is the speed of the
wind in kilometers per hour, then


 √
T if 0 ≤ v ≤ 6.5
10.45+5.29 v−0.279v
WCI = 33 − 22 (33 − T ) if 6.5 < v ≤ 72

1.6T − 19.8 if v > 72


(a) Determine the Wind Chill Index for

i. a temperature of 0◦ C and a wind speed of 49 km/hr
ii. a temperature of 11◦ C and a wind speed of 80 km/hr.
(b) Write a multiline function definition for the WCI if the temperature
is 11◦ C.

2. Use the graph of y = f ( x ) in the margin to evaluate f (0), f (1), f (2),

f (3), f (4) and f (5). Write a multiline function definition for f .

3. Use the graph of y = g( x ) in the margin to evaluate g(0), g(1), g(2),

g(3), g(4) and g(5). Write a multiline function definition for g.

4. Use the values given in the table below, along with h( x ) = 2x + 1, to

determine the missing values of f ◦ g, g ◦ f and h ◦ g.

x f (x) g( x ) f ◦ g( x ) g ◦ f (x) h ◦ g( x )
−1 2 0
0 1 2
1 −1 1
2 0 2

5. Use the graphs shown below and the function 6. Use the graphs shown below and the function
h( x ) = x − 2 to determine the values of h( x ) = 5 − 2x to determine the values of
(a) f ( f (1)), f ( g(2)), f ( g(0)), f ( g(1)) (a) h ( f (0)), f (h(1)), f ( g(2)), f ( f (3))
(b) g ( f (2)), g ( f (3)), g ( g(0)), g ( f (0)) (b) g ( f (0)), g ( f (1)), g (h(2)), h ( f (3))
(c) f (h(3)), f (h(4)), h ( g(0)), h ( g(1)) (c) f ( g(0)), f ( g(1)), f (h(2)), h ( g(3))
 40 contemporary calculus

7. Defining h( x ) = x − 2, f ( x ) as: 11. Write a multiline function definition for the curve
 y = f ( x ) shown below.

 3 if x < 1
f (x) = x − 2 if 1 ≤ x < 3

 1 if 3 ≤ x

and g( x ) as:
(
x2 − 3 if x < 0
g( x ) =
⌊x⌋ if 0 ≤ x

(a) evaluate f ( x ), g( x ) and h( x ) for x = −1, 0, 1,

2, 3 and 4.
(b) evaluate f ( g(1)), f (h(1)), h ( f (1)), f ( f (2)),
g ( g(3.5)).
(c) graph f ( x ), g( x ) and h( x ) for −5 ≤ x ≤ 5.
8. Defining h( x ) = 3, f ( x ) as: 12. Define B( x ) to be the area of the rectangle whose
 lower left corner is at the origin and whose up-
 x + 1 if x < 1

per right corner is at the point ( x, f ( x )) for the
f (x) = 1 if 1 ≤ x < 3

 2 − x if 3 ≤ x function f shown below. For example, B(3) = 6.
Evaluate B(1), B(2), B(4) and B(5).
and g( x ) as:
(
| x + 1| if x < 0
g( x ) =
2x if 0 ≤ x

(a) evaluate f ( x ), g( x ) and h( x ) for x = −1, 0, 1,

2, 3 and 4.
(b) evaluate f ( g(1)), f (h(1)), h ( f (1)), f ( f (2)),
g ( g(3.5)).
(c) graph f ( x ), g( x ) and h( x ) for −5 ≤ x ≤ 5.
9. You are planning to take a one-week vacation in
Europe, and the tour brochure says that Monday
and Tuesday will be spent in England, Wednes-
13. Define B( x ) to be the area of the rectangle whose
day in France, Thursday and Friday in Germany,
lower left corner is at the origin
 and whose upper
and Saturday and Sunday in Italy. Let L(d) be 1
right corner is at the point x, x .
the location of the tour group on day d and write
a multiline function definition for L(d). (a) Evaluate B(1), B(2) and B(3).
10. A state has just adopted the following state in- (b) Show that B( x ) = 1 for all x > 0.
come tax system: no tax on the first $10,000 √
14. For f ( x ) = |9 − x | and g( x ) = x − 1:
earned, 1% of the next $10,000 earned, 2% of
(a) evaluate f ◦ g(1), f ◦ g(3), f ◦ g(5), f ◦ g(7),
the next $20,000 earned, and 3% of all additional
f ◦ g (0).
earnings. Write a multiline function definition for
T ( x ), the state income tax due on earnings of x (b) evaluate f ◦ f (2), f ◦ f (5), f ◦ f (−2).
dollars. (c) Does f ◦ f ( x ) = | x | for all values of x?
 welcome to calculus 41

15. The function g( x ) is graphed below. Graph off x to the nearest tenth (the first decimal place).
(a) g( x ) − 1 What function will round off x to:
(b) g( x − 1) (a) the nearest hundredth (two decimal places)?
(c) | g( x )| (b) the nearest thousandth (three decimal places)?
(d) ⌊ g( x )⌋ 21. Modify the function in Example 6 to produce a
“square wave” graph with a “long on, short off,
long on, short off” pattern.
22. Many computer languages contain a “signum” or
“sign” function defined by


 1 if x > 0
sgn( x ) = 0 if x = 0

 −1 if x < 0

16. The function f ( x ) is graphed below. Graph

(a) Graph sgn( x ).
(a) f ( x ) − 2
(b) Graph sgn( x − 2).
(b) f ( x − 2)
(c) Graph sgn( x − 4).
(c) | f ( x )|
(d) Graph sgn( x − 2) · sgn( x − 4).
(d) ⌊ f ( x )⌋
(e) Graph 1 − sgn( x − 2) · sgn( x − 4).
(f) For real numbers a and b with a < b, describe
the graph of 1 − sgn( x − a) · sgn( x − b).
23. Define g( x ) to be the slope of the line tangent to
the graph of y = f ( x ) (shown below) at ( x, y).
(a) Estimate g(1), g(2), g(3) and g(4).
17. Find A and B so that f ( g( x )) = g ( f ( x )) when:
(b) Graph y = g( x ) for 0 ≤ x ≤ 4.
(a) f ( x ) = 3x + 2 and g( x ) = 2x + A
(b) f ( x ) = 3x + 2 and g( x ) = Bx − 1
18. Find C and D so that f ( g( x )) = g ( f ( x )) when:
(a) f ( x ) = Cx + 3 and g( x ) = Cx − 1
(b) f ( x ) = 2x + D and g( x ) = 3x + D
19. Graph y = f ( x ) = x − ⌊ x ⌋ for −1 ≤ x ≤ 3. This
function is called the “fractional part of x” and
its graph an example of a “sawtooth” graph.
20. The function f ( x ) = ⌊ x + 0.5⌋ rounds off x to the
⌊10x + 0.5⌋
nearest integer, while g( x ) = rounds
10
 42 contemporary calculus

24. Define h( x ) to be the slope of the line tangent to 26. Let f ( x ) = 1 + sin( x ).
the graph of y = f ( x ) (see figure below) at ( x, y). (a) What happens if you start with x = 1 and
(a) Estimate h(1), h(2), h(3) and h(4). repeatedly feed the output from f back into f ?
(b) Graph y = h( x ) for 0 ≤ x ≤ 4. (b) What happens if you start with x = 2 and
examine the iterates of f ?
(Be sure your calculator is in radian mode.)
2
27. Starting with x = 1, do the iterates of f ( x ) = x 2x+1
approach a number? What happens if you start
with x = 0.5 or x = 4?
x
28. Let f ( x ) = 2 + 3.
(a) What are the iterates of f if you start with
x = 2? x = 4? x = 6?
25. Using the cos (cosine) button on your calculator (b) Find a number c so that f (c) = c. This value
several times produces iterates of f ( x ) = cos( x ). of c is called a fixed point of f .
x
What number will the iterates approach if you (c) Find a fixed point of g( x ) = 2 + A.
use the cos button 20 or 30 times starting with 29. Let f ( x ) = x
+ 4.
3
(a) x = 1? (a) What are the iterates of f if you start with
(b) x = 2? x = 2? x = 4? x = 6?
(c) x = 10? (b) Find a number c so that f (c) = c.
x
(Be sure your calculator is in radian mode.) (c) Find a fixed point of g( x ) = 3 + A.

30. Some iterative procedures are geometric rather than numerical. Start with an equilateral triangle with sides
of length 1, as shown at left in the figure below.
• Remove the middle third of each line segment.
• Replace the removed portion with two segments with the same length as the removed segment.
The first two iterations of this procedure are shown at center and right in the figure below. Repeat these
steps several more times, each time removing the middle third of each line segment and replacing it with
two new segments. What happens to the length of the shape with each iteration? (The result of iterating
over and over with this procedure is called Koch’s Snowflake, named for Helga von Koch.)
 welcome to calculus 43

31. Sketch the graph of 32. Sketch the graph of

( (
3 − x if x is a rational number x2 if x is a rational number
p( x ) = q( x ) =
1 if x is an irrational number x + 11 if x is an irrational number

0.4 Practice Answers

1. C ( x ) is the cost for one night on date x:
(
$380 if x is between December 15 and April 30
C(x) =
$295 if x is any other date

2. For the graph, see the figure in the margin.

x g( x ) x g( x )
−3 −3 π
3 − π3
−1 2 2 −2
0 2 3 −3
1
2 2 4 undefined
1 undefined 5 1

3. Define f ( x ) as:


 1 if x ≤ −1
f (x) = 1−x if − 1 < x ≤ 1

 2 if 1 < x

4. f ◦ g(3) = f (2) = −21 = −2; f ◦ g(8) = f (3) is undefined; g ◦ f (4) =

g(4) = 5; f ◦ h(1) = f (2) = −21 = −2; f ◦ h(3) = f (2) = −2;
f ◦ h(2) = f (3) is undefined; h ◦ g(−1) =qh(0) = 0; f ◦ g( x ) =
√
f ( 1 + x ) = √ 1+ x , g ◦ f ( x ) = g x −
x x


1+ x −3 3 = 1 + x− 3

5. See the figure below:

44 contemporary calculus

9
+x
x
9 +1 9 +5
6. Using f ( x ) = , f (1) = 1 2 = 5, f (5) = 5 2 = 3.4, f (3.4) ≈
2
3.023529412 and f (3.023529412) ≈ 3.000091554. The next iteration
gives f (3.000091554) ≈ 3.000000001: these values are approaching 3,
the square root of 9.
6
x +x 6 +1
With A = 6, f ( x ) = , so f (1) = 1 2 = 3.5, f (3.5) =
2
6
3.5 +3.5
2 = 2.607142857, and the next iteration gives f (2.607142857) ≈
2.45425636. Then f (2.45425636) ≈ 2.449494372, f (2.449494372) ≈
2.449489743 and f (2.449489743) ≈ 2.449489743 (the output is the
same as the input to 9 decimal places): these values are approaching
√
2.449489743, an approximation of 6.
A
x +x
For any positive value A, the iterates of f ( x ) = (starting with
√ 2
any positive x) will approach A.

7. The figure below shows some intermediate steps and final graphs:

8. The figure in the margin shows the graph of y = x2 and the (thicker)
graph of y = x2 .
 

9. The figure below shows the “holey” graph of y = x with a hole at

each rational value of x and the “‘holey” graph of y = sin( x ) with a
hole at each irrational value of x. Together they form the graph of
r ( x ).

(This is a very crude image, since we can’t really see the individual
holes, which have zero width.)
 welcome to calculus 45

0.5 Mathematical Language

The calculus concepts we will explore in this book are simple and
powerful, but sometimes subtle. To succeed in calculus you will need
to master some techniques, but (more importantly) you will need to
understand ideas and be able to work with the ideas in words and
pictures — very clear words and pictures.
You also need to understand some of the common linguistic con-
structions used in mathematics. In this section, we will discuss a few of
the most common mathematical phrases, the meanings of these phrases,
and some of their equivalent forms.
Your calculus instructor is going to use these types of statements, and
it is very important that you understand exactly what your instructor
means. You have reached a level in mathematics where the precise use
of language is important.

Equivalent Statements
Two statements are equivalent if they always have the same logical
value (a logical value is either “true” or “false”): that is, they are both
true or are both false.
The statements “x = 3” and “x + 2 = 5” are equivalent statements,
because if one of them is true then so is the other — and if one of them
is false then so is the other.
The statements “x = 3” and “x2 − 4x + 3 = 0” are not equivalent,
because x = 1 makes the second statement true but the first one false.

AND and OR
In everyday language, we use the words “and” and “or” all the time,
but in mathematics we must use them very carefully.

The compound statement “A and B are true” is equivalent to

“both of A and B are true.”

If A or if B or if both are false, then the statement “A and B are true”

is false. The statement “x2 = 4 and x > 0” is true when x = 2 and is
false for every other value of x.

The compound statement “A or B is true” is equivalent to

“at least one of A or B is true.”

If both A and B are false, then the statement “A or B is true” is false.

46 contemporary calculus

The statement “x2 = 4 or x > 0” is true if x = −2 or x is any positive

number. The statement is false when x = −3 (and for lots of other
values of x).

Practice 1. Which values of x make each statement true?

(a) x < 5
(b) x + 2 = 6
(c) x2 − 10x + 24 = 0
(d) “(a) and (b)”
(e) “(a) or (c)”

Negation of a Statement
For some simple statements, we can construct the negation just by
adding the word “not.”

statement negation of the statement

x is equal to 3 (x = 3) x is not equal to 3 (x ̸= 3)
x is less than 5 (x < 5) x is not less than 5 (x ̸< 5)
x is greater than or equal to 5 (x ≥ 5)

When the statement contains words such as “all,” “no” or “some,” then
the negation becomes more complicated.

statement negation of the statement

All x satisfy A. At least one x does not satisfy A.
Every x satisfies A. There is an x that does not satisfy A.
Some x does not satisfy A.
No x satisfies A. At least one x satisfies A.
Every x does not satisfy A. Some x satisfies A.
There is an x that satisfies A. No x satisfies A.
At least one x satisfies A. Every x does not satisfy A.
Some x satisfies A.

We can also negate compound statements containing “and” and “or.”

statement negation of the statement

A and B are both true. At least one of A or B is not true.
A and B and C are all true. At least one of A or B or C is not true.
A or B is true. Both A and B are not true.
 welcome to calculus 47

Practice 2. Write the negation of each statement.

(a) x + 5 = 3
(b) All prime numbers are odd.
(c) x2 < 4
(d) x divides 2 and x divides 3.
(e) No mathematician can sing well.

If. . . Then. . . : A Very Common Structure in Mathematics

The most common and basic structure used in mathematical language is
the “If {some hypothesis} then {some conclusion}” sentence. Almost ev-
ery result in mathematics can be stated using one or more “If. . . then. . . ”
sentences.

“If A then B’’ means that when the hypothesis A is true,

the conclusion B must also be true.

When the hypothesis is false, the “If. . . then. . . ” sentence makes no

claim about the truth or falsity of the conclusion — the conclusion may
be either true or false.
Even in everyday life you have probably encountered “If. . . then. . . ”
statements for a long time. A parent might try to encourage a child
with a statement like “If you clean your room then I will buy you an
ice cream cone.”
To show that an “If. . . then. . . ” statement is not valid (not true), all
we need to do is find a single example where the hypothesis is true
and the conclusion is false. Such an example with a true hypothesis
and false conclusion is called a counterexample for the “If. . . then. . . ”
statement. A valid “If. . . then. . . ” statement has no counterexample.

A counterexample to the statement “If A then B”

is an example in which A is true and B is false.

The only way for the statement “If you clean your room then I will
buy you an ice cream cone” to be false is if the child cleaned the room
and the parent did not buy the ice cream cone. If the child did not
clean the room but the parent bought the ice cream cone anyway, we
would say that the statement was true.
The statement “If n is a positive integer, then n2 + 5n + 5 is a prime
number” has hypothesis “n is a positive integer” and conclusion “n2 +
5n + 5 is a prime number.” This “If. . . then. . . ” statement is false,
because replacing n with the number 5 will make the hypothesis true
48 contemporary calculus

and the conclusion false. The number 5 is a counterexample for the

statement.
Every invalid “If. . . then. . . ” statement has at least one counterexam-
ple, and the most convincing way to show that a statement is not valid
is to find a counterexample to the statement.
Several other language structures are equivalent to the “If. . . then. . . ”
form. The statements below all mean the same as “If {A} then {B}”:

• “All A are B.”

• “Every A is B.”
• “Each A is B.”
• “Whenever A, then B.”
• “B whenever A.”
• “A only if B.”
• “A implies B.”
• “A ⇒ B” (the symbol “⇒” means “implies”)

Practice 3. Restate “If {a shape is a square} then {the shape is a rectan-

gle}” as many ways as you can.

“If. . . then. . . ” statements occur hundreds of times in every math-

ematics book, including this one. It is important that you are able to
recognize the various forms of “If. . . then. . . ” statements and that you
are able to distinguish the hypotheses from the conclusions.

Contrapositive Form of an “If. . . Then. . . ” Statement

The statement “If A then B” means that if the hypothesis A is true, then
the conclusion B is guaranteed to be true.
Suppose we know that in a certain town the statement “If {a building
is a church} then {the building is green}” is a true statement. What can
we validly conclude about a red building? Based on the information we
have, we can validly conclude that the red building is “not a church.”
because every church is green. We can also conclude that a blue
building is not a church. In fact, we can conclude that every “not
green” building is “not a church.” That is, if the conclusion of a valid
“If. . . then. . . ” statement is false, then the hypothesis must also be false.

The contrapositive form of “If A then B” is

“If {negation of B} then {negation of A}”
or “If {B is false} then {A is false}.”
 welcome to calculus 49

The statement “If A then B” and its contrapositive

“If {not B} then {not A}” are equivalent.

What about a green building in the aforementioned town? The green

building may or may not be a church — perhaps every post office is
also painted green. Or perhaps every building in town is green, in
which case the statement “If {a building is a church} then {the building
is green}" is certainly true.

Practice 4. Write the contrapositive form of each statement.

(a) If a function is differentiable then it is continuous.

(b) All men are mortal.

(c) If x = 3 then x2 − 5x + 6 = 0

(d) If {2 divides x and 3 divides x} then {6 divides x}.

Converse of an “If. . . then. . . ” Statement

If we switch the hypothesis and the conclusion of an “If A then B”
statement, we get the converse “If B then A.” For example, the converse
of “If {a building is a church} then {the building is green}” is “If {a
building is green} then {the building is a church}.”
The converse of an “If. . . then. . . ” statement is not equivalent to the
original “If. . . then. . . ” statement. For example, the statement “If x = 2
then x2 = 4” is true, but the converse statement “If x2 = 4 then x = 2”
is not true because x = −2 makes the hypothesis of the converse true
and the conclusion false

The converse of “If A then B” is “If B then A.”

The statement “If A then B” and its converse “If B then A.”
are not equivalent.

Wrap-up
The precise use of language by mathematicians (and mathematics
books) is an attempt to clearly communicate ideas from one person to
another, but this requires that both people understand the use and rules
of the language. If you don’t understand this usage, the communication
of the ideas will almost certainly fail.
 50 contemporary calculus

0.5 Problems

In Problems 1–2, define the sets A, B and C 11. (a) If you love your country, you will vote for me.
as A = {1, 2, 3, 4, 5}, B = {0, 2, 4, 6} and C = (b) If guns are outlawed then only outlaws will
{−2, −1, 0, 1, 2, 3}. List all values of x that satisfy have guns.
each statement.
In 12–15, write the negation of each statement.
1. (a) x is in A and x is in B
12. (a) It is raining.
(b) x is in A or x is in C
(b) Some equations have solutions.
(c) x is not in B and x is in C
(c) f ( x ) and g( x ) are polynomials.
2. (a) x is not in B or C
13. (a) f ( x ) or g( x ) is positive.
(b) x is in B and C but not in A
(b) x is positive.
(c) x is not in A but is in B or C
(c) 8 is a prime number.
In Problems 3–5, list or describe all the values of x 14. (a) Some months have six Mondays.
that make each statement true. (b) All quadratic equations have solutions.
3. (a) x2 +3 > 1 (c) The absolute value of a number is positive.
(b) x3 + 3 > 1 15. (a) For all numbers a and b, | a + b| = | a| + |b|.
(c) ⌊ x ⌋ ≤ | x | (b) All snakes are poisonous.
x2 +3x
4. (a) x = x+3 (c) No dog can climb trees.
(b) x > 4 and x < 9 16. Write an “If. . . then. . . ” statement that is true but
(c) | x | = 3 and x < 0 whose converse is false.
5. (a) x + 5 = 3 or x2 = 9 17. Write an “If. . . then. . . ” statement that is true and
(b) x + 5 = 3 and x2 =9 whose converse is true.
(c) | x + 3| = | x | + 3 18. Write an “If. . . then. . . ” statement that is false and
In Problems 6–8, write the contrapositive of each whose converse is false.
statement. If false, give a counterexample. In 19–22, determine whether each statement is true
6. (a) If x > 3 then x2 > 9. or false. If false, give a counterexample.
(b) Every solution of x2 − 6x + 8 = 0 is even. 19. (a) If a and b are real numbers then:
7. (a) If x2 + x − 6 = 0 then x = 2 or x = −3. ( a + b )2 = a2 + b2
(b) All triangles have 3 sides.
(b) If a > b then a2 > b2 .
8. (a) Every polynomial has at least one zero.
(c) If a > b then a3 > b3 .
(b) If I exercise and eat right then I will be healthy.
20. (a) For all real numbers a and b, | a + b| = | a| + |b|
In Problems 9–11, write the contrapositive of each (b) For all real numbers a and b, ⌊ a⌋ + ⌊b⌋ = ⌊ a + b⌋.
statement. (c) If f ( x ) and g( x ) are linear functions, then
9. (a) If your car is properly tuned, it will get at least f ( g( x )) is a linear function.
24 miles per gallon. 21. (a) If f ( x ) and g( x ) are linear functions then
(b) You can have pie if you eat your vegetables. f ( x ) + g( x ) is a linear function.
10. (a) A well-prepared student will miss less than (b) If f ( x ) and g( x ) are linear functions then
15 points on the final exam. f ( x ) · g( x ) is a linear function.
(b) I feel good when I jog. (c) If x divides 6 then x divides 30.
 welcome to calculus 51

22. (a) If x divides 50 then x divides 10. 24. (a) Every square has 4 sides.
(b) If x divides yz then x divides y or z. (b) All 4-sided polygons are squares.
(c) If x divides a2 then x divides a. (c) Every triangle has 2 equal sides.
(d) Every 4-sided polygon with equal sides is a square.
In 23–26, rewrite each statement in the form of an
25. (a) Every solution of x + 5 = 9 is odd.
“If. . . then. . . ” statement and determine whether it is
(b) Every 3-sided polygon with equal sides is a
true or false. If the statement is false, give a coun-
triangle.
terexample.
(c) Every calculus student studies hard.
23. (a) The sum of two prime numbers is a prime. (d) All (real) solutions of x2 − 5x + 6 = 0 are even.
(b) The sum of two prime numbers is never a 26. (a) Every line in the plane intersects the x-axis.
prime number. (b) Every (real) solution of x2 + 3 = 0 is even.
(c) Every prime number is odd. (c) All birds can fly.
(d) Every prime number is even. (d) No mammal can fly.

0.5 Practice Answers

1. (a) All values of x less than 5. (b) x = 4 (c) Both x = 4 and x = 6.
(d) x = 4 (e) x = 6 and all x less than 5.

2. (a) x + 5 < 3 (b) At least one prime number is even. (There is an

even prime number.) (c) x2 = 4 (d) x does not divide 2 or x does not
divide 3. (e) At least one mathematician can sing well. (There is a
mathematician who can sing well.)

3. Here are several ways:

• All squares are rectangles.

• Every square is a rectangle.
• Each square is a rectangle.
• Whenever a shape is a square, then it is a rectangle.
• A shape is a rectangle whenever it is a square.
• A shape is a square only if it is a rectangle.
• A shape is a square implies that it is a rectangle.
• Being a square implies being a rectangle.

4. (a) If a function is not continuous then it is not differentiable.

(b) All immortals are not men.
(c) x2 − 5x + 6 ̸= 0 ⇒ x ̸= 3
(d) If {6 does not divide x} then {2 does not divide x or 3 does not divide x}.
1
Limits and Continuity

1.0 Tangent Lines, Velocities, Growth

In Section 0.2, we estimated the slope of a line tangent to the graph of

a function at a point. At the end of Section 0.3, we constructed a new
function that gave the slope of the line tangent to the graph of a given
function at each point. In both cases, before we could calculate a slope,
we had to estimate the tangent line from the graph of the given function,
a method that required an accurate graph and good estimating. In this
section we will begin to look at a more precise method of finding the
slope of a tangent line that does not require a graph or any estimation
by us. We will start with a non-applied problem and then look at two
applications of the same idea.

The Slope of a Line Tangent to a Function at a Point

Our goal is to find a way of exactly determining the slope of the line
that is tangent to a function (to the graph of the function) at a point in a
way that does not require us to actually have the graph of the function.
Let’s start with the problem of finding the slope of the line L (see
margin figure), which is tangent to f ( x ) = x2 at the point (2, 4). We
could estimate the slope of L from the graph, but we won’t. Instead,
we can see that the line through (2, 4) and (3, 9) on the graph of f is
an approximation of the slope of the tangent line, and we can calculate
that slope exactly:
∆y 9−4
m= = =5
∆x 3−2
But m = 5 is only an estimate of the slope of the tangent line — and
not a very good estimate. It’s too big. We can get a better estimate by
picking a second point on the graph of f closer to (2, 4) — the point
(2, 4) is fixed and it must be one of the two points we use. From the
figure in the margin, we can see that the slope of the line through the
points (2, 4) and (2.5, 6.25) is a better approximation of the slope of the
54 contemporary calculus

tangent line at (2, 4):

∆y 6.25 − 4 2.25
m= = = = 4.5
∆x 2.5 − 2 0.5
This is a better estimate, but still an approximation.
We can continue picking points closer and closer to (2, 4) on the
graph of f , and then calculating the slopes of the lines through each of
these points ( x, y) and the point (2, 4):

points to the left of (2, 4) points to the right of (2, 4)

x y = x2 slope x y = x2 slope
1.5 2.25 3.5 3 9 5
1.9 3.61 3.9 2.5 6.25 4.5
1.99 3.9601 3.99 2.01 4.0401 4.01

The only thing special about the x-values we picked is that they are
numbers close — and very close — to x = 2. Someone else might have
picked other nearby values for x. As the points we pick get closer and
closer to the point (2, 4) on the graph of y = x2 , the slopes of the lines
through the points and (2, 4) are better approximations of the slope of
the tangent line, and these slopes are getting closer and closer to 4.
Practice 1. What is the slope of the line through (2, 4) and ( x, y) for
y = x2 and x = 1.994? For x = 2.0003?
We can bypass much of the calculating by not picking the points one
at a time: let’s look at a general point near (2, 4). Define x = 2 + h so
h is the increment from 2 to x (see margin figure). If h is small, then
x = 2 + h is close to 2 and the point (2 + h, f (2 + h)) = 2 + h, (2 + h)2

is close to (2, 4). The slope m of the line through the points (2, 4) and
2 + h, (2 + h)2 is a good approximation of the slope of the tangent

line at the point (2, 4):

∆y (2 + h )2 − 4 (4 + 4h + h2 ) − 4
m= = =
∆x (2 + h ) − 2 h
4h + h2 h (4 + h )
= = = 4+h
h h
If h is very small, then m = 4 + h is a very good approximation to the
slope of the tangent line, and m = 4 + h also happens to be very close
to the value 4. The value m = 4 + h is called the slope of the secant
line through the two points (2, 4) and 2 + h, (2 + h)2 . The limiting

value 4 of m = 4 + h as h gets smaller and smaller is called the slope of

the tangent line to the graph of f at (2, 4).
Example 1. Find the slope of the line tangent to f ( x ) = x2 at the
point (1, 1) by evaluating the slope of the secant line through (1, 1) and
(1 + h, f (1 + h)) and then determining what happens as h gets very
small (see margin figure).
 limits and continuity 55

Solution. The slope of the secant line through the points (1, 1) and
(1 + h, f (1 + h)) is:
f (1 + h ) − 1 (1 + h )2 − 1 (1 + 2h + h2 ) − 1
m= = =
(1 + h ) − 1 h h
2h + h2 h (2 + h )
= = = 2+h
h h
As h gets very small, the value of m approaches the value 2, the slope
of tangent line at the point (1, 1). ◀

Practice 2. Find the slope of the line tangent to the graph of y =

f ( x ) = x2 at the point (−1, 1) by finding the slope of the secant line,
msec , through the points (−1, 1) and (−1 + h, f (−1 + h)) and then
determining what happens to msec as h gets very small.

Falling Tomato
Suppose we drop a tomato from the top of a 100-foot building (see
margin figure) and record its position at various times during its fall:

time (sec) height (ft)

0.0 100
0.5 96
1.0 84
1.5 64
2.0 36
2.5 0
Some questions are easy to answer directly from the table:
(a) How long did it take for the tomato to drop 100 feet?
(2.5 seconds)
(b) How far did the tomato fall during the first second?
(100 − 84 = 16 feet)
(c) How far did the tomato fall during the last second?
(64 − 0 = 64 feet)
(d) How far did the tomato fall between t = 0.5 and t = 1?
(96 − 84 = 12 feet)
Other questions require a little calculation:
(e) What was the average velocity of the tomato during its fall?
distance fallen ∆position −100 ft ft
average velocity = = = = −40
total time ∆time 2.5 s sec
(f) What was the average velocity between t = 1 and t = 2 seconds?
∆position 36 ft − 84 ft −48 ft ft
average velocity = = = = −48
∆time 2 s−1 s 1s sec
56 contemporary calculus

Some questions are more difficult.

(g) How fast was the tomato falling 1 second after it was dropped?

This question is significantly different from the previous two questions

about average velocity. Here we want the instantaneous velocity, the
velocity at an instant in time. Unfortunately, the tomato is not equipped
with a speedometer, so we will have to give an approximate answer.
One crude approximation of the instantaneous velocity after 1 second
is simply the average velocity during the entire fall, −40 secft . But the
tomato fell slowly at the beginning and rapidly near the end, so this
estimate may or may not be a good answer.
We can get a better approximation of the instantaneous velocity at
t = 1 by calculating the average velocities over a short time interval
near t = 1. The average velocity between t = 0.5 and t = 1 is:
−12 feet ft
= −24
0.5 sec sec
and the average velocity between t = 1 and t = 1.5 is
−20 feet ft
= −40
0.5 sec sec
so we can be reasonably sure that the instantaneous velocity is between
ft and −40 ft .
−24 sec sec
In general, the shorter the time interval over which we calculate the
average velocity, the better the average velocity will approximate the
instantaneous velocity. The average velocity over a time interval is:
∆position
∆time
which is the slope of the secant line through two points on the graph of
height versus time (see margin figure).

∆position
average velocity =
∆time
= slope of the secant line through two points

The instantaneous velocity at a particular time and height is the

slope of the tangent line to the graph at the point given by that time
and height.

instantaneous velocity = slope of the line tangent to the graph

Practice 3. Estimate the instantaneous velocity of the tomato 2 seconds

after it was dropped.
 limits and continuity 57

Growing Bacteria
Suppose we set up a machine to count the number of bacteria growing
on a Petri plate (see margin figure). At first there are few bacteria, so
the population grows slowly. Then there are more bacteria to divide,
so the population grows more quickly. Later, there are more bacteria
and less room and nutrients available for the expanding population, so
the population grows slowly again. Finally, the bacteria have used up
most of the nutrients and the population declines as bacteria die.
The population graph can be used to answer a number of questions:

(a) What is the bacteria population at time t = 3 days?

(about 500 bacteria)
(b) What is the population increment from t = 3 to t = 10 days?
(about 4, 000 bacteria)
(c) What is the rate of population growth from t = 3 to t = 10 days?

To answer this last question, we compute the average change in popula-

tion during that time:

change in population
average change in population =
change in time
∆population 4000 bacteria bacteria
= = ≈ 570
∆time 7 days day

This is the slope of the secant line through (3, 500) and (10, 4500).

∆population
average population growth rate =
∆time
= slope of the secant line through two points

Now for a more difficult question:

(d) What is the rate of population growth on the third day, at t = 3?

This question asks for the instantaneous rate of population change,

the slope of the line tangent to the population curve at (3, 500). If we
sketch a line approximately tangent to the curve at (3, 500) and pick
two points near the ends of the tangent line segment (see margin figure),
we can estimate that the instantaneous rate of population growth is
approximately 320 bacteria .
day

instantaneous population growth rate =

slope of the line tangent to the graph
58 contemporary calculus

Practice 4. Find approximate values for:

(a) the average change in population between t = 9 and t = 13.

(b) the rate of population growth at t = 9 days.

The tangent line problem, the instantaneous velocity problem and

the instantaneous growth rate problem are all similar. In each problem
we wanted to know how rapidly something was changing at an instant
in time, and each problem turned out to involve finding the slope
of a tangent line. The approach in each problem was also the same:
find an approximate solution and then examine what happens to the
approximate solution over shorter and shorter intervals. We will often
use this approach of finding a limiting value, but before we can use
it effectively we need to describe the concept of a limit with more
precision.

1.0 Problems

1. (a) What is the slope of the line through (3, 9) 4. (a) What is the slope of the line through (−1, −2)
and ( x, y) for y = x2 when: and ( x, y) for y = x2 + x − 2 when:
i. x = 2.97? i. x = −0.98?
ii. x = 3.001? ii. x = −1.03?
iii. x = 3 + h? iii. x = −1 + h?
(b) What happens to this last slope when h is very
(b) What happens to this last slope when h is very
small (close to 0)?
small (close to 0)?
(c) Sketch the graph of y = x2 for x near 3.
(c) Sketch the graph of y = x2 + x − 2 for x
2. (a) What is the slope of the line through (−2, 4) near −1.
and ( x, y) for y = x2 when:
5. The figure below shows the temperature during
i. x = −1.98?
a day in Ames.
ii. x = −2.03?
iii. x = −2 + h? (a) What was the average change in temperature
(b) What happens to this last slope when h is very from 9 a.m. to 1 p.m.?
small (close to 0)? (b) Estimate how fast the temperature was rising
(c) Sketch the graph of y = x2 for x near −2. at 10 a.m. and at 7 p.m.

3. (a) What is the slope of the line through (2, 4)

and ( x, y) for y = x2 + x − 2 when:
i. x = 1.99?
ii. x = 2.004?
iii. x = 2 + h?
(b) What happens to this last slope when h is very
small (close to 0)?
(c) Sketch the graph of y = x2 + x − 2 for x near 2.
 limits and continuity 59

6. The figure below shows the distance of a car from 8. The figure below shows the composite develop-
a measuring position located on the edge of a mental skill level of chessmasters at different ages
straight road. as determined by their performance against other
(a) What was the average velocity of the car from chessmasters. (From “Rating Systems for Human
t = 0 to t = 30 seconds? Abilities,” by W.H. Batchelder and R.S. Simpson,
(b) What was the average velocity from t = 10 to 1988. UMAP Module 698.)
t = 30 seconds? (a) At what age is the “typical” chessmaster play-
(c) About how fast was the car traveling at t = 10 ing the best chess?
seconds? At t = 20? At t = 30? (b) At approximately what age is the chessmas-
(d) What does the horizontal part of the graph ter’s skill level increasing most rapidly?
between t = 15 and t = 20 seconds tell you? (c) Describe the development of the “typical”
(e) What does the negative velocity at t = 25 rep- chessmaster’s skill in words.
resent? (d) Sketch graphs that you think would reason-
ably describe the performance levels versus
age for an athlete, a classical pianist, a rock
singer, a mathematician and a professional in
your major field.

7. The figure below shows the distance of a car from

a measuring position located on the edge of a
straight road.
9. Define A( x ) to be the area bounded by the t- (hor-
(a) What was the average velocity of the car from izontal) and y-axes, the horizontal line y = 3, and
t = 0 to t = 20 seconds? the vertical line at x (see figure below). For exam-
(b) What was the average velocity from t = 10 to ple, A(4) = 12 is the area of the 4 × 3 rectangle.
t = 30 seconds?
(a) Evaluate A(0), A(1), A(2), A(2.5) and A(3).
(c) About how fast was the car traveling at t = 10
(b) What area would A(4) − A(1) represent?
seconds? At t = 20? At t = 30?
(c) Graph y = A( x ) for 0 ≤ x ≤ 4.
 60 contemporary calculus

10. Define A( x ) to be the area bounded by the t-

(horizontal) and y-axes, the line y = t + 1, and
the vertical line at x (see figure). For example,
A(4) = 12.

(a) Evaluate A(0), A(1), A(2), A(2.5) and A(3).

(b) What area would A(3) − A(1) represent in the

figure?

(c) Graph y = A( x ) for 0 ≤ x ≤ 4.

 limits and continuity 61

1.0 Practice Answers

1. If x = 1.994, then y = 3.976036, so the slope between (2, 4) and ( x, y)
is:
4−y 4 − 3.976036 0.023964
= = ≈ 3.994
2−x 2 − 1.994 0.006
If x = 2.0003, then y ≈ 4.0012, so the slope between (2, 4) and ( x, y)
is:
4−y 4 − 4.0012 −0.0012
= = ≈ 4.0003
2−x 2 − 2.0003 −0.0003
2. Computing msec :

f (−1 + h) − (1) (−1 + h)2 − 1 1 − 2h + h2 − 1 h(−2 + h)

= = = = −2 + h
(−1 + h) − (−1) h h h

As h → 0, msec = −2 + h → −2.

3. The average velocity between t = 1.5 and t = 2.0 is:

36 − 64 feet feet
= −56
2.0 − 1.5 sec sec
The average velocity between t = 2.0 and t = 2.5 is:

0 − 36 feet feet
= −72
2.5 − 2.0 sec sec

The velocity at t = 2.0 is somewhere between −56 feet feet

sec and −72 sec ,
probably around the middle of this interval:

(−56) + (−72) feet

= −64
2 sec

4. (a) When t = 9 days, the population is approximately P = 4, 200

bacteria. When t = 13, P ≈ 5, 000. The average change in
population is approximately:

5000 − 4200 bacteria 800 bacteria bacteria

= = 200
13 − 9 days 4 days day

(b) To find the rate of population growth at t = 9 days, sketch

the line tangent to the population curve at the point (9, 4200)
and then use (9, 4200) and another point on the tangent line to
calculate the slope of the line. Using the approximate values
(5, 2800) and (9, 4200), the slope of the tangent line at the point
(9, 4200) is approximately:

4200 − 2800 bacteria 1400 bacteria bacteria

= ≈ 350
9 − 5 days 4 days day
62 contemporary calculus

1.1 The Limit of a Function

Calculus has been called the study of continuous change, and the limit
is the basic concept that allows us to describe and analyze such change.
An understanding of limits is necessary to understand derivatives,
integrals and other fundamental topics of calculus.

The Idea (Informally)

The limit of a function at a point helps describe the behavior of the
function when the input variable is near — but does not equal — a
specified number (see margin figure). If the values of f ( x ) are all “very
close” — as close as we want — to one number L as we restrict values
of x to be “very close” to (but not equal to) a number c, then

we say: “the limit of f ( x ), as x approaches c, is L”

The symbol → means “approaches” or

and we write: lim f ( x ) = L
x →c
“gets very close to.”

It is very important to note that:

f (c) is a single number that describes the behavior (value)

of f at the point x = c

while:

lim f ( x ) is a single number that describes the behavior

x →c
of f near, but not at the point x = c

If we have a graph of the function f ( x ) near x = c, then it is often

easy to estimate lim f ( x ).
x →c

Example 1. Use the graph of y = f ( x ) given in the margin to estimate

the following limits:

(a) lim f ( x ) (b) lim f ( x ) (c) lim f ( x ) (d) lim f ( x )

x →1 x →2 x →3 x →4

Solution. Each of these limits involves a different issue, as you may be

able to tell from the graph.

(a) lim f ( x ) = 2: When x is very close to 1, the values of f ( x )

x →1
appear to be very close to y = 2. In this example, it happens that
f (1) = 2, but that is irrelevant for the limit. The only thing that
matters is what happens for x close to 1 but with x ̸= 1.
 limits and continuity 63

(b) f (2) is undefined, but we only care about the behavior of f ( x )

for x close to 2, where the values of f ( x ) appear to be close to 3.
If we restrict x close enough to 2, the values of f ( x ) will be as
close to 3 as we want, so lim f ( x ) = 3.
x →2

(c) When x is close to 3, the values of f ( x ) appear to be close to 1,

so lim f ( x ) = 1. For this limit it is completely irrelevant that
x →3
f (3) = 2: we only care about what happens to f ( x ) for x close
to (but not equal to) 3.

(d) This one is harder and we need to be careful. When x is close to

4 and slightly less than 4 (x is just to the left of 4 on the x-axis)
then the values of f ( x ) are close to 2. But if x is close to 4 and
slightly larger than 4 then the values of f ( x ) are close to 3.

If we know only that x is very close to 4, then we cannot say

whether y = f ( x ) will be close to 2 or close to 3 — it depends on
whether x is on the right or the left side of 4. In this situation,
the f ( x ) values are not all close to a single number when x is
close to 4, so we say that lim f ( x ) does not exist.
x →4

In (d), it is irrelevant that f (4) = 1. The limit, as x approaches 4, would

still cease to exist if f (4) was 3 or 2 or anything else. ◀

Practice 1. Use the graph of y = f ( x ) in the margin to estimate the

following limits:

(a) lim f ( x ) (b) lim f (t) (c) lim f ( x ) (d) lim f (w)
x →1 t →2 x →3 w →4

2x2 − x − 1
Example 2. Determine the value of lim .
x →3 x−1

2
Solution. We need to investigate the values of f ( x ) = 2x x−−x1−1 when
x is close to 3. If the f ( x ) values are all arbitrarily close to — or even
equal to — some number L, then L will be the limit.
One way to keep track of both the x and the f ( x ) values is to set up
a table and to pick several x values that get closer and closer (but not
equal) to 3.
We can pick some values of x that approach 3 from the left, say
x = 2.91, 2.9997, 2.999993 and 2.9999999, and some values of x that
approach 3 from the right, say x = 3.1, 3.004, 3.0001 and 3.000002. The
only thing important about these particular values for x is that they get
closer and closer to 3 without actually equaling 3. You should try some
other values “close to 3” to see what happens. Our table of values is:
64 contemporary calculus

x f (x) x f (x)
2.9 6.82 3.1 7.2
2.9997 6.9994 3.004 7.008
2.999993 6.999986 3.0001 7.0002
2.9999999 6.9999998 3.000002 7.000004
↓ ↓ ↓ ↓
3 7 3 7

As the x values get closer and closer to 3, the f ( x ) values are all close to
7. In fact, we can get f ( x ) as close to 7 as we want (“arbitrarily close”)
by taking the values of x very close (“sufficiently close”) to 3. We write:

2x2 − x − 1
lim =7
x →3 x−1

Instead of using a table of values, we could have graphed y = f ( x )

for x close to 3 (see margin) and used the graph to answer the limit
question. This graphical approach is easier, particularly if you have a
calculator or computer do the graphing work for you, but it is really
very similar to the “table of values” method: in each method you need
to evaluate y = f ( x ) at many values of x near 3. ◀

In the previous example, you might have noticed that if we just

evaluate f (3), then we get the correct answer, 7. That works for this
particular problem, but it often fails. The next example (identical to the
previous one, except x → 1) illustrates one such difficulty.

2x2 − x − 1
Example 3. Find lim .
x →1 x−1

2
Solution. You might try to evaluate f ( x ) = 2x x−−x1−1 at x = 1, but
f (1) = 00 , so f is not defined at x = 1.
It is tempting — but wrong — to conclude that this function does not
have a limit as x approaches 1.
Table Method: Trying some “test” values for x that get closer and
closer to 1 from both the left and the right, we get:

x f (x) x f (x)
0.9 2.82 1.1 3.2
0.9998 2.9996 1.003 3.006
0.999994 2.999988 1.0001 3.0002
0.9999999 2.9999998 1.000007 3.000014
↓ ↓ ↓ ↓
1 3 1 3
 limits and continuity 65

The function f is not defined at x = 1, but when x gets close to 1, the

values of f ( x ) are very close to 3. We can ensure that f ( x ) is as close to
3 as we want by restricting x to be very close to 1, so:
2x2 − x − 1
lim =3
x →1 x−1
2
Graph Method: We can graph y = f ( x ) = 2x x−−x1−1 for x close to
1 (see margin) and notice that whenever x is close to 1, the values of
y = f ( x ) are close to 3; f is not defined at x = 1, so the graph has a
hole above x = 1, but we only care about what f ( x ) is doing for x close
to but not equal to 1.
Algebra Method: We could have found the same result by noting:
2x2 − x − 1 (2x + 1)( x − 1)
f (x) = = = 2x + 1
x−1 x−1
as long as x ̸= 1. The “x → 1” part of the limit means that x is close to 1
but not equal to 1, so our division step is valid and:
2x2 − x − 1
lim = lim [2x + 1] = 3
x →1 x−1 x →1

which is the same answer we obtained using the first two methods. ◀

Three Methods for Evaluating Limits

The previous example utilized three different methods, each of which
led us to the same answer for the limit.

The Algebra Method

The algebra method involves algebraically simplifying the function
before trying to evaluate its limit. Often, this simplification just means
factoring and dividing, but sometimes more complicated algebraic or
even trigonometric steps are needed.

The Table Method

To evaluate a limit of a function f ( x ) as x approaches c, the table
method involves calculating the values of f ( x ) for “enough” values of x
very close to c so that we can “confidently” determine a limiting value
of f ( x ). If f ( x ) is well behaved, we may not need to use very many
values for x. However, this method is usually used with complicated
functions, and then we need to evaluate f ( x ) for lots of values of x.
A computer or calculator can often make the function evaluations
easier, but their calculations are subject to “round off” errors. The result
of any computer calculation that involves both large and small numbers
should be viewed with some suspicion. For example, the function
((0.1) x + 1) − 1 (0.1) x
f (x) = = =1
(0.1) x (0.1) x
66 contemporary calculus

for every value of x, and my calculator gives the correct answer for
some values of x: f (3) = 1, and f (8) and f (9) both equal 1.
But my calculator says (0.1)10 + 1 − 1 = 0, so it evaluates f (10) to

be 0, definitely an incorrect value.

Your calculator may evaluate f (10) correctly, but try f (35) or f (107).

Calculators are too handy to be ignored, but they are too prone to
these types of errors to be believed uncritically. Be careful.

The Graph Method

The graph method is closely related to the table method, but we
create a graph of the function instead of a table of values, and then we
use the graph to determine the limiting value of f ( x ) (if there is one).

Which Method Should You Use?

In general, the algebraic method is preferred because it is precise
and does not depend on which values of x we chose or the accuracy
of our graph or precision of our calculator. If you can evaluate a limit
algebraically, you should do so. Sometimes, however, it will be very
difficult to evaluate a limit algebraically, and the table or graph methods
offer worthwhile alternatives. Even when you can algebraically evaluate
the limit of a function, it is still a good idea to graph the function or
evaluate it at a few points just to check that your algebraic answer is
reasonable.
The table and graph methods have the same advantages and disad-
vantages. Both can be used on complicated functions that are difficult
to handle algebraically or whose algebraic properties you don’t know.
Often both methods can be easily programmed on a calculator or
computer. However, these two methods are very time-consuming by
hand and are prone to round-off errors on computers. You need to
know how to use these methods when you can’t figure out how to use
the algebraic method, but you need to use these two methods warily.
Example 4. Evaluate each limit.

x2 + 5x + 6 x2 + 5x + 6
(a) lim (b) lim
x →0 x2 + 3x + 2 x →−2 x2 + 3x + 2

Solution. The function in each limit is the same but x is approaching

a different number in each of them.
(a) Because x → 0, we know that x is getting closer and closer to
0, so the values of the x2 , 5x and 3x terms get as close to 0 as
we want. The numerator approaches 6 and the denominator
approaches 2, so the values of the whole function get arbitrarily
close to 62 = 3, the limit.
 limits and continuity 67

(b) As x approaches −2, the numerator and denominator approach

0, and a small number divided by a small number can be almost
anything — the ratio depends on the size of the top compared to
the size of the bottom. More investigation is needed.
Table Method: If we pick some values of x close to (but not
equal to) −2, we get the table:

x2 +5x +6
x x2 + 5x + 6 x2 + 3x + 2 x2 +3x +2

−1.97 0.0309 −0.0291 −1.061856

−2.005 −0.004975 0.005025 −0.990050
−1.9998 0.00020004 −0.00019996 −1.00040008
−2.00003 −0.00002999 0.0000300009 −0.9996666
↓ ↓ ↓ ↓
−2 0 0 −1

Even though the numerator and denominator are each getting

closer and closer to 0, their ratio is getting arbitrarily close to −1,
which appears to be the limit.
2
+5x +6 in the margin
Graph Method: The graph of y = f ( x ) = xx2 + 3x +2
appears to show that the values of f ( x ) are very close to −1
when the x-values are close to −2.
Algebra Method: Factoring the numerator and denominator:

x2 + 5x + 6 ( x + 2)( x + 3)
f (x) = =
x2 + 3x + 2 ( x + 2)( x + 1)

We know x → −2 so x ̸= −2 and we can divide the top and

bottom by ( x + 2). Then

( x + 3) 1
f (x) = → = −1
( x + 1) −1

as x → −2. ◀

You should remember the technique used in the previous example:

polynomial 0
If lim = ,
x →c another polynomial 0
try dividing the top and bottom by x − c.
68 contemporary calculus

Practice 2. Evaluate each limit.

x2 − x − 2
(a) lim
x →2 x−2
t · sin(t)
(b) lim
t →0 t2 + 3t
w−2
(c) lim
w →2 ln( w2 )

One-Sided Limits
Sometimes, what happens to us at a place depends on the direction
we use to approach that place. If we approach Niagara Falls from
the upstream side, then we will be 182 feet higher and have different
worries than if we approach from the downstream side. Similarly, the
values of a function near a point may depend on the direction we use
to approach that point.
If we let x approach 3 from the left (x is close to 3 and x < 3) then
the values of ⌊ x ⌋ = INT( x ) equal 2 (see margin).
If we let x approach 3 from the right (x is close to 3 and x > 3) then
the values of ⌊ x ⌋ = INT( x ) equal 3.
On the number line we can approach a point from the left or the
right, and that leads to one-sided limits.

Definition of Left and Right Limits:

The left limit as x approaches c of f ( x ) is L if the values of f ( x ) are
as close to L as we want when x is very close to but left of c (x < c):

lim f ( x ) = L
x →c−

The right limit, lim f ( x ), requires that x lie to the right of c (x > c).
x →c+

Example 5. Evaluate lim x − ⌊ x ⌋ and lim x − ⌊ x ⌋.

x →2− x →2+

Solution. The left-limit notation x → 2− requires that x be close to 2

and that x be to the left of 2, so x < 2. If 1 < x < 2, then ⌊ x ⌋ = 1 and:

lim x − ⌊ x ⌋ = lim x − 1 = 2 − 1 = 1
x →2− x →2−

If x is close to 2 and is to the right of 2, then 2 < x < 3, so ⌊ x ⌋ = 2 and:

lim x − ⌊ x ⌋ = lim x − 2 = 2 − 2 = 0
x →2+ x →2+

A graph of f ( x ) = x − ⌊ x ⌋ appears in the margin. ◀

 limits and continuity 69

If the left and right limits of f ( x ) have the same value at x = c:

lim f ( x ) = lim f ( x ) = L
x →c− x →c+

then the value of f ( x ) is close to L whenever x is close to c, and it does

not matter whether x is left or right of c, so

lim f ( x ) = L
x →c

Similarly, if:
lim f ( x ) = L
x →c

then f ( x ) is close to L whenever x is close to c and less than c, and

whenever x is close to c and greater than c, so:

lim f ( x ) = lim f ( x ) = L
x →c− x →c+

We can combine these two statements into a single theorem.

One-Sided Limit Theorem:

lim f ( x ) = L if and only if lim f ( x ) = lim f ( x ) = L

x →c x →c− x →c+

This theorem has an important corollary.

Corollary:

If lim f ( x ) ̸= lim f ( x ), then lim f ( x ) does not exist.

x →c− x →c+ x →c

One-sided limits are particularly useful for describing the behavior

of functions that have steps or jumps.
To determine the limit of a function involving the greatest integer or
absolute value or a multiline definition, definitely consider both the left
and right limits.

Practice 3. Use the graph in the margin to evaluate the one- and two-
sided limits of f at x = 0, 1, 2 and 3.

Practice 4. Defining f ( x ) as:


 1 if x < 1

f (x) = x if 1 < x < 3

 2 if 3 < x

find the one- and two-sided limits of f at 1 and 3.

70 contemporary calculus

1.1 Problems

1. Use the graph below to estimate the limits. 4. Use the graph below to estimate the limits.

(a) lim f ( x ) (b) lim f ( x ) (a) lim f (3x ) (b) lim f ( x + 1)

x →1 x →2 x →1 x →2

(c) lim f ( x ) (d) lim f ( x ) (c) lim f (2x − 4) (d) lim | f (4 + x )|

x →3 x →4
x →3 x →0

2. Use the graph below to estimate the limits.

(a) lim f ( x ) (b) lim f ( x ) In Problems 5–11, evaluate (or estimate) each limit.
x →1 x →2

(c) lim f ( x ) (d) lim f ( x ) x2 + 3x + 3 x2 + 3x + 3

x →3 x →4 5. (a) lim (b) lim
x →1 x−2 x →2 x−2
x+7 x+7
6. (a) lim (b) lim
x →0 x2 + 9x + 14 x →3 x2 + 9x + 14
x+7 x+7
(c) lim (d) lim
x →−4 x2 + 9x + 14 x →−7 x2 + 9x + 14
cos( x ) cos( x )
7. (a) lim (b) lim
x →1 x x →π x
cos( x )
(c) lim
3. Use the graph below to estimate the limits. x →−1 x
√ √
8. (a) lim x − 3 (b) lim x−3
(a) lim f (2x ) (b) lim f ( x − 1) x →7 x →9
x →1 x →2
√
(c) lim f (2x − 5) (d) lim f (4 + x ) x−3
x →3 x →0
(c) lim
x →9 x−9

9. (a) lim | x | (b) lim | x |

x →0− x →0+

(c) lim | x |
x →0

|x| |x|
10. (a) lim (b) lim
x →0− x x →0+ x
|x|
(c) lim
x →0 x
 limits and continuity 71

11. (a) lim | x − 5| | x − 5|

x →5 (b) lim
x →3 x−5

| x − 5|
(c) lim
x →5 x−5

12. Find the one- and two-sided limits of:



 x if x < 0
f (x) = sin( x ) if 0 < x ≤ 2

 1 if 2 < x

as x → 0, 1 and 2.
13. Find the one- and two-sided limits of: 19. Define A( x ) to be the area bounded by the t- and
y-axes, the line y = 21 t + 2 and the vertical line t = x


 1 if x ≤ 2
g( x ) = 8 (See figure below). For example, A(4) = 12.
 x if 2 < x < 4
 6 − x if 4 < x
(a) Evaluate A(0), A(1), A(2) and A(3).

as x → 1, 2, 4 and 5. (b) Graph y = A( x ) for 0 ≤ x ≤ 4.

In 14–17, use a calculator or computer to get approxi- (c) What area does A(3) − A(1) represent?
mate answers accurate to 2 decimal places.

2x − 1 log10 ( x )
14. (a) lim (b) lim
x →0 x x →1 x−1

3x − 1 ln( x )
15. (a) lim (b) lim
x →0 x x →1 x−1

√
x−1−2 sin(3x )
16. (a) lim (b) lim
x →5 x−5 x →0 5x

√
x−4 sin(7x ) √
17. (a) lim (b) lim
x →16 x − 16 x →0 2x 20. Sketch the graph of f (t) = 4t − t2 for 0 ≤ t ≤ 4
(you should get a semicircle). Define A( x ) to be
18. Define A( x ) to be the area bounded by the t- and the area bounded below by the t-axis, above by the
y-axes, the “bent line” in the figure below, and the graph y = f (t) and on the right by the vertical line
vertical line t = x. For example, A(4) = 10. at t = x.

(a) Evaluate A(0), A(1), A(2) and A(3). (a) Evaluate A(0), A(2) and A(4).
(b) Graph y = A( x ) for 0 ≤ x ≤ 4. (b) Sketch a graph y = A( x ) for 0 ≤ x ≤ 4.
(c) What area does A(3) − A(1) represent? (c) What area does A(3) − A(1) represent?
72 contemporary calculus

1.1 Practice Answers

1. (a) 2
(b) 2
(c) does not exist (no limit)
(d) 1
( x + 1)( x − 2)
2. (a) lim = lim ( x + 1) = 3
x →2 x−2 x →2

t sin(t) sin(t) 0
(b) lim = lim = =0
t →0 t ( t + 3) t →0 t + 3 3
w−2
(c) lim = 2 To see this, make a graph or a table:
w →2 ln( w2 )
w −2 w −2
w ln( w2 )
w ln( w2 )

2.2 2.098411737 1.9 1.949572575

2.01 2.004995844 1.99 1.994995823
2.003 2.001499625 1.9992 1.999599973
2.0001 2.00005 1.9999 1.99995
↓ ↓ ↓ ↓
2 2 2 2

3. lim f ( x ) = 1 lim f ( x ) = 2 lim f ( x ) DNE

x →0− x →0+ x →0

lim f ( x ) = 1 lim f ( x ) = 1 lim f ( x ) = 1

x →1− x →1+ x →1

lim f ( x ) = −1 lim f ( x ) = −1 lim f ( x ) = −1

x →2− x →2+ x →2

lim f ( x ) = −1 lim f ( x ) = 1 lim f ( x ) DNE

x →3− x →3+ x →3

4. lim f ( x ) = 1 lim f ( x ) = 1 lim f ( x ) = 1

x →1− x →1+ x →1

lim f ( x ) = 3 lim f ( x ) = 2 lim f ( x ) DNE

x →3− x →3+ x →3
 limits and continuity 73

1.2 Properties of Limits

This section presents results that make it easier to calculate limits of

combinations of functions or to show that a limit does not exist. The
main result says we can determine the limit of “elementary combina-
tions” of functions by calculating the limit of each function separately
and recombining these results to get our final answer.

Main Limit Theorem:

If lim f ( x ) = L and lim g( x ) = M
x→a x→a

then (a) lim [ f ( x ) + g( x )] = L + M

x→a

(b) lim [ f ( x ) − g( x )] = L − M
x→a

(c) lim k · f ( x ) = k · L
x→a

(d) lim f ( x ) · g( x ) = L · M
x→a

f (x) L
(e) lim = (if M ̸= 0)
x→a g( x ) M

(f) lim [ f ( x )]n = Ln

x→a
q
n
√
n
(g) lim f (x) = L When n is an even integer in part (g) of
x→a the Main Limit Theorem, we need L ≥ 0
and f ( x ) ≥ 0 for x near a.

The Main Limit Theorem says we get the same result if we first
perform the algebra and then take the limit or if we take the limits first
and then perform the algebra: for example, (a) says that the limit of the
sum equals the sum of the limits.
A proof of the Main Limit Theorem is not inherently difficult, but it
requires a more precise definition of the limit concept than we have at
the moment, and it then involves a number of technical difficulties.

Practice 1. For f ( x ) = x2 − x − 6 and g( x ) = x2 − 2x − 3, evaluate:

(a) lim [ f ( x ) + g( x )] (e) lim f ( x ) · g( x )

x →1 x →3

(b) lim f ( x ) · g( x ) f (x)

x →1 (f) lim
x →3 g( x )
f (x)
(c) lim (g) lim [ f ( x )]3
x →1 g( x ) x →2
q
(d) lim [ f ( x ) + g( x )] (h) lim 1 − g( x )
x →3 x →2
74 contemporary calculus

Limits of Some Very Nice Functions: Substitution

As you may have noticed in the previous example, for some functions
f ( x ) it is possible to calculate the limit as x approaches a simply by
substituting x = a into the function and then evaluating f ( a), but
sometimes this method does not work. The following results help to
(partially) answer the question about when such a substitution is valid.

Two Easy Limits:

lim k = k and lim x = a

x→a x→a

We can use the preceding Two Easy Limits and the Main Limit
Theorem to prove the following Substitution Theorem.

Substitution Theorem For Polynomial and Rational Functions:

If P( x ) and Q( x ) are polynomials and a is any number

P( x ) P( a)
then lim P( x ) = P( a) and lim =
x→a x→a Q( x ) Q( a)
as long as Q( a) ̸= 0.

The Substitution Theorem says that we can calculate the limits of

polynomials and rational functions by substituting (as long as the
substitution does not result in a division by 0).

Practice 2. Evaluate each limit.

x2 − 2x
h i
(a) lim 5x3 − x2 + 3 (c) lim
x →2
x →2 x2 − x − 2
x3 − 7x
(b) lim
x →2 x2 + 3x

Limits of Other Combinations of Functions

So far we have concentrated on limits of single functions and elementary

combinations of functions. If we are working with limits of other com-
binations or compositions of functions, the situation becomes slightly
more difficult, but sometimes these more complicated limits have useful
geometric interpretations.

Example 1. Use the graph in the margin to estimate each limit.

(a) lim [3 + f ( x )] (c) lim f (3 − x )

x →1 x →0

(b) lim f (2 + x ) (d) lim [ f ( x + 1) − f ( x )]

x →1 x →2
 limits and continuity 75

Solution. (a) lim [3 + f ( x )] requires a straightforward application of

x →1
part (a) of the Main Limit Theorem:

lim [3 + f ( x )] = lim 3 + lim f ( x ) = 3 + 2 = 5

x →1 x →1 x →1

(b) We first need to examine what happens to the quantity 2 + x as

x → 1 before we can consider the limit of f (2 + x ). When x is
very close to 1, the value of 2 + x is very close to 3, so the limit
of f (2 + x ) as x → 1 is equivalent to the limit of f (w) as w → 3
(where w = 2 + x) and it is clear from the graph that lim f (w) = 1,
w →3
so:
lim f (2 + x ) = lim f (w) = 1
x →1 w →3

In most situations it is not necessary to formally substitute a new

variable w for the quantity 2 + x, but it is still necessary to think
about what happens to the quantity 2 + x as x → 1.

(c) As x → 0 the quantity 3 − x will approach 3, so we want to

know what happens to the values of f when the input variable is
approaching 3:
lim f (3 − x ) = 1
x →0

(d) Using part (b) of the Main Limit Theorem:

lim [ f ( x + 1) − f ( x )] = lim f ( x + 1) − lim f ( x )

x →2 x →2 x →2
= lim f (w) − lim f ( x ) = 1 − 3 = −2
w →3 x →2

Notice the use of the substitution w = x + 1 above. ◀

Practice 3. Use the graph in the margin to estimate each limit.

(a) lim f (2x ) (c) lim 3 · f (4 + x )

x →1 x →0

(b) lim f ( x − 1) (d) lim f (3x − 2)

x →2 x →2

Example 2. Use the graph in the margin to estimate each limit.

(a) lim f (3 + h) (c) lim [ f (3 + h) − f (3)]

h →0 h →0

f (3 + h ) − f (3)
(b) lim f (3) (d) lim
h →0 h →0 h

Solution. The last limit is a special type of limit we will encounter

often in this book, while the first three parts are the steps we need to
evaluate it.
76 contemporary calculus

(a) As h → 0, the quantity w = 3 + h will approach 3, so

lim f (3 + h) = lim f (w) = 1

h →0 w →3

(b) f (3) is a constant (equal to 1) and does not depend on h in any

way, so:
lim f (3) = f (3) = 1
h →0

(c) This limit is just an algebraic combination of the first two limits:

lim [ f (3 + h) − f (3)] = lim f (3 + h) − lim f (3) = 1 − 1 = 0

h →0 h →0 h →0

The quantity f (3 + h) − f (3) also has a geometric interpretation:

it is the change in the y-coordinates, the ∆y, between the points
(3, f (3)) and (3 + h, f (3 + h)) (see margin figure).
f (3 + h ) − f (3)
(d) As h → 0, the numerator and denominator of
h
both approach 0, so we cannot immediately determine the value
of the limit. But if we recognize that f (3 + h) − f (3) = ∆y for
the two points (3, f (3)) and (3 + h, f (3 + h)) and that h = ∆x for
f (3 + h ) − f (3)
the same two points, then we can interpret as
h
∆y
∆x , which is the slope of the secant line through the two points:

f (3 + h ) − f (3)
lim = lim [slope of the secant line]
h →0 h ∆x →0
= slope of the tangent line at (3, f (3))
≈ −1

This last limit represents the slope of line tangent to the graph of
f at the point (3, f (3)).
It is a pattern we will encounter often. ◀

Tangent Lines as Limits

If we have two points on the graph of the function y = f ( x ):

( x, f ( x )) and ( x + h, f ( x + h))

then ∆y = f ( x + h) − f ( x ) and ∆x = ( x + h) − ( x ) = h, so the slope of

the secant line through those points is:
∆y
msec =
∆x
and the slope of the line tangent to the graph of f at the point ( x, f ( x ))
is, by definition,
f ( x + h) − f ( x )
mtan = lim [slope of the secant line] = lim
∆x →0 h →0 h
 limits and continuity 77

Example 3. Give a geometric interpretation for the following limits

and estimate their values for the function whose graph appears in the
margin.

f (1 + h ) − f (1) f (2 + h ) − f (2)
(a) lim (b) lim
h →0 h h →0 h

Solution. (a) The limit represents the slope of the line tangent to the
f (1 + h ) − f (1)
graph of f ( x ) at the point (1, f (1)), so lim ≈ 1. (b) The
h →0 h
limit represents the slope of the line tangent to the graph of f ( x ) at the
f (2 + h ) − f (2)
point (2, f (2)), so lim ≈ −1. ◀
h →0 h

Practice 4. Give a geometric interpretation for the following limits

and estimate their values for the function whose graph appears in the
margin.

g (1 + h ) − g (1) g ( h ) − g (0)
(a) lim (c) lim
h →0 h h →0 h
g (3 + h ) − g (3)
(b) lim
h →0 h

Comparing the Limits of Functions

Sometimes it is difficult to work directly with a function. However, if we
can compare our complicated function with simpler ones, then we can
use information about the simpler functions to draw conclusions about
the complicated one. If the complicated function is always between
two functions whose limits are equal, then we know the limit of the
complicated function.

Squeezing Theorem:

If g( x ) ≤ f ( x ) ≤ h( x ) for all x near (but not equal to) c

and lim g( x ) = lim h( x ) = L

x →c x →c

then lim f ( x ) = L.
x →c

The margin figure shows the idea behind the proof of this theorem: the
function f ( x ) gets “squeezed” between the smaller function g( x ) and
the bigger function h( x ). Because g( x ) and h( x ) converge to the same
limit, L, so must f ( x ).
We can use the Squeezing Theorem to evaluate some “hard” limits by
squeezing a “complicated” function in between two “simpler” functions
with “easier” limits.
78 contemporary calculus

Example 4. Use the inequality − | x | ≤ sin( x ) ≤ | x | to determine:

(a) lim sin( x ) (b) lim cos( x )

x →0 x →0

Solution. (a) lim | x | = 0 and lim − | x | = 0 so, by the Squeezing The-

x →0 x →0 q
orem, lim sin( x ) = 0. (b) If − π2 < x < π2 , then cos( x ) = 1 − sin2 ( x ),
x →0 q p
so lim cos( x ) = lim 1 − sin2 ( x ) = 1 − 02 = 1. ◀
x →0 x →0
 
1
Example 5. Evaluate lim x · sin .
x →0 x
 
Solution. In the graph of sin 1x (see margin), the y-values change
very rapidly for values of x near 0, but they all lie between −1 and 1:
 
1
−1 ≤ sin ≤1
x
so, if x > 0, multiplying this inequality by x we get:
 
1
− x ≤ x · sin ≤x
x
which we can rewrite as:
 
1
− | x | ≤ x · sin ≤ |x|
x
because | x | = x when x > 0.
If x < 0, when we multiply the original inequality by x we get:
   
1 1
− x ≥ x · sin ≥ x ⇒ | x | ≥ x · sin ≥ − |x|
x x
because | x | = − x when x < 0. Either way we have:
 
1
− | x | ≤ x · sin ≤ |x|
x
for all x ̸= 0, and in particular for x near 0. Both “simple” functions
(− | x | and | x |) approach 0 as x → 0, so
 
1
lim x · sin =0
x →0 x
by the Squeezing Theorem. ◀

Practice 5. If f ( x ) is always between x2 + 2 and 2x + 1, what can you

say about lim f ( x )?
x →1
sin( x )
Problem 27 guides you through the steps Practice 6. Use the relation cos( x ) ≤ x ≤ 1 to show that:
to prove this relation.
sin( x )
lim =1
x →0 x
 limits and continuity 79

List Method for Showing that a Limit Does Not Exist

If the limit of f ( x ), as x approaches c, exists and equals L, then we
can guarantee that the values of f ( x ) are as close to L as we want by
restricting the values of x to be very, very close to c. To show that a
limit, as x approaches c, does not exist, we need to show that no matter
how closely we restrict the values of x to c, the values of f ( x ) are not
all close to a single, finite value L.
One way to demonstrate that lim f ( x ) does not exist is to show that
x →c
the left and right limits exist but are not equal.
Another method of showing that lim f ( x ) does not exist uses two
x →c
(infinite) lists of numbers, { a1 , a2 , a3 , a4 , . . .} and {b1 , b2 , b3 , b4 , . . .}, with
entries that become arbitrarily close to the value c as the subscripts
get larger, but for which the corresponding lists of function values,
{ f ( a1 ), f ( a2 ), f ( a3 ), f ( a4 ), . . .} and { f (b1 ), f (b2 ), f (b3 ), f (b4 ), . . .} approach
two different numbers as the subscripts get larger.
Example 6. For f ( x ) defined as:

 1 if x < 1

f (x) = x if 1 < x < 3

 2 if 3 < x

show that lim f ( x ) does not exist.

x →3

Solution. We could use one-sided limits to show that this limit does
not exist, but instead we will use the list method.
One way to define values of { a1 , a2 , a3 , a4 , . . .} that approach 3 from
the right is to define a1 = 3 + 1, a2 = 3 + 12 , a3 = 3 + 13 , a4 = 3 + 14 and,
in general, an = 3 + n1 . Then an > 3 so f ( an ) = 2 for all subscripts n,
and the values in the list { f ( a1 ), f ( a2 ), f ( a3 ), f ( a4 ), . . .} are all close to
2 — in fact, all of the f ( an ) values equal 2.
We can define values of {b1 , b2 , b3 , b4 , . . .} that approach 3 from the
left by b1 = 3 − 1, b2 = 3 − 12 , b3 = 3 − 13 , b4 = 3 − 14 , and, in gen-
eral, bn = 3 − n1 . Then bn < 3 so f (bn ) = bn = 3 − n1 for each sub-
script
n n, and the values in the o list { f (b1 ), f (b2 ), f (b3 ), f (b4 ), . . .} =
2, 2.5, 2 32 , 2 34 , 2 45 , . . . , 3 − n1 , . . . are all close to 3 for large values of n.
Because the values in the lists { f ( a1 ), f ( a2 ), f ( a3 ), f ( a4 ), . . .} and
{ f (b1 ), f (b2 ), f (b3 ), f (b4 ), . . .} have two different limiting values, we
can conclude that lim f ( x ) does not exist. ◀
x →3

Example 7. Define h( x ) as:

(
2 if x is a rational number
h( x ) =
1 if x is an irrational number
(the “holey” function introduced in Section 0.4). Use the list method to
show that lim h( x ) does not exist.
x →3
80 contemporary calculus

Solution. Let { a1 , a2 , a3 , a4 , . . .} be a list of rational numbers that ap-

proach 3: for example, a1 = 3 + 1, a2 = 3 + 12 , a3 = 3 + 13 ,. . . , an = 3 + n1 .
Then f ( an ) = 2 for all n, so:

{ f ( a1 ), f ( a2 ), f ( a3 ), f ( a4 ), . . .} = {2, 2, 2, 2, . . .}

and the f ( an ) values are all “close to” (in fact, equal) 2.
If {b1 , b2 , b3 , b4 , . . .} is a list of irrational numbers that approach 3
(for example, b1 = 3 + π, b2 = 3 + π2 ,. . . ,bn = 3 + πn ) then:

{ f (b1 ), f (b2 ), f (b3 ), f (b4 ), . . .} = {1, 1, 1, 1, . . .}

and the f (bn ) values are all close to 1 for large values of n.
Because the f ( an ) and f (bn ) values become close to two different
numbers, the limit of f ( x ) as x → 3 does not exist. A similar argument
will work as x approaches any number c, so for every c we can show
that lim ( x ) does not exist. The “holey” function does not have a limit
x →c
as x approaches any value c. ◀

1.2 Problems

1. Use the functions f and g defined by the graphs 3. Use the function h defined by the graph below to
below to determine the following limits. determine the following limits.

(a) lim [ f ( x ) + g( x )] (b) lim f ( x ) · g( x ) (a) lim h(2x − 2) (b) lim [ x + h( x )]

x →1 x →1 x →2 x →2
x
f (x) (d) lim f ( g( x )) (c) lim h(1 + x ) (d) lim h
(c) lim x →1 x →2 x →3 2
x →1 g( x )

2. Use the functions f and g defined by the graphs 4. Use the function h defined by the graph above to
above to determine the following limits. determine the following limits.

(a) lim [ f ( x ) + g( x )] (b) lim f ( x ) · g( x ) (a) lim h(5 − x ) (b) lim [h(3 + x ) − h(3)]
x →2 x →2 x →2 x →0

f (x) (d) lim f ( g( x )) (c) lim x · h( x − 1) h (3 + x ) − h (3)

(c) lim x →2 x →2 (d) lim
x →2 g( x ) x →0 x
 limits and continuity 81

5. Label the parts of the graph of f (below) that are (g) lim f ( x ) (h) lim f ( x ) (i) lim f ( x )
x →−1+ x →−1− x →−1
described by
8. Use the graph from Problem 7 to estimate:
(a) 2 + h (b) f (2)
(a) lim f ( x ) (b) lim f ( x ) (c) lim f ( x )
(c) f (2 + h) (d) f (2 + h) − f (2) x →2+ x →2− x →2

f (2 + h ) − f (2) f (2 − h ) − f (2) (d) lim f ( x ) (e) lim f ( x ) (f) lim f ( x )

(e) (f) x →4+ x →4− x →4
(2 + h ) − 2 (2 − h ) − 2
(g) lim f ( x ) (h) lim f ( x ) (i) lim f ( x )
x →−2+ x →−2− x →−2

9. The Lorentz contraction formula in relativity the-

ory says the length L of an object moving at v
miles per second with respect to an observer is:
r
v2
L = A· 1− 2
6. Label the parts of the graph of g (below) that are c
described by where c is the speed of light (a constant).
(a) Determine the object’s “rest length” (v = 0).
(a) a + h (b) g( a)
(b) Determine: lim L
v→c−
(c) g( a + h) (d) g( a + h) − g( a)
10. Evaluate each limit.
g( a + h) − g( a) g( a − h) − g( a)
(e) (f) (a) lim ⌊ x ⌋ (b) lim ⌊ x ⌋
( a + h) − a ( a − h) − a x →2+ x →2−

(c) lim ⌊x⌋ (d) lim ⌊x⌋

x →−2+ x →−2−
jxk
(e) lim ⌊x⌋ (f) lim
x →−2.3 x →3 2

⌊x⌋ ⌊2 + x ⌋ − ⌊2⌋
(g) lim (h) lim
x →3 2 x →0+ x

11. For f ( x ) and g( x ) defined as:

7. Use the graph below estimate:
( (
1 if x < 1 x if x ̸= 2
f (x) = g( x ) =
x if 1 < x 3 if x = 2

determine the following limits:

(a) lim [ f ( x ) + g( x )] f (x)

x →2 (b) lim
x →2 g( x )

(c) lim f ( g( x )) g( x )
x →2 (d) lim
(a) lim f ( x ) (b) lim f ( x ) (c) lim f ( x ) x →0 f (x)
x →1+ x →1− x →1
f (x) (f) lim g( f ( x ))
(d) lim f ( x ) (e) lim f ( x ) (f) lim f ( x ) (e) lim x →1
x →3+ x →3− x →3 x →1 g( x )
 82 contemporary calculus

12. Give geometric interpretations for each limit and 16. (a) For h > 0, find the slope of the line through
use a calculator to estimate its value. the points (h, |h|) and (0, 0).
arctan(0 + h) − arctan(0) (b) For h < 0, find the slope of the line through
(a) lim
h →0 h
the points (h, |h|) and (0, 0).
arctan(1 + h) − arctan(1) |h| |h| |h|
(b) lim (c) Evaluate lim , lim and lim .
h →0 h h →0− h h →0+ h h →0 h
arctan(2 + h) − arctan(2)
(c) lim
h →0 h In 17–18, describe the behavior at each integer
of the function y = f ( x ) in the figure provided,
using one of these phrases:
• “connected and smooth”
• “connected with a corner”
• “not connected because of a simple hole that
could be plugged by adding or moving one
point”
• “not connected because of a vertical jump that
could not be plugged by moving one point”

cos(h) − 1
13. (a) What does lim represent in rela-
h →0 h
tion to the graph of y = cos( x )? It may help 17.
to recognize that:
cos(h) − 1 cos(0 + h) − cos(0)
=
h h
(b) Graphically and using your calculator, esti-
cos(h) − 1
mate lim .
h →0 h
ln(1 + h)
14. (a) What does the ratio represent in re-
h
lation to the graph of y = ln( x )? It may help 18.
to recognize that:
ln(1 + h) ln(1 + h) − ln(1)
=
h h
(b) Graphically and using your calculator, deter-
ln(1 + h) | x − 2|
mine lim . 19. Use the list method to show that lim does
h x →2 x−2
h →0 not exist .
15. Use your calculator (to generate a table of values)  
1
to help you estimate the value of each limit. 20. Show that lim sin does not exist. (Sugges-
x →0 x 
eh − 1
(a) lim tion: Let f ( x ) = sin 1x and let an = nπ 1
so
h →0 h  
tan(1 + c) − tan(1) that f ( an ) = sin a1n = sin(nπ ) = 0 for ev-
(b) lim
c →0 c ery n. Then pick bn = 2nπ1+ π so that f (bn ) =
2
g (2 + t ) − g (2)  
(c) lim when g(t) = t2 − 5. 1
sin bn = sin(2nπ + 2 ) = sin( π2 ) = 1 for all n.)
π
t →0 t
 limits and continuity 83

   
In Problems 21–26, use the Squeezing Theorem to 1 p 1
23. lim 3 + x2 sin 24. lim 1 − x2 cos
help evaluate each limit. x →0 x x →1− x−1

√ 1
     
21. lim x2 cos
1
22. lim 3
x sin
1
25. lim x2 ·
1 26. lim (−1)⌊ x ⌋ (1 − cos( x ))
x →0 x2 x →0 x3 x →0 x2 x →0

sin(θ )
27. This problem outlines the steps of a proof that lim = 1.
θ θ →0+
π
Refer to the margin figure, assume that 0 < θ < 2 , and justify why
each statement must be true.

(a) Area of △OPB = 12 (base)(height) = 21 sin(θ )

area of the sector (the pie shaped region) OPB θ
(b) =
area of the whole circle 2π
θ θ
(c) area of the sector OPB = π · =
2π 2
(d) The line L through the points (0, 0) and P = (cos(θ ), sin(θ )) has
sin(θ ) sin(θ )
slope m = , so C = (1, )
cos(θ ) cos(θ )
1 1 sin(θ )
(e) area of △OCB = (base)(height) = (1)
2 2 cos(θ )
(f) area of △OPB < area of sector OPB < area of △OCB
1 θ 1 sin(θ ) sin(θ )
(g) sin(θ ) < < (1) ⇒ sin(θ ) < θ <
2 2 2 cos(θ ) cos(θ )
θ 1 sin(θ )
(h) 1 < < ⇒ 1> > cos(θ )
sin(θ ) cos(θ ) θ
(i) lim 1 = 1 and lim cos(θ ) = 1.
θ →0+ θ →0+
sin(θ )
(j) lim =1
θ →0+ θ

1.2 Practice Answers

3 5
1. (a) −10 (b) 24 (c) 2 (d) 0 (e) 0 (f) 4 (g) −64 (h) 2

2. (a) 39 (b) − 53 (c) 2

3 3. (a) 0 (b) 2 (c) 3 (d) 1

4. (a) slope of the line tangent to the graph of g at the point (1, g(1));
estimated slope ≈ −2
(b) slope of the line tangent to the graph of g at the point (3, g(3));
estimated slope ≈ 0
(c) slope of the line tangent to the graph of g at the point (0, g(0));
estimated slope ≈ 1
h i
5. lim x2 + 2 = 3 and lim [2x + 1] = 3 so lim f ( x ) = 3
x →1 x →1 x →1

sin( x )
6. lim cos( x ) = 1 and lim 1 = 1 so lim =1
x →0 x →0 x →0 x
84 contemporary calculus

1.3 Continuous Functions

In Section 1.2 we saw a few “nice” functions whose limits as x → a

simply involved substituting a into the function: lim f ( x ) = f ( a).
x→a
Functions whose limits have this substitution property are called con-
tinuous functions and such functions possess a number of other useful
properties.
In this section we will examine what it means graphically for a func-
tion to be continuous (or not continuous), state some properties of con-
tinuous functions, and look at a few applications of these properties —
2x + 1
including a way to solve horrible equations such as sin( x ) = .
x−2

Definition of a Continuous Function

We begin by formally stating the definition of this new concept.

Definition of Continuity at a Point:

A function f is continuous at x = a if and only if

lim f ( x ) = f ( a).
x→a

The graph in the margin illustrates some of the different ways a

function can behave at and near a point, and the accompanying table
contains some numerical information about the example function f and
its behavior. We can conclude from the information in the table that f
is continuous at 1 because lim f ( x ) = 2 = f (1).
x →1
We can also conclude that f is not continuous at 2 or 3 or 4, because
lim f ( x ) ̸= f (2), lim f ( x ) ̸= f (3) and lim f ( x ) ̸= f (4).
x →2 x →3 x →4

a f ( a) lim f ( x )
x→a
Graphical Meaning of Continuity
1 2 2
When x is close to 1, the values of f ( x ) are close to the value f (1), and
2 1 2
3 2 DNE the graph of f does not have a hole or break at x = 1. The graph of f is
4 undefined 2 “connected” at x = 1 and can be drawn without lifting your pencil. At
x = 2 and x = 4 the graph of f has “holes,” and at x = 3 the graph has
a “break.” The function f is also continuous at 1.7 (why?) and at every
point shown except at 2, 3 and 4.

Informally, we can say:

• A function is continuous at a point if the graph of the function

is connected there.

• A function is not continuous at a point if its graph has a hole or

break at that point.
 limits and continuity 85

Sometimes the definition of “continuous” (the substitution condition

for limits) is easier to use if we chop it into several smaller pieces and
then check whether or not our function satisfies each piece.

f is continuous at a if and only if:

(i) f is defined at a

(ii) the limit of f ( x ), as x → a, exists

(so the left limit and right limits exist and are equal)

(iii) the value of f at a equals the value of the limit as x → a:

lim f ( x ) = f ( a)
x→a

If f satisfies conditions (i), (ii) and (iii), then f is continuous at a. If

f does not satisfy one or more of the three conditions at a, then f is not
continuous at a.
For f ( x ) in the figure on the previous page, all three conditions are
satisfied for a = 1, so f is continuous at 1. For a = 2, conditions (i)
and (ii) are satisfied but not (iii), so f is not continuous at 2. For a = 3,
condition (i) is satisfied but (ii) is violated, so f is not continuous at 3.
For a = 4, condition (i) is violated, so f is not continuous at 4.
A function is continuous on an interval if it is continuous at every
point in the interval.
A function f is continuous from the left at a if lim f ( x ) = f ( a)
x → a−
and is continuous from the right at a if lim f ( x ) = f ( a).
x → a+

Example 1. Is the function


 x+1
 if x ≤ 1
f (x) = 2 if 1 < x ≤ 2
1


x −3 if x > 2

continuous at x = 1? At x = 2? At x = 3?

Solution. We could answer these questions by examining a graph of

f ( x ), but let’s try them without the benefit of a graph. At x = 1,
f (1) = 2 and the left and right limits are equal:

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

x →1− x →1− x →1+ x →1+

and their common limit matches the value of the function at x = 1:

lim f ( x ) = 2 = f (1)
x →1

so f is continuous at 1.
86 contemporary calculus

At x = 2, f (2) = 2, but the left and right limits are not equal:
1
lim f ( x ) = lim 2 = 2 ̸= −1 = lim = lim f ( x )
x →2− x →1− x →2+ x−3 x →2+

so f fails condition (ii), hence is not continuous at 2. We can, however,

say that f is continuous from the left (but not from the right) at 2.
1
At x = 3, f (3) = , which is undefined, so f is not continuous at 3
0
because it fails condition (i). ◀

Example 2. Where is f ( x ) = 3x2 − 2x continuous?

Solution. By the Substitution Theorem for Polynomial and Rational

Functions, lim P( x ) = P( a) for any polynomial P( x ) at any point a,
x→a
so every polynomial is continuous everywhere. In particular, f ( x ) =
3x2 − 2x is continuous everywhere. ◀
x+5
Example 3. Where is the function g( x ) = continuous? Where is
x−3
x2 + 4x − 5
h( x ) = continuous?
x2 − 4x + 3
Solution. Because g( x ) is a rational function, the Substitution Theorem
for Polynomial and Rational Functions says it is continuous everywhere
except where its denominator is 0: g is continuous everywhere except
at x = 3. The graph of g (see margin) is “connected” everywhere except
at x = 3, where it has a vertical asymptote.
We can rewrite the rational function h( x ) as:
( x − 1)( x + 5)
h( x ) =
( x − 1)( x − 3)
and note that its denominator is 0 at x = 1 and x = 3, so h is continuous
everywhere except 3 and 1. The graph of h (see margin) is “connected”
everywhere except at 3, where it has a vertical asymptote, and 1, where
it has a hole: f (1) = 00 is undefined. ◀

Example 4. Where is f ( x ) = ⌊ x ⌋ continuous?

Solution. The graph of y = ⌊ x ⌋ seems to be “connected” except at

each integer, where there is a “jump” (see margin).
If a is an integer, then lim ⌊ x ⌋ = a − 1 and lim ⌊ x ⌋ = a so lim ⌊ x ⌋
x → a− x → a+ x→a
is undefined, and ⌊ x ⌋ is not continuous at x = a.
If a is not an integer, then the left and right limits of ⌊ x ⌋, as x → a,
both equal ⌊ a⌋ so: lim ⌊ x ⌋ = ⌊ a⌋, hence ⌊ x ⌋ is continuous at x = a.
x→a
Summarizing: ⌊ x ⌋ is continuous everywhere except at the integers.
In fact, f ( x ) = ⌊ x ⌋ is continuous from the right everywhere and is
continuous from the left everywhere except at the integers. ◀
|x|
Practice 1. Where is f ( x ) = continuous?
x
 limits and continuity 87

Why Do We Care Whether a Function Is Continuous?

There are several reasons for us to examine continuous functions and
their properties:

• Many applications in engineering, the sciences and business are

continuous or are modeled by continuous functions or by pieces of
continuous functions.

• Continuous functions share a number of useful properties that do

not necessarily hold true if the function is not continuous. If a
result is true of all continuous functions and we have a continuous
function, then the result is true for our function. This can save us
from having to show, one by one, that each result is true for each
particular function we use. Some of these properties are given in the
remainder of this section.

• Differential calculus has been called the study of continuous change,

and many of the results of calculus are guaranteed to be true only for
continuous functions. If you look ahead into Chapters 2 and 3, you
will see that many of the theorems have the form “If f is continuous
and (some additional hypothesis), then (some conclusion).”

Combinations of Continuous Functions

Not only are most of the basic functions we will encounter continuous
at most points, so are basic combinations of those functions.

Theorem:
If f ( x ) and g( x ) are continuous at a
and k is any constant

then the elementary combinations of f and g

• k · f (x)

• f ( x ) + g( x )

• f ( x ) − g( x )

• f ( x ) · g( x )
f (x)
• (as long as g( a) ̸= 0)
g( x )

are continuous at a.

The continuity of a function is defined using limits, and all of these

results about simple combinations of continuous functions follow from
the results about combinations of limits in the Main Limit Theorem.
88 contemporary calculus

Our hypothesis is that f and g are both continuous at a, so we can

assume that

lim f ( x ) = f ( a) and lim g( x ) = g( a)

x→a x→a

and then use the appropriate part of the Main Limit Theorem.
For example,

lim [ f ( x ) + g( x )] = lim f ( x ) + lim g( x ) = f ( a) + g( a)

x→a x→a x→a

so f + g is continuous at a.

Practice 2. Prove: If f and g are continuous at a, then k · f and f − g

are continuous at a (where k a constant).

Composition of Continuous Functions:

If g( x ) is continuous at a and
f ( x ) is continuous at g( a)

then lim f ( g( x )) = f ( lim g( x )) = f ( g( a))

x→a x→a
so f ◦ g( x ) = f ( g( x )) is continuous at a.

The proof of this result involves some technical details, but just
formalizes the following line of reasoning:
The hypothesis that “g is continuous at a” means that if x is close
to a then g( x ) will be close to g( a). Similarly, “ f is continuous at g( a)”
means that if g( x ) is close to g( a) then f ( g( x )) = f ◦ g( x ) will be close
to f ( g( a)) = f ◦ g( a). Finally, we can conclude that if x is close to a,
then g( x ) is close to g( a) so f ◦ g( x ) is close to f ◦ g( a) and therefore
f ◦ g is continuous at x = a.
The next theorem presents an alternate version of the limit condition
for continuity, which we will use occasionally in the future.

Theorem:

lim f ( x ) = f ( a) if and only if lim f ( a + h) = f ( a)

x→a h →0

Proof. Let’s define a new variable h by h = x − a so that x = a + h

(see margin figure). Then x → a if and only if h = x − a → 0, so
lim f ( x ) = lim f ( a + h) and therefore lim f ( x ) = f ( a) if and only if
x→a h →0 x→a
lim f ( a + h) = f ( a).
h →0

We can restate the result of this theorem as:

A function f is continuous at a if and only if lim f ( a + h) = f ( a).

h →0
 limits and continuity 89

Which Functions Are Continuous?

Fortunately, the functions we encounter most often are either continu-
ous everywhere or continuous everywhere except at a few places.

Theorem: The following functions are continuous everywhere

(a) polynomials (b) sin( x ) and cos( x ) (c) | x |

Proof. (a) This follows from the Substitution Theorem for Polynomial
and Rational Functions and the definition of continuity.

(b) The graph of y = sin( x ) (see margin) clearly has no holes or breaks,
so it is reasonable to think that sin( x ) is continuous everywhere.
Justifying this algebraically, for every real number a:
Recall the angle addition formula for
lim sin( a + h) = lim [sin( a) cos(h) + cos( a) sin(h)] sin(θ ) and the results from Section 1.2
h →0 h →0 that lim cos(h) = 1 and lim sin(h) = 0.
h →0 h →0
= lim sin( a) · lim cos(h) + lim cos( a) · lim sin(h)
h →0 h →0 h →0 h →0
= sin( a) · 1 + cos( a) · 0 = sin( a)
so f ( x ) = sin( x ) is continuous at every point. The justification for
f ( x ) = cos( x ) is similar.

(c) For f ( x ) = | x |, when x > 0, then | x | = x and its graph (see margin)
is a straight line and is continuous because x is a polynomial.
When x < 0, then | x | = − x and it is also continuous. The only
questionable point is the “corner” on the graph when x = 0, but
the graph there is only bent, not broken:

lim |0 + h| = lim h = 0
h →0+ h →0+

and:
lim |0 + h| = lim −h = 0
h →0− h →0−
so:
lim |0 + h| = 0 = |0|
h →0
and f ( x ) = | x | is also continuous at 0.

A continuous function can have corners but not holes or breaks.

Even functions that fail to be continuous at some points are often

continuous most places:
• A rational function is continuous except where the denominator is 0.

• The trig functions tan( x ), cot( x ), sec( x ) and csc( x ) are continuous
except where they are undefined.
90 contemporary calculus

• The greatest integer function ⌊ x ⌋ is continuous except at each integer.

• But the “holey” function

(
2 if x is a rational number
h( x ) =
1 if x is an irrational number

is discontinuous everywhere.

Intermediate Value Property of Continuous Functions

Because the graph of a continuous function is connected and does not
have any holes or breaks in it, the values of the function can not “skip”
or “jump over” a horizontal line (see margin figure). If one value of the
continuous function is below the line and another value of the function
is above the line, then somewhere the graph will cross the line. The
next theorem makes this statement more precise. The result seems
obvious, but its proof is technically difficult and is not given here.

Intermediate Value Theorem for Continuous Functions:

If f is continuous on the interval [ a, b]
and V is any value between f ( a) and f (b)

then there is a number c between a and b so that

f (c) = V. (That is, f actually takes on each
intermediate value between f ( a) and f (b).)

If the graph of f connects the points ( a, f ( a)) and (b, f (b)) and V
is any number between f ( a) and f (b), then the graph of f must cross
the horizontal line y = V somewhere between x = a and x = b (see
margin figure). Since f is continuous, its graph cannot “hop” over the
line y = V.
We often take this theorem for granted in some common situations:

• If a child’s temperature rose from 98.6◦ F to 101.3◦ F, then there was

an instant when the child’s temperature was exactly 100◦ F. (In fact,
every temperature between 98.6◦ F and 101.3◦ F occurred at some
instant.)

• If you dove to pick up a shell 25 feet below the surface of a lagoon,

then at some instant in time you were 17 feet below the surface.
(Actually, you want to be at 17 feet twice. Why?)

• If you started driving from a stop (velocity = 0) and accelerated to a

velocity of 30 kilometers per hour, then there was an instant when
your velocity was exactly 10 kilometers per hour.
 limits and continuity 91

But we cannot apply the Intermediate Value Theorem if the function

is not continuous:

• In 1987 it cost 22¢ to mail a first-class letter inside the United States,
and in 1990 it cost 25¢ to mail the same letter, but we cannot conclude
that there was a time when it cost 23¢ or 24¢ or 24.7¢ to send the
letter. (Postal rates did not increase in a continuous fashion. They
jumped directly from 22¢ to 25¢.)

• Prices, taxes and rates of pay change in jumps — discrete steps —

without taking on the intermediate values.

The Intermediate Value Theorem (IVT) is an example of an “existence

theorem”: it concludes that something exists (a number c so that
f (c) = V). But like many existence theorems, it does not tell us how to
find the the thing that exists (the value of c) and is of no use in actually
finding those numbers or objects.

Bisection Algorithm for Approximating Roots

The IVT can help us finds roots of functions and solve equations. If f
is continuous on [ a, b] and f ( a) and f (b) have opposite signs (one is
positive and one is negative), then 0 is an intermediate value between
f ( a) and f (b) so f will have a root c between x = a and x = b where
f (c) = 0.
While the IVT does not tell us how to find c, it lays the groundwork
for a method commonly used to approximate the roots of continuous
functions.

Bisection Algorithm for Finding a Root of f ( x )

1. Find two values of x (call them a and b) so that f ( a) and f (b) have
opposite signs. (The IVT will then guarantee that f ( x ) has a root
between a and b.)

2. Calculate the midpoint (or bisection point) of the interval [ a, b],

a+b
using the formula m = , and evaluate f (m).
2
3. (a) If f (m) = 0, then m is a root of f and we are done.
(b) If f (m) ̸= 0, then f (m) has the sign opposite f ( a) or f (b):
i. if f ( a) and f (m) have opposite signs, then f has a root in
[ a, m] so put b = m
ii. if f (b) and f (m) have opposite signs, then f has a root in
[m, b] so put a = m

4. Repeat steps 2 and 3 until a root is found exactly or is approximated

closely enough.
92 contemporary calculus

The length of the interval known to contain a root is cut in half

each time through steps 2 and 3, so the Bisection Algorithm quickly
“squeezes” in on a root (see margin figure).
The steps of the Bisection Algorithm can be done “by hand,” but
it is tedious to do very many of them that way. Computers are very
good with this type of tedious repetition, and the algorithm is simple
to program.

Example 5. Find a root of f ( x ) = − x3 + x + 1.

Solution. f (0) = 1 and f (1) = 1 so we cannot conclude that f has a

root between 0 and 1. f (1) = 1 and f (2) = −5 have opposite signs, so
by the IVT (this function is a polynomial, so it is continuous everywhere
and the IVT applies) we know that there is a number c between 1 and
2 such that f (c) = 0 (see figure). The midpoint of the interval [1, 2] is
m = 1+2 2 = 32 = 1.5 and f ( 32 ) = − 78 so f changes sign between 1 and 1.5
and we can be sure that there is a root between 1 and 1.5. If we repeat
the operation for the interval [1, 1.5], the midpoint is m = 1+21.5 = 1.25,
19
and f (1.25) = 64 > 0 so f changes sign between 1.25 and 1.5 and we
know f has a root between 1.25 and 1.5.
Repeating this procedure a few more times, we get:

b+ a
a b m= 2 f ( a) f (b) f (m) root between
1 2 1 −5 1 2
1 2 1.5 1 −5 −0.875 1 1.5
1 1.5 1.25 1 −0.875 0.2969 1.25 1.5
1.25 1.5 1.375 0.2969 −0.875 −0.2246 1.25 1.375
1.25 1.375 1.3125 0.2969 −0.2246 0.0515 1.3125 1.375
1.3125 1.375 1.34375

If we continue the table, the interval containing the root will squeeze
around the value 1.324718. ◀

The Bisection Algorithm has one major drawback: there are some
roots it does not find. The algorithm requires that the function take
on both positive and negative values near the root so that the graph
actually crosses the x-axis. The function f ( x ) = x2 − 6x + 9 = ( x − 3)2
has the root x = 3 but is never negative (see margin figure). We cannot
find two starting points a and b so that f ( a) and f (b) have opposite
signs, so we cannot use the Bisection Algorithm to find the root x = 3.
In Chapter 2 we will see another method — Newton’s Method — that
does find roots of this type.
The Bisection Algorithm requires that we supply two starting x-
values, a and b, at which the function has opposite signs. These values
can often be found with a little “trial and error,” or we can examine the
graph of the function to help pick the two values.
 limits and continuity 93

Finally, the Bisection Algorithm can also be used to solve equations,

because the solution of any equation can always be transformed into an
equivalent problem of finding roots by moving everything to one side
of the equal sign. For example, the problem of solving the equation
x3 = x + 1 can be transformed into the equivalent problem of solving
x3 − x − 1 = 0 or of finding the roots of f ( x ) = x3 − x − 1, which is
equivalent to the problem we solved in the previous example.
2x + 1
Example 6. Find all solutions of sin( x ) = (with x in radians.)
x−2
Solution. We can convert this problem of solving an equation to the
problem of finding the roots of
2x + 1
f ( x ) = sin( x ) − =0
x−2
The function f ( x ) is continuous everywhere except at x = 2, and the
graph of f ( x ) (in the margin) can help us find two starting values
for the Bisection Algorithm. The graph shows that f (−1) is negative
and f (0) is positive, and we know f ( x ) is continuous on the interval
[−1, 0]. Using the algorithm with the starting interval [−1, 0], we know
that a root is contained in the shrinking intervals [−0.5, 0], [−0.25, 0],
[−0.25, −0.125],. . . , [−0.238281, −0.236328],. . . , [−0.237176, −0.237177]
so the root is approximately −0.237177.
We might notice that f (0) = 0.5 > 0 while f (π ) = 0 − 2π +1
π −2 ≈
−6.38 < 0. Why is it wrong to conclude that f ( x ) has another root
between x = 0 and x = π? ◀

1.3 Problems
1. At which points is the function in the graph below discontinuous?

2. At which points is the function in the graph below discontinuous?

94 contemporary calculus

3. Find at least one point at which each function is 6. A continuous function g has the values:
not continuous and state which of the three con-
x 1 2 3 4 5 6 7
ditions in the definition of continuity is violated
at that point. g( x ) −3 1 4 −1 3 −2 −1
(a) g has at least roots between 1 and 5.
x+5 x2 + x − 6
(a) (b)
x−3 x−2 (b) g( x ) = 3.2 in at least places between
q j k x = 1 and x = 7.
(c) cos( x ) (d) x2
(c) g( x ) = −0.7 in at least places between
x x
(e) (f) x = 3 and x = 7.
sin( x ) x

(g) ln( x2 )
π (d) g( x ) = 1.3 in at least places between
(h)
x2 − 6x + 9 x = 2 and x = 6.
(i) tan( x ) (e) Is it possible for g( x ) to equal π for some x-
4. Which two of the following functions are not con- value(s) between 5 and 6?
tinuous? Use appropriate theorems to justify that
each of the other functions is continuous. 7. This problem asks you to verify that the Interme-
7 (b) cos2 ( x5 − 7x + π ) diate Value Theorem is true for some particular
(a) p
2 + sin( x ) functions, intervals and intermediate values. In
each problem you are given a function f , an inter-
x2 − 5 x2 − 5
(c) (d) val [ a, b] and a value V. Verify that V is between
1 + cos2 ( x ) 1 + cos( x )
f ( a) and f (b) and find a value of c in the given
(e) ⌊3 + 0.5 sin( x )⌋ (f) ⌊0.3 sin( x ) + 1.5⌋ interval so that f (c) = V.
q p (a) f ( x ) = x2 on [0, 3], V = 2
(g) cos(sin( x )) (h) x2 − 6x + 10
q (b) f ( x ) = x2 on [−1, 2], V = 3
sin( x )
(i) 3 cos( x ) (j) 2
1
 
(c) f ( x ) = sin( x ) on 0, π2 , V = 2
(k) 1 − 3− x (l) arctan(1 − x2 ) (d) f ( x ) = x on [0, 1], V = 1
3

5. A continuous function f has the values: (e) f ( x ) = x2 − x on [2, 5], V = 4

x 0 1 2 3 4 5
(f) f ( x ) = ln( x ) on [1, 10], V = 2
f (x) 5 3 −2 −1 3 −2
(a) f has at least roots between 0 and 5. 8. Two students claim that they both started with the
(b) f ( x ) = 4 in at least places between points x = 1 and x = 9 and applied the Bisection
x = 0 and x = 5. Algorithm to the function graphed below. The
first student says that the algorithm converged to
(c) f ( x ) = 2 in at least places between the root near x = 8, but the second claims that
x = 0 and x = 5. the algorithm will converge to the root near x = 4.
(d) f ( x ) = 3 in at least places between Who is correct?
x = 0 and x = 5.

(e) Is it possible for f ( x ) to equal 7 for some x-

value(s) between 0 and 5?
 limits and continuity 95

9. Two students claim that they both started with the 12. f ( x ) = x2 − 2 on [0, 3]
points x = 0 and x = 5 and applied the Bisection 13. g( x ) = x3 − 3x2 + 3 on [−1, 0], [1, 2], [2, 4]
Algorithm to the function graphed below. The
14. h(t) = t5 − 3t + 1 on [1, 3]
first student says that the algorithm converged
to the root labeled A, but the second claims that 15. r ( x ) = 5 − 2x on [1, 3]
the algorithm will converge to the root labeled B. 16. s( x ) = sin(2x ) − cos( x ) on [0, π ]
Who is correct?
17. p(t) = t3 + 3t + 1 on [−1, 1]
18. Explain what is wrong with this reasoning:
If f ( x ) = 1x then

f (−1) = −1 < 0 and f (1) = 1 > 0

so f must have a root between x = −1 and x = 1.

10. If you apply the Bisection Algorithm to the func- 19. Each of the following statements is false for some
tion graphed below, which root does the algo- functions. For each statement, sketch the graph
rithm find if you use: of a counterexample.
(a) starting points 0 and 9? (a) If f (3) = 5 and f (7) = −3, then f has a root
(b) starting points 1 and 5? between x = 3 and x = 7.
(c) starting points 3 and 5? (b) If f has a root between x = 2 and x = 5, then
f (2) and f (5) have opposite signs.
(c) If the graph of a function has a sharp corner,
then the function is not continuous there.
20. Define A( x ) to be the area bounded by the t- and
y-axes, the curve y = f (t), and the vertical line
11. If you apply the Bisection Algorithm to the func- t = x (see figure below). It is clear that A(1) < 2
tion graphed below, which root does the algo- and A(3) > 2. Do you think there is a value of x
rithm find if you use: between 1 and 3 so that A( x ) = 2? If so, justify
(a) starting points 3 and 7? your conclusion and estimate the location of the
(b) starting points 5 and 6? value of x that makes A( x ) = 2. If not, justify
(c) starting points 1 and 6? your conclusion.

In 12–17, use the IVT to verify each function has a

root in the given interval(s). Then use the Bisection
Algorithm to narrow the location of that root to an
interval of length less than or equal to 0.1.
 96 contemporary calculus

21. Define A( x ) to be the area bounded by the t- and g( x ) = f ( x ) − x and start by considering the
y-axes, the curve y = f (t), and the vertical line values g(0) and g(1).
t = x (see figure below). (c) What does part (b) have to do with part (a)?
(a) Shade the part of the graph represented (d) Is the theorem in part (b) true if we replace
by A(2.1) − A(2) and estimate the value of the closed interval [0, 1] with the open interval
A(2.1) − A(2) (0, 1)?
.
0.1
23. A piece of string is tied in a loop and tossed onto
(b) Shade the part of the graph represented
quadrant I enclosing a single region (see figure
by A(4.1) − A(4) and estimate the value of
A(4.1) − A(4) below).
.
0.1 (a) Is it always possible to find a line L passing
through the origin so that L divides the region
into two equal areas? (Justify your answer.)
(b) Is it always possible to find a line L parallel to
the x-axis so that L divides the region into two
equal areas? (Justify your answer.)
(c) Is it always possible to find two lines, L paral-
lel to the x-axis and M parallel to the y-axis, so
that L and M divide the region into four equal
areas? (Justify your answer.)
22. (a) A square sheet of paper has a straight line
drawn on it from the lower-left corner to the
upper-right corner. Is it possible for you to
start on the left edge of the sheet and draw
a “connected” line to the right edge that does
not cross the diagonal line?
(b) Prove: If f is continuous on the interval [0, 1]
and 0 ≤ f ( x ) ≤ 1 for all x, then there is a
number c with 0 ≤ c ≤ 1 such that f (c) = c.
(The number c is called a “fixed point” of f
because the image of c is the same as c: f does
not “move” c.) Hint: Define a new function
 limits and continuity 97

1.3 Practice Answers

|x|
1. f ( x ) = (see margin figure) is continuous everywhere except at
x
x = 0, where this function is not defined.
|x|
If a > 0, then lim = 1 = f ( a) so f is continuous at a.
x→a x
|x|
If a < 0, then lim = −1 = f ( a) so f is continuous at a.
x→a x
But f (0) is not defined and

|x| |x|
lim = −1 ̸= 1 = lim
x →0− x x →0+ x

|x|
so lim does not exist.
x→a x
2. (a) To prove that k · f is continuous at a, we need to prove that k · f
satisfies the definition of continuity at a: lim k · f ( x ) = k · f ( a).
x→a
Using results about limits, we know

lim k · f ( x ) = k · lim f ( x ) = k · f ( a)
x→a x→a

(because f is continuous at a) so k · f is continuous at a.

(b) To prove that f − g is continuous at a, we need to prove that f − g
satisfies the definition of continuity at a: lim [ f ( x ) − g( x )] =
x→a
f ( a) − g( a). Again using information about limits:

lim [ f ( x ) − g( x )] = lim f ( x ) − lim g( x ) = f ( a) − g( a)

x→a x→a x→a

(because f and g are both continuous at a) so f − g is continuous

at a.
98 contemporary calculus

1.4 Definition of Limit

It may seem strange that we have been using and calculating the values
of limits for quite a while without having a precise definition of “limit,”
but the history of mathematics shows that many concepts — including
limits — were successfully used before they were precisely defined or
even fully understood. We have chosen to follow the historical sequence,
emphasizing the intuitive and graphical meaning of limit because most
This intuitive and graphical understand- students find these ideas and calculations easier than the definition.
ing of limit was sufficient for the first 100- Mathematics, however, is more than a collection of useful tools, and
plus years of the development of calcu-
lus (from Newton and Leibniz in the late part of its power and beauty comes from the fact that in mathematics
1600s to Cauchy in the early 1800s) and it terms are precisely defined and results are rigorously proved. Math-
is sufficient for using and understanding
the results in beginning calculus. ematical tastes (what is mathematically beautiful, interesting, useful)
change over time, but because of careful definitions and proofs, the
results remain true — everywhere and forever. Textbooks seldom give
all of the definitions and proofs, but it is important to mathematics that
such definitions and proofs exist.
The goal of this section is to provide a precise definition of the limit
of a function. The definition will not help you calculate the values
of limits, but it provides a precise statement of what a limit is. The
definition of limit is then used to verify the limits of some functions
and prove some general results.

The Intuitive Approach

The precise (“formal”) definition of limit carefully states the ideas that
we have already been using graphically and intuitively. The following
side-by-side columns show some of the phrases we have been using to
describe limits, and those phrases — particularly the last ones — provide
the basis on which to build the definition of limit.

A Particular Limit General Limit

lim 2x − 1 = 5 lim f ( x ) = L
x →3 x→a

“as the values of x approach 3, the values of 2x − 1 “as the values of x approach a, the values of f ( x )
approach (are arbitrarily close to) 5” approach (are arbitrarily close to) L”

“when x is close to 3 (but not equal to 3), the value “when x is close to a (but not equal to a), the value
of 2x − 1 is close to 5” of f ( x ) is close to L”

“we can guarantee that the values of 2x − 1 are as “we can guarantee that the values of f ( x ) are as close
close to 5 as we want by restricting the values of x to L as we want by restricting the values of x to be
to be sufficiently close to 3 (but not equal to 3)” sufficiently close to a (but not equal to a)”

Let’s examine what the last phrase (“we can. . . ”) means for the Particu-
lar Limit in the previous discussion.
 limits and continuity 99

Example 1. We know lim 2x − 1 = 5 and need to show that we can

x →3
guarantee that the values of f ( x ) = 2x − 1 are as close to 5 as we want
by restricting the values of x to be sufficiently close to 3.
What values of x guarantee that f ( x ) = 2x − 1 is within:

(a) 1 unit of 5? (b) 0.2 units of 5? (c) E units of 5?

Solution. (a) “Within 1 unit of 5” means between 5 − 1 = 4 and 5 + 1 =

6, so the question can be rephrased as “for what values of x is
y = 2x − 1 between 4 and 6: 4 < 2x − 1 < 6?” We want to know
which values of x ensure the values of y = 2x − 1 are in the the
shaded band in the uppermost margin figure. The algebraic process
is straightforward:

4 < 2x − 1 < 6 ⇒ 5 < 2x < 7 ⇒ 2.5 < x < 3.5

We can restate this result as follows: “If x is within 0.5 units of 3,

then y = 2x − 1 is within 1 unit of 5.” (See second margin figure)
Any smaller distance also satisfies the guarantee: for example, “If x
is within 0.4 units of 3, then y = 2x − 1 is within 1 unit of 5.” (See
third margin figure)
(b) “Within 0.2 units of 5” means between 5 − 0.2 = 4.8 and 5 + 0.2 =
5.2, so the question can be rephrased as “for which values of x is
y = 2x − 1 between 4.8 and 5.2: 4.8 < 2x − 1 < 5.2?” Solving for x,
we get 5.8 < 2x < 6.2 and 2.9 < x < 3.1. “If x is within 0.1 units
of 3, then y = 2x − 1 is within 0.2 units of 5.” (See fourth margin
figure.) Any smaller distance also satisfies the guarantee.
Rather than redoing these calculations for every possible distance
from 5, we can do the work once, generally:
(c) “Within E unit of 5” means between 5 − E and 5 + E, so the question
becomes, “For what values of x is y = 2x − 1 between 5 − E and
5 + E: 5 − E < 2x − 1 < 5 + E?” Solving 5 − E < 2x − 1 < 5 + E
for x, we get:
E E
6 − E < 2x < 6 + E ⇒ 3− < x < 3+
2 2
“If x is within E2 units of 3, then y = 2x − 1 is within E units of 5.”
(See last figure.) Any smaller distance also works. ◀

Part (c) of Example 1 illustrates the power of general solutions

in mathematics. Rather than redoing similar calculations every time
someone demands that f ( x ) = 2x − 1 be within some given distance of
5, we did the calculations once. And then we can quickly respond for
any given distance. For the question “What values of x guarantee that
f ( x ) = 2x − 1 is within 0.4, 0.1 or 0.006 units of 5?” we can answer, “If
x is within 0.2 (= 0.4 0.1 0.006
2 ), 0.05 (= 2 ) or 0.003 (= 2 ) units of 3.”
100 contemporary calculus

Practice 1. Knowing that lim 4x − 5 = 3, determine which values of x

x →2
guarantee that f ( x ) = 4x − 5 is within

(a) 1 unit of 3. (b) 0.08 units of 3. (c) E units of 3.

The same ideas work even if the graphs of the functions are not
straight lines, but the calculations become more complicated.

Example 2. Knowing that lim x2 = 4, determine which values of x

x →2
guarantee that f ( x ) = x2 is within:

(a) 1 unit of 4. (b) 0.2 units of 4.

State each answer in the form: “If x is within units of 2, then

f ( x ) is within units of 4.”

Solution. (a) If x2 is within 1 unit of 4 (and x is near 2, hence positive)

√ √
then 3 < x2 < 5 so 3 < x < 5 or 1.732 < x < 2.236. The interval
√
containing these x values extends from 2 − 3 ≈ 0.268 units to the
√
left of 2 to 5 − 2 ≈ 0.236 units to the right of 2. Because we want
to specify a single distance on each side of 2, we can pick the smaller
of the two distances, 0.236, and say: “If x is within 0.236 units of 2,
then f ( x ) is within 1 unit of 4.”
(b) Similarly, if x2 is within 0.2 units of 4 (and x is near 2, so x > 0)
√ √
then 3.8 < x2 < 4.2 so 3.8 < x < 4.2 or 1.949 < x < 2.049. The
√
interval containing these x values extends from 2 − 3.8 ≈ 0.051
√
units to the left of 2 to 4.2 − 2 ≈ 0.049 units to the right of 2.
Again picking the smaller of the two distances, we can say: “If x is
within 0.049 units of 2, then f ( x ) is within 1 unit of 4.” ◀

The situation in Example 2 — with different distances on the left and

right sides — is very common, and we always pick our single distance to
be the smaller of the distances to the left and right. By using the smaller
distance, we can be certain that if x is within that smaller distance on
either side, then the value of f ( x ) is within the specified distance of the
value of the limit.
√
Practice 2. Knowing that lim x = 3, determine which values of x
√ x → 9
guarantee that f ( x ) = x is within:

(a) 1 unit of 3. (b) 0.2 units of 3.

State each answer in the form: “If x is within units of 9, then

f ( x ) is within units of 3.”

The same ideas can also be used when the function and the specified
distance are given graphically, and in that case we can give the answer
graphically.
 limits and continuity 101

Example 3. In the margin figure, lim f ( x ) = 3. Which values of x

x →2
guarantee that y = f ( x ) is within E units (given graphically) of 3? State
your answer in the form: “If x is within (show a distance D graphically)
of 2, then f ( x ) is within E units of 3.”

Solution. The solution process requires several steps:

(i) Use the given distance E to find the values 3 − E and 3 + E on

the y-axis. (See margin.)
(ii) Sketch the horizontal band with lower edge at y = 3 − E and
upper edge at y = 3 + E.
(iii) Find the first locations to the right and left of x = 2 where the
graph of y = f ( x ) crosses the lines y = 3 − E and y = 3 + E,
and at these locations draw vertical line segments extending to
the x-axis.
(iv) On the x-axis, graphically determine the distance from 2 to the
vertical line on the left (labeled DL ) and from 2 to the vertical
line on the right (labeled DR ).
(v) Let the length D be the smaller of the lengths DL and DR .

If x is within D units of 2, then f ( x ) is within E units of 3. ◀

Practice 3. In the last margin figure, lim f ( x ) = 1.8. Which values of x

x →3
guarantee that y = f ( x ) is within E units (given graphically) of 1.8?

The Formal Definition of Limit

The ideas from the previous Examples and Practice problems, restated
for general functions and limits, provide the basis for the definition of
limit given below. The use of the lowercase Greek letters ϵ (epsilon) and
δ (delta) in the definition is standard, and this definition is sometimes
called the “epsilon-delta” definition of a limit.

Definition of lim f ( x ) = L:
x→a
For every given number ϵ > 0 there is a number δ > 0 so that

if x is within δ units of a (and x ̸= a)

then f (x) is within ϵ units of L

Equivalently: | f ( x ) − L| < ϵ whenever 0 < | x − a| < δ

In this definition, ϵ represents the given distance on either side of

the limiting value y = L, and δ is the distance on each side of the
point x = a on the x-axis that we have been finding in the previous
102 contemporary calculus

examples. This definition has the form of a “challenge and response”:

for any positive challenge ϵ (make f ( x ) within ϵ of L), there is a positive
response δ (start with x within δ of a and x ̸= a).

Example 4. As seen in the second margin figure, lim f ( x ) = L, with

x→a
a value for ϵ given graphically as a length. Find a length for δ that
satisfies the definition of limit (so “if x is within δ of a, and x ̸= a, then
f ( x ) is within ϵ of L”).

Solution. Follow the steps outlined in Example 3. The length for δ

is shown in the third margin figure, and any shorter length for δ also
satisfies the definition. ◀

Practice 4. In the bottom margin figure, lim f ( x ) = L, with a value for

x→a
ϵ given graphically. Find a length for δ that satisfies the definition of
limit.

Example 5. Prove that lim 4x − 5 = 7.

x →3

Solution. We need to show that “for every given ϵ > 0 there is a δ > 0
so that if x is within δ units of 3 (and x ̸= 3) then 4x − 5 is within ϵ
units of 7.”
Actually, there are two things we need to do. First, we need to find a
value for δ (typically depending on ϵ) and, second, we need to show
that our δ really does satisfy the “if. . . then. . . ” part of the definition.
Finding δ is similar to part (c) in Example 1 and Practice 1: Assume
4x − 5 is within ϵ units of 7 and solve for x. If 7 − ϵ < 4x − 5 < 7 + ϵ
then 12 − ϵ < 4x < 12 + ϵ ⇒ 3 − 4ϵ < x < 3 + 4ϵ so x is within 4ϵ units
of 3. Put δ = 4ϵ .
To show that δ = 4ϵ satisfies the definition, we merely reverse the
order of the steps in the previous paragraph. Assume that x is within δ
units of 3. Then 3 − δ < x < 3 + δ, so:
ϵ ϵ
3− < x < 3+ ⇒ 12 − ϵ < 4x < 12 + ϵ
4 4
⇒ 7 − ϵ < 4x − 5 < 7 + ϵ

so we can conclude that f ( x ) = 4x − 5 is within ϵ units of 7. This

formally verifies that lim 4x − 5 = 7. ◀
x →3

Practice 5. Prove that lim 5x + 3 = 23.

x →4

The method used to prove the values of the limits for these particular
linear functions can also be used to prove the following general result
about the limits of linear functions.

Theorem: lim mx + b = ma + b
x→a
 limits and continuity 103

Proof. Let f ( x ) = mx + b.
Case 1: m = 0. Then f ( x ) = 0x + b = b is simply a constant function,
and any value for δ > 0 satisfies the definition. Given any value of
ϵ > 0, let δ = 1 (any positive value for δ works). If x is is within 1 unit
of a, then f ( x ) − f ( a) = b − b = 0 < ϵ, so we have shown that for any
ϵ > 0 there is a δ > 0 that satisfies the limit definition.
Case 2: m ̸= 0. For any ϵ > 0, put δ = |mϵ | > 0. If x is within δ = |mϵ |
of a then
ϵ ϵ ϵ ϵ ϵ
a− < x < a+ ⇒ < x−a < ⇒ | x − a| <
|m| |m| |m| |m| |m|
Then the distance between f ( x ) and L = ma + b is:

| f ( x ) − L| = |(mx + b) − (ma + b)| = |mx − ma|

ϵ
= |m| · | x − a| < |m| =ϵ
|m|
so f ( x ) is within ϵ of L = ma + b.
In each case, we have shown that “given any ϵ > 0, there is a δ > 0”
that satisfies the rest of the limit definition.

If there is even a single value of ϵ for which there is no δ, then we

say that the limit “does not exist.”

Example 6. With f ( x ) defined as:

(
2 if x < 1
f (x) =
4 if x > 1

use the limit definition to prove that lim f ( x ) does not exist.
x →1

Solution. One common proof technique in mathematics is called “proof

by contradiction” and that is the method we use here:

• We assume that the limit does exist and equals some number L.
• We show that this assumption leads to a contradiction
• We conclude that the assumption must have been false.

We therefore conclude that the limit does not exist.

First, assume that the limit exists: lim f ( x ) = L for some value for L. The definition says “for every ϵ” so we
x →1
1 can certainly pick 12 as our ϵ value; why
Let ϵ = 2. Then, because we are assuming that the limit exists, there is
we chose this particular value for ϵ shows
a δ > 0 so that if x is within δ of 1 then f ( x ) is within ϵ of L. up later in the proof.
Next, let x1 be between 1 and 1 + δ. Then x1 > 1 so f ( x1 ) = 4. Also,
x1 is within δ of 1 so f ( x1 ) = 4 is within 12 of L, which means that L is
between 3.5 and 4.5: 3.5 < L < 4.5.
Let x2 be between 1 and 1 − δ. Then x2 < 1, so f ( x2 ) = 2. Also, x2
is within δ of 1 so f ( x2 ) = 2 is within 12 of L, which means that L is
between 1.5 and 2.5: 1.5 < L < 2.5.
104 contemporary calculus

These inequalities provide the contradiction we hoped to find. There

is no value L that satisfies both 3.5 < L < 4.5 and 1.5 < L < 2.5, so our
assumption must be false: f ( x ) does not have a limit as x → 1. ◀
1
Practice 6. Use the limit definition to prove that lim does not exist.
x →0 x

Proofs of Two Limit Theorems

We conclude with proofs of two parts of the Main Limit Theorem so
you can see how such proofs proceed — you have already used these
theorems to evaluate limits.

Theorem:
There are rigorous proofs of all of the
other limit properties in the Main Limit If lim f ( x ) = L
x→a
Theorem, but they are somewhat more then lim k · f ( x ) = kL
complicated than the proofs given here. x→a

Proof. Case k = 0: The theorem is true but not very interesting:

lim k · f ( x ) = lim 0 · f ( x ) = lim 0 = 0 = 0 · L = kL
x→a x→a x→a

Case k ̸= 0: Because lim f ( x ) = L, then, by the definition, for every

x→a
ϵ > 0 there is a δ > 0 so that | f ( x ) − L| < ϵ whenever | x − a| < δ.
For any ϵ > 0, we know |ϵk| > 0, so pick a value of δ that satisfies
| f ( x ) − L| < |ϵk| whenever | x − a| < δ.
When | x − a| < δ (“x is within δ of a”) then | f ( x ) − L| < |ϵk| (“ f ( x )
is within |ϵk| of L”) so |k| · | f ( x ) − L| < ϵ ⇒ |k · f ( x ) − k · L| < ϵ
(that is, k · f ( x ) is within ϵ of k · L).

Theorem:
If lim f ( x ) = L and lim g( x ) = M
x→a x→a
then lim [ f ( x ) + g( x )] = L + M.
x→a

Proof. Given any ϵ > 0, we know 2ϵ > 0, so there is a number δ f > 0

such that when | x − a| < δ f then | f ( x ) − L| < 2ϵ (“if x is within δ f of a,
then f ( x ) is within 2ϵ of L”).
Likewise, there is a number δg > 0 such that when | x − a| < δg then
| g( x ) − M| < 2ϵ (“if x is within δg of a, then g( x ) is within 2ϵ of M”).
Let δ be the smaller of δ f and δg . If | x − a| < δ then | f ( x ) − L| < 2ϵ
and | g( x ) − M | < 2ϵ so:
|( f ( x ) + g( x )) − ( L + M))| = |( f ( x ) − L) + ( g( x ) − M)|
Here we use the “triangle inequality”:
ϵ ϵ
| a + b| ≤ | a| + |b|
≤ | f ( x ) − L| + | g( x ) − M| <
+ =ϵ
2 2
so f ( x ) + g( x ) is within ϵ of L + M whenever x is within δ of a.
 limits and continuity 105

1.4 Problems

In Problems 1–4, state each answer in the form “If x 8. Knowing that lim x2 = 9, what values of x guar-
x →3
is within units of. . . ” antee that f ( x ) = x2 is within:
1. Knowing that lim 2x + 1 = 7, what values of x
x →3 (a) 1 unit of 9? (b) 0.2 units of 9?
guarantee that f ( x ) = 2x + 1 is within:
9. Knowing that lim x3 = 8, what values of x guar-
(a) 1 unit of 7? (b) 0.6 units of 7? x →2
antee that f ( x ) = x3 is within:
(c) 0.04 units of 7? (d) ϵ units of 7?
(a) 0.5 units of 8? (b) 0.05 units of 8?
2. Knowing that lim 3x + 2 = 5, what values of x √
x →1 10. Knowing that lim x = 4, what values of x
x →16√
guarantee that f ( x ) = 3x + 2 is within:
guarantee that f ( x ) = x is within:
(a) 1 unit of 5? (b) 0.6 units of 5? (a) 1 unit of 4? (b) 0.1 units of 4?
√
(c) 0.09 units of 5? (d) ϵ units of 5? 11. Knowing that lim 1 + x = 2, what values of x
x →3 √
3. Knowing that lim 4x − 3 = 5, what values of x guarantee that f ( x ) = 1 + x is within:
x →2
guarantee that f ( x ) = 4x − 3 is within: (a) 1 unit of 2? (b) 0.0002 units of 2?

(a) 1 unit of 5? (b) 0.4 units of 5? 12. You must cut four pieces of wire (all the same
length) and form them into a square. If the area
(c) 0.08 units of 5? (d) ϵ units of 5?
of the square must be within 0.06 in2 of 100 in2 ,
then each piece of wire must be within how many
4. Knowing that lim 5x − 3 = 2, what values of x
x →1 inches of 10 in?
guarantee that f ( x ) = 5x − 3 is within:
13. You need to cut four pieces of wire (all the same
(a) 1 unit of 2? (b) 0.5 units of 2? length) and form them into a square. If the area
of the square must be within 0.06 in2 of 25 in2 ,
(c) 0.01 units of 2? (d) ϵ units of 2? then each piece of wire must be within how many
inches of 5 in?
5. For Problems 1–4, list the slope of each function
Problems 14–17 give lim f ( x ) = L, the function f
f and the δ (as a function of ϵ). For these linear x→a
and a value for ϵ graphically. Find a length for δ that
functions f , how is δ related to the slope?
satisfies the limit definition for the given function
6. You have been asked to cut two boards (exactly
and value of ϵ.
the same length after the cut) and place them end
to end. If the combined length must be within
0.06 inches of 30 inches, then each board must be
within how many inches of 15?
7. You have been asked to cut three boards (exactly
14.
the same length after the cut) and place them end
to end. If the combined length must be within
0.06 inches of 30 inches, then each board must be
within how many inches of 10?
 106 contemporary calculus

18. Redo each of Problems 14–17 taking a new value

of ϵ to be half the value of ϵ given in the problem.

In Problems 19–22, use the limit definition to prove

15. that the given limit does not exist. (Find a value
for ϵ > 0 for which there is no δ that satisfies the
definition.)

19. With f ( x ) defined as:

(
4 if x < 2
f (x) =
3 if x > 2

show that lim f ( x ) does not exist.

x →2
20. Show that lim ⌊ x ⌋ does not exist.
x →3
16.
21. With f ( x ) defined as:
(
x if x < 2
f (x) =
6−x if x > 2

show that lim f ( x ) does not exist.

x →2
22. With f ( x ) defined as:
(
x+1 if x < 2
f (x) =
x2 if x > 2
17.
show that lim f ( x ) does not exist.
x →2
23. Prove: If lim f ( x ) = L and lim g( x ) = M then
x→a x→a
lim [ f ( x ) − g( x )] = L − M.
x→a

1.4 Practice Answers

1. (a) 3 − 1 < 4x − 5 < 3 + 1 ⇒ 7 < 4x < 9 ⇒ 1.75 < x < 2.25:
“x within 14 unit of 2.”
(b) 3 − 0.08 < 4x − 5 < 3 + 0.08 ⇒ 7.92 < 4x < 8.08 ⇒
1.98 < x < 2.02: “x within 0.02 units of 2.”
(c) 3 − E < 4x − 5 < 3 + E ⇒ 8 − E < 4x < 8 + E ⇒
E
2− 4 < x < 2 + E4 : “x within E
4 units of 2.”
 limits and continuity 107

√
2. “Within 1 unit of 3”: If 2 < x < 4, then 4 < x < 16, which extends
from 5 units to the left of 9 to 7 units to right of 9. Using the smaller
√
of these two distances from 9: “If x is within 5 units of 9, then x is
within 1 unit of 3.”
√
“Within 0.2 units of 3”: If 2.8 < x < 3.2, then 7.84 < x < 10.24,
which extends from 1.16 units to the left of 9 to 1.24 units to the right
of 9. “If x is within 1.16 units of 9, then x is within 0.2 units of 3.

3. 4.

ϵ ϵ
5. Given any ϵ > 0, take δ = 5. If x is within δ = 5 of 4, then
4 − 5ϵ < x < 4 + 5ϵ so:
ϵ ϵ
− < x − 4 < ⇒ −ϵ < 5x − 20 < ϵ ⇒ −ϵ < (5x + 3) − 23 < ϵ
5 5
so, finally, f ( x ) = 5x + 3 is within ϵ of L = 23.
We have shown that “for any ϵ > 0, there is a δ > 0 (namely δ = 5ϵ )”
so that the rest of the definition is satisfied.
This is a much more sophisticated
6. Using “proof by contradiction” as in the solution to Example 6: (= harder) problem.

1
• Assume that the limit exists: lim = L for some value of L. Let
x →0 x
ϵ = 1. Since we’re assuming that the limit exists, there is a δ > 0 The definition says “for every ϵ” so we
can pick ϵ = 1. For this particular limit,
so that if x is within δ of 0 then f ( x ) = 1x is within ϵ = 1 of L.
the definition fails for every ϵ > 0.
• Let x1 be between 0 and 0 + δ and also require that x1 < 12 . Then
0 < x1 < 12 so f ( x1 ) = x1 > 2. Because x1 is within δ of 0,
1
f ( x1 ) > 2 is within ϵ = 1 of L, so L > 2 − ϵ = 1: that is, 1 < L.
Let x2 be between 0 and 0 − δ and also require x2 > − 12 . Then
0 > x2 > 21 so f ( x2 ) = x12 < −2. Since x2 is within δ of 0,
f ( x2 ) < −2 is within ϵ = 1 of L, so L < −2 + ϵ = −1 ⇒ −1 > L.
• The two inequalities derived above provide the contradiction we
were hoping to find. There is no value L that satisfies both 1 < L
and L < −1, so we can conclude that our assumption was false
and that f ( x ) = 1x does not have a limit as x → 0.
2
The Derivative

The two previous chapters have laid the foundation for the study of
calculus. They provided a review of some material you will need and
started to emphasize the various ways we will view and use functions:
functions given by graphs, equations and tables of values.
Chapter 2 will focus on the idea of tangent lines. We will develop a
definition for the derivative of a function and calculate derivatives of
some functions using this definition. Then we will examine some of the
properties of derivatives, see some relatively easy ways to calculate the
derivatives, and begin to look at some ways we can use them.

2.0 Introduction to Derivatives

This section begins with a very graphical approach to slopes of tangent

lines. It then examines the problem of finding the slopes of the tangent
lines for a single function, y = x2 , in some detail — and illustrates how
these slopes can help us solve fairly sophisticated problems.

Slopes of Tangent Lines: Graphically

The figure in the margin shows the graph of a function y = f ( x ). We
can use the information in the graph to fill in the table:

x y = f (x) m( x )
0 0 1
1 1 0
2 0 −1
3 −1 0
4 1 1
1
5 2 2

where m( x ) is the (estimated) slope of the line tangent to the graph

of y = f ( x ) at the point ( x, y). We can estimate the values of m( x ) at
some non-integer values of x as well: m(0.5) ≈ 0.5 and m(1.3) ≈ −0.3,
110 contemporary calculus

for example. We can even say something about the behavior of m( x )

over entire intervals: if 0 < x < 1, then m( x ) is positive, for example.
The values of m( x ) definitely depend on the values of x (the slope
varies as x varies, and there is at most one slope associated with each
value of x) so m( x ) is a function of x. We can use the results in the table
to help sketch a graph of the function m( x ) (see top margin figure).

Practice 1. A graph of y = f ( x ) appears in the margin. Set up a table

of (estimated) values for x and m( x ), the slope of the line tangent to
the graph of y = f ( x ) at the point ( x, y), and then sketch a graph of
the function m( x ).

In some applications, we need to know where the graph of a function

f ( x ) has horizontal tangent lines (that is, where the slope of the tangent
line equals 0). The slopes of the lines tangent to graph of y = f ( x ) in
Practice 1 are 0 when x = 2 or x ≈ 4.25.

Practice 2. At what values of x does the graph of y = g( x ) (in the

margin) have horizontal tangent lines?

Example 1. The graph of the height of a rocket at time t appears in

the margin. Sketch a graph of the velocity of the rocket at time t.
(Remember that instantaneous velocity corresponds to the slope of the
line tangent to the graph of position or height function.)

Solution. The penultimate margin figure shows some sample tangent

line segments, while the bottom margin figure shows the velocity of
the rocket. (What so you think happened at time t = 8?) ◀

Practice 3. The graph below shows the temperature during a summer

day in Chicago. Sketch a graph of the rate at which the temperature is
changing at each moment in time. (As with instantaneous velocity, the
instantaneous rate of change for the temperature corresponds to the
slope of the line tangent to the temperature graph.)

The function m( x ), the slope of the line tangent to the graph of

y = f ( x ) at ( x, f ( x )), is called the derivative of f ( x ).
 the derivative 111

We used the idea of the slope of the tangent line all throughout
Chapter 1. In Section 2.1, we will formally define the derivative of a
function and begin to examine some of its properties, but first let’s see
what we can do when we have a formula for f ( x ).

Tangents to y = x2
When we have a formula for a function, we can determine the slope
of the tangent line at a point ( x, f ( x )) by calculating the slope of the
secant line through the points ( x, f ( x )) and ( x + h, f ( x + h)):

f ( x + h) − f ( x )
msec =
( x + h) − ( x )

and then taking the limit of msec as h approaches 0:

f ( x + h) − f ( x )
mtan = lim msec = lim
h →0 h →0 ( x + h) − ( x )

Example 2. Find the slope of the line tangent to the graph of the
function y = f ( x ) = x2 at the point (2, 4).

Solution. In this example, x = 2, so x + h = 2 + h and f ( x + h) =

f (2 + h) = (2 + h)2 . The slope of the tangent line at (2, 4) is

f (2 + h ) − f (2)
mtan = lim msec = lim
h →0 h →0 (2 + h ) − (2)
(2 + h) − 22
2 4 + 4h + h2 − 4
= lim = lim
h →0 h h →0 h
4h + h 2
= lim = lim [4 + h] = 4
h →0 h h →0

The line tangent to y = x2 at the point (2, 4) has slope 4. ◀

We can use the point-slope formula for a line to find an equation of

this tangent line:

y − y0 = m( x − x0 ) ⇒ y − 4 = 4( x − 2) ⇒ y = 4x − 4

Practice 4. Use the method of Example 2 to show that the slope of

the line tangent to the graph of y = f ( x ) = x2 at the point (1, 1) is
mtan = 2. Also find the values of mtan at (0, 0) and (−1, 1).

It is possible to compute the slopes of the tangent lines one point

at a time, as we have been doing, but that is not very efficient. You
should have noticed in Practice 4 that the algebra for each point was
very similar, so let’s do all the work just once, for an arbitrary point
( x, f ( x )) = ( x, x2 ) and then use the general result to find the slopes at
the particular points we’re interested in.
112 contemporary calculus

The slope of the line tangent to the graph of y = f ( x ) = x2 at the

arbitrary point ( x, x2 ) is:

f ( x + h) − f ( x )
mtan = lim msec = lim
h →0 h →0 ( x + h) − ( x )
(x + h) − x2
2 x2 + 2xh + h2 − x2
= lim = lim
h →0 h h →0 h
2xh + h 2
= lim = lim [2x + h] = 2x
h →0 h h →0

The slope of the line tangent to the graph of y = f ( x ) = x2 at the point

( x, x2 ) is mtan = 2x. We can use this general result at any value of x
without going through all of the calculations again. The slope of the
line tangent to y = f ( x ) = x2 at the point (4, 16) is mtan = 2(4) = 8
and the slope at ( p, p2 ) is mtan = 2( p) = 2p. The value of x determines
the location of our point on the curve, ( x, x2 ), as well as the slope of the
line tangent to the curve at that point, mtan = 2x. The slope mtan = 2x
is a function of x and is called the derivative of y = x2 .
Simply knowing that the slope of the line tangent to the graph of
y = x2 is mtan = 2x at a point ( x, y) can help us quickly find an
equation of the line tangent to the graph of y = x2 at any point and
answer a number of difficult-sounding questions.

Example 3. Find equations of the lines tangent to y = x2 at the points

(3, 9) and ( p, p2 ).

Solution. At (3, 9), the slope of the tangent line is 2x = 2(3) = 6, and
the equation of the line is y − 9 = 6( x − 3) ⇒ y = 6x − 9.
At ( p, p2 ), the slope of the tangent line is 2x = 2( p) = 2p, and the
equation of the line is y − p2 = 2p( x − p) ⇒ y = 2px − p2 . ◀

Example 4. A rocket has been programmed to follow the path y = x2 in

space (from left to right along the curve, as seen in the margin figure),
but an emergency has arisen and the crew must return to their base,
which is located at coordinates (3, 5). At what point on the path y = x2
should the captain turn off the engines so that the ship will coast along
a path tangent to the curve to return to the base?

Solution. You might spend a few minutes trying to solve this problem
without using the relation mtan = 2x, but the problem is much easier
if we do use that result.
Let’s assume that the captain turns off the engine at the point ( p, q)
on the curve y = x2 and then try to determine what values p and q
must have so that the resulting tangent line to the curve will go through
the point (3, 5). The point ( p, q) is on the curve y = x2 , so q = p2 and
the equation of the tangent line, found in Example 3, must then be
y = 2px − p2 .
 the derivative 113

To find the value of p so that the tangent line will go through the
point (3, 5), we can substitute the values x = 3 and y = 5 into the
equation of the tangent line and solve for p:

y = 2px − p2 ⇒ 5 = 2p(3) − p2 ⇒ p2 − 6p + 5 = 0
⇒ ( p − 1)( p − 5) = 0

The only solutions are p = 1 and p = 5, so the only possible points

are (1, 1) and (5, 25). You can verify that the tangent lines to y = x2 at
(1, 1) and (5, 25) both go through the point (3, 5). Because the ship is
moving from left to right along the curve, the captain should turn off
the engines at the point (1, 1). (Why not at (5, 25)?) ◀

Practice 5. Verify that if the rocket engines in Example 4 are shut off at
(2, 4), then the rocket will go through the point (3, 8).

2.0 Problems

1. Use the function f ( x ) graphed below to fill in the 2. Use the function g( x ) graphed below to fill in the
table and then graph m( x ), the estimated slope table and then graph m( x ), the estimated slope
of the tangent line to y = f ( x ) at the point ( x, y). of the tangent line to y = g( x ) at the point ( x, y).

x f (x) m( x ) x f (x) m( x ) x g( x ) m( x ) x g( x ) m( x )
0.0 2.5 0.0 2.5
0.5 3.0 0.5 3.0
1.0 3.5 1.0 3.5
1.5 4.0 1.5 4.0
2.0 2.0
114 contemporary calculus

3. (a) At what values of x does the graph of f 6. Match the situation descriptions with the corre-
(shown below) have a horizontal tangent line? sponding time-velocity graphs shown below.
(b) At what value(s) of x is the value of f the (a) A car quickly leaving from a stop sign.
largest? Smallest? (b) A car sedately leaving from a stop sign.
(c) Sketch a graph of m( x ), the slope of the line (c) A student bouncing on a trampoline.
tangent to the graph of f at the point ( x, f ( x )). (d) A ball thrown straight up.
(e) A student confidently striding across campus
to take a calculus test.
(f) An unprepared student walking across cam-
pus to take a calculus test.

4. (a) At what values of x does the graph of g

(shown below) have a horizontal tangent line?
(b) At what value(s) of x is the value of g the
largest? Smallest?
Problems 7–10 assume that a rocket is following the
(c) Sketch a graph of m( x ), the slope of the line path y = x2 , from left to right.
tangent to the graph of g at the point ( x, g( x )).
7. At what point should the engine be turned off in
order to coast along the tangent line to a base at
(5, 16)?

8. At (3, −7)? 9. At (1, 3)?

10. Which points in the plane can not be reached by

the rocket? Why not?
In Problems 11–16, perform these steps:
(a) Calculate and simplify:
f ( x + h) − f ( x )
msec =
5. (a) Sketch the graph of f ( x ) = sin( x ) on the in- ( x + h) − ( x )
terval −3 ≤ x ≤ 10.
(b) Determine mtan = lim msec .
h →0
(b) Sketch a graph of m( x ), the slope of the line
(c) Evaluate mtan at x = 2.
tangent to the graph of sin( x ) at the point
(d) Find an equation of the line tangent to the graph
( x, sin( x )).
of f at (2, f (2)).
(c) Your graph in part (b) should look familiar.
What function is it? 11. f ( x ) = 3x − 7 12. f ( x ) = 2 − 7x
 the derivative 115

13. f ( x ) = ax + b where a and b are constants 17. Given f ( x ) = x2 + 2x, at which point(s) ( p, f ( p))
does the line tangent to the graph at that point
14. f ( x ) = x2 + 3x 15. f ( x ) = 8 − 3x2 also go through the point (3, 6)?

16. f ( x ) = ax2 + bx + c where a, b and c are constants 18. (a) If a ̸= 0, then what is the shape of the graph
of y = f ( x ) = ax2 + bx + c?
In Problems 17–18, use the result:
(b) At what value(s) of x is the line tangent to the
f ( x ) = ax2 + bx + c ⇒ mtan = 2ax + b graph of f ( x ) horizontal?

2.0 Practice Answers

1. Approximate values of m( x ) appear in the table in the margin; the
margin figure shows a graph of m( x ). x f (x) m( x )
0 2 −1
2. The tangent lines to the graph of g are horizontal (slope = 0) when 1 1 −1
x ≈ −1, 1, 2.5 and 5. 2 1
3 0
3 1 1
3 1
3. The figure below shows a graph of the approximate rate of tempera- 4 2 2
5 1 −2
ture change (slope).

4. At (1, 1), the slope of the tangent line is:

f (1 + h ) − f (1)
mtan = lim msec = lim
h →0 h →0 (1 + h ) − (1)
( 1 + h ) − 12
2 1 + 2h + h2 − 1
= lim = lim
h →0 h h →0 h
2h + h2
= lim = lim [2 + h] = 2
h →0 h h →0
116 contemporary calculus

so the line tangent to y = x2 at the point (1, 1) has slope 2. At (0, 0):

f (0 + h ) − f (1)
mtan = lim msec = lim
h →0 h →0 (0 + h ) − (0)
( 0 + h ) − 02
2 h2
= lim = lim = lim h = 0
h →0 h h →0 h h →0

so the line tangent to y = x2 at (0, 0) has slope 0. At (−1, 1):

f (−1 + h) − f (−1)
mtan = lim msec = lim
h →0 h→0 (−1 + h ) − (−1)
(−1 + h)2 − (−1)2 1 − 2h + h2 −1
= lim = lim
h →0 h h →0 h
−2h + h2
= lim = lim [−2 + h] = −2
h →0 h h →0

so the line tangent to y = x2 at the point (−1, 1) has slope −2.

5. From Example 4 we know the slope of the tangent line is mtan = 2x,
so the slope of the tangent line at (2, 4) is mtan = 2x = 2(2) =
4. The tangent line has slope 4 and goes through the point (2, 4),
so an equation of the tangent line (using y − y0 = m( x − x0 )) is
y − 4 = 4( x − 2) or y = 4x − 4. The point (3, 8) satisfies the equation
y = 4x − 4, so the point (3, 8) lies on the tangent line.
 the derivative 117

2.1 The Definition of Derivative

The graphical idea of a slope of a tangent line is very useful, but for
some purposes we need a more algebraic definition of the derivative
of a function. We will use this definition to calculate the derivatives
of several functions and see that these results agree with our graphical
understanding. We will also look at several different interpretations for
the derivative, and obtain a theorem that will allow us to easily and
quickly determine the derivative of any fixed power of x.
In the previous section we found the slope of the tangent line to
the graph of the function f ( x ) = x2 at an arbitrary point ( x, f ( x )) by
calculating the slope of the secant line through the points ( x, f ( x )) and
( x + h, f ( x + h)):
f ( x + h) − f ( x )
msec =
( x + h) − ( x )
and then taking the limit of msec as h approached 0 (see margin). That
approach to calculating slopes of tangent lines motivates the definition
of the derivative of a function.

Definition of the Derivative:

The derivative of a function f is a new function,
f ′ (pronounced “eff prime”), whose value at x is:

f ( x + h) − f ( x )
f ′ ( x ) = lim
h →0 h
if this limit exists and is finite.

This is the definition of differential calculus, and you must know

it and understand what it says. The rest of this chapter and all of
Chapter 3 are built on this definition, as is much of what appears in
later chapters. It is remarkable that such a simple idea (the slope of a
tangent line) and such a simple definition (for the derivative f ′ ) will
lead to so many important ideas and applications.

Notation
There are three commonly used notations for the derivative of y = f ( x ):

• f ′ ( x ) emphasizes that the derivative is a function related to f

• D( f ) emphasizes that we perform an operation on f to get f ′

df ∆f f ( x + h) − f ( x )
• emphasizes that the derivative is the limit of =
dx ∆x h
We will use all three notations so that you can become accustomed to
working with each of them.
118 contemporary calculus

The function f ′ ( x ) gives the slope of the tangent line to the graph of
y = f ( x ) at the point ( x, f ( x )), or the instantaneous rate of change of
the function f at the point ( x, f ( x )).
If, in the margin figure, we let x be the point a + h, then h = x − a.
As h → 0, we see that x → a and:
f ( a + h) − f ( a) f ( x ) − f ( a)
f ′ ( a) = lim = lim
h →0 h x→a x−a
We will use whichever of these two forms is more convenient alge-
braically in a particular situation.

Calculating Some Derivatives Using the Definition

Fortunately, we will soon have some quick and easy ways to calculate
most derivatives, but first we will need to use the definition to deter-
mine the derivatives of a few basic functions. In Section 2.2, we will use
those results and some properties of derivatives to calculate derivatives
of combinations of the basic functions. Let’s begin by using the graphs
and then the definition to find a few derivatives.

Example 1. Graph y = f ( x ) = 5 and estimate the slope of the tangent

line at each point on the graph. Then use the definition of the derivative
to calculate the exact slope of the tangent line at each point. Your
graphical estimate and the exact result from the definition should
agree.

Solution. The graph of y = f ( x ) = 5 is a horizontal line (see margin),

which has slope 0, so we should expect that its tangent line will also
have slope 0.
Using the definition: With f ( x ) = 5, then f ( x + h) = 5 no matter
what h is, so:
f ( x + h) − f ( x ) 5−5
D( f ( x )) = lim = lim = lim 0 = 0
h →0 h h →0 h h →0

and this agrees with our graphical estimate of the derivative. ◀

Using similar steps, it is easy to show that the derivative of any

constant function is 0.

Theorem: If f ( x ) = k, then f ′ ( x ) = 0.

Practice 1. Graph y = f ( x ) = 7x and estimate the slope of the tangent

line at each point on the graph. Then use the definition of the derivative
to calculate the exact slope of the tangent line at each point.

Example 2. Describe the derivative of y = f ( x ) = 5x3 graphically and

compute it using the definition. Find an equation of the line tangent to
y = 5x3 at the point (1, 5).
 the derivative 119

Solution. It appears from the graph of y = f ( x ) = 5x3 (see margin)

that f ( x ) is increasing, so the slopes of the tangent lines are positive
except perhaps at x = 0, where the graph seems to flatten out.
With f ( x ) = 5x3 we have:

f ( x + h) = 5( x + h)3 = 5( x3 + 3x2 h + 3xh2 + h3 )

and using this last expression in the definition of the derivative:

f ( x + h) − f ( x ) 5( x3 + 3x2 h + 3xh2 + h3 ) − 5x3

f ′ ( x ) = lim = lim
h →0 h h →0 h
15x2 h + 15xh2 + 5h3
= lim = lim (15x2 + 15xh + 5h2 ) = 15x2
h →0 h h →0

so D(5x3 ) = 15x2 , which is positive except when x = 0 (as we predicted

from the graph).
The function f ′ ( x ) = 15x2 gives the slope of the line tangent to
the graph of f ( x ) = 5x3 at the point ( x, f ( x )). At the point (1, 5),
the slope of the tangent line is f ′ (1) = 15(1)2 = 15. From the point-
slope formula, an equation of the tangent line to f at that point is
y − 5 = 15( x − 1) or y = 15x − 10. ◀

Practice 2. Use the definition to show that the derivative of y = x3 is

dy
= 3x2 . Find an equation of the line tangent to the graph of y = x3
dx
at the point (2, 8).

If f has a derivative at x, we say that f is differentiable at x. If we

have a point on the graph of a differentiable function and a slope (the
derivative evaluated at the point), it is easy to write an equation of the
tangent line.

Tangent Line Formula:

If f ( x ) is differentiable at x = a
then an equation of the line tangent to f at ( a, f ( a)) is:

y = f ( a) + f ′ ( a)( x − a)

Proof. The tangent line goes through the point ( a, f ( a)) with slope
f ′ ( a) so, using the point-slope formula, y − f ( a) = f ′ ( a)( x − a) or
y = f ( a) + f ′ ( a)( x − a).

Practice 3. The derivatives D( x ) = 1, D( x2 ) = 2x, D( x3 ) = 3x2 exhibit

the start of a pattern. Without using the definition of the derivative,
what do you think the following derivatives will be? D( x4 ), D( x5 ),
√ 1
D( x43 ), D( x ) = D( x 2 ) and D( x π ). (Just make an intelligent “guess”
based on the pattern of the previous examples. )
120 contemporary calculus

Before further investigating the “pattern” for the derivatives of pow-

ers of x and general properties of derivatives, let’s compute the deriva-
tives of two functions that are not powers of x: sin( x ) and | x |.

Theorem: D(sin( x )) = cos( x )

The graph of y = f ( x ) = sin( x ) (see margin) should be very familiar

to you. The graph has horizontal tangent lines (slope = 0) when
x = ± π2 and x = ± 3π π
2 and so on. If 0 < x < 2 , then the slopes of
the tangent lines to the graph of y = sin( x ) are positive. Similarly,
if π2 < x < 3π 2 , then the slopes of the tangent lines are negative.
Finally, because the graph of y = sin( x ) is periodic, we expect that the
derivative of y = sin( x ) will also be periodic. Note that the function
cos( x ) possesses all of those desired properties for the slope function.

Proof. With f ( x ) = sin( x ), apply an angle addition formula to get:

f ( x + h) = sin( x + h) = sin( x ) cos(h) + cos( x ) sin(h)

and use this formula in the definition of the derivative:

f ( x + h) − f ( x )
f ′ ( x ) = lim
h →0 h
(sin( x ) cos(h) + cos( x ) sin(h)) − sin( x )
= lim
h →0 h
This limit looks formidable, but just collect the terms containing sin( x ):

(sin( x ) cos(h) − sin( x )) + cos( x ) sin(h)

lim
h →0 h
so you can factor out sin( x ) from the first two terms, rewriting as:

cos(h) − 1
 
sin(h)
lim sin( x ) · + cos( x ) ·
h →0 h h

Now calculate the limits separately:

cos(h) − 1 sin(h)
lim sin( x ) · lim + lim cos( x ) · lim
h →0 h →0 h h →0 h →0 h
The first and third limits do not depend on h, and we calculated the
second and fourth limits in Section 1.2:

sin( x ) · 0 + cos( x ) · 1 = cos( x )

So D(sin( x )) = cos( x ) and the various properties we expected of the

derivative of y = sin( x ) by examining its graph are true of cos( x ).
You will need the angle addition formula
for cosine to rewrite cos( x + h) as:
Practice 4. Show that D(cos( x )) = − sin( x ) using the definition.
cos( x ) · cos(h) − sin( x ) · sin(h)
 the derivative 121

The derivative of cos( x ) resembles the situation for sin( x ) but differs
by an important negative sign. You should memorize both of these
important derivatives.
dy
Example 3. For y = | x |, find .
dx
Solution. The graph of y = f ( x ) = | x | (see margin) is a “V” shape
with its vertex at the origin. When x > 0, the graph is just y = | x | = x,
which is part of a line with slope +1, so we should expect the derivative
of | x | to be +1. When x < 0, the graph is y = | x | = − x, which is part of
a line with slope −1, so we expect the derivative of | x | to be −1. When
x = 0, the graph has a corner, and we should expect the derivative of
| x | to be undefined at x = 0, as there is no single candidate for a line
tangent to the graph there.
Using the definition, consider the same three cases discussed previ-
ously: x > 0, x < 0 and x = 0.
If x > 0, then, for small values of h, x + h > 0, so:
| x + h| − | x | x+h−x h
D( f ( x )) = lim = lim = lim = 1
h →0 h h →0 h h →0 h

If x < 0, then, for small values of h, x + h < 0, so:

| x + h| − | x | −( x + h) − (− x ) −h
D( f ( x )) = lim = lim = lim = −1
h →0 h h →0 h h →0 h

When x = 0, the situation is a bit more complicated:

|0 + h | − |0| |h|
f ′ (0) = lim = lim
h →0 h h →0 h

|h| |h|
This is undefined, as lim = +1 and lim = −1, so:
h →0+ h h →0− h


 1 if x > 0
D(| x |) = undefined if x = 0 The derivative of | x | agrees with the func-
tion sgn( x ) defined in Chapter 0, except

 −1 if x < 0
at x = 0: D(| x |) is undefined at x = 0
but sgn(0) = 0.
|x|
or, equivalently, D(| x |) = . ◀
x

Practice 5. Graph y = | x − 2| and y = |2x | and use the graphs to

determine D(| x − 2|) and D(|2x |).

So far we have emphasized the derivative as the slope of the line

tangent to a graph. That very visual interpretation is very useful when
examining the graph of a function, and we will continue to use it.
Derivatives, however, are employed in a wide variety of fields and
applications, and some of these fields use other interpretations. A few
commonly used interpretations of the derivative follow.
122 contemporary calculus

Interpretations of the Derivative

General
Rate of Change The function f ′ ( x ) is the rate of change of the function
at x. If the units for x are years and the units for f ( x ) are people, then
df people
the units for dx are year , a rate of change in population.

Graphical
Slope f ′ ( x ) is the slope of the line tangent to the graph of f at ( x, f ( x )).

Physical
Velocity If f ( x ) is the position of an object at time x, then f ′ ( x ) is the
velocity of the object at time x. If the units for x are hours and f ( x ) is
distance, measured in miles, then the units for f ′ ( x ) = dx are miles ,
df
hour
miles per hour, which is a measure of velocity.

Acceleration If f ( x ) is the velocity of an object at time x, then f ′ ( x ) is

the acceleration of the object at time x. If the units for x are hours and
f ( x ) has the units miles , then the units for the acceleration f ′ ( x ) = dx
df
hour
are miles/hour = miles2 , “miles per hour per hour.”
hour hour

Magnification f ′ ( x ) is the magnification factor of the function f for

points close to x. If a and b are two points very close to x, then the
distance between f ( a) and f (b) will be close to f ′ ( x ) times the original
distance between a and b: f (b) − f ( a) ≈ f ′ ( x )(b − a).

Business
Marginal Cost If f ( x ) is the total cost of producing x objects, then f ′ ( x )
is the marginal cost, at a production level of x: (approximately) the
additional cost of making one more object once we have already made x
objects. If the units for x are bicycles and the units for f ( x ) are dollars,
then the units for f ′ ( x ) = dx are dollars , the cost per bicycle.
df
bicycle

Marginal Profit If f ( x ) is the total profit from producing and selling x

objects, then f ′ ( x ) is the marginal profit: the profit to be made from
producing and selling one more object. If the units for x are bicycles
df
and the units for f ( x ) are dollars, then the units for f ′ ( x ) = dx are
dollars , the profit per bicycle.
bicycle

In financial contexts, the word “marginal” usually refers to the

derivative or rate of change of some quantity. One of the strengths of
calculus is that it provides a unity and economy of ideas among diverse
applications. The vocabulary and problems may be different, but the
ideas and even the notations of calculus remain useful.
 the derivative 123

Example 4. A small cork is bobbing up and down, and at time t seconds

it is h(t) = sin(t) feet above the mean water level (see margin). Find
the height, velocity and acceleration of the cork when t = 2 seconds.
(Include the proper units for each answer.)

Solution. h(t) = sin(t) represents the height of the cork at any time t,
so the height of the cork when t = 2 is h(2) = sin(2) ≈ 0.91 feet above
the mean water level.
d
The velocity is the derivative of the position, so v(t) = dt h(t) =
d ∆h
dt sin ( t ) = cos ( t ) . The derivative of position is the limit of ∆t so the
,
units are feet . After 2 seconds, the velocity is v(2) = cos(2) ≈
seconds
−0.42 feet per second.
d
The acceleration is the derivative of the velocity, so a(t) = dt v(t) =
d ∆v
dt cos( t ) = − sin( t ). The derivative of velocity is the limit of ∆t , so the
units are feet/second or feet 2 . After 2 seconds the acceleration is
seconds second
a(2) = − sin(2) ≈ −0.91 sec ft . ◀
2

Practice 6. Find the height, velocity and acceleration of the cork in the
previous example after 1 second.

A Most Useful Formula: D( x n )

Functions that include powers of x are very common (every polynomial
is a sum of terms that include powers of x) and, fortunately, it is easy to
calculate the derivatives of such powers. The “pattern” emerging from
the first few examples in this section is, in fact, true for all powers of
x. We will only state and prove the “pattern” here for positive integer
powers of x, but it is also true for other powers (as we will prove later).

Theorem: If n is a positive integer, then: D( x n ) = n · x n−1

This theorem is an example of the power of generality and proof in

mathematics. Rather than resorting to the definition when we encounter
a new exponent p in the form x p (imagine using the definition to
calculate the derivative of x307 ), we can justify the pattern for all positive
integer exponents n, and then simply apply the result for whatever
exponent we have. We know, from the first examples in this section,
that the theorem is true for n = 1, 2 and 3, but no number of examples
would guarantee that the pattern is true for all exponents. We need a
proof that what we think is true really is true.

Proof. With f ( x ) = x n , f ( x + h) = ( x + h)n , and in order to simplify

f ( x + h) − f ( x ) = ( x + h)n − x n , we will need to expand ( x + h)n . How-
ever, we really only need to know the first two terms of the expansion
124 contemporary calculus

and to know that all of the other terms of the expansion contain a
power of h of at least 2.
You may also be familiar with Pascal’s The Binomial Theorem from algebra says (for n > 3) that:
triangle:
1 ( x + h ) n = x n + n · x n −1 h + a · x n −2 h 2 + b · x n −3 h 3 + · · · + h n
1 1

1 2 1 where a and b represent numerical coefficients. (Expand ( x + h)n for a

1 3 3 1
few different values of n to convince yourself of this result.) Then:

1 4 6 4 1 f ( x + h) − f ( x ) ( x + h)n − x n
D( f ( x )) = lim = lim
Among many beautiful and amazing h →0 h h →0 h
properties, the numbers in row n of the
triangle (counting the first row as row 0)
Now expand ( x + h)n to get:
give the coefficients in the expansion of
( A + B)n . Notice that each entry in the x n + n · x n −1 h + a · x n −2 h 2 + b · x n −3 h 3 + · · · + h n − x n
lim
interior of the triangle is the sum of the h →0 h
two numbers immediately above it.
Eliminating x n − x n we get:

n · x n −1 h + a · x n −2 h 2 + b · x n −3 h 3 + · · · + h n
lim
h →0 h
and we can then factor h out of the numerator:
h ( n · x n −1 + a · x n −2 h + b · x n −3 h 2 + · · · + h n −1 )
lim
h →0 h
and divide top and bottom by the factor h:
h i
lim n · x n−1 + a · x n−2 h + b · x n−3 h2 + · · · + hn−1
h →0

We are left with a polynomial in h and can now compute the limit by
simply evaluating the polynomial at h = 0 to get D( x n ) = n · x n−1 .
d d 31
Practice 7. Calculate D( x5 ), 2
dx ( x ), D( x100 ), dt ( t ) and D( x0 ).

We will occasionally use the result of the theorem for the derivatives
of all constant powers of x even though it has only been proven for
positive integer powers, so far. A proof of a more general result (for all
rational powers of x) appears in Section 2.9
d √
 
1
Example 5. Find D and ( x ).
x dx
Solution. Rewriting the fraction using a negative exponent:
 
1 1
D = D ( x −1 ) = − 1 · x −1−1 = − x −2 = − 2
x x
Rewriting the square root using a fractional exponent:
d √ 1 1 1 1 1 1
( x ) = D ( x 2 ) = · x 2 −1 = x − 2 = √
dx 2 2 2 x
These results can also be obtained by using the definition of the deriva-
tive, but the algebra involved is slightly awkward. ◀
 the derivative 125

3 d 1
  d
Practice 8. Find D( x 2 ), ( x 3 ), D √1x and (tπ ).
dx dt
√
Example 6. It costs x hundred dollars to run a training program for
x employees.

(a) How much does it cost to train 100 employees? 101 employees?
If you already need to train 100 employees, how much additional
money will it cost to add 1 more employee to those being trained?
√
(b) For f ( x ) = x, calculate f ′ ( x ) and evaluate f ′ at x = 100. How
does f ′ (100) compare with the last answer in part (a)?
√ 1
Solution. (a) Put f ( x ) = x = x 2 hundred dollars, the cost to train
x employees. Then f (100) = $1000 and f (101) = $1004.99, so it costs
1
$4.99 additional to train the 101st employee. (b) f ′ ( x ) = 21 x − 2 = 2√
1
x
so
f ′ (100) = √1 = 20 1
hundred dollars = $5.00. Clearly f ′ (100) is very
2 100
close to the actual additional cost of training the 101st employee. ◀

Important Information and Results

This section contains a great deal of important information that we will So it is worthwhile to collect here some
of those important ideas.
continue to use throughout the rest of the course.
f ( x + h) − f ( x )
Definition of Derivative: f ′ ( x ) = lim Valid if the limit exists and is finite.
h →0 h
df
Notations for the Derivative: f ′ ( x ), D( f ( x )), dx
An equation of the line tangent to the
Tangent Line Equation: y = f ( a) + f ′ ( a) · ( x − a) graph of f at ( a, f ( a)).
Formulas:

• D(constant) = 0
Proved for n = positive integer, but true
• D ( x n ) = n · x n −1 for all constants n.

• D(sin( x )) = cos( x ) and D(cos( x )) = − sin( x )


 1 if x > 0
 |x|
• D(| x |) = undefined if x = 0 =
 x
 −1 if x < 0

Interpretations of f ′ ( x ):

• Slope of a line tangent to a graph

• Instantaneous rate of change of a function at a point

• Velocity or acceleration

• Magnification factor

• Marginal change
126 contemporary calculus

2.1 Problems

1. Match the functions f , g and h shown below with 6. f ( x ) = x3 + 4x

the graphs of their derivatives (show in the bot- (a) (1, 5), (1 + h, (1 + h)3 + 4(1 + h))
tom row).
(b) ( x, x3 + 4x ), ( x + h, ( x + h)3 + 4( x + h))
7. Use the graph below to estimate the values of
these limits. (It helps to recognize what the limit
represents.)

f (0 + h ) − f (0) f (1 + h ) − f (1)
(a) lim (b) lim
h →0 h h →0 h

f (2 + w ) − 1 f (3 + h ) − f (3)
(c) lim (d) lim
w →0 w h →0 h

2. The figure below shows six graphs, three of which f (4 + h ) − f (4) f (5 + s ) − f (5)
(e) lim (f) lim
are derivatives of the other three. Match the func- h →0 h s →0 s
tions with their derivatives.

8. Use the graph below to estimate the values of

these limits.

In Problems 3–6, find the slope msec of the secant g (0 + h ) − g (0) g (1 + h ) − g (1)
(a) lim (b) lim
line through the two given points and then calculate h →0 h h →0 h
mtan = lim msec .
h →0 g (2 + w ) − 2 g (3 + h ) − g (3)
(c) lim (d) lim
3. f ( x ) = x2 w →0 w h →0 h

(a) (−2, 4), (−2 + h, (−2 + h)2 ) g (4 + h ) − g (4) g (5 + s ) − g (5)

(e) lim (f) lim
(b) (0.5, 0.25), (0.5 + h, (0.5 + h)2 ) h →0 h s →0 s
4. f ( x ) = 3 + x2
(a) (1, 4), (1 + h, 3 + (1 + h)2 )
(b) ( x, 3 + x2 ), ( x + h, 3 + ( x + h)2 )
5. f ( x ) = 7x − x2
(a) (1, 6), (1 + h, 7(1 + h) − (1 + h)2 )
(b) ( x, 7x − x2 ), ( x + h, 7( x + h) − ( x + h)2 )
 the derivative 127

In Problems 9–12, use the definition of the derivative 21. (a) Find the angle that the line tangent to y = x2
to calculate f ′ ( x ) and then evaluate f ′ (3). at (1, 1) makes with the x-axis.
(b) Find the angle that the line tangent to y = x3
9. f ( x ) = x2 +8 10. f ( x ) = 5x2 − 2x at (1, 1) makes with the x-axis.
(c) The curves y = x2 and y = x3 intersect at the
11. f ( x ) = 2x3 − 5x 12. f ( x ) = 7x3 + x point (1, 1). Find the angle of intersection of
the two curves (actually the angle between
their tangent lines) at the point (1, 1).
13. Graph f ( x ) = x2 , g( x ) = x2 + 3 and h( x ) =
22. The figure below shows the graph of y = f ( x ).
x2 − 5. Calculate the derivatives of f , g and h.
Sketch a graph of y = f ′ ( x ).
14. Graph f ( x ) = 5x, g( x ) = 5x + 2 and h( x ) =
5x − 7. Calculate the derivatives of f , g and h.

In Problems 15–18, find the slopes and equations of

the lines tangent to y = f ( x ) at the given points.

15. f ( x ) = x2 + 8 at (1, 9) and (−2, 12).

16. f ( x ) = 5x2 − 2x at (2, 16) and (0, 0).

17. f ( x ) = sin( x ) at (π, 0) and ( π2 , 1). 23. The figure below shows the graph of the height of
an object at time t. Sketch a graph of the object’s
18. f ( x ) = | x + 3| at (0, 3) and (−3, 0).
upward velocity. What are the units for each axis
19. (a) Find an equation of the line tangent to the on the velocity graph?
graph of y = x2 + 1 at the point (2, 5).
(b) Find an equation of the line perpendicular to
the graph of y = x2 + 1 at (2, 5).
(c) Where is the line tangent to the graph of
y = x2 + 1 horizontal?
(d) Find an equation of the line tangent to the
graph of y = x2 + 1 at the point ( p, q).
(e) Find the point(s) ( p, q) on the graph of y =
x2 + 1 so the tangent line to the curve at ( p, q) 24. Fill in the table with units for f ′ ( x ).
goes through the point (1, −7).

20. (a) Find an equation of the line tangent to the units for x units for f ( x ) units for f ′ ( x )
graph of y = x3 at the point (2, 8).
hours miles
(b) Where, if ever, is the line tangent to the graph
people automobiles
of y = x3 horizontal?
dollars pancakes
(c) Find an equation of the line tangent to the days trout
graph of y = x3 at the point ( p, q). seconds miles per second
(d) Find the point(s) ( p, q) on the graph of y = x3 seconds gallons
so the tangent line to the curve at ( p, q) goes study hours test points
through the point (16, 0).
 128 contemporary calculus

25. A rock dropped into a deep hole will drop 29. Define A( x ) to be the area bounded by the t-axis,
d( x ) = 16x2 feet in x seconds. the line y = t, and a vertical line at t = x (see
figure below).
(a) How far into the hole will the rock be after 4 (a) Evaluate A(0), A(1), A(2) and A(3).
seconds? After 5 seconds? (b) Find a formula for A( x ) valid for x ≥ 0.
(b) How fast will it be falling at exactly 4 seconds? (c) Determine A′ ( x ).
After 5 seconds? After x seconds? (d) What does A′ ( x ) represent?

26. It takes T ( x ) = x2 hours to weave x small rugs.

What is the marginal production time to weave
a rug? (Be sure to include the units with your
answer.)

√
27. It costs C ( x ) = x dollars to produce x golf balls.
What is the marginal production cost to make a
golf ball? What is the marginal production cost
when x = 25? When x = 100? (Include units.) 30. Compute each derivative.

28. Define A( x ) to be the area bounded by the t- and (a) D( x12 ) d √

(b) ( 7 x)
y-axes, the line y = 5 and a vertical line at t = x dx
(see figure below). 
1

d c
(c) D (d) (x )
x3 dx
(a) Evaluate A(0), A(1), A(2) and A(3).
(e) D(| x − 2|)
(b) Find a formula for A( x ) valid for x ≥ 0.
31. Compute each derivative.
(c) Determine A′ ( x ).
(a) D( x9 ) d 2
(b) (x 3 )
(d) What does A′ ( x ) represent? dx
 
1 d π
(c) D (d) (x )
x4 dx

(e) D(| x + 5|)

In Problems 32–37, find a function f that has the

given derivative. (Each problem has several correct
answers, just find one of them.)

32. f ′ ( x ) = 4x + 3 33. f ′ ( x ) = 3x2 + 8x

34. D( f ( x )) = 12x2 − 7 35. f ′ (t) = 5 cos(t)

d
36. dx f ( x ) = 2x − sin( x ) 37. D( f ( x )) = x + x2
 the derivative 129

2.1 Practice Answers

1. The graph of f ( x ) = 7x is a line through the origin. The slope of the
line is 7. For all x:
f ( x + h) − f ( x ) 7( x + h) − 7x 7h
mtan = lim = lim = lim = lim 7 = 7
h →0 h h →0 h h →0 h h →0

2. f ( x ) = x3 ⇒ f ( x + h) = ( x + h)3 = x3 + 3x2 h + 3xh2 + h3 so:

f ( x + h) − f ( x ) x3 + 3x2 h + 3xh2 + h3 − x3
f ′ ( x ) = lim = lim
h →0 h h →0 h
2 2
3x h + 3xh + h 3
= lim = lim 3x2 + 3xh + h2 = 3x2
h →0 h h →0

At the point (2, 8), the slope of the tangent line is 3(2)2 = 12 so an
equation of the tangent line is y − 8 = 12( x − 2) or y = 12x − 16.

3. D( x4 ) = 4x3 , D( x5 ) = 5x4 , D( x43 ) = 43x42 ,

√ 1 1
D( x ) = D( x 2 ) = 12 x − 2 = 2√1
x
, D( x π ) = πx π −1

4. Proceeding as we did to find the derivative to sin( x ):

cos( x + h) − cos( x ) cos( x ) cos(h) − sin( x ) sin(h) − cos( x )

D(cos( x )) = lim = lim
h →0 h h →0 h
cos(h) − 1
 
sin(h)
= lim cos( x ) · − sin( x ) = cos( x ) · 0 − sin( x ) · 1 = − sin( x )
h →0 h h

5. See margin figure for the graphs of y = | x − 2| and y = |2x |.


 1 if x > 2
 | x − 2|
D(| x − 2|) = undefined if x = 2 =
 x−2
 −1 if x < 2

 2 if x > 0
 2 |x|
D(|2x |) = undefined if x = 0 =
 x
 −2 if x < 0

6. h(t) = sin(t) so h(1) = sin(1) ≈ 0.84 ft;

v(t) = cos(t) so v(1) = cos(1) ≈ 0.54 ft/sec;
a(t) = − sin(t) so a(1) = − sin(1) ≈ −0.84 ft/sec2 .
d d 31
7. D( x5 ) = 5x4 , dx ( x2 ) = 2x1 = 2x, D( x100 ) = 100x99 , dt ( t ) = 31t30
and D( x0 ) = 0x −1 = 0 or D( x0 ) = D(1) = 0
3 3 21 d 1 1 2 1 3
8. D( x 2 ) = 2x , ( x 3 ) = x − 3 , D( √1x ) = D( x − 2 ) = − 21 x − 2 ,
dx 3
d π
(t ) = πtπ −1 .
dt
130 contemporary calculus

2.2 Derivatives: Properties and Formulas

This section begins with a look at which functions have derivatives.

Then we’ll examine how to calculate derivatives of elementary combi-
nations of basic functions. By knowing the derivatives of some basic
functions and just a few differentiation patterns, you will be able to
calculate the derivatives of a tremendous variety of functions. This
section contains most — but not quite all — of the general differentiation
patterns you will ever need.

Which Functions Have Derivatives?

A function must be continuous in order to be differentiable.

It is vital to understand what this theo- Theorem:

rem tells us and what it does not tell us:
If a function is differentiable at a point, If a function is differentiable at a point
then the function is automatically contin-
then it is continuous at that point.
uous there. If the function is continuous
at a point, then the function may or may
not be differentiable there.
Proof. Assume that the hypothesis ( f is differentiable at the point c) is
f (c + h) − f (c)
true. Then lim must exist and be equal to f ′ (c). We
h →0 h
want to show that f must necessarily be continuous at c, so we need to
show that lim f (c + h) = f (c).
h →0
It’s not yet obvious why we want to do so, but we can write:
f (c + h) − f (c)
f (c + h) = f (c) + ·h
h
and then compute the limit of both sides of this expression:
f (c + h) − f (c)
 
lim f (c + h) = lim f (c) + ·h
h →0 h →0 h
f (c + h) − f (c)
 
= lim f (c) + lim ·h
h →0 h →0 h
f (c + h) − f (c)
 
= lim f (c) + lim · lim h
h →0 h →0 h h →0

= f (c) + f ′ (c) · 0 = f (c)

Therefore f is continuous at c.

We often use the contrapositive form of this theorem, which tells us

about some functions that do not have derivatives.

Contrapositive Form of the Theorem:

If f is not continuous at a point

then f is not differentiable at that point.
 the derivative 131

Example 1. Show that f ( x ) = ⌊ x ⌋ is not continuous and not differen-

tiable at x = 2 (see margin figure).

Solution. The one-sided limits lim ⌊ x ⌋ = 2 and lim ⌊ x ⌋ = 1 have

x →2+ x →2−
different values, so lim ⌊ x ⌋ does not exist. Therefore f ( x ) = ⌊ x ⌋ is not
x →2
continuous at 2, and as a result it is not differentiable at 2. ◀

Lack of continuity implies lack of differentiability, but the next exam-

ples show that continuity is not enough to guarantee differentiability.

Example 2. Show that f ( x ) = | x | is continuous but not differentiable

at x = 0 (see margin figure).

Solution. We know that lim | x | = 0 = |0|, so f is continuous at 0, but

x →0
in Section 2.1 we saw that | x | was not differentiable at x = 0. ◀

A function is not differentiable at a cusp or a “corner.”

√ 1
Example 3. Show that f ( x ) = 3 x = x 3 is continuous but not differen-
tiable at x = 0 (see margin figure).
√ √ √
Solution. We can verify that lim 3 x = lim 3 x = 0, so lim 3 x =
x →0+ x →0− x →0
√
3 ′ 1 −2 1
0 = 0 so f is continuous at 0. But f ( x ) = x 3 = √ 3
, which is
3 3 x2
undefined at x = 0, so f is not differentiable at 0. ◀

A function is not differentiable where its tangent line is vertical.

Practice 1. At which integer values of x is the graph of f in the margin

figure continuous? Differentiable?

Graphically, a function is continuous if and only if its graph is

“connected” and does not have any holes or breaks. Graphically, a
function is differentiable if and only if it is continuous and its graph is
“smooth” with no corners or vertical tangent lines.

Derivatives of Elementary Combinations of Functions

We now begin to compute derivatives of more complicated functions
built from combinations of simpler functions.

Example 4. The derivative of f ( x ) = x is D( f ( x )) = 1 and the deriva-

tive of g( x ) = 5 is D( g( x )) = 0. What are the derivatives of the
f
elementary combinations: 3 · f , f + g, f − g, f · g and ?
g
132 contemporary calculus

Solution. The first three derivatives follow “nice” patterns:

D(3 · f ( x )) = D(3x ) = 3 = 3 · 1 = 3 · D( f ( x ))
D( f ( x ) + g( x )) = D( x + 5) = 1 = 1 + 0 = D( f ( x )) + D( g( x ))
D( f ( x ) − g( x )) = D( x − 5) = 1 = 1 − 0 = D( f ( x )) − D( g( x ))

yet the other two derivatives fail to follow the same “nice” patterns:
D(f ( x ) · 
g( x )) = D(5x ) = 5 but D( f ( x )) · D( g( x )) = 1 · 0 = 0, and
f (x) x 1 D( f ( x )) 1
D =D = but = is undefined. ◀
g( x ) 5 5 D( g( x )) 0
The two very simple functions in the previous
  example show that,
f D( f )
in general, D( f · g) ̸= D( f ) · D( g) and D ̸= .
g D( g)
Practice 2. For f ( x ) = 6x + 8 and g( x ) = 2, compute the derivatives of
f
3 · f , f + g, f − g, f · g and .
g

Main Differentiation Theorem:

If f and g are differentiable at x, then:

(a) Constant Multiple Rule:

D(k · f ( x )) = k · D( f ( x ))

(b) Sum Rule:

D( f ( x ) + g( x )) = D( f ( x )) + D( g( x ))

(c) Difference Rule:

D( f ( x ) − g( x )) = D( f ( x )) − D( g( x ))

(d) Product Rule:

D( f ( x ) · g( x )) = f ( x ) · D( g( x )) + g( x ) · D( f ( x ))

(e) Quotient Rule:

g( x ) · D( f ( x )) − f ( x ) · D( g( x ))
 
Part (e) requires that g( x ) ̸= 0. f (x)
D =
g( x ) [ g( x )]2

This theorem says that the simple patterns in the previous example
for constant multiples of functions and sums and differences of func-
tions are true for all differentiable functions. It also includes the correct
patterns for derivatives of products and quotients of differentiable
functions.
 the derivative 133

The proofs of parts (a), (b) and (c) of this theorem are straightforward,
but parts (d) and (e) require some clever algebraic manipulations. Let’s
look at some examples before tackling the proof.

Example 5. Recall that D( x2 ) = 2x and D(sin( x )) = cos( x ). Find

d
D(3 sin( x )) and dx (5x2 − 7 sin( x )).

Solution. Computing D(3 sin( x )) requires part (a) of the theorem with
k = 3 and f ( x ) = sin( x ) so D(3 · sin( x )) = 3 · D(sin( x )) = 3 cos( x ),
d
while (5x2 − 7 sin( x )) uses part (c) of the theorem with f ( x ) = 5x2
dx
and g( x ) = 7 sin( x ) so:

d d d
(5x2 − 7 sin( x )) = (5x2 ) − (7 sin( x ))
dx dx dx
d 2 d
= 5 · ( x ) − 7 · (sin( x ))
dx dx
= 5(2x ) − 7(cos( x ))

which simplifies to 10x − 7 cos( x ). ◀

d
Practice 3. Find D( x3 − 5 sin( x )) and (sin( x ) − 4x3 ).
dx
Practice 4. The table below gives the values of functions f and g, as
well as their derivatives, at various points. Fill in the missing values for
D(3 · f ( x )), D(2 · f ( x ) + g( x )) and D(3 · g( x ) − f ( x )).

x f (x) f ′ (x) g( x ) g′ ( x ) D(3 f ( x )) D(2 f ( x ) + g( x )) D(3g( x ) − f ( x ))

0 3 −2 −4 3
1 2 −1 1 0
2 4 2 3 1

Practice 5. Use the Main Differentiation Theorem to complete the table.

   
f (x) g( x )
x f (x) f ′ (x) g( x ) g′ ( x ) D( f ( x ) · g( x )) D D
g( x ) f (x)
0 3 −2 −4 3
1 2 −1 1 0
2 4 2 3 1

x3
 
d
Example 6. Determine D( x2 · sin( x )) and .
dx sin( x )

Solution. (a) Use the Product Rule with f ( x ) = x2 and g( x ) = sin( x ):

Many calculus students find it easier to
2
remember the Product Rule in words:
D( x · sin( x )) = D( f ( x ) · g( x )) = f ( x ) · D( g( x )) + g( x ) · D( f ( x )) “the first function times the derivative of
the second plus the second function times
= x2 · D(sin( x )) + sin( x ) · D( x2 ) the derivative of the first.”
= x2 · cos( x ) + sin( x ) · 2x = x2 cos( x ) + 2x sin( x )
134 contemporary calculus

(b) Use the Quotient Rule with f ( x ) = x3 and g( x ) = sin( x ):

x3
   
d d f (x)
=
dx sin( x ) dx g( x )
The Quotient Rule in words: “the bottom g( x ) · D( f ( x )) − f ( x ) · D( g( x ))
times the derivative of the top minus the =
top times the derivative of the bottom, all [ g( x )]2
over the bottom squared.”
sin( x ) · D( x3 ) − x3 · D(sin( x ))
=
[sin( x )]2
sin( x ) · 3x2 − x3 · cos( x )
=
sin2 ( x )
3x2 sin( x ) − x3 · cos( x )
=
sin2 ( x )

which could also be rewritten in terms of csc( x ) and cot( x ). ◀

3t − 2
   
d cos( x )
Practice 6. Find D(( x2 + 1)(7x − 3)), and D .
dt 5t + 1 x
Now that we’ve seen how to use the theorem, let’s prove it.

Proof. The only general fact we have about derivatives is the definition
as a limit, so our proofs here will need to recast derivatives as limits
and then use some results about limits. The proofs involve applications
of the definition of the derivative and results about limits.

(a) Using the derivative definition and the limit laws:

k · f ( x + h) − k · f ( x )
D(k · f ( x )) = lim
h →0 h
f ( x + h) − f ( x )
= lim k ·
h →0 h
f ( x + h) − f ( x )
= k · lim = k · D( f ( x ))
h →0 h

(b) You try it (see Practice problem that follows).

(c) Once again using the derivative definition and the limit laws:

[ f ( x + h) − g( x + h)] − [ f ( x ) − g( x )]
D( f ( x ) − g( x )) = lim
h →0 h
[ f ( x + h) − f ( x )] − [ g( x + h) − g( x )]
= lim
h →0 h
f ( x + h) − f ( x ) g( x + h) − g( x )
= lim − lim
h →0 h h →0 h
= D( f ( x )) − D( g( x ))

The proofs of parts (d) and (e) of the theorem are more complicated
but only involve elementary techniques, used in just the right way.
 the derivative 135

Sometimes we will omit such computational proofs, but the Prod-

uct and Quotient Rules are fundamental techniques you will need
hundreds of times.

(d) By the hypothesis, f and g are differentiable, so:

f ( x + h) − f ( x )
lim = f ′ (x)
h →0 h
and:
g( x + h) − g( x )
lim = g′ ( x )
h →0 h
Also, both f and g are continuous (why?) so lim f ( x + h) = f ( x )
h →0
and lim g( x + h) = g( x ).
h →0
Let P( x ) = f ( x ) · g( x ). Then P( x + h) = f ( x + h) · g( x + h) and:
P( x + h) − P( x )
D( f ( x ) · g( x )) = D( P( x )) = lim
h →0 h
f ( x + h) · g( x + h) − f ( x ) · g( x )
= lim
h →0 h
At this stage we need to use some cleverness to add and subtract
f ( x ) · g( x + h) from the numerator (you’ll see why shortly):
f ( x + h) · g( x + h) + [− f ( x ) · g( x + h) + f ( x ) · g( x + h)] − f ( x ) g( x )
lim
h →0 h
We can then split this giant fraction into two more manageable limits:
f ( x + h) g( x + h) − f ( x ) g( x + h) f ( x ) g( x + h) − f ( x ) g( x )
lim + lim
h →0 h h →0 h
and then factor out a common factor from each numerator:
f ( x + h) − f ( x ) g( x + h) − g( x )
lim g( x + h) · + lim f ( x ) ·
h →0 h h →0 h
Taking limits of each piece (and using the continuity of g( x )) we get:

D( f ( x ) · g( x )) = g( x ) · f ′ ( x ) + f ( x ) · g′ ( x ) = g · D( f ) + f · D( g)

The steps for a proof of the Quotient Rule appear in Problem 69.

Practice 7. Prove the Sum Rule: D( f ( x ) + g( x )) = D( f ( x )) + D( g( x )).

(Refer to the proof of part (c) for guidance.)

Using the Differentiation Rules

You definitely need to memorize the differentiation rules, but it is vi-
tally important that you also know how to use them. Sometimes it is
clear that the function we want to differentiate is a sum or product of
two obvious functions, but we commonly need to differentiate func-
tions that involve several operations and functions. Memorizing the
differentiation rules is only the first step in learning to use them.
136 contemporary calculus

Example 7. Calculate D( x5 + x · sin( x )).

Solution. This function is more difficult because it involves both an

addition and a multiplication. Which rule(s) should we use — or, more
importantly, which rule should we use first?
First apply the Sum Rule to trade one derivative for two easier ones:

D( x5 + x · sin( x )) = D( x5 ) + D( x · sin( x ))
= 5x4 + [ x · D(sin( x )) + sin( x ) · D( x )]
= 5x4 + x · cos( x ) + sin( x )

This last expression involves no more derivatives, so we are done. ◀

If instead of computing the derivative you were evaluating the func-

tion x5 + x sin( x ) for some particular value of x, you would:
• raise x to the 5th power

• calculate sin( x )

• multiply sin( x ) by x and, finally,

• add (sum) the values of x5 and x sin( x )

Notice that the final step of your evaluation of f indicates the first rule
to use to calculate the derivative of f .
Practice 8. Which differentiation rule should you apply first for each
of the following?

(a) x · cos( x ) − x3 · sin( x ) (b) (2x − 3) cos( x )

(c) 2 cos( x ) − 7x2 cos( x ) + 3x

√(d)
x
 2  2
x −5 t −5
 
d
Practice 9. Calculate D and .
sin( x ) dt t · sin(t)

Example 8. A mass attached to a spring oscillates up and down but

the motion becomes “damped” due to friction and air resistance. The
sin(t)
height of the mass after t seconds is given by h(t) = 5 + (in feet).
1+t
Find the height and velocity of the mass after 2 seconds.
sin(2) 0.909
Solution. The height is h(2) = 5 + ≈ 5+ = 5.303 feet
1+2 3
above the ground. The velocity is h′ (2), so we must first compute h′ (t)
and then evaluate the derivative at time t = 2:
(1 + t) · cos(t) − sin(t) · 1
h′ (t) =
(1 + t )2
3 cos(2) − sin(2) −2.158
so h′ (2) = ≈ ≈ −0.24 feet per second. ◀
9 9
 the derivative 137

Practice 10. What are the height and velocity of the weight in the
previous example after 5 seconds? What are the height and velocity of
the weight be after a “long time” has passed?
Example 9. Calculate D( x · sin( x ) · cos( x )).

Solution. Clearly we need to use the Product Rule, because the only
operation in this function is multiplication. But the Product Rule deals
with a product of two functions and here we have the product of three:
x and sin( x ) and cos( x ). If, however, we think of our two functions as
f ( x ) = x · sin( x ) and g( x ) = cos( x ), then we do have the product of
two functions and:

D( x · sin( x ) · cos( x )) = D( f ( x ) · g( x ))
= f ( x ) · D( g( x )) + g( x ) · D( f ( x ))
= x sin( x ) · D(cos( x )) + cos( x ) · D( x sin( x ))
We are not done, but we have traded one hard derivative for two
easier ones. We know that D(cos( x )) = − sin( x ) and we can use the
Product Rule (again) to calculate D( x sin( x )). Then the last line of our
calculation above becomes:

x sin( x ) · [− sin( x )] + cos( x ) · [ x D(sin( x )) + sin( x ) D( x )]

and then:
− x sin2 ( x ) + cos( x ) [ x cos( x ) + sin( x )(1)]
which simplifies to − x sin2 ( x ) + x cos2 ( x ) + cos( x ) sin( x ). ◀

Evaluating a Derivative at a Point

The derivative of a function f ( x ) is a new function f ′ ( x ) that tells us
the slope of the line tangent to the graph of f at each point x. To find
the slope of the tangent line at a particular point (c, f (c)) on the graph
of f , we should first calculate the derivative f ′ ( x ) and then evaluate
the function f ′ ( x ) at the point x = c to get the number f ′ (c). If you
mistakenly evaluate f first, you get a number f (c), and the derivative
of a constant is always equal to 0.
Example 10. Determine the slope of the line tangent to the graph of
f ( x ) = 3x + sin( x ) at (0, f (0)) and (1, f (1)).

Solution. f ′ ( x ) = D(3x + sin( x )) = D(3x ) + D(sin( x )) = 3 + cos( x ).

When x = 0, the graph of y = 3x + sin( x ) goes through the point
(0, 3(0) + sin(0)) = (0, 0) with slope f ′ (0) = 3 + cos(0) = 4. When
x = 1, the graph goes through the point (1, 3(1) + sin(1)) ≈ (1, 3.84)
with slope f ′ (1) = 3 + cos(1) ≈ 3.54. ◀

Practice 11. Where do f ( x ) = x2 − 10x + 3 and g( x ) = x3 − 12x have

horizontal tangent lines?
138 contemporary calculus

Important Information and Results

This section, like the last one, contains a Differentiability and Continuity: If a function is differentiable then
great deal of important information that
we will continue to use throughout the it must be continuous. If a function is not continuous then it cannot
rest of the course, so we collect here some be differentiable. A function may be continuous at a point and not
of those important results.
differentiable there.
Graphically: Continuous means “connected”; differentiable means “con-
tinuous, smooth and not vertical.”
Differentiation Patterns:

• [k · f ( x )]′ = k · f ′ ( x )

• [ f ( x ) + g( x )]′ = f ′ ( x ) + g′ ( x )

• [ f ( x ) − g( x )]′ = f ′ ( x ) − g′ ( x )

• [ f ( x ) · g( x )]′ = f ( x ) · g′ ( x ) + g( x ) · f ′ ( x )

f (x) ′ g( x ) · f ′ ( x ) − f ( x ) · g′ ( x )
 
• =
g( x ) [ g( x )]2
• The final step used to evaluate a function f indicates the first rule
used to differentiate f .

Evaluating a derivative at a point: First differentiate and then evaluate.

2.2 Problems

1. Use the graph of y = f ( x ) below to determine: 2. Use the graph of y = g( x ) below to determine:
(a) at which integers f is continuous. (a) at which integers g is continuous.
(b) at which integers f is differentiable. (b) at which integers g is differentiable.

 
f f
3. Use the values given in the table to determine the values of f · g, D( f · g), gand D g .
 
f (x) f (x)
x f (x) f ′ (x) g( x ) g′ ( x ) f ( x ) · g( x ) D( f ( x ) · g( x )) g( x )
D g( x )

0 2 3 1 5
1 −3 2 5 −2
2 0 −3 2 4
3 1 −1 0 3
 the derivative 139

 
f f
4. Use the values given in the table to determine the values of f · g, D( f · g), g and D g .
 
f (x) f (x)
x f (x) f ′ (x) g( x ) g′ ( x ) f ( x ) · g( x ) D( f ( x ) · g( x )) g( x )
D g( x )

0 4 2 3 −3
1 0 3 2 1
2 −2 5 0 −1
3 −1 −2 −3 4
   
5. Use the information in the figure below to plot d 1 d t
19. 20.
f
the values of the functions f + g, f · g and g and dx 1 + x2 dt 1 + t3
their derivatives at x = 1, 2 and 3.    
d 1 d 1
21. 22.
dθ cos(θ ) dθ sin(θ )
   
d sin(θ ) d cos(θ )
23. 24.
dθ cos(θ ) dθ sin(θ )
 
25. D 8x5 − 3x4 + 2x3 + 7x2 − 12x + 147

6. Use the information in the figure above to plot 26. (a) D(sin( x )) (b) D (sin( x ) + 7)
g
the values of the functions 2 f , f − g and f and (c) D(sin( x ) − 8000) (d) D(sin( x ) + k )
their derivatives at x = 1, 2 and 3. 27. Find values for the constants a, b and c so that
7. Calculate D(( x − 5)(3x + 7)) by: the parabola f ( x ) = ax2 + bx + c has f (0) = 0,
f ′ (0) = 0 and f ′ (10) = 30.
(a) using the Product Rule.
(b) expanding and then differentiating. 28. If f is a differentiable function, how are the:

Verify that both methods give the same result. (a) graphs of y = f ( x ) and y = f ( x ) + k related?
 3
x − 3x + 2 (b) derivatives of f ( x ) and f ( x ) + k related?

8. Calculate D √ by:
x 29. If f and g are differentiable functions that always
(a) using the Quotient Rule. differ by a constant ( f ( x ) − g( x ) = k for all x)
(b) rewriting and then differentiating. then what can you conclude about their graphs?
Verify that both methods give the same result. Their derivatives?
30. If f and g are differentiable functions whose sum
In Problems 9–26, compute each derivative. is a constant ( f ( x ) + g( x ) = k for all x) then
what can you conclude about their graphs? Their
d   d  π
9. 19x3 − 7 10. 5 cos(t) + derivatives?
dx dt 2
31. If the product of f and g is a constant (that is,
11. D(sin( x ) + cos( x )) 12. D(7 sin( x ) − 3 cos( x )) D( f ( x ))
f ( x ) · g( x ) = k for all x) then how are
√ f (x)
13. D( x2 · cos( x )) 14. D( x · sin( x )) D( g( x ))
and related?
g( x )
15. D(sin2 ( x )) 16. D(cos2 ( x ))
f (x)
32. If the quotient of f and g is a constant ( =k
g( x )
   
d cos( x ) d sin(t)
17.
dx x2
18.
dt t3 for all x) then how are g · f ′ and f · g′ related?
 140 contemporary calculus

In Problems 33–40: 52. For what value(s) of A is

(a) calculate f ′ (1)
(
Ax − 4 if x < 2
(b) determine where f ′ ( x ) = 0. f (x) =
x2 + x if x ≥ 2
33. f ( x ) = x2 − 5x + 13 differentiable at x = 2?
34. f ( x ) = 5x2 − 40x + 73 53. For what values of A and B is
35. f ( x ) = 3x − 2 cos( x ) (
Ax + B if x < 1
36. f ( x ) = | x + 2| f (x) =
x2 + x if x ≥ 1
37. f ( x ) = x3 + 9x2 + 6
differentiable at x = 1?
38. f ( x ) = x3 + 3x2 + 3x − 1
54. An arrow shot straight up from ground level (get
39. f ( x ) = x3 + 2x2 + 2x − 1 out of the way!) with an initial velocity of 128 feet
7x per second will be at height h( x ) = −16x2 + 128x
40. f ( x ) =
x2 +4 feet after x seconds (see figure below).
41. f ( x ) = x · sin( x ) and 0 ≤ x ≤ 5. (You may (a) Determine the velocity of the arrow when
need to use the Bisection Algorithm or the “trace” x = 0, 1 and 2 seconds.
option on a calculator to approximate where
(b) What is the velocity of the arrow, v( x ), at any
f ′ ( x ) = 0.)
time x?
42. f ( x ) = Ax2 + Bx + C, where B and C are con- (c) At what time x will the velocity of the arrow
stants and A ̸= 0 is constant. be 0?
43. f ( x ) = x3 + Ax2 + Bx + C with constants A, B (d) What is the greatest height the arrow reaches?
and C. Can you find conditions on the constants (e) How long will the arrow be aloft?
A, B and C that will guarantee that the graph
(f) Use the answer for the velocity in part (b) to
of y = f ( x ) has two distinct “turning points”?
determine the acceleration, a( x ) = v′ ( x ), at
(Here a “turning point” means a place where the
any time x.
curve changes from increasing to decreasing or
from decreasing to increasing, like the vertex of a
parabola.)

In 44–51, where are the functions differentiable?

44. f ( x ) = | x | cos( x ) 45. f ( x ) = tan( x )

x−5 x2 + x
46. f ( x ) = 47. f ( x ) =
x+3 x2 − 3x

48. f ( x ) =  x2 − 4 49. f ( x ) =  x3 − 1
   

( 55. If an arrow is shot straight up from ground level

0 if x < 0
50. f ( x ) = on the moon with an initial velocity of 128 feet per
sin( x ) if x ≥ 0
( second, its height will be h( x ) = −2.65x2 + 128x
x if x < 0 feet after x seconds. Redo parts (a)–(e) of problem
51. f ( x ) =
sin( x ) if x ≥ 0 40 using this new formula for h( x ).
 the derivative 141

56. In general, if an arrow is shot straight upward 60. (a) Find an equation for the line L that is tangent
with an initial velocity of 128 feet per second from 1
to the curve y = at the point (1, 1).
ground level on a planet with a constant gravita- x
(b) Determine where L intersects the x-axis and
tional acceleration of g feet per second2 then its
g the y-axis.
height will be h( x ) = − 2 x2 + 128x feet after x
seconds. Answer the questions in problem 40 for (c) Determine the area of the region in the first
arrows shot on Mars and Jupiter. quadrant bounded by L, the x-axis and the
y-axis (see figure below).
object g (ft/sec2 ) g (cm/sec2 )
Mercury 11.8 358
Venus 20.1 887
Earth 32.2 981
moon 5.3 162
Mars 12.3 374
Jupiter 85.3 2601
Saturn 36.6 1117
Uranus 34.4 1049 61. (a) Find an equation for the line L that is tangent
Neptune 1
43.5 1325 to the curve y = at the point (2, 12 ).
x
Source: CRC Handbook of Chemistry and Physics 1
(b) Graph y = x and L and determine where L
57. If an object on Earth is propelled upward from intersects the x-axis and the y-axis.
ground level with an initial velocity of v0 feet per (c) Determine the area of the region in the first
second, then its height after x seconds will be quadrant bounded by L, the x-axis and the
h( x ) = −16x2 + v0 x. y-axis.
(a) Find the object’s velocity after x seconds. 62. (a) Find an equation for the line L that is tan-
1
(b) Find the greatest height the object will reach. gent to the curve y = at the point ( p, 1p )
x
(c) How long will the object remain aloft? (assuming p ̸= 0).
58. In order for a 6-foot-tall basketball player to dunk (b) Determine where L intersects the x-axis and
the ball, the player must achieve a vertical jump of the y-axis.
about 3 feet. Use the information in the previous (c) Determine the area of the region in the first
problems to answer the following questions. quadrant bounded by L, the x-axis and the
(a) What is the smallest initial vertical velocity the y-axis.
player can have and still dunk the ball? (d) How does the area of the triangle in part (c)
(b) With the initial velocity achieved in part (a), depend on the initial point ( p, 1p )?
how high would the player jump on the moon? 63. Find values for the coefficients a, b and c so that
59. The best high jumpers in the world manage to lift the parabola f ( x ) = ax2 + bx + c goes through
their centers of mass approximately 3.75 feet. the point (1, 4) and is tangent to the line y =
(a) What is the initial vertical velocity these high 9x − 13 at the point (3, 14).
jumpers attain? 64. Find values for the coefficients a, b and c so that
(b) How long are these high jumpers in the air? the parabola f ( x ) = ax2 + bx + c goes through
(c) How high would they lift their centers of mass the point (0, 1) and is also tangent to the line
on the moon? y = 3x − 2 at the point (2, 4).
 142 contemporary calculus

65. (a) Find a function f so that D( f ( x )) = 3x2 . 68. The graph of y = g′ ( x ) appears below. Assume
(b) Find another function g with D( g( x )) = 3x2 . that g is continuous.
(c) Can you find more functions whose deriva- (a) Assume g(0) = 0 and sketch a graph of
tives are 3x2 ? y = g ( x ).
66. (a) Find a function f so that f ′ ( x ) = 6x + cos( x ).
(b) Assume g(0) = 1 and graph y = g( x ).
(b) Find another function g with g′ ( x ) = f ′ ( x ).
67. The graph of y = f ′ ( x ) appears below.
(a) Assume f (0) = 0 and sketch a graph of
y = f ( x ).
(b) Assume f (0) = 1 and graph y = f ( x ).

69. Assume that f and g are differentiable functions

and that g( x ) ̸= 0. State why each step in the
following proof of the Quotient Rule is valid.

Proof of the Quotient Rule

1 f ( x + h) g( x ) − g( x + h) f ( x )
     
f (x) 1 f ( x + h) f (x)
D = lim − = lim
g( x ) h →0 h g( x + h) g( x ) h →0 h g( x + h) g( x )
f ( x + h) g( x ) + (− f ( x ) g( x ) + f ( x ) g( x )) − g( x + h) f ( x )
 
1
= lim
h →0 g ( x + h ) g ( x ) h
f ( x + h) − f ( x ) g( x ) − g( x + h)
 
1
= lim g( x ) + f (x)
h →0 g ( x + h ) g ( x ) h h
1
g( x ) · f ′ ( x ) − f ( x ) · g′ ( x )
 
=
[ g( x )]2
g( x ) · f ′ ( x ) − f ( x ) · g′ ( x )
=
[ g( x )]2

Practice Answers
1. f is continuous at x = −1, 0, 2, 4, 6 and 7.
f is differentiable at x = −1, 2, 4, and 7.

2. f ( x ) = 6x + 8 and g( x ) = 2 so D( f ( x )) = 6 and D( g( x )) = 0.
D(3 · f ( x )) = 3 · D( f ( x )) = 3(6) = 18
D( f ( x ) + g( x )) = D( f ( x )) + D( g( x )) = 6 + 0 = 6
D( f ( x ) − g( x )) = D( f ( x )) − D( g( x )) = 6 − 0 = 6
· g( x )) = f ( x ) g′ ( x ) + g( x ) f ′ ( x ) = (6x + 8)(0) + (2)(6) = 12
D(f ( x ) 
f (x) g( x ) f ′ ( x )− f ( x ) g′ ( x ) (2)(6)−(6x +8)(0) 12
D = = = =3
g( x ) [ g( x )]2 22 4
 the derivative 143

3. D( x3 − 5 sin( x )) = D( x3 ) − 5 · D(sin( x )) = 3x2 − 5 cos( x )

d   d d 3
sin( x ) − 4x3 = sin( x ) − 4 · x = cos( x ) − 12x2
dx dx dx

D(3 f ( x )) D(2 f ( x ) + g( x )) D(3g( x ) − f ( x ))

4. −6 −1 11
−3 −2 1
6 5 1

   
f (x) g( x )
D( f ( x ) · g( x )) D D
g( x ) f (x)
−4(−2)−(3)(3) 1 (3)(3)−(−4)(−2)
5. 3 · 3 + (−4)(−2) = 17 (−4)2
= − 16 32
= 19
1(−1)−(2)(0) 2(0)−1(−1)
2 · 0 + 1(−1) = −1 12
= −1 22
= 14
3(2)−(4)(1) 4(1)−3(2)
4 · 1 + 3 · 2 = 10 32
= 29 42
= − 18

6. D(( x2 + 1)(7x − 3)) = ( x2 + 1) D(7x − 3) + (7x − 3) D( x2 + 1)

= ( x2 + 1)(7) + (7x − 3)(2x ) = 21x2 − 6x + 7
or: D(( x2 +1)(7x − 3)) = D(7x3 − 3x2 + 7x ) = 21x2 − 6x + 7
d 3t − 2 (5t + 1) D(3t − 2) − (3t − 2) D(5t + 1) (5t + 1)(3) − (3t − 2)(5) 13
= = =
dt  5t + 1  (5t + 1)2 (5t + 1)2 (5t + 1)2
cos( x ) x D(cos( x )) − cos( x ) D( x ) x (− sin( x )) − cos( x )(1) − x · sin( x ) − cos( x )
D = = =
x x2 x2 x2
7. Mimicking the proof of the Difference Rule:

[ f ( x + h) + g( x + h)] − [ f ( x ) + g( x )]
D( f ( x ) + g( x )) = lim
h →0 h
[ f ( x + h) − f ( x )] + [ g( x + h) − g( x )]
= lim
h →0 h
f ( x + h) − f ( x ) g( x + h) − g( x )
= lim + lim
h →0 h h →0 h
= D( f ( x )) + D( g( x ))

8. (a) difference rule (b) product rule (c) difference rule (d) quotient rule
  sin( x ) D( x2 − 5) − ( x2 − 5) D(sin( x )) sin( x )(2x ) − ( x2 − 5) cos( x )
9. D x2 − 5sin( x ) = 2
=
(sin( x )) sin2 ( x )
d 2
  t · sin(t) D(t2 − 5) − (t2 − 5) D(t · sin(t)) t · sin(t)(2t) − (t2 − 5) [t cos(t) + sin(t)]
t − 5t · sin(t) = 2
=
dt (t · sin(t)) t2 · sin2 (t)
sin(5) (1+5) cos(5)−sin(5)
10. h(5) = 5 + 1+5 ≈ 4.84 ft.; v(5) = h′ (5) = (1+5)2
≈ 0.074 ft/sec.
sin(t)
“long time”: h(t) = 5 + ≈ 5 feet when t is very large;
1+ t
(1 + t) cos(t) − sin(t) cos(t) sin(t)
h′ (t) = = − ≈ 0 ft/sec when t is very large.
(1 + t )2 1+t (1 + t )2
11. f ′ ( x ) = 2x − 10 so f ′ ( x ) = 0 ⇒ 2x − 10 = 0 ⇒ x = 5.
g′ ( x ) = 3x2 − 12 so g′ ( x ) = 0 ⇒ 3x2 − 12 = 0 ⇒ x2 = 4 ⇒ x = ±2.
144 contemporary calculus

2.3 More Differentiation Patterns

Polynomials are very useful, but they are not the only functions we
need. This section uses the ideas of the two previous sections to develop
techniques for differentiating powers of functions, and to determine the
derivatives of some particular functions that occur often in applications:
the trigonometric and exponential functions.
As you focus on learning how to differentiate different types and
combinations of functions, it is important to remember what derivatives
are and what they measure. Calculators and computers are available
to calculate derivatives. Part of your job as a professional will be to
decide which functions need to be differentiated and how to use the
resulting derivatives. You can succeed at that only if you understand
what a derivative is and what it measures.

A Power Rule for Functions: D( f n ( x ))

If we apply the Product Rule to the product of a function with itself, a
pattern emerges.

D( f 2 ) = D( f · f ) = f · D( f ) + f · D( f ) = = 2 f · D( f )
3 2 2 2 2
D( f ) = D( f · f ) = f · D( f ) + f · D( f ) = f · D( f ) + f · 2 f · D( f ) = 3 f 2 · D( f )
D( f 4 ) = D( f 3 · f ) = f 3 · D( f ) + f · D( f 3 ) = f 3 · D( f ) + f · 3 f 2 · D( f ) = 4 f 3 · D( f )

Practice 1. What is the pattern here? What do you think the results
will be for D( f 5 ) and D( f 13 )?
We could keep differentiating higher and higher powers of f ( x ) by
writing them as products of lower powers of f ( x ) and using the Product
Rule, but the Power Rule for Functions guarantees that the pattern we
just saw for the small integer powers also works for all constant powers
of functions.
The Power Rule for Functions is a spe- Power Rule for Functions:
cial case of a more general theorem, the
Chain Rule, which we will examine in If p is any constant
Section 2.4, so we will wait until then to
prove the Power Rule for Functions. then D( f p ( x )) = p · f p−1 ( x ) · D( f ( x )).

Example 1. Use the Power Rule for Functions to find:

Remember: sin2 ( x ) = [sin( x )]2

d p 
(a) D(( x3 − 5)2 ) (b) 2x + 3x5 (c) D(sin2 ( x ))
dx
Solution. (a) To match the pattern of the Power Rule for D(( x3 − 5)2 ),
let f ( x ) = x3 − 5 and p = 2. Then:
Check that you get the same answer by D(( x3 − 5)2 ) = D( f p ( x )) = p · f p−1 ( x ) · D( f ( x ))
first expanding ( x3 − 5)2 and then taking
the derivative. = 2( x3 − 5)1 D( x3 − 5) = 2( x3 − 5)(3x2 ) = 6x2 ( x3 − 5)
 the derivative 145

d p  d  1

(b) To match the pattern for 2x + 3x5 = (2x + 3x5 ) 2 , let
dx dx
f ( x ) = 2x + 3x5 and take p = 21 . Then:
d p  d d
2x + 3x5 = ( f p ( x )) = p · f p−1 ( x ) · ( f ( x ))
dx dx dx
1 1 d
= (2x + 3x5 )− 2 (2x + 3x5 )
2 dx
1 1 2 + 15x4
= (2x + 3x5 )− 2 (2 + 15x4 ) = √
2 2 2x + 3x5

(c) To match the pattern for D(sin2 ( x )), let f ( x ) = sin( x ) and p = 2:

D(sin2 ( x )) = D( f p ( x )) = p · f p−1 ( x ) · D( f ( x ))
= 2 sin1 ( x ) D(sin( x )) = 2 sin( x ) cos( x )

We could also rewrite this last expression as sin(2x ). ◀

Practice 2. Use the Power Rule for Functions to find:

d  5  p 
(a) (2x − π )2 (b) D x + 7x2 (c) D(cos4 ( x ))
dx
Example 2. Use calculus to show that the line tangent to the circle
x2 + y2 = 25 at the point (3, 4) has slope − 43 .
√
Solution. The top half of the circle is the graph of f ( x ) = 25 − x2 so:
 1
 1 1 −x
f ′ ( x ) = D (25 − x2 ) 2 ) = (25 − x2 )− 2 · D(25 − x2 ) = √
2 25 − x2
−3 3
and f ′ (3) = √ = − . As a check, you can verify that the slope
25 − 32 4
of the radial line through the center of the circle (0, 0) and the point
(3, 4) has slope 43 and is perpendicular to the tangent line that has a
slope of − 34 . ◀

Derivatives of Trigonometric Functions

We have some general rules that apply to any elementary combination
of differentiable functions, but in order to use the rules we still need to
know the derivatives of some basic functions. Here we will begin to
add to the list of functions whose derivatives we know.
We already know the derivatives of the sine and cosine functions, and
each of the other four trigonometric functions is just a ratio involving
sines or cosines. Using the Quotient Rule, we can easily differentiate
the rest of the trigonometric functions.

Theorem:
D(tan( x )) = sec2 ( x ) D(sec( x )) = sec( x ) · tan( x )
D(cot( x )) = − csc2 ( x ) D(csc( x )) = − csc( x ) · cot( x )
146 contemporary calculus

sin( x ) cos( x )
Proof. From trigonometry, we know tan( x ) = , cot( x ) = ,
cos( x ) sin( x )
1 1
sec( x ) = and csc( x ) = . From calculus, we already know
cos( x ) sin( x )
D(sin( x )) = cos( x ) and D(cos( x )) = − sin( x ). So:

cos( x ) · D(sin( x )) − sin( x ) · D(cos( x ))

 
sin( x )
D(tan( x )) = D =
cos( x ) (cos( x ))2
cos( x ) · cos( x ) − sin( x )(− sin( x ))
=
cos2 ( x )
cos2 ( x ) + sin2 ( x ) 1
= 2
= = sec2 ( x )
cos ( x ) cos2 ( x )

Similarly:

cos( x ) · D(1) − 1 · D(cos( x ))

 
1
D(sec( x )) = D =
cos( x ) (cos( x ))2
cos( x ) · 0 − (− sin( x ))
=
cos2 ( x )
sin( x ) 1 sin( x )
= 2
= · = sec( x ) · tan( x )
cos ( x ) cos( x ) cos( x )

Instead of the Quotient Rule, we could have used the Power Rule to
calculate D(sec( x )) = D((cos( x ))−1 ).
cos( x )
Practice 3. Use the Quotient Rule on f ( x ) = cot( x ) = to prove
sin( x )
that f ′ ( x ) = − csc2 ( x ).

Practice 4. Prove that D(csc( x )) = − csc( x ) · cot( x ). The justification

of this result is very similar to the justification for D(sec( x )).
 
5 d sec(t) p 
Practice 5. Find: (a) D( x tan( x )) (b) (c) D cot( x ) − x
dt t

Derivatives of Exponential Functions

We can estimate the value of a derivative of an exponential function (a
function of the form f ( x ) = a x where a > 0) by estimating the slope of
the line tangent to the graph of such a function, or we can numerically
approximate those slopes.

Example 3. Estimate the value of the derivative of f ( x ) = 2x at the

point (0, 20 ) = (0, 1) by approximating the slope of the line tangent to
f ( x ) = 2x at that point.

Solution. We can get estimates from the graph of f ( x ) = 2x by care-

fully graphing y = 2x for small values of x (so that x is near 0), sketch-
ing secant lines, and then measuring the slopes of the secant lines (see
margin figure).
 the derivative 147

We can also estimate the slope numerically by using the definition

of the derivative:
f (0 + h ) − f (0) 20+h − 20 2h − 1
f ′ (0) = lim = lim = lim
h →0 h h →0 h h →0 h
2h − 1
and evaluating for some very small values of h. From the table
h
below we can see that f ′ (0) ≈ 0.693. ◀

2h −1 3h −1 e h −1
h h h h

+0.1 0.717734625
−0.1 0.669670084
+0.01 0.695555006
−0.01 0.690750451
+0.001 0.693387463
−0.001 0.692907009
↓ ↓ ↓ ↓
0 ≈ 0.693 ≈ 1.099 1

3h − 1
Practice 6. Fill in the table for and show that the slope of the
h
x
line tangent to g( x ) = 3 at (0, 1) is approximately 1.099.

At (0, 1), the slope of the tangent to y = 2x is less than 1 and the
slope of the tangent to y = 3x is slightly greater than 1. You might
expect that there is a number b between 2 and 3 so that the slope of
the tangent to y = b x is exactly 1. Indeed, there is such a number,
e ≈ 2.71828182845904, with
eh − 1
lim =1
h →0 h In fact, e is a “transcendental” number,
which means that it is not the root of
The number e is irrational and plays a very important role in calculus
any polynomial equation with rational
and applications. coefficients.
We have not proved that this number e with the desired limit prop- Don’t worry — we’ll tie up some of these
loose ends in Chapter 7.
erty actually exists, but if we assume it does, then it becomes relatively
straightforward to calculate D(e x ).

Theorem: D(e x ) = e x

Proof. Using the definition of the derivative:

e x+h − e x e x · eh − e x
D(e x ) = lim = lim
h →0 h h →0 h
e h−1 eh − 1
= lim e x · = lim e x · lim
h →0 h h →0 h →0 h
= ex · 1 = ex

The function f ( x ) = e x is its own derivative: f ′ ( x ) = f ( x ).

148 contemporary calculus

Notice that the limit property of e that Graphically: the height of f ( x ) = e x at any point and the slope of
we assumed was true actually says that
the tangent to f ( x ) = e x at that point are the same: as the graph gets
for f ( x ) = e x , f ′ (0) = 1. So knowing
the derivative of f ( x ) = e x at a single higher, its slope gets steeper.
point (x = 0) allows us to determine its  x 
derivative at every other point. d t e
(c) D(e5x )


Example 4. Find: (a) t · e (b) D
dt sin( x )

Solution. (a) Using the Product Rule with f (t) = t and g(t) = et :

d
t · e t = t · D ( e t ) + e t · D ( t ) = t · e t + e t · 1 = ( t + 1) e t


dt
(b) Using the Quotient Rule with f ( x ) = e x and g( x ) = sin( x ):
 x 
e sin( x ) · D(e x ) − e x · D(sin( x ))
D =
sin( x ) [sin( x )]2
sin( x ) · e x − e x (cos( x ))
=
sin2 ( x )

(c) Using the Power Rule for Functions with f ( x ) = e x and p = 5:

D((e x )5 ) = 5(e x )4 · D(e x ) = 5e4x · e x = 5e5x

where we have rewritten e5x as (e x )5 . ◀

Practice 7. Find: (a) D( x3 e x ) (b) D((e x )3 ).

Higher Derivatives: Derivatives of Derivatives

The derivative of a function f is a new function f ′ and if this new
function is differentiable we can calculate the derivative of this new
function to get the derivative of the derivative of f , denoted by f ′′ and
called the second derivative of f .
For example, if f ( x ) = x5 then f ′ ( x ) = 5x4 and:

f ′′ ( x ) = ( f ′ ( x ))′ = (5x4 )′ = 20x3

Definitions: Given a differentiable function f ,

• the first derivative is f ′ ( x ), the rate of change of f .

• the second derivative is f ′′ ( x ) = ( f ′ ( x ))′ , the rate of change of f ′ .

• the third derivative is f ′′′ ( x ) = ( f ′′ ( x ))′ , the rate of change of f ′′ .

dy
For y = f ( x ), we write f ′ ( x ) = , so we can extend that notation to
dx
 
d dy 2
d y ′′′ d d2 y
 
d3 y
′′
write f ( x ) = = 2 , f (x) = = and so on.
dx dx dx dx dx2 dx3
 the derivative 149

Practice 8. Find f ′ , f ′′ and f ′′′ for f ( x ) = 3x7 , f ( x ) = sin( x ) and

f ( x ) = x · cos( x ).

If f ( x ) represents the position of a particle at time x, then v( x ) =

f ′ (x)will represent the velocity (rate of change of the position) of the
particle and a( x ) = v′ ( x ) = f ′′ ( x ) will represent the acceleration (the
rate of change of the velocity) of the particle.

Example 5. The height (in feet) of a particle at time t seconds is given

by t3 − 4t2 + 8t. Find the height, velocity and acceleration of the particle
when t = 0, 1 and 2 seconds.

Solution. f (t) = t3 − 4t2 + 8t so f (0) = 0 feet, f (1) = 5 feet and

f (2) = 8 feet. The velocity is given by v(t) = f ′ (t) = 3t2 − 8t + 8
so v(0) = 8 ft/sec, v(1) = 3 ft/sec and v(2) = 4 ft/sec. At each of
these times the velocity is positive and the particle is moving upward
(increasing in height). The acceleration is a(t) = 6t − 8 so a(0) = −8
ft/sec2 , a(1) = −2 ft/sec2 and a(2) = 4 ft/sec2 . ◀

We will examine the geometric (graphical) meaning of the second

derivative in the next chapter.

A Really “Bent” Function

In Section 1.2 we saw that the “holey” function
(
2 if x is a rational number
h( x ) =
1 if x is an irrational number

is discontinuous at every value of x, so h( x ) is not differentiable any-

where. We can create graphs of continuous functions that are not
differentiable at several places just by putting corners at those places,
but how many corners can a continuous function have? How badly can
a continuous function fail to be differentiable?
In the mid-1800s, the German mathematician Karl Weierstrass sur-
prised and even shocked the mathematical world by creating a function
that was continuous everywhere but differentiable nowhere — a func-
tion whose graph was everywhere connected and everywhere bent! He
used techniques we have not investigated yet, but we can begin to see
how such a function could be built.
Start with a function f 1 (see margin) that zigzags between the values
1 1
2 and − 2 and has a “corner” at each integer. This starting function f 1
is continuous everywhere and is differentiable everywhere except at
the integers. Next create a list of functions f 2 , f 3 , f 4 , . . . , each of which
is “shorter” than the previous one but with many more “corners” than
the previous one. For example, we might make f 2 zigzag between the
150 contemporary calculus

values 14 and − 14 and have “corners” at ± 12 , ± 32 , ± 52 , etc.; f 3 zigzag

between 19 and − 19 and have “corners” at ± 13 , ± 23 , ± 33 = ±1, etc.
If we add f 1 and f 2 , we get a continuous function (because the sum of
two continuous functions is continuous) with corners at 0, ± 12 , ±1, ± 32 ,
. . . . If we then add f 3 to the previous sum, we get a new continuous
function with even more corners. If we continue adding the functions
in our list “indefinitely,” the final result will be a continuous function
that is differentiable nowhere.
We haven’t developed enough mathematics here to precisely describe
what it means to add an infinite number of functions together or
to verify that the resulting function is nowhere differentiable — but
we will. You can at least start to imagine what a strange, totally
“bent” function it must be. Until Weierstrass created his “everywhere
continuous, nowhere differentiable” function, most mathematicians
thought a continuous function could only be “bad” in a few places.
Weierstrass’ function was (and is) considered “pathological,” a great
example of how bad something can be. The mathematician Charles
Hermite expressed a reaction shared by many when they first encounter
the Weierstrass function: “I turn away with fright and horror from this
lamentable evil of functions which do not have derivatives.”

Important Results
Power Rule for Functions: D( f p ( x )) = p · f p−1 ( x ) · D( f ( x ))

Derivatives of the Trigonometric Functions:

D(sin( x )) = cos( x ) D(cos( x )) = − sin( x )

D(tan( x )) = sec2 ( x ) D(cot( x )) = − csc2 ( x )
D(sec( x )) = sec( x ) tan( x ) D(csc( x )) = − csc( x ) cot( x )

Derivative of the Exponential Function: D(e x ) = e x

2.3 Problems

1. Let f (1) = 2 and f ′ (1) = 3. Find the values of 2. Let f (2) = −2 and f ′ (2) = 5. Find the values of
each of the following derivatives at x = 1. each of the following derivatives at x = 2.
(a) D( f 2 ( x )) (a) D( f 2 ( x ))
(b) D( f 5 ( x )) (b) D( f −3 ( x ))
p p
(c) D( f ( x )) (c) D( f ( x ))
 the derivative 151

3. For x = 1 and x = 3 estimate the values of f ( x ) 12. A mass attached to a spring is at a height of
(whose graph appears below), f ′ ( x ) and 2 sin(t)
h(t) = 3 − feet above the floor t sec-
d  2  1 + 0.1t2
(b) D f 3 ( x ) (c) D f 5 ( x )



(a) f (x) onds after it is released.
dx
(a) Graph h(t).
(b) At what height is the mass when it is released?
(c) Determine the velocity of the mass at time t.
(d) What happens to the height and the velocity
of the mass a “long time” after it is released?
13. The kinetic energy K of an object of mass m and
velocity v is 12 mv2 .
(a) Find the kinetic energy of an object with mass
m and height h(t) = 5t feet at t = 1 and t = 2
seconds.
(b) Find the kinetic energy of an object with mass
4. For x = 0 and x = 2 estimate the values of f ( x )
m and height h(t) = t2 feet at t = 1 and t = 2
(whose graph appears above), f ′ ( x ) and
seconds.
d  3  d  5 
(a) D f 2 ( x )


(b) f (x) (c) f (x) 14. An object of mass m is attached to a spring and
dx dx
In Problems 5–10, find f ′ ( x ). has height h(t) = 3 + sin(t) feet at time t seconds.
(a) Find the height and kinetic energy of the object
5. f ( x ) = (2x − 8)5
when t = 1, 2 and 3 seconds.
6. f ( x ) = (6x − x2 )10 (b) Find the rate of change in the kinetic energy of
7. f ( x ) = x · (3x + 7)5 the object when t = 1, 2 and 3 seconds.
dK
8. f ( x ) = (2x + 3)6 · ( x − 2)4 (c) Can K ever be negative? Can ever be nega-
√ dt
9. f ( x ) = x2 + 6x − 1 tive? Why?
x−5 In Problems 15–20, compute f ′ ( x ).
10. f ( x ) =
( x + 3)4
15. f ( x ) = x · sin( x )
11. A mass attached to the end of a spring is at a
height of h(t) = 3 − 2 sin(t) feet above the floor t 16. f ( x ) = sin5 ( x )
seconds after it is released. 17. f ( x ) = e x − sec( x )
p
(a) Graph h(t). 18. f ( x ) = cos( x ) + 1
(b) At what height is the mass when it is released? 19. f ( x ) = e− x + sin( x )
√
(c) How high does above the floor and how close 20. f ( x ) = x2 − 4x + 3
to the floor does the mass ever get?
In Problems 21–26, find an equation for the line
(d) Determine the height, velocity and acceleration
tangent to the graph of y = f ( x ) at the given point.
at time t. (Be sure to include the correct units.)
(e) Why is this an unrealistic model of the motion 21. f ( x ) = ( x − 5)7 at (4, −1)
of a mass attached to a real spring? 22. f ( x ) = e x at (0, 1)
 152 contemporary calculus

√
23. f ( x ) = 25 − x2 at (3, 4) 43. The function f ( x ) defined as
3
24. f ( x ) = sin ( x ) at (π, 0) (
x · sin( 1x ) if x ̸= 0
25. f ( x ) = ( x − a)5 at ( a, 0) f (x) =
0 if x = 0
26. f ( x ) = x · cos5 ( x ) at (0, 0)
shown below is continuous at 0 because we can
27. (a) Find an equation for the line tangent to f ( x ) =
show (using the Squeezing Theorem) that
e x at the point (3, e3 ).
(b) Where will this tangent line intersect the x- lim f ( x ) = 0 = f (0)
h →0
axis?
(c) Where will the tangent line to f ( x ) = e x at the Is f differentiable at 0? To answer this question,
point ( p, e p ) intersect the x-axis? use the definition of f ′ (0) and consider

In Problems 28–33, calculate f ′ and f ′′ . f (0 + h ) − f (0)

lim
h →0 h
28. f ( x ) = 7x2 + 5x − 3
29. f ( x ) = cos( x )
30. f ( x ) = sin( x )
31. f ( x ) = x2 · sin( x )
32. f ( x ) = x · sin( x )
33. f ( x ) = e x · cos( x )
34. Calculate the first 8 derivatives of f ( x ) = sin( x ).
What is the pattern? What is the 208th derivative
of sin( x )?
35. What will the second derivative of a quadratic
polynomial be? The third derivative? The fourth
derivative? 44. The function f ( x ) defined as
36. What will the third derivative of a cubic polyno- (
x2 · sin( 1x ) if x ̸= 0
mial be? The fourth derivative? f (x) =
0 if x = 0
37. What can you say about the n-th and (n + 1)-st
derivatives of a polynomial of degree n? (shown at the top of the next page) is continuous
at 0 because we can show (using the Squeezing
In Problems 38–42, you are given f ′ . Find a function
Theorem) that
f with the given derivative.
38. f ′ ( x ) = 4x + 2 lim f ( x ) = 0 = f (0)
h →0
39. f ′ ( x ) = 5e x
Is f differentiable at 0? To answer this question,
40. f ′ ( x ) = 3 · sin2 ( x ) · cos( x ) use the definition of f ′ (0) and consider
41. f ′ ( x ) = 5(1 + e x )4 · e x
f (0 + h ) − f (0)
42. f ′ ( x ) = e x + sin( x ) lim
h →0 h
 the derivative 153

47. Define n! to be the product of all positive integers

from 1 through n. For example, 2! = 1 · 2 = 2,
3! = 1 · 2 · 3 = 6 and 4! = 1 · 2 · 3 · 4 = 24.
(a) Calculate the value of the sums:

1
s1 = 1 +
1!
1 1
s2 = 1 + +
1! 2!
1 1 1
s3 = 1 + + +
1! 2! 3!
1 1 1 1
s4 = 1 + + + +
1! 2! 3! 4!
The number e appears in a variety of unusual situa- 1 1 1 1 1
s5 = 1 + + + + +
tions. Problems 45–48 illustrate a few of these. 1! 2! 3! 4! 5!
1 1 1 1 1 1
s6 = 1 + + + + + +
45. Use your  calculator to examine the values of 1! 2! 3! 4! 5! 6!
1 x

f (x) = 1 + when x is relatively large (for (b) What value do the sums in part (a) seem to be
x
example, x = 100, 1000 and 10, 000. Try some approaching?
other large values for x. If x is large, the value of (c) Calculate s7 and s8 .
f ( x ) is close to what number? 48. If it is late at night and you are tired of study-
46. If you put $1 into a bank account that pays 1% ing calculus, try the following experiment with
interest per year and compounds the interest x a friend. Take the 2 through 10 of hearts from a
times a year, then after one year you will have regular deck of cards and shuffle these nine cards
  x
1+ 0.01
dollars in the account. well. Have your friend do the same with the 2
x
through 10 of spades. Now compare your cards
(a) How much money will you have after one year
one at a time. If there is a match, for example you
if the bank calculates the interest once a year?
both play a 5, then the game is over and you win.
(b) How much money will you have after one year If you make it through the entire nine cards with
if the bank calculates the interest twice a year? no match, then your friend wins. If you play the
(c) How much money will you have after one year game many times, then the ratio:
if the bank calculates the interest 365 times a
year? total number of games played
number of times your friend wins
(d) How does your answer to part (c) compare
with e0.01 ? will be approximately equal to e.

2.3 Practice Answers

1. The pattern is D( f n ( x )) = n · f n−1 ( x ) · D( f ( x )):
D( f 5 ( x )) = 5 f 4 ( x ) · D( f ( x )) and D( f 13 ( x )) = 13 f 12 ( x ) · D( f ( x ))
d
2. (2x5 − π )2 = 2(2x5 − π )1 D(2x5 − π ) = 2(2x5 − π )(10x4 ) = 40x9 − 20πx4
dx
1
 1 1 1 + 14x
D ( x + 7x2 ) 2 = ( x + 7x2 )− 2 D( x + 7x2 ) = √
  2 2 x + 7x2
D (cos( x )) = 4(cos( x )) D(cos( x )) = 4(cos( x ))3 (− sin( x )) = −4 cos3 ( x ) sin( x )
4 3
154 contemporary calculus

3. Mimicking the proof for the derivative of tan( x ):

sin( x ) · D(cos( x )) − cos( x ) · D(sin( x ))
 
cos( x )
D =
sin( x ) (sin( x ))2
sin( x )(− sin( x )) − cos( x )(cos( x ))
=
sin2 ( x )
−(sin2 ( x ) + cos2 ( x )) −1
= 2
= = − csc2 ( x )
sin ( x ) sin2 ( x )

4. Mimicking the proof for the derivative of sec( x ):

sin( x ) · D(1) − 1 · D(sin( x )
 
1
D(csc( x )) = D =
sin( x ) sin2 ( x )
sin( x ) · 0 − cos( x ) 1 cos( x )
= 2
=− · = − cot( x ) csc( x )
sin ( x ) sin ( x ) sin( x )

5. D( x5 · tan( x )) = x5 D(tan( x )) + tan( x ) D( x5 ) = x5 sec2 ( x ) + tan( x )(5x4 )

t D(sec(t)) − sec(t) D(t) t sec(t) tan(t) − sec(t)

 
d sec(t)
= 2
=
dt t t t2
 1
 1 1
D (cot( x ) − x ) 2 = (cot( x ) − x )− 2 D(cot( x ) − x )
2
1 1 − csc2 ( x ) − 1
= (cot( x ) − x )− 2 (− csc2 ( x ) − 1) = p
2 2 cot( x ) − x

6. Filling in values for both 3x and e x :

2h −1 3h −1 e h −1
h h h h

+0.1 0.717734625 1.161231740 1.0517091808

−0.1 0.669670084 1.040415402 0.9516258196
+0.01 0.695555006 1.104669194 1.0050167084
−0.01 0.690750451 1.092599583 0.9950166251
+0.001 0.693387463 1.099215984 1.0005001667
−0.001 0.692907009 1.098009035 0.9995001666
↓ ↓ ↓ ↓
0 ≈ 0.693 ≈ 1.0986 1

7. D(x3 e x ) 
= x3 D(e x ) + e x D( x3 ) = x3 e x + e x · 3x2 = x2 e x ( x + 3)
D (e x )3 = 3 (e x )2 D(e x ) = 3e2x · e x = 3e3x

8. f ( x ) = 3x7 ⇒ f ′ ( x ) = 21x6 ⇒ f ′′ ( x ) = 126x5 ⇒ f ′′′ ( x ) = 630x4

f ( x ) = sin( x ) ⇒ f ′ ( x ) = cos( x ) ⇒ f ′′ ( x ) = − sin( x )
⇒ f ′′′ ( x ) = − cos( x )
f ( x ) = x · cos( x ) ⇒ f ′ ( x ) = − x sin( x ) + cos( x )
⇒ f ′′ ( x ) = − x cos( x ) − 2 sin( x ) ⇒ f ′′′ ( x ) = x sin( x ) − 3 cos( x )
 the derivative 155

2.4 The Chain Rule

The Chain Rule is the most important and most often used of the
differentiation patterns. It enables us to differentiate composites of
functions such as y = sin( x2 ). It is a powerful tool for determining
the derivatives of some new functions such as logarithms and inverse
trigonometric functions. And it leads to important applications in a
variety of fields. You will need the Chain Rule hundreds of times in
this course. Practice with it now will save you time — and points —
later. Fortunately, with practice, the Chain Rule is also easy to use. We
already know how to differentiate the composition of some functions.

Example 1. For f ( x ) = 5x − 4 and g( x ) = 2x + 1, find f ◦ g( x ) and

D( f ◦ g( x )).

Solution. Writing f ◦ g( x ) = f ( g( x )) = 5(2x + 1) − 4 = 10x + 1, we

can compute that D( f ◦ g( x )) = D(10x + 1) = 10. ◀

Practice 1. For f ( x ) = 5x − 4 and g( x ) = x2 , find f ◦ g( x ), D( f ◦ g( x )),

g ◦ f ( x ) and D( g ◦ f ( x )).

Some compositions, however, are still very difficult to differentiate.

We know the derivatives of g( x ) = x2 and h( x ) = sin( x ), and we know
how to differentiate certain combinations of these functions, such as
x2 + sin( x ), x2 · sin( x ) and even sin2 ( x ) = (sin( x ))2 . But the derivative
of the simple composition f ( x ) = h ◦ g( x ) = sin( x2 ) is hard — until we To see just how difficult, try using the
definition of derivative on it.
know the Chain Rule.

Example 2. (a) Suppose amplifier Y doubles the strength of the output

signal from amplifier U, and U triples the strength of the original
signal x. How does the final signal out of Y compare with the
original signal x?

(b) Suppose y changes twice as fast as u, and u changes three times as

fast as x. How does the rate of change of y compare with the rate of
change of x?

Solution. In each case we are comparing the result of a composition,

and the answer to each question is 6, the product of the two amplifica-
tions or rates of change. In part (a), we have that:

signal out of Y signal out of Y signal out of U

= · = 2·3 = 6
signal x signal out of U signal x

In part (b):
∆y ∆y ∆u
= · = 2·3 = 6
∆x ∆u ∆x
These examples are simple cases of the Chain Rule for differentiating a
composition of functions. ◀
156 contemporary calculus

The Chain Rule

We can express the chain rule using more than one type of notation.
Each will be useful in various situations.

Chain Rule (Leibniz notation form):

If y is a differentiable function of u and
u is a differentiable function of x
then y is a differentiable function of x and
dy dy du
= · .
dx du dx

Idea for a proof. If ∆u ̸= 0 then:

∆y ∆y ∆u ∆y ∆u
  
dy
= lim = lim · = lim lim
dx ∆x →0 ∆x ∆x →0 ∆u ∆x ∆x →0 ∆u ∆x →0 ∆x
∆y ∆u
  
dy du
= lim lim = ·
∆u→0 ∆u ∆x →0 ∆x du dx
The key step here is to argue that ∆x → 0 implies ∆u → 0, which
follows from the continuity of u as as function of x.
Although this nice short argument gets to the heart of why the Chain
Rule works, it is not quite valid. If du
dx ̸ = 0, then it is possible to show
A justification that holds true for all cases that ∆u ̸= 0 for all “very small” values of ∆x, and the “idea for a
is more complicated and provides no new
proof” becomes a real proof. There are, however, functions for which
conceptual insight. Problem 84 at the
end of this section guides you through a ∆u = 0 for infinitely many small values of ∆x (no matter how close to
rigorous proof of the Chain Rule. 0 we restrict ∆x) and this creates problems with the simple argument
outlined above.
dy du
The symbol is a single symbol, as is , so we cannot eliminate
du dx
dy du
du from the product in the Chain Rule by “cancelling” du as we
du dx
∆y ∆u
can with ∆u in the fractions · . It is, however, perfectly fine to
∆u ∆x
use the idea of cancelling du to help you remember the proper statement
of the Chain Rule.

Example 3. Write y = cos( x2 + 3) as y = cos(u) with u = x2 + 3 and

dy
find dx .
dy du
Solution. y = cos(u) ⇒ = − sin(u) and u = x2 + 3 ⇒ dx = 2x.
du
Using the Chain Rule:

dy dy du
= · = − sin(u) · 2x = −2x · sin( x2 + 3)
dx du dx
Notice that in the last step we have eliminated the intermediate variable
u to express the derivative only in terms of x. ◀
dy
Practice 2. Find dx for y = sin(4x + e x ).
 the derivative 157

We can also state the Chain Rule in terms of composition of functions.

The notation is different, but the meaning is precisely the same.

Chain Rule (composition form):

You may find it easier to think of the re-
If g is differentiable at x and sult of the composition form of the Chain
f is differentiable at g( x ) Rule in words: “the derivative of the out-
side function (evaluated at the original
then the composite f ◦ g is differentiable at x and
inside function) times the derivative of
( f ◦ g)′ ( x ) = D( f ( g( x ))) = f ′ ( g( x )) · g′ ( x ). the inside function” where f is the out-
side function and g is the inside function.

Example 4. Differentiate sin( x2 ).

Solution. We can write the function sin( x2 ) as the composition f ◦ g

of two simple functions: f ( x ) = sin( x ) and g( x ) = x2 : f ◦ g( x ) =
f ( g( x )) = f ( x2 ) = sin( x2 ). Both f and g are differentiable functions
with derivatives f ′ ( x ) = cos( x ) and g′ ( x ) = 2x, so the Chain Rule says:
If you tried using the definition of deriva-
D(sin( x2 )) = ( f ◦ g)′ ( x ) = f ′ ( g( x )) · g′ ( x ) = cos( g( x )) · 2x tive to calculate the derivative of this
function at the beginning of this section,
= cos( x2 ) · 2x = 2x cos( x2 )
you can really appreciate the power of
the Chain Rule for differentiating com-
Check that you get the same answer using the Leibniz notation. ◀ positions of functions, even simple ones
like these.
Example 5. The table below gives values for f , f ′ , g and g′ at various
points. Use these values to determine ( f ◦ g)( x ) and ( f ◦ g)′ ( x ) at
x = −1 and x = 0.

x f (x) g( x ) f ′ (x) g′ ( x ) ( f ◦ g)( x ) ( f ◦ g)′ ( x )

−1 2 3 1 0
0 −1 1 3 2
1 1 0 −1 3
2 3 −1 0 1
3 0 2 2 −1

Solution. ( f ◦ g)(−1) = f ( g(−1)) = f (3) = 0, ( f ◦ g)(0) = f ( g(0)) =

f (1) = 1, ( f ◦ g)′ (−1) = f ′ ( g(−1)) · g′ (−1) = f ′ (3) · 0 = 2 · 0 = 0 and
( f ◦ g)′ (0) = f ′ ( g(0)) · g′ (0) = f ′ (1) · 2 = (−1)(2) = −2. ◀

Practice 3. Fill in the table in Example 5 for ( f ◦ g)( x ) and ( f ◦ g)′ ( x )

at x = 1, 2 and 3.

Neither form of the Chain Rule is inherently superior to the other —

use the one you prefer or the one that appears most useful in a particular
situation. The Chain Rule will be used hundreds of times in the rest of
this book, and it is important that you master its usage. The time you
spend now mastering and understanding how to use the Chain Rule
will be paid back tenfold over the next several chapters.
158 contemporary calculus

 
Example 6. Determine D ecos( x) using each form of the Chain Rule.

Solution. Using the Leibniz notation: y = eu and u = cos( x ) so we

dy
have du = eu and du
dx = − sin( x ). Applying the Chain Rule:

dy dy du
= · = eu · (− sin( x )) = − sin( x ) · ecos(x)
dx du dx

We can also write the function ecos( x) as the composition of f ( x ) = e x

with g( x ) = cos( x ), so the Chain Rule says:

D(ecos( x) ) = f ′ ( g( x )) · g′ ( x ) = e g( x) · (− sin( x )) = − sin( x ) · ecos( x)

because D(e x ) = e x and D (cos( x )) = − sin( x ). ◀

d d  3t 
Practice 4. Calculate D (sin(7x − 1)), (sin( ax + b)) and e .
dx dt
Practice 5. Use the graph of g given in the margin along with the Chain
Rule to estimate D (sin( g( x ))) and D ( g(sin( x ))) at x = π.

The Chain Rule is a general differentiation pattern that can be used

along with other general patterns like the Product and Quotient Rules.

d
Example 7. Determine D e3x · sin(5x + 7) and (cos( x · e x )).


dx

Solution. The function e3x sin(5x + 7) is a product of two functions so

we need the Product Rule first:

D(e3x · sin(5x + 7)) = e3x · D(sin(5x + 7)) + sin(5x + 7) · D(e3x )

= e3x · cos(5x + 7) · 5 + sin(5x + 7) · e3x · 3
= 5e3x cos(5x + 7) + 3e3x sin(5x + 7)

The function cos( x · e x ) is a composition of cosine with a product so we

need the Chain Rule first:
d d
(cos( x · e x )) = − sin( x · e x ) · ( x · e x )
dx  dx 
x d x x d
= − sin( xe ) · x · (e ) + e · (x)
dx dx
= − sin( xe x ) · ( xe x + e x )

We could also write this last answer as −( x + 1)e x sin(e x ). ◀

Sometimes we want to differentiate a composition of more than two

functions. We can do so if we proceed in a careful, step-by-step way.
√
Example 8. Find D(sin( x3 + 1)).
 the derivative 159

√
Solution. The function sin( x3 + 1) can be viewed as a composition
√
f ◦ g of f ( x ) = sin( x ) and g( x ) = x3 + 1. Then:
p
(sin( x3 + 1))′ = f ′ ( g( x )) · g′ ( x ) = cos( g( x )) · g′ ( x )
p p
= cos( x3 + 1) · D( x3 + 1)
√
For the derivative of x3 + 1, we can use the Chain Rule again or its
special case, the Power Rule:
p 1 1 3 1
D( x3 + 1) = D(( x3 + 1) 2 ) = ( x + 1) − 2 · D ( x 3 + 1)
2
1 3 1
=( x + 1)− 2 · 3x2
2
 p  p 1 1
Finally, D sin( x3 + 1) = cos( x3 + 1) · ( x3 + 1)− 2 · 3x2 , which
√ 2
3x2 cos( x3 + 1)
can be rewritten as √ . ◀
2 x3 + 1
This example was more complicated than the earlier ones, but it is
just a matter of applying the Chain Rule twice, to a composition of
a composition. If you proceed step by step and don’t get lost in the
details of the problem, these multiple applications of the Chain Rule
are relatively straightforward.
We can also use the Leibniz form of the Chain Rule for a composition
√
of more than two functions. If y = sin( x3 + 1), then y = sin(u) with
√
u = w and w = x3 + 1. The Leibniz form of the Chain Rule says:

dy dy du dw 1
= · · = cos(u) · √ · 3x2
dx du dw dx 2 w
p 1
= cos( x3 + 1) · √ · 3x2
2 x3 + 1
which agrees with our previous answer.
dy
Practice 6. (a) Find D(sin(cos(5x ))). (b) For y = ecos(3x) , find dx .

The Chain Rule and Tables of Derivatives

With the Chain Rule, the derivatives of all sorts of strange and won-
derful functions become available. If we know f ′ and g′ , then we also
know the derivatives of their composition: ( f ( g( x ))′ = f ′ ( g( x )) · g′ ( x ).
We have begun to build a list of derivatives of “basic” functions,
such as x n , sin( x ) and e x . We will continue to add to that list later
in the course, but if we peek ahead at the rest of that list — spoiler
1
alert! — to (for example) see that D(arctan( x )) = , then we can
1 + x2
use the Chain Rule to compute derivatives of compositions of those
functions.
160 contemporary calculus

1
Example 9. Given that D(arcsin( x )) = √ , compute the deriva-
1 − x2
d
tives D(arcsin(5x )) and (arcsin(e x )).
dx
Solution. Write arcsin(5x ) as the composition of f ( x ) = arcsin( x )
1
with g( x ) = 5x. We know g′ ( x ) = 5 and f ′ ( x ) = √ , so we have
1 − x2
1 1
f ′ ( g( x )) = p = √ . Then:
1 − ( g( x ))2 1 − 25x2
1 5
D(arcsin(5x )) = f ′ ( g( x )) · g′ ( x ) = p ·5 = √
1 − (5x )2 1 − 25x2
We can write y = arcsin(e x ) as y = arcsin(u) with u = e x , and we know
dy 1 du
that = √ and = e x so:
du 1−u 2 dx
dy dy du 1 ex
= · = √ · ex = √
dx du dx 1 − u2 1 − e2x
f ′ (x)
We can generalize this result to say that D(arcsin( f ( x ))) = p
1 − ( f ( x ))2
d 1 du
or, in Leibniz notation, (arcsin(u)) = √ · . ◀
du 1 − u dx
2

1
Practice 7. Given that D(arctan( x )) = , compute the derivatives
1 + x2
d
D(arctan( x3 )) and (arctan(e x )).
dx
Appendix D in the back of this book shows the derivative patterns
for a variety of functions. You may not know much about some of the
functions, but with the given differentiation patterns and the Chain
Rule you should be able to calculate derivatives of compositions that
involve these new functions. It is just a matter of following the pattern.
Practice 8. Use the patterns D(sinh( x )) = cosh( x ) and D(ln( x )) = 1x
to determine:
d  
(a) D(sinh(5x − 7)) (b) ln(3 + e2x ) (c) D(arcsin(1 + 3x ))
dx
d  3 
Example 10. If D( F ( x )) = e x · sin( x ), find D( F (5x )) and F (t ) .
dt
Solution. D( F (5x )) = D( F ( g( x )) with g( x ) = 5x and we know that
F ′ ( x ) = e x · sin( x ) so:
D( F (5x )) = F ′ ( g( x )) · g′ ( x ) = e g( x) · sin( g( x )) · 5 = e5x · sin(5x ) · 5
dy du
With y = F (u) and u = t3 we know = eu · sin(u) and = 3t2 so:
du dt
dy dy du 3
= · = eu · sin(u) · 3t2 = et · sin(t3 ) · 3t2
dt du dt
Notice that we have eliminated the intermediate variable u (which
didn’t appear in the original problem) from the final answer. ◀
 the derivative 161

Proof of the Power Rule For Functions

We started using the Power Rule For Functions in Section 2.3. Now we
can easily prove it.

Power Rule For Functions:

If p is any constant
then D( f p ( x )) = p · f p−1 ( x ) · D( f ( x )).

dy
Proof. Write y = f p ( x ) as y = u p with u = f ( x ). Then = p · u p −1
du
du
and = f ′ ( x ) so:
dx
dy dy du
= · = p · u p −1 · f ′ ( x ) = p · f p −1 ( x ) · f ′ ( x )
dx du dx
by the Chain Rule.

2.4 Problems
In Problems 1–6 , find two functions f and g so that the given function is the composition of f and g.

2. y = sin4 (3x − 8)
p
1. y = ( x3 − 7x )5 3. y = (2 + sin( x ))5

1 √
5. y =  x2 − 4
 
4. y = √ 6. y = tan( x )
x2 + 9

7. For each function in Problems 1–6, write y as a function of u for some u that is a function of x.
For Problems 8–9, use the values given in this table to determine the indicated quantities:

x f (x) g( x ) f ′ (x) g′ ( x ) ( f ◦ g)( x ) ( f ◦ g)′ ( x )

x f (x) g( x ) f ′ (x) g′ ( x ) ( f ◦ g)( x ) ( f ◦ g)′ ( x )
−2 2 −1 1 1
−1 1 2 0 2
0 −2 1 2 −1
1 0 −2 −1 2
2 1 0 1 −1

8. ( f ◦ g)( x ) and ( f ◦ g)′ ( x ) at x = 1 and x = 2.

9. ( f ◦ g)( x ) and ( f ◦ g)′ ( x ) at x = −2, −1 and 0.

10. Using the figure in the margin, estimate the values of g( x ), g′ ( x ),

( f ◦ g)( x ), f ′ ( g( x )) and ( f ◦ g)′ ( x ) at x = 1.

11. Using the figure in the margin, estimate the values of g( x ), g′ ( x ),

( f ◦ g)( x ), f ′ ( g( x )) and ( f ◦ g)′ ( x ) at x = 2.
 162 contemporary calculus

In Problems 12–22, compute the derivative. (d) If you are in a balloon that is 2,000 feet above
4 ! the Pacific Ocean and is rising at 500 feet per
12. D ( x2 + 2x + 3)87


3 minute, how fast is the air pressure on the
13. D 1−
x balloon changing?
 5 ! (e) If the temperature of the gas in the balloon
d 1 5
14. x+ 15. D remained constant during this ascent, what
dx x
p
2 + sin( x ) would happen to the volume of the balloon?
d d  2 
Find the indicated derivatives in Problems 26–33.
16. (t · sin(3t + 2)) 17. x · sin( x2 + 3)
dt dx
(2x + 3)2 d
  q
d 26. D 27. 1 + cos2 (z)
18. (sin(2x ) · cos(5x + 1)) (5x − 7)3 dz
dx

7

d

5

28. D (sin(3x + 5)) d
19. D 20. 29. tan(3x + 5)
cos( x3 − x ) dt 3 + et dx
√
d


cos(7t2 ) 31. D sin( x + 1)
21. D (e x + e− x ) 22. D (e x − e− x ) 30.
dt
√
d  sin( x) 


23. An object attached to a spring is at a height of 32. D sec( x + 1) 33. e
dx
h(t) = 3 − cos(2t) feet above the floor t seconds
after it is released. In Problems 34–37 , calculate f ′ ( x ) · x ′ (t) when t = 3
(a) At what height was it released? and use these values to determine the value of
d
(b) Determine its height, velocity and acceleration ( f ( x (t))) when t = 3.
dt
at any time t.
34. f ( x ) = cos( x ), x = t2 − t + 5
(c) If the object has mass m, determine its kinetic √ 21
energy K = 12 mv2 and dK 35. f ( x ) = x, x = 2 +
dt at any time t. t
24. An employee with d days of production expe- 36. f ( x ) = e x , x = sin(t)
rience will be able to produce approximately 37. f ( x ) = tan3 ( x ), x = 8
P(d) = 3 + 15(1 − e−0.2d ) items per day. In 38–43, find a function that has the given function
(a) Graph P(d). as its derivative. (You are given a function f ′ ( x ) and
(b) Approximately how many items will a begin- are asked to find a corresponding function f ( x ).)
ning employee be able to produce each day?
38. f ′ ( x ) = (3x + 1)4 39. f ′ ( x ) = (7x − 13)10
(c) How many items will a very experienced em-
ployee be able to produce each day? √
40. f ′ ( x ) = 3x − 4 41. f ′ ( x ) = sin(2x − 3)
(d) What is the marginal production rate of an
employee with 5 days of experience? (Include 42. f ′ ( x ) = 6e3x 43. f ′ ( x ) = cos( x )esin( x)
units for your answer. What does this mean?)
If two functions are equal, then their derivatives are
25. The air pressure P(h), in pounds per square inch,
also equal. In 44–47 , differentiate each side of the
at an altitude of h feet above sea level is approxi-
trigonometric identity to get a new identity.
mately P(h) = 14.7e−0.0000385h .
1
(a) What is the air pressure at sea level? 44. sin2 ( x ) = 2 − 21 cos(2x )
(b) What is the air pressure at 30,000 feet? 45. cos(2x ) = cos2 ( x ) − sin2 ( x )
(c) At what altitude is the air pressure 10 pounds 46. sin(2x ) = 2 sin( x ) · cos( x )
per square inch? 47. sin(3x ) = 3 sin( x ) − 4 sin3 ( x )
 the derivative 163

Derivatives of Families of Functions use that pattern to compute derivatives of associated

So far we have emphasized derivatives of particular composite functions.
functions, but sometimes we want to investigate the
73. D arctan( x2 )


derivatives of a whole family of functions all at once. 72. D (arctan(7x ))
In 48–71, A, B, C and D represent constants and the
d d
given formulas describe families of functions. 74. (arctan(ln(t))) 75. (arctan(e x ))
dt dx
dy d d 
For Problems 48–65, calculate y′ =

dx . 76. (arcsin(4w)) 77. arcsin( x3 )
dw dx
48. y = Ax3 − B 49. y = Ax3 + Bx2 + C
79. D arcsin(et )


78. D (arcsin(ln( x )))

50. y = sin( Ax + B) 51. y = sin( Ax2 + B)

80. D (ln(3x + 1)) 81. D (ln(sin( x )))
√
52. y = Ax3 + cos( Bx ) 53. y = A + Bx2 d d
82. (ln(arctan( x ))) 83. (ln(es ))
√ dx ds
54. y = A − Bx2 55. y = A − cos( Bx )
84. To prove the Chain Rule, assume g( x ) is differ-
56. y = cos( Ax + B) 57. y = cos( Ax2 + B) entiable at x = a and f ( x ) is differentiable at
x = g( a). We need to show that
58. y = A · e Bx 59. y = x · e Bx
f ( g( x )) − f ( g( a))
lim
60. y = e Ax + e− Ax 61. y = e Ax − e− Ax x→a x−a
exists and is equal to f ′ ( g( a)) · g′ ( a). To do this,
sin( Ax ) Ax
62. y = 63. y = define a new function F as:
x sin( Bx )
( f (y)− f ( g(a))
1 Ax + B y− g( a)
if y ̸= g( a)
64. y = 65. y = F (y) =
′
f ( g( a)) if y = g( a)
Ax + B Cx + D

In 66–71, (a) find y′ (b) find the value(s) of x so that and justify each of the following statements.
y′ = 0 and (c) find y′′ . Typically your answer in part (a) F (y) is continuous at y = g( a) because:
(b) will contain A’s, B’s and (sometimes) C’s.
f (y) − f ( g( a))
lim F (y) = lim = F ( g( a))
66. y = Ax2 + Bx + C y→ g( a) y→ g( a) y − g( a)
67. y = Ax ( B − x ) = ABx − Ax2
(b) By considering separately the cases g( x ) =
68. y = Ax ( B − x2 ) = ABx − Ax3 g( a) and g( x ) ̸= g( a):
69. y = Ax2 ( B − x ) = ABx2 − Ax3 f ( g( x )) − f ( g( a)) g( x ) − g( a)
= F ( g( x )) ·
70. y = Ax2 + Bx x−a x−a
71. y = Ax3 + Bx2 + C for all x ̸= a.
f ( g( x )) − f ( g( a)) g( x ) − g( a)
In Problems 72–83, use the differentiation patterns (c) lim = lim F ( g( x )) ·
1 1 x→a x−a x→a x−a
D(arctan( x )) = , D(arcsin( x )) = √ g( x ) − g( a)
1 + x2 1 − x2 (d) lim F ( g( x )) · = F ( g( a)) · g′ ( a)
1 x→a x−a
and D(ln( x )) = . We have not derived the deriva- f ( g( x )) − f ( g( a))
x (e) lim = f ′ ( g( a)) · g′ ( a)
tives for these functions (yet), but if you are handed x→a x−a
the derivative pattern then you should be able to (f) ( f ◦ g)′ ( a) = f ′ ( g( a)) · g′ ( a)
164 contemporary calculus

2.4 Practice Answers

1. f ( x ) = 5x − 4 and g( x ) = x2 ⇒ f ′ ( x ) = 5 and g′ ( x ) = 2x, so
f ◦ g( x ) = f ( g( x )) = f ( x2 )) = 5x2 − 4 and D(5x2 − 4) = 10x or:

D( f ◦ g( x )) = f ′ ( g( x )) · g′ ( x ) = 5 · 2x = 10x

g ◦ f ( x ) = g( f ( x )) = g(5x − 4) = (5x − 4)2 = 25x2 − 40x + 16 and

D(25x2 − 40x + 16) = 50x − 40 or:

D( g ◦ f ( x )) = g′ ( f ( x )) · f ′ ( x ) = 2(5x − 4) · 5 = 50x − 40

d
2. (sin(4x + e x )) = cos(4x + e x ) · D(4x + e x ) = cos(4x + e x ) · (4 + e x )
dx
3. To fill in the last column, compute:

f ′ ( g(1)) · g′ (1) = f ′ (0) · 3 = (3)(3) = 9

f ′ ( g(2)) · g′ (2) = f ′ (−1) · 1 = (1)(1) = 1
f ′ ( g(3) · g′ (3) = f ′ (2) · (−1) = (0)(−1) = 0

x f (x) g( x ) f ′ (x) g′ ( x ) ( f ◦ g)( x ) ( f ◦ g)′ ( x )

1 1 0 −1 3 −1 9
2 3 −1 0 1 2 3
3 0 2 2 −1 3 0

4. D (sin(7x − 1)) = cos(7x − 1) · D(7x − 1) = 7 · cos(7x − 1)

d
(sin( ax + b)) = cos( ax + b) · D( ax + b) = a · cos( ax + b)
dx  
d d
e3t = e3t · (3t) = 3 · e3t
dt dt
5. D (sin( g( x ))) = cos( g( x )) · g′ ( x ). At x = π, cos( g(π )) · g′ (π ) ≈
cos(0.86) · (−1) ≈ −0.65. D ( g(sin( x ))) = g′ (sin( x )) · cos( x ). At
x = π, g′ (sin(π )) · cos(π ) = g′ (0) · (−1) ≈ −2

6. D (sin(cos(5x ))) = cos(cos(5x )) · D(cos(5x ))

= cos(cos(5x )) · (− sin(5x )) · D(5x ) = −5 · sin(5x ) · cos(cos(5x ))
d  cos(3x) 
e = ecos(3x) · D(cos(3x )) = ecos(3x) (− sin(3x )) D(3x )
dx
= −3 · sin(3x ) · ecos(3x)
  1 3x2
7. D arctan( x3 ) = · D ( x 3
) =
1 + ( x 3 )2 1 + x6
d 1 ex
(arctan(e x )) = · D ( e x
) =
dx 1 + ( e x )2 1 + e2x
8. D(sinh(5x − 7)) = cosh(5x − 7) · D(5x − 7) = 5 · cosh(5x − 7)
d   1 2e2x
ln(3 + e2x = · D ( 3 + e 2x
) =
dx 3 + e2x 3 + e2x
1 3
D(arcsin(1 + 3x )) = p · D(1 + 3x ) = p
1 − (1 + 3x ) 2 1 − (1 + 3x )2
 the derivative 165

2.5 Applications of the Chain Rule

The Chain Rule can help us determine the derivatives of logarithmic

functions like f ( x ) = ln( x ) and general exponential functions like
f ( x ) = a x . We will also use it to answer some applied questions and to
find slopes of graphs given by parametric equations.

Derivatives of Logarithms
You know from precalculus that the natural logarithm ln( x ) is defined
as the inverse of the exponential function e x : eln( x) = x for x > 0.
We can use this identity along with the Chain Rule to determine the
derivative of the natural logarithm.

1 g′ ( x )
D(ln( x )) = and D (ln( g( x ))) =
x g( x )

u u
Proof.
 We  know that D(e ) = e , so using the Chain Rule we have
D e f ( x) = e f ( x) · f ′ ( x ). Differentiating each side of the identity
eln( x) = x, we get:
 
D eln( x) = D( x ) ⇒ eln( x) · D(ln( x )) = 1
1
⇒ x · D(ln( x )) = 1 ⇒ D(ln( x )) =
x
The function ln( g( x )) is the composition of f ( x ) = ln( x ) with g( x ) so
the Chain Rule says:
You can remember the differentiation
1 g′ ( x ) pattern for the the natural logarithm in
D (ln( g( x )) = D ( f ( g( x ))) = f ′ ( g( x )) · g′ ( x ) = · g′ ( x ) = words as: “one over the inside times the
g( x ) g( x ) the derivative of the inside.”

1
Graph f ( x ) = ln( x ) along with f ′ ( x ) =
and compare the behavior of
x
the function at various points with the values of its derivative at those
1
points. Does y = possess the properties you would expect to see
x
from the derivative of f ( x ) = ln( x )?

Example 1. Find D(ln(sin( x ))) and D(ln( x2 + 3)).

g′ ( x )
Solution. Using the pattern D(ln( g( x )) = with g( x ) = sin( x ):
g( x )

g′ ( x ) D(sin( x )) cos( x )
D(ln(sin( x ))) = = = = cot( x )
g( x ) sin( x ) sin( x )

g′ ( x ) 2x
With g( x ) = x2 + 3, D(ln( x2 + 3)) = = 2 . ◀
g( x ) x +3
166 contemporary calculus

We can use the Change of Base Formula from precalculus to rewrite

any logarithm as a natural logarithm, and then we can differentiate the
resulting natural logarithm.

Change of Base Formula for Logarithms:

logb ( x )
loga ( x ) = for all positive a, b and x.
logb ( a)
Your calculator likely has two logarithm
buttons: ln for the natural logarithm Example 2. Use the Change of Base formula and your calculator to
(base e) and log for the common log-
arithm (base 10). Be careful, however, find logπ (7) and log2 (8).
as more advanced mathematics texts (as
ln(7) 1.946
well as the Web site Wolfram|Alpha) use Solution. logπ (7) = ≈ ≈ 1.700. (Check that π 1.7 ≈ 7.)
log for the (base e) natural logarithm. ln(π ) 1.145
ln(8)
Likewise, log2 (8) = = 3. ◀
ln(2)
Practice 1. Find the values of log9 20, log3 20 and logπ e.
loge ( x )
Putting b = e in the Change of Base Formula, loga ( x ) = =
loge ( a)
ln( x )
, so any logarithm can be written as a natural logarithm divided
ln( a)
by a constant. This makes any logarithmic function easy to differentiate.

1 f ′ (x) 1
D (loga ( x )) = and D (loga ( f ( x ))) = ·
x ln( a) f ( x ) ln( a)
 
ln x 1 1 1 1
Proof. D (loga ( x )) = D = · D(ln x ) = · = .
ln a ln( a) ln( a) x x ln( a)
The second differentiation formula follows from the Chain Rule.

Practice 2. Calculate D log10 (sin( x )) and D (logπ (e x )).

The number e might seem like an “unnatural” base for a natural

logarithm, but of all the possible bases, the logarithm with base e has
the nicest and easiest derivative. The natural logarithm is even related
to the distribution of prime numbers. In 1896, the mathematicians
Hadamard and Vallée-Poussin proved the following conjecture of Gauss
(the Prime Number Theorem): For large values of N,
N
number of primes less than N ≈
ln( N )

Derivative of a x
Once we know the derivative of e x and the Chain Rule, it is relatively
easy to determine the derivative of a x for any a > 0.

D( a x ) = a x · ln( a) for a > 0.

 the derivative 167

x
Proof. If a > x
 a >
 0, then
x
 0 and a = e
ln( a )
= e x·ln(a) , so we have:
x
D( a x ) = D eln(a ) = D e x·ln(a) = e x·ln(a) · D( x · ln( a)) = a x · ln( a).

d  sin(t) 
Example 3. Calculate D(7x ) and 2 .
dt
Solution. D(7x ) = 7x · ln(7) ≈ (1.95)7x . We can write y = 2sin(t) as
dy dy du
y = 2u with u = sin(t). Using the Chain Rule: = · =
dt du dt
2u · ln(2) cos(t) = 2sin(t) · ln(2) · cos(t). ◀
d  2

Practice 3. Calculate D (sin(2x )) and 3t .
dt

Some Applied Problems

Let’s examine some applications involving more complicated functions.

Example 4. A ball at the end of a rubber band (see margin) is oscillating

up and down, anditsheight (in feet) above the floor at time t seconds
t
is h(t) = 5 + 2 sin (with t in radians).
2
(a) How fast is the ball traveling after 2 seconds? After 4 seconds? After
60 seconds?

(b) Is the ball moving up or down after 2 seconds? After 4 seconds?

After 60 seconds?

(c) Is the vertical velocity of the ball ever 0?

    
t t 1
Solution. (a) v(t) = h′ (t) = D 5 + 2 sin = 2 cos · so
2 2 2
v(t) = cos 2t feet/second: v(2) = cos( 22 ) ≈ 0.540 ft/s, v(4) =

cos( 42 ) ≈ −0.416 ft/s, and v(60) = cos( 602 ) ≈ 0.154 ft/s.

(b) The ball is moving up at t = 2 and t = 60, down when t = 4.

 
t t π
(c) v(t) = cos = 0 when = ± k · π ⇒ t = π ± 2πk for any
2 2 2
integer k. ◀

Example 5. If 2,400 people now have a disease, and the number of

people with the disease appears to double every 3 years, then the
t
number of people expected to have the disease in t years is y = 2400 · 2 3 .

(a) How many people are expected to have the disease in 2 years?

(b) When are 50,000 people expected to have the disease?

(c) How fast is the number of people with the disease growing now?
How fast is it expected to be growing 2 years from now?
168 contemporary calculus

2
Solution. (a) In 2 years, y = 2400 · 2 3 ≈ 3,810 people.
t
(b) We know y = 50000 and need to solve 50000 = 2400 · 2 3 for
t. Taking logarithms of each side of the equation: ln(50000) =
 2
 t
ln 2400 · 2 3 = ln(2400) + · ln(2) so 10.819 ≈ 7.783 + 0.231t and
3
t ≈ 13.14 years. We expect 50,000 people to have the disease about
13 years from now.
dy
(c) This question asks for when t = 0 and t = 2.
dt
dy d  t
 t 1 t
= 2400 · 2 3 = 2400 · 2 3 · ln(2) · ≈ 554.5 · 2 3
dt dt 3
Now, at t = 0, the rate of growth of the disease is approximately
554.5 · 20 ≈ 554.5 people/year. In 2 years, the rate of growth will be
2
approximately 554.5 · 2 3 ≈ 880 people/year. ◀

Example 6. You are riding in a balloon, and at time t (in minutes) you
are h(t) = t + sin(t) thousand feet above sea level. If the temperature
72
at an elevation h is T (h) = degrees Fahrenheit, then how fast is
1+h
the temperature changing when t = 5 minutes?

Solution. As t changes, your elevation will change. And, as your

elevation changes, so will the temperature. It is not difficult to write
the temperature as a function of time, and then we could calculate
dT
= T ′ (t) and evaluate T ′ (5). Or we could use the Chain Rule:
dt
dT dT dh 72
= · =− · (1 + cos(t))
dt dh dt (1 + h )2

At t = 5, h(5) = 5 + sin(5) ≈ 4.04 so T ′ (5) ≈ − (1+72

4.04)2
· (1 + 0.284) ≈
◦
−3.64 /minute. ◀

Practice 4. Write the temperature T in the previous example as a

function of the variable t alone and then differentiate T to determine
dT
the value of when t = 5 minutes.
dt
Example 7. A scientist has determined that, under optimum conditions,
an initial population of 40 bacteria will grow “exponentially” to f (t) =
t
40 · e 5 bacteria after t hours.

(a) Graph y = f (t) for 0 ≤ t ≤ 15. Calculate f (0), f (5) and f (10).

(b) How fast is the population increasing at time t? (Find f ′ (t).)

(c) Show that the rate of population increase, f ′ (t), is proportional to

the population, f (t), at any time t. (Show f ′ (t) = K · f (t) for some
constant K.)
 the derivative 169

Solution. (a) The graph of y = f (t) appears in the margin. f (0) =

0 5
40 · e 5 = 40 bacteria, f (5) = 40 · e 5 = 40e ≈ 109 bacteria and
10
f (10) = 40 · e 5 ≈ 296 bacteria.
 t
 t t t
(b) f ′ (t) = dt
d
( f (t)) = dtd 40 · e 5 = 40 · e 5 · dtd 5t = 40 · e 5 · 15 = 8 · e 5

bacteria/hour.
t t
(c) f ′ (t) = 8 · e 5 = 51 · 40e 5 = 15 f (t) so f ′ (t) = K · f (t) with K = 15 . The
rate of change of the population is proportional to its size. ◀

Parametric Equations
Suppose a robot has been programmed to move in the xy-plane so at
time t its x-coordinate will be sin(t) and its y-coordinate will be t2 . Both
x and y are functions of the independent parameter t: x (t) = sin(t) and
y(t) = t2 . The path of the robot (see margin) can be found by plotting
( x, y) = ( x (t), y(t)) for lots of values of t.

t x (t) = sin(t) y ( t ) = t2 point

0 0 0 (0, 0)
0.5 0.48 0.25 (0.48, 0.25)
1.0 0.84 1 (0.84, 1)
1.5 1.00 2.25 (1, 2.25)
2.0 0.91 4 (0.91, 4)

dy
Typically we know x (t) and y(t) and need to find , the slope of
dx
the tangent line to the graph of ( x (t), y(t)). The Chain Rule says:
dy dy dx
= ·
dt dx dt
dy
so , algebraically solving for , we get:
dx
dy
dy dt
= dx
dx
dt
dy dx
If we can calculate and , the derivatives of y and x with respect
dt dt
dy
to the parameter t, then we can determine , the rate of change of y
dx
with respect to x.

If x = x (t) and y = y(t) are differentiable

dx
with respect to t and ̸= 0
dt
dy
dy dt
then = dx .
dx
dt
170 contemporary calculus

Example 8. Find the slope of the tangent line to the graph of ( x, y) =

sin(t), t2 when t = 2.

dx dy
Solution. = cos(t) and = 2t. When t = 2, the object is at the
dt dt
2
point (sin(2), 2 ) ≈ (0.91, 4) and the slope of the tangent line is:
dy
dy dt 2t 2·2 4
= dx
= = ≈ ≈ −9.61
dx
dt
cos(t) cos(2) −0.42

Notice in the figure that the slope of the tangent line to the curve at
(0.91, 4) is negative and very steep. ◀

Practice 5. Graph ( x, y) = (3 cos(t), 2 sin(t)) and find the slope of the

tangent line when t = π2 .
dy
When we calculated , the slope of the tangent line to the graph of
dx
dx dy
( x (t), y(t)), we used the derivatives and . Each of these also has a
dt dt
dx
geometric meaning: measures the rate of change of x (t) with respect
dt
to t: it tells us whether the x-coordinate is increasing or decreasing as
dy
the t-variable increases (and how fast it is changing), while measures
dt
the rate of change of y(t) with respect to t.

Example 9. For the parametric graph in the margin, determine whether

dx dy dy
, and are positive or negative when t = 2.
dt dt dx
Solution. As we move through the point B (where t = 2) in the di-
rection of increasing values of t, we are moving to the left, so x (t) is
dx dy
decreasing and < 0. The values of y(t) are increasing, so > 0.
dt dt
dy
Finally, the slope of the tangent line, , is negative. ◀
dx
dy
As a check on the sign of in the previous example:
dx
dy
dy dt positive
= dx
= = negative
dx negative
dt

Practice 6. For the parametric graph in the previous example, tell

dx dy dy
whether , and are positive or negative at t = 1 and t = 3.
dt dt dx

Speed
If we know the position of an object at any time, then we can determine
its speed. The formula for speed comes from the distance formula and
looks a lot like it, but involves derivatives.
 the derivative 171

If x = x (t) and y = y(t) give the location of an object

at time t and both are differentiable functions of t
then the speed of the object is
s
 2  2
dx dy
+
dt dt

Proof. The speed of an object is the limit, as ∆t → 0, of (see margin):

s
(∆x )2 + (∆y)2 (∆x )2 + (∆y)2
p
change in position
= =
change in time ∆t (∆t)2
s s
∆x 2
 2  2  2
∆y
 
dx dy
= + → +
∆t ∆t dt dt

as ∆t → 0.

Example 10. Find the speed of the object whose location at time t is
( x, y) = sin(t), t2 when t = 0 and t = 1.

dx dy
Solution. = cos(t) and = 2t so:
dt dt
q q
speed = (cos(t))2 + (2t)2 = cos2 (t) + 4t2
p √
When t = 0, speed = cos2 (0) + 4(0)2 = 1 + 0 = 1. When t = 1,
p √
speed = cos2 (1) + 4(1)2 ≈ 0.29 + 4 ≈ 2.07. ◀

Practice 7. Show that an object located at ( x, y) = (3 sin(t), 3 cos(t)) at

time t has a constant speed. (This object is moving on a circular path.)

Practice 8. Is the object at ( x, y) = (3 cos(t), 2 sin(t)) at time t traveling

faster at the top of the ellipse (t = π2 ) or at the right edge (t = 0)?

2.5 Problems

In Problems 1–27, differentiate the given function. 9. ln(sin( x )) 10. ln(kx )

1. ln(5x ) 2. ln( x2 ) 11. log2 (sin( x )) 12. ln(e x )

 
3. ln( x k ) 4. ln( x x ) = x · ln( x ) 13. log5 (5x ) 14. ln e f ( x)

5. ln(cos( x )) 6. cos(ln( x )) 15. x · ln(3x ) 16. e x · ln( x )

q
ln( x )
7. log2 (5x ) 8. log2 (kx ) 17. 18. x + ln(3x )
x
 172 contemporary calculus

√

19. ln 5x − 3 20. ln(cos(t)) In 33–41, find a function with the given derivative.

21. cos(ln(w)) 22. ln( ax + b) 8 3

33. f ′ ( x ) = 34. h′ ( x ) =
√ x 3x + 5
24. 3x


23. ln t+1
cos( x ) x
35. f ′ ( x ) = 36. g′ ( x ) =
25. 5sin( x) 26. x · ln( x ) − x 3 + sin( x ) 1 + x2

27. ln (sec( x ) + tan( x )) 37. g′ ( x ) = 3e5x 38. h′ ( x ) = e2

28. Find the slope of the line tangent to f ( x ) = ln( x ) 2

39. f ′ ( x ) = 2x · e x 40. g′ ( x ) = cos( x )esin( x)
at the point (e, 1). Find the slope of the line tan-
gent to g( x ) = e x at the point (1, e). How are the cos( x )
41. h′ ( x ) = cot( x ) =
slopes of f and g at these points related? sin( x )
29. Find a point P on the graph of f ( x ) = ln( x ) so 42. Define A( x ) to be the area bounded between the
the tangent line to f at P goes through the origin. t-axis, the graph of y = f (t) and a vertical line
at t = x (see figure below). The area under each
30. You are moving from left to right along the graph
“hump” of f is 2 square inches.
of y = ln( x ) (see figure below).
(a) If the x-coordinate of your location at time t (a) Graph A( x ) for 0 ≤ x ≤ 9.
seconds is x (t) = 3t + 2, then how fast is your (b) Graph A′ ( x ) for 0 ≤ x ≤ 9.
elevation increasing?
(b) If the x-coordinate of your location at time t
seconds is x (t) = et , then how fast is your
elevation increasing?

Problems 43–48 involve parametric equations.

43. At time t minutes, robot A is at (t, 2t + 1) and
robot B is at (t2 , 2t2 + 1).
(a) Where is each robot when t = 0 and t = 1?
31. The percent of a population, p(t), who have
(b) Sketch the path each robot follows during the
heard a rumor by time t is often modeled by
100
−1 first minute.
p(t) = = 100 1 + Ae−t for some dy
1 + Ae−t (c) Find the slope of the tangent line, , to the
′
positive constant A. Calculate p (t), the rate at dx
path of each robot at t = 1 minute.
which the rumor is spreading.
(d) Find the speed of each robot at t = 1 minute.
32. If we start with A atoms of a radioactive material
(e) Discuss the motion of a robot that follows the
that has a “half-life” (the time it takes for half
path (sin(t), 2 sin(t) + 1) for 20 minutes.
of the material to decay) of 500 years, then the
number of radioactive atoms left after t years is 44. Let x (t) = t + 1 and y(t) = t2 .
ln(2) (a) Graph ( x (t), y(t)) for −1 ≤ t ≤ 4.
r (t) = A · e−Kt where K = . Calculate r ′ (t)
500 dx dy dy
and show that r ′ (t) is proportional to r (t) (that is, (b) Find , , the tangent slope , and speed
dt dt dx
r ′ (t) = b · r (t) for some constant b). when t = 1 and t = 4.
 the derivative 173

45. For the parametric graph shown below, determine 50. After t seconds, a projectile hurled with initial ve-
dx dy dy locity v and angle θ will be at x (t) = v · cos(θ ) · t
whether , and are positive, negative or
dt dt dx feet and y(t) = v · sin(θ ) · t − 16t2 feet (see figure
0 when t = 1 and t = 3.
below). (This formula neglects air resistance.)
(a) For an initial velocity of 80 feet/second and an
angle of π4 , find T > 0 so that y( T ) = 0. What
does this value for t represent physically? Eval-
uate x ( T ).
dy
(b) For v and θ in part (a), calculate . Find T so
dx
dy
that = 0 at t = T, and evaluate x ( T ). What
dx
does x ( T ) represent physically?
46. For the parametric graph shown below, determine (c) What initial velocity is needed so a ball hit at
dx dy dy an angle of π4 ≈ 0.7854 will go over a 40-foot-
whether , and are positive, negative or
dt dt dx high fence 350 feet away?
0 when t = 1 and t = 3.
(d) What initial velocity is needed so a ball hit at
an angle of 0.7 radians will go over a 40-foot-
high fence 350 feet away?

47. The parametric graph ( x (t), y(t)) defined by 51. Use the method from the proof that D(ln( x )) = 1
x
x (t) = R · (t − sin(t)) and y(t) = R · (1 − cos(t)) to compute the derivative D(arctan( x )):
is called a cycloid, the path of a light attached to
(a) Rewrite y = arctan( x ) as tan(y) = x.
the edge of a rolling wheel with radius R.
(b) Differentiate both sides using the Chain Rule
(a) Graph ( x (t), y(t)) for 0 ≤ t ≤ 4π.
dx dy dy and solve for y′ .
(b) Find , , the tangent slope , and speed
dt dt dx (c) Use the identity 1 + tan2 (θ ) = sec2 (θ ) and the
π
when t = 2 and t = π. 1
fact that tan(y) = x to show that y′ = .
48. Describe the motion of particles whose locations 1 + x2
at time t are (cos(t), sin(t)) and (cos(t), − sin(t)). 1
52. Use the method from the proof that D(ln( x )) = x
49. (a) Describe the path of a robot whose location at to compute the derivative D(arcsin( x )):
time t is (3 · cos(t), 5 · sin(t)). (a) Rewrite y = arcsin( x ) as sin(y) = x.
(b) Describe the path of a robot whose location at (b) Differentiate both sides using the Chain Rule
time t is ( A · cos(t), B · sin(t)). and solve for y′ .
(c) Give parametric equations so the robot will
(c) Use the identity cos2 (θ ) + sin2 (θ ) = 1 and the
move along the same path as in part (a) but in 1
the opposite direction. fact that sin(y) = x to show that y′ = √ .
1 − x2
174 contemporary calculus

2.5 Practice Answers

log(20) ln(20)
1. log9 (20) = ≈ 1.3634165 ≈
log(9) ln(9)
log(20) ln(20)
log3 (20) = ≈ 2.726833 ≈
log(3) ln(3)
log(e) ln(e) 1
logπ (e) = ≈ 0.8735685 ≈ =
log(π ) ln(π ) ln(π )

1 cos( x )
2. D log10 (sin( x )) = D(sin( x )) =
sin( x ) · ln(10) sin( x ) · ln(10)
1 ex 1
D (logπ (e x )) = x D(e x ) = x =
e · ln(π ) e · ln(π ) ln(π )
3. D (sin(2x )) = cos(2x ) D (2x ) = cos(2x ) · 2x · ln(2)
d  t2  2 2
3 = 3t ln(3) D(t2 ) = 3t ln(3) · 2t
dt
72 72
4. T = = ⇒
1+h 1 + t + sin(t)
dT (1 + t + sin(t)) · 0 − 72 · D(1 + t + sin(t)) −72(1 + cos(t))
= =
dt (1 + t + sin(t))2 (1 + t + sin(t))2
dT −72(1 + cos(5))
When t = 5, = ≈ −3.63695.
dt (1 + 5 + sin(5))2
dx
5. x (t) = 3 cos(t) ⇒ = −3 sin(t), y(t) = 2 sin(t) ⇒ dy
dt = 2 cos( t ):
dt
dy
2 cos( π2 )

dy dt 2 cos(t) dy  2·0
= dx
= ⇒ = π = =0
dx
dt
−3 sin(t) dx t= π
 −3 sin( 2 ) −3 · 1
2

(See margin for graph.)

6. x = 1: positive, positive, positive. x = 3: positive, negative, negative.

dy
7. x (t) = 3 sin(t) ⇒ dx
dt = 3 cos( t ) and y ( t ) = 3 cos( t ) ⇒ dt =
−3 sin(t). So:
s
 2  2 q
dx dy
speed = + = (3 cos(t))2 + (−3 sin(t))2
dt dt
q √
= 9 · cos2 (t) + 9 · sin2 (t) = 9 = 3 (a constant)

dy
8. x (t) = 3 cos(t) ⇒ dx dt = −3 sin( t ) and y ( t ) = 2 sin( t ) ⇒ dt =
2 cos(t) so:
s
 2  2 q
dx dy
speed = + = (−3 sin(t))2 + (2 cos(t))2
dt dt
q
= 9 · sin2 (t) + 4 · cos2 (t)
√
When t = 0, the speed is 9 · 02 + 4 · 12 = 2.
√
When t = π2 , the speed is 9 · 12 + 4 · 02 = 3 (faster).
 the derivative 175

2.6 Related Rates

Throughout the next several sections we’ll look at a variety of applica-

tions of derivatives. Probably no single application will be of interest
or use to everyone, but at least some of them should be useful to you.
Applications also reinforce what you have been practicing: they require
that you recall what a derivative means and require you to use the
differentiation techniques covered in the last several sections. Most
people gain a deeper understanding and appreciation of a tool as they
use it, and differentiation is both a powerful concept and a useful tool.

The Derivative as a Rate of Change

In Section 2.1, we discussed several interpretations of the derivative
of a function. Here we will examine the “rate of change of a function”
interpretation. If several variables or quantities are related to each other
and some of the variables are changing at a known rate, then we can
use derivatives to determine how rapidly the other variables must be
changing.

Example 1. The radius of a circle is increasing at a rate of 10 feet each

second (see margin figure) and we want to know how fast the area of
the circle is increasing when the radius is 5 feet. What can we do?

Solution. We could get an approximate answer by calculating the area

of the circle when the radius is 5 feet:

A = πr2 = π (5 feet)2 ≈ 78.6 feet2

and the area 1 second later when the radius is 10 feet larger than before:

A = πr2 = π (15 feet)2 ≈ 706.9 feet2

and then computing:

∆area 706.9 feet2 − 78.6 feet2 ft2

= = 628.3
∆time 1 second sec
This approximate answer represents the average change in area during
the 1-second period when the radius increased from 5 feet to 15 feet. It
is also the slope of the secant line through the points P and Q in the
margin figure, and it is clearly not a very good approximation of the
instantaneous rate of change of the area, the slope of the tangent line at
the point P.
∆A
We could get a better approximation by calculating over a shorter
∆t
time interval, say ∆t = 0.1 seconds. In this scenario, the original area
176 contemporary calculus

is still 78.6 ft2 but the new area (after t = 0.1 seconds has passed) is
A = π (6 feet)2 ≈ 113.1 ft2 (why is the new radius 6 feet?) so:

∆A 113.1 feet2 − 78.6 feet2 ft2

= = 345
∆t 0.1 second sec
This is the slope of the secant line through the points P and Q in the
margin figure, which represents a much better approximation of the
slope of the tangent line at P — but it is still only an approximation.
Using derivatives, we can get an exact answer without doing very much
work at all.
We know that the two variables in this problem, the radius r and
the area A, are related to each other by the formula A = πr2 . We also
know that both r and A are changing over time, so each of them is a
function of an additional variable t (time, in seconds): r (t) and A(t).
We want to know the rate of change of the area “when the radius is
5 feet” so if t = 0 corresponds to the particular moment in time when
the radius is 5 feet, we can write r (0) = 5.
The statement that “the radius is increasing at a rate of 10 feet each
second” can be translated into a mathematical statement about the
rate of change, the derivative of r (radius) with respect to t (time):
if t = 0 corresponds to the moment when the radius is 5 feet, then
dr
r ′ (0) = = 10 ft/sec.
dt
The question about the rate of change of the area is a question about
dA
A′ (t) = .
dt
Collecting all of this information. . .

• variables: r (t) = radius at time t, A(t) = area at time t

• we know: r (0) = 5 feet and r ′ (0) = 10 ft/sec

• we want to know: A′ (0), the rate of change of area with respect to

time at the moment when r = 5 feet

• connecting equation: A = πr2 or A(t) = π [r (t)]2

To find A′ (0) we must first find A′ (t) and then evaluate this derivative
at t = 0. Differentiating both sides of the connecting equation, we get:

A(t) = π [r (t)]2 ⇒ A′ (t) = 2π [r (t)]1 · r ′ (t) ⇒ A′ (t) = 2π · r (t) · r ′ (t)

Notice that we have used the Power Rule Now we can plug in t = 0 and use the information we know:
for Functions (or, more generally, the
Chain Rule) because the area is a func-
A′ (0) = 2π · r (0) · r ′ (0) = 2π · 5 · 10 = 100π
tion of the radius, which is a function of
time.
When the radius is 5 feet, the area is increasing at 100π ft2 /sec ≈ 314.2
square feet per second. ◀
 the derivative 177

Before considering other examples, let’s review the solution to the We should take care in future problems to
consider whether the information we are
previous example. The statement “the radius is increasing at a rate of given about rates of change holds true all
10 feet each second” implies that this rate of change is the same at t = 0 the time or just at a particular moment
(the moment in time we were interested in) as at any other time during in time. That didn’t matter in our first
example, but it might in other situations.
this process, say t = 1.5 or t = 98: r ′ (0) = r ′ (1.5) = r ′ (98) = 10. But
we only used the fact that r ′ (0) = 10 in our solution.
Next, notice that we let t = 0 correspond to the particular moment
in time the question asked about (the moment when r = 5). But this
choice was arbitrary: we could have let this moment correspond to
t = 2.8 or t = 7π and the eventual answer would have been the same.
Finally, notice that we explicitly wrote each variable (and their deriva-
tives) as a function of the time variable, t: A(t), r (t), A′ (t) and r ′ (t).
Consequently, we used the composition form of the Chain Rule:

( A ◦ r )′ (t) = A′ (r (t)) · r ′ (t)

Let’s redo the previous example using the Leibniz form of the Chain
Rule, keeping the above observations in mind.

Solution. We know that the two variables in this problem, the radius r
and the area A, are related to each other by the formula A = πr2 . We
also know that both r and A are changing over time, so each of them is
a function of an additional variable t (time, in seconds).
We want to know the rate of change of the area “when the radius is
dA
5 feet,” which translates to evaluating at the moment when r = 5.
dt
We write this in Leibniz notation as:

dA 
dt  r =5

The statement that “the radius is increasing at a rate of 10 feet each

dr
second” translates into = 10. From the connecting equation A = πr2
dt
dA
we know that = 2πr. Furthermore,the Chain Rule tells us that:
dr
dA dA dr
= ·
dt dr dt
dA dr
We know that = 2πr and = 10 are always true, so we can rewrite
dr dt
the Chain Rule statement above as:
dA
= 2πr · 10 = 20πr
dt
Finally, we evaluate both sides at the moment in time we are interested
in (the moment when r = 5):

dA  
= 20πr  = 20π · 5 = 100π ≈ 314.2

dt r=5
 r =5

which is the same answer we found in the original solution. ◀

178 contemporary calculus

The key steps in finding the rate of change of the area of the circle were:

• write the known information in a mathematical form, expressing

dr
rates of change as derivatives: = 10 ft/sec
dt
dA
• write the question in a mathematical form: =?
dt
• find an equation connecting or relating the variables: A = πr2

• differentiate both sides of the connecting equation using the Chain

dA dA dr
Rule (and other differentiation patterns as necessary): =
dt dr dt
• substitute all of the known values that are always true into the
equation resulting from the previous step and (if necessary) solve
dA
for the desired quantity in the resulting equation: = 2πr · 10
dt
• substitute all of the known values that are true at the particular
moment in time the question asks about into the equation resulting
dA  
from the previous step: = 2πr · 10 = 100π

dt r=5

r =5

Example 2. Divers’ lives depend on understanding situations involving

related rates. In water, thepressure at a depth of x feet is approx-
 x
imately P( x ) = 15 1 + pounds per square inch (compared to
33
approximately P(0) = 15 pounds per square inch at sea level). Volume
k
is inversely proportional to the pressure, V = , so doubling the pres-
P
sure will result in half the original volume. Remember that volume is a
function of the pressure: V = V ( P).

(a) Suppose a diver’s lungs, at a depth of 66 feet, contained 1 cubic foot

of air and the diver ascended to the surface without releasing any
air. What would happen?

(b) If a diver started at a depth of 66 feet and ascended at a rate of 2 feet

per second, how fast would the pressure be changing?

(Dives deeper than 50 feet also involve a risk of the “bends,” or decom-
pression sickness, if the ascent is too rapid. Tables are available that
show the safe rates of ascent from different depths.)

Solution. (a) The diver would risk rupturing his or her lungs. The
1 cubic foot of air at a depth of 66 feet would be at a pressure of
66


P(66) = 15 1 + 33 = 45 pounds per square inch (psi). Because the
pressure at sea level, P(0) = 15 psi, is only 13 as great, each cubic foot
of air would expand to 3 cubic feet, and the diver’s lungs would be
in danger. Divers are taught to release air as they ascend to avoid
this danger. (b) The diver is ascending at a rate of 2 feet/second
 the derivative 179

so the rate of change of the diver’s depth with respect to time is

dx
= −2 ft/s. (Why is this rate of change negative?) The pressure is
dt
x
= 15 + 15


P = 15 1 + 33 33 x, a function of x, so using the Chain Rule:
 
dP dP dx 15 psi ft 30 psi psi
= · = · −2 =− ≈ −0.91
dt dx dt 33 ft sec 33 sec sec
The rates of change in this problem are constant (they hold true at any
moment in time during the ascent) so we are done. ◀

Example 3. The height of a cylinder is increasing at 7 meters per second

and the radius is increasing at 3 meters per second. How fast is the
volume changing when the cylinder is 5 meters high and has a radius
of 6 meters? (See margin.)

Solution. First we need to translate our known information into a

mathematical format. The height and radius are given: at the particular
moment in time the question asks about, h = height = 5 m and r =
radius = 6 m. We are also told how fast h and r are changing at this
moment in time: dh dr
dt = 7 m/sec and dt = 3 m/sec. Finally, we are asked
∆V
to find dt , and we should expect the units of dV
dV
dt to be the same as ∆t ,
which are m3 /sec.
• variables: h(t) = height at time t seconds, r (t) = radius at time t,
V (t) = volume at time t.
dh
• we know: at a particular moment in time, h = 5 m, dt = 7 m/sec,
r = 6 m and dr
dt = 3 m/sec
dV
• we want to know: dt at this particular moment in time
We also need an equation that relates the variables h, r and V (all of
which are functions of time t) to each other:
• connecting equation: V = πr2 h
Differentiating each side of this equation with respect to t (remembering
that h, r and V are functions of t), we have:
dV d  2  dh d  2
= πr h = πr2 · +h· πr
dt dt dt dt
dh dr
= πr2 · + h · 2πr ·
dt dt
using the Product Rule (on the product πr2 · h) and the Power Rule for
Functions (on πr2 , remembering that r is actually a function of t).
The rest of the solution just involves substituting values and doing
some arithmetic. At the particular moment in time we’re interested in:
dV m m
= π · 62 m2 · 7 + 5 m · 2π · 6 m · 3
dt sec sec
m3 m3
= 432π ≈ 1357.2
sec sec
180 contemporary calculus

The volume of the cylinder is increasing at a rate of 1,357.2 cubic meters

per second. (It is always encouraging when the units of our answer are
the ones we expect.) ◀

Practice 1. How fast is the surface area of the cylinder changing in

the previous example? (Assume that h, r, dh dr
dt and dt have the same
values as in the example and use the figure in the margin to help you
determine an equation relating the surface area of the cylinder to the
variables h and r. The cylinder includes a top and bottom.)

Practice 2. How fast is the volume of the cylinder in the previous

example changing if the radius is decreasing at a rate of 3 meters per
second? (The height, radius and rate of change of the height are the
same as in the previous example: 5 m, 6 m and 7 m/sec respectively.)

Usually, the most difficult part of Related Rates problems is to find

an equation that relates or connects all of the variables. In the previous
problems, the relating equations required a knowledge of geometry and
formulas for areas and volumes (or knowing where to look them up).
Other Related Rates problems may require information about similar
triangles, the Pythagorean Theorem or trigonometric identities: the
information required varies from problem to problem.
It is a good idea — a very good idea — to draw a picture of the
physical situation whenever possible. It is also a good idea, particularly
if the problem is very important (your next raise depends on getting
the right answer), to calculate at least one approximate answer as a check
of your exact answer.

Example 4. Water is flowing into a conical tank at a rate of 5 m3 /sec.

If the radius of the top of the cone is 2 m, the height is 7 m, and the
depth of the water is 4 m, then how fast is the water level rising?

Solution. Let’s define our variables to be h = height (or depth) of the

water in the cone and V = the volume of the water in the cone. Both h
and V are changing, and both of them are functions of time t. We are
told in the problem that h = 4 m and dV 3
dt = 5 m /sec, and we are asked
∆h
to find dh dh
dt . We expect that the units of dt will be the same as ∆t , which
are meters/second.

• variables: h(t) = height at time t seconds, r (t) = radius of the top

surface of the water at time t, V (t) = volume of water at time t

dV
• we know: dt = 5 m3 /sec (always true) and h = 4 m (at a particular
moment)

dh
• we want to know: dt at this particular moment
 the derivative 181

Unfortunately, the equation for the volume of a cone, V = 13 πr2 h, also

involves an additional variable r, the radius of the cone at the top of
the water. This is a situation in which a picture can be a great help by
suggesting that we have a pair of similar triangles:

r top radius 2m 2 2
= = = ⇒ r= h
h total height 7m 7 7

Knowing this, we can rewrite the volume of the water contained in the
cone, V = 13 πr2 h, as a function of the single variable h:
 2
1 2 1 2 4
• connecting equation: V = πr h = π h h= πh3
3 3 7 147
The rest of the solution is reasonably straightforward.
 
dV dV dh d 4 3 dh
= · = πh ·
dt dh dt dh 147 dt

dV
We know = 5 always holds, and the derivative is easy to compute:
dt
4 dh
5= πh2 ·
49 dt
At the particular moment in time we want to know about (when h = 4):

4  dh  64π dh 
5= πh2  ·  ⇒ 5= · 

49 h=4 dt h=4 49 dt h=4
and we can now solve for the quantity of interest:

dh  5 245 m
= 64π = ≈ 1.22
dt h=4 64π sec

49

This example was a bit more challenging because we needed to use

similar triangles to get an equation relating V to h and because we
dh
eventually needed to do some arithmetic to solve for . ◀
dt
Practice 3. A rainbow trout has taken the fly at the end of a 60-foot
fishing line, and the line is being reeled in at a rate of 30 feet per minute.
If the tip of the rod is 10 feet above the water and the trout is at the
surface of the water, how fast is the trout being pulled toward the
angler? (Hint: Draw a picture and use the Pythagorean Theorem.)

Example 5. When rain is falling vertically, the amount (volume) of rain

collected in a cylinder is proportional to the area of the opening of the
cylinder. If you place a narrow cylindrical glass and a wide cylindrical
glass out in the rain:

(a) which glass will collect water faster?

(b) in which glass will the water level rise faster?

182 contemporary calculus

Solution. Let’s assume that the smaller glass has a radius of r and the
larger glass has a radius of R, so that R > r. The areas of their openings
are πr2 and πR2 , respectively. Call the volume of water collected in
each glass v (for the smaller glass) and V (for the larger glass).
dv
(a) The smaller glass will collect water at the rate = K · πr2 and
dt
dV dV dv
the larger at the rate = K · πR2 so > and the larger glass
dt dt dt
will collect water faster than the smaller glass.
(b) The volume of water in each glass is a function of the radius
of the glass and the height of the water in the glass: v = πr2 h and
V = πR2 H where h and H are the heights of the water levels in the
smaller and larger glasses, respectively. The heights h and H vary with
t (in other words, they are each functions of t) while the radii (r and R)
remain constant, so:
dv
dv d  2  dh dh Kπr2
= πr h = πr2 ⇒ = dt2 = =K
dt dt dt dt πr πr2
Similarly:
dV
dV d  2  dH dH KπR2
= πR H = πR2 ⇒ = dt 2 = =K
dt dt dt dt πR πR2
dh dH
So =K= , which tells us the water level in each glass is rising at
dt dt
the same rate. In a one-minute period, the larger glass will collect more
rain, but the larger glass also requires more rain to raise its water level
by a fixed amount. How do you think the volumes and water levels
would change if we placed a small glass and a large plastic (rectangular)
box side by side in the rain? ◀

2.6 Problems
1. An expandable sphere is being filled with liquid at a constant rate
from a tap (imagine a water balloon connected to a faucet). When
the radius of the sphere is 3 inches, the radius is increasing at 2
inches per minute. How fast is the liquid coming out of the tap?
(V = 43 πr3 )

2. The 12-inch base of a right triangle is growing at 3 inches per hour,

and the 16-inch height of the triangle is shrinking at 3 inches per
hour (see figure in the margin).

(a) Is the area increasing or decreasing?

(b) Is the perimeter increasing or decreasing?
(c) Is the hypotenuse increasing or decreasing?
 the derivative 183

3. One hour later the right triangle in the previous 5. The length of a 12-foot by 8-foot rectangle is in-
problem is 15 inches long and 13 inches high creasing at a rate of 3 feet per second and the
(see figure below) and the base and height are width is decreasing at 2 feet per second (see fig-
changing at the same rate as in Problem 2. ure below).
(a) Is the area increasing or decreasing now? (a) How fast is the perimeter changing?
(b) Is the hypotenuse increasing or decreasing? (b) How fast is the area changing?
(c) Is the perimeter increasing or decreasing?

6. A circle of radius 3 inches is inside a square with

12-inch sides (see figure below). How fast is the
4. A young woman and her boyfriend plan to elope, area between the circle and square changing if the
but she must rescue him from his mother, who radius is increasing at 4 inches per minute and
has locked him in his room. The young woman the sides are increasing at 2 inches per minute?
has placed a 20-foot long ladder against his house
and is knocking on his window when his mother
begins pulling the bottom of the ladder away
from the house at a rate of 3 feet per second (see
figure below). How fast is the top of the ladder
(and the young couple) falling when the bottom
of the ladder is:
(a) 12 feet from the bottom of the wall?
(b) 16 feet from the bottom of the wall?
(c) 19 feet from the bottom of the wall? 7. An oil tanker in Puget Sound has sprung a leak,
and a circular oil slick is forming. The oil slick is
4 inches thick everywhere, is 100 feet in diameter,
and the diameter is increasing at 12 feet per hour.
Your job, as the Coast Guard commander or the
tanker’s captain, is to determine how fast the oil
is leaking from the tanker.
 184 contemporary calculus

8. A mathematical species of slug has a semicircular 11. Redo the previous problem if the person is 20 feet
cross section and is always 5 times as long as it is from the lamp post.
high (see figure below). When the slug is 5 inches
12. Water is being poured at a rate of 15 cubic feet
long, it is growing at 0.2 inches per week.
per minute into a conical reservoir that is 20 feet
(a) How fast is its volume increasing?
deep and has a top radius of 10 feet (see below).
(b) How fast is the area of its “foot” (the part of the
slug in contact with the ground) increasing? (a) How long will it take to fill the empty reser-
voir?
(b) How fast is the water level rising when the
water is 4 feet deep?
(c) How fast is the water level rising when the
water is 16 feet deep?
9. Lava flowing from a hole at the top of a hill is
forming a conical mountain whose height is al-
ways the same as the width of its base (see figure
below). If the mountain is increasing in height at
2 feet per hour when it is 500 feet high, how fast
is the lava flowing (that is, how fast is the volume
of the mountain increasing)? (V = 31 πr2 h)

10. A 6-foot-tall person is walking away from a 14- 13. The string of a kite is perfectly taut and always
foot lamp post at 3 feet per second. When the makes an angle of 35◦ above horizontal.
person is 10 feet away from the lamp post:
(a) If the kite flyer has let out 500 feet of string,
(a) how fast is the length of the shadow changing?
how high is the kite?
(b) how fast is the tip of the shadow moving away
from the lamp post? (b) If the string is let out at a rate of 10 feet per
second, how fast is the kite’s height increasing?
 the derivative 185

14. A small tracking telescope is viewing a hot-air 18. If the rate at which water vapor condenses onto a
balloon rise from a point 1,000 meters away from spherical raindrop is proportional to the surface
a point directly under the balloon. area of the raindrop, show that the radius of the
raindrop will increase at a constant rate.
(a) When the viewing angle is 20◦ , it is increasing
at a rate of 3◦ per minute. How high is the 19. Define A( x ) to be the area bounded by the t- and
balloon, and how fast is it rising? y-axes, and the lines y = 5 and t = x.
(b) When the viewing angle is 80◦ , it is increasing (a) Find a formula for A as a function of x.
at a rate of 2◦ per minute. How high is the (b) Determine A′ ( x ) when x = 1, 2, 4 and 9.
balloon, and how fast is it rising? (c) If x is a function of time, x (t) = t2 , find a
formula for A as a function of t.
(d) Determine A′ (t) when t = 1, 2 and 3.
(e) Suppose instead x (t) = 2 + sin(t). Find a for-
mula for A(t) and determine A′ (t).

15. The 8-foot diameter of a spherical gas bubble 20. The point P is going around the circle x2 + y2 = 1
is increasing at 2 feet per hour, and the 12-foot- twice a minute. How fast is the distance between
long edges of a cube containing the bubble are the point P and the point (4, 3) changing:
increasing at 3 feet per hour. Is the volume con-
(a) when P = (1, 0)?
tained between the spherical bubble and the cube
increasing or decreasing? At what rate? (b) when P = (0, 1)?
(c) when P = (0.8, 0.6)?
16. In general, the strength S of an animal is propor-
tional to the cross-sectional area of its muscles, (Suggestion: Write x and y as parametric func-
and this area is proportional to the square of its tions of time t.)
height H, so the strength S = aH 2 . Similarly,
the weight W of the animal is proportional to
the cube of its height, so W = bH 3 . Finally, the
relative strength R of an animal is the ratio of its
strength to its weight. As the animal grows, show
that its strength and weight increase, but that the
relative strength decreases.
17. The snow in a hemispherical pile melts at a rate
proportional to its exposed surface area (the sur-
face area of the hemisphere). Show that the height
of the snow pile is decreasing at a constant rate.
 186 contemporary calculus

21. You are walking along a sidewalk toward a 40-

foot-wide sign adjacent to the sidewalk and per-
pendicular to it. If your viewing angle θ is 10◦ :
(a) how far are you from the corner of the sign?
(b) how fast is your viewing angle changing if you
are walking at 25 feet per minute?
(c) how fast are you walking if the angle is increas-
ing at 2◦ per minute?

2.6 Practice Answers

1. The surface area is S = 2πrh + 2πr2 . From the Example, we know
that dh dr
dt = 7 m/sec and dt = 3 m/sec, and we want to know how
fast the surface area is changing when h = 5 m and r = 6 m.
dS dh dr dr
= 2πr · + 2π · h + 2π · 2r ·
dt dt  dt  m  dt
m  m
= 2π (6 m) 7 + 2π 3 (5m) + 2π (2 · 6 m) 3
sec sec sec
Note that the units represent a rate of m2 m2
change of area. = 186π ≈ 584.34
sec sec

dr
2. The volume is V = πr2 h. We know that dt = −3 m/sec and that
h = 5 m, r = 6 m and dh
dt = 7 m/sec.
dV dh dr
= πr2 · + π · 2r ·
dt dt  dt
2 m  m
= π (6 m) 7 + π (2 · 6 m) −3
sec sec
Note that the units represent a rate of m3 m3
change of volume. = 72π ≈ 226.19
sec sec

3. See margin figure. We know dL ft

dt = −30 min (always true); F rep-
resents the distance from the fish to a point directly below the tip
of the rod, and the distance from that point to the angler remains
constant, so dF
dt will equal the rate at which the fish is moving toward
the angler. We want to know dF . The Pythagorean Theorem

dt  L=60
connects F and L: F2 + 102 = L2 . Differentiating with respect to t
and using the Power Rule for Functions:
dF dL dF L dL
2F · + 0 = 2L · ⇒ = ·
dt dt dt F dt
At a particular moment in time, L = 60 ⇒ F2 + 102 = 602 ⇒ F =
√ √ √
3600 − 100 = 3500 = 10 35 so:
dF  60 180 ft
= −30 · √ = − √ ≈ −30.43
dt L=60 min

10 35 35
 the derivative 187

2.7 Newton’s Method

Newton’s method is a process that can find roots of functions whose

graphs cross or just “kiss” the x-axis. Although this method is a bit
harder to apply than the Bisection Algorithm, it often finds roots that
the Bisection Algorithm misses, and it usually finds them faster.

Off on a Tangent
The basic idea of Newton’s Method is remarkably simple and graphical:
at a point ( x, f ( x )) on the graph of f , the tangent line to the graph
“points toward” a root of f , a place where the graph touches the x-axis.
To find a root of f , we just pick a starting value x0 , go to the point
( x0 , f ( x0 )) on the graph of f , build a tangent line there, and follow the
tangent line to where it crosses the x-axis, say at x1 .
If x1 is a root of f , we are done. If x1 is not a root of f , then x1 is
usually closer to the root than x0 was, and we can repeat the process,
using x1 as our new starting point. Newton’s method is an iterative
procedure — that is, the output from one application of the method
becomes the starting point for the next application.
Let’s begin with the function f ( x ) = x2 − 5, whose roots we already
√
know (x = ± 5 ≈ ±2.236067977), to illustrate Newton’s method.
First, pick some value for x0 , say x0 = 4, and move to the point
( x0 , f ( x0 )) = (4, 11) on the graph of f . The tangent line to the graph of
f at (4, 11) “points to” a location on the x-axis that is closer to the root
of f than the point we started with. We calculate this location on the
x-axis by finding an equation of the line tangent to the graph of f at
(4, 11) and then finding where this line intersects the x-axis.
At (4, 11), the line tangent to f has slope f ′ (4) = 2(4) = 8, so an
equation of the tangent line is y − 11 = 8( x − 4). Setting y = 0, we can
find where this line crosses the x-axis:
11 21
0 − 11 = 8( x − 4) ⇒ x = 4 − = = 2.625
8 8
Call this new value x1 : The point x1 = 2.625 is closer to the actual
√
root 5, but it certainly does not equal the actual root. So we can use
this new x-value, x1 = 2.625, to repeat the procedure:
• move to the point ( x1 , f ( x1 )) = (2.625, 1.890625)

• find an equation of the tangent line at ( x1 , f ( x1 )):

y − 1.890625 = 5.25( x − 2.625)

• find x2 , the x-value where this new line intersects the x-axis:

y − 1.890625 = 5.25( x − 2.625) ⇒ 0 − 1.890625 = 5.25( x2 − 2.625)

⇒ x2 = 2.264880952
188 contemporary calculus

Repeating this process, each new estimate for the root of f ( x ) = x2 − 5

becomes the starting point to calculate the next estimate. We get:

x0 =4 (0 correct digits)
x1 = 2.625 (1 correct digit)
x2 = 2.262880952 (2 correct digits)
x3 = 2.236251252 (4 correct digits)
x4 = 2.236067985 (8 correct digits)

It only took 4 iterations to get an approximation within 0.000000008 of

√
the exact value of 5 . One more iteration gives an approximation x5
that has 16 correct digits. If we start with x0 = −2 (or any negative
√
number), then the values of xn approach − 5 ≈ −2.23606.

Practice 1. Find where the tangent line to f ( x ) = x3 + 3x − 1 at (1, 3)

intersects the x-axis.

Practice 2. A starting point and a graph of f appear in the margin.

Label the approximate locations of the next two points on the x-axis
that will be found by Newton’s method.

The Algorithm for Newton’s Method

Rather than deal with each particular function and starting point, let’s
find a pattern for a general function f .
For the starting point x0 , the slope of the tangent line at the point
( x0 , f ( x0 )) is f ′ ( x0 ) so the equation of the tangent line is y − f ( x0 ) =
f ′ ( x0 ) · ( x − x0 ). This line intersects the x-axis at a point ( x1 , 0), so:

The process for Newton’s Method, start- f ( x0 )

0 − f ( x0 ) = f ′ ( x0 ) · ( x1 − x0 ) ⇒ x1 = x0 −
ing with x0 and graphically finding the f ′ ( x0 )
locations on the x-axis of x1 , x2 and x3 .
f ( x1 )
Starting with x1 and repeating this process we get x2 = x1 − ,
f ′ ( x1 )
f ( x2 )
x3 = x2 − and so on. In general, starting with xn , the line
f ′ ( x2 )
tangent to the graph of f at ( xn , f ( xn )) intersects the x-axis at ( xn+1 , 0)
f ( xn )
with xn+1 = xn − ′ , our new estimate for the root of f .
f ( xn )

Algorithm for Newton’s Method:

1. Pick a starting value x0 (preferably close to a root of f ( x )).

f ( xn )
2. For each xn , calculate a new estimate xn+1 = xn −
f ′ ( xn )
3. Repeat step 2 until the estimates are “close enough” to a root or
until the method “fails.”
 the derivative 189

When we use Newton’s method with f ( x ) = x2 − 5, the function in

our first example, we have f ′ ( x ) = 2x so
Problem 16 helps you show this pattern —
f ( xn ) xn 2 − 5 2xn 2 − xn 2 + 5 called Heron’s method — approximates
x n +1 = xn − ′ = xn − =
f ( xn ) 2xn 2xn the square root of any positive number:
2   just replace 5 with the number whose
xn + 5 1 5 square root you want to find.
= = xn +
2xn 2 xn

The new approximation, xn+1 , is the average of the previous approxi-

mation, xn , and 5 divided by the previous approximation, x5n .

Example 1. Use Newton’s method to approximate the root(s) of f ( x ) =

2x + x · sin( x + 3) − 5.

Solution. f ′ ( x ) = 2 + x cos( x + 3) + sin( x + 3) so:

f ( xn ) 2xn + xn · sin( xn + 3) − 5
x n +1 = x n − ′
= xn −
f ( xn ) 2 + xn · cos( xn + 3) + sin( xn + 3)

The graph of f ( x ) (see margin) indicates only one root of f , which

is near x = 3, so pick x0 = 3. Then Newton’s method yields the
values x0 = 3, x1 = 2.96484457, x2 = 2.96446277, x3 = 2.96446273 (the
underlined digits agree with the exact answer). ◀

If we had picked x0 = 4 in the previous example, Newton’s method

would have required 4 iterations to get 9 digits of accuracy. For x0 = 5,
7 iterations are needed to get 9 digits of accuracy. If we pick x0 = 5.1,
then the values of xn are not close to the actual root after even 100
iterations: x100 ≈ −49.183. Picking a “good” value for x0 can result in
values of xn that get close to the root quickly. Picking a “poor” value
for x0 can result in xn values that take many more iterations to get close
to the root — or that don’t approach the root at all.

The graph of the function can help you pick a “good” x0 .

Practice 3. Put x0 = 3 and use Newton’s method to find the first two
iterates, x1 and x2 , for the function f ( x ) = x3 − 3x2 + x − 1.

Example 2. The function graphed in the margin has roots at x = 3 and

x = 7. If we pick x0 = 1 and apply Newton’s method, which root do
the iterates (the values of xn ) approach?

Solution. The iterates of x0 = 1 are labeled in the margin graph. They

are approaching the root at 7. ◀

Practice 4. For the function graphed in the margin, which root do the
iterates of Newton’s method approach if:

(a) x0 = 2? (b) x0 = 3? (c) x0 = 5?

190 contemporary calculus

Iteration
We have been emphasizing the geometric nature of Newton’s method,
but Newton’s method is also an example of iterating a function. If
f (x)
N (x) = x − ′ , the “pattern” in the algorithm, then:
f (x)
f ( x0 )
x1 = x0 − = N ( x0 )
f ′ ( x0 )
f (x )
x2 = x1 − ′ 1 = N ( x1 ) = N ( N ( x0 )) = N ◦ N ( x0 )
f ( x1 )
f ( x2 )
x3 = x2 − ′ = N ( x2 ) = N ( N ( N ( x0 ))) = N ◦ N ◦ N ( x0 )
f ( x2 )
and, in general:

xn = N ( xn−1 ) = nth iteration of N starting with x0

At each step, we use the output from N as the next input into N.

What Can Go Wrong?

When Newton’s method works, it usually works very well and the
values of xn approach a root of f very quickly, often doubling the
number of correct digits with each iteration. There are, however, several
things that can go wrong.
An obvious problem with Newton’s method is that f ′ ( xn ) can be
0. Then the algorithm tells us to divide by 0 and xn+1 is undefined.
Geometrically, if f ′ ( xn ) = 0, the tangent line to the graph of f at xn is
horizontal and does not intersect the x-axis at any point. If f ′ ( xn ) = 0,
just pick another starting value x0 and begin again. In practice, a second
or third choice of x0 usually succeeds.
There are two other less obvious difficulties that are not as easy
to overcome — the values of the iterates xn may become locked into
an infinitely repeating loop (see margin), or they may actually move
farther away from a root (see lower margin figure).
Example 3. Put x0 = 1 and use Newton’s method to find the first two
iterates, x1 and x2 , for the function f ( x ) = x3 − 3x2 + x − 1.

Solution. This is the function from the previous Practice Problem, but
with a different starting value for x0 : f ′ ( x ) = 3x2 − 6x + 1 so,
f ( x0 ) f (1) −2
x1 = x0 − = 1− = 1− =0
f ′ ( x0 ) f ′ (1) −2
f (x ) f (0) −1
and x2 = x1 − ′ 1 = 0 − = 0− =1
f ( x1 ) f ′ (0) 1
which is the same as x0 , so x3 = x1 = 0 and x4 = x2 = 1. The values of
xn alternate between 1 and 0 and do not approach a root. ◀
 the derivative 191

Newton’s method behaves badly at only a few starting points for this
particular function — for most starting points, Newton’s method con-
verges to the root of this function. There are some functions, however,
that defeat Newton’s method for almost every starting point.
√ 1
Practice 5. For f ( x ) = 3 x = x 3 and x0 = 1, verify that x1 = −2,
x2 = 4 and x3 = −8. Also try x0 = −3 and verify that the same pattern
holds: xn+1 = −2xn . Graph f and explain why the Newton’s method
iterates get farther and farther away from the root at 0.
Newton’s method is powerful and quick and very easy to program
on a calculator or computer. It usually works so well that many people
routinely use it as the first method they apply. If Newton’s method
fails for their particular function, they simply try some other method.

Chaotic Behavior and Newton’s Method

An algorithm leads to chaotic behavior if two starting points that
are close together generate iterates that are sometimes far apart and
sometimes close together: | a0 − b0 | is small but | an − bn | is large for lots
(infinitely many) of values of n and | an − bn | is small for lots of values
of n. The iterates of the next simple algorithm exhibit chaotic behavior.
A Simple Chaotic Algorithm: Starting with any number between 0
and 1, double the number and keep the fractional part of the result:
x1 is the fractional part of 2x0 , x2 is the fractional part of 2x1 , and in
general, xn+1 = 2xn − ⌊2xn ⌋.
If x0 = 0.33, then the iterates of this algorithm are 0.66, 0.32 =
fractional part of 2 · 0.66, 0.64, 0.28, 0.56, . . . The iterates for two other
starting values close to 0.33 are given below as well as the iterates of
0.470 and 0.471:
x0 0.32 0.33 0.34 0.470 0.471
x1 0.64 0.66 0.68 0.940 0.942
x2 0.28 0.32 0.36 0.880 0.884
x3 0.56 0.64 0.72 0.760 0.768
x4 0.12 0.28 0.44 0.520 0.536
x5 0.24 0.56 0.88 0.040 0.072
x6 0.48 0.12 0.76 0.080 0.144
x7 0.96 0.24 0.56 0.160 0.288
x8 0.92 0.48 0.12 0.320 0.576
x9 0.84 0.96 0.24 0.640 0.152

There are starting values as close together as we want whose iterates

are far apart infinitely often.
Many physical, biological and financial phenomena exhibit chaotic
behavior. Atoms can start out within inches of each other and several
192 contemporary calculus

weeks later be hundreds of miles apart. The idea that small initial
differences can lead to dramatically diverse outcomes is sometimes
called the “butterfly effect” from the title of a talk (“Predictability: Does
the Flap of a Butterfly’s Wings in Brazil Set Off a Tornado in Texas?”)
given by Edward Lorenz, one of the first people to investigate chaos.
The “butterfly effect” has important implications about the possibility —
or rather the impossibility — of accurate long-range weather forecasting.
Chaotic behavior is also an important aspect of studying turbulent air
and water flows, the incidence and spread of diseases, and even the
fluctuating behavior of the stock market.
Newton’s method often exhibits chaotic behavior and — because it
is relatively easy to study — is often used as a model to investigate
the properties of chaotic behavior. If we use Newton’s method to
approximate the roots of f ( x ) = x3 − x (with roots 0, +1 and −1),
then starting points that are very close together can have iterates that
converge to different roots. The iterates of 0.4472 and 0.4473 converge
to the roots 0 and +1, respectively. The iterates of the median value
0.44725 converge to the root −1, and the iterates of another nearby
1 1 1
point, √ ≈ 0.44721, simply cycle between − √ and + √ and do
5 5 5
not converge at all.

Practice 6. Find the first four Newton’s method iterates of x0 = 0.997

and x0 = 1.02 for f ( x ) = x2 + 1. Try two other starting values very
close to 1 (but not equal to 1) and find their first four iterates. Use the
graph of f ( x ) = x2 + 1 to explain how starting points so close together
can quickly have iterates so far apart.

2.7 Problems

1. The graph of y = f ( x ) appears below. Estimate 2. The graph of y = g( x ) appears below. Estimate
the locations of x1 and x2 when you apply New- the locations of x1 and x2 when you apply New-
ton’s method with the given starting value x0 . ton’s method starting value with the value x0
shown in the graph.
 the derivative 193

3. The function graphed below has several roots. In Problems 12–15, use Newton’s method to find all
Which root do the iterates of Newton’s method roots or solutions, accurate to 2 decimal places, of
converge to if we start with x0 = 1? With x0 = 5? the given equation. It is helpful to examine a graph
to determine a “good” starting value x0 .

12. 2 + x = e x
x
13. = x2 − 2
x+3
14. x = sin( x )
4. The function graphed below has several roots. √
Which root do the iterates of Newton’s method 15. x = 5 3
converge to if we start with x0 = 2? With x0 = 6? 16. Show that if you apply Newton’s method to
√
f ( x ) = x2 − A to approximate A, then
 
1 A
x n +1 = xn +
2 xn

so the new estimate of the square root is the aver-

5. What happens to the iterates if we apply New- age of the previous estimate and A divided by the
ton’s method to the function graphed below and previous estimate. This method of approximating
start with x0 = 1? With x0 = 5? square roots is called Heron’s method.
17. Use Newton’s method to devise an algorithm for
approximating the cube root of a number A.
18. Use Newton’s method to devise an algorithm for
approximating the n-th root of a number A.

6. What happens if we apply Newton’s method to a Problems 19–22 involve chaotic behavior.
function f and start with x0 = a root of f ?
19. The iterates of numbers using the Simple Chaotic
7. What happens if we apply Newton’s method to a
Algorithm have some interesting properties.
function f and start with x0 = a maximum of f ?
(a) Verify that the iterates starting with x0 = 0 are
In Problems 8–9, a function and a value for x0 are
all equal to 0.
given. Apply Newton’s method to find x1 and x2 .
(b) Verify that if x0 = 12 , 14 , 18 and, in general, 21n ,
8. f ( x ) = x3 + x − 1 and x0 = 1
then the n-th iterate of x0 is 0 (and so are all
9. f ( x ) = x4 − x3 − 5 and x0 = 2 iterates beyond the n-th iterate.)
In Problems 10–11, use Newton’s method to find
20. When Newton’s method is applied to the func-
a root, accurate to 2 decimal places, of the given
tion f ( x ) = x2 + 1, most starting values for x0
functions using the given starting points.
lead to chaotic behavior for xn . Find a value for
10. f ( x ) = x3 − 7 and x0 = 2 x0 so that the iterates alternate: x1 = − x0 and
11. f ( x ) = x − cos( x ) and x0 = 0.7 x2 = − x1 = x0 .
 194 contemporary calculus

21. The function f ( x ) defined as: 22. (a) After many iterations (50 is fine) what hap-
( pens when you apply Newton’s method start-
2x if 0 ≤ x < 21
f (x) = ing with x0 = 0.5 to:
2 − 2x if 21 ≤ x ≤ 1
i. f ( x ) = 2x (1 − x )
is called a “stretch and fold” function.
ii. f ( x ) = 3.3x (1 − x )
(a) Describe what f does to the points in the inter-
val [0, 1]. iii. f ( x ) = 3.83x (1 − x )
(b) Examine and describe the behavior of the iter- (b) What do you think happens to the iterates of
ates of 32 , 25 , 27 and 29 . f ( x ) = 3.7x (1 − x )? What actually happens?
(c) Examine and describe the behavior of the iter- (c) Repeat parts (a)–(b) with some other starting
ates of 0.10, 0.105 and 0.11. values x0 between 0 and 1 (0 < x0 < 1). Does
(d) Do the iterates of f lead to chaotic behavior? the starting value seem to effect the eventual
behavior of the iterates?
(The behavior of the iterates of f depends in a
strange way on the numerical value of the leading
coefficient. The behavior exhibited in part (b) is
an example of “chaos.”)

2.7 Practice Answers

1. f ′ ( x ) = 3x2 + 3, so the slope of the tangent line at (1, 3) is f ′ (1) = 6

and an equation of the tangent line is y − 3 = 6( x − 1) or y = 6x − 3.
The y-coordinate of a point on the x-axis is 0 so putting y = 0 in
this equation: 0 = 6x − 3 ⇒ x = 12 . The line tangent to the graph
of f ( x ) = x3 + 3x + 1 at the point (1, 3) intersects the x-axis at the
point ( 21 , 0).

2. The approximate locations of x1 and x2 appear in the margin.

3. Using f ′ ( x ) = 3x2 + 3 and x0 = 3:

f ( x0 ) f (3) 2
x1 = x0 − ′
= 3− ′ = 3− = 2.8
f ( x0 ) f (3) 10
f (x ) f (2.8) 0.232
x2 = x1 − ′ 1 = 2.8 − ′ = 2.8 − ≈ 2.769948187
f ( x1 ) f (2.8) 7.72
f ( x2 )
x3 = x2 − ′ ≈ 2.769292663
f ( x2 )

4. The margin figure shows the first iteration of Newton’s Method for
x0 = 2, 3 and 5: If x0 = 2, the iterates approach the root at a; if
x0 = 3, they approach the root at c; and if x0 = 5, they approach the
root at a.
 the derivative 195

1 2
5. f ( x ) = x 3 ⇒ f ′ ( x ) = 13 x − 3 . If x0 = 1, then:

f (1) 1
x1 = 1 − = 1 − 1 = 1 − 3 = −2
f ′ (1) 3
1
f (−2) (−2) 3 −2
x2 = −2 − ′ = −2 − 2 = −2 − 1 = 4
f (−2) 1 −
3 (−2)
3 3
1
f (4) (4) 3 4
x3 = 4 − ′ = 4− 2 = 4 − 1 = −8
f (4) 1 −
3 (4)
3 3

and so on. If x0 = −3, then:

1
f (−3) (−3) 3
x1 = −3 − ′
= −3 − = −3 + 9 = 6
f (−3) 1 − 32
3 (−3)
1
f (6) (6) 3 6
x2 = 6 − ′ = 6− 2 = 6 − 1 = −12
f (6) 1 −
3 (6)
3 3
√
The graph of f ( x ) = 3 x has a shape similar to the margin figure
and the behavior of the iterates is similar to the pattern shown in that
figure. Unless x0 = 0 (the only root of f ) the iterates alternate in sign
and double in magnitude with each iteration: they get progressively
farther from the root with each iteration.

6. If x0 = 0.997, then x1 ≈ −0.003, x2 ≈ 166.4, x3 ≈ 83.2, x4 ≈ 41.6.

If x0 = 1.02, then x1 ≈ 0.0198, x2 ≈ −25.2376 , x3 ≈ −12.6 and
x4 ≈ −6.26.
196 contemporary calculus

2.8 Linear Approximation and Differentials

Newton’s method used tangent lines to “point toward” a root of a

function. In this section we examine and use another geometric charac-
teristic of tangent lines:

If f is differentiable at a, c is close to a
and y = L( x ) is the line tangent to f ( x ) at x = a
then L(c) is close to f (c).

We can use this idea to approximate the values of some commonly

used functions and to predict the “error” or uncertainty in a compu-
tation if we know the “error” or uncertainty in our original data. At
the end of this section, we will define a related concept called the
differential of a function.

Linear Approximation
Because this section uses tangent lines extensively, it is worthwhile to
recall how we find the equation of the line tangent to f ( x ) where x = a:
the tangent line goes through the point ( a, f ( a)) and has slope f ′ ( a) so,
using the point-slope form y − y0 = m( x − x0 ) for linear equations, we
have y − f ( a) = f ′ ( a) · ( x − a) ⇒ y = f ( a) + f ′ ( a) · ( x − a).

If f is differentiable at x = a
then an equation of the line L tangent to f at x = a is:
L( x ) = f ( a) + f ′ ( a) · ( x − a)

Example 1. Find a formula for L( x ), the linear function tangent to the

√
graph of f ( x ) = x at the point (9, 3). Evaluate L(9.1) and L(8.88) to
√ √
approximate 9.1 and 8.88.
√ 1 1
Solution. f ( x ) = x = x 2 ⇒ f ′ ( x ) = 12 x − 2 = 2√
1
x
so f (9) = 3 and
f ′ (9) = 1
√ = 16 . Thus:
2 9

1
L ( x ) = f (9) + f ′ (9) · ( x − 9) = 3 + ( x − 9)
6
If x is close to 9, then the value of L( x ) should be a good approximation
√
of the value of x. The number 9.1 is close to 9 so 9.1 = f (9.1) ≈
√
L(9.1) = 3 + 16 (9.1 − 9) ≈ 3.016666. Similarly, 8.88 = f (8.88) ≈
√
L(8.88) = 3 + 16 (8.88 − 9) = 2.98. In fact, 9.1 ≈ 3.016621, so our
√
estimate using L(9.1) is within 0.000045 of the exact answer; 8.88 ≈
2.979933 (accurate to 6 decimal places) and our estimate is within
0.00007 of the exact answer. ◀
 the derivative 197

In each case in the previous example, we got a good estimate of a

square root with very little work. The graph in the margin indicates
the graph of the tangent line y = L( x ) lies slightly above the graph
of y = f ( x ); consequently (as we observed), each estimate is slightly
larger than the exact value.

Practice 1. Find a formula for L( x ), the linear function tangent to the

√
graph of f ( x ) = x at the point (16, 4). Evaluate L(16.1) and L(15.92)
√ √
to approximate 16.1 and 15.92. Are your approximations using L
larger or smaller than the exact values of the square roots?

Practice 2. Find a formula for L( x ), the linear function tangent to the

graph of f ( x ) = x3 at the point (1, 1) and use L( x ) to approximate
(1.02)3 and (0.97)3 . Do you think your approximations using L are
larger or smaller than the exact values?

The process we have used to approximate square roots and cubics

can be used to approximate values of any differentiable function, and
the main result about the linear approximation follows from the two
statements in the boxes. Putting these two statements together, we have
the process for Linear Approximation.

Linear Approximation Process:

If f is differentiable at a and L( x ) = f ( a) + f ′ ( a) · ( x − a)
then (geometrically) the graph of L( x ) is close to the graph of
f ( x ) when x is close to a
and (algebraically) the values of the L( x ) approximate the
values of f ( x ) when x is close to a:
f ( x ) ≈ L( x ) = f ( a) + f ′ ( a) · ( x − a)

Sometimes we replace “x − a” with “∆x” in the last equation, and

the statement becomes f ( x ) ≈ f ( a) + f ′ ( a) · ∆x.

Example 2. Use the linear approximation process to approximate the

value of e0.1 .

Solution. f ( x ) = e x ⇒ f ′ ( x ) = e x so we need to pick a value a near

x = 0.1 for which we know the exact value of f ( a) = e a and f ′ ( a) = e a :
a = 0 is an obvious choice. Then:

e0.1 = f (0.1) ≈ L(0.1) = f (0) + f ′ (0) · (0.1 − 0)

= e0 + e0 · (0.1) = 1 + 1 · (0.1) = 1.1

You can use your calculator to verify that this approximation is within
0.0052 of the exact value of e0.1 . ◀
198 contemporary calculus

Practice 3. Approximate the value of (1.06)4 , the amount $1 becomes

after 4 years in a bank account paying 6% interest compounded annually.
(Take f ( x ) = x4 and a = 1.)

Practice 4. Use the linear approximation process and the values in the
table below to estimate the value of f when x = 1.1, 1.23 and 1.38.

x f (x) f ′ (x)
1.0 0.7854 0.5
1.2 0.8761 0.4098
1.4 0.9505 0.3378

We can approximate functions as well as numbers (specific values of

those functions).
√
Example 3. Find a linear approximation formula L( x ) for 1 + x when
√ √
x is small. Use your result to approximate 1.1 and 0.96.
√ 1 1
Solution. f ( x ) = 1 + x = (1 + x ) 2 ⇒ f ′ ( x ) = 12 (1 + x )− 2 = √1 ,
2 1+ x
so because “x is small,” we know that x is close to 0 and we can pick
a = 0. Then f ( a) = f (0) = 1 and f ′ ( a) = f ′ (0) = 12 so
√ 1 x
1 + x ≈ L ( x ) = f (0) + f ′ (0) · ( x − 0) = 1 + x = 1 +
2 2
√ √
Taking x = 0.1, 1.1 = 1 + 0.1 ≈ 1 + 0.1 2 = 1.05; taking x = −0.04,
√
1 + (−.04) ≈ 1 + −0.04
p
0.96 = 2 = 0.98. Use your calculator to
determine by how much each estimate differs from the true value. ◀

Applications of Linear Approximation to Measurement “Error”

Most scientific experiments use instruments to take measurements, but
these instruments are not perfect, and the measurements we get from
them are only accurate up to a certain level of precision. If we know
this level of accuracy of our instruments and measurements, we can
use the idea of linear approximation to estimate the level of accuracy
of results we calculate from our measurements.
If we measure the side x of a square to be 8 inches, then we would of
course calculate its area to be 82 = 64 square inches. Suppose, as would
reasonable with a real measurement, that our measuring instrument
could only measure or be read to the nearest 0.05 inches. Then our
measurement of 8 inches would really mean some number between
8 − 0.05 = 7.95 inches and 8 + 0.05 = 8.05 inches, so the true area of the
square would be between 7.952 = 63.2025 and 8.052 = 64.8025 square
inches. Our possible “error” or “uncertainty,” because of the limitations
of the instrument, could be as much as 64.8025 − 64 = 0.8025 square
inches, so we could report the area of the square to be 64 ± 0.8025
 the derivative 199

square inches. We can also use the linear approximation method to

estimate the “error” or uncertainty of the area. For a function as simple as the area of a
For a square with side x, the area is A( x ) = x2 and A′ ( x ) = 2x. square, this linear approximation method
really isn’t needed, but it serves as a use-
If ∆x represents the “error” or uncertainty of our measurement of ful and easily understood illustration of
the side then, using the linear approximation technique for A( x ), the technique.
A( x ) ≈ A( a) + A′ ( a) · ∆x so the uncertainty of our calculated area is
A( x ) − A( a) ≈ A′ ( a) · ∆x. In this example, a = 8 inches and ∆x = 0.05
inches, so A(8.05) ≈ A(8) + A′ (8) · (0.05) = 64 + 2(8) · (0.05) = 64.8
square inches, and the uncertainty in our calculated area is approxi-
mately A(8 + 0.05) − A(8) ≈ A′ (8) · ∆x = 2(8 inches)(0.05 inches) =
0.8 square inches. (Compare this approximation of the biggest pos-
sible error with the exact answer of 0.8025 square inches computed
previously.) This process can be summarized as:

Linear Approximation Error:

If the value of the x-variable is measured to be x = a with

a maximum “error” of ∆x units
then ∆ f , the “error” in estimating f ( x ), is:
∆ f = f ( x ) − f ( a) ≈ f ′ ( a) · ∆x.

Practice 5. If we measure the side of a cube to be 4 cm with an uncer-

tainty of 0.1 cm, what is the volume of the cube and the uncertainty of
our calculation of the volume? (Use linear approximation.)

Example 4. We are using a tracking telescope to follow a small rocket.

Suppose we are 3,000 meters from the launch point of the rocket, and, 2
seconds after the launch we measure the angle of the inclination of the
rocket to be 64◦ with a possible “error” of 2◦ . How high is the rocket
and what is the possible error in this calculated height?

Solution. Our measured angle is x = 1.1170 radians with ∆x = 0.0349

radians (all trigonometric work should be in radians), and the height of
the rocket at an angle x is f ( x ) = 3000 tan( x ) so f (1.1170) ≈ 6151 m.
Our uncertainty in the height is ∆ f ≈ f ′ ( x ) · ∆x ≈ 3000 · sec2 ( x ) · ∆x =
3000 sec2 (1.1170) · 0.0349 ≈ 545 m. If our measured angle of 64◦ can be
in error by as much as 2◦ , then our calculated height of 6,151 m can be
in error by as much as 545 m. The height is 6151 ± 545 meters. ◀

Practice 6. Suppose we measured the angle of inclination in the pre-

vious Example to be 43◦ ± 1◦ . Estimate the height of the rocket in the
form “height ± error.”

In some scientific and engineering applications, the calculated result

must be within some given specification. You might need to determine
200 contemporary calculus

how accurate the initial measurements must be in order to guarantee

the final calculation is within that specification. Added precision usu-
ally costs time and money, so it is important to choose a measuring
instrument good enough for the job but which is not too expensive.

Example 5. Your company produces ball bearings (small metal spheres)

with a volume of 10 cm3 and the volume must be accurate to within 0.1
cm3 . What radius should the bearings have — and what error can you
tolerate in the radius measurement to meet the accuracy specification
for the volume?

Solution. We want V = 10 and we know that the volume of a sphere

is V = 34 πr3 , so solve 10 = 43 πr3 for r to get r = 1.3365 cm. V (r ) =
′ ′
3 πr ⇒ V (r ) = 4πr so ∆V ≈ V (r ) · ∆r. In this case we know that
4 3 2

∆V = 0.1 cm and we have calculated r = 1.3365 cm, so 0.1 cm3 =

3

V ′ (1.3365 cm) · ∆r = (22.45 cm2 ) · ∆r. Solving for ∆r, we get ∆r ≈

0.0045 cm. To meet the specification for allowable error in volume, we
must allow the radius to vary no more than 0.0045 cm. If we instead
measure the diameter of the sphere, then we want the diameter to be
d = 2r = 2(1.3365 ± 0.0045) = 2.673 ± 0.009 cm. ◀

Practice 7. You want to determine the height of a rocket to within 10

meters when it is 4,000 meters high (see margin figure). How accurate
must your angle of measurement be? (Do your calculations in radians.)

Relative Error and Percentage Error

The “error” we’ve been examining is called the absolute error to dis-
tinguish it from two other terms, the relative error and the percentage
error, which compare the absolute error with the magnitude of the
number being measured. An “error” of 6 inches in measuring the
Earth’s circumference would be extremely small, but a 6-inch error in
measuring your head for a hat would result in a very bad fit.

Definitions:

error of f ∆f
The Relative Error of f is =
value of f f
∆f
The Percentage Error of f is · 100%.
f

Example 6. If the relative error in the calculation of the area of a circle

must be less than 0.4, then what relative error can we tolerate in the
measurement of the radius?
 the derivative 201

Solution. A(r ) = πr2 ⇒ A′ (r ) = 2πr and ∆A ≈ A′ (r ) · ∆r = 2πr∆r.

The Relative Error of A is:
∆A 2πr∆r ∆r
≈ =2
A πr2 r
∆A
We can guarantee that the Relative Error of A, , is less than 0.4 if
A
∆r 1 ∆A
the Relative Error of r, = , is less than 12 (0.4) = 0.2. ◀
r 2 A
Practice 8. If you can measure the side of a cube with a percentage error
less than 3%, then what will the percentage error for your calculation
of the surface area of the cube be?

The Differential of f
As shown in the margin, the change in value of the function f near
the point ( x, f ( x )) is ∆ f = f ( x + ∆x ) − f ( x ) and the change along
the tangent line is f ′ ( x ) · ∆x. If ∆x is small, then we have used the
approximation that ∆ f ≈ f ′ ( x ) · ∆x. This leads to the definition of a
new quantity, d f , called the differential of f .

Definition:

The differential of f is d f = f ′ ( x ) · dx where dx is any real number.

The differential of f represents the change in f , as x changes from

x to x + dx, along the tangent line to the graph of f at the point
( x, f ( x )). If we take dx to be the number ∆x, then the differential is an
approximation of ∆ f : ∆ f ≈ f ′ ( x ) · ∆x = f ′ ( x ) · dx = d f .

Example 7. Determine the differential for the functions f ( x ) = x3 − 7x,

g( x ) = sin( x ) and h(r ) = πr2 .

Solution. d f = f ′ ( x ) · dx = (3x2 − 7) dx, dg = g′ ( x ) · dx = cos( x ) dx, While we will do very little with differen-
and dh = h′ (r ) dr = 2πr dr. ◀ tials for a while, we will use them exten-
√ sively in integral calculus.
Practice 9. Determine the differentials of f ( x ) = ln( x ), u = 1 − 3x
and r = 3 cos(θ ).

The Linear Approximation “Error” | f ( x ) − L( x )|

An approximation is most valuable if we also have have some measure
of the size of the “error,” the distance between the approximate value
and the value being approximated. Typically, we will not know the
exact value of the error (why not?), but it is useful to know an upper
bound for the error. For example, if one scale gives the weight of a gold
202 contemporary calculus

pendant as 10.64 grams with an error less than 0.3 grams (10.64 ± 0.3
grams) and another scale gives the weight of the same pendant as 10.53
grams with an error less than 0.02 grams (10.53 ± 0.02 grams), then we
can have more faith in the second approximate weight because of the
smaller “error” guarantee. Before finding a guarantee on the size of the
error of the linear approximation process, we will check how well the
linear approximation process approximates values of some functions
we can compute exactly. Then we will prove one bound on the possible
error and state a somewhat stronger bound.

Example 8. Given the function f ( x ) = x2 , evaluate the expressions

f (2 + ∆x ), L(2 + ∆x ) and | f (2 + ∆x ) − L(2 + ∆x )| for ∆x = 0.1, 0.05,
0.01, 0.001 and for a general value of ∆x.

Solution. f (2 + ∆x ) = (2 + ∆x )2 = 22 + 4∆x + (∆x )2 and L(2 + ∆x ) =

f (2) + f ′ (2) · ∆x = 22 + 4 · ∆x. Then:

∆x f (2 + ∆x ) L(2 + ∆x ) | f (2 + ∆x ) − L(2 + ∆x )|
0.1 4.41 4.4 0.01
0.05 4.2025 4.2 0.0025
0.01 4.0401 4.04 0.0001
0.001 4.004001 4.004 0.000001

Cutting the value of ∆x in half makes the error one fourth as large.
Cutting ∆x to 101 1
as large makes the error 100 as large. In general:
   
| f (2 + ∆x ) − L(2 + ∆x )| =  22 + 4 · ∆x + (∆x )2 − 22 + 4 · ∆x 
 

= (∆x )2

This function and error also have a nice geometric interpretation (see
margin): f ( x ) = x2 is the area of a square of side x so f (2 + ∆x ) is
the area of a square of side 2 + ∆x, and that area is the sum of the
pieces with areas 22 , 2 · ∆x, 2 · ∆x and (∆x )2 . The linear approximation
L(2 + ∆x ) = 22 + 4 · ∆x to the area of the square includes the three
largest pieces, 22 , 2 · ∆x and 2 · ∆x, but omits the small square with area
(∆x )2 so the approximation is in error by the amount (∆x )2 . ◀

Practice 10. Given f ( x ) = x3 , evaluate f (4 + ∆x ), L(4 + ∆x ) and

| f (4 + ∆x ) − L(4 + ∆x )| for ∆x = 0.1, 0.05, 0.01, 0.001 and for a general
value of ∆x. Use the margin figure to give a geometric interpretation of
f (4 + ∆x ), L(4 + ∆x ) and | f (4 + ∆x ) − L(4 + ∆x )|.

In the previous Example and previous Practice problem, the error

| f ( a + ∆x ) − L( a + ∆x )| was very small, proportional to (∆x )2 , when
∆x was small. In general, this error approaches 0 as ∆x → 0.
 the derivative 203

Theorem:
If f ( x ) is differentiable at a
and L( a + ∆x ) = f ( a) + f ′ ( a) · ∆x
then lim | f ( a + ∆x ) − L( a + ∆x )| = 0
∆x →0
| f ( a + ∆x ) − L( a + ∆x )|
and lim = 0.
∆x →0 ∆x

Proof. First rewrite the quantity inside the absolute value as:

f ( a + ∆x ) − L( a + ∆x ) = f ( a + ∆x ) − f ( a) − f ′ ( a) · ∆x
f ( a + ∆x ) − f ( a)
 
= − f ′ ( a) · ∆x
∆x

f ( a + ∆x ) − f ( a)
But f is differentiable at x = a so lim = f ′ ( a ),
∆x →0 ∆x
f ( a + ∆x ) − f ( a)
 
which we can rewrite as lim − f ′ ( a) = 0. Thus:
∆x →0 ∆x

f ( a + ∆x ) − f ( a)
 
lim [ f ( a + ∆x ) − L( a + ∆x )] = lim − f ′ ( a) · lim ∆x = 0 · 0 = 0
∆x →0 ∆x →0 ∆x ∆x →0

Not only does the difference f ( a + ∆x ) − L( a + ∆x ) approach 0, but

this difference approaches 0 so fast that we can divide it by ∆x, another
quantity approaching 0, and the quotient still approaches 0.

In the next chapter we will be able to prove that the error of the
linear approximation process is in fact proportional to (∆x )2 . For now,
we just state the result.

Theorem:
If f is differentiable at a
and | f ′′ ( x )| ≤ M for all x between a and a + ∆x
then |“error”| = | f ( a + ∆x ) − L( a + ∆x )| ≤ 21 M · (∆x )2 .

2.8 Problems
1. The figure in the margin shows the tangent line to a function g at
the point (2, 2) and a line segment ∆x units long.

(a) On the figure, label the locations of

i. 2 + ∆x on the x-axis
ii. the point (2 + ∆x, g(2 + ∆x ))
iii. the point (2 + ∆x, g(2) + g′ (2) · ∆x )
(b) How large is the “error,” ( g(2) + g′ (2) · ∆x ) − ( g(2 + ∆x ))?
204 contemporary calculus

2. In the figure below, is the linear approximation 8. The height of a triangle is exactly 4 inches, and the
L( a + ∆x ) larger or smaller than the value of base is measured to be 7±0.5 inches (see figure
f ( a + ∆x ) when: below). Shade a part of the figure that represents
the “error” in the calculation of the area of the
(a) a = 1 and ∆x = 0.2?
triangle.
(b) a = 2 and ∆x = −0.1?
(c) a = 3 and ∆x = 0.1?
(d) a = 4 and ∆x = 0.2?
(e) a = 4 and ∆x = −0.2?

In Problems 3–4, find a formula for the linear func- 9. A rectangle has one side on the x-axis, one side on
tion L( x ) tangent to the given function f at the given the y-axis and a corner on the graph of y = x2 + 1
point ( a, f ( a)). Use the value L( a + ∆x ) to approxi- (see figure below).
mate the value of f ( a + ∆x ). (a) Use Linear Approximation of the area formula
√ to estimate the increase in the area of the rect-
3. (a) f ( x ) = x, a = 4, ∆x = 0.2
√ angle if the base grows from 2 to 2.3 inches.
(b) f ( x ) = x, a = 81, ∆x = −1
(b) Calculate exactly the increase in the area of
(c) f ( x ) = sin( x ), a = 0, ∆x = 0.3 the rectangle as the base grows from 2 to 2.3
4. (a) f ( x ) = ln( x ), a = 1, ∆x = 0.3 inches.
(b) f ( x ) = e x , a = 0, ∆x = 0.1
(c) f ( x ) = x5 , a = 1, ∆x = 0.03
5. Show that (1 + x )n ≈ 1 + nx if x is “close to” 0.
(Suggestion: Put f ( x ) = (1 + x )n and a = 0 and
then replace ∆x with x.)

In 6–7, use the linear approximation process to ob-

tain each formula for x “close to” 0.
6. (a) (1 − x )n ≈ 1 − nx
(b) sin( x ) ≈ x
(c) e x ≈ 1 + x
7. (a) ln(1 + x ) ≈ x
(b) cos( x ) ≈ 1
(c) tan( x ) ≈ x


(d) sin π2 + x ≈ 1
 the derivative 205

π
10. You know that you can measure the diameter of (a) How far away will the ball land if θ = 4 and
a circle to within 0.3 cm of the exact value. v = 80 feet/second?
(a) How large is the “error” in the calculated area (b) Which will result in a greater change in the
of a circle with a measured diameter of 7.4 cm? distance: a 5% error in the angle θ or a 5%
(b) How large is the “error” with a measured di- error in the initial velocity v?
ameter of 13.6 cm? 17. For the function graphed below, estimate the
(c) How large is the percentage error in the calcu- value of d f when
lated area with a measured diameter of d?
(a) x = 2 and dx = 1
11. You are minting gold coins that must have a vol-
(b) x = 4 and dx = −1
ume of 47.3±0.1 cm3 . If you can manufacture the
(c) x = 3 and dx = 2
coins to be exactly 2 cm high, how much variation
can you allow for the radius?
12. If F is the fraction of carbon-14 remaining in
a plant sample Y years after it died, then Y =
5700 ln(0.5) · ln( F ).
(a) Estimate the age of a plant sample in which
83±2% (0.83 ± 0.02) of the carbon-14 remains.
(b) Estimate the age of a plant sample in which
13±2% (0.13 ± 0.02) of the carbon-14 remains.
18. For the function graphed below, estimate the
13. Your company is making dice (cubes) and speci-
value of d f when
fications require that their volume be 87±2 cm3 .
How long should each side be and how much (a) x = 1 and dx = 2
variation can be allowed? (b) x = 2 and dx = −1
14. If the specifications require a cube with a surface (c) x = 3 and dx = 1
area of 43±0.2 cm2 , how long should each side be
and how much variation can be allowed in order
to meet the specifications?
15. The period P, in seconds, for a pendulum to make
one complete swing
s and return to the release
L
point is P = 2π where L is the length of the
g
pendulum in feet and g is 32 feet/sec2 .
(a) If L = 2 feet, what is the period?
(b) If P = 1 second, how long is the pendulum? 19. Calculate the differentials d f for the following
(c) Estimate the change in P if L increases from 2 functions:
feet to 2.1 feet.
(a) f ( x ) = x2 − 3x
(d) The length of a 24-foot pendulum is increasing
(b) f ( x ) = e x
2 inches per hour. Is the period getting longer
or shorter? How fast is the period changing? (c) f ( x ) = sin(5x )

16. A ball thrown at an angle θ (with the horizontal) (d) f ( x ) = x3 + 2x with x = 1 and dx = 0.2
2
with an initial velocity v will land vg · sin(2θ ) feet (e) f ( x ) = ln( x ) with x = e and dx = −0.1
√
from the thrower. (f) f ( x ) = 2x + 5 with x = 22 and dx = 3.
206 contemporary calculus

2.8 Practice Answers

1
1. f ( x ) = x 2 ⇒ f ′ ( x ) = 1
√
2 x
. At the point (16, 4) on the graph of f ,
the slope of the tangent line is f ′ (16) = √1 = 1 . An equation of
8
2 16
the tangent line is y − 4 = 81 ( x − 16) or y = 18 x + 2: L( x ) = 81 x + 2.
So:
√ 1
16.1 ≈ L(16.1) = (16.1) + 2 = 4.0125
8
√ 1
15.92 ≈ L(15.92) = (15.92) + 2 = 3.99
8

2. f ( x ) = x3 ⇒ f ′ ( x ) = 3x2 . At (1, 1), the slope of the tangent line is

f ′ (1) = 3. An equation of the tangent line is y − 1 = 3( x − 1) or
y = 3x − 2: L( x ) = 3x − 2. So:

(1.02)3 ≈ L(1.02) = 3(1.02) − 2 = 1.06

(0.97)3 ≈ L(0.97) = 3(0.97) − 2 = 0.91

3. f ( x ) = x4 ⇒ f ′ ( x ) = 4x3 . Taking a = 1 and ∆x = 0.06:

(1.06)4 = f (1.06) ≈ L(1.06) = f (1) + f ′ (1) · (0.06)

= 14 + 4(13 )(0.06) = 1.24

4. Using values given in the table:

f (1.1) ≈ f (1) + f ′ (1) · (0.1)

= 0.7854 + (0.5)(0.1) = 0.8354
f (1.23) ≈ f (1.2) + f ′ (1.2) · (0.03)
= 0.8761 + (0.4098)(0.03) = 0.888394
f (1.38) ≈ f (1.4) + f ′ (1.4) · (−0.02)
= 0.9505 + (0.3378)(−0.02) = 0.943744

5. f ( x ) = x3 ⇒ f ′ ( x ) = 3x2 so f (4) = 43 = 64 cm3 and the “error” is:

∆ f ≈ f ′ ( x ) · ∆x = 3x2 · ∆x

When x = 4 and ∆x = 0.1, ∆ f ≈ 3(4)2 (0.1) = 4.8 cm3 .

6. 43◦ ± 1◦ is equivalent to 0.75049 ± 0.01745 radians, so with f ( x ) =

3000 tan( x ) we have f (0.75049) = 3000 tan(0.75049) ≈ 2797.5 m and
f ′ ( x ) = 3000 sec2 ( x ). So:

∆ f ≈ f ′ ( x ) · ∆x = 3000 sec2 ( x ) · ∆x
= 3000 sec2 (0.75049) · (0.01745) = 97.9 m

The height of the rocket is 2797.5 ± 97.9 m.

 the derivative 207

7. f (θ ) = 2000 tan(θ ) ⇒ f ′ (θ ) = 2000 sec2 (θ ) and we know f (θ ) =

4000, so:

4000 = 2000 tan(θ ) ⇒ tan(θ ) = 2 ⇒ θ ≈ 1.10715 (radians)

and thus f ′ (1.10715) = 2000 sec2 (1.10715) ≈ 10000. Finally, the

“error” is given by ∆ f ≈ f ′ (θ ) · ∆θ so:

10
10 ≈ 10000 · ∆θ ⇒ ∆θ ≈ = 0.001 (radians) ≈ 0.057◦
10000

8. A(r ) = 6r2 ⇒ A′ (r ) = 12r ⇒ ∆A ≈ A′ (r ) · ∆r = 12r · ∆r and we

also know that ∆r
r < 0.03, so the percentage error is:

∆A 12r · ∆r 2∆r
· 100% = 2
· 100% = · 100% < 200(0.03)% = 6%
A 6r r

9. Computing differentials:

1
d f = f ′ ( x ) · dx =
dx
x
du −3
du = · dx = √ dx
dx 2 1 − 3x
dr
dr = dθ = −3 sin(θ ) dθ
dθ

10. f ( x ) = x3 ⇒ f ′ ( x ) = 3x2 so:

L(4 + ∆x ) = f (4) + f ′ (4)∆x = 43 + 3(4)2 ∆x = 64 + 48∆x

Evaluating the various quantities at the indicated points:

∆x f (4 + ∆x ) L(4 + ∆x ) | f (4 + ∆x ) − L(4 + ∆x )|
0.1 68.921 68.8 0.121
0.05 66.430125 66.4 0.030125
0.01 64.481201 64.48 0.001201
0.001 64.048012 64.048 0.000012

f (4 + ∆x ) is the actual volume of the cube with side length 4 + ∆x.

L(4 + ∆x ) is the volume of the cube with side length 4 (V = 64) plus
the volume of the three “slabs” (V = 3 · 42 · ∆x).
| f (4 + ∆x ) − L(4 + ∆x )| is the volume of the “leftover” pieces from
L: the three “rods” (V = 3 · 4 · (∆x )2 ) and the tiny cube (V = (∆x )3 ).
208 contemporary calculus

2.9 Implicit and Logarithmic Differentiation

This short section presents two more differentiation techniques, both

more specialized than the ones we have already seen — and conse-
quently used on a smaller class of functions. For some functions,
however, one of these techniques may be the only method that works.
The idea of each method is straightforward, but actually using each of
them requires that you proceed carefully and practice.

Implicit Differentiation
In our work up until now, the functions we needed to differentiate were
either given explicitly as a function of x, such as y = f ( x ) = x2 + sin( x ),
or it was fairly straightforward to find an explicit formula, such as
√3
solving y3 − 3x2 = 5 to get y = 5 + 3x2 . Sometimes, however, we will
have an equation relating x and y that is either difficult or impossible
to solve explicitly for y, such as y2 + 2y = sin( x ) + 4 (difficult) or
y + sin(y) = x3 − x (impossible). In each case, we can still find y′ =
f ′ ( x ) by using implicit differentiation.
The key idea behind implicit differentiation is to assume that y is a
function of x even if we cannot explicitly solve for y. This assumption
does not require any work, but we need to be very careful to treat y
as a function when we differentiate and to use the Chain Rule or the
Power Rule for Functions.
Example 1. Assume y is a function of x and compute each derivative:
d  3 2
(a) D(y3 ) (b) x y (c) (sin(y))′
dx
Solution. (a) We need the Power Rule for Functions because y is a
function of x:
D(y3 ) = 3y2 · D(y) = 3y2 · y′

(b) We need to use the Product Rule and the Chain Rule:
d  3 2 d  2 d  3 dy
x y = x3 · y + y2 · x = x3 · 2y · + y2 · 3x2
dx dx dx dx
(c) We just need to remember that D(sin(u)) = cos(u) and then use
the Chain Rule: (sin(y))′ = cos(y) · y′ . ◀

Practice 1. Assume that y is a function of x. Calculate:

d
(a) D x2 + y2


(b) (sin(2 + 3y)).
dx

Implicit Differentiation:

To determine y′ , differentiate each side of the defining equation,

treating y as a function of x, and then algebraically solve for y′ .
 the derivative 209

Example 2. Find the slope of the tangent line to the circle x2 + y2 = 25

at the point (3, 4) with and without implicit differentiation.
√
Solution. Explicitly: We can solve x2 + y2 = 25 for y: y = ± 25 − x2
but because the point (3, 4) is on the top half of the circle, we just need
√
y = 25 − x2 so:
p  1 − 1 −x
2
D(y) = D 25 − x2 = 25 − x2 · (−2x ) = √
2 25 − x2

Replacing x with 3, we have y′ = √ −3 = − 34 .

25−32
Implicitly: We differentiate each side of the equation x2 + y2 = 25
treating y as a function of x and then solve for y′ :
  −2x x
D x2 + y2 = D(25) ⇒ 2x + 2y · y′ = 0 ⇒ y′ = =−
2y y

so at the point (3, 4), y′ = − 34 , the same answer we found explicitly. ◀

Practice 2. Find the slope of the tangent line to y3 − 3x2 = 15 at the

point (2, 3) with and without implicit differentiation.

In the previous Example and Practice problem, it was easy to explic-

itly solve for y, and then we could differentiate y to get y′ . Because we
could explicitly solve for y, we had a choice of methods for calculating
y′ . Sometimes, however, we cannot explicitly solve for y and the only
way to determine y′ is with implicit differentiation.

Example 3. Determine y′ at (0, 2) for y2 + 2y = sin( x ) + 8.

Solution. Assuming that y is a function of x and differentiating each

side of the equation, we get:
 
D y2 + 2y = D (sin( x ) + 8) ⇒ 2y · y′ + 2y′ = cos( x )
cos( x ) We could have first solved the equation
⇒ (2y + 2)y′ = cos( x ) ⇒ y′ = explicitly for y using the quadratic for-
2y + 2
mula. Do you see how? Would that make
cos(0) 1 the problem easier or harder than using
so, at the point (0, 2), y′ = = . ◀ implicit differentiation?
2(2) + 2 6

Practice 3. Determine y′ at (1, 0) for y + sin(y) = x3 − x.

In practice, the equations may be rather complicated, but if you pro-

ceed carefully and step by step, implicit differentiation is not difficult.
Just remember that y must be treated as a function so every time you
differentiate a term containing a y you should use the Chain Rule and
get something that has a y′ . The algebra needed to solve for y′ is always
easy — if you differentiated correctly, the resulting equation will be a
linear equation in the variable y′ .
210 contemporary calculus

Example 4. Find an equation of the tangent line L to the “tilted”

parabola graphed below at the point (1, 2).

Solution. The line goes through the point (1, 2) so we need to find the
slope there. Differentiating each side of the equation, we get:
 
D x2 + 2xy + y2 + 3x − 7y + 2 = D(0)

which yields:

2x + 2x · y′ + 2y + 2y · y′ + 3 − 7y′ = 0
⇒ (2x + 2y − 7)y′ = −2x − 2y − 3
−2x − 2y − 3
⇒ y′ =
2x + 2y − 7
−2 − 4 − 3
so the slope at (1, 2) is m = y′ = = 9. Finally, an equation
2+4−7
for the line is y − 2 = 9( x − 1) so y = 9x − 7. ◀

Practice 4. Find the points where the parabola graphed above crosses
the y-axis, and find the slopes of the tangent lines at those points.
Implicit differentiation provides an alternate method for differenti-
ating equations that can be solved explicitly for the function we want,
and it is the only method for finding the derivative of a function that
we cannot describe explicitly.

Logarithmic Differentiation
f ′ (x)
In Section 2.5 we saw that D (ln( f ( x ))) = . If we simply multiply
f (x)
′
each side by f ( x ), we have: f ( x ) = f ( x ) · D (ln( f ( x ))). When the
logarithm of a function is simpler than the function itself, it is often
easier to differentiate the logarithm of f than to differentiate f itself.
 the derivative 211

Logarithmic Differentiation:

f ′ ( x ) = f ( x ) · D (ln( f ( x )))

In words: The derivative of f is f times the derivative of the natural

logarithm of f . Usually it is easiest to proceed in three steps:

• Calculate ln ( f ( x )) and simplify.

• Calculate D (ln( f ( x ))) and simplify

• Multiply the result in the previous step by f ( x ).

Let’s examine what happens when we use this process on an “easy”

function, f ( x ) = x2 , and a “hard” one, f ( x ) = 2x . Certainly we
don’t need to use logarithmic differentiation to find the derivative of
f ( x ) = x2 , but sometimes it is instructive to try a new algorithm on
a familiar function. Logarithmic differentiation is the easiest way to
find the derivative of f ( x ) = 2x (if we don’t remember the pattern for
differentiating a x from Section 2.5).

f ( x ) = x2 f ( x ) = 2x
ln ( f ( x )) = ln( x2 ) = 2 · ln( x ) ln ( f ( x )) = ln(2x ) = x · ln(2)
D (ln ( f ( x ))) = D (2 · ln( x )) = 2x D (ln ( f ( x ))) = D ( x · ln(2)) = ln(2)
f ′ ( x ) = f ( x ) · D (ln ( f ( x ))) = x2 · 2
x = 2x f ′ ( x ) = f ( x ) · D (ln ( f ( x ))) = 2x · ln(2)

Example 5. Use the pattern f ′ ( x ) = f ( x ) · D (ln( f ( x ))) to find the

derivative of f ( x ) = (3x + 7)5 sin(2x ).

Solution. Apply the natural logarithm to both sides and rewrite:

 
ln ( f ( x )) = ln (3x + 7)5 · sin(2x ) = 5 ln(3x + 7) + ln (sin(2x ))

so:

D (ln( f ( x ))) = D (5 ln(3x + 7) + ln (sin(2x )))

3 cos(2x )
= 5· +2·
3x + 7 sin(2x )

Then:

f ′ ( x ) = f ( x ) · D (ln( f ( x )))
 
15 cos(2x )
= (3x + 7)5 sin(2x ) +2·
3x + 7 sin(2x )
= 15(3x + 7)4 sin(2x ) + 2(3x + 7)5 cos(2x )

the same result we would obtain using the Product Rule. ◀

212 contemporary calculus

Practice 5. Use logarithmic differentiation to find the derivative of

f ( x ) = (2x + 1)3 (3x2 − 4)7 ( x + 7)4 .

We could have differentiated the functions in the previous Example

and Practice problem without logarithmic differentiation. There are,
however, functions for which logarithmic differentiation is the only
method we can use. We know how to differentiate x raised to a constant
power, D ( x p ) = p · x p−1 , and a constant to a variable power, D (b x ) =
b x ln(b), but the function f ( x ) = x x has both a variable base and a
variable power, so neither differentiation rule applies. We need to use
logarithmic differentiation.

Example 6. Find D ( x x ), assuming that x > 0.

Solution. Apply the natural logarithm to both sides and rewrite:

ln ( f ( x )) = ln ( x x ) = x · ln( x )

so:

D (ln ( f ( x ))) = D ( x · ln( x )) = x · D (ln( x )) + ln( x ) · D( x )

1
= x· + ln( x ) · 1 = 1 + ln( x )
x
Then D ( x x ) = f ′ ( x ) = f ( x ) D (ln ( f ( x ))) = x x (1 + ln( x )). ◀
 
Practice 6. Find D xsin( x) assuming that x > 0.

Logarithmic differentiation is an alternate method for differentiating

some functions such as products and quotients, and it is the only
method we’ve seen for differentiating some other functions such as
variable bases to variable exponents.

2.9 Problems

dy
In Problems 1–10 find dx in two ways: (a) by differ- x2 y2
5. + = 1, point: (0, 4)
entiating implicitly and (b) by explicitly solving for 9 16
dy
y and then differentiating. Then find the value of dx x2 y2
6. + = 1, point: (3, 0)
at the given point using your results from both the 9 16
implicit and the explicit differentiation. 7. ln(y) + 3x − 7 = 0, point: (2, e)
1. x2 + y2 = 100, point: (6, 8)
8. x2 − y2 = 16, point: (5, 3)
2. x2 + 5y2 = 45, point: (5, 2)
9. x2 − y2 = 16, point: (5, −3)
3. x2 − 3xy + 7y = 5, point: (2, 1)
√ √
4. x + y = 5, point: (4, 9) 10. y2 + 7x3 − 3x = 8, point: (1, 2)
 the derivative 213

11. Find the slopes of the lines tangent to the graph

 2 √
20. x 2 + y2 + 1 − 4x2 = 81, point: (0, 2 2)
below at the points (3, 1), (3, 3) and (4, 2).
2 2
21. x 3 + y 3 = 5, point: (8, 1)
22. x + cos( xy) = y + 3, point: (2, 0)
23. Find the slope of the line tangent to the ellipse
shown in the figure below at the point (1, 2).

12. Find the slopes of the lines tangent to the graph

in the figure above at the points where the graph
crosses the y-axis.
13. Find the slopes of the lines tangent to the graph
below at the points (5, 0), (5, 6) and (−4, 3).
24. Find the slopes of the tangent lines at the points
where the ellipse shown above crosses the y-axis.
25. Find y′ for y = Ax2 + Bx + C and for the equation
x = Ay2 + By + C.
26. Find y′ for y = Ax3 + B and for x = Ay3 + B.
27. Find y′ for Ax2 + Bxy + Cy2 + Dx + Ey + F = 0.
28. In Chapter 1 we assumed that the tangent line to
a circle at a point was perpendicular to the radial
line passing through that point and the center of
the circle. Use implicit differentiation to prove
that the line tangent to the circle x2 + y2 = r2 (see
below) at an arbitrary point ( x, y) is perpendicu-
14. Find the slopes of the lines tangent to the graph
lar to the line passing through (0, 0) and ( x, y).
in the figure above at the points where the graph
crosses the y-axis.

dy
In Problems 15–22 , find dx using implicit differenti-
ation and then find the slope of the line tangent to
the graph of the equation at the given point.
15. y3 − 5y = 5x2 + 7, point: (1, 3)
16. y2 − 5xy + x2 + 21 = 0, point: (2, 5)
17. y2 + sin(y) = 2x − 6, point: (3, 0)
18. y + 2x2 y3 = 4x + 7, point: (3, 1)
19. ey + sin(y) = x2 − 3, point: (2, 0)
 214 contemporary calculus

Problems 29–31 use the figure from Problems 23–24. x f (x) ln ( f ( x )) D (ln ( f ( x ))) f ′ (x)
29. Find the coordinates of point A where the tangent 1 5 1.6 −1
49.
line to the ellipse is horizontal. 2 2 0.7 0
30. Find the coordinates of point B where the tangent 3 7 1.9 2
line to the ellipse is vertical.
x g( x ) ln ( g( x )) D (ln ( g( x ))) g′ ( x )
31. Find the coordinates of points C and D.
dy 50. 2 1.4 0.3 1.2
In 32–40, find dx in two ways: (a) by using the
3 3.3 1.2 0.6
“usual” differentiation patterns and (b) by using log-
7 13.6 2.6 0.2
arithmic differentiation.
Problems 51–55 illustrate how logarithmic differen-
32. y = x · sin(3x ) 33. y = ( x2 + 5)7 ( x 3 − 1)4 tiation can be used to verify some differentiation
patterns we already know (51–52, 54) and to derive
sin(3x − 1) 35. y = x5 · (3x + 2)4 some new patterns (53, 55). Assume that all of the
34. y =
x+7 functions are differentiable and that the function
36. y = 7x 37. y = esin( x) combinations are defined.

√ 51. Use logarithmic differentiation on f · g to re-

38. y = cos7 (2x + 5) 39. y = 25 − x2 derive the Product Rule: D ( f · g) = f · g′ + g · f ′ .
f
x · cos( x ) 52. Use logarithmic differentiation on to re-derive
40. y = g
x2 + 1  
f g · f ′ − f · g′
dy the quotient rule: D = .
In 41–46, use logarithmic differentiation to find dx .
g g2
53. Use logarithmic differentiation to obtain a prod-
41. y = xcos( x)
uct rule for three functions: D ( f · g · h) = ?.
42. y = (cos( x )) x 54. Use logarithmic differentiation on the exponential
43. y = x4 · ( x − 2)7 · sin(3x ) function a x (with a > 0) to show that its deriva-
√
x + 10 tive is a x ln( a).
44. y =
(2x + 3)3 · (5x − 1)7 55. Use logarithmic differentiation to determine a
45. y = (3 + sin( x )) x pattern for the derivative of f g : D ( f g ) = ?.
56. In Section 2.1 we proved the Power Rule D( x n ) =
s
x2 + 1
46. y = n · x n−1 for any positive integer n.
x2 − 1
(a) Why does this formula hold for n = 0?
In 47–50, use the values in each table to calculate the
(b) Use the Quotient Rule to prove that D( x −m ) =
values of the derivative in the last column.
−m · x −m−1 for any positive integer m and con-
x f (x) ln ( f ( x )) D (ln ( f ( x ))) f ′ (x) clude that the Power Rule holds for all integers.
p

47. 1 1 0.0 1.2 (c) Now let y = x q where p and q are integers so
2 9 2.2 1.8 that yq = x p . Use implicit differentiation to
3 64 4.2 2.1 show that the Power Rule holds for all rational
exponents. (We still have not considered the
x g( x ) ln ( g( x )) D (ln ( g( x ))) g′ ( x ) case where y = x a with a an irrational num-
1 5 1.6 0.6 ber, because we haven’t actually defined what
48.
2 10 2.3 0.7 x a means for a irrational. We will take care of
3 20 3.0 0.8 that — and the extension of the Power Rule to
all real exponents — in Chapter 7.)
 the derivative 215

2.9 Practice Answers

1. D( x2 + y2 ) = 2x + 2y · y′
d ′
dx (sin(2 + 3y )) = cos(2 + 3y ) · D (2 + 3y ) = cos(2 + 3y ) · 3y

1 2 − 2   1 2 − 2
3 3
2. Explicitly: y′ = 3x + 15 D 3x2 + 15 = 3x + 15 (6x ).
3 3
2 2
= 13 3(2)2 + 15 3 (6 · 2) = 4 (27)− 3 = 49 .
When ( x, y) =(2, 3), y′ 

Implicitly: D y3 − 3x2 = D(15) ⇒ 3y2 · y′ − 6x = 0 so y′ = 2x y2

.
2 · 2 4
When ( x, y) = (2, 3), y′ = 2 = .
3 9
3. D (y + sin(y)) = D x − x ⇒ y′ + cos(y) · y′ = 3x2 − 1 ⇒ y′ ·
3


2
(1 + cos(y)) = 3x2 − 1, so we have y′ = 1+3xcos−(1y) . When ( x, y) =
3(1)2 − 1
(1, 0), y′ = = 1.
1 + cos(0)

√ set x = 0 and
4. To find where the parabola crosses the y-axis, we can
7± (−7)2 −4(1)(2)
solve for the values of y: y2 − 7y + 2 = 0 ⇒ y = 2(1)
=
√
7± 41
2 ≈ 0.3 and 6.7. The parabola crosses the y-axis (approximately)
at the points (0, 0.3) and (0, 6.7). From Example 4, we know that
−2x − 2y − 3
y′ = , so at the point (0, 0.3), the slope is approxi-
2x + 2y − 7
0 − 0.6 − 3
mately ≈ 0.56, and at the point (0, 6.7), the slope is
0 + 0.6 − 7
0 − 13.4 − 3
approximately ≈ −2.56.
0 + 13.4 − 7
5. Applying the formula f ′ ( x ) = f ( x ) · D (ln ( f ( x ))) to the function
f ( x ) = (2x + 1)3 (3x2 − 4)7 ( x + 7)4 , we have:

ln ( f ( x )) = 3 · ln(2x + 1) + 7 · ln(3x2 − 4) + 4 · ln( x + 7)

so:
3 7 4
D (ln ( f ( x ))) = (2) + 2 (6x ) + (1)
2x + 1 3x − 4 x+7
and thus:
 
6 42x 4
f ′ ( x ) = f ( x ) · D (ln ( f ( x ))) = (2x + 1)3 (3x2 − 4)7 ( x + 7)4 · + +
2x + 1 3x2 − 4 x + 7

6. Using f ′ ( x ) = f ( x ) · D (ln ( f ( x ))) with f ( x ) = xsin( x) :

 
ln ( f ( x )) = ln xsin( x) = sin( x ) · ln( x )

so:
1
D (ln ( f ( x ))) = D (sin( x ) · ln( x )) = sin( x ) · D (ln( x )) + ln( x ) · D(sin( x )) = sin( x ) · + ln( x ) · cos( x )
x
and thus:
 
sin( x )
f ′ ( x ) = f ( x ) · D (ln ( f ( x ))) = xsin( x) · + ln( x ) · cos( x )
x
3
Derivatives and Graphs

In this chapter, we explore what the first and second derivatives of a

function tell us about the graph of that function and apply this graphical
knowledge to locate the extreme values of a function.

3.1 Finding Maximums and Minimums

In theory and applications, we often want to maximize or minimize

some quantity. An engineer may want to maximize the speed of a
new computer or minimize the heat produced by an appliance. A
manufacturer may want to maximize profits and market share or mini-
mize waste. A student may want to maximize a grade in calculus or
minimize the hours of study needed to earn a particular grade.
Many natural objects follow minimum or maximum principles, so
if we want to model natural phenomena we may need to maximize or
minimize. A light ray travels along a “minimum time” path. The shape
and surface texture of some animals tend to minimize or maximize
heat loss. Systems reach equilibrium when their potential energy is
minimized. A basic tenet of evolution is that a genetic characteristic
that maximizes the reproductive success of an individual will become
more common in a species.
Calculus provides tools for analyzing functions and their behavior
and for finding maximums and minimums.

Methods for Finding Maximums and Minimums

We can try to find where a function f is largest or smallest by evaluating
f at lots of values of x, a method that is not very efficient and may
not find the exact place where f achieves its extreme value. If we try
hundreds or thousands of values for x, however, then we can often find
a value of f that is close to the maximum or minimum. In general, this
type of exhaustive search is only practical if you have a computer do
the work.
218 contemporary calculus

The graph of a function provides a visual way of examining lots

of values of f , and it is a good method, particularly if you have a
computer to do the work for you. It is still inefficient, however, as
you (or a computer) still need to evaluate the function at hundreds or
thousands of inputs in order to create the graph — and we still may not
find the exact location of the maximum or minimum.
Calculus provides ways to drastically narrow the number of points
we need to examine to find the exact locations of maximums and
minimums. Instead of examining f at thousands of values of x, calculus
can often guarantee that the maximum or minimum must occur at one
of three or four values of x, a substantial improvement in efficiency.

A Little Terminology
Before we examine how calculus can help us find maximums and
minimums, we need to carefully define these concepts.

Definitions:

• f has a maximum or global maximum at x = a

if f ( a) ≥ f ( x ) for all x in the domain of f .

• The maximum value of f is then f ( a)

and this maximum value of f occurs at a.

• The maximum point on the graph of f is ( a, f ( a)).

The previous definition involves the overall biggest value a function

attains on its entire domain. We are sometimes interested in how a
function behaves locally rather than globally.

Definition: f has a local or relative maximum at x = a if f ( a) ≥ f ( x )

for all x “near” a, (that is, in some open interval that contains a).

Global and local minimums are defined similarly by replacing the

≥ symbol with ≤ in the previous definitions.

Definition:

f has a global extreme at x = a

if f ( a) is a global maximum or minimum.

See the margin figure for graphical examples of local and global
extremes of a function.
 derivatives and graphs 219

You should notice that every global extreme is also a local extreme,
but there are local extremes that are not global extremes. If h( x ) is the
height of the earth above sea level at location x, then the global max-
imum of h is h(summit of Mt. Everest) = 29,028 feet. The local maxi-
mum of h for the United States is h(summit of Mt. McKinley) = 20,320
feet. The local minimum of h for the United States is h(Death Valley) =
-282 feet.

Finding Maximums and Minimums of a Function

One way to narrow our search for a maximum value of a function f is
to eliminate those values of x that, for some reason, cannot possibly
make f maximum.

Theorem:
If f ′ ( a) > 0 or f ′ ( a) < 0
then f ( a) is not a local maximum or minimum.

Proof. Assume that f ′ ( a) > 0. By definition:

f ( a + ∆x ) − f ( a)
f ′ ( a) = lim
∆x →0 ∆x

f ( a + ∆x ) − f ( a)
so f ′ ( a) = lim > 0. This means that the right and
∆x →0 ∆x
f ( a + ∆x ) − f ( a)
left limits are both positive: f ′ ( a) = lim > 0 and
∆x →0 + ∆x
f ( a + ∆x ) − f ( a)
f ′ ( a) = lim > 0.
∆x →0 − ∆x
Considering the right limit, we know that if we restrict ∆x > 0 to
f ( a + ∆x ) − f ( a)
be sufficiently small, we can guarantee that > 0 so,
∆x
multiplying each side of this last inequality by the positive number ∆x,
we have f ( a + ∆x ) − f ( a) > 0 ⇒ f ( a + ∆x ) > f ( a) for all sufficiently
small values of ∆x > 0, so any open interval containing x = a will also
contain values of x with f ( x ) > f ( a). This tell us that f ( a) is not a
maximum.
Considering the left limit, we know that if we restrict ∆x < 0 to
f ( a + ∆x ) − f ( a)
be sufficiently small, we can guarantee that > 0 so,
∆x
multiplying each side of this last inequality by the negative number ∆x,
we have f ( a + ∆x ) − f ( a) < 0 ⇒ f ( a + ∆x ) < f ( a) for all sufficiently
small values of ∆x < 0, so any open interval containing x = a will also
contain values of x with f ( x ) < f ( a). This tell us that f ( a) is not a
minimum.
The argument for the “ f ′ ( a) < 0” case is similar.
220 contemporary calculus

When we evaluate the derivative of a function f at a point x = a,

there are only four possible outcomes: f ′ ( a) > 0, f ′ ( a) < 0, f ′ ( a) = 0
or f ′ ( a) is undefined. If we are looking for extreme values of f , then
we can eliminate those points at which f ′ is positive or negative, and
only two possibilities remain: f ′ ( a) = 0 or f ′ ( a) is undefined.

Theorem:
If f is defined on an open interval
and f ( a) is a local extreme of f
then either f ′ ( a) = 0 or f is not differentiable at a.

Example 1. Find the local extremes of f ( x ) = x3 − 6x2 + 9x + 2.

Solution. An extreme value of f can occur only where f ′ ( x ) = 0 or

where f is not differentiable; f ( x ) is a polynomial, so it is differentiable
for all values of x, and we can restrict our attention to points where
f ′ ( x ) = 0.

f ′ ( x ) = 3x2 − 12x + 9 = 3( x2 − 4x + 3) = 3( x − 1)( x − 3)

so f ′ ( x ) = 0 only at x = 1 and x = 3.
The only possible locations of local extremes of f are at x = 1 and
x = 3. We don’t know yet whether f (1) or f (3) is a local extreme of
f , but we can be certain that no other point is a local extreme. The
graph of f (see margin) shows that (1, f (1)) = (1, 6) appears to be a
local maximum and (3, f (3)) = (3, 2) appears to be a local minimum.
This function does not have a global maximum or minimum. ◀

Practice 1. Find the local extremes of f ( x ) = x2 + 4x − 5 and of g( x ) =

2x3 − 12x2 + 7.

It is important to recognize that the two conditions “ f ′ ( a) = 0” or “ f

not differentiable at a” do not guarantee that f ( a) is a local maximum
or minimum. They only say that f ( a) might be a local extreme or that
f ( a) is a candidate for being a local extreme.

Example 2. Find all local extremes of f ( x ) = x3 .

Solution. f ( x ) = x3 is differentiable for all x, and f ′ ( x ) = 3x2 equals

0 only at x = 0, so the only candidate is the point (0, 0). But if x > 0
then f ( x ) = x3 > 0 = f (0), so f (0) is not a local maximum. Similarly,
if x < 0 then f ( x ) = x3 < 0 = f (0) so f (0) is not a local minimum.
The point (0, 0) is the only candidate to be a local extreme of f , but
this candidate did not turn out to be a local extreme of f . The function
f ( x ) = x3 does not have any local extremes. ◀
 derivatives and graphs 221

If f ′ ( a) = 0 or f is not differentiable at a
then the point ( a, f ( a)) is a candidate to be a local extreme
but may not actually be a local extreme.

Practice 2. Sketch the graph of a differentiable function f that satisfies

the conditions: f (1) = 5, f (3) = 1, f (4) = 3 and f (6) = 7; f ′ (1) = 0,
f ′ (3) = 0, f ′ (4) = 0 and f ′ (6) = 0; the only local maximums of f are
at (1, 5) and (6, 7); and the only local minimum is at (3, 1).

Is f ( a) a Maximum or Minimum or Neither?

Once we have found the candidates ( a, f ( a)) for extreme points of f , we
still have the problem of determining whether the point is a maximum,
a minimum or neither.
One method involves graphing (or letting a calculator graph) the
function near a, and then drawing a conclusion from the graph. All
of the graphs in the margin have f (2) = 3, and on each of the graphs
f ′ (2) either equals 0 or is undefined. It is clear from the graphs that the
point (2, 3) is: a local maximum in (a) and (d); a local minimum in (b)
and (e); and not a local extreme in (c) and (f).
In Sections 3.3 and 3.4, we will investigate how information about
the first and second derivatives of f can help determine whether the
candidate ( a, f ( a)) is a maximum, a minimum or neither.

Endpoint Extremes
So far we have discussed finding extreme values of functions over the
entire real number line or on an open interval, but in practice we may
need to find the extreme of a function over some closed interval [c, d].
If an extreme value of f occurs at x = a between c and d (c < a < d)
then the previous reasoning and results still apply: either f ′ ( a) = 0 or
f is not differentiable at a. On a closed interval, however, there is one
more possibility: an extreme can occur at an endpoint of the closed
interval (see margin): at x = c or x = d.
We can extend our definition of a local extreme at x = a (which
requires f ( a) ≥ f ( x ) [or f ( a) ≤ f ( x )] for all x in some open interval
containing a) to include x = a being the endpoint of a closed interval:
f ( a) ≥ f ( x ) [or f ( a) ≤ f ( x )] for all x in an interval of the form [ a, a + h)
(for left endpoints) or ( a − h, a] (for right endpoints), where h > 0 is a
number small enough to guarantee the “half-open” interval is in the
domain of f ( x ). Using this extended definition, the function in the
margin has a local maximum (which is also a global maximum) at x = c
and a local minimum (also a global minimum) at x = d.
222 contemporary calculus

Practice 3. List all of the extremes ( a, f ( a)) of the function in the

margin figure on the interval [1, 4] and state whether f ′ ( a) = 0, f is not
differentiable at a, or a is an endpoint.

Example 3. Find the extreme values of f ( x ) = x3 − 3x2 − 9x + 5 for

−2 ≤ x ≤ 6.

Solution. We need to find investigate points where f ′ ( x ) = 0, points

where f is not differentiable, and the endpoints:

• f ′ ( x ) = 3x2 − 6x − 9 = 3( x + 1)( x − 3), so f ′ ( x ) = 0 only at x = −1

and x = 3.

• f is a polynomial, so it is differentiable everywhere.

• The endpoints of the interval are x = −2 and x = 6.

Altogether we have four points in the interval to examine, and any

extreme values of f can only occur when x is one of those four points:
f (−2) = 3, f (−1) = 10, f (3) = −22 and f (6) = 59. The (global)
minimum of f on [−2, 6] is −22 when x = 3, and the (global) maximum
of f on [−2, 6] is 59 when x = 6. ◀

Sometimes the function we need to maximize or minimize is more

complicated, but the same methods work.
√
Example 4. Find the extreme values of f ( x ) = 13 64 + x2 + 15 (10 − x )
for 0 ≤ x ≤ 10.

Solution. This function comes from an application we will examine in

section 3.5. The only possible locations of extremes are where f ′ ( x ) = 0
or f ′ ( x ) is undefined or where x is an endpoint of the interval [0, 10].
 
1 1 1

2
f ′ (x) = D 64 + x2 + (10 − x )
3 5
1 1 1 1
= · (64 + x2 )− 2 · 2x −
3 2 5
x 1
= √ −
3 64 + x2 5

To find where f ′ ( x ) = 0, set the derivative equal to 0 and solve for x:

x 1 x 1 x2 1
√ − =0 ⇒ √ = ⇒ 2
=
3 64 + x2 5 3 64 + x 2 5 576 + 9x 25
2
⇒ 16x = 576 ⇒ x = ±6

but only x = 6 is in the interval [0, 10]. Evaluating f at this point gives
f (6) ≈ 4.13.
We can evaluate the formula for f ′ ( x ) for any value of x, so the
derivative is always defined.
 derivatives and graphs 223

Finally, the interval [0, 10] has two endpoints, x = 0 and x = 10, and
f (0) ≈ 4.67 while f (10) ≈ 4.27.
The maximum of f on [0, 10] must occur at one of the points (0, 4.67),
(6, 4.13) and (10, 4.27), and the minimum must occur at one of these
three points as well.
The maximum value of f is 4.67 at x = 0, and the minimum value
of f is 4.13 at x = 6. ◀

Practice 4. Rework the previous Example to find the extreme values of

√
f ( x ) = 13 64 + x2 + 15 (10 − x ) for 0 ≤ x ≤ 5.

Critical Numbers
The points at which a function might have an extreme value are called
critical numbers.
Definitions: A critical number for a function f is a value x = a
in the domain of f so that:

• f ′ ( a) = 0 or

• f is not differentiable at a or

• a is an endpoint of a closed interval to which f is restricted.

If we are trying to find the extreme values of f on an open interval

c < x < d or on the entire number line, then the set of inputs to which
f is restricted will not include any endpoints, so we will not need to
worry about any endpoint critical numbers.
We can now give a very succinct description of where to look for
extreme values of a function.

An extreme value of f can only occur at a critical number.

The critical numbers only give possible locations of extremes; some

critical numbers are not locations of extremes. In other words, criti-
cal numbers are the candidates for the locations of maximums and
minimums. Section 3.5 is devoted entirely to translat-
ing and solving maximum and minimum
problems.
224 contemporary calculus

Which Functions Have Extremes?

Some functions don’t have extreme values: Example 2 showed that
f ( x ) = x3 (defined on the entire number line) did not have a maximum
or minimum.

Example 5. Find the extreme values of f ( x ) = x.

Solution. Because f ′ ( x ) = 1 > 0 for all x, the first theorem in this

section guarantees that f has no extreme values. The function f ( x ) = x
does not have a maximum or minimum on the real number line. ◀

With the previous function, the domain was so large that we could
always make the function output larger or smaller than any given value
by choosing an appropriate input x. The next example shows that we
can encounter the same difficulty even on a “small” interval.
1
Example 6. Show that f ( x ) = does not have a maximum or minimum
x
on the interval (0, 1).

Solution. f is continuous for all x ̸= 0 so f is continuous on the

1
interval (0, 1). For 0 < x < 1, f ( x ) = > 0 and for any number a
x
strictly between 0 and 1, we can show that f ( a) is neither a maximum
nor a minimum of f on (0, 1), as follows.
Pick b to be any number between 0 and a: 0 < b < a. Then
1 1
f (b) = > = f ( a), so f ( a) is not a maximum. Similarly, pick c to be
How would the situation change if we b a
changed the interval in this example to 1 1
(0, 1]? To [1, 2]? any number between a and 1: a < c < 1. Then f ( a) = > = f (c),
a c
so f ( a) is not a minimum. The interval (0, 1) is not “large,” yet f does
not attain an extreme value anywhere in (0, 1). ◀

The Extreme Value Theorem provides conditions that guarantee a

function to have a maximum and a minimum.

Extreme Value Theorem:

If f is continuous on a closed interval [ a, b]
then f attains both a maximum and minimum on [ a, b].

The proof of this theorem is difficult, so we omit it. The margin figure
illustrates some of the possibilities for continuous and discontinuous
functions on open and closed intervals.
The Extreme Value Theorem guarantees that certain functions (con-
tinuous ones) on certain intervals (closed ones) must have maximums
and minimums. Other functions on other intervals may or may not
have maximums and minimums.
 derivatives and graphs 225

3.1 Problems

1. Label all of the local maximums and minimums 23. Sketch the graph of a continuous function f with:
of the function in the figure below. Also label all (a) f (1) = 3, f ′ (1) = 0 and the point (1, 3) a rela-
of the critical points. tive maximum of f .
(b) f (2) = 1, f ′ (2) = 0 and the point (2, 1) a rela-
tive minimum of f .
(c) f (3) = 5, f is not differentiable at x = 3, and
the point (3, 5) a relative maximum of f .
(d) f (4) = 7, f is not differentiable at x = 4, and
the point (4, 7) a relative minimum of f .
(e) f (5) = 4, f ′ (5) = 0 and the point (5, 4) not a
2. Label the local extremes and critical points of the relative minimum or maximum of f .
function graphed below.
(f) f (6) = 3, f not differentiable at 6, and (6, 3)
not a relative minimum or maximum of f .

In Problems 24–37, find all critical points and local

extremes of each function on the given intervals.

24. f ( x ) = x2 − 6x + 5 on the entire real number line

25. f ( x ) = x2 − 6x + 5 on [−2, 5]
26. f ( x ) = 2 − x3 on the entire real number line
In Problems 3–22, find all of the critical points and
local maximums and minimums of each function. 27. f ( x ) = 2 − x3 on [−2, 1]
28. f ( x ) = x3 − 3x + 5 on the entire real number line
3. f ( x ) = x2 + 8x + 7 4. f ( x ) = 2x2 − 12x + 7
29. f ( x ) = x3 − 3x + 5 on [−2, 1]
5. f ( x ) = sin( x ) 6. f ( x ) = x3 − 6x2 + 5
30. f ( x ) = x5 − 5x4 + 5x3 + 7 on (−∞, ∞)
√
3
7. f ( x ) = x 8. f ( x ) = 5x − 2 31. f ( x ) = x5 − 5x4 + 5x3 + 7 on [0, 2]
√3 1
9. f ( x ) = xe5x 10. f ( x ) = 1 + x2 32. f ( x ) = on (−∞, ∞)
x2
+1
11. f ( x ) = ( x − 1)2 ( x − 3) 33. f ( x ) = 2
1
on [1, 3]
x +1
12. f ( x ) = ln( x2 − 6x + 11) p
34. f ( x ) = 3 x2 + 4 − x on (−∞, ∞)
13. f ( x ) = 2x3 − 96x + 42 p
35. f ( x ) = 3 x2 + 4 − x on [0, 2]
14. f ( x ) = 5x + cos(2x + 1)
36. f ( x ) = xe−5x on (−∞, ∞)
−( x −2)2 16. f ( x ) = | x + 5|
15. f ( x ) = e
h i
37. f ( x ) = x3 − ln( x ) on 12 , 2
x x3
17. f ( x ) = 18. f ( x ) = 38. (a) Find two numbers whose sum is 22 and whose
1 + x2 1 + x4 product is as large as possible. (Suggestion:
2
23 call the numbers x and 22 − x).
19. f ( x ) = ( x − 2) 3 20. f ( x ) = x2 − 1
(b) Find two numbers whose sum is A > 0 and
√
3
√
3
21. f ( x ) = x2 −4 22. f ( x ) = x−2 whose product is as large as possible.
 226 contemporary calculus

39. Find the coordinates of the point in the first quad- 42. Find the radius and height of the cylinder that
rant on the circle x2 + y2 = 1 so that the rectangle has the largest volume (V = πr2 h) if the sum of
in the figure below has the largest possible area. the radius and height is 9.

43. Suppose you are working with a polynomial of

degree 3 on a closed interval.

(a) What is the largest number of critical points

the function can have on the interval?
(b) What is the smallest number of critical points
it can have?
(c) What are the patterns for the most and fewest
40. Find the coordinates of the point in the first quad-
critical points a polynomial of degree n on a
rant on the ellipse 9x2 + 16y2 = 144 so that the
closed interval can have?
rectangle in the figure below has:
(a) the largest possible area. 44. Suppose you have a polynomial of degree 3 di-
(b) The smallest possible area. vided by a polynomial of degree 2 on a closed
interval.

(a) What is the largest number of critical points

the function can have on the interval?
(b) What is the smallest number of critical points
it can have?

45. Suppose f (1) = 5 and f ′ (1) = 0. What can we

conclude about the point (1, 5) if:

(a) f ′ ( x ) < 0 for x < 1 and f ′ ( x ) > 0 for x > 1?

(b) f ′ ( x ) < 0 for x < 1 and f ′ ( x ) < 0 for x > 1?

41. Find the value for x so the box shown below has: (c) f ′ ( x ) > 0 for x < 1 and f ′ ( x ) < 0 for x > 1?

(a) the largest possible volume. (d) f ′ ( x ) > 0 for x < 1 and f ′ ( x ) > 0 for x > 1?
(b) The smallest possible volume.
46. Suppose f (2) = 3 and f is continuous but not
differentiable at x = 2. What can we conclude
about the point (2, 3) if:

(a) f ′ ( x ) < 0 for x < 2 and f ′ ( x ) > 0 for x > 2?

(b) f ′ ( x ) < 0 for x < 2 and f ′ ( x ) < 0 for x > 2?
(c) f ′ ( x ) > 0 for x < 2 and f ′ ( x ) < 0 for x > 2?
(d) f ′ ( x ) > 0 for x < 2 and f ′ ( x ) > 0 for x > 2?
 derivatives and graphs 227

47. The figure below shows the graph of f ′ ( x ), which 52. Imagine the graph of f ( x ) = x2 . Does f have a
is continuous on (0, 12) except at x = 8. maximum value for x in the given interval?
(a) Which values of x are critical points of f ( x )?
(b) At which values of x does f attain a local max- (a) [−2, 3] (b) [−2, 3) (c) (−2, 3]
imum?
(c) At which values of x does f attain a local min- (d) [−2, 1) (e) (−2, 1]
imum?
53. Imagine the graph of f ( x ) = x2 . Does f have a
minimum value for x in the interval I?

(a) [−2, 3] (b) [−2, 3) (c) (−2, 3]

(d) [−2, 1) (e) (−2, 1]

54. Define A( x ) to be the area bounded between the

48. The figure below shows the graph of f ′ ( x ), which t-axis, the graph of y = f (t) and a vertical line at
is continuous on (0, 13) except at x = 7. t = x (see figure below).
(a) Which values of x are critical points? (a) At what value of x is A( x ) minimum?
(b) At which values of x does f attain a local max- (b) At what value of x is A( x ) maximum?
imum?
(c) At which values of x does f attain a local min-
imum?

55. Define S( x ) to be the slope of the line through

49. State the contrapositive form of the Extreme Value
the points (0, 0) and ( x, f ( x )) in the figure below.
Theorem.
50. Imagine the graph of f ( x ) = 1 − x. Does f have (a) At what value of x is S( x ) minimum?
a maximum value for x in the given interval? (b) At what value of x is S( x ) maximum?

(a) [0, 2] (b) [0, 2) (c) (0, 2]

(d) (0, 2) (e) (1, π ]

51. Imagine the graph of f ( x ) = 1 − x. Does f have

a minimum value for x in the given interval?

(a) [0, 2] (b) [0, 2) (c) (0, 2]

(d) (0, 2) (e) (1, π ]

228 contemporary calculus

3.1 Practice Answers

1. f ( x ) = x2 + 4x − 5 is a polynomial so f is differentiable for all x
and f ′ ( x ) = 2x + 4; f ′ ( x ) = 0 when x = −2 so the only candidate
for a local extreme is x = −2. Because the graph of f is a parabola
opening up, the point (−2, f (−2)) = (−2, −9) is a local minimum.
g( x ) = 2x3 − 12x2 + 7 is a polynomial so g is differentiable for all
x and g′ ( x ) = 6x2 − 24x = 6x ( x − 4) so g′ ( x ) = 0 when x = 0 or
4, so the only candidates for a local extreme are x = 0 and x = 4.
The graph of g (see margin) indicates that g has a local maximum at
(0, 7) and a local minimum at (4, −57).

2. See the margin figure.

x f (x) f ′ (x) max/min
1 5 0 local max
3 1 0 local min
4 3 0 neither
6 7 0 local max

3. (1, f (1)) is a global minimum; x = 1 is an endpoint

(2, f (2)) is a local maximum; f ′ (2) = 0
(3, f (3)) is a local/global minimum; f is not differentiable at x = 3
(4, f (4)) is a global maximum; x = 4 is an endpoint

4. This is the same function used in Example 4, but now the interval is
[0, 5] instead of [0, 10]. See the Example for the calculations.
Critical points:

• endpoints: x = 0 and x = 5
• f is differentiable for all 0 < x < 5: none
• f ′ ( x ) = 0: none in [0, 5]

f (0) ≈ 4.67 is the maximum of f on [0, 5];

f (5) ≈ 4.14 is the minimum of f on [0, 5].
 derivatives and graphs 229

3.2 Mean Value Theorem

If you averaged 30 miles per hour during a trip, then at some instant
during the trip you were traveling exactly 30 miles per hour.
That relatively obvious statement is the Mean Value Theorem as it
applies to a particular trip. It may seem strange that such a simple
statement would be important or useful to anyone, but the Mean Value
Theorem is important and some of its consequences are very useful in
a variety of areas. Many of the results in the rest of this chapter depend
on the Mean Value Theorem, and one of the corollaries of the Mean
Value Theorem will be used every time we calculate an “integral” in
later chapters. A truly delightful aspect of mathematics is that an idea
as simple and obvious as the Mean Value Theorem can be so powerful.
Before we state and prove the Mean Value Theorem and examine
some of its consequences, we will consider a simplified version called
Rolle’s Theorem.

Rolle’s Theorem
Pick any two points on the x-axis and think about all of the differ-
entiable functions that pass through those two points. Because our
functions are differentiable, they must be continuous and their graphs
cannot have any holes or breaks. Also, since these functions are dif-
ferentiable, their derivatives are defined everywhere between our two
points and their graphs can not have any “corners” or vertical tangents.
The graphs of the functions in the margin figure can still have all
sorts of shapes, and it may seem unlikely that they have any common
properties other than the ones we have stated, but Michel Rolle (1652–
1719) found one. He noticed that every one of these functions has one
or more points where the tangent line is horizontal (see margin), and
this result is named after him.
Rolle’s Theorem:
If f ( a) = f (b)
and f ( x ) is continuous for a ≤ x ≤ b
and differentiable for a < x < b
then there is at least one number c between a and b so that
f ′ (c) = 0.

Proof. We consider three cases: when f ( x ) = f ( a) for all x in ( a, b),

when f ( x ) > f ( a) for some x in ( a, b), and when f ( x ) < f ( a) for some
x in ( a, b).
Case I: If f ( x ) = f ( a) for all x between a and b, then the graph of
f is a horizontal line segment and f ′ (c) = 0 for all values of c strictly
between a and b.
230 contemporary calculus

Case II: Suppose f ( x ) > f ( a) for some x in ( a, b). Because f is

continuous on the closed interval [ a, b], we know from the Extreme
Value Theorem that f must attain a maximum value on the closed
interval [ a, b]. Because f ( x ) > f ( a) for some value of x in [ a, b], then
Notice that Rolle’s Theorem tells us that the maximum of f must occur at some value c strictly between a and
(at least one) number c with the required
b: a < c < b. (Why can’t the maximum be at a or b?) Because f (c) is a
properties exists, but does not tell us how
to find c. local maximum of f , c is a critical number of f , meaning f ′ (c) = 0 or
f ′ (c) is undefined. But f is differentiable at all x between a and b, so
the only possibility is that f ′ (c) = 0.
Case III: Suppose f ( x ) < f ( a) for some x in ( a, b). Then, arguing
as we did in Case II, f attains a minimum at some value x = c strictly
between a and b, and so f ′ (c) = 0.
In each case, there is at least one value of c between a and b so that
f ′ (c) = 0.

Example 1. Show that f ( x ) = x3 − 6x2 + 9x + 2 satisfies the hypotheses

of Rolle’s Theorem on the interval [0, 3] and find a value of c that the
theorem tells you must exist.

Solution. Because f is a polynomial, it is continuous and differentiable

everywhere. Furthermore, f (0) = 2 = f (3), so Rolle’s Theorem applies.
Differentiating:

f ′ ( x ) = 3x2 − 12x + 9 = 3( x − 1)( x − 3)

so f ′ ( x ) = 0 when x = 1 and when x = 3. The value c = 1 is between

0 and 3. The margin figure shows a graph of f . ◀

Practice 1. Find the value(s) of c for Rolle’s Theorem for the functions
graphed below.

The Mean Value Theorem

Geometrically, the Mean Value Theorem is a “tilted” version of Rolle’s
Theorem (see margin). In each theorem we conclude that there is a
number c so that the slope of the tangent line to f at x = c is the same
as the slope of the line connecting the two ends of the graph of f on the
interval [ a, b]. In Rolle’s Theorem, the two ends of the graph of f are
at the same height, f ( a) = f (b), so the slope of the line connecting the
ends is zero. In the Mean Value Theorem, the two ends of the graph
 derivatives and graphs 231

of f do not have to be at the same height, so the line through the two
ends does not have to have a slope of zero.

Mean Value Theorem:

If f ( x ) is continuous for a ≤ x ≤ b
and differentiable for a < x < b
then there is at least one number c between a and b so the
line tangent to the graph of f at x = c is parallel to
the secant line through ( a, f ( a)) and (b, f (b)):

f (b) − f ( a)
f ′ (c) =
b−a

Proof. The proof of the Mean Value Theorem uses a tactic common in
mathematics: introduce a new function that satisfies the hypotheses of
some theorem we already know and then use the conclusion of that
previously proven theorem. For the Mean Value Theorem we introduce
a new function, h( x ), which satisfies the hypotheses of Rolle’s Theorem.
Then we can be certain that the conclusion of Rolle’s Theorem is true for
h( x ) and the Mean Value Theorem for f will follow from the conclusion
of Rolle’s Theorem for h.
First, let g( x ) be the linear function passing through the points
( a, f ( a)) and (b, f (b)) of the graph of f . The function g goes through
the point ( a, f ( a)) so g( a) = f ( a). Similarly, g(b) = f (b). The slope of
f (b) − f ( a) f (b) − f ( a)
the linear function g( x ) is so g′ ( x ) = for all x
b−a b−a
between a and b, and g is continuous and differentiable. (The formula
f (b)− f ( a)
for g is g( x ) = f ( a) + m( x − a) with m = b− a .)
Define h( x ) = f ( x ) − g( x ) for a ≤ x ≤ b (see margin). The function
h satisfies the hypotheses of Rolle’s theorem:

• h( a) = f ( a) − g( a) = 0 and h(b) = f (b) − g(b) = 0

• h( x ) is continuous for a ≤ x ≤ b because both f and g are continuous

there

• h( x ) is differentiable for a < x < b because both f and g are differ-

entiable there

so the conclusion of Rolle’s Theorem applies to h: there is a c between

a and b so that h′ (c) = 0.
The derivative of h( x ) = f ( x ) − g( x ) is h′ ( x ) = f ′ ( x ) − g′ ( x ) so we
know that there is a number c between a and b with h′ (c) = 0. But:
f (b) − f ( a)
0 = h′ (c) = f ′ (c) − g′ (c) ⇒ f ′ (c) = g′ (c) =
b−a
which is exactly what we needed to prove.
232 contemporary calculus

Graphically, the Mean Value Theorem says that there is at least one
point c where the slope of the tangent line, f ′ (c), equals the slope of
the line through the end points of the graph segment, ( a, f ( a)) and
(b, f (b)). The margin figure shows the locations of the parallel tangent
lines for several functions and intervals.
The Mean Value Theorem also has a very natural interpretation
if f ( x ) represents the position of an object at time x: f ′ ( x ) repre-
f (b) − f ( a)
sents the velocity of the object at the instant x and =
b−a
change in position
represents the average (mean) velocity of the ob-
change in time
ject during the time interval from time a to time b. The Mean Value
Theorem says that there is a time c (between a and b) when the instan-
taneous velocity, f ′ (c), is equal to the average velocity for the entire
f (b) − f ( a)
trip, . If your average velocity during a trip is 30 miles per
b−a
hour, then at some instant during the trip you were traveling exactly 30
miles per hour.

Practice 2. For f ( x ) = 5x2 − 4x + 3 on the interval [1, 3], calculate

f (b) − f ( a)
m= and find the value(s) of c so that f ′ (c) = m.
b−a

Some Consequences of the Mean Value Theorem

If the Mean Value Theorem was just an isolated result about the exis-
tence of a particular point c, it would not be very important or useful.
However, the Mean Value Theorem is the basis of several results about
the behavior of functions over entire intervals, and it is these conse-
quences that give it an important place in calculus for both theoretical
and applied uses.
The next two corollaries are just the first of many results that follow
from the Mean Value Theorem.
We already know, from the Main Differentiation Theorem, that the
derivative of a constant function f ( x ) = K is always 0, but can a non-
constant function have a derivative that is always 0? The first corollary
says no.

Corollary 1:

If f ′ ( x ) = 0 for all x in an interval I

then f ( x ) = K, a constant, for all x in I.

Proof. Assume f ′ ( x ) = 0 for all x in an interval I. Pick any two points

a and b (with a ̸= b) in the interval. Then, by the Mean Value Theorem,
f (b) − f ( a)
there is a number c between a and b so that f ′ (c) = . By
b−a
′ ′
our assumption, f ( x ) = 0 for all x in I, so we know that 0 = f (c) =
 derivatives and graphs 233

f (b) − f ( a)
and thus f (b) − f ( a) = 0 ⇒ f (b) = f ( a). But a and b were
b−a
two arbitrary points in I, so the value of f ( x ) is the same for any two
values of x in I, and f is a constant function on the interval I.

We already know that if two functions are “parallel” (differ by a

constant), then their derivatives are equal, but can two non-parallel
functions have the same derivative? The second corollary says no.

Corollary 2:

If f ′ ( x ) = g′ ( x ) for all x in an interval I

then f ( x ) − g( x ) = K, a constant, for all x in I, so the
graphs of f and g are “parallel” on the interval I.

Proof. This corollary involves two functions instead of just one, but
we can imitate the proof of the Mean Value Theorem and introduce
a new function h( x ) = f ( x ) − g( x ). The function h is differentiable
and h′ ( x ) = f ′ ( x ) − g′ ( x ) = 0 for all x in I so, by Corollary 1, h( x ) is a
constant function and K = h( x ) = f ( x ) − g( x ) for all x in the interval.
Thus f ( x ) = g( x ) + K.

We will use Corollary 2 hundreds of times in Chapters 4 and 5 when

we work with “integrals.” Typically you will be given the derivative
of a function, f ′ ( x ), and be asked to find all functions f that have that
derivative. Corollary 2 tells us that if we can find one function f that
has the derivative we want, then the only other functions that have the
same derivative are of the form f ( x ) + K where K is a constant: once
you find one function with the right derivative, you have essentially
found all of them.
Example 2. (a) Find all functions whose derivatives equal 2x. (b) Find
a function g( x ) with g′ ( x ) = 2x and g(3) = 5.

Solution. (a) Observe that f ( x ) = x2 ⇒ f ′ ( x ) = 2x, so one func-

tion with the derivative we want is f ( x ) = x2 . Corollary 2 guar-
antees that every function g whose derivative is 2x has the form
g( x ) = f ( x ) + K = x2 + K. (b) Because g′ ( x ) = 2x, we know that
g must have the form g( x ) = x2 + K, but this gives a whole “family”
of functions (see margin) and we want to find one member of that
family. We also know that g(3) = 5 so we want to find the member
of the family that passes through the point (3, 5). Replacing g( x ) with
5 and x with 3 in the formula g( x ) = x2 + K, we can solve for the
value of K: 5 = g(3) = (3)2 + K ⇒ K = −4. The function we want is
g( x ) = x2 − 4. ◀

Practice 3. Restate Corollary 2 as a statement about the positions and

velocities of two cars.
234 contemporary calculus

3.2 Problems

1. In the figure below, find the number(s) “c” that 9. If I take off in an airplane, fly around for a while
Rolle’s Theorem promises (guarantees). and land at the same place I took off from, then
my starting and stopping heights are the same
but the airplane is always moving. Why doesn’t
this violate Rolle’s theorem, which says there is
an instant when my velocity is 0?
10. Prove the following corollary of Rolle’s Theorem:
If P( x ) is a polynomial, then between any two
roots of P there is a root of P′ .
For Problems 2–4, verify that the hypotheses of
Rolle’s Theorem are satisfied for each of the func- 11. Use the corollary in Problem 10 to justify the con-
tions on the given intervals, and find the value of clusion that the only root of f ( x ) = x3 + 5x − 18
the number(s) “c” that Rolle’s Theorem promises. is 2. (Suggestion: What could you conclude about
f ′ if f had another root?)
2. (a) f ( x ) = x2 on [−2, 2]
(b) f ( x ) = x2 − 5x + 8 on [0, 5] 12. In the figure below, find the location(s) of the “c”
that the Mean Value Theorem promises.
3. (a) f ( x ) = sin( x ) on [0, π ]
(b) f ( x ) = sin( x ) on [π, 5π ]
4. (a) f ( x ) = x3 − x + 3 on [−1, 1]
(b) f ( x ) = x · cos( x ) on [0, π2 ]

5. Suppose you toss a ball straight up and catch it

when it comes down. If h(t) is the height of the
ball t seconds after you toss it, what does Rolle’s
Theorem say about the velocity of the ball? Why
is it easier to catch a ball that someone on the
ground tosses up to you on a balcony, than for
you to be on the ground and catch a ball that
In Problems 13–15, verify that the hypotheses of the
someone on a balcony tosses down to you?
Mean Value Theorem are satisfied for each of the
1
6. If f ( x ) = 2 , then f (−1) = 1 and f (1) = 1 but functions on the given intervals, and find the num-
x
2 ber(s) “c” that the Mean Value Theorem guarantees.
f ′ ( x ) = − 3 is never equal to 0. Why doesn’t
x
this function violate Rolle’s Theorem? 13. (a) f ( x ) = x2 on [0, 2]
7. If f ( x ) = | x |, then f (−1) = 1 and f (1) = 1 but (b) f ( x ) = x2 − 5x + 8 on [1, 5]
f ′ ( x ) is never equal to 0. Why doesn’t this func-
14. (a) f ( x ) = sin( x ) on [0, π2 ]
tion violate Rolle’s Theorem?
(b) f ( x ) = x3 on [−1, 3]
8. If f ( x ) = x2 , then f ′ ( x ) = 2x is never 0 on the √
interval [1, 3]. Why doesn’t this function violate 15. (a) f ( x ) = 5 − x on [1, 9]
Rolle’s Theorem? (b) f ( x ) = 2x + 1 on [1, 7]
 derivatives and graphs 235

16. For the quadratic functions in parts (a) and (b) of 24. Sketch the graphs of several members of the “fam-
Problem 13, the number c turned out to be the ily” of functions whose derivatives always equal
midpoint of the interval: c = a+2 b . 3. Give a formula that defines every function in
this family.
(a) For f ( x ) = 3x2 + x − 7 on [1, 3], show that
f (3) − f (1) 25. Sketch the graphs of several members of the “fam-
f ′ (2) = . ily” of functions whose derivatives always equal
3−1
(b) For f ( x ) = x2 − 5x + 3 on [2, 5], show that 3x2 . Give a formula that defines every function
7 f (5) − f (2) in this family.
f′ = .
2 5−2 26. At t seconds after takeoff, the upward velocity
2
(c) For f ( x ) =
 Ax + Bx + C on [ a, b], show that of a helicopter was v(t) = 3t2 + 2t feet/second.
a+b f (b) − f ( a) Two seconds after takeoff, the helicopter was 80
f′ = .
2 b−a feet above sea level. Find a formula for the height
17. If f ( x ) = | x |, then f (−1) = 1 and f (3) = 3 but of the helicopter at every time t.
f (3) − f (−1) 1 27. Assume that a rocket is fired from the ground
f ′ ( x ) is never equal to = . Why
3 − (−1) 2 and has the upward velocity shown in the figure
doesn’t this violate the Mean Value Theorem?
below. Estimate the height of the rocket when
t = 1, 2 and 5 seconds.
In Problems 18–19, you are a traffic-court judge. In
each case, a driver has challenged a speeding ticket
and you need to decide if the ticket is appropriate.

18. A tolltaker says, “Your Honor, based on the

elapsed time from when the car entered the toll
road until the car stopped at my booth, I know
the average speed of the car was 83 miles per
hour. I did not actually see the car speeding, but
I know it was and I gave the driver a ticket.”

19. The driver in the next case says, “Your Honor, my

average velocity on that portion of the toll road 28. The figure below shows the upward velocity of
was only 17 miles per hour, so I could not have a rocket. Use the information in the graph to
been speeding.” estimate the height of the rocket when t = 1, 2
20. Find three different functions ( f , g and h) so that and 5 seconds.
f ′ ( x ) = g′ ( x ) = h′ ( x ) = cos( x ).

21. Find a function f so that f ′ ( x ) = 3x2 + 2x + 5

and f (1) = 10.

22. Find g( x ) so that g′ ( x ) = x2 + 3 and g(0) = 2.

23. Find values for A and B so that the graph of the

parabola f ( x ) = Ax2 + B is:

(a) tangent to y = 4x + 5 at the point (1, 9).

(b) tangent to y = 7 − 2x at the point (2, 3).
(c) tangent to y = x2 + 3x − 2 at the point (0, 2).
 236 contemporary calculus

29. Determine a formula for f ( x ) if you know: 33. Define A( x ) to be the area bounded by the t-axis,
f ′′ ( x ) = 6, f ′ (0) = 4 and f (0) = −5. the line y = 2t + 1 and a vertical line at t = x.
30. Determine a formula for g( x ) if you know: (a) Find a formula for A( x ).
g′′ ( x ) = 12x, g′ (1) = 9 and g(2) = 30. (b) Determine A′ ( x ).
31. Define A( x ) to be the area bounded by the t-axis,
the line y = 3 and a vertical line at t = x.
(a) Find a formula for A( x ).
(b) Determine A′ ( x ).

In Problems 34–36, given a list of numbers a1 , a2 , a3 ,

a4 , . . . , the consecutive differences between num-
bers in the list are: a2 − a1 , a3 − a2 , a4 − a3 , . . .

34. If a1 = 5 and the consecutive difference is always

32. Define A( x ) to be the area bounded by the t-axis, 0, what can you conclude about the numbers in
the line y = 2t and a vertical line at t = x. the list?
(a) Find a formula for A( x ). 35. If a1 = 5 and the consecutive difference is always
(b) Determine A′ ( x ). 3, find a formula for an .
36. Suppose the “a” list starts with 3, 4, 7, 8, 6, 10,
13,. . . , and there is a “b” list that has the same
consecutive differences as the “a” list.
(a) If b1 = 5, find the next six numbers in the “b”
list. How is bn related to an ?
(b) If b1 = 2, find the next six numbers in the “b”
list. How is bn related to an ?
(c) If b1 = B, find the next six numbers in the “b”
list. How is bn related to an ?
 derivatives and graphs 237

3.2 Practice Answers

1. f ′ ( x ) = 0 when x = 2 and 6, so c = 2 and c = 6.
g′ ( x ) = 0 when x = 2, 4 and 6, so c = 2, c = 4 and c = 6.

2. With f ( x ) = 5x2 − 4x + 3 on [1, 3], f (1) = 4 and f (3) = 36 so:

f (b) − f ( a) 36 − 4
m= = = 16
b−a 3−1
f ′ ( x ) = 10x − 4 so f ′ (c) = 10c − 4 = 16 ⇒ 10c = 20 ⇒ c = 2. The
graph of f showing the location of c appears below.

3. If two cars have the same velocities during an interval of time (so that
f ′ (t) = g′ (t) for t in I) then the cars are always a constant distance
apart during that time interval. (Note: “Same velocity” means same
speed and same direction. If two cars are traveling at the same
speed but in different directions, then the distance between them
changes and is not constant.)
238 contemporary calculus

3.3 The First Derivative and the Shape of f

This section examines some of the interplay between the shape of the
graph of a function f and the behavior of its derivative, f ′ . If we have a
graph of f , we will investigate what we can conclude about the values
of f ′ . And if we know values of f ′ , we will investigate what we can
conclude about the graph of f .

Definitions: Given any interval I, a function f is . . .

In this definition, I can be of the form increasing on I if, for all x1 and x2 in I, x1 < x2 ⇒ f ( x1 ) < f ( x2 )
( a, b), [ a, b), ( a, b], [ a, b], (−∞, b), (−∞, b],
decreasing on I if, for all x1 and x2 in I, x1 < x2 ⇒ f ( x1 ) > f ( x2 )
( a, ∞), [ a, ∞) or (−∞, ∞), where a < b.
monotonic on I if f is increasing or decreasing on I

Graphically, f is increasing (decreasing) if, as we move from left to

right along the graph of f , the height of the graph increases (decreases).
These same ideas make sense if we consider h(t) to be the height (in
feet) of a rocket at time t seconds. We naturally say that the rocket is
rising or that its height is increasing if the height h(t) increases over a
period of time, as t increases.

Example 1. List the intervals on which the function graphed below is

increasing or decreasing.

Solution. f is increasing on the intervals [0, 0.3] (approximately), [2, 3]

and [4, 6]. f is decreasing on (approximately) [0.3, 2] and [6, 8]. On
the interval [3, 4] the function is not increasing or decreasing — it is
constant. It is also valid to say that f is increasing on the intervals
[0.5, 0.8] and (0.5, 0.8) as well as many others, but we usually talk about
the longest intervals on which f is monotonic. ◀

Practice 1. List the intervals on which the function graphed below is

increasing or decreasing.
 derivatives and graphs 239

If we have an accurate graph of a function, then it is relatively easy

to determine where f is monotonic, but if the function is defined by a
formula, then a little more work is required. The next two theorems
relate the values of the derivative of f to the monotonicity of f . The
first theorem says that if we know where f is monotonic, then we also
know something about the values of f ′ . The second theorem says that
if we know about the values of f ′ then we can draw conclusions about
where f is monotonic.

First Shape Theorem:

For a function f that is differentiable on an interval ( a, b):

• if f is increasing on ( a, b) then f ′ ( x ) ≥ 0 for all x in ( a, b)

• if f is decreasing on ( a, b) then f ′ ( x ) ≤ 0 for all x in ( a, b)

• if f is constant on ( a, b), then f ′ ( x ) = 0 for all x in ( a, b)

Proof. Most people find a picture such as the one in the margin to be
a convincing justification of this theorem: if the graph of f increases
near a point ( x, f ( x )), then the tangent line is also increasing, and the
slope of the tangent line is positive (or perhaps zero at a few places). A
more precise proof, however, requires that we use the definitions of the
derivative of f and of “increasing” (given above).
Case I: Assume that f is increasing on ( a, b). We know that f is
differentiable, so if x is any number in the interval ( a, b) then

f ( x + h) − f ( x )
f ′ ( x ) = lim
h →0 h

and this limit exists and is a finite value. If h is any small enough
positive number so that x + h is also in the interval ( a, b), then x <
x + h ⇒ f ( x ) < f ( x + h) (by the definition of “increasing”). We know
that the numerator, f ( x + h) − f ( x ), and the denominator, h, are both
positive, so the limiting value, f ′ ( x ), must be positive or zero: f ′ ( x ) ≥ 0.
Case II: Assume that f is decreasing on ( a, b). If x < x + h, then
f ( x ) > f ( x + h) (by the definition of “decreasing”). So the numerator The proof of this part is very similar to
of the limit, f ( x + h) − f ( x ), will be negative but the denominator, h, the “increasing” proof.

will still be positive, so the limiting value, f ′ ( x ), must be negative or

zero: f ′ ( x ) ≤ 0.
Case III: The derivative of a constant is 0, so if f is constant on ( a, b)
then f ′ ( x ) = 0 for all x in ( a, b).

The previous theorem is easy to understand, but you need to pay

attention to exactly what it says and what it does not say. It is possible
for a differentiable function that is increasing on an interval to have
horizontal tangent lines at some places in the interval (see margin). It is
240 contemporary calculus

also possible for a continuous function that is increasing on an interval

to have an undefined derivative at some places in the interval. Finally,
it is possible for a function that is increasing on an interval to fail to be
continuous at some places in the interval (see margin).
The First Shape Theorem has a natural interpretation in terms of the
height h(t) and upward velocity h′ (t) of a helicopter at time t. If the
height of the helicopter is increasing (h(t) is an increasing function),
then the helicopter has a positive or zero upward velocity: h′ (t) ≥ 0. If
the height of the helicopter is not changing, then its upward velocity is
0: h′ (t) = 0.

Example 2. A figure in the margin shows the height of a helicopter

during a period of time. Sketch the graph of the upward velocity of the
dh
helicopter, .
dt
dh
Solution. The graph of v(t) = appears in the margin. Notice that
dt
h(t) has a local maximum when t = 2 and t = 5, and that v(2) = 0
and v(5) = 0. Similarly, h(t) has a local minimum when t = 3, and
v(3) = 0. When h is increasing, v is positive. When h is decreasing, v is
negative. ◀

Practice 2. A figure in the margin shows the population of rabbits

on an island during a 6-year period. Sketch the graph of the rate of
dR
population change, , during those years.
dt

Example 3. A graph of f appears in the margin; sketch a graph of f ′ .

Solution. It is a good idea to look first for the points where f ′ ( x ) = 0

or where f is not differentiable (the critical points of f ). These locations
are usually easy to spot, and they naturally break the problem into
several smaller pieces. The only numbers at which f ′ ( x ) = 0 are x = −1
and x = 2, so the only places the graph of f ′ ( x ) will cross the x-axis
are at x = −1 and x = 2: we can therefore plot the points (−1, 0) and
(2, 0) on the graph of f ′ . The only place where f is not differentiable is
at the “corner” above x = 5, so the graph of f ′ will not be defined for
x = 5. The rest of the graph of f is relatively easy to sketch:

• if x < −1 then f ( x ) is decreasing so f ′ ( x ) is negative

• if −1 < x < 2 then f ( x ) is increasing so f ′ ( x ) is positive

• if 2 < x < 5 then f ( x ) is decreasing so f ′ ( x ) is negative

• if 5 < x then f ( x ) is decreasing so f ′ ( x ) is negative

 derivatives and graphs 241

A graph of f ′ appears on the previous page: f ( x ) is continuous at

x = 5, but not differentiable at x = 5 (indicated by the “hole”). ◀

Practice 3. A graph of f appears in the margin. Sketch a graph of f ′ .

(The graph of f has a “corner” at x = 5.)

The next theorem is almost the converse of the First Shape Theorem
and explains the relationship between the values of the derivative and
the graph of a function from a different perspective. It says that if
we know something about the values of f ′ , then we can draw some
conclusions about the shape of the graph of f .

Second Shape Theorem:

For a function f that is differentiable on an interval I:

• if f ′ ( x ) > 0 for all x in the interval I, then f is increasing on I

• if f ′ ( x ) < 0 for all x in the interval I, then f is decreasing on I

• if f ′ ( x ) = 0 for all x in the interval I, then f is constant on I

Proof. This theorem follows directly from the Mean Value Theorem,
and the last part is just a restatement of the First Corollary of the Mean
Value Theorem.
Case I: Assume that f ′ ( x ) > 0 for all x in I and pick any points a
and b in I with a < b. Then, by the Mean Value Theorem, there is a
f (b) − f ( a)
point c between a and b so that = f ′ (c) > 0 and we can
b−a
conclude that f (b) − f ( a) > 0, which means that f (b) > f ( a). Because
a < b ⇒ f ( a) < f (b), we know that f is increasing on I.
Case II: Assume that f ′ ( x ) < 0 for all x in I and pick any points
a and b in I with a < b. Then there is a point c between a and b so
f (b) − f ( a)
that = f ′ (c) < 0, and we can conclude that f (b) − f ( a) =
b−a
(b − a) f ′ (c) < 0 so f (b) < f ( a). Because a < b ⇒ f ( a) > f (b), we
know f is decreasing on I.

Practice 4. Rewrite the Second Shape Theorem as a statement about the

height h(t) and upward velocity h′ (t) of a helicopter at time t seconds.

The value of a function f at a number x tells us the height of the

graph of f above or below the point ( x, 0) on the x-axis. The value
of f ′ at a number x tells us whether the graph of f is increasing or
decreasing (or neither) as the graph passes through the point ( x, f ( x ))
on the graph of f . If f ( x ) is positive, it is possible for f ′ ( x ) to be
positive, negative, zero or undefined: the value of f ( x ) has absolutely
nothing to do with the value of f ′ . The margin figure illustrates some
of the possible combinations of values for f and f ′ .
242 contemporary calculus

Practice 5. Graph a continuous function that satisfies the conditions on

f and f ′ given below:

x −2 −1 0 1 2 3
f (x) 1 −1 −2 −1 0 2
f ′ (x) −1 0 1 2 −1 1

The Second Shape Theorem can be particularly useful if we need to

graph a function f defined by a formula. Between any two consecutive
critical numbers of f , the graph of f is monotonic (why?). If we can
find all of the critical numbers of f , then the domain of f will be broken
naturally into a number of pieces on which f will be monotonic.

Example 4. Use information about the values of f ′ to help graph

f ( x ) = x3 − 6x2 + 9x + 1.

Solution. f ′ ( x ) = 3x2 − 12x + 9 = 3( x − 1)( x − 3) so f ′ ( x ) = 0 only

when x = 1 or x = 3; f ′ is a polynomial, so it is always defined. The
only critical numbers, x = 1 and x = 3, break the real number line into
three pieces on which f is monotonic: (−∞, 1), (1, 3) and (3, ∞).

• x < 1 ⇒ f ′ ( x ) = 3(negative)(negative) > 0 ⇒ f increasing

• 1 < x < 3 ⇒ f ′ ( x ) = 3(positive)(negative) < 0 ⇒ f is decreasing

• 3 < x ⇒ f ′ ( x ) = 3(positive)(positive) > 0 ⇒ f is increasing

Although we don’t yet know the value of f anywhere, we do know a lot

about the shape of the graph of f : as we move from left to right along
the x-axis, the graph of f increases until x = 1, then decreases until
x = 3, after which the graph increases again (see margin). The graph
of f “turns” when x = 1 and x = 3. To plot the graph of f , we still
need to evaluate f at a few values of x, but only at a very few values:
f (1) = 5, and (1, 5) is a local maximum of f ; f (3) = 1, and (3, 1) is a
local minimum of f . A graph of f appears in the margin. ◀

Practice 6. Use information about the values of f ′ to help graph the

function f ( x ) = x3 − 3x2 − 24x + 5.

Example 5. Use the graph of f ′ in the margin to sketch the shape of the
graph of f . Why isn’t the graph of f ′ enough to completely determine
the graph of f ?

Solution. Several functions that have the derivative we want appear

in the margin, and each provides a correct answer. By the Second
Corollary to the Mean Value Theorem, we know there is a whole family
of “parallel” functions that share the derivative we want, and each
 derivatives and graphs 243

of these functions provides a correct answer. If we had additional

information about the function — such as a point it passes through —
then only one member of the family would satisfy the extra condition
and there would be only one correct answer. ◀

Practice 7. Use the graph of g′ provided in the margin to sketch the

shape of a graph of g.

Practice 8. A weather balloon is released from the ground and sends

back its upward velocity measurements (see margin). Sketch a graph
of the height of the balloon. When was the balloon highest?

Using the Derivative to Test for Extremes

The first derivative of a function tells about the general shape of the
function, and we can use that shape information to determine whether
an extreme point is a (local) maximum or minimum or neither.

First Derivative Test for Local Extremes:

Let f be a continuous function with f ′ (c) = 0 or f ′ (c) undefined.

• If f ′ (left of c) > 0 and f ′ (right of c) < 0

then (c, f (c)) is a local maximum.

• If f ′ (left of c) < 0 and f ′ (right of c) > 0

then (c, f (c)) is a local minimum.

• If f ′ (left of c) > 0 and f ′ (right of c) > 0

then (c, f (c)) is not a local extreme.

• If f ′ (left of c) < 0 and f ′ (right of c) < 0

then (c, f (c)) is not a local extreme.

Practice 9. Find all extremes of f ( x ) = 3x2 − 12x + 7 and use the First
Derivative Test to classify them as maximums, minimums or neither.

3.3 Problems
In Problems 1–3, sketch the graph of the derivative of each function.

1. 2. 3.
 244 contemporary calculus

Problems 4–6 show the graph of the height of a helicopter; sketch a graph of its upward velocity.

4. 5. 6.

7. In the figure below, match the graphs of the func- 8. Match the graphs showing the heights of rockets
tions with those of their derivatives. with those showing their velocities.

9. Use the Second Shape Theorem to show that 11. A student is working with a complicated function
f ( x ) = ln( x ) is monotonic increasing on the in- f and has shown that the derivative of f is always
terval (0, ∞). positive. A minute later the student also claims
10. Use the Second Shape Theorem to show that that f ( x ) = 2 when x = 1 and when x = π. With-
g( x ) = e x is monotonic increasing on the entire out checking the student’s work, how can you be
real number line. certain that it contains an error?
 derivatives and graphs 245

12. The figure below shows the graph of the deriva- 13. The figure below shows the graph of the deriva-
tive of a continuous function f . tive of a continuous function g.
(a) List the critical numbers of f . (a) List the critical numbers of g.
(b) What values of x result in a local maximum? (b) What values of x result in a local maximum?
(c) What values of x result in a local minimum?
(c) What values of x result in a local minimum?

Problems 14–16 show the graphs of the upward velocities of three helicopters. Use the graphs to determine
when each helicopter was at a (relative) maximum or minimum height.

14. 15. 16.

In 17–22, use information from the derivative of each 25. h( x ) = x3 + 9x − 10 has a root at x = 1. Without
function to help you graph the function. Find all graphing h, show that h has no other roots.
local maximums and minimums of each function.
26. Sketch the graphs of monotonic decreasing func-
17. f ( x ) = x3 − 3x2 − 9x − 5
tions that have exactly (a) no roots (b) one root
18. g( x ) = 2x3 − 15x2 + 6
and (c) two roots.
19. h( x ) = x4 − 8x2 + 3
20. s(t) = t + sin(t) 27. Each of the following statements is false. Give (or
2 sketch) a counterexample for each statement.
21. r (t) = 2
t +1
x2 + 3 (a) If f is increasing on an interval I, then f ′ ( x ) >
22. f ( x ) =
x 0 for all x in I.
23. f ( x ) = 2x + cos( x ) so f (0) = 1. Without graph-
(b) If f is increasing and differentiable on I, then
ing the function, you can be certain that f has
f ′ ( x ) > 0 for all x in I.
how many positive roots?
24. g( x ) = 2x − cos( x ) so g(0) = −1. Without graph- (c) If cars A and B always have the same speed,
ing the function, you can be certain that g has how then they will always be the same distance
many positive roots? apart.
 246 contemporary calculus

28. (a) Find several different functions f that all have 30. Sketch functions with the given properties to help
the same derivative f ′ ( x ) = 2. determine whether each statement is true or false.
(b) Determine a function f with derivative (a) If f ′ (7) > 0 and f ′ ( x ) > 0 for all x near 7, then
f ′ ( x ) = 2 that also satisfies f (1) = 5. f (7) is a local maximum of f on [1, 7].
(c) Determine a function g with g′ ( x ) = 2 for (b) If g′ (7) < 0 and g′ ( x ) < 0 for all x near 7, then
which the graph of g goes through (2, 1). g(7) is a local minimum of g on [1, 7].
29. (a) Find several different functions h that all have (c) If h′ (1) > 0 and h′ ( x ) > 0 for all x near 1, then
the same derivative h′ ( x ) = 2x. h(1) is a local minimum of h on [1, 7].
(b) Determine a function f with derivative (d) If r ′ (1) < 0 and r ′ ( x ) < 0 for all x near 1, then
f ′ ( x ) = 2x that also satisfies f (3) = 20. r (1) is a local maximum of r on [1, 7].
(c) Determine a function g with g′ ( x ) = 2x for (e) If s′ (7) = 0, then s(7) is a local maximum of s
which the graph of g goes through (2, 7). on [1, 7].

3.3 Practice Answers

1. g is increasing on [2, 4] and [6, 8]; g is decreasing on [0, 2] and [4, 5];
g is constant on [5, 6].
dR
2. The graph in the margin shows the rate of population change, .
dt
3. A graph of f ′ appears below. Notice how the graph of f ′ is 0 where
f has a maximum or minimum.

4. The Second Shape Theorem for helicopters:

• If the upward velocity h′ is positive during time interval I then

the height h is increasing during time interval I.
• If the upward velocity h′ is negative during time interval I then
the height h is decreasing during time interval I.
• If the upward velocity h′ is zero during time interval I then the
height h is constant during time interval I.
 derivatives and graphs 247

5. A graph satisfying the conditions in the table appears in the margin.

x −2 −1 0 1 2 3
f (x) 1 −1 −2 −1 0 2
f ′ (x) −1 0 1 2 −1 1

6. f ′ ( x ) = 3x2 − 6x − 24 = 3( x − 4)( x + 2) so f ′ ( x ) = 0 if x = −2 or
x = 4.

• x < −2 ⇒ f ′ ( x ) = 3(negative)(negative) > 0 ⇒ f increasing

• −2 < x < 4 ⇒ f ′ ( x ) = 3(negative)(positive) < 0 ⇒ f decreasing
• x > 4 ⇒ f ′ ( x ) = 3(positive)(positive) > 0 ⇒ f increasing

Thus f has a relative maximum at x = −2 and a relative minimum

at x = 4. A graph of f appears in the margin.

7. The figure below left shows several possible graphs for g. Each has
the correct shape to give the graph of g′ . Notice that the graphs of g
are “parallel” (differ by a constant).

8. The figure above right shows the height graph for the balloon. The
balloon was highest at 4 p.m. and had a local minimum at 6 p.m.

9. f ′ ( x ) = 6x − 12 so f ′ ( x ) = 0 only if x = 2.

• x < 2 ⇒ f ′ ( x ) < 0 ⇒ f decreasing

• x > 2 ⇒ f ′ ( x ) > 0 ⇒ f increasing

From this we can conclude that f has a minimum when x = 2 and

has a shape similar to graph provided in the margin.
We could also have noticed that the graph of the quadratic function
f ( x ) = 3x2 − 12x + 7 must be an upward-opening parabola.
248 contemporary calculus

3.4 The Second Derivative and the Shape of f

The first derivative of a function provides information about the shape

of the function, so the second derivative of a function provides informa-
tion about the shape of the first derivative, which in turn will provide
additional information about the shape of the original function f .
In this section we investigate how to use the second derivative (and
the shape of the first derivative) to reach conclusions about the shape of
the original function. The first derivative tells us whether the graph of
f is increasing or decreasing. The second derivative will tell us about
the “concavity” of f : whether f is curving upward or downward.

Concavity
Graphically, a function is concave up if its graph is curved with the
opening upward (see margin); similarly, a function is concave down if
its graph opens downward. The concavity of a function can be impor-
tant in applied problems and can even affect billion-dollar decisions.
An Epidemic: Suppose you, as an official at the CDC, must decide
whether current methods are effectively fighting the spread of a dis-
ease — or whether more drastic measures are required. In the margin
figure, f ( x ) represents the number of people infected with the disease
at time x in two different situations. In both cases the number of people
with the disease, f (now), and the rate at which new people are getting
sick, f ′ (now), are the same. The difference is the concavity of f , and
that difference might have a big effect on your decision. In (a), f is
concave down at “now,” and it appears that the current methods are
starting to bring the epidemic under control; in (b), f is concave up,
and it appears that the epidemic is growing out of control.
Usually it is easy to determine the concavity of a function by exam-
ining its graph, but we also need a definition that does not require a
graph of the function, a definition we can apply to a function described
by a formula alone.

Definition: Let f be a differentiable function.

• f is concave up at a if the graph of f is above the tangent line L

to f for all x close to (but not equal to) a:

f ( x ) > L( x ) = f ( a) + f ′ ( a)( x − a)

• f is concave down at a if the graph of f is below the tangent line

L to f for all x close to (but not equal to) a:

f ( x ) < L( x ) = f ( a) + f ′ ( a)( x − a)
 derivatives and graphs 249

The margin figure shows the concavity of a function at several points.

The next theorem provides an easily applied test for the concavity of a
function given by a formula.

The Second Derivative Condition for Concavity:

Let f be a twice differentiable function on an interval I.

(a) f ′′ ( x ) > 0 on I ⇒ f ′ ( x ) increasing on I ⇒ f concave up on I

(b) f ′′ ( x ) < 0 on I ⇒ f ′ ( x ) decreasing on I ⇒ f concave down on I

(c) f ′′ ( a) = 0 ⇒ no information
( f ( x ) may be concave up or concave down or neither at a)

Proof. (a) Assume that f ′′ ( x ) > 0 for all x in I, and let a be any point
in I. We want to show that f is concave up at a, so we need to prove
that the graph of f (see margin) is above the tangent line to f at a:
f ( x ) > L( x ) = f ( a) + f ′ ( a)( x − a) for x close to a. Assume that x
is in I and apply the Mean Value Theorem to f on the interval with
endpoints a and x: there is a number c between a and x so that

f ( x ) − f ( a)
f ′ (c) = ⇒ f ( x ) = f ( a) + f ′ (c)( x − a)
x−a

Because f ′′ > 0 on I, we know that f ′′ > 0 between a and x, so the

Second Shape Theorem tells us that f ′ is increasing between a and
x. We will consider two cases: x > a and x < a.

• If x > a then x − a > 0 and c is in the interval [ a, x ] so a < c.

Because f ′ is increasing, a < c ⇒ f ′ ( a) < f ′ (c). Multiplying each
side of this last inequality by the positive quantity x − a yields
f ′ ( a)( x − a) < f ′ (c)( x − a). Adding f ( a) to each side of this last
inequality, we have:

L( x ) = f ( a) + f ′ ( a)( x − a) < f ( a) + f ′ (c)( x − a) = f ( x )

• If x < a then x − a < 0 and c is in the interval [ x, a] so c < a.

Because f ′ is increasing, c < a ⇒ f ′ (c) < f ′ ( a). Multiplying each
side of this last inequality by the negative quantity x − a yields
f ′ (c)( x − a) > f ′ ( a)( x − a) so:

f ( x ) = f ( a) + f ′ (c)( x − a) > f ( a) + f ′ ( a)( x − a) = L( x )

In each case we see that f ( x ) is above the tangent line L( x ).

(b) The proof of this part is similar.

250 contemporary calculus

(c) Let f ( x ) = x4 , g( x ) = − x4 and h( x ) = x3 (see margin). The second

derivative of each of these functions is zero at a = 0, and at (0, 0)
they all have the same tangent line: L( x ) = 0 (the x-axis). However,
at (0, 0) they all have different concavity: f is concave up, while g is
concave down and h is neither concave up nor concave down.

Practice 1. Use the graph of f in the lower margin figure to finish filling
in the table with “+” for positive, “−” for negative or “0.”

x f (x) f ′ (x) f ′′ ( x ) concavity

1 + + − down
2 +
3 −
4

“Feeling” the Second Derivative

Earlier we saw that if a function f (t) represents the position of a car at
time t, then f ′ (t) gives the velocity and f ′′ (t) the acceleration of the car
at the instant t.
If we are driving along a straight, smooth road, then what we feel is
the acceleration of the car:
• a large positive acceleration feels like a “push” toward the back of
the car

• a large negative acceleration (a deceleration) feels like a “push”

toward the front of the car

• an acceleration of 0 for a period of time means the velocity is constant

and we do not feel pushed in either direction
In a moving vehicle it is possible to measure these “pushes,” the
acceleration, and from that information to determine the velocity of the
vehicle, and from the velocity information to determine the position.
Inertial guidance systems in airplanes use this tactic: they measure
front–back, left–right and up–down acceleration several times a second
and then calculate the position of the plane. They also use computers
to keep track of time and the rotation of the earth under the plane.
After all, in six hours the Earth has made a quarter of a revolution, and
Dallas has rotated more than 5,000 miles!
Example 1. The upward acceleration of a rocket was a(t) = 30 m/sec2
during the first six seconds of flight, 0 ≤ t ≤ 6. The velocity of the
rocket at t = 0 was 0 m/sec and the height of the rocket above the
ground at t = 0 was 25 m. Find a formula for the height of the rocket
at time t and determine the height at t = 6 seconds.
 derivatives and graphs 251

Solution. v′ (t) = a(t) = 30 ⇒ v(t) = 30t + K for some constant K. We

also know v(0) = 0 so 30(0) + K = 0 ⇒ K = 0 and this v(t) = 30t.
Similarly, h′ (t) = v(t) = 30t ⇒ h(t) = 15t2 + C for some constant
C. We know that h(0) = 25 so 15(0)2 + C = 25 ⇒ C = 25. Thus
h(t) = 15t2 + 25 so h(6) = 15(6)2 + 25 = 565 m. ◀

f ′′ and Extreme Values of f

The concavity of a function can also help us determine whether a critical
point is a maximum or minimum or neither. For example, if a point is
at the bottom of a concave-up function then that point is a minimum.

The Second Derivative Test for Extremes:

Let f be a twice differentiable function.

(a) If f ′ (c) = 0 and f ′′ (c) < 0

then f is concave down and has a local maximum at x = c.

(b) If f ′ (c) = 0 and f ′′ (c) > 0

then f is concave up and has a local minimum at x = c.

(c) If f ′ (c) = 0 and f ′′ (c) = 0 then f may have a local maximum,

a local minimum or neither at x = c.

Proof. (a) Assume that f ′ (c) = 0. If f ′′ (c) < 0 then f is concave

down at x = c so the graph of f will be below the tangent line
L( x ) for values of x near c. The tangent line, however, is given by
L( x ) = f (c) + f ′ (c)( x − c) = f (c) + 0( x − c) = f (c), so if x is close
to c then f ( x ) < L( x ) = f (c) and f has a local maximum at x = c.

(b) The proof for a local minimum of f is similar.

(c) If f ′ (c) = 0 and f ′′ (c) = 0, then we cannot immediately conclude

anything about local maximums or minimums of f at x = c. The
functions f ( x ) = x4 , g( x ) = − x4 and h( x ) = x3 all have their first
and second derivatives equal to zero at x = 0, but f has a local
minimum at 0, g has a local maximum at 0, and h has neither a local
maximum nor a local minimum at x = 0.

The Second Derivative Test for Extremes is very useful when f ′′ is

easy to calculate and evaluate. Sometimes, however, the First Derivative
Test — or simply a graph of the function — provides an easier way to
determine if the function has a local maximum or a local minimum: it
depends on the function and on which tools you have available.

Practice 2. f ( x ) = 2x3 − 15x2 + 24x − 7 has critical numbers x = 1

and x = 4. Use the Second Derivative Test for Extremes to determine
whether f (1) and f (4) are maximums or minimums or neither.
252 contemporary calculus

Inflection Points
Maximums and minimums typically occur at places where the second
derivative of a function is positive or negative, but the places where the
second derivative is 0 are also of interest.

Definition:

An inflection point is a point on the graph of a function

where the concavity of the function changes, from concave up
to concave down or from concave down to concave up.

Practice 3. Which of the labeled points in the margin figure are inflec-
tion points?

To find the inflection points of a function we can use the second

derivative of the function. If f ′′ ( x ) > 0, then the graph of f is concave
up at the point ( x, f ( x )) so ( x, f ( x )) is not an inflection point. Similarly,
if f ′′ ( x ) < 0 then the graph of f is concave down at the point ( x, f ( x ))
and the point is not an inflection point. The only points left that can
possibly be inflection points are the places where f ′′ ( x ) = 0 or where
f ′′ ( x ) does not exist (in other words, where f ′ is not differentiable). To
find the inflection points of a function we need only check the points
where f ′′ ( x ) is 0 or undefined. If f ′′ (c) = 0 or is undefined, then the
point (c, f (c)) may or may not be an inflection point — we need to
check the concavity of f on each side of x = c. The functions in the
next example illustrate what can happen at such a point.
√
Example 2. Let f ( x ) = x3 , g( x ) = x4 and h( x ) = 3 x (see margin). For
which of these functions is the point (0, 0) an inflection point?

Solution. Graphically, it is clear that the concavity of f ( x ) = x3 and

√
h( x ) = 3 x changes at (0, 0), so (0, 0) is an inflection point for f and h.
The function g( x ) = x4 is concave up everywhere, so (0, 0) is not an
inflection point of g.
f ( x ) = x3 ⇒ f ′ ( x ) = 3x2 ⇒ f ′′ ( x ) = 6x so the only point at which
f ( x ) = 0 or is undefined ( f ′ is not differentiable) is at x = 0. If x < 0
′′

then f ′′ ( x ) < 0 so f is concave down; if x > 0 then f ′′ ( x ) > 0 so f

is concave up. Thus at x = 0 the concavity of f changes so the point
(0, f (0)) = (0, 0) is an inflection point of f ( x ) = x3 .
g( x ) = x4 ⇒ g′ ( x ) = 4x3 ⇒ g′′ ( x ) = 12x2 so the only point at
which g′′ ( x ) = 0 or is undefined is at x = 0. But g′′ ( x ) > 0 (so g is
concave up) for any x ̸= 0. Thus the concavity of g never changes, so
the point (0, g(0)) = (0, 0) is not an inflection point of g( x ) = x4 .
√ 1 1 2 2 5
h( x ) = 3 x = x 3 ⇒ h′ ( x ) = x − 3 ⇒ h′′ ( x ) = − x − 3 so h′′ is not
3 9
defined if x = 0 (and h′′ ( x ) ̸= 0 elsewhere); h′′ (negative number) > 0
 derivatives and graphs 253

and h′′ (positive number) < 0, so h changes concavity at (0, 0) and (0, 0)
√
is an inflection point of h( x ) = 3 x. ◀

Practice 4. Find all inflection points of f ( x ) = x4 − 12x3 + 30x2 + 5x − 7.

Example 3. Sketch a graph of a function with f (2) = 3, f ′ (2) = 1 and

an inflection point at (2, 3).

Solution. Two solutions appear in the margin. ◀

Using f ′ and f ′′ to Graph f

Today you can easily graph most functions of interest using a graphing
calculator — and create even nicer graphs using an app on your phone
or a Web-based graphing utility. Earlier generations of calculus students
did not have these tools, so they relied on calculus to help them draw
graphs of unfamiliar functions by hand. While you can create a graph in
seconds that your predecessors may have labored over for half an hour
or longer, you can still use calculus to help you select an appropriate
graphing “window,” and to be confident that your window has not
missed any points of interest on the graph of a function.
2
Example 4. Create a graph of f ( x ) = xe−9x that shows all local and
global extrema and all inflection points.

Solution. If you graph f ( x ) on a calculator using the standard window

(−10 ≤ x ≤ 10 and −10 ≤ y ≤ 10) you will likely see nothing other
than the coordinate axes (see margin). You might consult a table of
values for the function to help adjust the window, but this trial-and-
error technique will still not guarantee that you have displayed all
points of interest. Computing the first derivative of f , we get:
h 2
i 2
h i 2
f ′ ( x ) = x −18xe−9x + e−9x · 1 = 1 − 18x2 e−9x

which is defined for all values of x; f ′ ( x ) = 0 ⇒ 1 − 18x2 = 0 ⇒

1
x2 = 18 ⇒ x = ± √1 , so the only critical numbers are x = − √1 and
3 2 3 2
1
x= √ . Computing the second derivative of f , we get:
3 2
h i h 2
i 2
f ′′ ( x ) = 1 − 18x2 · −18xe−9x + e−9x · [−36x ]
h i 2
h i 2
= 324x3 − 54x e−9x = 54x 6x2 − 1 e−9x
  q
We can check that f ′′ − 1
√ = 12 2
e > 0, so f must have a local
3 2   q
minimum at x = − 1
√ ; similarly, f ′′ 1
√ = −12 2e < 0, so f must
3 2 3 2
1
have a local maximum at x = √ .
3 2
254 contemporary calculus

Furthermore, f ′′ ( x ) = 0 only when x = 0 or when 6x2 − 1 = 0 ⇒

x = ± √1 , so these three values are candidates for locations of inflection
6
points of f . Noting that:

1 1 1 1
−1 < − √ < − √ < 0 < √ < √ < 1
6 3 2 3 2 6
  q
and that f ′′ (−1) = −270e−9 < 0 and f ′′ − √
1
= 12 2e > 0, we
3 2
observe that f is concave down to the left of x = − √1 and concave up
6
to the right of x = − √1 , so f does in fact have an inflection point at
6  q
x = − √1 . Likewise, f ′′ √ 1
= −12 2e < 0 and f ′′ (1) = 270e−9 > 0,
6 3 2
so f ′′ ( x ) switches sign at x = 0 and at x = √1 , and therefore f ( x )
6
changes concavity at those points as well.
We have now identified two local extrema of f and three inflection
points of f . Equally important, we have used calculus to show that these
five points of interest are the only places where extrema or inflection
points can occur. If we create a graph of f that includes these five points,
our graph is guaranteed to include all “interesting” features of the graph
of f . A window with −1 ≤  x ≤ 1 and −0.2 < y < 0.2 (because the
1
local extreme values are f ± √ ≈ ±0.14) should provide a graph
2 3
(see margin) that includes all five points of interest. ◀

Practice 5. Compute the first and second derivatives of the function

g( x ) = x4 + 4x3 − 90x2 + 13, locate all extrema and inflection points of
g( x ), and create a graph of g( x ) that shows these points of interest.

Most problems in calculus textbooks are Even with calculus, we will typically need calculators or computers
set up to make solving these equations
relatively straightforward, but in general
to help solve the equations f ′ ( x ) = 0 and f ′′ ( x ) = 0 that we use to find
this will not be the case. critical numbers and candidates for inflection points.

3.4 Problems

In Problems 1–2, each statement describes a quan- 2. (a) “The child’s temperature is still rising, but
tity f (t) changing over time. For each statement, tell more slowly than it was a few hours ago.”
what f represents and whether the first and second (b) “The number of whales is decreasing, but at a
derivatives of f are positive or negative.
slower rate than last year.”
1. (a) “Unemployment rose again, but the rate of
(c) “The number of people with the flu is rising
increase is smaller than last month.”
and at a faster rate than last month.”
(b) “Our profits declined again, but at a slower
rate than last month.” 3. Sketch the graphs of functions that are defined
(c) “The population is still rising and at a faster and concave up everywhere and have exactly:
rate than last year.” (a) no roots. (b) 1 root. (c) 2 roots. (d) 3 roots.
 derivatives and graphs 255

4. On which intervals is the function graphed below: 12. At which values of x labeled in the figure below
is the point ( x, g( x )) an inflection point?
(a) concave up? (b) concave down?

13. How many inflection points can a:

(a) quadratic polynomial have?
(b) cubic polynomial have?
5. On which intervals is the function graphed below: (c) polynomial of degree n have?
14. Fill in the table with “+,” “−,”or “0” for the func-
(a) concave up? (b) concave down? tion graphed below.

x f (x) f ′ (x) f ′′ ( x )
0
1
2
3

Problems 6–10 give a function and values of x so

that f ′ ( x ) = 0. Use the Second Derivative Test to
determine whether each point ( x, f ( x )) is a local
maximum, a local minimum or neither.

6. f ( x ) = 2x3 − 15x2 + 6; x = 0, 5
7. g( x ) = x3 − 3x2 − 9x + 7; x = −1, 3
8. h( x ) = x4 − 8x2 − 2; x = −2, 0, 2 15. Fill in the table with “+,” “−,”or “0” for the func-
3π
tion graphed below.
9. f ( x ) = sin5 ( x ); x = π
2, π, 2
1 x g( x ) g′ ( x ) g′′ ( x )
10. f ( x ) = x · ln( x ); x = e
0
11. At which values of x labeled in the figure below
1
is the point ( x, f ( x )) an inflection point?
2
3
 256 contemporary calculus

16. Sketch functions f for x-values near 1 so that In Problems 19–36, locate all critical numbers, local
f (1) = 2 and: extrema and inflection points of the given function,
(a) f ′ (1) > 0, f ′′ (1) > 0 and use these results to sketch a graph of the func-
(b) f ′ (1) > 0, f ′′ (1) < 0 tion showing all points of interest.
(c) f ′ (1) < 0, f ′′ (1) > 0 19. f ( x ) = x3 − 21x2 + 144x − 350
(d) f ′ (1) > 0, f ′′ (1) = 0, f ′′ (1− ) < 0, f ′′ (1+ ) >0
1 3 45
(e) f ′ (1) > 0, f ′′ (1) = 0, f ′′ (1− ) > 0, f ′′ (1+ ) < 0 20. g( x ) = x + x2 − x + 100
6 2
17. Some people like to think of a concave-up graph 21. f ( x ) = e7x − 5x
as one that will “hold water” and of a concave-
down graph as one which will “spill water.” That 22. g( x ) = e7x − 5x
description is accurate for a concave-down graph, 23. f ( x ) = e−3x + x
but it can fail for a concave-up graph. Sketch
24. g( x ) = e−3x − x
the graph of a function that is concave up on an
interval but will not “hold water.” 25. f ( x ) = xe−3x
1 − (x−c2)2
18. The function f ( x ) = e 2b defines the Gaus- 26. g( x ) = xe5x
2π
sian distribution used extensively in statistics 4 1
27. f ( x ) = x 3 − x 3
and probability; its graph (see below) is a “bell-
4 1
shaped” curve. 28. g( x ) = 6x 3 + 3x 3
 
29. f ( x ) = ln 1 + x2
 
30. g( x ) = ln x2 − 6x + 10
p3
31. f ( x ) = x2 + 2x + 2
p
32. g( x ) = x2 + 2x + 2
2 1
(a) Show that f has a maximum at x = c. (The 33. f ( x ) = x 3 (1 − x ) 3
1 2
value c is called the mean of this distribution.) 34. g( x ) = x 3 (1 − x ) 3
(b) Show that f has inflection points where x =
35. f (θ ) = sin(θ ) + sin2 (θ )
c + b and x = c − b. (The value b is called the
standard deviation of this distribution.) 36. g(θ ) = cos(θ ) − sin2 (θ )

3.4 Practice Answers

1. See the margin figure for reference.

x f (x) f ′ (x) f ′′ ( x ) concavity

1 + + − down
2 + − − down
3 − − + up
4 − 0 − down
 derivatives and graphs 257

2. f ′ ( x ) = 6x2 − 30x + 24, which is defined for all x. f ′ ( x ) = 0 if x = 1

or x = 4 (critical values). f ′′ ( x ) = 12x − 30 so f ′′ (1) = −18 < 0 tells
us that f is concave down at the critical value x = 1, so (1, f (1)) =
(1, 4) is a relative maximum; and f ′′ (4) = 18 > 0 tells us that f is
concave up at the critical value x = 4, so (4, f (4)) = (4, −23) is a
relative minimum. A graph of f appears in the margin.

3. The points labeled b and g are inflection points.

4. f ′ ( x ) = 4x3 − 36x2 + 60x + 5 ⇒ f ′′ ( x ) = 12x2 − 72x + 60 = 12( x2 −

6x + 5) = 12( x − 1)( x − 5) so the only candidates to be inflection
points are x = 1 and x = 5.

• If x < 1 then f ′′ ( x ) = 12(neg)(neg) > 0

• If 1 < x < 5 then f ′′ ( x ) = 12(pos)(neg) < 0
• If 5 < x then f ′′ ( x ) = 12(pos)(pos) > 0

f changes concavity at x = 1 and x = 5, so x = 1 and x = 5 are both

inflection points. A graph of f appears in the margin.

5. g( x ) = x4 + 4x3 − 90x2 + 13 ⇒ g′ ( x ) = 4x3 + 12x2 − 180x ⇒

g′′ ( x ) = 12x2 + 24x − 180; because g′ ( x ) and g′′ ( x ) are polyno-
mials, they exist everywhere. The critical numbers for g( x ) oc-
cur where g′ ( x ) = 0 ⇒ 4x3 + 12x2 − 180x = 4x ( x2 + 3x − 45) =
4x ( x + 9)( x − 5) = 0 ⇒ x = −9, x = 0 or x = 5. Using the Second
Derivative Test: g′′ (−9) = 576 > 0, so g( x ) has a local minimum at
x = −9; g′′ (0) = −180 < 0, so g( x ) has a local maximum at x = 0;
and g′′ (5) = 240 > 0, so g( x ) has a local minimum at x = 5.
Candidates for inflection points occur where g′′ ( x ) = 0:

12x2 + 24x − 180 = 12( x2 + 2x − 15) = 12( x − 3)( x + 5) = 0

⇒ x = −5 or x = 3

Observing that g′′ ( x ) > 0 for x < −5, g′′ ( x ) < 0 for −5 < x < 3
and g′′ ( x ) > 0 for x > 3 confirms that both candidates are in
fact inflection points. A graphing window with −12 ≤ x ≤ 8
(this is only one reasonable possibility) should include all points
of interest. Checking that g(−9) = −3632, g(0) = 13 and g(5) =
−1112 suggests that a graphing window with −4000 ≤ y ≤ 1000
should work (see margin).
258 contemporary calculus

3.5 Applied Maximum and Minimum Problems

We have used derivatives to find maximums and minimums of functions

given by formulas, but it is very unlikely that someone will simply
hand you a function and ask you to find its extreme value(s). Typically,
someone will describe a problem and ask your help to maximize or
minimize a quantity: “What is the largest volume of a package that the
post office will accept?”; “What is the quickest way to get from here
to there?”; or “What is the least expensive way to accomplish some
task?” These problems often involve restrictions — or constraints — and
sometimes neither the problem nor the constraints are clearly stated.
Before we can use calculus or other mathematical techniques to solve
these “max/min” problems, we need to understand the situation at hand
and translate the problem into mathematical form. After solving the
problem using calculus (or other mathematical techniques) we need to
check that our mathematical solution really solves the original problem.
Often, the most challenging part of this procedure is understanding the
problem and translating it into mathematical form.
In this section we examine some problems that require understand-
ing, translation, solution and checking. Most will not be as complicated
as those a working scientist, engineer or economist needs to solve, but
they represent a step toward developing the required skills.
Example 1. The company you own has a large supply of 8-inch by
15-inch rectangular pieces of tin, and you decide to use them to make
boxes by cutting a square from each corner and folding up the sides
(see margin). For example, if you cut a 1-inch square from each corner,
the resulting 6-inch by 13-inch by 1-inch box has a volume of 78 cubic
inches. The amount of money you can charge for a box depends on
how much the box holds, so you want to make boxes with the largest
possible volume. What size square should you cut from each corner?

Solution. To help understand the problem, first drawing a diagram

can be very helpful. Then we need to translate it into a mathematical
problem:
• identify the variables

• label the variable and constant parts of the diagram

• write the quantity to be maximized as a function of the variables

If we label the side of the square to be removed as x inches, then the
box is x inches high, 8 − 2x inches wide and 15 − 2x inches long, so the
volume is:

(length)(width)(height) = (15 − 2x )(8 − 2x ) · x

= 4x3 − 46x2 + 120x cubic inches
 derivatives and graphs 259

Now we have a mathematical problem, to maximize the function

V ( x ) = 4x3 − 46x2 + 120x, so we use existing calculus techniques,
computing V ′ ( x ) = 12x2 − 92x + 120 to find the critical points.

• Set V ′ ( x ) = 0 and solve by factoring or using the quadratic formula:

5
V ′ ( x ) = 12x2 − 92x + 120 = 4(3x − 5)( x − 6) = 0 ⇒ x = or x = 6
3
5
so x = 3 and x = 6 are critical points of V.

• V ′ ( x ) is a polynomial so it is defined everywhere and there are no

critical points resulting from an undefined derivative.

• What are the endpoints for x in this problem? A square cannot have
a negative length, so x ≥ 0. We cannot remove more than half of the
width, so 8 − 2x ≥ 0 ⇒ x ≤ 4. Together, these two inequalities say
that 0 ≤ x ≤ 4, so the endpoints are x = 0 and x = 4. (Note that
the value x = 6 is not in this interval, so x = 6 cannot maximize the
volume and we do not consider it further.)

The maximum volume must occur at the critical point x = 53 or at one

of the endpoints (x = 0 and x = 4): V (0) = 0, V ( 53 ) = 2450
27 ≈ 90.74
cubic inches, and V (4) = 0, so the maximum volume of the box occurs
when we remove a 53 -inch by 53 -inch square from each corner, resulting
in a box 53 inches high, 8 − 2( 53 ) = 14 5 35
3 inches wide and 15 − 2( 3 ) = 3
inches long. ◀

Practice 1. If you start with 7-inch by 15-inch pieces of tin, what size
square should you remove from each corner so the box will have as large
a volume as possible? [Hint: 12x2 − 88x + 105 = (2x − 3)(6x − 35)]

We were fortunate in the previous Example and Practice problem

because the functions we created to describe the volume were functions
of only one variable. In other situations, the function we get will have
more than one variable, and we will need to use additional information
to rewrite our function as a function of a single variable. Typically, the
constraints will contain the additional information we need.

Example 2. We want to fence a rectangular area in our backyard for a

garden. One side of the garden is along the edge of the yard, which is
already fenced, so we only need to build a new fence along the other
three sides of the rectangle (see margin). If a neighbor gives us 80 feet
of fencing left over from a home-improvement project, what dimensions
should the garden have in order to enclose the largest possible area
using all of the available material?

Solution. As a first step toward understanding the problem, we draw

a diagram or picture of the situation. Next, we identify the variables:
260 contemporary calculus

in this case, the length (call it x) and width (call it y) of the garden. The
margin figure shows a labeled diagram, which we can use to write a
formula for the function that we want to maximize:

A = area of the rectangle = (length)(width) = x · y

Unfortunately, our function A involves two variables, x and y, so we

need to find a relationship between them (an equation containing both
x and y) that we can solve for wither x or y. The constraint says that we
have 80 feet of fencing available, so x + 2y = 80 ⇒ y = 40 − 2x . Then:

 x x2
A = x · y = x 40 − = 40x −
2 2
which is a function of a single variable (x). We want to maximize A.
A′ ( x ) = 40 − x so the only way A′ ( x ) = 0 is to have x = 40, and
A( x ) is differentiable for all x so the only critical number (other than
the endpoints) is x = 40. Finally, 0 ≤ x ≤ 80 (why?) so we also need
to check x = 0 and x = 80: the maximum area must occur at x = 0,
x = 40 or x = 80.

02
A(0) = 40(0) − = 0 square feet
2
402
A(40) = 40(40) − = 800 square feet
2
802
A(80) = 40(80) − = 0 square feet
2
so the largest rectangular garden has an area of 800 square feet, with
dimensions x = 40 feet by y = 40 − 40
2 = 20 feet. ◀

Practice 2. Suppose you decide to create the rectangular garden in a

corner of your yard. Then two sides of the garden are bounded by the
existing fence, so you only need to use the available 80 feet of fencing
to enclose the other two sides. What are the dimensions of the new
garden of largest area? What are the dimensions if you have F feet of
new fencing available?

Example 3. You need to reach home as quickly as possible, but you are
in a rowboat on a lake 4 miles from shore and your home is 2 miles up
the shore (see margin). If you can row at 3 miles per hour and walk
at 5 miles per hour, toward which point on the shore should you row?
What if your home is 7 miles up the coast?

Solution. The margin figure shows a labeled diagram with the variable
x representing the distance along the shore from point A, the nearest
point on the shore to your boat, to point P, the point you row toward.
 derivatives and graphs 261

The total time — rowing and walking — is:

T = total time
= (rowing time from boat to P) + (walking time from P to B)
distance from boat to P distance from P to B
= +
rate rowing boat rate walking along shore
√ √
x2 + 42 2−x x2 + 16 2 − x
= + = +
3 5 3 5

It is not reasonable to row to a point below A and then walk home,

so x ≥ 0. Similarly, we can conclude that x ≤ 2, so our interval is
0 ≤ x ≤ 2 and the endpoints are x = 0 and x = 2.
To find the other critical numbers of T between x = 0 and x = 2, we
need the derivative of T:

1 1 2 − 1 1 x 1
2
T ′ (x) = · x + 16 (2x ) − = √ −
3 2 5 3 x + 16 5
2

This derivative is defined for all values of x (and in particular for all
values in the interval 0 ≤ x ≤ 2). To find where T ′ ( x ) = 0 we solve:

x 1 p
√ − = 0 ⇒ 5x = 3 x2 + 16
3 x2 + 16 5
⇒ 25x2 = 9x2 + 144
⇒ 16x2 = 144 ⇒ x2 = 9 ⇒ x = ±3

Neither of these numbers, however, is in our interval 0 ≤ x ≤ 2, so

neither of them gives a minimum time. The only critical numbers for T
on this interval are the endpoints, x = 0 and x = 2:
√
0 + 16 2 − 0 4 2
T (0) = + = + ≈ 1.73 hours
√ 3 5 3 5
√
2
2 + 16 2 − 2 20
T (2) = + = ≈ 1.49 hours
3 5 3

The quickest route has P 2 miles down the coast: you should row
directly toward home.
If your home is 7 miles down the coast, then the interval for x is
0 ≤ x ≤ 7, which has endpoints x = 0 and x = 7. Our function for the
travel time is now:
√
x2 + 16 7 − x x 1
T (x) = + ⇒ T ′ (x) = √ −
3 5 3 x + 16 5
2

so the only point in our interval where T ′ ( x ) = 0 is at x = 3 and the

derivative is defined for all values in this interval. So the only critical
262 contemporary calculus

numbers for T are x = 0, x = 3 and x = 7:

√
0 + 16 7 − 0 4 7
T (0) = + = + ≈ 2.73 hours
√ 3 5 3 5
√
2
3 + 16 7 − 3 65 4
T (3) = + = + ≈ 2.47 hours
√ 3 5 3 5
2
7 + 16 7 − 7 5
T (7) = + = ≈ 2.68 hours
3 5 3
The quickest way home is to aim for a point P that is 3 miles down the
shore, row directly to P, and then walk along the shore to home. ◀

One challenge of max/min problems is that they may require geom-

etry, trigonometry or other mathematical facts and relationships.

Example 4. Find the height and radius of the least expensive closed
cylinder that has a volume of 1,000 cubic inches. Assume that the
materials needed to construct the cylinder are free, but that it costs
80¢ per inch to weld the top and bottom onto the cylinder and to weld
the seam up the side of the cylinder (see margin).

Solution. If we let r be the radius of the cylinder and h be its height,

then the volume is V = πr2 h = 1000. The quantity we want to minimize
is cost, and

C = (top seam cost) + (bottom seam cost) + (side seam cost)

 ¢ 
= (total seam length) 80
inch
= (2πr + 2πr + h) (80) = 320πr + 80h

Unfortunately, C is a function of two variables, r and h, but we can use

the information in the constraint (V = πr2 h = 1000) to solve for h and
then substitute this expression for h into the formula for C:
 
1000 1000
1000 = πr2 h ⇒ h = ⇒ C = 320πr + 80h = 320πr + 80
πr2 πr2
which is a function of a single variable. Differentiating:
160000
C ′ (r ) = 320π −
πr3
which is defined except when r = 0 (a value that does not make sense
in the original problem) and there are no restrictions on r (other than
r > 0) so there are no endpoints to check. Thus C will be at a minimum
when C ′ (r ) = 0:
r
160000 3 500 3 500
320π − =0 ⇒ r = 2 ⇒ r=
πr3 π π2
1000 1000
so r ≈ 3.7 inches and h = =  q 2 ≈ 23.3 inches. ◀
πr2 3 500
π π2
 derivatives and graphs 263

Practice 3. Find the height and radius of the least expensive closed
cylinder that has a volume of 1,000 cubic inches, assuming that the
only cost for this cylinder is the price of the materials: the material for
the top and bottom costs 5¢ per square inch, while the material for the
sides costs 3¢ per square inch (see margin).

Example 5. Find the dimensions of the least expensive rectangular box

that is three times as long as it is wide and which holds 100 cubic
centimeters of water. The material for the bottom costs 7¢ per cm2 , the
sides cost 5¢ per cm2 and the top costs 2¢ per cm2 .

Solution. Label the box so that w = width, l = length and h = height.

Then our cost function C is:

C = (bottom cost) + (cost of front and back) + (cost of ends) + (top cost)
= (bottom area)(7) + (front and back area)(5) + (ends area)(5) + (top area)(2)
= (wl )(7) + (2lh)(5) + (2wh)(5) + (wl )(2)
= 7wl + 10lh + 10wh + 2wl
= 9wl + 10lh + 10wh
Unfortunately, C is a function of three variables (w, l and h) but we
can use the information from the constraints to eliminate some of the
variables: the box is “three times as long as it is wide” so l = 3w and

C = 9wl + 10lh + 10wh = 9w(3w) + 10(3w)h + 10wh = 27w2 + 40wh

We also know the volume V is 100 in3 and V = lwh = 3w2 h (because
100
l = 3w), so h = . Then:
3w2
 
2 2 100 4000
C = 27w + 40wh = 27w + 40w 2
= 27w2 +
3w 3w
which is a function of a single variable. Differentiating:
4000
C ′ (w) = 54w −
3w2
which is defined everywhere except w = 0 (yielding a box of volume 0)
and therer is no constraint interval, so C is minimized when C ′ (w) =
3 4000 100
0⇒w= ≈ 2.91 inches ⇒ l = 3w ≈ 8.73 inches ⇒ h = 3w 2 ≈
162
3.94 inches. The minimum cost is approximately $6.87. ◀

Problems described in words are usually more difficult to solve

because we first need to understand and “translate” a real-life problem
into a mathematical problem. Unfortunately, those skills only seem to
come with practice. With practice, however, you will start to recognize
patterns for understanding, translating and solving these problems, and
you will develop the skills you need. So read carefully, draw pictures,
think hard — and do the best you can.
264 contemporary calculus

3.5 Problems

1. (a) You have 200 feet of fencing to enclose a rect- design in part (a). Design a pen that uses only
angular vegetable garden. What should the 120 feet of fencing and has four equal-sized
dimensions of your garden be in order to en- stalls but encloses more than 400 ft2 . (Hint:
close the largest area? Don’t use rectangles and squares.)
(b) Show that if you have P feet of fencing avail-
able, the garden of greatest area is a square.
(c) What are the dimensions of the largest rectan-
gular garden you can enclose with P feet of
fencing if one edge of the garden borders a
straight river and does not need to be fenced?
(d) Just thinking — calculus will not help: What
do you think is the shape of the largest garden 4. (a) You need to form a 10-inch by 15-inch piece
that can be enclosed with P feet of fencing if of tin into a box (with no top) by cutting a
we do not require the garden to be rectangu- square from each corner and folding up the
lar? What if one edge of the garden borders a sides. How much should you cut so the re-
(straight) river? sulting box has the greatest volume?
2. (a) You have 200 feet of fencing available to con-
struct a rectangular pen with a fence divider
down the middle (see below). What dimen-
sions of the pen enclose the largest total area?

(b) If you need two dividers, what dimensions of

the pen enclose the largest area?
(c) What are the dimensions in parts (a) and (b)
(b) If the piece of tin is A inches by B inches, how
if one edge of the pen borders on a river and
much should you cut from each corner so the
does not require any fencing?
resulting box has the greatest volume?
3. You have 120 feet of fencing to construct a pen
5. Find the dimensions of a box with largest vol-
with four equal-sized stalls.
ume formed from a 10-inch by 10-inch piece of
(a) If the pen is rectangular and shaped like the
cardboard cut and folded as shown below.
one shown below, what are the dimensions of
the pen of largest area and what is that area?

(b) The square pen below uses 120 feet of fencing

but encloses a larger area (400 ft2 ) than the best
 derivatives and graphs 265

6. (a) You must construct a square-bottomed box 9. You are a lifeguard standing at the edge of the
with no top that will hold 100 cubic inches of water when you notice a swimmer in trouble (see
water. If the bottom and sides are made from figure below) 40 m out in the water from a point
the same material, what are the dimensions of 60 m down the beach. Assuming you can run
the box that uses the least material? (Assume at a speed of 8 meters per second and swim at
that no material is wasted.) a rate of 2 meters per second, how far along the
(b) Suppose the box in part (a) uses different ma- shore should you run before diving into the wa-
terials for the bottom and the sides. If the bot- ter in order to reach the swimmer as quickly as
tom material costs 5¢ per square inch and the possible?
side material costs 3¢ per square inch, what
are the dimensions of the least expensive box
that will hold 100 cubic inches of water?
(This is a “classic” problem with many variations.
We could require that the box be twice as long
as it is wide, or that the box have a top, or that
the ends cost a different amount than the front
and back, or even that it costs a certain amount
to weld each edge. You should be able to set up
the cost equations for these variations.)
7. (a) Determine the dimensions of the least expen-
sive cylindrical can that will hold 100 cubic
inches if the materials cost 2¢, 5¢ and 3¢ per
square inch, respectively, for the top, bottom 10. You have been asked to determine the least ex-
and sides. pensive route for a telephone cable that connects
(b) How do the dimensions of the least expensive Andersonville with Beantown (see figure below).
can change if the bottom material costs more
than 5¢ per square inch?
8. You have 100 feet of fencing to build a pen in the
shape of a circular sector, the “pie slice” shown
rs
below. The area of such a sector is .
2
(a) What value of r maximizes the enclosed area?
(b) What central angle maximizes the area?

(a) If it costs $5000 per mile to lay the cable on

land and $8000 per mile to lay the cable across
the river (with the cost of the cable included),
find the least expensive route.
(b) What is the least expensive route if the cable
costs $7000 per mile in addition to the cost to
lay it?
 266 contemporary calculus

11. You have been asked to determine where a wa- (a) Find the dimensions of the acceptable box with
ter works should be built along a river between a square end that has the largest volume.
Chesterville and Denton (see below) to minimize (b) Find the dimensions of the acceptable box that
the total cost of the pipe to the towns. has the largest volume if its end is a rectangle
twice as long as it is wide.
(c) Find the dimensions of the acceptable box with
a circular end that has the largest volume.
14. Just thinking — you don’t need calculus for this
problem: A spider and a fly are located on op-
posite corners of a cube (see below). What is the
shortest path along the surface of the cube from
the spider to the fly?

(a) Assume that the same size (and cost) pipe is

used to each town. (This part can be done
quickly without using calculus.)
(b) Assume instead that the pipe to Chesterville
costs $3000 per mile and to Denton it costs
$7000 per mile.
12. Light from a bulb at A is reflected off a flat mirror
to your eye at point B (see below). If the time
(and length of the path) from A to the mirror and
then to your eye is a minimum, show that the
angle of incidence equals the angle of reflection.
(Hint: This is similar to the previous problem.) 15. Two sides of a triangle are 7 and 10 inches long.
What is the length of the third side so the area
of the triangle will be greatest? (This problem
can be done without using calculus. How? If you
do use calculus, consider the angle θ between the
two sides.)
16. Find the shortest distance from the point (2, 0) to
the curve:
13. U.S. postal regulations state that the sum of the
(a) y = 3x − 1 (b) y = x2
length and girth (distance around) of a package
must be no more than 108 inches (see below). (c) x2 + y2 = 1 (d) y = sin( x )

17. Find the dimensions of the rectangle with the

largest area if the base must be on the x-axis and
its other two corners are on the graph of:
(a) y = 16 − x2 , −4 ≤ x ≤ 4
(b) x2 + y2 = 1
(c) | x | + |y| = 1
(d) y = cos( x ), − π2 ≤ x ≤ π
2
 derivatives and graphs 267

18. The strength of a wooden beam is proportional 21. You have a 6-inch-diameter paper disk that you
to the product of its width and the square of its want to form into a drinking cup by removing a
height (see figure below). What are the dimen- pie-shaped wedge (sector) and then forming the
sions of the strongest beam that can be cut from remaining paper into a cone (see below). Find
a log with diameter: the height and top radius of the cone so the that
(a) 12 inches? the volume of the cup is as large as possible.
(b) d inches?

19. You have a long piece of 12-inch-wide metal that

you plan to fold along the center line to form a 22. (a) What value of b minimizes the sum of the
V-shaped gutter (see below). What angle θ will squares of the vertical distances from y =
yield a gutter that holds the most water (that is, 2x + b to the points (1, 1), (1, 2) and (2, 2)?
has the largest cross-sectional area)?

20. You have a long piece of 8-inch-wide metal that

you plan to make into a gutter by bending up 3
inches on each side (see below). What angle θ
will yield a gutter that holds the most water?
(b) What slope m minimizes the sum of the
squares of the vertical distances from the line
y = mx to the points (1, 1), (1, 2) and (2, 2)?
(c) What slope m minimizes the sum of the
squares of the vertical distances from the line
y = mx to the points (2, 1), (4, 3), (−2, −2)
and (−4, −2)?
 268 contemporary calculus

 
23. You own a small airplane that holds a maximum 1 b
ln .” Use calculus to show that Si-
of 20 passengers. It costs you $100 per flight from b−a a
monton is correct.
St. Thomas to St. Croix for gas and wages plus
an additional $6 per passenger for the extra gas Note: Models of this type have uses for describing
required by the extra weight. The charge per pas- the behavior of groups, but it is dangerous — and
senger is $30 each if 10 people charter your plane usually invalid — to apply group descriptions or
(10 is the minimum number you will fly), and comparisons to individuals in a group. (Scientific
this charge is reduced by $1 per passenger for Genius by Dean Simonton, Cambridge University
each passenger over 10 who travels (that is, if 11 Press, 1988, pp. 69–73)
fly they each pay $29, if 12 fly they each pay $28, 27. After the table was wiped and the potato chips
etc.). What number of passengers on a flight will dried off, the question remained: “Just how far
maximize your profit? could a can of cola be tipped before it fell over?”
24. Prove: If f and g are differentiable functions and
(a) For a full can or an empty can the answer was
if the vertical distance between f and g is greatest
easy: the center of gravity (CG) of the can is
at x = c, then f ′ (c) = g′ (c) and the tangent lines
at the middle of the can, half as high as the
to f and g are parallel when x = c.
height of the can, and we can tilt the can until
the CG is directly above the bottom rim (see
below left). Find θ if the height of the can is 12
cm and the diameter is 5 cm.

25. Profit = revenue − expenses. Assuming revenue

and expenses are differentiable functions, show
that when  profit
 is maximized, then marginal
 
dR dE (b) For a partly filled can, more thinking was
revenue equals marginal expense .
dx dx needed. Some ideas you will see in Chapter
26. Dean Simonton claims the “productivity levels” 5 tell us that the CG of a can holding x cm of
of people in various fields can be described as 360 + 9.6x2
a function of their “career age” t by p(t) = cola is C ( x ) = cm above the bot-
60 + 19.2x
e−at − e−bt where a and b are constants depend- tom of the can. Find the height x of cola that
ing on the field, and career age is approximately will make the CG as low as possible.
20 less than the actual age of the individual. (c) Assuming that the cola is frozen solid (so the
(a) Based on this model, at what ages do math- top of the cola stays parallel to the bottom of
ematicians (a = 0.03, b = 0.05), geologists the can), how far can we tilt a can containing
(a = 0.02, b = 0.04) and historians (a = 0.02, x cm of cola? (See above right.)
b = 0.03) reach their maximum productivity? (d) If the can contained x cm of liquid cola, could
(b) Simonton says, “With a little calculus we can we tilt it farther or less far than the frozen cola
show that the curve (p(t)) maximizes at t = before it would fall over?
 derivatives and graphs 269

28. Just thinking — calculus will not help with this one.

(a) Four towns are located at the corners of a square. What is the
shortest length of road we can construct so that it is possible to
travel along the road from any town to any other town?
The problem of finding the shortest path
connecting several points in the plane is
called the “Steiner problem.” It is impor-
tant for designing computer chips and
telephone networks to be as efficient as
possible.

(b) What is the shortest connecting path for five towns located on the
corners of a pentagon?

Generalized Max/Min Problems

The previous max/min problems were mostly numerical problems: the
amount of fencing in Problem 2 was 200 feet, the lengths of the piece
of tin in Problem 4 were 10 and 15, and the parabola in Problem 17(a)
was y = 16 − x2 . In working those problems, you might have noticed
some patterns among the numbers in the problem and the numbers
in your answers, and you might have wondered if the pattern was a
coincidence or if there really was a general pattern at work. Rather
than trying several numerical examples to see if the “pattern” holds,
mathematicians, engineers, scientists and others sometimes resort to
generalizing the problem. We free the problem from the particular
numbers by replacing the numbers with letters, and then we solve the
generalized problem. In this way, relationships between the values
in the problem and those in the solution can become more obvious.
Solutions to these generalized problems are also useful if you want to
program a computer to quickly provide numerical answers.

29. (a) Find the dimensions of the rectangle with the 30. (a) Find the dimensions of the rectangle with the
greatest area that can be built so the base of greatest area that can be built so the base of
the rectangle is on the x-axis between 0 and 1 the rectangle is on the x-axis between 0 and 1
(0 ≤ x ≤ 1) and one corner of the rectangle is and one corner of the rectangle is on the curve
on the curve y = x2 (see above right). What is y = x3 . What is the area of this rectangle?
the area of this rectangle? (b) Generalize the problem in part (a) for the
curve y = Cx3 with C > 0 and 0 ≤ x ≤ 1.
(b) Generalize the problem in part (a) for the
(c) Generalize for the curve y = Cx3 with C > 0
parabola y = Cx2 with C > 0 and 0 ≤ x ≤ 1.
and 0 ≤ x ≤ B.
(c) Generalize for the parabola y = Cx2 with (d) Generalize for the curve y = Cx n with C > 0,
C > 0 and 0 ≤ x ≤ B. n a positive integer, and 0 ≤ x ≤ B.
 270 contemporary calculus

31. (a) The base of a right triangle is 50 and the height (b) You have T dollars to buy fencing material to
is 20. Find the dimensions and area of the enclose a rectangular plot of land. The fence
rectangle with the greatest area that can be for the top and bottom costs $A per foot and
enclosed in the triangle if the base of the rect- for the sides it costs $B per foot. Find the
angle must lie on the base of the triangle. dimensions of the plot with the largest area.
For this largest plot, how much money was
used for the top and bottom (together), and
for the sides (together)?
(c) You have T dollars to buy fencing material to
enclose a rectangular plot of land. The fence
costs $A per foot for the top, $B/foot for the
bottom, $C/ft for the left side and $D/ft for
(b) The base of a right triangle is B and the height
the right side. Find the dimensions of the plot
is H. Find the dimensions and area of the
with the largest area. For this largest plot, how
rectangle with the greatest area that can be
much money was used for the top and bottom
enclosed in the triangle if the base of the rect-
(together), and for the sides (together)?
angle must lie on the base of the triangle.
(c) State your general conclusion from part (b) in 33. Determine the dimensions of the least expensive
words. cylindrical can that will hold V cubic inches if the
top material costs $A per square inch, the bottom
32. (a) You have T dollars to buy fencing material to
material costs $B per square inch, and the side
enclose a rectangular plot of land. The fence
material costs $C per square inch.
for the top and bottom costs $5 per foot and
for the sides it costs $3 per foot. Find the di- 34. Find the location of C in the figure below so that
mensions of the plot with the largest area. For the sum of the distances from A to C and from C
this largest plot, how much money was used to B is a minimum.
for the top and bottom, and for the sides?

3.5 Practice Answers

1. V ( x ) = x (15 − 2x )(7 − 2x ) = 4x3 − 44x2 + 105x so:

V ′ ( x ) = 12x2 − 88x + 105 = (2x − 3)(6x − 35)

which is defined for all x: the only critical numbers are the endpoints
x = 0 and x = 72 and where V ′ ( x ) = 0: x = 23 and x = 35 35
6 (but 6 is
 derivatives and graphs 271

not in the interval [0, 72 ] so it is not practical). The maximum volume

must occur when x = 0, x = 32 or x = 27 :
V (0) = 0 · (15 − 2 · 0) · (7 − 2 · 0) = 0
     
3 3 3 3 3
V = · 15 − 2 · · 7−2· = (12)(4) = 72
2 2 2 2 2
     
7 7 7 7 7
V = 15 − 2 · · 7−2· = (8)(0) = 0
2 2 2 2 2
The maximum-volume box will result from cutting a 1.5-by-1.5 inch
square from each corner. A graph of V ( x ) appears in the margin.

2. (a) We have 80 feet of fencing (see margin). Our assignment is

to maximize the area of the garden: A = x · y (two variables).
Fortunately, we have the constraint that x + y = 80, so y = 80 − x
and our assignment reduces to maximizing a function of one
variable:
A = x · y = x · (80 − x ) = 80x − x2 ⇒ A′ ( x ) = 80 − 2x
so A′ ( x ) = 0 ⇒ x = 40. Because A′′ ( x ) = −2 < 0, the graph
of A is concave down, hence A has a maximum at x = 40. The
maximum area is A(40) = 40 · 40 = 1600 ft2 when x = 40 feet
and y = 40 feet. The maximum-area garden is a square.
(b) This is similar to part (a) except we have F feet of fencing instead
of 80 feet: x + y = F ⇒ y = F − x and we want to maximize
A = xy = x ( F − x ) = Fx − x2 . Differentiating, A′ ( x ) =F −
 2x so2
A′ ( x ) = 0 ⇒ x = F2 ⇒ y = F2 . The maximum area is A F2 = F4
square feet when the garden is a square with half of the new
fence used on each of the two new sides.

3. The cost C is given by:

C = 5(area of top) + 3(area of sides) + 5(area of bottom)
= 5(πr2 ) + 3(2πrh) + 5(πr2 )
so our assignment is to minimize C = 10πr2 + 6πrh, a function of
two variables (r and h). Fortunately, we also have the constraint that
volume = 1000 in3 = πr2 h ⇒ h = 1000πr2
. So:
 
1000 6000 6000
C = 10πr2 + 6πr = 10πr2 + ⇒ C ′ (r ) = 20πr − 2
πr2 r r
which exists for r ̸= 0 (r = 0 is not in the domain of C (r )).
r
′ 6000 3 3 6000
C (r ) = 0 ⇒ 20πr − 2 = 0 ⇒ 20πr = 6000 ⇒ r = ≈ 4.57 inches
r 20π
1000
When r = 4.57, h = π (4.57)2
≈ 15.24 inches. Examining the second
derivative, C ′′ (r )
= 12000
20π + r3 > 0 for all r > 0 so C is concave up
and we have found the minimum cost.
272 contemporary calculus

3.6 Asymptotic Behavior of Functions

When you turn on an automobile or a light bulb or a computer, many

things happen. Some of them are uniquely part of the start-up process
of the system. These “transient” things occur only during start up,
and then the system settles down to its steady-state operation. This
start-up behavior can be very important, but sometimes we want to
investigate the steady-state — or long-term — behavior: how does the
system behave “after a long time?” In this section we investigate and
describe the long-term behavior of functions and the systems they
model: how does a function behave “when x (or − x) is arbitrarily
large?”

Limits as x Becomes Arbitrarily Large (“Approaches Infinity”)

The same type of questions we considered about a function f as x
approached a finite number can also be asked about f as x “becomes
arbitrarily large” (or “increases without bound”) — that is, eventually
becomes larger than any fixed number.
5x
Example 1. What happens to the values of f ( x ) = and g( x ) =
2x + 3
sin(7x + 1)
as x becomes arbitrarily large (increases without bound)?
3x
5x sin(7x +1)
x 2x +3 3x
Solution. One approach is numerical: evaluate f ( x ) and g( x ) for some
10 2.17 0.031702
“large” values of x and see if there is a pattern to the values of f ( x ) and
100 2.463 −0.001374
1000 2.4962 0.000333 g( x ). The margin table shows the values of f ( x ) and g( x ) for several
10, 000 2.4996 0.000001 large values of x. When x is very large, it appears that the values of
f ( x ) are close to 2.5 = 52 and the values of g( x ) are close to 0. In fact,
we can guarantee that the values of f ( x ) are as close to 52 as someone
5x
wants by taking x to be “big enough.” The values of f ( x ) =
2x + 3
may or may not ever equal 52 (they never do), but if x is “large,” then
f ( x ) is “very close to” 52 . Similarly, we can guarantee that the values of
g( x ) are as close to 0 as someone wants by taking x to be “big enough.”
The graphs of f and g for “large” values of x appear in the margin. ◀

3x + 4
Practice 1. What happens to the values of f ( x ) = and g( x ) =
x−2
cos(5x )
as x becomes arbitrarily large?
2x + 7
We can express the answers to Example 1 using limits. “As x becomes
5x
arbitrarily large, the values of approach 52 ” can be written:
2x + 3
5x 5
lim =
x →∞ 2x + 3 2
 derivatives and graphs 273

sin(7x + 1)
and “the values of approach 0” can be written:
3x During this discussion — and throughout
sin(7x + 1) this book — we do not treat “infinity” or
lim =0 “∞” as a number, but only as a useful no-
x →∞ 3x tation. “Infinity” is not part of the real
We read lim as “the limit as x approaches infinity,” meaning “the limit number system, and we use the common
x →∞ notation “x → ∞” and the phrase “x ap-
as x becomes arbitrarily large” or “as x increases without bound.” proaches infinity” only to mean that “x
The notation “x → −∞,” read as “x approaches negative infinity,” becomes arbitrarily large.”
means that the values of − x become arbitrarily large.
Practice 2. Rewrite your answers to Practice 1 using limit notation.
The expression lim f ( x ) asks about the behavior of f ( x ) as the
x →∞
values of x get larger and larger without any bound. One way to
determine this behavior is to look at the values of f ( x ) for some values
of x that are very “large.” If the values of the function get arbitrarily A more formal definition of the limit as
“x → ∞” appears at the end of this sec-
close to a single number as x gets larger and larger, then we will say
tion.
that number is the limit of the function as x approaches infinity.
6x + 7 sin(3x )
Practice 3. Fill in the table for f ( x ) = and g( x ) = and
3 − 2x x
use those values to estimate lim f ( x ) and lim g( x ).
x →∞ x →∞

6x +7 sin(3x )
x 3−2x x

10
200
500
20, 000

1
Example 2. How large must x be to guarantee that f ( x ) = < 0.1?
x
That f ( x ) < 0.001? That f ( x ) < E (with E > 0)?
1 1 1 1
Solution. If x > 10, then < = 0.1. If x > 1000, then < =
x 10 x 1000
0.001. In general, if E is any positive number, then we can guarantee
1 1
that | f ( x )| < E by picking only values of x > > 0: if x > , then
E E
1 1
< E. From this we can conclude that lim = 0. ◀
x x →∞ x

1
Practice 4. How large must x be to guarantee that f ( x ) = < 0.1?
x2
1
That f ( x ) < 0.001? That f ( x ) < E (with E > 0)? Evaluate lim 2 .
x →∞ x
The Main Limit Theorem (Section 1.2) about limits of combinations
of functions still holds true if the limits as “x → a” are replaced with
limits as “x → ∞” but we will not prove those results.
Polynomials arise regularly in applications, and we often need the
limit, as “x → ∞,” of ratios of polynomials or functions containing
powers of x. In these situations the following technique is often helpful:
274 contemporary calculus

• factor the highest power of x in the denominator from both the

numerator and the denominator

• cancel the common factor from the numerator and denominator

The limit of the new denominator is a constant, so the limit of the

resulting ratio is easier to determine.

7x2 + 3x − 4 9x + 2
Example 3. Determine lim and lim .
x →∞ 3x2 − 5 x →∞ 3x2 − 5x + 1

Solution. Factoring x2 out of the numerator and the denominator of

the first rational function results in:

7x2 + 3x − 4 x2 (7 + 3x − x42 ) 7 + 3x − x42 7

lim = lim = lim =
x →∞ 3x2 − 5 x →∞ x2 (3 − x52 ) x →∞ 3 − x52 3

where we used the facts that 3

x → 0, 4
x2
→ 0 and 5
x2
→ 0 as x → ∞.
Similarly:

9x + 2 x2 ( 9x + x22 ) 9 2
x + x2 0
lim = lim = lim = =0
x →∞ 3x2 − 5x + 1 x → ∞ x 2 (3 − 5 + 1 ) x →∞ 3 − 5 + 1 3
x x2 x x2

because k
x → 0 and c
x2
→ 0 as x → ∞ for any constants k and c. ◀

If we need to evaluate a more difficult limit as x → ∞, it is often

useful to algebraically manipulate the function into the form of a ratio
and then use the previous technique.
If the values of the function oscillate and do not approach a single
number as x becomes arbitrarily large, then the function does not have
a limit as x approaches ∞: the limit does not exist.

Example 4. Evaluate lim sin( x ) and lim x − ⌊ x ⌋

x →∞ x →∞

Solution. As x → ∞, f ( x ) = sin( x ) and g( x ) = x − ⌊ x ⌋ do not have

limits. As x grows without bound, the values of f ( x ) = sin( x ) oscillate
between −1 and +1 (see margin), and these values do not approach a
single number. Similarly, g( x ) = x − ⌊ x ⌋ continues to take on all values
between 0 and 1, and these values never approach a single number. ◀

Using Calculators to Help Find Limits as “x → ∞” or “x → −∞”

Calculators only store a limited number of digits for each quantity.
This becomes a severe limitation when we deal with extremely large
quantities.

Example 5. The value of f ( x ) = ( x + 1) − x is clearly equal to 1 for all

values of x, and your calculator will give the right answer if you use it
to evaluate f (4) or f (5). Now use it to evaluate f for a big value of x,
 derivatives and graphs 275

say x = 1040 : f (1040 ) = (1040 + 1) − 1040 = 1, but most calculators do

not store 40 digits of a number, and they will respond that f (1040 ) = 0,
which is wrong. In this example the calculator’s error is obvious, but
similar errors can occur in less obvious ways when using calculators
for computations involving very large numbers.

You should be careful with — and somewhat suspicious of — the

answers your calculator gives you.

Calculators can still be helpful for examining limits as x → ∞ and

x → −∞ as long as we don’t place too much faith in their responses.
Even if you have forgotten some of the properties of the natural loga-
√
rithm function ln( x ) and the cube root function 3 x, a little experimen-
ln( x )
tation on your calculator can help convince you that lim √ = 0.
x →∞ 3
x

The Limit Is Infinite

1
The function f ( x ) = is undefined at x = 0, but we can still ask
x2
about the behavior of f ( x ) for values of x “close to” 0. The margin
figure indicates that if x is very small (close to 0) then f ( x ) is very large.
As the values of x get closer to 0, the values of f ( x ) grow larger and
can be made as large as we want by picking x to be close enough to 0.
Even though the values of f are not approaching any one number, we
use the “infinity” notation to indicate that the values of f are growing
1
without bound, and write: lim 2 = ∞.
x →0 x
1 1
The values of 2 do not equal “infinity”: the notation lim 2 = ∞
x x →0 x
1
means that the values of 2 can be made arbitrarily large by picking
x
values of x very close to 0.
1
The limit, as x → 0, of is slightly more complicated. If x is close to
x
1
0, then the value of f ( x ) = can be a large positive number or a large
x
1
negative number, depending on the sign of x. The function f ( x ) =
x
does not have a (two-sided) limit as x approaches 0, but we can still
investigate one-sided limits:

1 1
lim =∞ and lim = −∞
x →0+ x x →0− x
x−5 x−5
Example 6. Determine lim and lim .
x →3+ x−3 x →3 − x −3
Solution. As x → 3+ , x − 5 → −2 and x − 3 → 0. Because the
denominator is approaching 0, we cannot use the Main Limit Theorem,
276 contemporary calculus

and we need to examine the function more carefully. When x → 3+ ,

we know that x > 3 so x − 3 > 0. So if x is close to 3 and slightly larger
than 3, then the ratio of x − 5 to x − 3 is:
a number close to − 2
= large negative number
small positive number
As x > 3 gets closer to 3:
x−5 a number closer to − 2
= = larger negative number
x−3 positive and closer to 0
By taking x > 3 even closer to 3, the denominator gets closer to 0
but remains positive, so the ratio gets arbitrarily large and negative:
x−5
lim = −∞.
x →3+ x − 3
As x → 3− , x − 5 → −2 and x − 3 → 0 as before, but now we know
that x < 3 so x − 3 < 0. So if x is close to 3 and slightly smaller than 3,
then the ratio of x − 5 to x − 3 is:
a number close to − 2
= large positive number
small negative number
x−5
so lim = ∞. ◀
x →3− x−3
7 3x 3x2 − 6x
Practice 5. Find: (a) lim (b) lim (c) lim .
x →2+ 2−x x →2+ 2x − 4 x →2+ x−2

Horizontal Asymptotes
The limits of f , as “x → ∞” and “x → −∞,” provide information about
horizontal asymptotes of f .

Definition: The line y = K is a horizontal asymptote of f if:

lim f ( x ) = K or lim f (x) = K

x →∞ x →−∞

2x + sin( x )
Example 7. Find any horizontal asymptotes of f ( x ) = .
x
Solution. Computing the limit as x → ∞:
   
2x + sin( x ) 2x sin( x ) sin( x )
lim = lim + = lim 2 +
x →∞ x x →∞ x x x →∞ x
sin( x )
= 2 + lim = 2+0 = 2
x →∞ x
so the line y = 2 is a horizontal asymptote of f . The limit, as “x → −∞,"
is also 2 so y = 2 is the only horizontal asymptote of f . The graphs of f
and y = 2 appear in the margin. A function may or may not cross its
asymptote. ◀
 derivatives and graphs 277

You likely explored horizontal asymptotes in a previous course using

terms like “end behavior” and investigating only rational functions.
The tools of calculus allow us to make the the notion of “end behavior”
more precise and investigate a wider variety of functions.

Vertical Asymptotes
As with horizontal asymptotes, you have likely studied vertical asymp-
totes before (at least for rational functions). We can now define vertical
asymptotes using infinite limits.

Definition: The vertical line x = a is a vertical asymptote of the

graph of f if either or both of the one-sided limits of f , as x → a−
or x → a+ , is infinite.

If our function f is the ratio of a polynomial P( x ) and a polynomial

P( x )
Q ( x ), f ( x ) = , then the only candidates for vertical asymptotes
Q( x )
are the values of x where Q( x ) = 0. However, the fact that Q( a) = 0 is
not enough to guarantee that the line x = a is a vertical asymptote of f ;
we also need to evaluate P( a).
If Q( a) = 0 and P( a) ̸= 0, then the line x = a must be a vertical
asymptote of f . If Q( a) = 0 and P( a) = 0, then the line x = a may or
may not be a vertical asymptote.

x2 − x − 6
Example 8. Find the vertical asymptotes of f ( x ) = and
x2 − x
x2 − 3x
g( x ) = .
x2 − x

Solution. Factoring the numerator and denominator of f ( x ) yields

( x − 3)( x + 2)
f (x) = so the only values of x that make the denomi-
x ( x − 1)
nator 0 are x = 0 and x = 1, and these are the only candidates to be
vertical asymptotes. Because lim f ( x ) = +∞ and lim f ( x ) = −∞,
x →0+ x →1+
both x = 0 and x = 1 are vertical asymptotes of f .
x ( x − 3)
Factoring the numerator and denominator of g( x ) yields
x ( x − 1)
so the only candidate to be vertical asymptotes are x = 0 and x = 1.
x ( x − 3) x−3
Because lim g( x ) = lim = lim = −∞ the line
x →1+ x →1+ x ( x − 1 ) x →1+ x − 1
x−3
x = 1 must be a vertical asymptote of g. But lim g( x ) = lim =
x →0 x →0 x − 1
3 ̸= ±∞ so x = 0 is not a vertical asymptote of g. ◀

x2 + x
Practice 6. Find the vertical asymptotes of f ( x ) = and
x2 +x−2
x2 − 1
g( x ) = .
x−1
278 contemporary calculus

Other Asymptotes as “x → ∞” and “x → −∞”

If the limit of f ( x ), as x → ∞ or x → −∞, is a constant K, then
the graph of f gets arbitrarily close to the horizontal line y = K, in
which case we call y = K a horizontal asymptote of f . Some functions,
however, approach lines that are not horizontal.
x2 + 2x + 1 1
Example 9. Find all asymptotes of f ( x ) = = x+2+ .
x x
1
Solution. If x is a large positive (or negative) number, then is very
x
close to 0, and the graph of f ( x ) is very close to the line y = x + 2 (see
margin). The line y = x + 2 is an asymptote of the graph of f .
1
If x is a large positive number, then is positive, and the graph of f
x
is slightly above the graph of y = x + 2. If x is a large negative number,
1
then is negative, and the graph of f will be slightly below the graph
x
1
of y = x + 2. The piece of f never equals 0, so the graph of f never
x
crosses or touches the graph of the asymptote y = x + 2.
The graph of f also has a vertical asymptote at x = 0 because
lim f ( x ) = ∞ and lim f ( x ) = −∞. ◀
x →0+ x →0−

2x2 − x − 1 2
Practice 7. Find all asymptotes of g( x ) = = 2x − 3 + .
x+1 x+1
Some functions even have nonlinear asymptotes: asymptotes that
are not straight lines. The graphs of these functions approach some
nonlinear function when the values of x become arbitrarily large.
x4 + 3x3 + x2 + 4x + 5
Example 10. Find all asymptotes of f ( x ) = =
x2 + 1
x+5
x2 + 3x + .
x2 + 1
x+5
Solution. When x is very large, positive or negative, then is
x2 + 1
very close to 0 and the graph of f is very close to the graph of g( x ) =
x2 + 3x. The function g( x ) = x2 + 3x is a nonlinear asymptote of f .
The denominator of f is never 0 and f has no vertical asymptotes. ◀
x3 + 2 sin( x ) 2 sin( x )
Practice 8. Find all asymptotes of f ( x ) = = x2 + .
x x
If we can write f ( x ) as a sum of two functions, f ( x ) = g( x ) + r ( x ),
with lim r ( x ) = 0, then the graph of f is asymptotic to the graph of
x →±∞
g, and g is an asymptote of f . In this situation:
• if g( x ) = K, then f has a horizontal asymptote y = K

• if g( x ) = ax + b, then f has a linear asymptote y = ax + b

• otherwise f has a nonlinear asymptote y = g( x )

 derivatives and graphs 279

Formal Definition of lim f ( x ) = K

x →∞

The following definition states precisely what we mean by the phrase

“we can guarantee that the values of f ( x ) are arbitrarily close to K by
restricting the values of x to be sufficiently large.”

Definition: lim f ( x ) = K means that, for every given ϵ > 0, there

x →∞
is a number N so that:
if x is larger than N
then f ( x ) is within ϵ units of K.

Equivalently: | f ( x ) − K | < ϵ whenever x > N.

x 1
Example 11. Show that lim = .
x →∞ 2x + 1 2
Solution. Typically, we need to do two things. First we need to find a
value of N, often depending on ϵ. Then we need to show that the value
of N we found satisfies the conditions of the definition.
Assume that | f ( x ) − K | is less than ϵ and solve for x > 0:
     
 x   2x − (2x + 1)   −1  1
ϵ>   =  =  =
2x + 1  2(2x + 1)  4x + 2  4x + 2
 
1 1 1
⇒ 4x + 2 > ⇒ x > −2
ϵ 4 ϵ
 
1 1
So, given any ϵ > 0, take N = −2 .
4 ϵ
Now we can just reverse the order of the steps above  to show
 that
1 1
this N satisfies the limit definition. If x > 0 and x > − 2 then:
4 ϵ
 
1 1  x 1 
4x + 2 > ⇒ ϵ > = − = | f (x) − K|
ϵ 4x + 2 2x + 1 2 
We have shown that “for every given ϵ, there is an N” that satisfies the
definition. ◀

3.6 Problems
1. The margin figure shows f ( x ) and g( x ) for 0 ≤ x ≤ 5. Define a new
f (x)
function h( x ) = .
g( x )
(a) At what value of x does h( x ) have a root?
(b) Determine the limits of h( x ) as x → 1+ , x → 1− , x → 3+ and
x → 3− .
(c) Where does h( x ) have a vertical asymptote?
 280 contemporary calculus

3x2 + 5x
 
2. The figure below shows f ( x ) and g( x ) on the
22. ln
f (x) x2 − 4
interval 0 ≤ x ≤ 5. Let h( x ) = .
g( x ) 23. ln( x + 8) − ln( x − 5)
(a) At what value(s) of x does h( x ) have a root?
24. ln(3x + 8) − ln(2x + 5)
(b) Where does h( x ) have vertical asymptotes?
25. Salt water with a concentration of 0.2 pounds of
salt per gallon flows into a large tank that initially
contains 50 gallons of pure water.
(a) If the flow rate of salt water into the tank is 4
gallons per minute, what is the volume V (t) of
water and the amount A(t) of salt in the tank
t minutes after the flow begins?
3. The figure below shows f ( x ) and g( x ) for 0 ≤ (b) Show that the salt concentration C (t) at time t
f (x) 0.8t
x ≤ 5. Let h( x ) = . Determine the limits of is C (t) = .
g( x ) 4t + 50
h( x ) as x → 2+ , x → 2− , x → 4+ and x → 4− . (c) What happens to the concentration C (t) after
a “long” time?
(d) Redo parts (a)–(c) for a large tank that initially
contains 200 gallons of pure water.
26. Under certain laboratory conditions, an agar plate
contains B(t) = 100 2 − e−t bacteria t hours af-

ter the start of the experiment.

For Problems 4–24, calculate the limit of each ex-
(a) How many bacteria are on the plate at the start
pression as “x → ∞.”
of the experiment (t = 0)?
6 28 (b) Show that the population is always increasing.
4. 5.
x+2 3x − 5 (Show B′ (t) > 0 for all t > 0.)
7x + 12 4 − 3x (c) What happens to the population B(t) after a
6. 7.
3x − 2 x+8 “long” time?
5 sin(2x ) cos(3x ) (d) Redo parts (a)–(c) for B(t) = A(2 − e−t ).
8. 9.
2x 5x − 1
For Problems 27–41 , calculate the limits.
2x − 3 sin( x ) 4 + x · sin( x )
10. 11.
5x − 1 2x − 4 x+5 x−1
27. lim 28. lim
x2
− 5x + 2 2x2 −9 x →0 x2 x →3 ( x − 3)2
12. 13.
x2 + 8x − 4 3x2 + 10x
√ x−7 x−1
x+5 5x2 − 7x + 2 29. lim 30. lim
14. √ 15. x →5 ( x − 5)2 x →2+ x−2
4x − 2 2x3 + 4x
x−1 x−1
x + sin( x ) 7x2 + x · sin( x ) 31. lim 32. lim
16. 17. x →2− x−2 x →3+ x−2
x − sin( x ) 3 − x2 + sin(7x2 )
√ x+3 x2 + 5
7x143 + 734x − 2 9x2 + 16 33. lim 34. lim
18. 150 19. √ x →4+ 4−x x →1− 1−x
x − 99x83 + 25 2 + x2 + 1

3x + 5
 
7x + 4
 x2 − 4 x2 − x − 2
20. sin 21. cos 35. lim 36. lim
2x − 1 x2 + x + 1 x →3+ x2 − 2x − 3 x →2 x2 − 4
 derivatives and graphs 281

x−2 x3 + 7x − 4 cos( x ) 3−x

37. lim 38. lim 48. f ( x ) = 49. f ( x ) = 2 +
x →0 1 − cos( x ) x →∞ x2 + 11x x2 x−1

sin( x − 5) x+1 x · sin( x ) 2x2 + x + 5

39. lim 40. lim 50. f ( x ) = 51. f ( x ) =
x →5 ( x − 5) x →0 sin2 ( x ) x2 − x x
1 + cos( x )
41. lim x2 + x 53. f ( x ) =
1
+ sin( x )
x →0+ 1 − ex 52. f ( x ) =
x+1 x−2
In Problems 42–59, write an equation of each asymp-
x x
tote for each function and state whether it is a verti- 54. f ( x ) = x + 55. f ( x ) = x2 +
x2 + 1 x2 + 1
cal, horizontal or slant asymptote.
x x · cos( x )
x+2 x−3 56. f ( x ) = x2 + 57. f ( x ) =
42. f ( x ) = 43. f ( x ) = x+1 x−3
x−1 x2
x−1 x+5 x3 − x2 + 2x − 1
44. f ( x ) = 45. f ( x ) = 58. f ( x ) =
x2 − x x2 − 4x + 3 x−1
s
x + sin( x ) x2 − 4 x2 + 3x + 2
46. f ( x ) = 47. f ( x ) = 59. f ( x ) =
3x − 3 x+1 x+3

3.6 Practice Answers

1. As x becomes arbitrarily large, the values of f ( x ) approach 3 and
the values of g( x ) approach 0.
3x + 4 cos(5x )
2. lim = 3 and lim =0
x →∞ x−2 x →∞ 2x + 7

3. The completed table appears in the margin. x 6x +7 sin(3x )

3−2x x
√ 1 10 −3.94117647 −0.09880311
4. If x > 10 ≈ 3.162 then f ( x ) =
< 0.1. −3.04030227
x2 200 0.00220912
√ 1 500 −3.00160048 0.00017869
If x > 1000 ≈ 31.62 then f ( x ) = 2 < 0.001. 20, 000 −3.00040003 0.00004787
x ↓ ↓
r
1 1 1 −3 0
If x > = √ then f ( x ) = 2 < E.
E E x
5. (a) As x → 2+ , 2 − x → 0, and x > 2 so 2 − x < 0: 2 − x takes on
small negative values.
7 7
= = large negative number
2−x small negative number
7
so we represent the limit as: lim = −∞.
2−xx →2+
(b) As x → 2+ , 2x − 4 → 0, and x > 2 so 2x − 4 > 0: 2x − 4 takes on
small positive values. And as x → 2+ , 3x → 6 so:
3x number near 6
= = large positive number
2x − 4 small positive number
3x
so we represent the limit as: lim = +∞.
x →2+ 2x − 4
282 contemporary calculus

(c) As x → 2+ , 3x2 − 6x → 0 and x − 2 → 0 so we need to do more

work. Factoring the numerator as 3x2 − 6x = 3x ( x − 2):

3x2 − 6x 3x ( x − 2)
lim = lim = lim 3x = 6
x →2+ x−2 x →2 + x−2 x →2+

where we were able to cancel the x − 2 factor because the limit

involves values of x close to (but not equal to) 2.

x2 + x x ( x + 1)
6. (a) f ( x ) =2
= so f has vertical asymptotes
x +x−2 ( x − 1)( x + 2)
at x = 1 and x = −2.
x2 − 1 ( x + 1)( x − 1)
(b) g( x ) = = so the value x = 1 is not in the
x−1 x−1
domain of g. If x ̸= 1, then g( x ) = x + 1: g has a “hole” when
x = 1 and no vertical asymptotes.
2
7. g( x ) = 2x − 3 + has a vertical asymptote at x = −1 and
x+1
2
no horizontal asymptotes, but lim = 0 so g has the linear
x →∞ x + 1
asymptote y = 2x − 3.
2 sin( x )
8. f ( x ) = x2 + is not defined at x = 0, so f has a vertical
x
2 sin( x )
asymptote or a “hole” there; lim x2 + = 0 + 2 = 2 so f has
x →0 x
2 sin( x )
a “hole” when x = 0. Because lim = 0, f has the nonlinear
x →∞ x
2
asymptote y = x (but no horizontal asymptotes).
 derivatives and graphs 283

3.7 L’Hôpital’s Rule

f ( x + h) − f ( x )
When taking limits of slopes of secant lines, msec =
h
as h → 0, we frequently encountered one difficulty: both the numerator
and the denominator approached 0. And because the denominator
approached 0, we could not apply the Main Limit Theorem. In many
situations, however, we managed to get past this “ 00 ” difficulty by using
algebra or geometry or trigonometry to rewrite the expression and then
take the limit. But there was no common approach or pattern. The alge-
(2 + h )2 − 4
braic steps we used to evaluate lim seem quite different
h →0 h
sin(2 + h) − sin(2)
from the trigonometric steps needed for lim . Although discovered by Johann Bernoulli,
h →0 h this rule was named for the Marquis
In this section we consider a single technique, called l’Hôpital’s Rule, de l’Hôpital (pronounced low-pee-TALL),
that enables us to quickly and easily evaluate many limits of the form who published it in his 1696 calculus text-
book, Analysis of the Infinitely Small for the
“ 00 ” as well as several other challenging indeterminate forms.
Understanding of Curved Lines.

A Linear Example
The graphs of two linear functions appear in the margin and we want
f (x)
to find lim . Unfortunately, lim f ( x ) = 0 and lim g( x ) = 0 so we
x →5 g ( x ) x →5 x →5
cannot apply the Main Limit Theorem. We do know, however, that f
and g are linear, so we can calculate their slopes, and we know that they
both lines go through the point (5, 0) so we can find their equations:
f ( x ) = −2( x − 5) and g( x ) = 3( x − 5).
Now the limit is easier to compute:
f (x) −2( x − 5) −2 2 slope of f
lim = lim = lim =− =
x →5 g( x ) x →5 3 ( x − 5 ) x →5 3 3 slope of g
In fact, this pattern works for any two linear functions: If f and g are
linear functions with slopes m ̸= 0 and n ̸= 0 and a common root
at x = a, then f ( x ) − f ( a) = m( x − a) and g( x ) − g( a) = n( x − a) so
f ( x ) = m( x − a) and g( x ) = n( x − a). Then:
f (x) m( x − a) m m slope of f
lim = lim = lim = =
x→a g( x ) x→a n( x − a) x→a n n slope of g
A more powerful result—that the same pattern holds true for differen-
tiable functions even if they are not linear—is called l’Hôpital’s Rule.

L’Hôpital’s Rule (“ 00 ” Form)

If f and g are differentiable at x = a,

f ( a) = 0, g( a) = 0 and g′ ( a) ̸= 0
f (x) f ′ ( a) slope of f at a
then lim = ′ =
x→a g( x ) g ( a) slope of g at a
284 contemporary calculus

Idea for a proof: Even though f and g may not be linear functions,
they are differentiable. So at the point x = a they are “almost linear” in
the sense that we can approximate them quite well using their tangent
lines at that point (see margin).
Because f ( a) = g( a) = 0, f ( x ) ≈ f ( a) + f ′ ( a)( x − a) = f ′ ( a)( x − a)
and g( x ) ≈ g( a) + g′ ( a)( x − a) = g′ ( a)( x − a). So:

f (x) f ′ ( a)( x − a) f ′ ( a) f ′ ( a)
lim ≈ lim ′ = lim ′ = ′
x→a g( x ) x → a g ( a )( x − a ) x→a g ( a) g ( a)

Unfortunately, we have ignored some

subtle difficulties, such as g( x ) or g′ ( x ) x2 + sin(5x ) ln( x )
Example 1. Determine lim and lim x .
possibly being 0 when x is close to, but x →0 3x x →1 e − e
not equal to, a. Because of these issues,
a full-fledged proof of l’Hôpital’s Rule is Solution. We could evaluate the first limit without l’Hôpital’s Rule,
omitted. but let’s use it anyway. We can match the pattern of l’Hôpital’s Rule
by letting a = 0, f ( x ) = x2 + sin(5x ) and g( x ) = 3x. Then f (0) = 0,
g(0) = 0, and f and g are differentiable with f ′ ( x ) = 2x + 5 cos(5x )
and g′ ( x ) = 3, so:

x2 + sin(5x ) f ′ (0) 2 · 0 + 5 cos(5 · 0) 5

lim = ′ = =
x →0 3x g (0) 3 3

For the second limit, let a = 1, f ( x ) = ln( x ) and g( x ) = e x − e. Then

f (1) = 0, g(1) = 0, f and g are differentiable for x near 1 (when x > 0),
1
and f ′ ( x ) = and g′ ( x ) = e x . Then:
x
1
ln( x ) f ′ (1) 1 1
lim x
= ′
= 1
=
x →1 e −e g (1) e e

Here no simplification was possible, so we needed l’Hôpital’s Rule. ◀

1 − cos(5x ) x2 + x − 6
Practice 1. Evaluate lim and lim 2 .
x →0 3x x →2 x + 2x − 8

Strong Version of l’Hôpital’s Rule

We can strengthen L’Hôpital’s Rule to include cases when g′ ( a) = 0,
and the indeterminate form “ ∞
∞ ” when f and g increase without bound.

L’Hôpital’s Rule (Strong “ 00 ” and “ ∞

∞ ” Forms)

If f and g are differentiable on an open interval I con-

taining a, g′ ( x ) ̸= 0 on I except possibly at a, and
f (x) 0 ∞
lim = “ ” or “ ”
x→a g( x ) 0 ∞
f (x) f ′ (x)
then lim = lim ′ if the limit on the right exists.
x→a g( x ) x→a g ( x )

(Here “a” can represent a finite number or “∞.”)

 derivatives and graphs 285

e7x
Example 2. Evaluate lim .
x →∞ 5x

Solution. As “x → ∞,” both e7x and 5x increase without bound, so

we have an “ ∞
∞ ” indeterminate form and can use the Strong Version of
e7x 7e7x
l’Hôpital’s Rule: lim = lim = ∞. ◀
x →∞ 5x x →∞ 5

f′
The limit of may also be an indeterminate form, in which case
g′
f′
we can apply l’Hôpital’s Rule again to the ratio ′ . We can continue
g
using l’Hôpital’s Rule at each stage as long as we have an indeterminate
quotient.

x3
Example 3. Compute lim .
x →0 x − sin( x )

Solution. As x → 0, f ( x ) = x3 → 0 and g( x ) = x − sin( x ) → 0 so:

x3 3x2 6x 6
lim = lim = lim = lim =6
x →0 x − sin( x ) x →0 1 − cos( x ) x →0 sin( x ) x →0 cos( x )

where we have used l’Hôpital’s Rule three times in succession. (At each
stage, you should verify the conditions for l’Hôpital’s Rule hold.) ◀

x2 + e x
Practice 2. Use l’Hôpital’s Rule to find lim .
x →∞ x3 + 8x

Which Function Grows Faster?

Sometimes we want to compare the asymptotic behavior of two systems
or functions for large values of x. L’Hôpital’s Rule can be useful in such
situations. For example, if we have two algorithms for sorting names,
and each algorithm takes longer and longer to sort larger collections
of names, we may want to know which algorithm will accomplish the
task more efficiently for really large collections of names.

Example 4. Algorithm A requires n · ln(n) steps to sort n names and

algorithm B requires n1.5 steps. Which algorithm will be better for
sorting very large collections of names?

Solution. We can compare the ratio of the number of steps each algo-
n · ln(n)
rithm requires, , and then take the limit of this ratio as n grows
n1.5
n · ln(n)
arbitrarily large: lim .
n→∞ n1.5
If this limit is infinite, we say that n · ln(n) “grows faster” than n1.5 .
If the limit is 0, we say that n1.5 grows faster than n · ln(n).
286 contemporary calculus

Because n · ln(n) and n1.5 both grow arbitrarily large when n becomes
ln(n)
large, we can simplify the ratio to 0.5 and then use l’Hôpital’s Rule:
n
1
ln(n) n 2
lim = lim = lim √ = 0
n→∞ n0.5 n→∞ 0.5n−0.5 n→∞ n

We conclude that n1.5 grows faster than n · ln(n) so algorithm A requires

fewer steps for really large sorts. ◀

Practice 3. Algorithm A requires en operations to find the shortest path

connecting n towns, while algorithm B requires 100 · ln(n) operations
for the same task and algorithm C requires n5 operations. Which
algorithm is best for finding the shortest path connecting a very large
number of towns? The worst?

Other Indeterminate Forms

We call “ 00 ” an indeterminate form because knowing that f approaches
f
0 and g approaches 0 is not enough to determine the limit of g , even
if that limit exists. The ratio of a “small” number divided by a “small”
number can be almost anything as three simple “ 00 ” examples show:

3x x2 5x
lim =3 while lim =0 and lim =∞
x →0 x x →0 x x →0 x3
Similarly, “ ∞
∞ ” is an indeterminate form because knowing that f and
g both grow arbitrarily large is not enough to determine the value of
f
the limit of g or even if the limit exists:

3x x2 5x
lim =3 while lim =∞ and lim =0
x →∞ x x →∞ x x →∞ x3

In addition to the indeterminate quotient forms “ 00 ” and “ ∞

∞ ” there
are several other “indeterminate forms.” In each case, the resulting
limit depends not only on each function’s limit but also on how quickly
each function approaches its limit.

• Product: If f approaches 0 and g grows arbitrarily large, the product

f · g has the indeterminate form “0 · ∞.”

• Exponent: If f and g both approach 0, the function f g has the

indeterminate form “00 .”

• Exponent: If f approaches 1 and g grows arbitrarily large, the

function f g has the indeterminate form “1∞ .”

• Exponent: If f grows arbitrarily large and g approaches 0, the

function f g has the indeterminate form “∞0 .”
 derivatives and graphs 287

• Difference: If f and g both grow arbitrarily large, the function f − g

has the indeterminate form “∞ − ∞.”

Unfortunately, l’Hôpital’s Rule can only be used directly with an

indeterminate quotient ( 00 or “ ∞
∞ ”), but we can algebraically manipulate
these other forms into quotients and then apply l’Hôpital’s Rule.

Example 5. Evaluate lim x · ln( x ).

x →0+

Solution. This limit involves an indeterminate product (of the form

“0 · −∞”) but we need a quotient in order to apply l’Hôpital’s Rule. If
we rewrite the product x · ln( x ) as a quotient:
1
ln( x ) x
lim x · ln( x ) = lim = lim = lim − x = 0
x →0+ x →0+ 1 x →0+ −21 x →0+
x x

results from applying the “ ∞

∞ ” version of l’Hôpital’s Rule. ◀

To use l’Hôpital’s Rule on a product f · g with indeterminate form

f g
“0 · ∞,” first rewrite f · g as a quotient: 1 or 1 . Then apply
g f
l’Hôpital’s Rule.

Example 6. Evaluate lim x x .

x →0+

Solution. This limit involves the indeterminate form 00 . We can con-

vert it to a product by recalling a property of exponential and logarith-
mic functions: for any positive number a, a = eln(a) so:
g)
f g = eln( f = e g·ln( f )

Applying this to x x :
x)
lim x x = lim eln( x = lim e x·ln(x)
x →0+ x →0+ x →0+

This last limit involves the indeterminate product x · ln( x ). From the
previous example we know that lim x · ln( x ) = 0 so we can conclude
x →0+
that:
lim x x = lim e x·ln( x) = elimx→0+ x ·ln( x )
= e0 = 1
x →0+ x →0+
because the function f (u) = eu is continuous everywhere. ◀

To use l’Hôpital’s Rule on an expression involving exponents, f g

with the indeterminate form “00 ,” “1∞ ” or “∞0 ,” first convert it to
an expression involving an indeterminate product by recognizing
that f g = e g·ln( f ) and then determining the limit of g · ln( f ). The
final result is elimit of g·ln( f ) .
 288 contemporary calculus

 a x
Example 7. Evaluate lim 1+ .
x →∞ x
Solution. This expression has the form 1∞ so we first use logarithms
to convert the problem into a limit involving a product:
a x
 a
lim = lim e x·ln(1+ x )
1+
x →∞ x x →∞
 a
so now we need to compute lim x · ln 1 + . This limit has the form
x →∞ x
“∞ · 0” so we now convert the product to a quotient:

ln 1 + xa


 a
lim x · ln 1 + = lim 1
x →∞ x x →∞
x

This last limit has the form “ 00 ” so we can finally apply l’Hôpital’s Rule:
−a
a

x2
ln 1 + x 1+ xa a a
lim 1
= lim −1
= lim a = =a
x →∞ x →∞ x →∞ 1+ 1
x x2 x

and conclude that:

 a x a x ·ln(1+ xa )
lim 1 + = lim e x·ln(1+ x ) = elimx→∞ = ea
x →∞ x x →∞

where we have again used the continuity of the function f (u) = eu . ◀

3.7 Problems

In Problems 1–15, evaluate each limit. Be sure to 1 − cos( x )

15. lim
justify any use of l’Hôpital’s Rule. x →0 x · cos( x )
3x
x3 − 1 x4 − 16 16. Find a value for p so that lim = 2.
x →∞
px + 7
1. lim 2 2. lim 5
x →1 x − 1 x →2 x − 32 e px − 1
17. Find a value for p so that lim = 5.
ln(1 + 3x ) ex x →0 3x
3. lim 4. lim √
x →0 5x x →∞ x3 3x + 5
18. The limit lim √ has the indeterminate
x →∞ 2x − 1
x · ex x
2 −1 form “ ∞
5. lim 6. lim ∞ .” Why doesn’t l’Hôpital’s Rule work
x →0 1 − ex x →0 x with this limit? (Hint: Apply l’Hôpital’s Rule
ln( x ) ln( x ) twice and see what happens.) Evaluate the limit
7. lim 8. lim √
x →∞ x x →∞ x without using l’Hôpital’s Rule.
ex ex ex
ln( x ) e3x − e2x 19. (a) Evaluate lim , lim 2 and lim 5 .
x →∞ x x →∞ x x →∞ x
9. lim (p > 0) 10. lim
x →∞ xp x →0 4x (b) An algorithm is “exponential” if it requires
1 − cos(3x ) 1 − cos(2x ) a · ebn steps (a, and b are positive constants).
11. lim 12. lim An algorithm is “polynomial” if it requires
x →0 x2 x →0 x
c · nd steps. Show that polynomial algorithms
x m − am cos( a + x ) − cos( a)
13. lim 14. lim require fewer steps than exponential ones for
x→a x n − an x →0 x
large values of n.
 derivatives and graphs 289

x2 In Problems 21–30, evaluate each limit. Be sure to

20. The problem lim appeared on a test.
x →0 3x2
+x justify any use of l’Hôpital’s Rule.
One student determined the limit was an indeter-
minate “ 00 ” form and applied l’Hôpital’s Rule to 21. lim sin( x ) · ln( x ) 22. lim x3 e− x
get: x →0+ x →∞

√
x2 2x 2 1 23. lim x · ln( x ) 24. lim xsin( x)
x →0+ x →0+
lim = lim = lim =
x →0 3x2 + x x →0 6x + 1 x →0 6 3 x
26. lim (1 − cos(3x )) x

3
Another student also determined the limit was 25. lim 1− x →0
x →∞ x2
an indeterminate “ 00 ” form and wrote:  
1 1 28. lim [ x − ln( x )]
27. lim − x →∞
x2 2x 0 x →0 x sin( x )
lim 2
= lim = =0
x →0 3x + x x →0 6x + 1 0+1  1  2
x+5 x 3 x

Which student is correct? Why? 29. lim 30. lim 1+

x →∞ x x →∞ x

3.7 Practice Answers

1. Both numerator and denominator in the first limit are differentiable
and both equal 0 when x = 0, so we apply l’Hôpital’s Rule:

1 − cos(5x ) 5 sin(5x ) 0
lim = lim = =0
x →0 3x x →0 3 3

Both numerator and denominator in the second limit are differen-

tiable and both equal 0 when x = 0, so we apply l’Hôpital’s Rule:

x2 + x − 6 2x + 1 5
lim = lim =
x →2 x2 + 2x − 8 x →2 2x + 2 6

2. Both numerator and denominator are differentiable and both become

arbitrarily large as x → ∞, so we apply l’Hôpital’s Rule:

x2 + e x 2x + e x 2 + ex ex Note that we needed to apply l’Hôpital’s

lim = lim = lim = lim =∞
x →∞ x3 + 8x x →∞ 3x2 + 8 x →∞ 6x x →∞ 6 Rule three times and that each stage in-
volved an “ ∞
∞ ” indeterminate form.

3. Comparing A with en operations to B with 100 · ln(n) operations we

can apply l’Hôpital’s Rule:

en en n · en
lim = lim 1 = lim =∞
n→∞ 100 ln(n) n→∞ n→∞ 100
n

to show that B requires fewer operations than A.

Comparing B with 100 ln(n) operations to C with n5 operations, we
again apply l’Hôpital’s Rule:
100
100 ln(n) 20
lim 5
= lim n 4 = lim 5 = 0
n→∞ n n→∞ 5n n→∞ n
290 contemporary calculus

to show that B requires fewer operations than C. So B requires the

fewest operations of the three algorithms.
Comparing A to C we must apply l’Hôpital’s Rule repeatedly:

en en en en
lim = lim = lim = lim
n→∞ n5 n→∞ 5n4 n→∞ 20n3 n→∞ 60n2
e n e n
= lim = lim =∞
n→∞ 120n n→∞ 120

So A requires more operations than C and thus A requires the most

operations of the three algorithms.
A
Answers

Important Note about Precision of Answers: In many of the problems in this book you are required to read
information from a graph and to calculate with that information. You should take reasonable care to read the
graphs as accurately as you can (a small straightedge is helpful), but even skilled and careful people make
slightly different readings of the same graph. That is simply one of the drawbacks of graphical information.
When answers are given to graphical problems, the answers should be viewed as the best approximations we
could make, and they usually include the word “approximately” or the symbol “≈” meaning “approximately
equal to.” Your answers should be close to the given answers, but you should not be concerned if they differ
a little. (Yes those are vague terms, but it is all we can say when dealing with graphical information.)

Section 0.1 Section 0.2

1. approx. 1, 0, −1 1. (a) − 43 (b) 1
2 (c) 0 (d) 2 (e) undefined
70◦ − 150◦ ◦ 4
3. (a) (b) − 95(c) x + 2 (if x ̸= 2) (d) 4 + h (if h ̸= 0)
3. (a) ≈ = −8 3
10 min − 0 min min (e) a + x (if a ̸= x)
(b) ≈ 5 ◦ cooling; 3 ◦ cooling
min min 5. (a) t = 5: 5000 10 5000 5
(c) ≈ 5.5 ◦ cooling; 8.5 ◦ cooling 1500 = 3 ; t = 10: 3000 = 3 ; t = 20:
min min 5000 5 5000 50
6000 = 6 (b) any t > 0: 300t = 3t (c) Decreas-
(d) When t = 6 min.
ing, since the numerator remains constant at 5000
5. We estimate that the area is approximately (very while the denominator increases.
approximately) 9 cm2 .
7. The restaurant is 4 blocks south and 2 blocks east.
√ √
7. Method 1: Measure the diameter of the can, fill it The distance is 42 + 22 = 20 ≈ 4.47 blocks.
half full of water, measure the height of the water √ √ √
9. (a) 202 − 42 = 384 ≈ 19.6 feet (b) 384 4 ≈ 4.9
and calculate the volume. Submerge the bulb, √
(c) tan(θ ) = 384
4 ≈ 4.9 so θ ≈ 1.37 (≈ 78.5◦ )
measure the height of the water again, and calcu-
late the new volume. The volume of the bulb is 11. (a) The equation of the line through P = (2, 3)
the difference of the two calculated volumes. and Q = (8, 11) is y − 3 = 86 ( x − 2) ⇒ 6y − 8x = 2.
Method 2: Fill the can with water and weigh it. Substituting x = 2a + 8(1 − a) = 8 − 6a and
Submerge the bulb (displacing a volume of water y = 3a + 11(1 − a) = 11 − 8a into the equation
equal to the volume of the bulb), remove the bulb, for the line, we get:
and weigh the can again. By subtracting, find the 6(11 − 8a) − 8(8 − 6a) = 66 − 48a − 64 + 48a = 2
weight of the displaced water and use the fact
that the density of water is 1 gram per 1 cubic for all a, so the point with x = 2a + 8(1 − a) and
centimeter to determine the volume of the bulb. y = 3a + 11(1 − a) is on the line through P and
 A2 contemporary calculus

Q for any a. Furthermore, 2 ≤ 8 − 6a ≤ 8 for 23. A point P = ( x, y) lies on the circle if and only
0 ≤ a ≤ 1, so the point in question must be on if its distance from C = (h, k) is r: dist( P, C ) = r.
the line segment PQ. So P is on the circle if and only if:
√ q
(b) dist( P, Q) = 62 + 82 = 10, while: ( x − h )2 + ( y − k )2 = r
q
dist( P, R) = (8 − 6a − 2)2 + (11 − 8a − 3)2 if and only if:
q
= (6 − 6a)2 + (8 − 8a)2 ( x − h )2 + ( y − k )2 = r 2
q
= 62 (1 − a)2 + 82 (1 − a)2
5 12
q 25. (a) − 12 (b) undefined (vertical line) (c) 5 (d) 0
= 100(1 − a)2 = 10 · |1 − a|
(horizontal line)
= |1 − a| · dist( P, Q)
27. (a) ≈ 2.22 (b) ≈ 2.24 (c) (by inspection) 3 units,
13. (a) m1 · m2 = (1)(−1) = −1 which occurs at the point (5, 3)
29. (a) If B ̸= 0, we can solve for y: y = − A C
B x + B , so
A
the slope is m = − B .
B
(b) The required slope is A (the negative recip-
(b) A
rocal of − B ) and the y-intercept is 0, so the
B
equation is y = A x or Bx − Ay = 0.
(c) Solve the equations Ax + By = C and Bx −
Ay = 0 simultaneously to get:
(c) Because 20 units of x-values are physically
AC BC
wider on the screen than 20 units of y-values. x= and y=
A2+ B2 A2 + B2
(d) Set the window so that:
(d) The
r distance from this point to the origin is:
(xmax − xmin) ≈ 1.7(ymax − ymin)
AC BC
( 2 )2 + ( 2 )2
15. (a) y − 5 = 3( x − 2) or y = 3x − 1 A + B2 A + B2
(b) y − 2 = −2( x − 3) or y = 8 − 2x s
A2 C 2 B2 C 2
(c) y − 4 = − 12 ( x − 1) or y = − 12 x + 9
2 = 2 2 2
+ 2
(A + B ) ( A + B2 )2
17. (a) y − 5 = 23 ( x − 2) or y = 32 x + 2 s
(b) y − 2 = 23 ( x + 1) or y = 32 x + 7 ( A2 + B2 ) C 2
2 =
(c) x = 3 ( A2 + B2 )2
√ s
19. Distance between the centers = 62 + 82 = 10. C2 |C |
= = √
(a) 10 − 2 − 4 = 4 (b) 10 − 2 − 7 = 1 (c) 0 (they A2 + B2 A2 + B2
intersect) (d) 15 − 10 − 3 = 2 (e) 12 − 10 − 1 = 1
p
21. Find dist( P, C ) = ( x − h)2 + (y − k )2 and com- Section 0.3
pare the value to r:
 1. A: a, B: c, C: d, D:b
 inside the circle
 if dist( P, C ) < r
3. A: b, B: c, C: d, D: a
P is on the circle if dist( P, C ) = r

 outside the circle if dist( P, C ) > r 5. (a) C (b) A (c) B
 A3

7. (a) f (1) = 4, g(1) is undefined, H (1) = −1 21. Approximate values:

x f (x) g( x )
(b) 0 1.0 1.0
1 2.0 1.0
2 2.0 −1.0
(c) f (3x ) = (3x )2 + 3 = 9x2 + 3, g(3x ) = 3 1.0 0.0
√
3x − 5 (for x ≥ 53 ), H (3x ) = 3x3x−2 4 1.5 0.5
(d) f ( x + h) = ( x + h)2 + 3 = x2 + 2xh + h2 + 3,
√ x+h 23. On your own.
g( x + h) = x + h − 5, H ( x + h) = x+ h −2
9. (a) m = 2 (b) m = 2x + 3 + h (c) If x = 1.3, then
Section 0.4
m = 5.6 + h; if x = 1.1, then m = 5.2 + h; if
x = 1.002, then m = 5.004 + h. 1. (a) ≈ −18, ≈ −2.2 (b) If T = 11◦ C, WCI11 =
f ( a+h)− f ( a)
11. h = 2a + h − 2 (if h ̸= 0). If a = 1: h. If 
11 if 0 ≤ v ≤ 6.5
√

a = 2: 2 + h. If √a = 3:√4 + h. If a = x: √2x + h − 2.

22.55 − 5.29 v + 0.279v if 6.5 < v ≤ 72
g( a+h)− g( a)
h = a+hh− a . If a = 1: 1+hh−1 . If 
√ √
2+ h − 2
√ √
3+ h − 3
 −2.2 if v > 72
a = 2: h . If a = 3: h . If a = x:
√ √
x +h− x
h . 3. g(0) = 3, g(1) = 1, g(2) = 2, g(3) = 3, g(4) = 1,
13. (a) Approx. 250 miles, 375 miles. (b) Approx. 200 g(5) = 1.
miles/hour. (c) By flying along a circular arc 
 3 − x if x < 1

about 375 miles from the airport (or by landing
g( x ) = x if 1 ≤ x ≤ 3
at another airport). 
 1 if x > 3

5. (a) f ( f (1)) = 1, f ( g(2)) = 2,

f ( g(0)) = 2, f ( g(1)) = 3
(b) g( f (2)) = 0, g( f (3)) = 1,
15. (a)
g( g(0)) = 0, g( f (0)) = 0
(c) f (h(3)) = 3, f (h(4)) = 2,
h( g(0)) = 0, h( g(1)) = −1
x −1 0 1 2 3 4
(b) Largest: x = 2; smallest: x = 4.
(c) Largest: at x = 5; smallest at x = 3. f (x) 3 3 −1 0 1 1
7. (a)
g( x ) −2 0 1 2 3 4
17. The path of the slide is a straight line tangent to
h( x ) −3 −2 −1 0 1 2
the graph of the path at the point of fall:
(b) f ( g(1)) = −1, f (h(1)) = 3, h( f (1)) = −3,
f ( f (2)) = 3, g( g(3.5)) = 3

(c)
x +1
19. (a) s(1) = 2, s(3) = 43 , s(4) = 5
4 (b) s( x ) = x
 A4 contemporary calculus

9. If L(d) represents location on day d: 27. f (1) = 22 = 1. f (0.5) = 1.25, f (1.25) =

 1.025, f (1.025) ≈ 1.0003049, f (1.0003049) ≈
England if d = Monday or Tuesday
1.000000046,. . . f (4) = 2.125, f (2.125) ≈



 France if d = Wednesday
L(d) = 1.297794, f (1.297794) ≈ 1.034166, f (1.034166) ≈
 Germany if d = Thursday or Friday

 1.000564,. . .
Italy if d = Saturday

29. (a) f (2) = 14 14 50
3 ≈ 4.7, f ( 3 ) = 9 ≈ 5.6,
11. Assuming the left portion is part of a parabola: f ( 50 158 158 482
9 ) = 27 ≈ 5.85, f ( 27 ) = 81 ≈ 5.95
( . . . f (4) = 3 ≈ 5.3, f ( 3 ) = 52
16 16
9 ≈ 5.8,
x2 if x < 2 52 160 160 484
f (x) = f ( 9 ) = 27 ≈ 5.93, f ( 27 ) = 81 ≈ 5.975
x − 1 if x > 2
. . . f (6) = 6. (b) c = 6 (c) Solve c = g(c) = 3c + A
13. (a) B(1) = 1 · f (1) = 1 · 11 = 1, B(2) = 2 · 12 = 1, to get 3c = c + 3A ⇒ 2c = 3A ⇒ c = 3A 2 is a
B(3) = 3 · 13 = 1. (b) B( x ) = x · f ( x ) = x · 1x = 1 fixed point of g.
(if x > 0) 31. On your own.

Section 0.5
15.
1. (a) x = 2, 4 (b) x = −2, −1, 0, 1, 2, 3, 4, 5
(c) x = −2, −1, 1, 3
17. (a) f ( g( x )) = 6x + 2 + 3A, g( f ( x )) = g(3x + √
3. (a) all x (all real numbers) (b) x > 3 −2 (c) all x
2) = 6x + 4 + A. If f ( g( x )) = g( f ( x )), then
5. (a) x = −2, −3, 3 (b) no values of x (c) x ≥ 0
A = 1.
(b) f ( g( x )) = 3Bx − 1, g( f ( x )) = 3Bx + 2B − 1. 7. (a) If x ̸= 2 and x ̸= −3, then x2 + x − 6 ̸= 0.
If f ( g( x )) = g( f ( x )), then B = 0. True. (b) If an object does not have 3 sides, then
it is not a triangle. True.
19. Graph of f ( x ) = x − ⌊ x ⌋:
9. (a) If your car does not get at least 24 miles per
gallon, then it is not tuned properly. (b) If you
cannot have dessert, then you did not eat your
vegetables.
11. (a) If you will not vote for me, then you do not
love your country. (b) If not only outlaws have
guns, then guns are not outlawed. (poor English)
21. f ( x ) = ⌊1.3 + 0.5 sin( x )⌋ works. The value of A
If someone legally has a gun, then guns are not
in f ( x ) = ⌊ A + 0.5 sin( x )⌋ determines the rela-
illegal.
tive lengths of the long and short parts of the
pattern. 13. (a) Both f ( x ) and g( x ) are not positive. (b) x is
not positive. (x ≤ 0) (c) 8 is not a prime number.
23. (a) g(1) = 1, g(2) = 1, g(3) = 0, g(4) = −1.
15. (a) For some numbers a and b, | a + b| ̸= | a| + |b|.
(b) Some snake is not poisonous. (c) Some dog
can climb trees.
(b) 17. If x is an integer, then 2x is an even integer. True.
Converse: If 2x is an even integer, then x is an
integer. True. (It is not likely that these were the
statements you thought of; there are lots of other
25. ≈ 0.739 starting with x = 1, 2, 10 or any value examples.)
 A5

19. (a) False. If a = 3, b = 4, then ( a + b)2 = 72 = 49, (b) At 10 a.m., temperature was rising about 5◦ per
but a2 + b2 = 32 + 42 = 9 + 16 = 25. (b) False. If hour; at 7 p.m., its was rising about −10◦ /hr
a = −2, b = −3, then a > b, but a2 = 4 < 9 = b2 . (falling about 10◦ /hr).
(c) True. 7. All of these answers are approximate.
21. (a) True. (b) False. If f ( x ) = x + 1 and 300 ft − 0 ft ft
(a) average velocity ≈ = 15
g( x ) = x + 2, Then f ( x ) · g( x ) = x2 + 3x + 2 20 sec − 0 sec sec
100 ft − 200 ft ft
is not a linear function. (c) True. (b) average velocity ≈ = −5
30 sec − 10 sec sec
23. (a) If a and b are prime numbers, then a + b is (c) At t = 10 seconds, velocity ≈ 30 feet per second
prime. False: take a = 3 and b = 5. (between 20 and 35 ft/sec); at t = 20 seconds,
(b) If a and b are prime numbers, then a + b is not velocity ≈ −1 feet per second; at t = 30 sec-
prime. False: take a = 2 and b = 3. onds, velocity ≈ −40 feet per second.
(c) If x is a prime number, then x is odd. False: 9. (a) A(0) = 0, A(1) = 3, A(2) = 6, A(2.5) = 7.5,
take x = 2. (This is the only counterexample.) A(3) = 9 (b) The area of the rectangle bounded
(d) If x is a prime number, then x is even. False: below by the t-axis, above by the line y = 3, on
take x = 3 (or 5 or 7 or. . . ) the left by the vertical line t = 1 and on the right
25. (a) If x is a solution of x + 5 = 9, then x is odd. by the vertical line t = 4. (c) Graph of y = 3x.
False: take x = 4.
(b) If a 3-sided polygon has equal sides, then it is Section 1.1
a triangle. True. (We also have non-equilateral 1. (a) 2 (b) 1 (c) DNE (does not exist) (d) 1
triangles.)
3. (a) 1 (b) −1 (c) −1 (d) 2
(c) If a person is a calculus student, then that per- 13
5. (a) −7 (b) 0 (DNE)
son studies hard. False (unfortunately), but
we won’t mention names. 7. (a) 0.54 (radian mode!) (b) −0.318 (c) −0.54
(d) If x is a (real number) solution of x2 − 5x + 9. (a) 0 (b) 0 (c) 0
6 = 0, then x is even. False: take x = 3. 11. (a) 0 (b) −1 (c) DNE
13. The one- and two-sided limits agree at x = 1,
Section 1.0 x = 4 and x = 5, but not at x = 2:
y −9
1. (a) m = x−3 . If x = 2.97, m = −−0.1791
0.03 = 5.97. If lim g( x ) = 1 lim g( x ) = 1 lim g( x ) = 1
0.006001 x →1− x →1+ x →1
x = 3.001, m = 0.001 = 6.001. If x = 3 + h,
(3 + h )2 − 9 9 + 6h + h2 − 9 lim g( x ) = 1 lim g( x ) = 4 lim g( x ) DNE
m = = = 6+h x →2− x →2+ x →2
(3 + h ) − 3 h
(b) When h is close to 0, 6 + h is close to 6. lim g( x ) = 2 lim g( x ) = 2 lim g( x ) = 2
y −4 x →4− x →4+ x →4
3. (a) m = x−2 . If x = 1.99, m = −−0.04990.01 = 4.99.
0.020016
If x = 2.004, m = 0.004 = 5.004. If x = 2 + h: lim g( x ) = 1 lim g( x ) = 1 lim g( x ) = 1
x →5− x →5+ x →5
(2 + h )2 + (2 + h ) − 2 − 4
 
m= = 5+h
(2 + h ) − 2 15. (a) 1.0986 (b) 1
(b) When h is very small, 5 + h is very close to 5.
17. (a) 0.125 (b) 3.5
5. All of these answers are approximate. Your an- 19. (a) A(0) = 0, A(1) = 2.25, A(2) = 5, A(3) = 8.25
swers should be close to these numbers. (b) A( x ) = 2x + 14 x2 (c) The area of the trapezoid
(a) average rate of temperature change ≈ bounded below by the t-axis, above by the line
80◦ − 64◦ 16◦ ◦ y = 12 t + 2, on the left by the vertical line t = 1
= =4 and on the right by the vertical line t = 3.
1 p.m. − 9 a.m. 4 hours hour
 A6 contemporary calculus

Section 1.2 an integer (for example, at x = 1 or x = 2): fails

(ii). (e) Where sin( x ) = 0 (for example, at x = 0,
1. (a) 2 (b) 0 (c) DNE (d) 1.5
x = ±π, ±2π . . .): fails (i). (f) At x = 0: fails (i).
3. (a) 1 (b) 3 (c) 1 (d) ≈ 0.8 (g) At x = 0: fails (i). (h) At x = 3: fails (i). (i) At
x = π2 : fails (i).
5. (a) f ( x ) = 0 for at least 3 values of x in the inter-
val 0 ≤ x ≤ 5. (b) 1 (c) 3 (d) 2 (e) Yes. (It does not
have to happen, but it is possible.)
5. 7. (a) f (0) = 0, f (3) = 9 and 0 ≤ 2 ≤ 9;
√
c = 2 ≈ 1.414 (b) f (−1) = 1, f (2) = 4 and
√
1 ≤ 3 ≤ 4; c = 3 ≈ 1.732 (c) f (0) = 0,
f ( π2 = 1 and 0 ≤ 12 ≤ 1; c = arcsin( 12 ) ≈ 0.524
(d) f (0) = 0, f (1) = 1 and 0 ≤ 13 ≤ 1; c = 13
(e) f (2)√ = 2, f (5) = 20 and 2 ≤ 4 ≤ 20;
7. (a) 2 (b) −1 (c) DNE (d) 2
c = 1+ 2 17 ≈ 2.561 (f) f (1) = 0, f (10) ≈ 2.30
(e) 2 (f) 2 (g) 1 (h) 2 (i) DNE
and 0 ≤ 2 ≤ 2.30; c = e2 ≈ 7.389
9. (a) When v = 0, L = A. (b) 0
9. Neither student is correct. The bisection algo-
11. (a) 4 (b) 1 (c) 2 (d) 0 (e) 1 (f) 1 rithm converges to the root labeled C.
13. (a) Slope of the line tangent to the graph of 11. (a) D (b) D (c) hits B
y = cos( x ) at the point (0, 1). (b) slope = 0
13. [−0.9375, −0.875], root ≈ −0.879; [1.3125, 1.375],
15. (a) ≈ 1 (b) ≈ 3.43 (c) ≈ 4 root ≈ 1.347; [2.5, 2.5625], root ≈ 2.532
17. At x = −1: “connected and smooth”; at x = 0: 15. [2.3125, 2.375], root ≈ 2.32
“connected with a corner”; at x = 1: “simple
17. [−0.375, −0.3125], root ≈ −0.32
hole”; at x = 2: “vertical jump”; at x = 3: “sim-
ple hole”; at x = 4: “corner”; at x = 5: “smooth”
19. Many lists will work. Here is one example: Put
| a −2|
an = 2 + n1 so an approaches 2 and ann −2 = 1 for 19.
| bn − 2 |
all n. Put bn = 2 − n1 so bn → 2 and bn − 2 = −1
for all n.
 
1
21. − x2 ≤ x2 cos ≤ x2 so limit is 0 21. (a) A(2.1) − A(2) is the area of the region
x2
  bounded below by the t-axis, above by the
1
23. − x2 ≤ x2 sin ≤ x2 so limit is 3 graph of y = f (t), on the left by the vertical
x
  line t = 2, and on the right by the vertical
1 1 1 line t = 2.1.
A(2.1)− A(2)
≈ f (2) or f (2.1) ⇒
25. 2 − 1 < ≤ 2 so limit is 1 0.1
x x2 x A(2.1)− A(2)
0.1 ≈1
(b) A(4.1) − A(4) is the area of the region
Section 1.3
bounded below by the t-axis, above by the
1. Discontinuous at 1, 3 and 4. graph of y = f (t), on the left by the vertical
3. (a) Discontinuous at x = 3: fails condition (i) line t = 4, and on the right by the vertical line
A(4.1)− A(4)
there. (b) At x = 2: fails (i). (c) Where cos( x ) < 0 t = 4.1. 0.1 ≈ f (4) ≈ 2
(for example, at x = π): fails (i). (d) Where x2 is 23. (a) Yes (you justify). (b) Yes. (c) Try it.
 A7

Section 1.4 Similarly, f ( x ) = 3 for any x with 2 < x < 2 + δ,

so it must be true that:
1. (a) If x is within 12 unit of 3 then 2x + 1 is within
1 unit of 7. (b) 0.3 (c) 0.02 (d) 2ϵ 2 < x < 2 + δ ⇒ L − 0.1 < 3 < L + 0.1

3. (a) If x is within 14 unit of 2 then 4x − 3 is within ⇒ L < 3.1

1 unit of 5. (b) 0.1 (c) 0.02 (d) 4ϵ But no value of L can simultaneously satisfy
ϵ ϵ
5. In 1, m = 2, δ = in 2, m = 3, δ = in 3, m = 4, L > 3.9 and L < 3.1, so we have reached a
2; 3;
δ = 4ϵ ; in 4, m = 5, δ = 5ϵ . In general: δ = |mϵ | . contradiction and our assumption must be false:
lim f ( x ) does not exist.
7. 0.02 inches x →2
√
3 21. Assume lim f ( x ) exists. Let ϵ = 0.1, so there
9. (a) Any value of x between 7.5 ≈ 1.957 and x →2
√3 must be a δ so that:
8.5 ≈ 2.043: If x is within 0.043 units of 2 then
x3 will be within 0.5 units of 8. (b) Any x be- 2 − δ < x < 2 + δ ⇒ L − 0.1 < f ( x ) < L + 0.1
√ √
tween 3 7.95 ≈ 1.9958 and 3 8.05 ≈ 2.0042: If x
We can assume that δ ≤ 1 (if not, we can replace
is within 0.0042 units of 2 then x3 will be within
our initial δ with δ = 1 because any smaller value
0.05 units of 8.
will also work). Now, f ( x ) = x for any x with
11. (a) Any value of x between 0 and 8: If x is within 2 − δ < x < 2, so it must be true that:
√
3 units of 3 then 1 + x will be within 1 unit of
2. (b) Any x between 2.99920004 and 3.00080004: 2 − δ < x < 2 ⇒ L − 0.1 < x < L + 0.1
√
If x is within 0.00079996 units of 3 then 1 + x ⇒ L < 0.1 + x < 2.1
will be within 0.0002 units of 2. On the other hand, f ( x ) = 6 − x for any x with
13. 0.0059964 inches 2 < x < 2 + δ ≤ 3, so it must be true that:

2 < x < 2 + δ ⇒ L − 0.1 < 6 − x < L + 0.1

⇒ L > 5.9 − x ≥ 2.9
(because x ≤ 3). But no value of L can simulta-
neously satisfy L < 2.1 and L > 2.9, so we have
15.
reached a contradiction and our assumption must
be false: lim f ( x ) does not exist.
x →2
23. Given any ϵ > 0, we know 2ϵ > 0, so there
is a number δ f > 0 such that | x − a| < δ f ⇒
| f ( x ) − L| < 2ϵ . Likewise, there is a number
17. On your own.
δg > 0 such that | x − a| < δg ⇒ | g( x ) − L| < 2ϵ .
19. Assume lim f ( x ) exists. Let ϵ = 0.1, so there Let δ be the smaller of δ f and δg . If | x − a| < δ
x →2
must be a δ so that: then | f ( x ) − L| < 2ϵ and | g( x ) − M| < 2ϵ so:

2 − δ < x < 2 + δ ⇒ L − 0.1 < f ( x ) < L + 0.1 |( f ( x ) − g( x )) − ( L − M))|

= |( f ( x ) − L) − ( g( x ) − M)|
Now, f ( x ) = 4 for any x with 2 − δ < x < 2, so ≤ | f ( x ) − L| + | g( x ) − M|
it must be true that: ϵ ϵ
< + =ϵ
2 2
2 − δ < x < 2 ⇒ L − 0.1 < 4 < L + 0.1
so f ( x ) − g( x ) is within ϵ of L − M whenever x
⇒ L > 3.9 is within δ of a.
A8 contemporary calculus

Section 2.0 [3( x + h) − 7] − [3x − 7] 3h

11. (a) msec = = = 3
( x + h) − x h
1. Values are approximate; your answers may vary. (b) mtan = lim msec = lim 3 = 3. (c) At x = 2,
h →0 h →0
x f (x) m( x ) x f (x) m( x ) mtan = 3. (d) f (2) = −1 so the tangent line is
y − (−1) = 3( x − 2) or y = 3x − 7.
0.0 1.0 1 2.5 0.0 −2 [ a( x + h) + b] − [ ax + b] ah
0.5 1.4 1
3.0 −1.0 −2 13. (a) msec = = = a
2 ( x + h) − x h
1.0 1.6 0 3.5 −1.3 0 (b) mtan = lim msec = lim a = a. (c) At x = 2,
h →0 h →0
1.5 1.4 − 12 4.0 −1.0 1 mtan = a. (d) f (2) = 2a + b so the tangent line is
2.0 1.0 −2 y − (2a + b) = a( x − 2) or y = ax + b.
3. (a) At x = 1, 3 and 4. (b) Largest at x = 4; small- [8−3(x+h)2 ]−[8−3x2 ]
15. (a) msec = ( x +h)− x
= −6x − 3h
est at x = 3. (b) mtan = lim msec = lim [−6x − 3h] = −6x.
h →0 h →0
(c) At x = 2, mtan = −6(2) = −12 (d) f (2) = −4
so the tangent line is y − (−4) = −12( x − 2) or
y = −12x + 20.
17. a = 1, b = 2, c = 0 ⇒ mtan = (2)(1)( x ) + 2 =
(c)
2x + 2 so we need p to satisfy 6 − ( p2 + 2p) =
(2p + 2)(3 − p) ⇒ p2 − 6p = 0 ⇒ p = 0 or p = 6
so the points are (0, 0) and (6, 48).

Section 2.1
1. derivative (a) of g (b) of h (c) of f
3. (a) msec = h − 4 ⇒ mtan = lim msec = −4
5. (a) h →0
(b) msec = h + 1 ⇒ mtan = lim msec = 1
h →0
5. (a) msec = 5 − h ⇒ mtan = lim msec = 5
h →0
(b) msec = 7 − 2x − h ⇒ mtan = 7 − 2x
(b)
7. (a) −1 (b) −1 (c) 0 (d) +1 (e) DNE (f) DNE
9. Using the definition:
( x + h )2 + 8 − x 2 + 8
   
(c) m( x ) = cos( x ) ′
f ( x ) = lim
7. Assume we turn off the engine at the point ( p, q) h →0 h
on the curve y = x2 and then find values of 2xh + h2
= lim = lim [2x + h] = 2x
p and q so the tangent line to y = x2 at the h →0 h h →0

point ( p, q) goes through the given point (5, 16): so f ′ (3) = 2 · 3 = 6

( p, q) is on y = x2 so q = p2 and the tan- 11. Using the definition:
gent line to y = x2 at ( p, p2 ) is y = 2px − p2 
2( x + h)3 − 5( x + h) − 2x3 − 5x
  
′
so, substituting x = 5 and y = 16, we have f ( x ) = lim
h →0 h
16 = 2p(5) − p2 ⇒ p2 − 10p + 16 = 0 ⇒ p = 2
6x2 h + 6xh2 + 2h3 − 5h
or p = 8. The solution we want (moving left to = lim
h →0 h
right along the curve) is p = 2, q = p2 = 4 (p = 8, h i
= lim 6x + 6xh + 2h2 − 5 = 6x2 − 5
2
q = 64 would be the right-to-left solution). h →0

9. Impossible: (1, 3) sits “inside” the parabola. so f ′ (3) = 6 · 32 − 5 = 49

 A9

13. For any constant C, if f ( x ) = x2 + C, then: 25. (a) d(4) = 256 ft; d(5) = 400 ft
f ( x + h) − f ( x ) (b) d′ ( x ) = 32x ⇒ d′ (4) = 128 ft/sec and
f ′ ( x ) = lim
h →0 h d′ (5) = 160 ft/sec.
( x + h )2 + C − x 2 + C
   
27. Marginal production cost is C ′ ( x ) = 1
√ dollars
= lim 2 x
h →0 h per golf ball, so C ′ (25) = √1 = 1 = $0.10 per
2 25 10
2xh + h2
= lim = lim [2x + h] = 2x ball and C ′ (100) = √1 1
= 20 = $0.05 per ball.
h →0 h h →0 2 100

The graphs of f ( x ) = x2 , g( x ) = x2 + 3 and 29. (a) A(0) = 0, A(1) = 12 , A(2) = 2 and A(3) = 92 .
x2
h( x ) = x2 − 5 are “parallel” parabolas: g is f (b) A( x ) = (x ≥ 0) (c) A′ ( x ) = x (d) A′ ( x )
2
shifted up 3 units; h is f shifted down 5 units. represents the rate at which A( x ) is increasing
15. f ′ ( x ) = 2x ⇒ f ′ (1) = 2 so an equation of the tan- (the rate at which area is accumulating).
gent line at (1, 9) is y − 9 = 2( x − 1) or y = 2x + 7; 2 −4
31. (a) 9x8 (b) (c) 5 (d) πx π −1 (e) 1 if x > −5
f ′ (−2) = −4 so an equation of the tangent line at 1
3x 3 x
(−2, 12) is y − 12 = −4( x + 2) or y = −4x + 4. and −1 if x < −5
17. f ′ ( x ) = cos( x ) ⇒ f ′ (π ) = cos(π ) = −1 so an 33. f ( x ) = x3 + 4x2 (plus any constant)
equation of the tangent line at (π, 0) is y − 0 =
35. f (t) = 5 · sin(t) 37. f ( x ) = 21 x2 + 13 x3
−1( x − π ) or y = − x + π; f ′ π2 = cos π2 = 0





so an equation of the tangent line at π2 , 1 is


y − 1 = 0 x − π2 or y = 1. Section 2.2
19. (a) y − 5 = 4( x − 2) or y = 4x − 3 (b) x + 4y = 22 1. (a) 0, 1, 2, 3, 5 (b) 0, 3, 5
or y = −0.25x + 5.5 (c) f ′ ( x ) = 2x, so the 3. In the table below, und. means “undefined”:
tangent line is horizontal only if x = 0: at
f (x) ′
 
f (x)
the point (0, 1). (d) f ′ ( p) = 2p (the slope x f ( x ) · g( x ) ( f ( x ) · g( x ))′ g( x ) g( x )
of the tangent line) so y − q = 2p( x − p) or
0 2 13 2 −7
y = 2px + (q − 2p2 ). Since q = p2 + 1, the tangent
1 −15 16 − 35 4
25
line is y = 2px + ( p2 + 1 − 2p2 ) = 2px − p2 + 1.
2 0 −6 0 − 23
(e) We need p such that −7 = 2p(1) − p2 + 1, so
3 0 3 und. und.
p2 − 2p − 8 = 0 ⇒ p = −2 or p = 4. There are
 ′
two such points: (−2, 5) and (4, 17).
( f · g)′
f f
x f +g ( f + g)′ f ·g g g
21. (a) y′ ( x ) = 2x, so y′ (1) = 2 and the angle is
3
arctan(2) ≈ 1.107 radians ≈ 63◦ (b) y′ ( x ) = 3x2 5. 1 5 0 6 2 2 − 10
4
1 1 1 1
so y′ (1) = 3 and the angle is arctan(3) ≈ 1.249 2 4 2 3 2 3 − 18
radians ≈ 72◦ (c) 1.249 − 1.107 radians = 0.142 3 4 0 4 0 1 1
radians (or 71.57◦ − 63.43◦ ≈ 8.1◦ )
7. (a) D (( x − 5)(3x + 7)) = ( x − 5) · 3 + (3x + 7) · 1
23. Units on the horizontal axis are “seconds”; units
(b) D 3x2 − 8x − 35 = 6x − 8 (the same result)


on the vertical axis are “feet per second.”
9. 19 · 3x2 = 57x2 11. cos( x ) − sin( x )

13. x2 · [− sin( x )] + cos( x ) · 2x

15. Applying the Product Rule:
 
D sin2 ( x ) = sin( x ) · cos( x ) + sin( x ) · cos( x )

= 2 sin( x ) cos( x ) = sin(2x )

 A10 contemporary calculus

x2 (− sin( x )) − (cos( x )) (2x )

 
d cos( x ) 49. Everywhere except at x = 1.
17. 2
= 2
=
dx x ( x2 ) 51. Everywhere. The only possible difficulty is at
− x [ x sin( x ) + 2 cos( x )] x sin( x ) + 2 cos( x ) x = 0: the definition of the derivative gives
=−
x4 x3 f ′ (0) = 1 and the derivatives of the “two pieces”
2
(1 + x ) · 0 − 1 · 2x 2x
19. =− of f match (and equal 1) at x = 0.
2 2 2
(1 + x ) (1 + x 2 )
53. Continuity of f at x = 1 requires A + B = 2. The
cos(θ ) · 0 − 1 · (− sin(θ )) sin(θ )
21. = “left derivative” of f at x = 1 is D( Ax + B) = A
2 cos 2 (θ )
(cos(θ )) and the “right derivative” of f at x = 1 is
cos(θ ) · cos(θ ) − sin(θ ) · (− sin(θ )) 1 3 (if x > 1 then D( x2 + x ) = 2x + 1), so to
23. 2
=
(cos(θ )) cos2 (θ ) achieve differentiability we need A = 3 and
25. 40x4 − 12x3 + 6x2 + 14x − 12 B = 2 − A = −1.
27. f ( x ) = ax2 + bx + c so f (0) = c, hence f (0) = 0 55. (a) h( x ) = 128x − 2.65x2 ⇒ h′ ( x ) = 128 − 5.3x
⇒ c = 0. f ′ ( x ) = 2ax + b so f ′ (0) = b, hence so h′ (0) = 128 ft/sec, h′ (1) = 122.7 ft/sec
f ′ (0) = 0 ⇒ b = 0. Finally, f ′ (10) = 20a + b = and h′ (2) = 117.4 ft/sec
20a so f ′ (10) = 30 ⇒ 20a = 30 ⇒ a = 32 ⇒ (b) v( x ) = h′ ( x ) = 128 − 5.3x ft/sec
f ( x ) = 32 x2 + 0x + 0 = 23 x2 . (c) v( x ) = 0 when x = 128
≈ 24.15 sec
5.3
29. Their graphs are vertical shifts of each other, and (d) h(24.15) ≈ 1545.66 ft
their derivatives are equal. (e) about 48.3 seconds: 24.15 up and 24.15 down
31. f ( x ) · g( x ) = k ⇒ D ( f ( x ) · g( x )) = D (k) = 0 so 57. (a) h( x ) = vo x − 16x2 ⇒ h′ ( x ) = vo − 32x ft/sec
f ( x ) · g′ ( x ) + g( x ) · f ′ ( x ) = 0. If f ( x ) ̸= 0 and (b) Max height when x = 32 vo
: max height =
f ′ (x) ′
g( x ) ̸= 0, then f (x)
= − gg((xx)) . vo

vo


h 32 = vo 32 − 16 32 = 64 ft vo 2

(vo )2

33. (a) f ′ ( x ) = 2x − 5, so f ′ (1) = −3 vo vo

(c) Time aloft = 2 32 = 16 sec
(b) f ′ (x) = 0 only if x = 52 . (vo ) 2 √
59. (a) 64 = 3.75 ⇒ vo = 8 3.75 ≈ 15.5 ft/sec
√
35. (a) f ′ (x) = 3 + 2 sin( x ); f ′ (1)
= 3 + 2 sin(1) ≈ 4.7 vo
(b) 2 32 = 8 16

3.75
≈ 0.97 sec
√
(b) f ′ (x) > 0 because sin( x ) > − 32 for all x. ( v o )2 (8 3.75)2 240
(c) Max lift = 2g = 2(5.3) = 10.6 ≈ 22.64 ft
37. (a) f ′ ( x ) = 3x2 + 18x = 3x ( x + 6), so f ′ (1) = 21
61. (a) y′ = − x12 so y′ (2) = − 14 and the tangent line
(b) f ′ ( x ) = 0 only if x = 0 or x = −6.
is y − 1/2 = (− 14 )( x − 2) or y = − 14 x + 1
39. (a) f ′ ( x ) = 3x2 + 4x + 2, so f ′ (1) = 9
(b) x-intercept at (4, 0); y-intercept at (0, 1)
(b) f ′ ( x ) > 0 (the discriminant 42 − 4(3)(2) < 0).
(c) A = 12 (base)(height) = 21 (4)(1) = 2
41. (a) f ′ (x)
= x cos( x ) + sin( x ) ⇒ f ′ (1)
= cos(1) + 63. Because (1, 4) and (3, 14) are on the parabola, we
sin(1) ≈ 1.38
need a + b + c = 4 and 9a + 3b + c = 14. Subtract-
(b) The graph of f ′ ( x ) crosses the x-axis infinitely ing the first equation from the second, 8a + 2b =
often. The root of f ′ at x = 0 is easy to see; ap- 10; f ′ ( x ) = 2ax + b so f ′ (3) = 6a + b = 9, the
proximate others, such as those near x = 2.03 slope of y = 9x − 13. Now solve the system
and 4.91 and −2.03, using technology. 8a + 2b = 10 and 6a + b = 9 to get a = 2 and
43. f ′ ( x ) = 3x2 + 2Ax + B: the graph of y = f ( x ) has b = −3. Then use a + b + c = 4 to get c = 5:
two distinct “vertices” if f ′ ( x ) = 0 for two dis- a = 2, b = −3, c = 5.
tinct values of x. This occurs if the discriminant 65. (a) f ( x ) = x3
of 3x2 + 2Ax + B > 0: (2A)2 − 4(3)( B) > 0. (b) g( x ) = x3 + 1
45. Everywhere except at x = (2k + 1) π2 . (c) If h( x ) = x3 + C for any constant C, then
47. Everywhere except at x = 0 and x = 3. D (h( x )) = 3x2 .
 A11

67. (a) For 0 ≤ x ≤ 1, f ′ ( x ) = 1 so f ( x ) = x + C; 13. (a) If h(t) = 5t, then v(t) = h′ (t) = 5 and
because f (0) = 0, we know C = 0 and K (1) = K (2) = 12 m(52 ) = 12.5m. (b) If h(t) = t2 ,
f ( x ) = x. For 1 ≤ x ≤ 3, f ′ ( x ) = 2 − x then v(t) = h′ (t) = 2t so v(1) = 2 and
so f ( x ) = 2x − 21 x2 + K; because f (1) = 1, v(2) = 4, hence K (1) = 12 m(22 ) = 2m and
we know K = − 21 and f ( x ) = 2x − 12 x2 − 21 . K (2) = 12 m(42 ) = 8m.
For 3 ≤ x ≤ 4, f ′ ( x ) = x − 4 so f ( x ) = 15. x · (sin( x ))′ + sin( x ) · ( x )′ = x cos( x ) + sin( x )
1 2
2 x − 4x + L; because f (3) = 1, we know 17. f ′ ( x ) = e x − sec( x ) · tan( x )
L = 17 1 2 17
2 and f ( x ) = 2 x − 4x + 2 . 19. f ′ ( x ) = −e− x + cos( x )
(b) A vertical shift, up 1 unit, of the graph in (a).
21. f ′ ( x ) = 7( x − 5)6 (1) so f ′ (4) = 7(−1)6 (1) = 7.
Tangent line: y − (−1) = 7( x − 4) or y = 7x − 29.
Section 2.3
1
23. f ′ ( x ) = 12 (25 − x2 )− 2 (−2x ) = √ −x so f ′ (3) =
1. D( f 2 ( x )) = 2· f 1 ( x ) · f ′ ( x );
at x = 1, D( f 2 ( x ))
= 25− x2
√ −3 = − 34 . Tangent line: y − 4 = − 34 ( x − 3) or
2(2)(3) = 12. D( f 5 ( x )) = 5 · f 4 ( x ) · f ′ ( x ); at x = 25−9
1
1, D( f 5 ( x )) = 5(24 )(3) = 240. D( f 2 ( x )) = 12 · 3x + 4y = 25.
1 1 1
f − 2 ( x ) · f ′ ( x ); at x = 1, D( f 2 ( x )) = ( 12 )(2− 2 )(3) 25. f ′ ( x ) = 5( x − a)4 (1) so f ′ ( a) = 5( a − a)4 (1) = 0.
√
3 3 2 Tangent line: y − 0 = 0( x − a) or y = 0.
= √ = 4 .
2 2
27. (a) f ′ ( x ) = e x so f ′ (3) = e3 . Tangent line:
x f (x) f ′ (x) D( f 2 ) D( f 3 ) D( f 5 ) y − e3 = e3 ( x − 3) or y = e3 x − 2e3 .
3. 1 1 −1 −2 −3 −5 (b) 0 − e3 = e3 ( x − 3) ⇒ −1 = x − 3 ⇒ x = 2
3 2 −3 −12 −36 −240 (c) f ′ ( p) = e p so tangent line at ( p, e p ) is y − e p =
e p ( x − p); x-intercept: 0 − e p = e p ( x − p) ⇒
5. f ′ ( x ) = 5 · (2x − 8)4 · (2) = 10(2x − 8)4
−1 = x − p ⇒ x = p − 1
7. f ′ ( x ) = x · 5 · (3x + 7)4 · 3 + 1 · (3x + 7)5 =
29. f ′ ( x ) = − sin( x ) ⇒ f ′′ ( x ) = − cos( x )
(3x + 7)4 [15x + (3x + 7)] = (3x + 7)4 (18x + 7)
1 1 31. f ′ ( x ) = x2 cos( x ) + 2x sin( x ) ⇒
9. f ′ ( x ) = ( x2 + 6x − 1)− 2 · (2x + 6), which we f ′′ ( x ) = − x2 sin( x ) + 2x cos( x ) + 2x cos( x ) +
2
x+3 2 sin( x ) = − x2 sin( x ) + 4x cos( x ) + 2 sin( x )
can rewrite as √
2
x + 6x − 1 33. f ′ ( x ) = e x cos( x ) − e x sin( x ) ⇒ f ′′ ( x ) = −2e x sin( x )
11. (a) Graph of h(t) = 3 − 2 sin(t):
35. q′ = linear, q′′ = constant, q′′′ = q(4) = · · · = 0
37. p(n) = constant ⇒ p(n+1) = 0
39. f ( x ) = 5e x
41. f ( x ) = (1 + e x )5
43. No. Using the definition of the derivative:
 
1
f (0 + h ) − f (0) ( 0 + h ) · sin 0+ h − 0
lim = lim
h →0 h h →0 h
(b) When t = 0, h(0) = 3 feet.  
1
(c) highest: 5 feet above the floor which simplifies to lim sin ; to see that this
h →0 h  
lowest: 1 foot above the floor. last limit does not exist, graph sin 1h or evaluate
(d) v(t) = h′ (t) = −2 cos(t) ft/sec  
sin 1h for small values of h.
a(t) = v′ (t) = 2 sin(t) ft/sec2 .
(e) This spring oscillates forever. The motion of a  x
1
real spring would “damp out” due to friction. 45. 1+ ≈ 2.718 . . . = e when x is large.
x
 A12 contemporary calculus

47. (a) s2 = 2.5, s3 ≈ 2.67, s4 ≈ 2.708, s5 ≈ 2.716, (c) = 14.7(−0.0000385)e−0.0000385h psi/ft. At h =

dP
dh
s6 ≈ 2.718, s7 ≈ 2.71825, s8 ≈ 2.718178 (b) They 2,000 ft, dP −0.0000385(2000)
dh = 14.7(−0.0000385) e
are approaching e. psi/ft ≈ −0.000524 psi/ft. Finally, dP dt =
500(−0.000524) ≈ −0.262 psi/min.
Section 2.4 (d) If temperature is constant, then (from physics!)
we know (pressure)(volume) is constant, so
1. f ( x ) = x5 , g( x ) = x3 − 7x ⇒ f ◦ g( x ) = ( x3 − 7x )5
decreasing pressure means increasing volume.
5
3. Setting f ( x ) = x 2 and g( x ) = 2 + sin( x ) yields 2 cos(z) [− sin(z)] − sin(2z)
5 27. = p
f ◦ g( x ) = (2 + sin( x )) 2 . (The functions f ( x ) =
p
2
2 1 + cos (z) 2 1 + cos2 (z)
√
x and g( x ) = (2 + sin( x ))5 also work.) d
29. dxtan(3x + 5) = 3 · sec2 (3x + 5)
5. f ( x ) = | x |, g( x ) = x2 − 4 ⇒ f ◦ g( x ) =  x2 − 4 √ √
 
x + 1 · √1



31. D sin( x + 1) = cos
2 x +1
7. y = u5 , u = x3 − 7x; y = u4 , u = sin(3x − 8);
5 d  
y = u 2 , u = 2 + sin( x ); y = √1u , u = x2 + 9; 33. esin( x) = esin( x) · cos( x )
√ dx
y = |u|, u = x2 − 4; y = tan(u), u = x √
35. f ( x ) = x ⇒ f ′ ( x ) = 2√ 1
x
; x (t) = 2 + 21
t ⇒

x f ◦ g( x ) ( f ◦ g)′ ( x ) x ′ (t) = − 21
t2
. At t = 3, x = 9 and x ′ (t) = − 21
9 =
 
− 37 so dtd ( f ( x (t))) = √1 − 73 = − 18 7


−2 1 0 .
2 9
−1 1 2 37. f ( x ) = tan3 ( x )
⇒ ′
f (x) = 3 · tan2 ( x ) · sec2 ( x );
8. & 9.
0 0 1 ′
x (t) = 8 ⇒ x (t) = 0. When t = 3, x = 8 and
1 2 2 x ′ (t) = 0 so dt
d
( f ( x (t))) = 0.
2 −2 −2 1 11
39. f ( x ) = 77 (7x − 13)

11. g(2) ≈ 2, g′ (2) ≈ −1 41. f ( x ) = − 12 cos(2x − 3)

( f ◦ g)(2) = f ( g(2)) ≈ f (2) ≈ 1 43. f ( x ) = esin( x)
f ′ ( g(2)) ≈ f ′ (2) ≈ 0 45. −2 sin(2x ) = 2 cos( x ) [− sin( x )] − 2 sin( x ) · cos( x )
( f ◦ g)′ (2) = f ′ ( g(2)) · g′ (2) ≈ 0 or sin(2x ) = 2 sin( x ) · cos( x )

4 
3
13. D 1 − 3x = 4 1 − 3x ) · x32 47. 3 cos(3x ) = 3 cos( x ) − 12 sin2 ( x ) · cos( x ) so cos(3x )
3
= cos( x ) 1 − 4 sin2 ( x ) = cos( x ) 1 − 4 + 4 cos2 ( x )
     
−5 cos( x )
15. 5 − 12 (2 + sin( x ))− 2 · cos( x ) = 3
2(2+sin( x )) 2 = 4 cos3 ( x ) − 3 cos( x )
17. x2 · cos( x2 + 3) · 2x + sin( x2 + 3) · 2x =
 
49. y′ = 3Ax2 + 2Bx
2x x2 cos( x2 + 3) + sin( x2 + 3)
 
51. y′ = 2Ax · cos Ax2 + B


19. cos( x73 − x) = 7 sec x3 − x so D 7 sec x3 − x =




Bx
7(3x2 − 1) · sec( x3 − x ). tan( x3 − x ) 53. y′ = √
A + Bx2
21. D (e x + e− x ) = e x − e− x 55. y′ = B · sin( Bx )
23. (a) h(0) = 2 feet above the floor.
57. y′ = −2Ax · sin( Ax2 + B)
(b) h(t) = 3 − cos(2t) ft, v(t) = h′ (t) = 2 sin(2t)
59. y′ = x B · e Bx + e Bx = ( Bx + 1) · e Bx


ft/sec, a(t) = v′ (t) = 4 cos(2t) ft/sec2
(c) K = 21 mv2 = 12 m (2 sin(2t))2 = 2m sin2 (2t), 61. y′ = A · e Ax + A · e− Ax
dK A · sin( Bx ) − Ax · B · cos( Bx )
dt = 8m · sin(2t ) · cos(2t ) 63. y′ =
25. (a) P(0) = 14.7 psi (pounds per square inch), sin2 ( Bx )
P(30000) ≈ 4.63 psi (Cx + D ) A − ( Ax + B)C AD − BC
65. y′ = 2
=
(Cx + D ) (Cx + D )2
(b) 10 = 14.7e−0.0000385h ⇒ 10
 14.7 = e
−0.0000385h ⇒
67. (a) y′ = AB − 2Ax (b) x = AB
= B
(c) y′′ = −2A

1 10
h = −0.0000385 ln 14.7 ≈ 10,007 ft 2A 2
 A13

 
69. (a) y′ = 2ABx − 3Ax2 = Ax (2B − 3x ) 25. D 5sin( x) = 5sin( x) · ln(5) · cos( x )
2B
(b) x = 0, 3 1
27. · (sec( x ) tan( x ) + sec2 ( x )) = sec( x )
(c) y′′ = 2AB − 6Ax sec( x )+tan( x )

71. (a) y′ = 3Ax2 + 2Bx = x (3Ax + 2B) 29. f ( x ) = ln( x ), f ′ ( x ) = 1x . Let P = ( p, ln( p)). So
2B
(b) x = 0, − 3A we need y − ln( p) = 1p ( x − p) with x = 0 and
(c) y′′ = 6Ax + 2B y = 0: −ln( p) = −1 ⇒ p = e, hence P = (e, 1).
d   2x 100Ae−t
73. arctan( x2 ) = 31. 100(−1)(1 + Ae−t )−2 · ( Ae−t )(−1) =
dx 1 + x4 (1 + Ae−t )2
1 ex
75. D (arctan(e x )) = 2
· ex = 33. f ( x ) = 8 ln( x ) + any constant
1 + (e x ) 1 + e2x
35. f ( x ) = ln(3 + sin( x ))+ any constant
3x2
77. D(arcsin( x3 )) = √
1 − x6 37. g( x ) = 35 e5x + any constant
d 1 et 2
39. f ( x ) = e x + any constant
arcsin(et ) = p · et = √


79.
dt 1 − ( e t )2 1 − e2t
d 1 41. h( x ) = ln(sin( x )) + any constant
81. (ln(sin( x ))) = · cos( x ) = cot( x )
dx sin( x ) 43. (a) When t = 0, A is at (0, 1) and B is at (0, 1).
d 1 d d
83. ds (ln(es )) = es · es = 1 or ds (ln(es )) = ds ( s ) =1 When t = 1, A is at (1, 3) and B is at (1, 3).
(b) Both robots traverse the same path (the line
Section 2.5 segment y = 2x + 1 for 0 ≤ x ≤ 1):
1 1
1. D(ln(5x )) = ·5 =
5x x
k 1 k −1 k
3. D(ln( x )) = k · kx =
x x
1
5. D(ln(cos( x ))) = · (− sin( x )) = − tan( x )
cos( x )
1 1 1
7. D(log2 (5x )) = · · (5) =
5x ln(2) x ln(2)
1
9. D(ln(sin( x ))) = · cos( x ) = cot( x )
sin( x )
1 1 cot( x )
11. D(log2 (sin( x ))) = · · cos( x ) =
sin( x ) ln(2) ln(2)

13. D(log5 (5x )) = D( x ) = 1 dy

(c) = 2 for each, because y = 2x + 1.
dx
1 √ √
15. D( x · ln(3x )) = x · · 3 + ln(3x ) · 1 = 1 + ln(3x ) (d) A: dx
dy
12 + 22 = 5
3x dt = 1, dt = 2, speed =
B: dx = 2t, dy
p
x · 1x − ln( x ) · 1 1 − ln( x ) dt = 4t, speed = √ (2t)2 + (4t)2
 
ln( x ) √
dt
17. D = 2
=
x x x2 = 2 5t. At t = 1, B’s speed is 2 5.
  1
 1  1

(e) Moves along the same path y = 2x + 1, but
19. D ln (5x − 3) 2 = 1
· D (5x − 3) 2 =
(5x − 3) 2 to the right and up for about 1.57 minutes,
1 1 − 12 5 1 reverses direction and returns to its starting
1
· (5x − 3) · D(5x − 3) = ·
(5x − 3) 2 2 2 5x − 3 point, then continues left and down for 1.57
1 sin(ln(w)) minutes, reverses, and continues to oscillate.
21. D (cos(ln(w))) = − sin(ln(w)) · = −
w w dx dy dy
d
q 
1 45. When t = 1: dt = +, dt = −, dx = −.
23. ln(t + 1) = When t = 3, dx
= −, dy
= −, dy
= +.
dt 2( t + 1) dt dt dx
 A14 contemporary calculus

 
47. (a) x (t) = R(t − sin(t)), y(t) = R(1 − cos(t)): 1 dV 2π dr
7. Volume V = πr2 h = πr2 3 ⇒ dt = 3 r · dt .
So when r = 50 ft and dr
dt = 6 ft , we have
hr
3
ft3
 
dV 2π ft ft
dt = 3 (50 ft) 6 hr = 200π hr ≈ 628.32 hr .

9. w(t) = h(t) for all t so dw dt = dh . V = 13 πr2 h and

 2 dt
r = w2 = 2h so V = 31 π 2h h = 12 π 3
h ⇒ dV
dt =
dx dy dy
(b) dt = R (1 − cos( t )), dt = R sin( t ), so dx = π 2 dh dh ft
sin(t) dy 4 h dt . When h = 500 ft and dt = 2 hr , then
1−cos(t)
. When t = π2 , then dx
dt = R and dt = dV ft3
π 2
dy √ √ dt = 4 (500) (2) = 125000π hr .
R so dx = 1 and speed = R2 + R2 = R 2.
11. Let x be the distance from the lamp post to the
When t = π, dx = 2R and dy dy
dt = 0 so dx = 0
pdt person and L be the length of the shadow (both
and speed = (2R)2 + 02 = 2R.

2 in feet). By similar triangles, L6 = 8x ⇒ L = 34 x.
x 2
+ y5 = 1.


49. (a) The ellipse 3 We also know that dx = 3 ft .
(b) The ellipse x 2


+ By

2
1 if A ̸= 0 and
=  dt  sec
A (a) dL 3 dx 3 ft ft
dt = 4 dt = 4 3 sec = 2.25 sec .
B ̸= 0. If A = 0, the motion is oscillatory
d
( x + L) = dx dL ft
along the x-axis; if B = 0, the motion is oscil- (b) dt dt + dt = 5.25 sec .
latory along the y-axis. (We don’t actually need the value of x.)
(c) (3 · cos(t), −5 · sin(t)) works. 13. (a) sin(35◦ ) = ⇒ h = 500 sin(35◦ ) ≈ 287 ft
h
500
(b) L = length of the string soh = L sin ◦
 (35 ) and
Section 2.6 dh ◦ dL ft ft
dt = sin(35 ) dt ≈ (0.57) 10 sec = 5.7 sec

15. V = s3 − 34 πr3 ⇒ dV 2 ds 2 dr

dt = 3s dt − 4πr dt so when
dV
1. V = 43 πr3 ⇒ = 4πr2 · dr
dt dt . We know dr 
dt r =3 =
 r = 4 ft, dr = 1 ft , s = 12 ft and ds ft dV
dt = 3 hr : dt =
2 in/min, so dV = 4π (3 in)2 (2 in/min) = dt hr

dt  r =3
  3

72π in3 /min ≈ 226.19 in3 /min. 3(12 ft)2 3 ft − 4π (4 ft)2 1 ft ≈ 1094.94 ft .
hr hr hr
h i 3
3. (a) A = 21 bh ⇒ dA dt = 1
2 b dh
dt + h db
dt so with The volume is increasing at about 1,095 ft .
hr
b = 15 in, h = 13 in, db in dh in 17. Given: dV 2
dt = 3 hr , dt = −3 hr : dt = k · 2πr with k constant. We also
     have V = 3 πr ⇒ dV
2 3 2 dr
dt = 2πr dt so k · 2πr =
2
dA 1 in in dr dr
= (15 in) −3 + (13 in) 3 2πr2 dt ⇒ dt = k. The radius r is changing at a
dt 2 hr hr
constant rate.
which is < 0, so A is decreasing.
√ 19. (a) A = 5x (b) dA 2
(b) Hypotenuse C = b2 + h2 so: dx = 5 for all x > 0. (c) A = 5t
dA dA
(d) dt = 10t. When t = 1, dt = 10; when
dC b db + h dh 15(3) + 13(−3) t = 2, dA dA
= √dt dt
= √ dt = 20; when t = 3, dt = 30.
dt b2 + h2 152 + 132 dA
(e) A = 10 + 5 · sin(t) ⇒ dt = 5 · cos(t).
which is > 0 so C is increasing. 21. In (b) and (c) we must use radians because our for-
(c) Perimeter P = b + h + C so: mulas for derivatives of trigonometric functions
dP db dh dC 6 assume angles are measured in radians.
= + + = 3 + (−3) + √
dt dt dt dt 394 (a) tan(10◦ ) = 40
x ⇒x= 40
tan(10◦ )
≈ 226.9 ft.
which is > 0 so P is increasing. (b) x = tan40(θ ) = 40 cot(θ ) ⇒ dx 2 dθ
dt = −40 csc ( θ ) dt
dy
5. (a) Perimeter P = 2x + 2y ⇒ dP dx
dt = 2 dt + 2 dt = ⇒ dθ
2
sin (θ ) dx ◦
dt = − 40 dt so when θ = 10 ≈ 0.1745
2(3 ft/sec) + 2(−2 ft/sec) = 2 ft/sec. dx ft
dy radians and dt = −25 :
(b) Area A = xy ⇒ dA = x dt + y dx dA min
 dt dt ⇒ dt =
dθ sin2 (0.1745) (0.1736)2 (25)

ft + (8 ft) 3 ft = 0 ft2 /sec.
(12 ft) −2 sec sec =− (−25) ≈
dt 40 40
 A15

rad ≈ 1.079 ◦ .
≈ 0.0188 min 19. (a) 2(0) − ⌊2(0)⌋ = 0 − 0 = 0
min
(c) dx
= −40 csc2 (θ ) dθ
dt =
−40 dθ so when θ = (b) 2( 21 ) − ⌊2( 12 )⌋ = 1 − 1 = 0
dt sin2 (θ ) dt
◦ rad 2( 41 ) − ⌊2( 14 )⌋ = 21 − 0 = 12 → 0
10◦ ≈ 0.1745 rad and dθ
dt = 2 min ≈ 0.0349 min :
2( 81 ) − ⌊2( 18 )⌋ = 41 − 0 = 14 → 21 → 0
dθ −40 ft 2( 21n ) − ⌊2( 21n )⌋ = 2n1−1 − 0 = 2n1−1
≈ (0.0349) ≈ −46.3
dt (0.1736)2 min → 2n1−2 → · · · → 41 → 1
→0
2
(The “-” indicates the distance to the sign is
21. (a) If 0 ≤ x ≤ 12 , then f stretches x to twice its
decreasing: you are approaching the sign.)
value, 2x. If 12 < x ≤ 1, then f stretches
Your speed is 46.3 ft .
min x to twice its value (2x) and “folds” the
part above the value 1 (2x − 1) to below 1:
Section 2.7 1 − (2x − 1) = 2 − 2x.
1. The locations of x1 and x2 appear below: (b) f ( 32 ) = 23
f ( 52 ) = 45 , f ( 45 ) = 25 , and the values continue
to cycle.
f ( 27 ) = 47 , f ( 47 ) = 67 , f ( 76 ) = 27 , and the values
continue to cycle.
f ( 29 ) = 49 , f ( 49 ) = 89 , f ( 98 ) = 29 , and the values
continue to cycle.
(c) 0.1, 0.2, 0.4, 0.8, 0.4, 0.8, and the pair of values
0.4 and 0.8 continue to cycle.
0.105, 0.210, 0.42, 0.84, 0.32, 0.64, 0.72, 0.56,
0.88, 0.24, 0.48, 0.96, 0.08, 0.16, 0.32, . . .
0.11, 0.22, 0.44, 0.88, 0.24, 0.48, 0.96, 0.08, 0.16,
3. x0 = 1: a; x0 = 5: b
0.32, 0.64, 0.72, 0.56, 0.88, . . .
5. x0 = 1: 1, 2, 1, 2, 1, . . .
(d) Probably so.
x0 = 5: x1 is undefined because f ′ (5) = 0
7. If f is differentiable, then f ′ ( x0 ) = 0 and x1 =
f (x )
x0 − f ′ ( x0 ) is undefined. Section 2.8
0

9. f ( x ) = x4 − x3 − 5 ⇒ f ′ ( x ) = 4x3 − 3x2 so
3 37
1. (a) The desired points appear below:
x0 = 2 ⇒ x1 = 2 − 20 = 20 = 1.85 ⇒ x2 =
1.854 −1.853 −5
1.85 − 4(1.85)3 −3(1.85)2
≈ 1.824641
11. f ( x ) = x − cos( x ) ⇒ f ′ ( x ) = 1 + sin( x ) so x0 =
0.7 ⇒ x1 = 0.7394364978 ⇒ x2 = 0.7390851605,
so root ≈ 0.74.
x 2
13. To solve x+ 3 = x − 2 we search for roots of
x
f ( x ) = x2 − 2 − x+ 3 . If x0 = −4, then the iterates
xn → −3.3615; if x0 = −2, then xn → −1.1674; if
x0 = 2, then the iterates xn → 1.5289.
(b) The “error” appears in the figure above.
15. For x5 − 3 = 0 and x0 = 1, xn → 1.2457.
17. f ( x ) = x3 − A ⇒ f ′ ( x ) = 3x2 so: 3. (a) With f (4) = 2 and f ′ (4) = 1
√ = 14 , the tan-
2 4
xn3 − A gent line is y − 2 = 14 ( x − 4) or y = 14 x + 1
 
1 A √
x n +1 = x n − = 2xn + 2
3xn2 3 xn and 4.2 = f (4.2) ≈ 14 (4.2) + 1 = 2.05
A16 contemporary calculus

(b) f (81) = 9 and f ′ (81) = 18

1
so tangent line is (b) (base)(height) = (2.3)((2.3)2 + 1) = 14.467
1 1
y − 9 = 18 ( x − 81) or y = 18 ( x − 81) + 9 and so actual ∆A = 14.467 − (2)(22 + 1) = 4.467
√
80 = f (80) ≈ 18 (80 − 81) + 9 = 161
1
18 ≈ 8.944 11. V = πr2 h = 2πr2 and ∆V = 2π · 2r∆r = q4πr∆r.
(c) f (0) = 0 and f ′ (0)
= 1 so tangent line is Because V = 2πr2 = 47.3, we have r = 47.3 2π ≈
y = x and sin(0.3) = f (0.3) ≈ 0.3 2.7437 cm. We know |∆V | ≤ 0.1 so, using
∆V = 4πr∆r, we have 0.1 ≥ 4π (2.7437) |∆r | and
4. (a) f (1) = 0 and f ′ (1) = 1 so tangent line is y =
|∆r | ≤ 4π (2.7437
0.1
≈ 0.0029 cm. The required toler-
x − 1 and ln(1.3) = f (1.3) ≈ 1.3 − 1 = 0.3 )
ance is ±0.0029 cm. (Reality check: A coin 2 cm
(b) f (0) = 1 and f ′ (0) = 1 so tangent line is
high is quite unusual; 47.3 cm3 of gold weighs
y = 1 + x and e0.1 = f (0.1) ≈ 1.1
around 2 pounds!)
(c) f (1) = 1 and f ′ (1) = 5 so tangent line is √
13. V = x3 ⇒ ∆V ≈ 3x2 ∆x, V = 87 ⇒ x = 3 87 ≈
y − 1 = 5( x − 1) or y = 5x − 4 and (1.03)5 ≈ ∆V
4.431 cm so ∆x ≈ 3x 2
2 ≈ 3(4.431)2 ≈ 0.034 cm
5(0.03) − 4 = 1.15 q
15. With P = 2π Lg and g = 32 sec ft :
5. f ( x ) = (1 + x )n ⇒ f ′ ( x ) = n(1 + x )n−1 so 2

f (0) = 1 and f ′ (0) = n, hence tangent line is

q
2
(a) P = 2π 32 = π2 ≈ 1.57 sec
y − 1 = n( x − 0) or y = 1 + nx and (1 + x )n ≈ q
L
1 + nx (when x is close to 0) (b) 1 = 2π 32 ⇒ L = π82 ≈ 0.81 ft
2π
6. (a) f ( x ) = (1 − x )n ⇒ f ′ ( x ) = −n(1 − x )n−1 (c) dP = √ · √1 dL = 2π
16 (0.1) ≈ 0.039 sec
32 2 2
so f (0) = 1 and f ′ (0) = −n, hence tangent (d) 2 in = 16 ft = 21600
1 ft dP 2π 1
hr hr sec so dt = 4√2 · 2√24 ·
line is y − 1 = −n( x − 0) or y = 1 − nx and 1
≈ 5.25 × 10−6 > 0 (increasing)
(1 − x )n ≈ 1 − nx (when x is close to 0) 21600
17. (a) d f = f ′ (2) dx ≈ (0)(1) = 0
(b) f ( x ) = sin( x ) ⇒ f ′ ( x ) = cos( x ) so f (0) = 0
and f ′ (0) = 1, hence tangent line is y = x and (b) d f = f ′ (4) dx ≈ (0.3)(−1) = −0.3
sin( x ) ≈ x (for x close to 0) (c) d f = f ′ (3) dx ≈ (0.5)(2) = 1
(c) f ( x ) = e x ⇒ f ′ ( x ) = e x so f (0) = 1 and 19. (a) f ′ ( x ) = 2x − 3 ⇒ d f = (2x − 3) dx
f ′ (0) = 1, hence tangent line is y = x + 1 and (b) f ′ ( x ) = e x ⇒ d f = e x dx
e x ≈ x + 1 (for x near 0) (c) f ′ ( x ) = 5 cos(5x ) ⇒ d f = 5 cos(5x ) dx
7. (a) f ( x ) = ln(1 + x ) → f ′ ( x ) = 1+1 x so f (0) = 0 (d) f ′ ( x ) = 3x2 + 2 ⇒ d f = (3x2 + 2) dx so when
and f ′ (0) = 1, hence tangent line is y = x and x = 1 and dx = 0.2, d f = (3 · 12 + 2)(0.2) = 1
ln(1 + x ) ≈ x (e) f ′ ( x ) = 1x ⇒ d f = 1x dx so when x = e and
(b) f ( x ) = cos( x ) ⇒ f ′ ( x ) = − sin( x ) so f (0) = dx = −0.1, d f = 1e (−0.1) = − 10e1

1 and f ′ (0) = 0, hence tangent line is y = 1 (f) f ′ ( x ) = √ 1 ⇒ df = √ 1 dx so when

2x +5 2x +5
and cos( x ) ≈ 1 x = 22 and dx = 3, d f = √ (3) = 3
1
49 7
(c) f ( x ) = tan( x ) ⇒ f ′ ( x ) = sec2 ( x ), so f (0) =
0 and f ′ (0) = 1, hence tangent line is y = x
Section 2.9
and tan( x ) ≈ x
x
(d) f ( x ) = sin π2 + x ⇒ f ′ ( x ) = cos π2 + x so 1. (a) x2 + y2 = 100 ⇒ 2x + 2y · y′ = 0 ⇒ y′ = −



y
f (0) = 1 and f ′ (0) = 0, hence tangent line is 6 3
y = 1 and sin π2 + x ≈ 1.

so at (6, 8), y′ = − = −
8 4
p
2 ′ −x
9. (a) Area A( x ) = (base)(height) = x ( x2 + 1) = (b) y = 100 − x ⇒ y = √ so at
100 − x2
x3 + x ⇒ A′ ( x ) = 3x2 + 1 ⇒ A′ (2) = 13 so 6 6 3
∆A ≈ A′ (2) · ∆x = (13)(2.3 − 2) = 3.9 (6, 8), y′ = − √ =− =−
100 − 36 8 4
 A17

3. (a) x2 − 3xy + 7y = 5 ⇒ 2x − 3(y + xy′ ) + 7y′ = 2 2 2 −1 2 1

21. x3 + y3 = 5 ⇒ x 3 + y− 3 y′ = 0 ⇒ y′ =
3y − 2x 3−4 3 3
0 ⇒ y′ = ⇒ y′ (2,1) =

= −1 y1  1
7 − 3x 7−6 3 ′
 1 3 1
− ⇒ m = y (8,1) = − =−
5 − x2 ( 7 − 3x )(− 2x ) − (5 − x2 )(−3) x 8 2
(b) y = ⇒ y′ =
7 − 3x (7 − 3x )2 23. 2x + x · y′ + y + 2y · y′ + 3 − 7y′ = 0 ⇒ y′ =
′ ( 1 )(− 4 ) − ( 1 )(− 3) −2x − y − 3 −2(1) − (2) − 3 7
so at (2, 1), y = = −1 ⇒ y′ (1,2) =

(1)2 =
x + 2y − 7 (1) + 2(2) − 7 2
x2 y2 2x 2y ′ 25. Explicitly: y = Ax2 + Bx + C ⇒ y′ = 2Ax + B
5. (a) + =1⇒ + y =0⇒
9 16 9 16 Implicitly: x = Ay2 + By + C ⇒ 1 = 2Ayy′ + By′
16 x
y = − · ⇒ y′ (0,4) = 0
′

1
9 y ⇒ y′ =
r 2Ay + B
x2 4p
(b) y = 4 1 − = 9 − x2 ⇒ 27. Ax2 + Bxy + Cy2 + Dx + Ey + F = 0
9 3
4 −x ⇒ 2Ax + Bx · y′ + By + 2Cy · y′ + D + Ey′ = 0
y′ = · √ ⇒ y′ (0,4) = 0

−2Ax − By − D
3 9 − x2 ⇒ y′ =
1 Bx + 2Cy + E
7. (a) ln(y) + 3x − 7 = 0 ⇒ y′ + 3 = 0 ⇒ x
 y 28. x2 + y2 = r2 ⇒ 2x + 2y · y′ = 0 ⇒ y′ = − so
y′ = −3y ⇒ y′ (2,e) = −3e y
the slope of the tangent line is − yx . The slope of
(b) y = e7−3x ⇒ y′ = −3e7−3x ⇒ y
the line through the points (0, 0) and ( x, y) is x ,
y′ |(2,e) = −3e7−6 = −3e so the slopes of the lines are negative reciprocals
x
9. (a) x2 − y2 = 16 ⇒ 2x − 2yy′ = 0 ⇒ y′ = ⇒ of each other and the lines are perpendicular.
y
5 29. From the solution to problem 23, we know that
y′ (5,−3) = −

−2x −y−3
3 y′ = x+2y−7 so y′ = 0 when −2x − y − 3 = 0 ⇒
(b) The point (5, −3) is on the p bottom half of y = −2x − 3. Substituting y = −2x − 3 into the
the hyperbola so y = − x2 − 16 ⇒ y′ = original equation, we have:
x 5 5
⇒ y′ (5,−3) = − √

−√ =−
2
x − 16 25 − 16 3 x2 + x (−2x − 3) + (−2x − 3)2
1 + 3x − 7(−2x − 3) + 4 = 0
11. x = 4y − y2 ⇒ 1 = 4y′ − 2y · y′ ⇒ y′ = .
4 − 2y √
At (3, 1), y′ = 4−21(1) = 12 ; at (3, 3), y′ = 4−21(3) = ⇒ 3x2 + 26x + 34 = 0 ⇒ x =
−26± 262 −4(3)(34)
√ 2(3)
− 12 ; at (4, 2), y′ = 4−21(2) is undefined (the tan- = −13± 67
≈ −1.605 or −7.062.
3
gent line is vertical).
If x ≈ −1.605 (point A), then y = −2x − 3 ≈
13. x = y2 − 6y + 5 ⇒ 1 = 2y · y′ − 6y′ ⇒ y′ = −2(−1.605) − 3 = 0.21 so coordinates of A are
1 ′ 1 1
2y−6 . At (5, 0), y = 2(0)−6 = − 6 ; at (5, 6), (−1.605, 0.21); if x ≈ −7.062 (point C), then
y′ = 2(61)−6 = 16 ; at (−4, 3), y′ = 2(31)−6 is un- y = −2x − 3 ≈ −2(−7.062) − 3 = 11.124 so C
defined (vertical tangent line). is (−7.062, 11.124).
10x
15. 3y2 · y′ − 5y′ = 10x ⇒ y′ = ⇒ m = 31. From the solution to 29, point C is
3y2 − 5
10 5
(−7.062, 11.124). At D, y′ = −x2x − y −3
+2y−7 is unde-
y′ (1,3) =

= fined so x + 2y − 7 = 0 ⇒ x = 7 − 2y. Substitut-
22 11
17. y2 + sin(y) = 2x − 6 ⇒ 2y · y′ + cos(y) · y′ = 2 ⇒ ing x = 7 − 2y into the original equation:
2 2
y′ = ⇒ m = y′ (3,0) =

=2 (7 − 2y)2 + (7 − 2y)y + y2 + 3(7 − 2y) − 7y + 4 = 0
2y + cos(y) 0+1
√
19. ey + sin(y) = x2 − 3 ⇒ ey · y′ + cos(y) · y′ = so 3y2 − 34y + 74 = 0 ⇒ y =
34± 342 −4(3)(74)
=
2x √ 2(3)
2x ⇒ y′ = y ⇒ m = y′ (2,0) = 2

17± 67
e + cos(y) 3 ≈ 8.395 or 2.938.
 A18 contemporary calculus

If y ≈ 2.938 (point D), then x = 7 − 2y ≈ and solving for y′ yields:

7 − 2(2.938) = 1.124 so point D is (1.124, 2.938). p −x −x
Similarly, point B is (−9.79, 8.395). y′ = 25 − x2 · = √
25 − x2 25 − x2
33. (a) Using the Product Rule:
41. y = xcos( x) ⇒ ln(y) = cos( x ) · ln( x )
′ 2 7 3 3 2
y = ( x + 5) · 4( x − 1) · 3x ⇒ y′
= cos( x ) · 1
− ln( x ) · sin( x ) so:
y x
2 6 3 4
+ 7( x + 5) · 2x · ( x − 1)  
′ cos( x ) cos( x )
h i
2 6 3 3 3
= ( x + 5) ( x − 1) (2x ) 6x + 30x + 7x − 7 3 y = x − ln ( x ) · sin ( x )
x
h i
2 6 3 3 3
= ( x + 5) ( x − 1) (2x ) 13x + 30x − 7
43. ln(y) = 4 ln( x ) + 7 ln( x − 2) + ln(sin(3x ))
′
2 3
(b) ln(y) = 7 ln( x + 5) + 4 ln( x − 1) so: ⇒ yy = 4x + x−7 2 + 3sin cos(3x )
(3x )
so:
y′ 14x 12x2 
4 7

= 2 + 3 ′ 4 7
y = x ( x − 2) sin(3x ) + + 3 cot(3x )
y x +5 x −1 x x−2
and solving for y′ yields:
45. ln(y) = x · ln(3 + sin( x ))
12x2
 
14x ′
′
y =y 2 + 3 ⇒ yy = x · 3+ cos( x )
sin( x )
+ ln(3 + sin( x )) so:
x +5 x −1
12x2
   
2 7 3 4 14x ′ x x cos( x )
= ( x + 5) ( x − 1) + y = (3 + sin( x )) + ln(3 + sin( x ))
x2 + 5 x3 − 1 3 + sin( x )
which is the same as part (a). (Really! It is!)
47. f ′ (1) = 1(1.2) = 1.2; f ′ (2) = 9(1.8) = 16.2;
35. (a) y = x5 (3x + 2)4 so: f ′ (3) = 64(2.1) = 134.4
   
y′ = x5 D (3x + 2)4 + (3x + 2)4 D x5 49. f ′ (1) = 5(−1) = −5; f ′ (2) = 2(0) = 0;
f ′ (3) = 7(2) = 14
= x5 · 4(3x + 2)3 (3) + (3x + 2)4 · 5x4
D( f · g) f′ g′
(b) ln(y) = 5 ln( x ) + 4 ln(3x + 2) so: 51. ln( f · g) = ln( f ) + ln( g) ⇒ = +
h ′ i f ·g f g
f g′
y′ 5 12 ′
so D( f · g) = ( f · g) f + g = f · g + g · f ′
= +
y x 3x + 2  
f
 
f D g f′ g′
and solving for y′ yields: 52. ln = ln( f ) − ln( g) ⇒  f  = − so:
  g f g
5 12 g
y′ = y +
x 3x + 2     ′
g′ f′ f · g′

f f f
h i 5 12

D = − = − 2
= x5 (3x + 2)4 · + g g f g g g
x 3x + 2 ′ ′
f ·g −g· f
which is the same as in part (a). =
g2
37. (a) y = esin( x) ⇒ y′ = esin( x) · cos( x )
y′ 53. ln( f · g · h) = ln( f ) + ln( g) + ln(h)
(b) ln(y) = sin( x ) ⇒ = cos( x ) D( f · g · h) f′ g′ h′
y ⇒ = + + so:
f ·g·h f g h
⇒ y′ = y · cos( x ) = esin(x) · cos( x )
 ′
−x g′ h′

p f
39. (a) y = 25 − x2 ⇒ y′ = √ D( f · g · h) = ( f · g · h) + +
25 − x2 f g h
(b) ln(y) = 21 ln(25 − x2 ) so: = f ′ · g · h + f · g′ · h + f · g · h′
y′ 1 −2x −x
= · = 54. ln( a x ) = x ln( a) ⇒
D( a x )
= ln( a) ⇒ D( a x ) = a x ln( a)
y 2 25 − x2 25 − x2 ax
 A19

Section 3.1 which is a local maximum, and (4, −214), which

is a local minimum. When the interval is the en-
Local maximums at x = 3, x = 5, x = 9 and
tire real number line, this function does not have
x = 13; global maximum at x = 13. Local min-
a global maximum or global minimum.
imums at x = 1, x = 4.5, x = 7 and x = 10.5; 2 2
global minimum at x = 7. 15. f ( x ) = e−( x−2) ⇒ f ′ ( x ) = −2( x − 2)e−( x−2) ,
which is defined for all x; f ′ ( x ) = 0 only when
3. f ( x ) = x2 + 8x + 7 ⇒ f ′ ( x ) = 2x + 8, which is de-
1.
x = 2, so that is a critical number. There are no
fined for all values of x; f ′ ( x ) = 0 when x = −4,
endpoints. The only critical point is (2, 1), which
so x = −4 is a critical number. There are no end-
is a local and global maximum. When the interval
points. The only critical number is x = −4, so
is the entire real number line, this function does
the only critical point is (−4, f (−4)) = (−4, −9),
not have a local or global minimum.
which is the global (and local) minimum.
x 1 − x2
5. f ( x ) = sin( x ) ⇒ f ′ ( x ) = cos( x ), which is 17. f ( x ) = 2
⇒ f ′ (x) = , which is
1+x (1 + x 2 )2
defined for all values of x; f ′ ( x ) = 0 when defined for all x; f ′ ( x ) = 0 only when x = ±1, so
x = π2 + kπ so those values are critical numbers. these are the only critical numbers. There
  are no
There are no endpoints. f ( x ) = sin( x ) has local endpoints. The critical point −1, − 12is a local
and global maximums at x = π2 + 2kπ and global  
and global minimum; the critical point 1, 12 is
and local minimums at x = 3π 2 + 2kπ.
√ 1 2
a local and global maximum.
7. f ( x ) = 3 x = x 3 ⇒ f ′ ( x ) = 13 x − 3 = √ 1
3 2, 2 1
3 x 19. f ( x ) = ( x − 2) 3 ⇒ f ′ ( x ) = 23 ( x − 2)− 3 = √2 ,
which is defined for all values of x except x = 0, 3 3 x −2
which is defined for all values of x except x = 2,
and f ′ ( x ) ̸= 0 for all x, so x = 0 is the only criti-
and f ′ ( x ) ̸= 0 for all x so x = 2 is the only critical
cal number. There are no endpoints. the function
√ number. There are no endpoints. The function
f ( x ) = 3 x has no global or local extrema. 2
f ( x ) = ( x − 2) 3 has a global and local minimum
9. f ( x ) = xe5x ⇒ f ′ ( x ) = e5x [5x + 1], which is at x = 2 and no other extrema.
defined for all values of x; f ′ ( x ) = 0 when 1 2
21. f ( x ) = ( x2 − 4) 3 ⇒ f ′ ( x ) = 32 ( x2 − 4)− 3 · 2x =
5x + 1 = 0 ⇒ x = − 15 so this is the only critical 2x
√
3 2 2
, which is defined for all values of x
number. There are no endpoints. The function 3 ( x −4)
f ( x ) = xe5x has a local and global minimum at except x = ±2, and f ′ ( x ) = 0 only when x = 0,
x = − 51 and no other extrema. so x = −2, x = 0 and x = 2 are the only critical
numbers. There are no endpoints. The function
11. f ( x ) = ( x − 1)2 ( x − 3) ⇒ f ′ ( x ) = ( x − 1)2 + √
3
f ( x ) = x2 − 4 has a global and local minimum
2( x − 1)( x − 3) = ( x − 1)(3x − 7), which is de-
at x = 0 and no other global or local extrema.
fined for all values of x; f ′ ( x ) = 0 when x = 1
and x = 73 so those values are critical num-
bers. There are no endpoints. The only critical
points are (1, 0), which is a local maximum, and
7 32


3 , − 27 , which is a local minimum. When the in-
23.
terval is the entire real number line, this function
does not have any global extrema.
13. f ( x ) = 2x3 − 96x + 42 ⇒ f ′ ( x ) = 6x2 − 96,
which is defined for all values of x; f ′ ( x ) = 25. f ( x ) = x2 − 6x + 5 ⇒ f ′ ( x ) = 2x − 6, which is
6( x + 4)( x − 4) = 0 when x = −4 and x = 4 defined for all values of x; f ′ ( x ) = 0 only when
so those values are critical numbers. There are no x = 3, so that is a critical number. The endpoints
endpoints. The only critical points are (−4, 298), are x = −2 and x = 5. The critical points are
 A20 contemporary calculus

√
(3, −4), which is the local and global minimum, f (2) ≈ 6.49, so f ( x ) = 3 x2 + 4 − x has a local
(−2, 21), which is a local and global maximum, max at x = 0, a global and local min at x = √1 ,
2
and (5, 0), which is a local maximum. and a global and local max at x = 2.
27. f ( x ) = 2 − x3 ⇒ f ′ ( x ) = −3x2 , which is defined 37. f ( x ) = x3 − ln( x ) ⇒ f ′ ( x ) =h 3x2 i− 1x , which ex-
for all values of x; f ′ ( x ) = 0 only when x = 0, ists for all x in the interval 21 , 2 , and f ′ ( x ) =
so that is a critical number. The endpoints are 0⇒x= √1
. The endpoints are x = 1
and x = 2,
3 2
3
x = −2 and x = 1, which are also critical num-
so the three critical numbers are x = 21 , x = √
3
1
bers. The critical points are (−2, 10), which is a     3
1 1
local and global maximum, (0, 2), which is not a and x = 2: f 2 ≈ 0.82, f √ 3 ≈ 0.70 and
3
local or global maximum or minimum, and (1, 1), f (2) ≈ 7.3 so f ( x ) = x3 − ln( x ) has a local max
which is a local and global minimum. at x = 12 , a global and local min at x = √ 1
3 , and a
3
29. f ′ ( x ) = 3x2 − 3 = 3( x − 1)( x + 1), which is de- global and local max at x = 2.
fined for all x; f ′ ( x ) = 0 only when x = −1 √
39. A( x ) = 4x 1 − x2 for 0 < x < 1, so:
and x = 1, so these are critical numbers. The
− x2
 
endpoints x = −2 and x = 1 are also critical
p
′ 2
A (x) = 4 √ + 1−x
numbers. The critical points are (−2, 3), which is 1 − x2
1 − 2x2
 
a local and global minimum on [−2, 1], the point
=4 √
(−1, 7), which is a local and global maximum on 1 − x2
[−2, 1], and the point (1, 3), which is a local and
hence A′ ( x ) > 0 if 0 < x < √1 and A′ ( x ) < 0 if
global minimum on [−2, 1]. 2  
√1 < x < 1. Max when x = √1 : A √1 =2
31. f ′ ( x ) = 5x4 − 20x3 + 15x2 = 5x2 ( x2 − 4x + 3) = 2 2 2
5x2 ( x − 3)( x − 1), which is defined for all x; 41. V = x (8 − 2x )2 for 0 ≤ x ≤ 4, so:
f ′ ( x ) = 0 only when x = 0 or x = 1 in the
interval [0, 2], so each of these values is a crit- V ′ = x (2)(8 − 2x )(−2) + (8 − 2x )2
ical number. The endpoints x = 0 and x = 2 = (8 − 2x )(−4x + 8 − 2x )
are also critical numbers. The critical points are = (8 − 2x )(8 − 6x ) = 4(4 − x )(4 − 3x )
(0, 7), which is a local minimum, (1, 8) which is a
local and global maximum, and (2, −1), which is ′ 4 ′ 4
  V < 0 if x
>
hence
2
3 and V > 0 if 0 < x < 3 .
a local and global minimum. (It is also true that V 43 = 43 8 − 38 = 1024 27 ≈ 37.926 cubic units
f ′ (3) = 0, but x = 3 is not in the interval [0, 2].) is the largest volume. Smallest volume is 0, which
33. f ( x ) = 1
x 2 +1
⇒ f ′ ( x ) = (x− 2x
2 +1)2 , which is defined occurs when x = 0 and x = 4.
for all values of x; f ′ ( x ) = 0 only when x = 0, 43. (a) 4: the endpoints and two values of x for which
which is not in the interval [1, 3] so x = 0 is a f ′ ( x ) = 0 (b) 2: the endpoints (c) At most n + 1:
not a critical number. The endpoints x = 1 and the two endpoints and the n − 1 points x for
x = 3are critical numbers. The critical points are
 which f ′ ( x ) = 0; at least 2: the endpoints.
1, 1 , which is a local and global maximum, and
 2  45. (a) local minimum at (1, 5) (b) no extreme at (1, 5)
1
3, 10 , which is a local and global minimum. (c) local maximum at (1, 5) (d) no extreme at (1, 5)

− 1 47. (a) 0, 2, 6, 8, 11, 12 (b) 0, 6, 11 (c) 2, 8, 12
35. f ′ ( x ) = 32 x2 + 4 2 · 2x − 1 = √ 3x − 1, which
x 2 +4
′
exists for all x; f ( x ) = 0 only when 3x = 49. If f does not attain a maximum on [ a, b] or f does
√
x2 + 4 ⇒ x = √1 . The endpoints are x = 0 not attain a minimum on [ a, b], then f must have
2
and x = 2, so the critical numbers a discontinuity on [ a, b].
  are x = 0,
1
x= √ and x = 2: f (0) = 6, f √1 ≈ 5.66 and 51. (a) yes, −1 (b) no (c) yes, −1 (d) no (e) yes, 1 − π
2 2
 A21

53. (a) yes, 0 (b) yes, 0 (c) yes, 0 (d) yes, 0 (e) yes, 0 19. f ′ (c) = 17 at some time c does not prove the
55. (a) S( x ) is minimum when x ≈ 8 motorist “could not have been speeding.”

(b) S( x ) is maximum when x = 2. 21. f ( x ) = x3 + x2 + 5x + c, so f (1) = 7 + c = 10

when c = 3. Therefore f ( x ) = x3 + x2 + 5x + 3.

Section 3.2 23. (a) f ′ ( x ) = 2Ax. We need A · 12 + B = 9 and

2A · 1 = 4 so A = 2 and B = 7, hence
1. c ≈ 3, 10 and 13 f ( x ) = 2x2 + 7.
π 3π 5π 7π 9π
3. (a) c = (b) c = , , , (b) A · 22 + B = 3 and 2A · 2 = −2 so A = − 12
2 2 2 2 2
and B = 5, hence f ( x ) = − 12 x2 + 5.
5. Rolle’s Theorem asserts that the velocity h′ (t) will
equal 0 at some point between the time the ball is (c) A · 02 + B = 2 and 2A · 0 = 3, so there is no
tossed and the time you catch it. The ball is not such A. The point (0, 2) is not on the parabola
moving as fast when it reaches the balcony after y = x2 + 3x − 2.
being thrown from below. 25. f ( x ) = x3 + C, a family of “parallel” curves
7. The function does not violate Rolle’s Theorem 27. v(t) = 300: Assuming the rocket left the ground
because the function does not satisfy the hypothe- at t = 0, we have y(1) = 300 ft, y(2) = 600 ft,
ses of the theorem: f is not differentiable at 0, a y(5) = 1500 ft.
point in the interval −1 < x < 1. 29. f ′′ ( x ) = 6, f ′ (0) = 4, f (0) = −5 ⇒ f ( x ) =
9. No. Velocity is not the same as the rate of change 3x2 + 4x − 5
of altitude, which is only one of the components 31. (a) A( x ) = 3x (b) A′ ( x ) = 3
of position. Rolle’s Theorem only says there was
33. (a) A( x ) = x2 + x (b) A′ ( x ) = 2x + 1
a time when altitude was not changing.
35. a1 = 5, a2 = a1 + 3 = 5 + 3 = 8, a3 = a2 + 3 =
11. f ′ ( x ) = 3x2 + 5, so f ′ ( x ) = 0 has no real roots. If
(5 + 3) + 3 = 11, a4 = a3 + 3 = (5 + 3 + 3) + 3 =
f ( x ) = 0 for a value of x other than 2, then by
14: in general, an = 5 + 3(n − 1) = 2 + 3n
the corollary from Problem 10, we would have an
immediate contradiction.
Section 3.3
13. (a) f (0) = 0, f (2) = 4, f ′ (c) = 2c, so 42− 0
−0 = 2c
′
⇒ c = 1 (b) f (1) = 4, f (5) = 8, f (c) = 2c − 5,
so 85− 4
−1 = 2c − 5 ⇒ c = 3
14. (a) f (0) = 0, f π2 = 1, f ′ (c) = cos(c), so


1−0 2


π −0 = cos( c ) ⇒ c = arccos π ≈ 0.88
2

(b) f (−1) = −1, f (3) = 27, f ′ (c) = q3c2 , so

27+1 2 2 7 7
3+1 = 3c ⇒ c = 3 , hence c = (we
q3
7
know c > −1, which eliminates c = − 3 ) 1.
15. (a) f (1) = 4, f (9) = 2, f ′ (c) = −
√1 , so
2 c
2−4 −
√1
1−9 = 2 c
⇒ − 14 = −2 1
√
c
⇒ c = 4 (b) f (1) = 3,
15−3
f (7) = 15, f ′ (c)
= 2, so 7−1 = 2 and any c
between 1 and 7 will do.
17. The hypotheses are not all satisfied: f ′ ( x ) does
not exist at x = 0, which is between −1 and 3.
A22 contemporary calculus

3. 6.

7. A: Q, B: P, C: R
9. f ′ ( x ) = 1x > 0 for x > 0 so f ( x ) = ln( x ) is
increasing on (0, ∞).
11. If f is increasing then f (1) < f (π ) so f (1) and
f (π ) cannot both equal 2.
13. (a) x = 3, x = 8 (b) maximum at x = 8 (c) none
(or only at right endpoint)
4. 14. Relative maximum height at x = 2 and x = 7;
relative minimum height at x = 4.
15. Relative maximum height at x = 6; relative mini-
mum height at x = 8.
17. f ( x ) = x3 − 3x2 − 9x − 5 has a relative minimum
at (3, −32) and a relative maximum at (−1, 0).
19. h( x ) = x4 − 8x2 + 3 has a relative max at
(0, 3) and relative minimums at (2, −13) and
(−2, −13).
21. r (t) = 2(t2 + 1)−1 has a relative maximum at
(0, 2) and no relative minimums.
23. No positive roots. f ( x ) = 2x + cos( x ) is con-
tinuous and f (0) = 1 > 0. Because f ′ ( x ) =
2 − sin( x ) ≥ 1 > 0 for all x, f is increasing and
5.
never decreases back to the x-axis (a root).
24. One positive root. g( x ) = 2x − cos( x ) is contin-
uous, g(0) = −1 < 0 and g(1) = 2 − cos(1) ≥
1 > 0 so by the Intermediate Value Theorem,
g has a root between 0 and 1. Because g′ ( x ) =
2 + sin( x ) ≥ 1 > 0 for all x, g is always increasing
and can have only that one root.
 A23

25. h( x ) = x3 + 9x − 10 ⇒ h′ ( x ) = 3x2 + 9 = endpoints and points where the function is not

3( x2 + 3) > 0 for all x so h is always increas- differentiable from the intervals of concave up
ing and can cross the x-axis at most at one place. and concave down.)
Because the h(1) = 0, the graph of h crosses the 7. g′′ ( x ) = 6x − 6 ⇒ g′′ (−1) < 0 so (−1, 12) is a
x-axis at x = 1, which is the only root of h. local max; g′′ (3) > 0 so (3, −20) is a local min.
27. (a) See below left: f is always increasing but f ′ (1)
9. f ′′ ( x ) = 5 − sin5 ( x ) + 4 sin3 ( x ) · cos2 ( x ) ⇒
 
is undefined.
f ′′ π2



< 0 so π2 , 1 is a local maximum;
f ′′ 3π > 0 so 3π



2 2 , −1 is a local minimum;
f ′′ (π ) = 0 and f changes concavity at x = π
so (π, 0) is an inflection point.
11. d and e
13. (a) 0 (b) at most 1 (c) at most n − 2

(b) See above right: f ( x ) = x3 + 1 is always in- x f (x) f ′ (x) f ′′ ( x )

creasing but f ′ (0) = 0.
0 − + +
(c) If they travel at the same positive speed in 15. 1 + 0 −
different directions, then the distance between 2 − − +
them will not remain constant. 3 0 + +
29. (a) h( x ) = x2 or x2 + 1 or x2 − 7 or, in general,
x2 + C for any constant C.
(b) f ( x ) = x2 + C and 20 = f (3) = 32 + C ⇒
C = 20 − 9 = 11 so f ( x ) = x2 + 11. 17.
(c) g( x ) = x2 + C and 7 = g(2) = 22 + C ⇒ C =
7 − 4 = 3 so g( x ) = x2 + 3.

Section 3.4 19. f ′ ( x ) = 3x2 − 42x + 144 = 3( x2 − 14x + 48) =

3( x − 6)( x − 8), which exists everywhere and
1. (a) f (t) = number of workers unemployed at time
f ′ ( x ) = 0 ⇒ x = 6 or x = 8; meanwhile,
t; f ′ (t) > 0 and f ′′ (t) < 0
f ′′ ( x ) = 6x − 42 = 6( x − 7), so f ′′ (6) = −6 < 0
(b) f (t) = profit at time t; f ′ (t) < 0 and f ′′ (t) > 0
means that f has a local max at x = 6 and
(c) f (t) = population at time t; f ′ (t) > 0 and f ′′ (8) = 6 > 0 means that f has a local min
f ′′ (t) > 0 at x = 8. The only candidate for an inflection
point is where f ′′ ( x ) = 0 ⇒ x = 7; f ′′ ( x ) < 0
for x < 7 and f ′′ ( x ) > 0 for x > 7, so f does
3. have an inflection point at x = 7. A window with
2 ≤ x ≤ 12 and −50 ≤ y ≤ 10 should work.
21. f ′ ( x ) = 7e7x − 5 ⇒ f ′′ ( x ) = 49e7x so f ′′ ( x ) > 0
(d) Not possible. for all x, hence f is always concave up and has no
5. (a) Concave up on (0, 2), (2, 3+), (6, 9). (b) Con- inflection points. Its only critical number is where
7e7x − 5 = 0 ⇒ e7x = 75 ⇒ x = 17 · ln 57 ≈


cave down on (3+, 6). (A small technical note:
we have defined concavity only at points where −0.05, which must be a local min. A window
the function is differentiable, so we exclude the with −1 ≤ x ≤ 12 and 0 ≤ y ≤ 20 should work.
 A24 contemporary calculus

23. f ′ ( x ) = −3e−3x + 1 ⇒ f ′′ ( x ) = 9e−3x > 0 for (b) Using a rectangle with sides of length x and
all x, hence f is always concave up and has y (see figure above) 2x + 2y = P ⇒ y =
P
no inflection points. Its only critical number: 2 − x. We   to maximize A = x · y =
 want
−3e−3x + 1 = 0 ⇒ e−3x = 13 ⇒ x = 31 · ln (3) ≈ x P
−x = P
x − x2 and A′ ( x ) = P
− 2x
2 2 2
0.37, which must be a local min. Try a window so A′ ( x )
= 0 when x = P
⇒ y = P2 − = P4 ;
P
4 4
with −1 ≤ x ≤ 2 and 0 ≤ y ≤ 3. ′′
A ( x ) = −2 < 0 for all x, so x = P4 yields the
25. f ′ ( x ) = (1 − 3x )e−3x ⇒ f ′′ ( x ) = (9x − 6)e−3x so maximum enclosed area. This garden is a P4
the only critical number is x = 1
3 and f ′′ 1
3 − by P4 square.
3e−1 < 0, so it must be a local max. The only (c) Using a rectangle with sides of length x and y
inflection point is where x = 23 . Try a window (see figure below) 2x + y = P ⇒ y = P − 2x.
with −1 ≤ x ≤ 2 and −2 ≤ y ≤ 1. We want to maximize A = xy = x ( P − 2x ) =
27. f ′ ( x ) = 1 − 23
(4x − 1), so the critical numbers Px − 2x2 and A′ ( x ) = P 
− 4x ′
 so A ( x ) = 0
3x
P P
5
are x = 0 and x = 14 ; f ′′ ( x ) = 29 x − 3 (2x + 1) when x = 4 ⇒ y = P−2 4 = P2 .
so the only candidates for an inflection point
are x = 0 and x = − 12 . Because f ′′ ( x ) < 0
when x < − 12 and x > 0, the graph of f is con-
cave up on those intervals
 and concave down for
− 12 < x < 0; because f 14 > 0, f has a local min
at x = 41 . Try −1.5 ≤ x ≤ 2.5 and −1 ≤ y ≤ 3. (d) A circle. A semicircle.
29. Critical number at x = 0, which is a local min; in- 3. (a) If x is the horizontal length and y is the ver-
flection points at x = ±1; concave up on (−1, 1), tical length, then 120 = 2x + 5y ⇒ y =
concave down elsewhere; try a window with 24 − 25 x. We want to maximize A = xy =
−3 ≤ x ≤ 3 and −1 ≤ y ≤ 3. x 24 − 25 x = 24x − 25 x2 and A′ ( x ) = 24 − 45 x

31. Critical number at x = −1, which is a local min;

√ so A′ ( x ) = 0 when x = 30 (and y = 12);
inflection points at x = −1 ± 3; concave up be- A′′ ( x ) = − 45 < 0 for all x so x = 30
tween those values, concave down elsewhere; try yields the maximum enclosed area. Area is
a window with −6 ≤ x ≤ 6 and −1 ≤ y ≤ 4. (30 ft)(12 ft) = 360 square feet.
(b) A circular pen divided into four equal stalls
Section 3.5 by two diameters:
1. (a) Using a rectangle with sides of length x and
y (see figure below) 2x + 2y = 200 ⇒ y =
100 − x. We want to maximize A = x · y =
x (100 − x ) = 100x − x2 and A′ ( x ) = 100 − 2x
so A′ ( x ) = 0 ⇒ x = 50 ⇒ y = 100 − x = 50;
A′′ ( x ) = −2 < 0 for all x, so x = 50 yields
the maximum enclosed area. When x = 50,
A(50) = 50(100 − 50) = 2500 square feet.
does a better job than a square with area 400
square feet. If the radius is r, then 4r + 2πr =
120
120 ⇒ r = 4+ 2π ≈ 11.67; the resulting en-
closed area is A = πr2 ≈ π (11.67)2 ≈ 427.8
square feet.
 A25

The pen shown here: the least expensive cylindrical can decreases:
the least expensive can becomes narrower and
taller.
9. Recall that time = distance
rate . Let the run distance
= x (so 0 ≤ x ≤ 60; why?), hence the run time
= 8x and the swim distance = 402 + (60 − x )2
p

so the swim time = 12 402 + (60 − x )2 and the

p

total time is T ( x ) = 8x + 12 402 + (60 − x )2 . Dif-

p

ferentiating to compute T ′ ( x ) yields:

does even better. If each semicircle has radius
√
r, then the figure uses 4 2r + 4πr = 120 feet 1 1 1 2 − 1
2
of fencing so r = √120 ≈ 6.585 feet; the + · 40 + (60 − x )2 · 2(60 − x )(−1)
4 2+4π 8 2 2
resulting enclosed area is 1 60 − x
= − p
8 2 402 + (60 − x )2
A = (square) + (four semicircles)
so T ′ ( x ) = 0 when x = 60 ± √40 .
 
1 2 The value
= (2r )2 + 4 πr ≈ 445.90 ft2 15
2 x = 60 + √40 > 60 so the least total time occurs
15
5. The dimensions of the bottom of the box are when x = 60 − √40 ≈ 49.7 meters. In this situa-
15
10 − 2x and 5 − x so we want to maximize tion, the lifeguard should run about 56 of the way
V ( x ) = x (10 − 2x )(5 − x ) = 50x − 20x2 + 2x3 ; along the beach before going into the water.
V ′ ( x ) = 50 − 40x + 6x2 = 2(3x − 5)( x − 5) so
11. (a) Consider a similar problem with a new town
V ′ ( x ) = 0 when x = 5 or x = 53 . When x = 5,
D ∗ located at the “mirror image” of D across
then V = 0 (clearly not a maximum!) so x = 53 .
the river (below left):
The dimensions of the box with the largest vol-
ume are 53 inches by 10 20
3 inches by 3 inches.
7. (a) Let r and h be the radius and height of the
cylindrical can, so the volume is V = πr2 h =
100
100 ⇒ h = πr 2 . We want to minimize the cost:

C = 2(top area) + 5(bottom area) + 3(side area)

   
= 2 πr2 + 5 πr2 + 3 (2πrh)
  If the water works is built at any location
2 100 600 W along the river, then the distances are
= 7πr + 6πr = 7πr2 +
πr2 r the same from W to D and from W to D ∗ :
So C ′ (r ) = 14πr − 600 ⇒ C ′ (r ) = 0 when dist(W, D ) = dist(W, D ∗ ). Then dist(C, W ) +
r2
q
600
dist(W, D ) = dist(C, W ) + dist(W, D ∗ ) so the
r = 3 14π ≈ 2.39 ⇒ h ≈ π (100 ≈ 5.57.
2.39)2 shortest distance from C to D ∗ is a straight
(b) Let k = top material cost + bottom material line (see figure above right), and this straight
cost = 2¢ + bottom cost > 2¢ + 5¢ = 7¢. We line yields similar triangles with equal side
want to minimize the cost C = kπr2 + 600 r ; ratios: 3x = 105− x ⇒ x = 154 = 3.75 miles. A
C ′ (r ) = 2kπr − 600r2
so C ′ (r ) = 0 when r =
consequence of this “mirror image” view of
q
3 600
2kπ . If k = 8, then r ≈ 2.29; if k = 9, then the problem is that “at the best location W
r ≈ 2.20; if k = 10, then r ≈ 2.12. As the cost the angle of incidence α equals the angle of
of the bottom material increases, the radius of reflection β.”
 A26 contemporary calculus

(b) We want to minimize the cost: the triangle is a right triangle with sides 7 and
10. Using either approach, the maximum area
C ( x ) = 3000 dist(C, W ) + 7000 dist(W, D )
p q of the triangle is 12 (7)(10) = 35 square inches,
= 3000 x2 + 9 + 7000 (10 − x )2 + 25 and the third side is the hypotenuse with length
√ √
so, differentiating gives: 72 + 102 = 149 ≈ 12.2 inches.

3000x −7000(10 − x ) 17. (a) A( x ) = 2x (16 − x2 ) = 32x − 2x3 ⇒ A′ ( x ) =

′
C (x) = √ +p 32 − 6x2 so A′ (c) = 0 when x = 32 16
2
x +9 (10 − x )2 + 25 6 = 3 ≈
2.31. The dimensions of the rectangle are
and (after much algebra) it turns out that    2
C ′ ( x ) = 0 when x ≈ 7.82 miles. As it be- 2 16
3 ≈ 4.62 and 16 − 16
3 = 32
3 ≈ 10.67.
√
comes more expensive to build the pipe from (b) A( x ) = 2x 1 − x2 so, using the Product Rule:
W to D, the cheapest route tends to shorten
x2
p 
the distance from W to D. ′
A (x) = 2 1−x −2 √
1 − x2
13. (a) Let x be the length of one edge of the square
end. The volume is V ( x ) = x2 (108 − 4x ) = and A′ ( x ) = 0 when x = √1
≈ 0.707. The di-
2
  √
108x2 − 4x3 ⇒ V ′ ( x ) = 216x − 12x2 = mensions of the rectangle are 2 √1 = 2 ≈
6x (18 − x ) so V ′ ( x ) = 0 when x = 0 or x = 18. r  2
2

Clearly x = 0 results in a box of no volume, 1 1

1.414 and 1 − √ = √ ≈ 0.707.
2 2
so the dimensions of the acceptable box with
(c) The graph of | x | + |y| = 1 is a “diamond” (a
a square end and greatest volume are 18 in by
square) with corners at (1, 0), (0, 1), (−1, 0)
18 in by 36 in; the volume is 11,664 in3 .
and (0, −1). For 0 ≤ x ≤ 1:
(b) Let x be the length of the shorter edge of the
end. The volume V ( x ) = 2x2 (108 − 6x ) = A( x ) = 2x · 2(1 − x ) = 4x − 4x2
216x2 − 12x3 ⇒ V ′ ( x ) = 432x − 36x2 =
so A′ ( x ) = 4 − 8x and A′ ( x ) = 0 when x = 12 ;
36x (12 − x ) so V ′ ( x ) = 0 if x = 0 or x = 12.
A”( x ) = −8 < 0 for all x, so x = 12 must be
Clearly x = 0 results in a box of no volume, so
a local
 max. The dimensions  of the rectangle
the dimensions of the box with an acceptable  
1 1
shape and largest volume are 12 in by 24 in are 2 2 = 1 and 2 1 − 2 = 1.
by 36 in; the volume is 10,368 in3 . (d) A( x ) = 2x cos( x ), where 0 ≤ x ≤ π2 , so
(c) Let x be the radius of the circular end. A′ ( x ) = 2 cos( x ) − 2x · sin( x ) and A′ ( x ) = 0
The volume is V ( x ) = πx2 (108 − 2πx ) = when x ≈ 0.86. The dimensions of the rectan-
108πx2 − 2π 2 x3 ⇒ V ′ ( x ) = 216πx − 6π 2 x2 = gle are 2(0.86) = 1.72 and cos(0.86) ≈ 0.65.
6πx (36 − πx ) so V ′ ( x ) = 0 when x = 0 or 19. The cross-sectional area is:
x = 36π ≈ 11.46 in. The dimensions of the    
θ θ
acceptable box with circular end and largest A = 6 sin · 6 cos
2 2
volume are a radius of 11.46 in and a length
1
of 36 in; the volume is 14,851 in3 . = 36 · sin(θ ) = 18 sin(θ )
2
15. Without calculus: The area of the triangle is
1 1 which is a maximum when θ = π2 . Then the
2 (base)(height) = 2 (7)(height) and the height
maximum area is A = 18 sin π2 = 18 in2 . (This


is maximum when the angle between the sides
problem is similar to Problem 15.)
is a right angle. Using calculus: Let θ be the √
angle between the sides. Then the area of the tri- 21. V = 13 πr2 h and h = 9 − r2 so:
angle is A = 72 (10 sin(θ )) = 35 sin(θ ) ⇒ A′ (θ ) =
1 2p πp 4
35 cos(θ ) so A′ (θ ) = 0 when θ = π2 , hence V (r ) = πr 9 − r2 = 9r − r6
3 3
 A27

Differentiating and simplifying: 28. See below left for (a), below right for (b):

π 36r3 − 6r5
V ′ (r ) = ·√
6 9r4 − r6
so V ′ (r ) = 0 when 36r3 − 6r5 = 6r3 (6 − r2 ) = 0;
√
r = 0 results
q in√zero volume, so r = 6 ≈ 2.45 in
√
⇒ h = 9 − ( 6)2 = 3 ≈ 1.73 in.
23. Let n ≥ 10 be the number of passengers. The
income is I (n) = n(30 − (n − 10)) = 40n − n2 In the solutions to these “shortest path” problems,
and the cost is C (n) = 100 + 6n, so the profit is the roads all meet at 120◦ angles.
P(n) = I (n) − C (n) = (40n − n2 ) − (100 + 6n) =
29. (a) A = (base)(height) = (1 − x )( x2 ) = x2 − x3
34n − n2 − 100 ⇒ P′ (n) = 34 − 2n so P′ (n) = 0
for 0 ≤ x ≤ 1, so A′ ( x ) = 2x − 3x2 and
when n = 17: 17 passengers on the flight maxi-
A′ ( x ) = 0 ⇒ x = 23 (clearly the endpoints
mize your profit. (This is an example of treating
x = 0 and x = 1 will not yield the largest
a naturally discrete variable, the number of pas-
2
area). So A 23 = 31 23 = 27 4


sengers, as a continuous variable.) .
(b) A( x ) = (1 − x )(Cx2 ) = Cx2 − Cx3 where
25. Apply Problem 24 with R = f and E = g.
0 ≤ x ≤ 1, so A′ ( x ) = 2Cx − 3Cx2 and
27. (a) Let D be diameter of the can’s base and H = A′ ( x ) = 0 if x = 23 (as before, neither end-
the can’s height. Then:
2
point works). So A 32 = 13 · C 32 = 4C


  27 .
radius of can (c) A( x ) = ( B − x )(Cx2 ) = BCx2 − Cx3 , where
θ = arctan
height of CG 0 ≤ x ≤ B, so A′ ( x ) = 2BCx − 3Cx2 =
Cx (2B − 3x ) and A′ ( x ) = 0 if x = 23 B. So
   
D/2 D
= arctan = arctan  2
H/2 H    
B
A 2B C = 3 ( C ) 2B
3
4 3
= 27 B C.
For this can, D = 5 cm and H = 12 cm so 31. (a) Position the lower-left vertex of the triangle at
θ = arctan 2.5


6 ≈ 0.395 radians. The can can (0, 0) so the hypotenuse sits on y = 20 − 50 20
x
be tilted about 22.6◦ before it falls over. and the area of the rectangle is A = xy =
(b) Differentiating C ( x ) yields: x 20 − 25 x = 20x − 25 x2 ⇒ A′ ( x ) = 20 − 45 x

(60 + 19.2x )(19.2x ) − (360 + 9.6x2 )(19.2) and A′ ( x ) = 0 when x = 25. Then y = 10 and
(60 + 19.2x )2 the area of the rectangle is 25 · 10 = 250.
(b) Proceeding as in part (a), y = H− HBx
so C ′ ( x ) = 0 ⇒ (19.2)(9.6x2 + 60x − 360) = 0

H H 2
and A( x ) = x H − B x = Hx − B x ⇒
⇒ x = 3.75 or x = −10: the height of the cola
A′ ( x ) = H − 2H ′
B x and A ( x ) = 0 when
is 3.75 cm. B H
x = 2 . Then y = 2 and area of rectangle
(c) C (3.75) = 3.75 (The center of gravity is exactly
is B2 · H2 = BH
4 .
at the top edge of the cola. It turns out that
when the CG of a can-and-liquid system is as 33. Let r and h be radius and height of can, and F =
low as possible then the CG is at the top edge (top cost) + (bottom cost) + (cost of sides) =
2.5 A (πr )2 + B (πr )2 + C (2πrh). But volume V =


of the liquid.) Then θ = arctan 3.75 ≈ 0.588
V 2CV
radians: in this situation, the can can be tilted πr2 h ⇒ h = πr 2
2 so F (r ) = ( A + B ) πr + r ⇒
′ 2CV ′
F (r )2( A + B)πr − r2 , so F (r ) = 0 when r =
about 33.7◦ before it falls over. q
CV
(d) Less far. 3
π ( A+ B)
. Now you can find h and F.
 A28 contemporary calculus

1
Section 3.6 ln( x ) x 1
9. lim = lim = lim =0
x →∞ x p x →∞ px p−1 x →∞ px p
1. (a) h( x ) has a root at x = 1.
1 − cos(3x ) 3 sin(3x ) 9 cos(3x ) 9
(b) lim h( x ) = 0, lim h( x ) = 0 11. lim 2
= lim = lim =
x →1+ x →1− x →0 x x →0 2x x →0 2 2
lim h( x ) = −∞, lim h( x ) = +∞
x →3+ x →3− f′ m m
13. For a ̸= 0: ′
= x m−n → am−n . For a = 0:
(c) y = h( x ) has a vertical asymptote at x = 3 g n n
3. lim h( x ) = +∞, lim h( x ) = −∞ 
0 if m > n
x →2+ x →2− 

lim h( x ) = 0, lim h( x ) = 0 x m − am

 1 if m = n
x →4+ x →4− lim n =
x→a x − a n  +∞ if m < n and (m − n) is even

5. 0 7. −3 9. 0 11. DNE

DNE if m < n and (m − n) is odd


2
13. 3 15. 0 17. −7 19. 3 1 − cos( x ) sin( x )
15. lim = lim =0
x →0 x · cos( x ) x →0 − x · sin( x ) + cos( x )
21. cos(0) = 1 23. ln(1) = 0
f′ pe px p
17. ′ = → so p = 3(5) = 15.
25. (a) V (t) = 50 + 4t gals, A(t) = 0.8t lbs of salt g 3 3
(b) C (t) =
A(t)
=0.8t 19. (a) All three limits are +∞.
V (t) 50+4t
(b) After applying L’Hôpital’s Rule m times:
(c) as t → ∞, C (t) → 0.8
= 0.2 lbs per gal
4
(d) V (t) = 200 + 4t and A(t) = 0.8t, so C (t) = f (m) a · bm · ebn
0.8t 0.8 =
200+4t → 4 = 0.2 lbs per gal g(m) c(m)(m − 1)(m − 2) · · · (2)(1)
constant · ebn
27. +∞ 29. −∞ 31. −∞ 33. −∞ = → +∞
another constant
35. +∞ 37. −∞ 39. 1 41. −∞ ln( x ) − sin( x )
21. lim = lim · tan( x ) = −1 · 0 = 0
x →0+ csc( x ) x →0 + x
43. horizontal: y = 0; vertical: x = 0 1
ln( x ) x
√
44. horizontal: y = 0; vertical: x = 0; “hole”: (1, 1) 23. lim = lim = lim −2 x = 0
x →0+ x − 21 x →0+ − 12 x − 32 x →0+
45. horizontal: y = 0; vertical: x = 3 and x = 1  x  
3 x ·ln 1− 3
47. vertical: x = −1 25. Write 1− =e x2 and compute:
x2
49. horizontal: y = 1; vertical: x = 1  
3
51. slant: y = 2x + 1; vertical: x = 0 ln 1 − x2 −6
lim = lim =0
53. other: y = sin( x ); vertical: x = 2 x →∞ 1
x
x →∞ x − 3x
55. other: y = x2 so that the original limit equals e0 = 1.
57. other: y = cos( x ); vertical: x = 3 sin( x ) − x cos( x ) − 1
√ 27. lim = lim
59. y = x; x = −3 x →0 x · sin( x ) x →0 x · cos( x ) + sin( x )

Section 3.7 − sin( x )

= lim =0
x →0 − x · sin( x ) + 2 cos( x )
x3 − 1 3x2 3 3
1. lim 2
= lim = lim x = 1
x →1 x − 1 x →1 2x x →1 2 2  ln( x + 5
x )
x+5 x
ln(1 + 3x )
3
3 29. Write = e x and compute:
3. lim = lim 1+3x = x
x →0 5x x →0 5 5
ln x+ 5


x · ex x · ex + ex · 1 x+1 x −5
5. lim = lim = lim = −1 lim = lim =0
x →0 1 − e x x →0 −e x x →0 −1 x →∞ x x → ∞ x ( x + 5)
1
ln( x ) so that the original limit equals e0 = 1.
7. lim = lim x = 0
x →∞ x x →∞ 1
D
Derivative Facts

Basic Patterns

D(k) = 0 D(k · f ) = k · D( f ) k represents a constant

D( f + g) = D( f ) + D( g) D( f − g) = D( f ) − D( g)

g · D( f ) − f · D( g)
 
D( f · g) = f · D( g) + g · D( f ) f Product Rule and Quotient Rule
D =
g g2

Power Rules

D ( x p ) = p · x p −1 D ( f n ) = n · f n −1 · D ( f )

Chain Rule

D ( f ( g( x ))) = f ′ ( g( x )) · g′ ( x ) dy dy du
= ·
dx du dx

Exponential and Logarithmic Functions

D (eu ) = eu D ( au ) = au · ln( a)

1 1 f ′ (x)
D (ln(|u|)) = D (loga (|u|)) = D (ln( f ( x ))) =
u u · ln( a) f (x)

Trigonometric Functions

D (sin(u))) = cos(u)) D(tan(u)) = sec2 (u) D(sec(u)) = sec(u) · tan(u)

D(cos(u)) = − sin(u) D(cot(u)) = − csc2 (u) D(csc(u)) = − csc(u) · cot(u)

A30 contemporary calculus

Inverse Trigonometric Functions

1 1 1
D(arcsin(u)) = √ D(arctan(u)) = D(arcsec(u)) = √
1 − u2 1 + u2 | u | u2 − 1

−1 −1 −1
D(arccos(u)) = √ D(arccot(u)) = D(arccsc(u)) = √
1 − u2 1 + u2 | u | u2 − 1

Hyperbolic Functions

D (sinh(u))) = cosh(u)) D(cosh(u)) = sinh(u)

D(tanh(u)) = sech2 (u) D(coth(u)) = − csch2 (u)

D(sech(u)) = − sech(u) · tanh(u) D(csch(u)) = − csch(u) · coth(u)

Inverse Hyperbolic Functions

1 1
D(argsinh(u)) = √ D(argcosh(u)) = √ (for u > 1)
1 + u2 u2 −1
1 1
D(argtanh(u)) = (for |u| < 1) D(argcoth(u)) = (for |u| > 1)
1 − u2 1 − u2
−1 −1
D(argsech(u)) = √ (for 0 < u < 1) D(argcsch(u)) = √ (for u ̸= 0)
| u | 1 − u2 | u | u2 + 1
H
How to Succeed in Calculus

The following comments are based on over 30 years of watching stu-

dents succeed and fail in calculus courses at universities, colleges and
community colleges and of listening to their comments as they went
through their study of calculus. This is the best advice we can give to
help you succeed.
Calculus takes time. Almost no one fails calculus because they lack
sufficient “mental horsepower.” Most people who do not succeed are
unwilling (or unable) to devote the necessary time to the course. The
“necessary time” depends on how smart you are, what grade you want
to earn and on how competitive the calculus course is. Most calculus
teachers and successful calculus students agree that two (or three) hours
every weeknight and six or seven hours each weekend is a good way to
begin if you seriously expect to earn an A or B grade. If you are only
willing to devote five or 10 hours a week to calculus outside of class,
you should consider postponing your study of calculus.
Do NOT fall behind. The brisk pace of the calculus course is based
on the idea that “if you are in calculus, then you are relatively smart,
you have succeeded in previous mathematics courses, and you are
willing to work hard to do well.” It is terribly difficult to catch up and
keep up at the same time. A much safer approach is to work very
hard for the first month and then evaluate your situation. If you do fall
behind, spend a part of your study time catching up, but spend most
of it trying to follow and understand what is going on in class.
Go to class, every single class. We hope your calculus teacher makes
every idea crystal clear, makes every technique seem obvious and easy,
is enthusiastic about calculus, cares about you as a person, and even
makes you laugh sometimes. If not, you still need to attend class.
You need to hear the vocabulary of calculus spoken and to see how
mathematical ideas are strung together to reach conclusions. You need
to see how an expert problem-solver approaches problems. You need
to hear the announcements about homework and tests. And you need
to get to know some of the other students in the class. Unfortunately,
A32 contemporary calculus

when students get a bit behind or confused, they are most likely to
miss a class or two (or five). That is absolutely the worst time to miss
classes. Attend class anyway. Ask where on campus you can get some
free tutoring or counseling. Ask a classmate to help you for an hour
after class. If you must miss a class, ask a classmate what material was
covered and skim those sections before the next class. Even if you did
not read the material, return to class as soon as possible.
Work together. Study with a friend. Work in small groups. It is
much more fun and is very effective for doing well in calculus. Recent
studies — and our personal observations — indicate that students who
regularly work together in small groups are less likely to drop the
course and are more likely to get A’s or B’s. You need lots of time to
work on the material alone, but study groups of 3–5 students, working
together two or three times a week for a couple hours, seem to help
everyone in the group. Study groups offer you a way to get and give
help on the material and they can provide an occasional psychological
boost (“misery loves company?”); they are a place to use the mathemat-
ical language of the course, to trade mathematical tips, and to “cram”
for the next day’s test. Students in study groups are less likely to miss
important points in the course, and they get to know some very nice
people — their classmates.
Use the textbook effectively. There are a number of ways to use a
mathematics textbook:

• to gain an overview of the concepts and techniques,

• to gain an understanding of the material,

• to master the techniques, and

• to review the material and see how it connects with the rest of the
course.

The first time you read a section, just try to see what problems are
being discussed. Skip around, look at the pictures, and read some
of the problems and the definitions. If something looks complicated,
skip it. If an example looks interesting, read it and try to follow the
explanation. This is an exploratory phase. Don’t highlight or underline
at this stage — you don’t know what is important yet and what is just a
minor detail.
The next time through the section, proceed in a more organized
fashion, reading each introduction, example, explanation, theorem and
proof. This is the beginning of the “mastery” stage. If you don’t
understand the explanation of an example, put a question mark (in
pencil) in the margin and go on. Read and try to understand each step
in the proofs and ask yourself why that step is valid. If you don’t see
 A33

what justified moving from one step to another in the proof, pencil in
question marks in the margin. This second phase will go more slowly
than the first, but if you don’t understand some details, just keep going.
Don’t get bogged down yet.
Now worry about the details. Go quickly over the parts you already
understand, but slow down and try to figure out the parts marked with
question marks. Try to solve the example problems before you refer to
the explanations. If you now understand parts that were giving you
trouble, cross out the question marks. If you still don’t understand
something, put in another question mark and write down your question
to ask a teacher, tutor or classmate.
Finally, it is time to try the problems at the end of the section. Many
of them are similar to examples in the section, but now you need to
solve them. Some of the problems are more complicated than the
examples, but they still require the same basic techniques. Some of
the problems will require that you use concepts and facts from earlier
in the course, a combination of old and new concepts and techniques.
Working lots of problems is the “secret” of success in calculus.
Work the Problems. Many students read a problem, work it out and
check the answer in the back of the book. If their answer is correct, they
go on to the next problem. If their answer is wrong, they manipulate
(finagle, fudge, massage) their work until their new answer is correct,
and then they go on to the next problem. Do not try the next problem
yet! Before going on, spend a short time, just half a minute, thinking
about what you have just done in solving the problem. Ask yourself:
“What was the point of this problem?” “What big steps did I need to
take to solve this problem?” “What was the process?” Do not simply
review every single step of the solution process. Instead, look at the
outline of the solution, the process, the “big picture.” If your first
answer was wrong, ask yourself, “What about this problem should
have suggested the right process the first time?” As much learning and
retention can take place in the 30 seconds you spend reviewing the
process as took place in the 10 minutes you took to solve the problem.
A correct answer is important, but a correct process — carefully used —
will get you many correct answers.
There is one more step that too many students omit. Go back and
quickly look over the section one more time. Don’t worry about
the details, just try to understand the overall logic and layout of the
section. Ask yourself, “What was I expected to learn in this section?”
Typically, this last step — a review and overview — goes quickly, but it
is very valuable. It can help you see and retain the important ideas and
connections.
T
Trigonometry Facts

Right Angle Trigonometry

opp adj opp

sin(θ ) = cos(θ ) = tan(θ ) =
hyp hyp adj

hyp hyp adj

csc(θ ) = sec(θ ) = cot(θ ) =
opp adj opp

Trigonometric Functions

y x y
sin(θ ) = cos(θ ) = tan(θ ) =
r r x

x r r
cot(θ ) = sec(θ ) = csc(θ ) =
y x y

Fundamental Identities

1 sin(θ ) 1
sec(θ ) = tan(θ ) = =
cos(θ ) cos(θ ) cot(θ )

1 cos(θ ) 1
csc(θ ) = cot(θ ) = =
sin(θ ) sin(θ ) tan(θ )

sin2 (θ ) + cos2 (θ ) = 1 ⇒ 1 + tan2 (θ ) = sec2 (θ ) ⇒ cot2 (θ ) + 1 = csc2 (θ )

sin(−θ ) = − sin(θ ) cos(−θ ) = cos(θ ) tan(−θ ) = − tan(θ )

π  π  π 
sin − θ = cos(θ ) cos − θ = sin(θ ) tan − θ = cot(θ )
2 2 2

sin( A) sin( B) sin(C )

Law of Sines: = =
a b c

Law of Cosines: c2 = a2 + b2 − 2ab · cos(C )

A36 contemporary calculus

Angle Addition and Subtraction Formulas

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

sin( x − y) = sin( x ) · cos(y) − cos( x ) · sin(y)

cos( x + y) = cos( x ) · cos(y) − sin( x ) · sin(y)

cos( x − y) = cos( x ) · cos(y) + sin( x ) · sin(y)

tan( x ) + tan(y) tan( x ) + tan(y)

tan( x + y) = tan( x − y) =
1 − tan( x ) · tan(y) 1 + tan( x ) · tan(y)

Product-to-Sum Formulas

1
sin( x ) · sin(y) = 2 cos( x − y) − 12 cos( x + y)

1
cos( x ) · cos(y) = 2 cos( x − y) + 21 cos( x + y)

1
sin( x ) · cos(y) = 2 sin( x + y) + 12 sin( x − y)

Sum-to-Product Formulas

x−y
   
x+y
sin( x ) + sin(y) = 2 sin · cos
2 2
x−y
   
x+y
cos( x ) + cos(y) = 2 cos · cos
2 2
sin( x + y)
tan( x ) + tan(y) =
cos( x ) · cos(y)

Double-Angle Formulas

sin(2x ) = 2 sin( x ) · cos( x )

cos(2x ) = cos2 ( x ) − sin2 ( x ) = 2 cos2 ( x ) − 1 = 1 − 2 sin2 ( x )

2 tan( x )
tan(2x ) =
1 − tan2 ( x )
Half-Angle Formulas
x r
1
sin =± (1 − cos( x ))
2 2
x
The quadrant of determines the ±. x r
2 1
cos =± (1 + cos( x ))
2 2
x 1 − cos( x )
tan =
2 sin( x )