Mate Matic A

Math 221 1st Semester Calculus
Lecture Notes for Fall 2006.

Prof J. Robbin
December 21, 2006
All references to Thomas or the textbook in these notes refer to
Thomas CALCULUS 11th edition
published by Pearson Addison Wesley in 2005.
These notes may be downloaded from
http://www.math.wisc.edu/~robbin/221dir/lecs-221.pdf.
Some portions of these notes are adapted from
http://www.math.wisc.edu/~angenent/Free-Lecture-Notes/.
Some problems come from a list compiled by Arun Ram, others come from the
WES program, and others come from the aforementioned Thomas text or from
Stewart, Calculus (Early Transcendentals), 3rd Edition.
1
Contents
I What you need to know to take Calculus 221 5
1 Algebra 5
2 Coordinate Geometry 7
3 Functions 9
4 Trigonometry 12
5 Additional Exercises 15
II Limits 17
6 Tangent and Velocity 17
7 Limits 19
8 Two Limits in Trigonometry 26
9 Continuity 27
III Dierentiation 31
10 Derivatives Dened 31
11 Higher Derivatives and Dierential Notation 37
12 Implicit Functions 41
13 The Chain Rule 43
14 Inverse Functions 47
15 Dierentiating Trig Functions 50
16 Exponentials and Logarithms 53
17 Parametric Equations 59
18 Approximation* 63
19 Additional Exercises 67
2
IV Applications of Derivatives 68
20 The Derivative as A Rate of Change 68
21 Related Rates 71
22 Some Theorems about Derivatives 74
23 Curve Plotting 81
24 Max Min Word Problems 84
25 Exponential Growth 87
26 Indeterminate Forms (lHopitals Rule) 91
27 Antiderivatives 93
28 Additional Problems 95
V Integration 96
29 The Denite Integral 96
30 The Fundamental Theorem of Calculus 104
31 Averages 107
32 Change of Variables 108
VI Applications of Denite Integrals 112
33 Plane Area 112
34 Volumes 114
35 Arc Length 118
36 Surface Area 122
37 Center of Mass 125
VII Loose Ends 132
38 The Natural Log Again* 132
3
39 Taylor Approximation* 133
40 Newtons Method* 135
VIII More Problems 136
IX Notes for TAs 153
4
Chapter I
What you need to know to take
Calculus 221
In this chapter we will review material from high school mathematics.
We also teach some of this material at UW in Math 112, Math 113,
and Math 114, and some of it will be reviewed again as we need
it. This material should be familiar to you. If it is not, you may
not be ready for calculus. Pay special attention to the denitions.
Important terms are shown in boldface when they are rst dened.
1 Algebra
This section contains some things which should be easy for you. (If
they are not, you may not be ready for calculus.)
1.1. Answer these questions.
1. Factor x
2
6x + 8.
2. Find the values of x which satisfy x
2
7x + 9 = 0. (Quadratic formula.)
3. x
2
y
2
=? Does x
2
+y
2
factor?
4. True or False:

x
2
+ 4 = x + 2?
5. True or False: (9x)
1/2
= 3
x?
6. True or False:
x
2
x
8
x
3
= x
2+83
= x
7
?
7. Find x if 3 = log
2
(x).
8. What is log
7
(7
x
)?
9. True or False: log(x +y) = log(x) + log(y)?
10. True or False: sin(x +y) = sin(x) + sin(y)?
1.2. There are conventions about the order of operations. For example,
ab +c means (ab) +c and not a(b +c),
a
b
c
means a/(b/c) and not (a/b)/c,
a
b
c
means (a/b)/c and not a/(b/c),
log a +b means (log a) +b and not log(a +b).
If necessary, we use parentheses to indicate the order of doing the operations.
5
1.3. There is analogy between the laws of addition and the laws of multipli-
cation:
a +b = b +a ab = ba
(a +b) +c = a + (b +c) (ab)c = a(bc)
a + 0 = a a 1 = a
a + (a) = 0 a a
1
= 1
a b = a + (b) a/b = a b
1
a b = (a +c) (b +c) a/b = (ac)/(bc)
(a b) + (c d) = (a +c) (b +d) (a/b) (c/d) = (ac)/(bd)
(a b) (c d) = (a +d) (b +c) (a/b)/(c/d) = (ad)/(bc)
The last line explains why we invert and multiply to divide fractions. The only
other law of arithmetic is the distributive law
(a +b)c = ac +bc, c(a +b) = ca +cb.
Note that
(a +b)/c = (a/c) + (b/c), but
1
c/(a +b) ,= (c/a) + (c/b).
Exercises
Exercise 1.4. Answer the questions in 1.1
Exercise 1.5. Here is a list of some algebraic expressions that have been sim-
plied. Some steps in the simplication processes are correct and some of them
are WRONG! For each problem:
1. Determine if the simplied result is correct.
2. Determine if there are any mistakes made in the simplication process.
(NOTE: just because the result is correct does not mean there are no
mistakes).
3. If there are mistakes, redo the problem correctly. If there are no mistakes,
redo the problem with another correct method.
(i)
x
2
1
x + 1
=
x
2
+ (1)
x + 1
=
x
2
x
+
1
1
= x 1
(ii) (x +y)
2
(x y)
2
= x
2
+y
2
x
2
y
2
= 0
1
usually
6
(iii)
9(x 4)
2
3x 12
=
3
2
(x 4)
2
3x 12
=
(3x 12)
2
3x 12
= 3x 12
(iv)
x
2
y
5
2x
3
= x
2
y
5
2x
3
= 2x
6
y
5
(v)
(2x
3
+ 7x
2
+ 6) (2x
3
3x
2
17x + 3)
(x + 8) + (x 8)
=
4x
2
17x + 9
2x
= 2x17+
9
2x
(vi)
x
1
+y
1
x
1
y
1
=
(x +y)
1
(x y)
1
=
_
x +y
x y
_
1
=
x +y
x y
=
x +y
y x
.
2 Coordinate Geometry
The material in this section is crucial for understanding calculus. It
is reviewed in Thomas pages 9-18, but if you have not seen it before
you may not be ready for calculus.
2.1. Points and Slope. The notation P(x, y) is used as an abbreviation for
the more cumbersome phrase the point P whose coordinates are (x, y). The
slope of the line through the distinct points P
0
(x
0
, y
0
) and P
1
(x
1
, y
1
) is
m =
y
1
y
0
x
1
x
0
.
- x
6
y
P
0
P
1
x
1
x
0
y
1
y
0
2.2. Point-Slope Equation of a Line. If we use a dierent pair of points on
the line to compute the slope, we get the same answer. Hence, a point P(x, y)
lies on the line P
0
P
1
if and only if we get the same answer for the slope when
we use (x, y) in place of (x
1
, y
1
):
y y
0
x x
0
=
y
1
y
0
x
1
x
0
= m.
This equation has one minor aw; it doesnt work when (x, y) = (x
0
, y
0
) (never
divide by zero). To remedy this multiply by (x x
0
) and add y
0
to both sides:
7
y = y
0
+m(x x
0
).
This is the equation for the line through P
0
(x
0
, y
0
) with slope m; this form makes
it obvious that the point P
0
(x
0
, y
0
) lies on the line. For example, the equation
for the line through P
0
(2, 3) and P
1
(4, 11) is
y 3
x 2
=
11 3
4 2
= 4, or y = 3 + 4(x 2),
See Thomas page 11.
2.3. If a line has slope m, then any line perpendicular to it has slope 1/m.
(Proof: The slope of a line is the tangent of the angle it makes with the x-axis
and tan( +

2
) = 1/ tan().
2.4. Sometimes we use letters other than x and y. For this reason, always
label the axes when you draw a graph. For example the line 2u + 3v = 1 can
be written as v =
2
3
u +
1
3
and as u =
3
2
v +
1
2
. If the u-axis is horizontal, the
slope is
2
3
. If the v-axis is horizontal, the slope is
3
2
. The slope is always the
rise over the run, i.e. to nd the slope, take two points on the line, subtract
the vertical coordinates (this is the rise), subtract the horizontal coordinates
(this is the run), and divide the rise by the run.
2.5. Distance and Circles. By the Pythagorean theorem the distance from
P
0
(x
0
, y
0
) to P
1
(x
1
, y
1
) is
[P
0
P
1
[ =
_
(x
1
x
0
)
2
+ (y
1
y
0
)
2
The distance from a point P(x, y) to the point P
0
(x
0
, y
0
) is exactly a when
[P
0
P[
2
= a
2
, i.e.
(x x
0
)
2
+ (y y
0
)
2
= a
2
.
This is the equation for the circle centered at P
0
with radius a. For example,
the equation of the circle of radius 5 centered at P
0
(2, 3) is
(x 2)
2
+ (y 3)
2
= 5
2
, or x
2
+y
2
4x 6y 12 = 0.
The latter equation has the form of a general 2nd degree equation
Ax
2
+Bxy +Cy
2
+Dx +Ey +F = 0
(with A = C = 1, B = 0, D = 4, E = 6, F = 12). Ellipses (including
circles), parabolas, and hyperbolas all have equations of this form.
8
Exercises
Exercise 2.6. Let L
1
and L
2
be two lines. An equation for L
1
is 6x + 4y = 7.
The line L
2
is perpendicular to L
1
and goes through the point P
0
(1, 0). Find
an equation for the line L
2
and the point where the two lines intersect.
Exercise 2.7. Repeat the previous exercise replacing 6x+4y = 7 by ax+by = c
and P
0
(1, 0) by P
0
(x
0
, y
0
).
Exercise 2.8. Find the center P
0
(x
0
, y
0
) and radius a of the circle x
2
+ y
2
2x + 8y 20 = 0.
Exercise 2.9. Find the center P
0
(x
0
, y
0
) and radius a of the circle x
2
+ y
2
+
Ax +By +C = 0.
Exercise 2.10. Alice and Bob each graph the same line but Alice chooses one
of the variables to label the horizontal axes and Bob the other. What is the
relation between their values for the slope?
3 Functions
This material will be reviewed as we need it, but you should have
seen it before. It is explained in Thomas pages 19-27 .
Denition 3.1. A function is a rule which produces an output f(x) from
an input x. The set of inputs x for which the function is dened is called the
domain and f(x) (pronounced f of x) is the value of f at x. The set of
all possible outputs f(x) as x runs over the domain is called the range of the
function.
3.2. If a function f(x) is given by an expression in the variable x and the
domain is not explicitly specied the domain is understood to be the set of all
x for which the expression is meaningful. For example, for the function f(x) =
1/x
2
the domain is the set of all nonzero real numbers x (the value f(0) is not
dened because we dont divide by zero) and the range is the set of all positive
real numbers (the square of any nonzero number is positive). The domain and
range of the square root function

x is the set of all nonnegative numbers x.
The domain of the function y =

1 x
2
is the interval [1, 1] and the range is
the interval [0, 1], i.e.

1 x
2
is meaningful only if 1 x 1 (otherwise the
input to the square root function is negative) and 0
1 x
2
1.
Denition 3.3. The graph of a function
y = f(x)
is the set of all points P(x, y) whose coordinates (x, y) satsfy the equation y =
f(x). More generally, an equation of form
F(x, y) = 0
9
determines a set (graph) in the (x, y)-plane consisting of all points P(x, y) whose
coordinates (x, y) satisfy the equation. The graph of a function y = f(x) is a
special case: take F(x, y) = yf(x). To decide if a set is the graph of a function
we apply the
Vertical Line Test. A set in the (x, y)-plane is the graph of a function
if and only if every vertical line x = constant intersects the graph in
at most one point. [If the number a is in the domain, the vertical line
x = a intersects the graph y = f(x) in the point P(a, f(a)).]
3.4. For example, the graph of the equation x
2
+ y
2
= 1 is a circle; it is not
the graph of a function since the vertical line x = 0 (the y-axis) intersects the
graph in two points P
1
(0, 1) and P
2
(0, 1). This graph is however the union of
two dierent graphs each of which is the graph of a function:
2
x
2
+y
2
= 1 y =
_
1 x
2
or y =
_
1 x
2
.
3.5. Usually a function is dened by giving a formula as in f(x) =

1 x
2
.
Sometimes several formulas are used as in the denition
[x[ =
_
x if x 0
x if x < 0.
of the absolute value function.
Remark 3.6. In calculus, we learn to reason about a function even when we
cannot nd an explicit formula for it. For example, in theory the equation
x = y
5
+y (which has the form x = g(y)) can be rewritten in the form y = f(x)
but there is no formula for f(x) involving the mathematical operations studied
in high school. (Your TA will learn this in Math 742.)
Denition 3.7. When two functions are related by the condition
3
y = f(x) x = g(y)
(for all appropriate x and y) we say that the functions are inverse to one
another and write g = f
1
. The range of f is the domain of g and the range of
g is the domain of f. Often a function f has an inverse function g only after we
modify f by articially restricting its domain. The following table lists some
2
See Thomas Figure 1.28 page 4.
3
The notation is an abbreviation for if and only if.
10
common functions (suitably restricted) and their inverses.
y = f(x) x = f
1
(y)
y = y
0
+m(x x
0
) x = x
0
+ (y y
0
)/m
y = x
2
(x 0) x =

y (y 0)
y = a
x
x = log
a
(y) (y > 0)
y = sin (/2 /2) = sin
1
(y) (1 y 1)
x = cos (0 ) = cos
1
(x) (1 x 1)
u = tan (/2 < < /2) = tan
1
(u)
(The notations arcsiny = sin
1
y, arccos x = cos
1
x, and arctanu = tan
1
u
are also commonly used for the inverse trigonometric functions. The exponential
function y = a
x
is only dened for a positive.) To decide if a function has an
inverse we apply the
Horizontal Line Test. A function has an inverse if and only if every
horizontal line y = constant intersects the graph in at most one point.
[The horizontal line y = b intersects the graph y = f(x) in the point
P(f
1
(b), b).]
Remark 3.8. To nd an expression for f
1
(y) solve the equation y = f(x) for
x in terms of y. (It is not always possible to nd a nice formula.)
Remark 3.9. Two kinds of notation are used in calculus which I call func-
tional notation and variable notation. In functional notation the function
has a name, usually f or g, and is dened by an equation of form
f(x) = some formula in x.
Changing the variable in this notation does not change the function, e.g. the
function g dened by the formula
g(t) = t
2
is the same as the function g dened by the formula
g(x) = x
2
.
(Both formulas give g(3) = 9.) In variable notation we write
y = x
2
11
and say that y is a function of x. I like to write expressions like
y
x=3
= 9
to indicate that the value of y is 9 when the value of x is 3. The advantage of
functional notation is that it is makes clear where a function is being evaluated.
The disadvantage is that the notation tends to be cumbersome since we have to
introduce a name for each function we consider. I like functional notation for
proofs and variable notation for problems.
Exercises
Exercise 3.10. Let f(x) =
x 2
x + 4
.
(i) Find the domain and range of f.
(ii) Find the domain and range of f
1
.
(iii) Find a formula for f
1
(y).
Exercise 3.11. Repeat the previous exercise with f(x) =
ax +b
cx +d
where a, b, c, d
are constants with ad bc ,= 0.
Exercise 3.12. What is

x
2
? (
x)
2
?
Exercise 3.13. Draw the graph y = f(x) where f(x) is the function dened
by
f(x) =
_
3x for 0 x < 1;
4x 1 for 1 x.
Give a formula (like the formula for f(x)) for the inverse function x = f
1
(y).
4 Trigonometry
There is a review of trigonometry in Thomas pages 48-58 but you
should already be familiar with this material. You should also have
seen the inverse trigonometric functions before. These are reviewed
in Thomas section 7.7 (page 517) and well review it again later in
the course.
4.1. In calculus we always measure angles in radians rather than degrees. The
radian measure of an angle is the arclength of a circle of radius one (centered
at the vertex of the angle) cut out by the angle. Since the total length of the
circumference of a circle is 2 we get
2 radians = 360
o
.
12
(Since 2 = 6.283 . . . is about 6, this means that one radian is a little less than
60 degrees.) Let O(0, 0) denote the origin, P(x, y) be a point of the (x, y) plane
distinct from the origin, r = [OP[ be the distance from O to P, and denote
the angle between the positive x-axis and the ray OP. Then
sin =
y
r
tan =
y
x
sec =
r
x
cos =
x
r
cot =
x
y
csc =
r
y
Since r is positive (
always means the positive square root) these formulas

make it easy to remember the symmetries
sin() = sin(), cos() = cos ,
(which say that the sine is an odd function and the cosine is an even function)
and the sign reversals
sin( +) = sin, cos( +) = cos .
All the trigonometric functions have period 2:
f( + 2) = f(), f = sin, cos, tan, cot, sec, csc .
Because of the above sign reversal formulas for the sine and cosine and the
equations
tan =
sin
cos
, cot =
cos
sin
,
the tangent and cotangent have period :
tan( +) = tan, cot( +) = cot .
The cofunction of an angle is the function of its complement:
cos = sin
_
2

_
, cot = tan
_
2

_
, csc = sec
_
2

_
.
The trigonometric addition formulas are
sin( +) = sincos + cos sin,
cos( +) = cos cos sinsin.
13
Here is a proof of the addition formula for the sine
function. The area of a triangle POQ is
Area POQ =
ab
2
sin
where = POQ, a = [OP[ is the length of the
base, and b = [OQ[ so b sin is the altitude. In
the picture at the right the angle is obtuse but
this makes no dierence since
sin = sin( )
P
@
@
@
@
@
@
@
@
@
@
@
O
B
B
B
B
B
B
B
B
B
B
B
Q
a
b
Turn the picture on its side and drop the altitude
OR from the vertex O to the (new) base PQ. De-
ne
= POR, = ROQ,
so
+ = POQ =
The (new) altitude is
h = [OR[ = a cos = b cos
P Q
H
H
H
H
H
H
H
H
H
HH
O
R

a b
The areas of the right triangles are
Area POR =
1
2
(a sin)h =
ab
2
sincos ,
and
Area ROQ =
1
2
(b sin)h =
ab
2
cos sin.
From above
Area POQ =
ab
2
sin =
ab
2
sin( +)
But
Area POQ = Area POR + Area ROQ.
Substituting and dividing by ab/2 gives the trigonometric addition formula
sin( +) = sincos + cos sin
for the sine. The trigonometric addition formula for the cosine follows easily
from the formula for the sine and the principle that the cosine of an angle is the
14
sine of its complement:
cos( +) = sin
_
2
( +)
_
= sin
_
(
2
) + ()
_
= sin(
2
) cos() + cos(
2
) sin()
= cos cos sinsin.
Exercises
Exercise 4.2. Let u = tan be the tangent of an acute angle. Express sin,
cos , cot , sec , and csc as functions of u. (Hint: Draw a right triangle
whose legs are 1 and u.)
Exercise 4.3. Prove that
sin( +) = sin, cos( +) = cos , tan( +) = tan.
Exercise 4.4. Prove the trigonometric addition formula
tan( +) =
tan + tan
1 tantan
for the tangent function. (Hint: the tangent is the sine divided by the cosine.)
Exercise 4.5. True or false: sin
2
(x) = (sinx)
2
? sin
1
(x) = (sinx)
1
?
5 Additional Exercises
Exercise 5.1. Find an equation for the set of points in the plane which are
equidistant from the points P
1
(1, 3) and P
2
(2, 4). What is the geometric shape
of this set?
Exercise 5.2. Find an equation for the set of points in the plane which are
equidistant from the point P
0
(3, 1) and the line y = 2. What is the geometric
shape of this set?
Exercise 5.3. Find an equation for the set of points in the plane whose distance
to P
1
(1, 0) is twice their distance to P
2
(3, 0). What is the geometric shape of
this set?
Exercise 5.4. Find the distance from the point P
0
(4, 3) to the line y = 3x+1.
Exercise 5.5. Find the distance from the point P
0
(x
0
, y
0
) to the line y = mx+b.
Exercise 5.6. (i) Draw the graph of y = sin with the -axis horizontal and
the y-axis vertical.
(ii) Thicken that part of the graph where
2
. This is the graph of
= sin
1
(y).
15
(iii) Of course, sin
1
(sin) = for
2
but what is sin
1
(sin10)? Hint:
sin( + 2) = sin( ) = sin and 3.1 < < 3.2.
Exercise 5.7. A revolving beacon from a lighthouse shines on the straight
shore, and the closest point on the shore is a pier one half mile from the light-
house. Let denote the positive acute angle between the shore and the beam
of light. Write the distance from the pier to the point where the light shines on
the shore as a function of .
Exercise 5.8. Your town is threatened by ooding and must build a levy. The
cost C (in millions of dollars) of building a levy is a linear function C = f(H)
of the height H (in feet) of the levy. (This means that the graph of C = f(H)
is a straight line.) A one foot high levy costs $2.5 million to build; a half foot
high levy costs $1.5 million to build. The state government and the federal
government will subsidize the project in dierent ways. The state will pay
$400,000 ($0.4 million) independent of the height, whereas the feds will pay
$1.5 million per foot. Denote by f
1
(H) the net cost if the state subsidy is used,
f
2
(H) the net cost if the federal subsidy is used, and f
3
(H) the net cost if
both subsidies are used. Find expressions for all three functions, f
1
(H), f
2
(H),
f
3
(H). Draw and label all their graphs (on the same axes). Show where the
graphs meet the C axis. (The horizontal axis should be labeled as the H-axis
and the vertical axis should be labeled as the C-axis.)
16
Chapter II
Limits
6 Tangent and Velocity
This section is a warmup for calculus. It roughly corresponds to
Thomas pages 73-83 and 134-140.
6.1. Why is division by zero undened? If a/a = 1 why doesnt 0/0 = 1?
First answer: the laws of algebra would fail. For example,
a b
a
= b
if a ,= 0. If this were to work with a = 0 we could prove 2 = 3 as follows:
0 2 = 0 = 0 3 so
2 =
0 2
0
=
0
0
=
0 3
0
= 3!??!?
Second answer: Division would not be continuous. If a ,= 0 then y/x b/a
when y b and x a as in
y = 8.001 8 and x = 1.995 2 =
y
x
=
8.001
1.995

8
2
= 4.
(The notation x a means x is approximately equal to a.) But there is no
number which is approximately equal to y/x when y 0 and x 0. For
example,
y = 0.01 0 and x = 0.0001 0 =
y
x
=
0.0001
0.01
= 0.01,
but
y = 0.0001 0 and x = 0.01 0 =
y
x
=
0.01
0.0001
= 100.
6.2. The points P(a, a
2
) and Q(x, x
2
) lie on the parabola y = x
2
. Imagine
that these points are distinct but very close to each other, say (a, a
2
) = (3, 9)
and (x, x
2
) = (3.01, 9.0601). The slope of the secant line joining P and Q is
m
PQ
=
y
x
where
y = x
2
a
2
is the dierence between the vertical coordinate x
2
of Q and the vertical coor-
dinate a
2
of P and
x = x a
17
is the dierence of the horizontal coordinates of P and Q. Since y = x
2
a
2
=
(x a)(x +a) we get
m
PQ
=
y
x
=
x
2
a
2
x a
=
(x a)(x +a)
x a
= x +a.
As x gets closer and closer to a, the point Q gets closer and closer to the point
P and the secant line PQ gets closer and closer to the tangent line to the graph
y = x
2
at the point P. The slope of the tangent line is the limiting value
m
P
= lim
QP
m
PQ
of the slope of the secant line as the point Q gets closer and closer to the point
P. By the above formula for m
PQ
, this limit is
m
P
= lim
xa
(x +a) = a +a = 2a.
We will learn to write
f
(a) =
dy
dx
x=a
= lim
xa
y
x
for the slope of the tangent line to the curve y = f(x) at the point (a, f(a)) on
that curve.
And now for something completely dierent.
(But its not.)
6.3. What is the dierence between average velocity and instantaneous veloc-
ity? Consider a car moving on a highway. When we go from Madison to Mil-
waukee along highway I-94 the mile posts go from 240 to 300. If we complete the
trip in one hour, our average velocity is 60 miles per hour, but the speedometer
gives the instantaneous velocity and this will vary over the course of the trip.
Suppose we leave Madison at one PM and arrive at Milwaukee at two PM and
the quantity s = f(t) is our position at time t as determined by the mileposts.
(Imagine that there is a milepost every few feet.) Thus 240 f(t) 300 for
1 t 2. In a short time t we move a distance s = f(t + t) f(t). Our
average velocity over the tiny time interval from t to t + t is
v
av
(t, t + t) =
s
t
.
18
If t is very small (say t = one second) this is the speedometer reading at
time t. This is (almost exactly the same as) the instantaneous velocity
v
inst
(t) =
ds
dt
= lim
t0
s
t
.
7 Limits
The material in this section of these notes is treated in Thomas
pages 84-90, 102-114, and 115-123. The precise denition of the
notation lim
xa
F(x) = L is given in Thomas section 2.3 page 91,
but we will deemphasize it. If you want to learn it, you are probably
a very good student. Come to my oce hour. In the lectures we will
use the informal denitions of this section.
7.1. The notation
lim
xa
F(x) = L
is read the limit of F(x) as x approaches a is L. It means F(x) gets closer
and closer to L as x gets closer and closer to a. Sometimes the textbook writes
this as
F(x) L as x a;
this is read as F(x) approaches L as x approaches a or F(x) goes to L as x
goes to a. I like to explain limits by writing
F(x) L when x a (but x ,= a).
The notation A B means A is approximately equal to B. In Chapter III we
will study a limit
f
(a) = lim
xa
f(x) f(a)
x a
called the derivative. The derivative is a limit, but the concept of limit is more
general. Note that in the denition of limit we consider values of x close to a
but never plug in x = a. In the interesting examples (like the derivative) setting
x = a in F(x) leads to something undened like 0/0.
7.2. In the notation
lim
xa
F(x) = L
either a or L or both can be either or . The notation x means x
is large and positive (like x = 1000000), and the notation x means x
is large and negative (like x = 1000000). We also use one sided limits: The
notation
lim
xa+
F(x) = L
means F(x) L when x a but x > a. The notation
lim
xa
F(x) = L
19
means F(x) L when x a but x < a. (Think a+ means a + 0.0000000001
and a means a 0.0000000001.)
7.3. (Limit Laws)
4
If c is a constant, then
(I) lim
xa
c = c,
and
(II) lim
xa
x = a.
These laws are obvious. Assume that the limits
lim
xa
F
1
(x) = L
1
, lim
xa
F
2
(x) = L
2
,
exist and are nite. Then
(III) lim
xa
_
F
1
(x) +F
2
(x)
_
= lim
xa
F
1
(x) + lim
xa
F
2
(x),
(IV ) lim
xa
_
F
1
(x) F
2
(x)
_
= lim
xa
F
1
(x) lim
xa
F
2
(x),
(V ) lim
xa
_
F
1
(x) F
2
(x)
_
=
_
lim
xa
F
1
(x)
_
_
lim
xa
F
2
(x)
_
,
and, if lim
xa
F
2
(x) ,= 0,
(V I) lim
xa
F
1
(x)
F
2
(x)
=
lim
xa
F
1
(x)
lim
xa
F
2
(x)
.
In other words the limit of the sum is the sum of the limits, etc. These Laws
are true because if F
1
(x) L
1
and F
2
(x) L
2
, then
F
1
(x)+F
2
(x) L
1
+L
2
, F
1
(x)F
2
(x) L
1
L
2
, F
1
(x)F
2
(x) L
1
L
2
,
and (if L
2
,= 0)
F
1
(x)
F
2
(x)

L
1
L
2
.
7.4. The Limit Laws say that you can evaluate limits by just plugging in so
long as just plugging in doesnt give you something undened like 0/0 or /.
For example,
lim
x2
x
3
1
x
2
1
=
2
3
1
2
2
1
=
8 1
4 1
=
7
3
()
but if x 2 is replaced by x 1 we cant just plug in x = 1 because we will
get 0/0. We have to do some algebra rst as in
lim
x1
x
3
1
x
2
1
= lim
x1
(x 1)(x
2
+x + 1)
(x 1)(x + 1)
= lim
x1
(x
2
+x + 1)
(x + 1)
=
3
2
.
4
See Thomas section 2.2 page 84.
20
The point is that
x
3
1
x
2
1
=
(x
2
+x + 1)
(x + 1)
for x ,= 1 and the right hand side (but not the left) is meaningful even when
x = 1.
7.5. When there is no single L such that F(x) L for x a we write
lim
xa
F(x) D.N.E.
and say that the limit does not exist. This happens because there are two
distinct quantities L
1
and L
2
with F(x) L
1
for some values of x arbitrarily
near a and F(x) L
2
for some other values of x near a. Then if there were
a quantity L such that F(x) L whenever x a we would have that L = L
1
and L = L
2
which is absurd: L
1
,= L
2
. For example, x/[x[ = 1 if x > 0 and
x/[x[ = 1 if x < 0 so
lim
x0+
x
[x[
= 1, lim
x0
x
[x[
= 1,
but
lim
x0
x
[x[
D.N.E.
as there are values of x arbitrarily near 0 (like x = 0.001) where x/[x[ = 1 and
other values of x arbitrarily near 0 (like x = 0.001) where x/[x[ = 1. The
same thing can happen with innite limits:
lim
x0+
1
x
= , lim
x0
1
x
= ,
but
lim
x0
1
x
D.N.E.
Notice that we distinguish between an innite limit and one which does not
exist:
lim
x0
1
x
2
= .
Example 7.6. Sometimes even the one sided limits dont exist. For example
lim
x
cos(x) D.N.E.
as there are large values of x (like x = 1000) where cos(x) = 1 and other large
values of x (like x = 1001) where cos(x) = 1. The function cos(x) bounces
back and forth between these two values as x gets bigger and bigger. Similarly
lim
x0+
cos
_
1
x
_
D.N.E.
as there are small positive values of x where cos(1/x) = 1 (like x = (1000)
1
)
and other small positive values of x where cos(1/x) = 1 (like x = (1001)
1
).
The function cos(1/x) bounces back and forth between these two values as x
gets closer and closer to zero.
21
7.7. Dummy Variables. Beginners sometimes use the notation lim
xa
F(x)
incorrectly. The formula (see equation () 7.4)
lim
x1
x
2
1
x
3
1
=
2
3
says that if we take a number which is close to 1 (like 1.001), call it x, and
evaluate (x
2
1)/(x
3
1) we get an answer which is close to 2/3:
x
2
1
x
3
1
=
1.001
2
1
1.001
3
1

2
3
.
The formula
lim
u1
u
2
1
u
3
1
=
2
3
says that if we take a number which is close to 1 (like 1.001), call it u, and
evaluate (u
2
1)/(u
3
1) we get an answer which is close to 2/3:
u
2
1
u
3
1
=
1.001
2
1
1.001
3
1

2
3
.
Clearly these formulas have the same meaning:
lim
x1
x
2
1
x
3
1
= lim
u1
u
2
1
u
3
1
.
We express this formula by saying that the variable x on the left and the variable
u on the right are dummy variables. If a dummy variable in an expression
is systematically replaced by a completely dierent variable, the meaning of the
expression is unchanged.
7.8. Free Variables. The method used in computing equation () 7.4 shows
more generally that
lim
xa
x
2
a
2
x
3
a
3
=
2
3a
.
In this formula the variable x on the left is a dummy variable but the variable
a which appears on both sides of the formula is a free variable; the formula
asserts a fact that is true for all values of the free variable. Thus the formula
lim
xb
x
2
b
2
x
3
b
3
=
2
3b
conveys the same information whereas the formula
lim
xb
x
2
b
2
x
3
b
3
=
2
3a
is incorrect (unless a = b). If we substitute a number for the free variable a we
get the valid formula as in
lim
x7
x
2
7
2
x
3
7
3
=
2
21
whereas it is meaningless to replace the dummy variable x by a number.
22
7.9. Another kind of change of dummy variable is
lim
xa
F(x) = lim
h0
F(a +h).
Here the dummy variable x on the left is eectively replaced by a + h on the
right. If x = a + h the condition x a and h 0 are the same. Sometimes
these substitution makes the algebra more straight forward as in
lim
xa
x
2
a
2
x
3
a
3
= lim
h0
(a +h)
2
a
2
(a +h)
3
a
3
= lim
h0
3a
2
h + 3ah
2
+h
3
2ah +h
2
=
2
3a
. ()
Compare this with equation () of 7.4.
7.10. A rational function is a ratio of two polynomials, i.e. one which can
be written in the form
F(x) =
P(x)
Q(x)
.
where P(x) and Q(x) are polynomials. Evaluate
lim
xa
P(x)
Q(x)
as follows:
Case I. If a is nite and Q(a) ,= 0, then lim
xa
P(x)
Q(x)
=
P(a)
Q(a)
by the limit laws.
For example,
lim
x3
x
2
1
x
3
1
=
9 1
27 1
=
8
26
and
lim
x0
x
2
1
x
3
1
=
0 1
0 1
= 1.
Case II. The most common case is where a is nite and Q(a) = P(a) = 0. In
this case we make the change of dummy variable as in 7.9.
lim
xa
P(x)
Q(x)
= lim
h0
P(a +h)
Q(a +h)
There will be some cancellation and we can evaluate the limit. For exam-
ple,
lim
x1
x
2
1
x
3
1
= lim
h0
(1 +h)
2
1
(1 +h)
3
1
= lim
h0
(1 + 2h +h
2
) 1
(1 + 3h + 3h
2
+h
3
) 1
=
= lim
h0
2h +h
2
3h + 3h
2
+h
3
= lim
h0
2 +h
3 + 3h +h
2
=
2
3
23
Case III. If P(a) ,= 0 and Q(a) = 0 then we can factor out a a power of h
from the denominator Q(a +h), i.e.
Q(a +h) = h
n
M(a +h)
where M(a) ,= 0. If n is odd, the one sided limits exist and are innite
and opposite in sign so the two sided limit does not exist. For example,
lim
x1+
1
x 1
= lim
h0+
1
h
= , lim
x1
1
x 1
== lim
h0
1
h
= ,
and thus the two sided limit does not exist:
lim
x1
1
x 1
D.N.E.
If n is even, the one sided limits are innite and have the same sign so the
two sided limit is innite. For example,
lim
x1
1
(x 1)
2
= lim
h0
1
h
2
= .
Case IV. If a is innite, multiply top and bottom by the appropriate power of
x so that the limit on the bottom is nonzero and nite. Then use the law
that the limit of the quotient is the quotient of the limits. For example,
lim
x
x
2
1
x
3
1
= lim
x
x
2
1
x
3
1

x
3
x
3
= lim
x
x
1
x
3
1 x
3
=
0 0
1 0
= 0
Theorem 7.11. Sandwich Theorem.
5
Suppose that
f(x) g(x) h(x)
(for all x) and that
lim
xa
f(x) = lim
xa
h(x).
Then
lim
xa
f(x) = lim
xa
g(x) = lim
xa
h(x).
Proof: If f(x) L and h(x) L then g(x) L since g(x) is between f(x) and
h(x).
Example 7.12. The Sandwich Theorem says that if the function g(x) is sand-
wiched between two functions f(x) and h(x) and the limits of the outside func-
tions f and h exist and are equal, then the limit of the inside function g exists
and equals this common value. For example
[x[ xcos
_
1
x
_
[x[
5
See Thomas section 2.2 page 88. Some books call this the Squeeze Theorem.
24
since the cosine is always between 1 and 1. Since
lim
x0
[x[ = lim
x0
[x[ = 0
the Sandwich Theorem tells us that
lim
x0
xcos
_
1
x
_
= 0.
Recall that in Example 7.6 we saw that the limit lim
x0
cos(1/x) does not exist.
Exercises
Exercise 7.13. Find the following limits. (Write D.N.E. if the limit does not
exist.)
(a) lim
x7
(2x + 5) (b) lim
x7
(2x + 5) (c) lim
x
(2x + 5)
(d) lim
x4
(x + 3)
2006
(e) lim
x4
(x + 3)
2007
(f) lim
x
(x + 3)
2007
(g) lim
t1
t
2
+t 2
t
2
1
(h) lim
t1
t
2
+t 2
t
2
1
(i) lim
t1
t
2
+t 2
t
2
1
(j) lim
x
x
2
+ 3
x
2
+ 4
(k) lim
x
x
5
+ 3
x
2
+ 4
(l) lim
x
x
2
+ 1
x
5
+ 2
(m) lim
x
(2x + 1)
4
(3x
2
+ 1)
2
(n) lim
u
(2u + 1)
4
(3u
2
+ 1)
2
(o) lim
t0
(2t + 1)
4
(3t
2
+ 1)
2
(p) lim
x2
x
2
4
x
3
8
(q) lim
x
xsin(x) (q) lim
x
x
2
sinx
x + sinx
Exercise 7.14. Evaluate lim
x9
x 3
x 9
. Hint: Multiply top and bottom by
x + 3.
x2
1
x

1
2
x 2
. Hint: The function is rational.
x2
1
x

1
2
x 2
.
Exercise 7.17. Compute the one-sided limits lim
xc
f(x) and lim
xc+
f(x)
for c = 1 and c = 1 where the function f is dened by
f(x) =
_
_
_
x
3
for x < 1
ax +b for 1 x < 1
x
2
+ 2 for x 1.
where a and b are constants.
25
8 Two Limits in Trigonometry
In this section we prove two important limit formulas which we will
need in Section 15. The rst is proved as Thomas page 105 and the
second is an easy consequence.
8.1. The length of a circular arc is
given by
L = r
where r is the radius of the circle and
is the central angle of the circle. Con-
vince yourself of this by trying = 2,
= , = /2, = /4 etc. Similarly,
the area of a circular sector is given by
A = r
2
2
.
Theorem 8.2.
lim
0
sin
= 1, lim
0
1 cos
= 0.
Proof: Comparing the area of the cir-
cular sector with the areas of two right
triangles gives
sin cos
2
<

2
<
sin
2 cos
,
from which we get the limit formula
by algebra and the Sandwich Theorem.
The second formula is proved by multi-
plying top and bottom by 1 + cos .
Exercises
26
Exercise 8.3. Evaluate Find the limit or show that it does not exist. Distin-
guish between a limit which is innite from one which does not exist.
(a) lim
0
tan
. (b) lim
x0
1 cos x
xsinx
. (c) lim
x
2x
3
+ 3x
2
cos x
(x + 2)
3
.
(d) lim
x0
sin(x
2
)
x
2
. (e) lim
x0
x(1 cos x)
tan
3
x
. (f) lim
x0
sin(x
2
)
1 cos x
.
(g) lim
x0
cos x
x
2
+ 9
. (h) lim
x
sinx
x
. (i) lim
x0
sinx
x + sinx
.
9 Continuity
This material in this section comes from Thomas pages 124-133
and 147-149.
Denition 9.1. A real valued function f is continuousat a i
lim
xa
f(x) = f(a) (C)
A function is continuous i it is continuous at every a in its domain. A function
is continuous on a set (e.g. interval) I i it is dened and continuous at every
point a of I. Note that when we say that a function is continuous on some
interval it is understood that the domain of the function includes that interval.
For example, the function f(x) = 1/x
2
is continuous on the interval 1 < x < 5
but is not continuous on the interval 1 < x < 1.
Denition 9.2. The derivative of the function f is the function f
whose
value at the point a is given by
f
(a) = lim
xa
f(x) f(a)
x a
. (D)
A function f is said to be dierentiable on an interval I i the limit f
(a)
exists for every point a in I, i.e. i the domain of the derivative f
contains the
interval I.
Theorem 9.3. A dierentiable function is continuous.
27
Proof: Assume that the limit (D) exists. Then
_
lim
xa
f(x)
_
f(a) =
_
lim
xa
f(x)
_
_
lim
xa
f(a)
_
(I)
= lim
xa
_
f(x) f(a)
_
(III)
= lim
xa
f(x) f(a)
x a
(x a) (hsa)
= lim
xa
f(x) f(a)
x a
lim
xa
(x a) (V)
=
_
lim
xa
f(x) f(a)
x a
_
0 (I, IV, II)
= f
(a) 0 (denition)
= 0 (hsa)
so Equation (C) holds. In this proof the Roman numerals refer to the corre-
sponding Limit Law from7.3, (denition) means by denition, and (hsa) means
high school algebra.)
Remark 9.4. The converse of Theorem 9.3 is false. The function
f(x) = [x[
is continuous but not dierentiable at x = 0 since
f(x) f(0)
x 0
=
[x[
x
and lim
x0
|x|
x
does not exist by 7.5.
Example 9.5. A stupid way to make an example of a discontinuous function
is the following:
f(x) =
_
x
2
if x ,= 3,
47 if x = 3.
Then
lim
x3
f(x) = 9 ,= 47 = f(3).
The reason that the limit is 9 is that lim
xa
f(x) = L means that f(x) L
when x a but x ,= a; i.e. in the denition of limit (see 7.1) the actual value
of f(a) is explicitly excluded.
28
Example 9.6. The function
f(x) =
_
_
_
xcos
_
1
x
_
for x ,= 0,
0 for x = 0
is continuous by the Sandwich Theorem (see Example 7.12) but the function
g(x) =
_
_
_
cos
_
1
x
_
fot x ,= 0,
0 for x = 0
is not continuous at 0 (see Example 7.6).
Theorem 9.7 ( Intermediate Value Theorem). A continuous function f(s)
dened on the closed interval a x b takes every value between f(a) and f(b).
In other words, if f(a) < v < f(b) or f(b) < v < f(a) there is a c such that
a < c < b and f(c) = v.
9.8. Informally, the Intermediate Value Theorem says that the graph of a
continuous function is connected, i.e. you can draw the graph without lifting
your pencil. For example, the equation 23 = x
5
+x has a solution x satisfying
1 < x < 2 because the function f(x) = x
5
+x is continuous and
f(1) = 2 < 23 < 34 = f(2).
As you draw the graph of y = x
5
+x starting at the point (1, 2) and ending at
the point (2, 34) your pencil crosses over some point (x, 23). This is not true for
example for the discontinuous function
h(x) =
_
_
x
[x[
if x ,= 0,
0 if x = 0.
Since h(x) = 1 for x < 0 and h(x) = 1 for x > 0 you cannot draw the graph
y = h(x) without lifting your pencil.
Remark 9.9. The previous paragraph notwithstanding do not say that a
function is continuous i you can draw its graph without lifting your pencil.
A pencil is not a (formal) mathematical concept.
9.10. We say that f attains its maximum at c and its minimum at d and
that f(c) is the maximum value of f and f(d) is the minimum value of f
on the interval. Often the maximum is attained at an endpoint, i.e. c = a or
c = b and similarly for the minimum. For example, on the interval 2 x 3
the function f(x) = x
2
attains its minimum value 4 = f(2) at the left endpoint
and its maximum value 9 = f(3) at the right endpoint. When a < c < b we say
that c is an interior minimum; when a < d < b we say that d is an interior
29
maximum. For example, on the interval 2 x 1 the function f(x) = x
2
attains its maximum f(2) = 4 at the left endpoint and its minimum value
f(0) = 0 at the interior point 0. On an interval which is not closed a continuous
function need not assume its maximum or minimum. For example, the function
g(x) = 1/x is continuous on the interval 0 < x 1 but there is no number c
satisfying 0 < c 1 and g(x) g(c) for all x. This is because no matter what
c we pick in the interval 0 < x 1 we have that f(c
) > f(c) when c
is nearer
0 then c, e.g. when c
= c/2.
Theorem 9.11 (The Extreme Value Theorem). Let the function f(x) be
continuous on the closed nite interval a x b. Then there is at least one
number c such that a c b and f(x) f(c) for all x in the interval a x b
and there is at least one number d such that a d b and f(d) f(x) for all
x in the interval a x b.
Remark 9.12. The Intermediate Value Theorem and the Extreme Value Theo-
rem are stated but not proved in the textbook (see Thomas Theorem 11 page 130
and Theorem 1 page 246 ). They are normally proved in more advanced courses
like Math 521. Both theorems tell us that certain problems have a solution,
but the theorems dont tell us how to nd it. They will provide the theoretical
justication for the reasoning we employ in Chapter III.
Exercises
Exercise 9.13. Why doesnt the proof of Theorem 9.3 shows that every func-
tion is continuous?
Exercise 9.14. Find a constant k such that the function
f(x) =
_
3x + 2 for x < 2
x
2
+k for x 2.
is continuous. Hint: Compute the one-sided limits.
Exercise 9.15. Find constants a and b such that the function
f(x) =
_
_
_
x
3
for x < 1
ax +b for 1 x < 1
x
2
+ 2 for x 1.
is continuous for all x.
Exercise 9.16. Is there a constant k such that the function
f(x) =
_
sin(1/x) for x ,= 0
k for x = 0.
is continuous? If so, nd it; if not, say why.
Exercise 9.17. Let g(x) be continuous on the interval 1 x 1 and suppose
that g(1) = 3 and g(1) = 5. Can we necessarily nd an x between 1 and 1
with g(x) = 4? How about g(x) = 6? Explain.
Exercise 9.18. Prove that the equation x
5
4x+1 = sinx has a solution x in
the interval 1 x 2.
30
Chapter III
Dierentiation
10 Derivatives Dened
Derivatives and the dierentiation laws are explained in Thomas
pages 147-170.
Denition 10.1. The derivative of the function f is the function f
whose value at the point a is given by

f
(a) = lim
xa
f(x) f(a)
x a
.
A function f is said to be dierentiable at a i this limit exists and
dierentiable on an interval I i it is dierentiable every point a in I,
i.e. i the domain of the derivative f
contains the interval I; f is said to

be continuously dierentiable on an interval I i it is dierentiable
on I and its derivative f
is continuous on I.
10.2. Here are two ways of writing the denition of the derivative:
f
(x) = lim
h0
f(x +h) f(x)
h
, (1)
f
(a) = lim
xa
f(x) f(a)
x a
. (2)
To use equation (1) to estimate f
(4) we might take x = 4 and h = 0.001 0,

so x +h = 4.001 and
f
(4) = f
(x)
f(x +h) f(x)
h
=
f(4.001) f(4)
0.001
.
To use equation (2) to estimate f
(4) we might take a = 4 and x = 4.001 4,

so x a = 0.001 and
f
(4) = f
(a)
f(x) f(a)
x a
=
f(4.001) f(4)
0.001
.
This illustrates that the two formulas are dierent ways of expressing the same
thing. In equation (1) the variable h is a dummy variable whereas the variable
31
x is a free variable. If all occurrences of a dummy variable in an expression
are changed to another letter, the meaning of the expression is unchanged:
lim
k0
f(x +k) f(x)
k
= lim
h0
f(x +h) f(x)
h
.
A free variable is used to assert that an equation is valid for a range of values;
if the free variable is changed on one side of an equation it must be changed on
the other side as well. Thus (1) could be written
f
(y) = lim
h0
f(y +h) f(y)
h
.
One can substitute a number for a free variable as in
f
(4) = lim
h0
f(4 +h) f(4)
h
,
but substituting a number for a dummy variable yields nonsense. In equa-
tion (2), x is the dummy variable and a is the free variable.
10.3. We can calculate some derivatives using the method of equation ()
of 7.9. For example,
the derivative of the function
f(x) = x
2
is the function
f
(x) = 2x.
Here is the proof:
f
(x) = lim
h0
f(x +h) f(x)
h
= lim
h0
(x +h)
2
x
2
h
= lim
h0
(2x +h) = 2x.
the derivative of the identity function
g(x) = x
is the constant function
g
(x) = 1.
Here is the proof:
g
(x) = lim
h0
g(x +h) g(x)
h
= lim
h0
(x +h) x
h
= lim
h0
h
h
= 1.
the derivative of the constant function
k(x) = c
32
is the zero constant function
k
(x) = 0.
Here is the proof:
k
(x) = lim
h0
k(x +h) k(x)
h
= lim
h0
c c
h
= lim
h0
0 = 0.
10.4. The Dierentiation Rules. These are used to dierentiate expressions
in functions u and v when you know how to dierentiate u and v. In the following
Let c and n are constants, u and v are functions, and

denotes dierentiation.
The Dierentiation Rules are
(Constant Rule) c
= 0,
(Sum Rule) (u v)
= u
,
(Product Rule) (u v)
= u
v +u v
,
(Quotient Rule)
_
u
v
_
=
u/ v u v/
v
2
,
(Power Rule) (u
n
)
= nu
n1
u
.
Note that we already proved the Constant Rule above.
10.5. Proof of the Sum Rule. Suppose that f(x) = u(x) + v(x) for all x
where u and v are dierentiable. Then
f
(a) = lim
xa
f(x) f(a)
x a
(denition)
= lim
xa
_
u(x) +v(x)
_
_
u(a) +v(a)
_
x a
(hypothesis)
= lim
xa
_
u(x) u(a)
x a
+
v(x) v(a)
x a
_
(hsa)
= lim
xa
u(x) u(a)
x a
+ lim
xa
v(x) v(a)
x a
(limit law)
= u
(a) +v
(a) (denition)
33
10.6. Proof of the Product Rule. Suppose that f(x) = u(x)v(x) for all x
where u and v are dierentiable. Then
f
(a) = lim
xa
f(x) f(a)
x a
(denition)
= lim
xa
u(x) v(x) u(a) v(a)
x a
(hypothesis)
= lim
xa
__
u(x) u(a)
x a
_
v(a) +u(x)
_
v(x) v(a)
x a
__
(hsa)
=
_
lim
xa
u(x) u(a)
x a
_
v(a) +u(a)
_
lim
xa
v(x) v(a)
x a
_
(limit laws)
= u
(a)v(a) +v
(a)u(a) (denition)
(In the fourth step the theorem that a dierentiable function is continuous is
also used.)
10.7. Proof of the Quotient Rule. Suppose that f(x) = u(x)/v(x) for all
x where u and v are dierentiable and v(a) ,= 0. Then
f
(a) =
= lim
xa
f(x) f(a)
x a
(denition)
= lim
xa
_
u(x)/v(x)
_
_
u(a)/v(a)
_
x a
(hypothesis)
= lim
xa
u(x)v(a) u(a)v(x)
v(x)v(a)(x a)
(hsa)
= lim
xa
_
u(x) u(a)
x a

v(a)
v(x)v(a)

u(a)
v(x)v(a)

v(x) v(a)
x a
_
(hsa)
=
_
lim
xa
u(x) u(a)
x a
_
v(a)
v(a)
2

u(a)
v(a)
2

_
lim
xa
v(x) v(a)
x a
_
(lim law)
= u
(a)
v(a)
v(a)
2

u(a)
v(a)
2
v
(a) (denition)
=
u
(a)v(a) u(a)v
(a)
v(a)
2
(hsa)
(In the fth step the theorem that a dierentiable function is continuous is also
used.)
34
Remark 10.8. The Quotient Rule can be derived from the Product Rule as
follows: if w = u/v then w v = u so w
v +w v
= u
so
w
=
u
w v
v
=
u
(u/v) v
v
=
u
v u v
v
2
.
Unlike the argument above, this argument does not prove that w is dierentiable
if u and v are. Note that the rule
(cu)
= cu
is a trivial consequence of the Constant Rule and the Product Rule though of
course it can be quite easily proved directly. A special case of the Quotient Rule
is when u = 1. Then u
= 0 and the Quotient Rule reduces to

_
1
v
_
=
v
v
2
.
10.9. Proof of the Power Rule for Positive Integer Exponents. First
we prove the Power Rule when the exponent n is a nonnegative integer. When
n = 0 we have u
0
= 1 and the Power Rule is the Constant Rule. When n = 1
the formula says that u
= u
. Taking u = v in the Product Rule gives

(u
2
)
= u
u +u u
= 2u u
which is the Power Rule with n = 2. Using this and taking v = u

2
in the
Product Rule gives
(u
3
)
= (u u
2
)
= u
u
2
+u (u
2
)
= u
u
2
+u (2u u
) = 3u
2
u
.
Proceeding in this way we see that once we have proved the Power Rule for
some exponent n we can take v = u
n
in the Product Rule and get
(u
n+1
)
= (u u
n
)
= u
u
n
+u (u
n
)
= u
u
n
+u (nu
n1
u
) = (n+1)u
n
u
which proves the Power Rule for exponent n+1. Hence the Power Rule holds for
any nonnegative integer exponent n. (This style of proof is called Mathematical
Induction.)
10.10. Proof of the Power Rule for Negative Integer Exponents. We
can now prove the Power Rule for negative exponents using the the Quotient
Rule with u = 1 and v = u
n
:
(u
n
)
=
_
1
u
n
_
=
nu
n
u
u
2n
= nu
n1
u
.
10.11. Proof of the Power Rule for Rational Powers. We prove the
Power Rule when the exponent n is a rational number, i.e. a number of form
n = p/q where p and q are integers. For thus assume that
w = u
p/q
.
35
Raising both sides to the qth power gives
w
q
= u
p
.
Applying the Power Rule to both sides gives
qw
q1
w
= pu
p1
u
.
Dividing both sides by qw
q1
and substituting u
p/q
for w gives
w
=
pu
p1
u
qw
q1
=
pu
p1
u
qu
p(q1)/q
=
pu
p1
u
qu
p(p/q)
=
p
q
u
(p/q)1
u
which is the Power Rule for n = p/q. (In 38.4 we will prove the Power Rule
for any real exponent n.)
Example 10.12. Using the Dierentiation Rules you can easily dierentiate
any polynomial and hence any rational function. For example, using the Sum
Rule, the Power Rule with u(x) = x, the rule (cu)
= cu
, the derivative of the

polynomial
f(x) = 2x
4
x
3
+ 7
is
f
(x) = 8x
3
3x
2
.
By the Quotient Rule the derivative of rational function
g(x) =
2x
4
x
3
+ 7
1 +x
2
is
g
(x) =
(8x
3
3x
2
)(1 +x
2
) (2x
4
x
3
+ 7)2x
(1 +x
2
)
2
=
6x
5
5x
4
+ 8x
3
3x
2
+ 14x
(1 +x
2
)
2
.
Example 10.13. The derivative of f(x) =

x = x
1/2
is
f
(x) =
1
2
x
1/21
=
1
2
x
1/2
=
1
2x
1/2
=
1
2
x
where we used the power rule with n = 1/2 and u(x) = x.
Exercises
Exercise 10.14. Using the methods of Chapter II compute the limit
lim
x0
x + x
x + 2 + x

x
x + 2
x
36
Exercise 10.15. Using the denition of the derivative as a limit (only method
allowed), compute the derivative of f(x) =
x
x + 2
. Hint: See previous problem.
Exercise 10.16. Find f
(x) and g
(x) if f(x) = x
4
and g(x) = (1 +x
2
)
4
. Hint:
Use the power rule with u = x to nd f
(x) and with u = 1 +x

2
to nd g
(x).
Exercise 10.17. Let f(x) = (x
2
+1)(x
3
+3). Find f
(x) in two ways, rst by

multiplying and then dierentiating, and then using the product rule. Are the
answers the same?
Exercise 10.18. Let f(x) = (1+x
2
)
4
. Find dy/dx in two ways, rst by expand-
ing to get an expression for f(x) as a polynomial in x and then dierentiating,
and then by using the power rule. Are the answers the same?
Exercise 10.19. The equation
2x
x
2
1
=
1
x + 1
+
1
x 1
holds for all values of x so you should get the same answer if you dierentiate
both sides. Check this.
(x) and g
(x) if
f(x) =
1 +x
2
2x
4
x
3
+ 7
, g(x) =
2x
4
x
3
+ 7
1 +x
2
.
Note that f(x) = 1/g(x). Is it true that f
(x) = 1/g
(x)? What is the relation

between f
(x) and g
(x)? Hint: f(x) = g(x)

1
. What is the derivative of
f(x) g(x)?
(x) if f(x) = 1 +x +x
2
/2 +x
3
/3 +x
4
/4.
11 Higher Derivatives and Dierential Notation
Higher derivatives and Dierential Notation are explained in Thomas
page 168 and pages 221-231 respectively.
11.1. If y = f(x) we often use the notation, called Leibniz notation,
6
dy
dx
= f
(x)
for the derivative. This notation is very suggestive: changing the dummy vari-
able h to x (see 10.2) gives
dy
dx
= f
(x) = lim
h0
f(x +h) f(x)
h
= lim
x0
f(x + x) f(x)
x
= lim
x0
y
x
,
6
I called it variable notation in Remark 3.9
37
where
y = f(x + x) f(x).
Thus
y
x

dy
dx
when x 0.
When y = f(x), the equations
m = f
(x
0
) and m =
dy
dx
x=x0
are synonymous: both indicate that the derivative is to be evaluated at x = x
0
.
We use the former in functional notation and the latter in Leibniz notation.
Example 11.2. When we use functional notation as in
f(x) = x
3
x
we may write things like f(2) = 2
3
2 = 6, f(t) = t
3
t, f
(x) = 3x
2
1, and
f
(2) = 3(2)
2
1 = 11. When we use variable (Leibniz) notation as in
y = x
3
x
we may say things like the point (x, y) = (2, 6) lies on the curve y = x
3
x,
and write things like
dy
dx
= 3x
2
1,
dy
dx
x=2
= 11.
11.3. The second derivative of a function f is the derivative of the derivative of
f. It is denoted f
. The third derivative of a function f is the derivative of the

second derivative of f. It is denoted f
. The nth derivative of f is sometimes

denoted f
(n)
. Thus
f
(0)
= f, f
(1)
= f
, f
(2)
= f
, f
(3)
= f
, . . . .
Another notation for the nth derivative of y = f(x) is
d
n
y
dx
n
= f
(n)
(x).
11.4. For reasons which will become apparent when we study integration, we
sometimes also use dierential notation
dy = (3x
2
1) dx
instead of
dy
dx
= 3x
2
1. This handy notation reminds us that a derivative is
roughly a quotient of two innitely small quantities.
Warning: Never write something like dy = 3x
2
1, i.e. dont forget the dx.
The quantity dy is innitely small whereas 3x
2
1 is not.
38
11.5. Another common notation is operator notation as in
d
dx
(x
3
x) = 3x
2
1.
This allows us to avoiding introducing a name for x
3
x. It also explains why
we write
d
2
y
dx
2
=
_
d
dx
_
2
y
for the second derivative of y with respect to x. Be careful to distinguish the
second derivative from the square of the rst derivative. Usually
d
2
y
dx
2
,=
_
dy
dx
_
2
.
Denition 11.6. (Slopes and Tangents). The tangent line to a curve at
a P
0
(x
0
, y
0
) on a curve is the limit of the secant line connecting P
0
(x
0
, y
0
) to a
nearby point P(x, y) on the graph as the point P approaches the point P
0
; the
slope of the curve at P
0
is the slope of tangent line at P
0
. The slope of the
secant line is
y
x
=
y y
0
x x
0
so the slope of the curve at P
0
is
m = lim
x0
y
x
=
dy
dx
(x,y)=(x0,y0)
(see Section 6). More precisely:
When a curve is the graph of a function y = f(x), the slope of the curve
at the point P
0
(x
0
, f(x
0
)) on the curve is the derivative
m = f
(x
0
) = lim
xx0
f(x) f(x
0
)
x x
0
evaluated at the point x = x
0
so, by the point slope formula, an equation
for the tangent line at P
0
is y = y
0
+m(x x
0
), i.e.
y = f(x
0
) +f
(x
0
)(x x
0
).
Remark 11.7. The normal line to a curve at a point P
0
on the curve is the
line through P
0
which is perpendicular to the tangent line. Hence an equation
for the normal line to the curve y = f(x
0
) at the point P
0
(x
0
, f(x
0
) is
y = f(x
0
)
1
f
(x
0
)
(x x
0
).
39
(Because tan(+

2
) = 1/ tan, the slope of a line which is perpendicular to a
line of slope m is 1/m. Hence an equation for the line through P
0
(x
0
y
0
) which
is perpendicular to the line y = y
0
+m(x x
0
) is y = y
0
(1/m)(x x
0
).)
Example 11.8. We nd equations for the tangent and normal lines to the curve
y = x
2
at the point (x, y) = (3, 9). The slope of the curve at the point (3, 9) is
m =
dy
dx
x=3
= 2x
x=3
= 6
so the tangent line is y = 9 + 6(x 3) and the normal line is y = 9
1
6
(x 3).
Warning: A common mistake is to forget to evaluate the derivative. The
equation y = 9 +2x(x3) is not an equation for the tangent line, it is not even
an equation for a line. The correct answer is y = 9+6(x3) not y = 9+2x(x3).
Exercises
Exercise 11.9. Find the second derivative of x
7
with respect to x.
Exercise 11.10. Find the rst two derivatives f
(x) and f
(x) of the function

f(x) =
x
x + 2
.
Exercise 11.11. Find dy/dx and d
2
y/dx
2
if y = x/(x+2). Hint: See previous
problem.
Exercise 11.12. Find du/dt and d
2
u/dt
2
if u = t/(t + 2). Hint: See previous
problem.
Exercise 11.13. Find
d
dx
_
x
x + 2
_
and
d
2
dx
2
_
x
x + 2
_
. Hint: See previous
problem.
d
dx
_
x
x + 2
_
x=1
and
d
dx
_
1
1 + 2
_
.
Exercise 11.15. Find g(x) such that dy = g(x) dx if y = x/(x + 2).
Exercise 11.16. Let x = (1 t
2
)/(1 +t
2
), y = 2t/(1 +t
2
) and u = y/x. Find
dx/dt, dy/dt, and du/dt.
Exercise 11.17. Find d
2
y/dx
2
and (dy/dx)
2
if y = x
3
.
Exercise 11.18. Find equations for the tangent and normal lines to the curve
y = 4x/(1 +x
2
) at origin and at the point (1, 2).
Exercise 11.19. Find equations for the tangent and normal lines to the curve
y = 8/(4 +x
2
) at the point (2, 1) and at the point (0, 2).
Exercise 11.20. Does the curve y = x
4
2x
2
+2 have any horizontal tangents?
If so, where? Hint: A line is horizontal when its slope is zero.
40
Exercise 11.21. Let f(x) = 5+3x+
9
2
x
2
+
5
6
x
3
. Find f(0), f
(0), f
(0), f
(0),
and f
(k)
(0) for k 4.
Exercise 11.22. Let f(x) = 5 + 3(x 7) +
9
2
(x 7)
2
+
5
6
(x 7)
3
. Find f(7),
f
(7), f
(7), f
(7) and f
(k)
(7) for k 4.
Exercise 11.23. Let f(x) = c
0
+ c
1
x +
c
2
2
x
2
+
c
3
6
x
3
where c
0
, c
1
, c
2
, c
3
are
constants. Find f(0), f
(0), f
(0), f
(0) and f
(k)
(0) for k 4.
Exercise 11.24. Let f(x) = c
0
+ c
1
(x a) +
c
2
2
(x a)
2
+
c
3
6
(x a)
3
where
a, c
0
, c
1
, c
2
, c
3
are constants. Find f(a), f
(a), f
(a), f
(a) and f
(n)
(a) for
n 4.
Exercise 11.25. For each non-negative integer k, the notation k! is pronounced
k-factorial and is dened to be the product of the rst k positive integers, i.e.
k! = 1 2 3 (k 1) k.
Thus 1! = 1, 2! = 2, 3! = 6, 4! = 24, etc. We also dene 0! = 1. Find f
(k)
(x)
and f
(k)
(0) for all k = 0, 1, 2, . . . if f(x) = x
n
and n is a positive integer.
(k)
(x) and f
(k)
(a) for all k = 0, 1, 2, . . . if
f(x) =
c
0
0!
+
c
1
1!
(x a) +
c
2
2!
(x a)
2
+ +
c
n
n!
(x a)
n
where a, c
0
, c
1
, c
2
, . . . , c
n
are constants.
12 Implicit Functions
Implicit dierentiation is explained in Thomas pages 205-213.
12.1. When we say that the function y = f(x) is implicitly dened by an
equation in x and y we mean that if we substitute f(x) for y in that equation,
we get an equation (in x) that holds for all values of x. In this case, we can nd
the derivative by dierentiating the equation and solving for the derivative.
Example 12.2. The function y =

1 x
2
is implicitly dened by the equation
x
2
+y
2
= 1 (with the additional condition that y 0). We can nd the derivative
explicitly via
dy
dx
=
x
1 x
2
but it is easier to view x
2
+y
2
as a (constant) function of x, dierentiate to get
0 =
d
dx
(x
2
+y
2
) = 2x + 2y
dy
dx
,
and then solve to get
dy
dx
=
x
y
=
x
1 x
2
.
41
Example 12.3. Here is a more complicated example. A dierentiable function
y = f(x) is implicitly dened by the equation
y
2
+ 3xy + 7x
2
17 = 0. ()
and satises f(1) = 2. To nd f
(1) we can dierentiate () and solve:

2y
dy
dx
+ 3x
dy
dx
+ 3y + 14x = 0
(we used the product rule when we dierentiated 3xy) so
dy
dx
=
3y + 14x
2y + 3x
. (
)
Then
f
(1) =
dy
dx
x=1
=
3y + 14x
2y + 3x
(x,y)=(1,2)
=
6 + 14
4 + 3
.
Another (harder) way is to nd an explicit formula for y by using the quadratic
formula:
y =
B
B
2
4AC
2A
=
3x
_
9x
2
4(7x
2
17)
2
where A = 1, B = 3x, and C = 7x
2
17. Because f(1) = 2 we must take the
plus sign on the right and we see that y = f(x) is explicitly dened by
y =
3x +
_
9x
2
4(7x
2
17)
2
=
3x +
68 19x
2
2
()
We can nd f
(x) by dierentiating ():

dy
dx
=
3
2

19x
2
68 19x
2
. (
)
In even more complicated examples, it will be impossible (not merely dicult)
to nd a formula for the implicitly dened function. Nonetheless we can still
compute the derivative.
Example 12.4. A typical problem asks you to nd an equation for the tangent
line to a curve at a point on the curve. For example, to nd the equation for
the tangent line to the graph of () at the point (x, y) = (1, 2) we calculate the
slope by implicit dierentiation as before:
m =
dy
dx
(x,y)=(1,2)
=
3y + 14x
2y + 3x
(x,y)=(1,2)
=
20
7
.
Then, since the value of the derivative at a point is the slope of the tangent line
at that point, the equation of the tangent line is
y = 2
20
7
(x 1)
42
where we have used the point slope equation
y = y
0
+m(x x
0
)
(see 2.2) for the equation of the line of slope m through the point (x
0
, y
0
).
Exercises
Exercise 12.5. Check that the two formulas (
) and (
) for dy/dx in Exam-

ple 12.3 are actually equal.
Exercise 12.6. Find an equation for the tangent line to the curve
x
2
+xy y
2
= 1
at the point P
0
(2, 3). Answer: y 3 =
7
4
(x 2)
Exercise 12.7. The point P(1, 2) lies on the curve
y
5
+ 3xy + 7x
5
45 = 0.
Find equations for the tangent line at P via the method of Example 12.4. In this
case you must use implicit dierentiation: there is no analog of Equation ().
(Your TA will learn this in Math 742.)
Exercise 12.8. Find equations for the tangent line and the normal line to the
curve x
3
+y
3
= 9xy at the point (x, y) = (2, 4). Hint: The slope of the normal
line is the negative reciprocal of the slope of the tangent line.
13 The Chain Rule
The Chain Rule is explained in Thomas pages 190-194.
Denition 13.1. The composition of two functions f and g is the function
is f g dened by
(f g)(x) = f(g(x)).
Theorem 13.2 (Chain Rule). If f and g are dierentiable, so is the
composition f g and its derivative at the point x = a is
(f g)
(a) = f
(g(a))g
(a).
43
Proof: The idea is that
x a =
g(x) g(a)
x a
g
(a).
u g(a) =
f(u) f(g(a))
u g(a)
f
(g(a)),
and (because a dierentiable function is continuous)
x a = g(x) g(a).
Hence
(f g)(x) (f g)(a)
x a
=
f(g(x)) f(g(a))
g(x) g(a)

g(x) g(a)
x a
f
(g(a))g
(a)
when x a. This proof isnt quite a correct proof because if g(x) = g(a) the
expression in the middle is meaningless. In general we might have g(x) = g(a)
for some values of x which are arbitarily close to a and g(x) ,= g(a) for some
other values of x which are arbitrarily close to a. But if this happens we must
have g
(a) = 0 since (g(x) g(a))/(x a) g
(a) for x a. As before the

dierence quotient
_
f(g(x))f(g(a))
_
/(xa) f
(g(a))g
(a) = 0 for x a and

g(x) ,= g(a) whereas the dierence quotient exactly equal to zero if g(x) = g(a).
Example 13.3. Take f(u) = 5 +u
2
and g(x) = 7 +x
3
so f g is given by
(f g)(x) = f(g(x)) = 5 +g(x)
2
= 5 + (7 +x
3
)
2
= 54 + 14x
3
+x
6
.
Thus (f g)
(x) = 42x
2
+6x
5
(without the chain rule) while the chain rule give
the same answer:
(f g)
(x) = f
(g(x))g
(x) = 2(7 +x
3
)3x
2
= 42x
2
+ 6x
5
.
If we modify the example by taking f(u) = 5 +
u, the direct method doesnt

apply and we have to use the Chain Rule.
13.4. In dierential notation the chain rule looks like the cancellation rule for
multiplying fractions. Thus if y = f(u) and u = g(x) the chain rule is
dy
dx
=
dy
du

du
dx
. ()
For example, if y = 5 +u
2
and u = 7 +x
3
, then
dy
dx
=
dy
du

du
dx
= 2u 3x
2
= 2(7 +x
3
)3x
2
44
as in Example 13.3. The formula () is hardly surprising since the Chain Rule
is proved by taking the limit as x tend to zero in the formula
y
x
=
y
u

u
x
()
and the quantities in () really are fractions, not just limits of fractions.
Remark 13.5. In most problems you are given a complicated formula of y in
terms of x and asked to nd dy/dx. In such a case you must decide what to
take for u. For example, if you are told that y = 4 +
7 +x
3
and asked to nd
dy/dx you might take u = 7 +x
3
so y = 4 +
u and then compute

dy
dx
=
dy
du

du
dx
=
1
2
u
3x
2
=
1
2
7 +x
3
3x
2
.
The last step (where you replace u by its denition in terms of x) is important
because the problem was presented to you with only x and y as variables and u
was a variable you introduced yourself to do the problem. After awhile you will
be able to apply the Chain Rule without introducing new letters, and you will
simply think the derivative is the derivative of the outside with respect to the
inside times the derivative of the inside and write
d
dx
(4 +
_
7 +x
3
) =
1
2
7 +x
3
3x
2
.
Remark 13.6. When f(u) = u
n
, the Chain Rule becomes the Power Rule, i.e.
d
dx
u
n
= nu
n1
du
dx
.
Thus, for the functions weve encountered so far (rational functions and frac-
tional powers), the Chain Rule gives nothing new.
13.7. Usually we have to apply the Chain Rule more than once to compute a
derivative. Thus if y = f(u), u = g(v), and v = h(x) we have
dy
dx
=
dy
du

du
dv

dv
dx
.
In functional notation this is
(f g h)
(x) = f
(g(h(x)) g
(h(x)) h
(x).
Note that each of the three derivatives on the right is evaluated at a dierent
point. Thus if b = h(a) and c = g(b) the Chain Rule is
dy
dx
x=a
=
dy
du
u=c
du
dv
v=b
dv
dx
x=a
.
45
For example, if y =
1
1 +
9 +x
2
, then y = 1/(1 + u) where u = 1 +
v and
v = 9 +x
2
so
dy
dx
=
dy
du

du
dv

dv
dx
=
1
(1 +u)
2

1
2
v
2x.
so
dy
dx
x=4
=
dy
du
u=6
du
dv
v=25
dv
dx
x=4
=
1
7

1
10
8.
Exercises
Exercise 13.8. Let y =

1 +x
3
and nd dy/dx using the Chain Rule. Say
what plays the role of y = f(u) and u = g(x).
Exercise 13.9. Repeat the previous exercise with y = (1 +
1 +x)
3
.
Exercise 13.10. Alice and Bob dierentiated y =

1 +x
3
with respect to x
dierently. Alice wrote y =

u and u = 1+x
3
while Bob wrote y =

1 +v and
v = x
3
. Assuming neither one made a mistake, did they get the same answer?
Exercise 13.11. Let y = u
3
+ 1 and u = 3x + 7. Find
dy
dx
and
dy
du
. Express
the former in terms of x and the latter in terms of u.
Exercise 13.12. Suppose that f(x) =

x, g(x) = 1 + x
2
, v(x) = f g(x),
w(x) = g f(x). Find formulas for v(x), w(x), v
(x), and w
(x).
Exercise 13.13. Suppose that f(x) = x
2
+ 1, g(x) = x + 5, and
v = f g, w = g f, p = f g, q = g f.
Find v(x), w(x), p(x), and q(x).
Exercise 13.14. Suppose that the functions f and g and their derivatives with
respect to x have the following values at x = 0 and x = 1.
x f(x) g(x) f(x) g(x)
0 1 1 5 1/3
1 3 -4 -1/3 -8/3
Dene
v(x) = f(g(x)), w(x) = g(f(x)), p(x) = f(x)g(x), q(x) = g(x)f(x).
Evaluate v(0), w(0), p(0), q(0), v
(0) and w
(0), p
(0), q
(0). If there is insu-

cient information to answer the question, so indicate.
Exercise 13.15. A dierentiable function f satises f(3) = 5, f(9) = 7,
f
(3) = 11 and f
(9) = 13. Find an equation for the tangent line to the curve
y = f(x
2
) at the point (x, y) = (3, 7).
46
14 Inverse Functions
Inverse functions are explained in Thomas pages 466-475.
14.1. A graph of an equation of form y = f(x) satises the Vertical Line
Test: every vertical line x = a intersects the graph in at most one point. If a
is in the domain of f, then the vertical line x = a intersects the graph y = f(x)
in the point P(a, f(a)); if a is not in the domain of f, then the vertical line
x = a does not intersect the graph y = f(x) at all. A graph of an equation of
form x = g(y) satises the Horizontal Line Test: every horizontal line y = b
intersects the graph in at most one point. If b is in the domain of g, then the
horizontal line y = b intersects the graph in the point P(g(b), b); if b is not in
the domain of g, then the line y = b does not intersect the graph x = g(y) at
all.
Denition 14.2. When the graphs y = f(x) and x = g(y) are the
same, i.e. when
y = f(x) x = g(y)
we say that f and g are inverse functions and write g = f
1
. Thus
domain(f
1
) = range(f), range(f
1
) = domain(f),
and
y = f(x) x = f
1
(y) (#)
for x in the domain of f and y in the range of f.
Example 14.3. The graph y = x
2
does not satisfy the horizontal line test
since the horizontal line y = 9 intersects the graph in the two points (3, 9) and
(3, 9). Therefore this graph cannot be written in the form x = g(y). However,
if we restrict the the domain to x 0 the resulting graph does have the form
x = g(y):
For x 0: y = x
2
x =

y.
Let f(x) = x
2
(with the domain articially restricted to x 0); then f
1
(y) =
y. Thus f(3) = 9 so by (#) f

1
(9) = 3. Hence f(f
1
(9)) = 9 and
f
1
(f(3)) = 3. In general:
x
2
= x, (
y)
2
= y
for x 0. But there is nothing special about this example:
47
14.4. Cancellation Equations. If the graph y = f(x) satises the
horizontal line test (so that the inverse function f
1
is dened) then
f
1
(f(x)) = x and f(f
1
(y)) = y
for x in domain(f) = range(f
1
) and y in range(f) = domain(f
1
).
To see this choose x and let y = f(x). Then x = f
1
(y) by Equation (#)
in Denition 14.2. Substituting the former in the latter gives x = f
1
(f(x)).
Reversing the roles of f and f
1
proves the other cancellation equation.
Theorem 14.5 (Inverse Function Theorem). Suppose that f and
g are inverse functions, that f is dierentiable, and that f
(x) ,= 0 for
all x. Then g is dierentiable and
g
(y) =
1
f
(g(y))
.
Proof: The fact that g is dierentiable is normally proved in more advanced
courses like Math 521. Assuming this we prove the formula for g
(y) as follows.
By the Cancellation Equations of 14.4 we have
f(g(y)) = y.
Dierentiate with respect to y and use the Chain Rule to get
f
(g(y))g
(y) = 1.
Now divide both sides by f
(g(y)).
Remark 14.6. We can also write the Inverse Function Theorem as
(f
1
)
(y) =
1
f
(f
1
(y))
.
If we use this notation, we dont need a name for the inverse function.
Remark 14.7. A handy way to summarize the formula (f
1
)
(y) = 1/f
(f
1
(y))
from Theorem 14.5 is with Leibniz notation:
48
dx
dy
=
_
dy
dx
_
1
.
For example, for x > 0 and y = x
2
we have x =

y = y
1
2
so
dy
dx
= 2x,
dx
dy
=
1
2x
=
1
2
y
,
in agreement with the power rule
d
dy
y
1
2
=
dx
dy
=
1
2
y
=
1
2
y
1
2
1
.
Example 14.8. We nd the inverse function g = f
1
of the function
f(x) = x
3
+ 1
and its derivative. Since y = x
3
+ 1 x = (y 1)
1/3
, the inverse function
is g(y) = (y 1)
1/3
so
g
(y) =
(y 1)
2/3
3
.
The following calculation conrms that g
(y) = 1/f
(g(y)):
1
f
(g(y))
=
1
3g(y)
2
=
1
3
_
(y 1)
1/3
_
2
=
1
3(y 1)
2/3
=
(y 1)
2/3
3
.
Exercises
Exercise 14.9. Find the inverse function to f(x) = 3x + 6.
Exercise 14.10. Find the inverse function to f(x) = 7 + 5x
3
. Then nd its
derivative.
Exercise 14.11. Does the function f(x) = x
3
x have an inverse? (i.e. does
it satisfy the horizontal line test?) Hint: Factor x
3
x and draw the graph.)
Exercise 14.12. Find the inverse function to f(x) =

1 x
2
where the domain
is articially restricted to the interval 0 x 1. Draw a graph.
Exercise 14.13. Let f(x) = x
5
+ x and g(y) = f
1
(y). What is f(1)? g(2)?
f(2)? g(34)? Find f
(1), g
(2), f
(2), and g
(34). Warning: Dont try to nd a

formula for g(y).
49
Exercise 14.14. Assume that y = f(x). With the information given below
you can nd dx/dy for some values of y. Which values of y and what are the
corresponding values of dx/dy?
f(3) = 4, f(5) = 6, f
(3) = 1, f
(4) = 2, f
(5) = 3, f
(6) = 4.
Exercise 14.15. (i) For which constants c does is the function dened by
f(x) =
_
3x for 0 x < 1;
4x c for 1 x,
have an inverse function? (Hint: Horizontal Line Test.) (ii) For which value of c
is f(x) continuous? (iii) Draw a graph of y = f(x) for this value of c. (iv) Find
a formula (like the formula for f(x)) for the inverse function x = f
1
(y).
15 Dierentiating Trig Functions
Dierentiation of trig functions and of inverse trig functions is cov-
ered in Thomas pages 183-190 and 517-527. Ignore for the moment
the Integration Formulas on page 528. These formulas involve the
integral sign
_
which you dont need to understand till later in the
course.
15.1. We calculate
sin
() = lim
h0
sin( +h) sin()
h
.
By the trigonometric addition formula
sin( +) = sin() cos() + cos() sin().
with = and = h the dierence quotient is
sin( +h) sin()
h
=
sin() cos(h) + cos() sin(h) sin()
h
= cos()
sin(h)
h
+ sin()
cos(h) 1
h
Hence by the formulas
lim
h0
sin(h)
h
= 1 and lim
h0
cos(h) 1
h
= 0
from Section 8 we have
d
d
sin = cos .
50
By the principle that the cosine of an angle is the sine of its complement we
have
cos() = sin(
2
),
Dierentiating this gives
cos
() = sin
2
) = cos(
2
) = sin().
(we used the Chain Rule and
d
d
(
2
) = 1), i.e.
d
d
cos = sin.
Dierentiating the formula
tan =
y
x
=
sin
cos
using the quotient rule and the formulas just proved for the derivatives of the
sine and cosine gives
tan
=
dy
d
x y
dx
d
x
2
=
cos
2
+ sin
2
cos
2
=
1
cos
2
= sec
2
so
d
d
tan = sec
2
.
15.2. Inverse Trig Functions. The trig functions sine, cosine, tangent
etc. do not satisfy the horizontal line test: they are periodic. The inverse trig
functions are dened by articially restricting the domain of the corresponding
trig function. When we do this (see 3.7) we get
If
2

2
then y = sin = sin
1
(y)
If 0 then x = cos = cos
1
(y)
If
2

2
then u = tan = tan
1
(y).
Using these and the formula from Remark 14.7 we can dierentiate the inverse
trig functions.
d
dy
sin
1
(y) =
1
_
1 y
2
.
51
Proof: Let y = sin so = sin
1
(y). Then
d
dy
=
_
dy
d
_
1
=
1
cos
=
1
_
1 sin
2
=
1
_
1 y
2
where we used the Pythogorean Theorem cos
2
+ sin
2
= 1.
d
dx
cos
1
(x) =
1
1 x
2
.
Proof: Let x = cos so = cos
1
(y). Then
d
dx
=
_
dx
d
_
1
=
1
sin
=
1
1 cos
2
=
1
1 x
2
where we used the Pythogorean Theorem as before.
d
du
tan
1
(u) =
1
1 +u
2
.
Proof: Let u = tan so = tan
1
(u). Then
d
du
=
_
du
d
_
1
=
1
sec
2
=
1
1 + tan
2
=
1
1 +u
2
where we used the Pythogorean Theorem in the form 1 + tan
2
= sec
2
. This
follows from cos
2
+ sin
2
= 1 by dividing by cos
2
. (Recall that the secant
function is the reciprocal of the cosine.)
Example 15.3. We calculate the derivative of y = sin
1 +x
2
using the Chain
Rule:
dy
dx
=
_
cos
_
1 +x
2
_
_
1
2
1 +x
2
_
2x.
Exercises
Exercise 15.4. Find
d
d
cot and
d
dv
cot
1
(v).
Exercise 15.5. Find
d
d
sec and
d
dw
sec
1
(w).
Exercise 15.6. Find the second derivative of tan with respect to .
52
Exercise 15.7. In each of the following, nd dy/dx.
(a) y = sinx. (b) y = (sinx)
1
. (c) y = sin(x
1
). (d) y = sin
1
(x).
Exercise 15.8. Consider the following functions
f
1
(x) = sin(x
2
), f
2
(x) = (sinx)
2
, f
3
(x) = (sinx)x,
f
4
(x) = sin
2
x, f
5
(x) = sin(sinx).
Which (if any) of these functions are the same? Evaluate the derivative of each
of them. Use parentheses to make absolutely certain the order of evaluation is
unambiguous. When you use the Chain Rule to dierentiate a composition f g
say which function plays the role of g and which plays the role of f.
Exercise 15.9. Find the limit. Distinguish between an innite limit and one
which doesnt exist. (Give reasons!)
(a) lim
x0
sin3x
x
. (b) lim
x
sin3x
x
. (c) lim
x0+
sin3
x
.
(d) lim
h0
sin(3 +h) sin3
h
. (e) lim
x3
sinx sin3
x 3
.
16 Exponentials and Logarithms
The material in this section of the notes corresponds roughly to Sec-
tions 7.3 and 7.4 of Thomas pages 486-500 but Section 7.3 depends
on Section 7.2 which uses integration. Thomas postpones exponen-
tials and logarithms till late in the semester, but other books
7
do
it here. The theorems which are stated but not proved here will be
proved in Section 38 of these notes and in Section 7.3 of Thomas.
Theorem 16.1. For each positive number a there is a unique function called
the exponential function base a satisfying the following conditions:
(i) The domain of exp
a
is the set of all real numbers.
(ii) The range of exp
a
is the set of positive real numbers.
(iii) The function exp
a
is continuous.
(iv) The function exp
a
converts addition into multiplication, i.e.
exp
a
(x +y) = exp
a
(x) exp
a
(y).
(v) The value of exp
a
(x) when x = 1 is
exp
a
(1) = a.
7
My favorite is Stewart: Calculus: Early Transcendentals.
53
16.2. The more familiar notation for the exponential function is
exp
a
(x) = a
x
so parts (iv) and (v) of Theorem 16.1 take the more familiar form
a
x+y
= a
x
a
y
, a
1
= a. ()
This implies that, for integers, exponentiation is repeated multiplication,
8
e.g.
a
3
= a
1+1+1
= a
1
a
1
a
1
= a a a. Using () repeatedly gives
a
nx
= a
x +x + +x
. .
n
= a
x
a
x
a
x
. .
n
= (a
x
)
n
so taking x = 1/n and using () again proves that for any positive integer n
a
1
n
=
n
a
the nth root of a (i.e. the functions b = a
n
and a = b
1
n
are inverse functions).
It follows easily that parts (iv) and (v) of Theorem 16.1 determines the value
a
x
uniquely when x is a rational number
9
and, as every real number is a limit
of rational numbers, this shows that the conditions of Theorem 16.1 uniquely
determine the exponential function. In Section 38 we give a formula for exp
a
and show that it satises the conditions of Theorem 16.1.
16.3. The following familiar laws of algebra all follow easily from equation ().
a
0
= 1, a
x+y
= a
x
a
y
, a
x
=
1
a
x
,
a
1
= a, (ab)
x
= a
x
b
x
, (a
p
)
x
= a
px
.
For example, the reason why (ab)
x
= a
x
b
x
is that both sides satisfy the
conditions of Theorem 16.1 (reading ab for a). Similarly, to prove (a
p
)
x
= a
px
note that both sides (as functions of x and reading a
p
for a) satisfy the conditions
of Theorem 16.1 (at least if p ,= 0; if p = 0 both sides equal 1).
8
Just as multiplication is repeated addition.
9
A rational number is a ratio of integers, i.e. a fraction.
54
Theorem 16.4. For a > 1, the exponential function is dierentiable, strictly
increasing, and satises so
lim
x
a
x
= , and lim
x
a
x
= 0.
These will be proved in section 38. That the exponential function y = a
x
is
strictly increasing means that a
x1
< a
x2
for x
1
< x
2
. (For example, 2
3
< 2
4
.)
This, combined with the Intermediate Value Theorem 9.7 means that the range
of the exponential function is the set of all positive numbers and the graph of
the exponential function passes the vertical line test so the exponential function
has an inverse function.
Denition 16.5. The inverse function to the exponential function
exp
a
(x) = a
x
is called the logarithm function base a:
y = a
x
x = log
a
(y).
The range of log
a
is all real numbers, the domain is all positive rea
numbers, and
log
a
(a
x
) = x, a
log
a
(y)
= y
by the Cancellation Law for inverse functions.
Remark 16.6. Since (1/a)
x
= 1/a
x
= a
x
, similar statements hold if a < 1.
Thus for b < 1
lim
x
b
x
= 0 and lim
x
b
x
=
and the function b
x
is strictly decreasing.
16.7. Because the exponential and the logarithm are inverse functions the
properties the latter may be expressed in terms of the former. For example, the
exponential function converts addition into multiplication so the logarithm func-
tion coverts multiplication into addition. (This is why the logarithm function
was invented.) The following table summarizes this.
55
a
0
= 1, log
a
(1) = 0,
a
x1+x2
= a
x1
a
x2
, log
a
(y
1
y
2
) = log
a
(y
1
) + log
a
(y
2
),
a
x
= 1/a
x
, log
a
(1/y) = log
a
(y),
a
1
= a, log
a
(a) = 1,
(a
p
)
x
= a
px
, log
a
(y
p
) = p log
a
(y).
For example, to prove log
a
(y
1
y
2
) = log
a
(y
1
) + log
a
(y
2
) let x
1
= log
a
(y
1
) and
x
2
= log
a
(y
2
). Then y
1
y
2
= a
x1
a
x2
= a
x1+x2
so
log
a
(y
1
y
2
) = log
a
(a
x1+x2
) = x
1
+x
1
= log
a
(y
1
) + log
a
(y
2
).
To prove that log
a
(y
p
) = p log
a
(y) let x = log
a
(y). Then y = a
x
so
y
p
= (a
x
)
p
= a
xp
so
log
a
(y
p
) = log
a
(a
xp
) = x p = p x = p log
a
(y).
16.8. We compute the derivative of exp
a
(x) with respect to x:
exp
a
(x) = lim
h0
exp
a
(x +h) exp
a
(x)
h
= lim
h0
a
x+h
a
x
h
= a
x
lim
h0
a
h
1
h
.
This shows that
exp
a
(x) = a
x
exp
a
(0),
i.e.
d
dx
a
x
= c a
x
, c = exp
a
(0), ()
i.e. the exponential function and its derivative are proportional.
16.9. For the function y = a
x
= exp
a
(x) the formula () is
dy
dx
= cy, c = exp
a
(0) ()
so the inverse function x = log
a
(y) satises
dx
dy
=
1
cy
. ()
In Theorem 38.3 of Section 38 we will prove that there is a number
e = 2.71827182845904523536 . . .
56
such that
exp
e
(0) = 1.
We abbreviate exp
e
by exp so by denition
exp(x) = exp
e
(x) = e
x
.
When a = e the constant c in () is 1 so () simplies to
d
dx
e
x
= e
x
(
)
i.e. the function y = e
x
is its own derivative. The inverse function log
e
(y) is
called the natural logarithm and is usually written
ln(y) = log
e
(y).
By (
) its derivative is given by

d
dy
lny =
1
e
x
=
1
y
.
To summarize:
The derivatives of the exponential function y = e
x
and its inverse func-
tion x = ln(y) are given by
d
dx
e
x
= e
x
,
d
dy
ln(y) =
1
y
.
16.10. In science it is customary to express all exponentials in base e. Sub-
stituting a for y in the cancellation law e
ln y
= e
log
e
(y)
= y gives
a = e
ln a
so
a
x
= e
(ln a)x
so by the Chain Rule
57
d
dx
a
x
= e
(ln a)x
lna = a
x
lna,
i.e. the constant c in Equation () of 16.9 is c = lna. We express log
a
(y) in
terms of the natural logarithm as follows. Let
y = a
x
()
so by a
x
= e
(ln a)x
we have
y = e
(ln a)x
. ()
By () we have x = log
a
(y) and by () we have lny = xlna. Hence
log
a
(y) =
lny
lna
as both side equal x.
Example 16.11. We nd
d
dx
lnx using the Chain Rule as follows:

d
dx
lnx =
1
2
lnx
d
dx
lnx =
1
2
lnx
1
x
.
Example 16.12. If y = e
x
2
, then, by the Chain Rule,
dy
dx
= e
x
2
(2x). Note:
e
x
2
means e
(x
2
)
not (e
x
)
2
. The latter is e
2x
.
Exercises
Exercise 16.13. If x is large, which is bigger: 2
x
or x
2
? Hint: Try x =
1, 2, 3, . . . , 10.
Exercise 16.14. Find lim
x
e
x
1
e
x
+ 1
.
Exercise 16.15. Draw the graph of y = 2
x
. What is lim
x
2
x
? lim
x
2
x
?
Exercise 16.16. Find the second derivative of lnx with respect to x.
Exercise 16.17. Find the second derivative of 5
x
with respect to x.
Exercise 16.18. Find the second derivative of x
2
e
3x
with respect to x.
d
dx
x
2
and
d
dx
2
x
.
58
d
dx
sin(e
x
) and
d
dx
e
sin x
.
d
dx
ln(sinx) and
d
dx
sin(lnx).
Exercise 16.22. Let y = e
t
cos t. Show that
d
2
y
dt
2
= 2e
t
sint.
Exercise 16.23. Let y = xe
x
. Find
d
2
y
dx
2
2
dy
dx
+y.
Exercise 16.24. In Exercise 11.17 you were asked to compute d
2
y/dx
2
and
(dy/dx)
2
to reinforce the warning that these are usually not equal. Are they
ever equal? Hint: Try y = lnx.
d
dx
x
x
,
d
dx
x
x
x
, and
d
dx
(x
x
)
x
. Hint: x
x
= e
?
.
Exercise 16.26. Let y = (x + 1)
2
(x + 3)
4
(x + 5)
6
and u = lny. Find du/dx.
Hint: Use the fact that ln converts multiplication to addition before you dier-
entiate. It will simplify the calculation.
17 Parametric Equations
Parametric Equations are explained in Thomas pages 195-200.
Denition 17.1. A pair of equations
x = f(t), y = g(t)
assigns to each value of t a corresponding point P(x, y). The set of these points
is called a parametric curve and the equations are called parametric equa-
tions for the curve. The variable t is called the parameter and we say that the
equations trace out or parameterize the curve. Often t has the interpre-
tation of time and the parametric equations describe the position of a moving
particle at time t, i.e. the point corresponding to to the parameter value t is
the position of the particle at time t. Parameters other than time are also used.
The following examples show that sometimes (but not always) we can eliminate
the parameter and nd an equation of the form
F(x, y) = 0
which describes the curve.
Example 17.2. Rectilinear Motion. Heres a parametric curve:
x = 1 +t, y = 2 + 3t.
Both x and y are linear functions of time (i.e. the parameter t), so every time t
increases by an amount t (every time t seconds go by) the rst component
59
x increases by t, and the second component y increases by 3t. The point
at P(x, y) moves horizontally to the left with speed 1, and it moves vertically
upwards with speed 3.
Which curve is traced out by these equations? In this example we can nd
out by eliminating the parameter, i.e. solving one of the two equations for t, and
substituting the value of t you nd in the other equation. Here you can solve
x = 1 +t for t, with result t = x 1. From there you nd that
y = 2 + 3t = 2 + 3(x 1) = 3x 1.
So for any t the point P(x, y) is on the line y = 3x1. This particular parametric
curve traces out a straight line with equation y = 3x1, going from left to right.
Example 17.3. Rectilinear Motion (More Generally). Any constants x
0
,
y
0
, a, b such that either a ,= 0 or b ,= 0 give parametric equations
x = x
0
+a(t t
0
), y = y
0
+b(t t
0
) ()
which trace out the line
a(y y
0
) = b(x x
0
). ()
(Both sides equal ab(t t
0
).) At time t = t
0
the point P(x, y) is at P
0
(x
0
, y
0
).
The values corresponding to Example 17.2 are t
0
= 0, x
0
= 1, y
0
= 2, a = 1,
b = 3.
Example 17.4. Going back and forth on a straight line. Consider
x = x
0
+a sint, y = y
0
+b sint.
At each moment in time the point whose motion is described by this parametric
curve is on the straight line with equation () as in Example 17.3. However,
instead of moving along the line from one end to the other, the point at P(x, y)
keeps moving back and forth along the line () between the point P
1
corre-
sponding to time t = /2 and the point P
2
corresponding to time t = 3/2.
Example 17.5. Motion along a graph. Let y = f(x) be some function of
one variable (dened for x in some interval) and consider the parametric curve
given by
x = t, y = f(t).
At any moment in time the point P(x, y) has x coordinate equal to t, and
y = f(t) = f(x), since x = t. So this parametric curve describes motion on the
graph of y = f(x) in which the horizontal coordinate increases at a constant
rate.
Example 17.6. The standard parametrization of a circle. The paramet-
ric equations
x = cos , y = sin
satisfy
x
2
+y
2
= cos
2
+ sin
2
= 1,
60
so that P(x, y) always lies on the unit circle. As increases from to +the
point will move around the circle, going around innitely often. The point runs
around the circle in the counterclockwise direction, which is the mathematicians
favorite way of running around in circles. The more general equations
x = a cos t, y = b sint.
parameterize the ellipse
x
2
a
2
+
y
2
b
2
= 1.
Example 17.7. Another parametrization of a circle. The equations
x =
1 t
2
1 +t
2
, y =
2t
1 +t
2
also parameterize the unit circle. To see this divide both sides of the identity
10
(1 t
2
)
2
+ (2t)
2
= (1 +t
2
)
2
by (1 + t
2
)
2
to get x
2
+ y
2
= 1. However the point Q(1, 0) is left out since
y = 0 only when t = 0 and x = 1 ,= 1 when t = 0.
Example 17.8. A parametrization of a hyperbola. The functions
sinh(t) =
e
t
e
t
2
, cosh(t) =
e
t
+e
t
2
are called the hyperbolic sine and hyperbolic cosine respectively. This is
because the equations
x = cosh(t), y = sinh(t),
parameterise the part of the hyperbola
x
2
y
2
= 1
which to the right of the y-axis.
17.9. For parametric equations as in Denition 17.1 the chain rule gives
dy
dt
=
dy
dx

dx
dt
so dividing gives the formula
dy
dx
=
dy/dt
dx/dt
.
10
(1 t
2
)
2
+ (2t)
2
= (1 2t
2
+t
4
) + 4t
2
= 1 + 2t
2
+t
4
= (1 +t
2
)
2
61
We can use this formula to nd the slope of the tangent line at a point on the
curve. The following example illustrates this.
Example 17.10. The point P
0
_
3
2
,
1
2
_
lies on the unit circle x
2
+y
2
= 1. This
point corresponds to the parameter value = /6 in the standard parameteri-
zation x = cos , y = sin of Example 17.6. Since
dx
d
= sin,
dy
d
= cos
we get
dx
d
=/6
=
3
2
,
dy
d
=/6
=
1
2
,
and so the slope of the tangent line at P
0
is
m =
dy
dx
=/6
=
dy/d
dx/d
=/6
=
1
3
.
The point slope equation y = y
0
+m(x x
0
) for the tangent line is
y =
1
2

1
3
_
x
3
2
_
.
Remark 17.11. Let P
0
(x
0
, y
0
) be a point on a parametric curve corresponding
to a parameter value t = t
0
and let
a =
dx
dt
t=t0
and b =
dy
dt
t=t0
.
Then
x = x
0
+a(t t
0
), y = y
0
+b(t t
0
),
are parametric equations for the tangent line to the curve at P
0
. This is because
the point slope equation for the tangent line is
y = y
0
+m(x x
0
), where m =
dy
dx
(x,y)=(x0,y0)
.
and the slope is m = b/a. (See Example 17.3.)
Exercises
Exercise 17.12. Conrm Example 17.8 by showing that
_
cosh(t)
_
2
_
sinh(t)
_
2
= 1.
This is analogous to the Pythogorean Theorem
_
cos(t)
_
2
+
_
sin(t)
_
2
= 1.
62
Also show that the hyperbolic sine and hyperbolic cosine are derivatives of each
other. Thus we have the analogous equations
d
dt
sinh(t) = cosh(t),
d
dt
cosh(t) = sinh(t),
d
dt
sin(t) = cos(t),
d
dt
cos(t) = sin(t).
Note the signs!
Exercise 17.13. The point P
0
(
3
5
,
4
5
) lies on the unit circle x
2
+y
2
= 1. In the
parameterization of Example 17.7 it corresponds to the parameter value t = 2.
Use this parameterization to nd the equation of the tangent line at this point.
Then nd the (same) equation using y =

1 x
2
.
Exercise 17.14. Consider the parameterization
x =
1 t
2
1 +t
2
, y =
2t
1 +t
2
of the unit circlefrom 17.7. For which value of t is (x, y) = (1, 0)? (0, 1)?
(0, 1)?
_
3
5
,
4
5
_
?
_
2
2
,
2
2
_
? Is there a value of t for which (x, y) = (1, 0)?
18 Approximation*
Approximation is explained in Thomas pages 221-231 and 807-810.
In the latter reference Taylor Approximation is treated at a higher
level than here. This section is included as a warmup for innite
series which we study in Math 222. Before reading this section you
should do Exercises 11.21-11.26.
Denition 18.1. Let f(x) be a dierentiable function and a a point in its
domain. The linear approximation to f(x) at x = a is the linear function
L(x) whose graph is the tangent line to the curve y = f(x) at the point (a, f(a)),
i.e.
L(x) = f(a) +f
(a)(x a)
Note that L(a) = f(a) and L
(a) = f
(a).
Theorem 18.2 (Linear Approximation Theorem). The linear approxima-
tion L(x) is the linear function which best approximates f(x) near x = a in the
sense that
lim
xa
f(x) L(x)
x a
= 0.
Proof: lim
xa
f(x) L(x)
x a
= lim
xa
_
f(x) f(a)
x a
f
(a)
_
= 0.
63
Remark 18.3. The idea is that L(x) is a good approximation to f(x) when x
is close to a, i.e. the error f(x) L(x) is small. Theorem 18.2 says that not
only is the error small it is so small that even when it is divided by the small
number x a, the result is still small.
Denition 18.4. Given a number a in the domain of f and an integer
n 0, the polynomial
P
n
(x) =
n
k=0
f
(k)
(a)(x a)
k
k!
(#)
is called the degree n Taylor polynomial of f at the point a.
18.5. The letter

is the Greek S (for sum) and is pronounced sigma so the
notation used in (#) is called sigma notation. It is a handy notation but if
you dont like it you can indicate the summation with dots:
n
k=0
a
k
= a
0
+a
1
+ +a
n1
+a
n
.
Hence the rst few Taylor polynomials are
P
0
(x) = f(a),
P
1
(x) = f(a) +f
(a)(x a),
P
2
(x) = f(a) +f
(a)(x a) +
f
(a)(x a)
2
2
,
P
3
(x) = f(a) +f
(a)(x a) +
f
(a)(x a)
2
2
+
f
(a)(x a)
3
6
.
The linear approximation to f at a is the degree one Taylor Polynomial
L(x) = P
1
(x) = f(a) +f
(a)(x a).
The degree two Taylor Polynomial
Q(x) = P
2
(x) = f(a) +f
(a)(x a) +
f
(a)(x a)
2
2
is also called the quadratic approximation to f at a.
64
Theorem 18.6. The Taylor polynomial P
n
(x) is the unique polynomial
of degree n which has the same derivatives as f at a up to order n:
P
(k)
n
(a) = f
(k)
(a) for k = 0, 1, 2, . . . , n.
(Of course, P
(k)
n
(x) = 0 for k > n.)
18.7. The Taylor polynomial P
n
(x) for f(x) at a is the polynomial of degree
n which best approximates f(x) for x near a. To make this precise let
R
n
(x) = f(x) P
n
(x)
denote the nth Taylor Error of f at a. When R
n
(x) is small P
n
(x) is a good
approximation for f(x). How small is small? The answer is given by
Theorem 18.8 (Taylors Formula). Suppose that f is n + 1 times
dierentiable and that f
(n+1)
is continuous. Let a be a point in the
domain of f. P
n
(x) be the Taylor Polynomial for f at a, and R
n
(x) be
Taylor Error for f at a. Then
(i) f(x) = P
n
(x) +R
n
(x)
(ii) lim
xa
R
n
(x)
(x a)
n
= 0.
(iii) P
n
(x) is polynomial of degree n whch satises (iii).
Well study this more in Math 222 but for now well be satised to show how
it gives accurate approximations to complicated functions. (Your calculator uses
this method.)
Example 18.9. Take f(x) = x
1/3
and a = 8. Then
f
(x) = x
2/3
/3, f
(x) = 2x
5/3
/9,
so
f(a) = 2, f
(a) = 1/12, f
(a) = 1/144,
65
and hence
L(x) = 2 +
1
12
(x 8), Q(x) = 2 +
(x 8)
12

(x 8)
2
288
.
Notice that
L(9) = 2 +
1
12
= 2.083333333 . . .
is close to
f(9) = 9
1/3
= 2.080083823 . . .
and
Q(9) = L(9)
1
288
= 2.079861111 . . .
is even closer.
Exercises
Exercise 18.10. Evaluate

5
k=1
1
k
.
1/3
. Find the polynomial P(x) of degree three
which best approximates f(x) near x = 8. Calculate P(9) and compare it with
the value of 9
1/3
given by your calculator.

x. Find the polynomial P(x) of degree three
such that P
(k)
(4) = f
(k)
(4) for k = 0, 1, 2, 3. Calculate P(3), P(5), P(3.5),
P(4.5), P(3.9), P(4.1) and compare the results with the values

3,

5,

3.5,
4.5,

3.9,

4.1
Exercise 18.13. If f(x) = e
x
what is f
(k)
(x)? f
(k)
(0)? Find the rst ve
Taylor polynomials P
k
(x) (k = 1, 2, . . . , 5) at a = 0 for e
x
. Evaluate P
k
(1) and
compare the result with e = f(1).
Exercise 18.14. If f(x) = lnx what is f
(k)
(x)? f
(k)
(1)? Find the rst ve
Taylor polynomials P
k
(x) (k = 1, 2, . . . , 5) at a = 1 for lnx. Evaluate P
k
(2) and
compare the result with ln 2.
Exercise 18.15. If f(x) = sinx what is f
(k)
(x)? f
(k)
(0)? Hint: What is
f
(4)
(x)? Find the rst ve Taylor polynomials P
k
(x) (k = 1, 2, . . . , 5) at a =
0 for sinx. Evaluate P
k
(0.5) and compare the result with sin0.5 = f(0.5).
Warning: Make sure your calculator is computing in radians.
Exercise 18.16. Is the linear approximation the only linear function satisfying
the conclusion of Theorem 18.2. Why? Is the Taylor polynomial of degree n
the only polynomial of degree n to satisfy the conclusion of Theorem 18.8?
Exercise 18.17. Denote the linear approximations to f near a and to g near
b respectively by
L(x) = f(a) +f
(a)(x a), M(y) = g(b) +g
(b)(y b),
66
and assume b = f(a). Show that
M L(x) = (g f)(a) + (g f)
(a)(x a).
(This says that The linear approximation to the composition is the composition
of the linear approximations.)
19 Additional Exercises
_
(a +x)(b +x) where a and b are constants. Show
that
f
(x) =
(b a)
2
4f(x)
3
.
Exercise 19.2. Find all points on the parabola with the equation y = x
2
1
such that the normal line at the point goes through the origin.
Exercise 19.3. (i) Find c so that function
f(x) =
_
x +c for x < 1
3
x
for x 1
is continuous. (ii) Draw a crude graph of the equation y = f(x). (iii) Give a
formula (like the above formula for f(x)) for the inverse function x = f
1
(y) of
the function y = f(x).
67
%!TEX root = lecs-221.tex
Chapter IV
Applications of Derivatives
20 The Derivative as A Rate of Change
See Thomas pages 171-177.
20.1. The area of a circle of ra-
dius r is
A = r
2
and the circumference of a circle
is
C = 2r.
It is no coincidence that
dA
dr
= 2r = C.
The picture shows a circle of radius r and a slightly larger circle with the same
center of radius r + r. The dierence in the areas is
A = (r + r)
2
r
2
= 2rr +(r)
2
2rr.
The picture shows that A is the area between the two circles and illustrates
that it is roughly (and in the limit exactly) the circumference of the inner circle
times the change r in the radius. (Imagine cutting up the area between the
two circles in a bunch of small rectangles of height r and whose bases sum to
C.)
20.2. The volume of a sphere of radius r is
V =
4
3
r
3
and the area is
S = 4r
2
.
It is no coincidence that
dV
dr
= 4r
2
= S
To see this rotate the two circles in the picture about a common diameter to
make two concentric spheres. The spherical shell between the two spheres has
volume
V =
4
3
(r + r)
3
4
3
r
3
4r
2
r.
68
This volume is roughly (and in the limit exactly) the area of the sphere times
the change r in the radius. (Imagine cutting up the volume between the two
spheres in a bunch of small blocks of height r and whose bases have areas
which sum to S.)
20.3. The average rate of change of a function y = f(x) as x changes from
x
0
to x
0
+ x is
y
x
=
f(x
0
+ x) f(x
0
)
x
.
The instantaneuous rate of change of a function y = f(x) at x = x
0
dy
dx
x=x0
= lim
x0
y
x
= f
(x
0
).
For example, suppose that the position of a particle at time t is
s = f(t) = t
3
6t
2
+ 9t
where t is measured in seconds and t is measured in meters. The distance
travelled over a tiny time interval from t to t + t is
s = f(t + t) f(t).
The average velocity over that time interval iss
v
av
=
s
t
and the instantaneous velocity is the limit
v
inst
= lim
t0
s
t
as the size t of the time interval shrinks to zero. Thus
v
inst
=
ds
dt
= f
(t) = 3t
2
12t + 9 = 3(t 1)(t 3).
The particle is at rest when ds/dt = 0, i.e. when t = 1 and when t = 3. The
particle is moving in the positive direction when ds/dt is positive, i.e. for t < 1
and for t > 3 and is moving in the negative direction when ds/dt is negative,
i.e. for 1 < t < 3. Here is a schematic diagram:
t=0
s=0
t=3
s=0
t=1
s=4
69
Exercises
Exercise 20.4. The position in meters of a particle is given by s = t
3
4.5t
2
7t
where t is the time in seconds. When does the particle reach a velocity of 5
meters per second?
Exercise 20.5. Find the average rate of change of the area of a circle with
respect to its radius r as r changes from 2 to 3, from 2 to 2.5, and from 2 to
2.1. Then nd the instantantaneous rate when r = 2.
Exercise 20.6. Find the average rate of change of the volume of a cube with
respect to its edge length x as x changes from 5 to 6, from 5 to 5.1, and from 5
to 5.01. Then nd the instantantaneous rate when x = 5.
Exercise 20.7. If a tank holds for 5000 gallons of water which drains from the
bottom of the tank in 40 minutes then Toricellis law gives the volume V of
water remaining in the tank after t minutes as
V = 5000
_
1
t
40
_
2
, 0 t 40.
Find the rate at which water is draining from the tank after 5 minutes, after 10
minutes, and after 20 minutes.
Exercise 20.8. Water runs out of a cylindrical tank from a drain in the bottom.
The water level in the tank t hours after the tank starts to drain is
y = 6
_
1
t
12
_
2
meters
and the tank drains completely after 12 hours. Find the rate (measured in
meters per hour) at which the depth is decreasing after t hours. When is the
depth decreasing the fastest? The slowest? Show that the rate at which the
water level decreases is proportional to the square root of the water level.
Exercise 20.9. A heavy object is shot straight up from the Earths surface at
200 feet per second. Elementary physics tells us that its height after t seconds
is
y = 200t 16t
2
.
Find the velocity
v =
dy
dt
and the acceleration
a =
d
2
y
dt
2
as functions of t. When is the object (momentarily) at rest? When does it move
up? Down? When is it highest? When does it change direction? How high does
it go? When it it moving fastest?
70
Exercise 20.10. The size of a poulation of bacteria in a nutrient broth after t
hours is
N = 10
6
e
0.06t
.
How large is the population and how fast is it growing at time t = 0? t = 1?
t = 2?
Exercise 20.11. If f is the focal length of a lens and an object is placed at a
distance p from the lens then the image will be at a distance q from the lens
where
1
f
=
1
p
+
1
q
.
Find the rate of change of p with respect to q.
21 Related Rates
21.1. The rst step (and usually the hardest step for students) in solving
a word problem is to reformulate it in mathematical notation. This usually
means expressing some of the quantities in the problem as functions of the
other quantities. If the formulation of the problem does not explicitly assign
letter names to these quantities, you will rst have to name them yourself. In
these problems it is very helpful to keep track of the units.
21.2. The Balloon Problem. Air is pumped into a spherical balloon so that
its volume increases at a rate of 100 cubic centimeters per second. How fast is
the radius increasing at the instant when it is 25 centimeters?
Solution. To solve this problem we rst name everything in sight:
Let V denote the volume of the balloon, r denote the radius of the
balloon, and t denote the time in seconds.
The quantities are related by
V =
4
3
r
3
.
We write the given information in mathematical notation:
dV
dt
= 100cm
3
/sec.
We write what the problem asks us to nd in mathematical notation:
dr
dt
r=25
=?
Now we use the chain rule.
dV
dt
= 4r
2
dr
dt
71
so when r = 25 we have
dr
dt
r=25
=
dV/dt
4r
2
r=25
=
100
4625
.
The answer has the units of
cm
3
/sec
cm
2
= cm/sec.
21.3. The Ladder Problem. A ladder 10 feet long leans against a vertical
wall. The bottom of the ladder moves away from the wall at one foot per second.
How fast is the top sliding down the wall when it is 6 feet above the ground?
Solution. We name everything in
sight: Let x be the distance of the bot-
tom of the ladder from the wall, y be the
height of the top of the ladder above the
ground and t be the time in seconds. By
the Pythogorean Theorem x and y are
related by
x
2
+y
2
= 10
2
.
dy
dt
y=6
=?
dx
dt
= 1
Hence x =
_
10
2
y
2
so x = 8 when y = 6. Dierentiating gives
2x
dx
dt
+ 2y
dy
dt
= 0
so
dy
dt
y=6
=
2x dx/dt
2y
y=6
=
2 8 1
2 6
.
Exercises
Exercise 21.4. The radius r of a sphere at time t is given by the formula
r =

t
2
+ 1 1.
1. Give a formula for the rate of change in the radius with respect to time.
2. Give a formula for the rate of change of the volume of the sphere with
respect to time.
3. How fast is the volume changing when t = 2?
4. How fast is the volume changing when r = 2?
Exercise 21.5. Two cars, car A traveling west at 30 miles per hour and car B
traveling south at 22.5 miles per hour, are heading toward an intersection I.
72
1. At what rate is the angle IAB changing at the instant when cars A and
B are 300 feet and 400 feet, respectively, from the intersection?
2. At what rate is the angle IAB changing at the instant when cars A and
B are 300 meters and 400 meters, respectively, from the intersection?
Exercise 21.6. A point is moving on the curve y = x
2
. At the instant when it
is passing through (3, 9), y is changing so that
dy
dt
is 7. How fast is x changing
at that time?
Exercise 21.7. Sand is owing from a pipe at the constant rate of s cubic
meters per second, and is falling in a conical pile. The diameter of the base of
this pile is always three times the altitude. At what rate is the altitude of the
pile increasing when the altitude is h meters?
11
Exercise 21.8. A highway patrol plane ies 3 miles above (and along) a level
straight road at a steady 120 miles per hour. The pilot sees an oncoming car with
radar and determines that at the moment that the distance from the plane to
the car is 5 miles, this distance is decreasing at the rate of 160 miles per hour.
Find the cars speed along the highway. This problem comes from Thomas
page 221.
Exercise 21.9. If the highway patrol plane in the Exercise 21.8 were really
designed to catch speeders, it might have a device which works as follows. It
records the altitude h, the velocity u of the plane, the angle between the line
of sight to the car and the vertical, and the rate of change d/dt of . (The
values from the previous problem are h = 3, u = 120, and, at the moment the
measurement is taken, = cos
1
(3/5).) Then it computes the speed v of the
car. Find a formula expressing v in terms of h, u, and d/dt. As before
assume that h and u are constant, but of course and d/dt are not.
Exercise 21.10. A light at the top of a pole which is h feet high. A ball is
dropped from half the height of h at a point which is at a horizontal distance
a in feet from the pole. Assume that the ball falls a distance s = gt
2
/2 feet,
where t is the time in seconds since it was dropped and g is a constant. Find
how fast the tip of the shadow of the ball is moving along the ground t seconds
after it is dropped. (Express the answer in terms of h, a, g, and t.)
Exercise 21.11. A lighthouse is located 1000 feet from the nearest point on
shore and rotates three times per minute. How fast is the end of the beam of
light it emits moving along along the shore when it passes the closest point on
shore?
Exercise 21.12. You are videotaping a race from a stand 120 feet from the
track, following a car that is moving at a constant velocity along a straight track.
11
The volume of a right circular cone whose altitude is h and whose base has radious r is
V =
r
2
h
3
. We will prove this using calculus in Chapter VI.
73
When the car is directly in front of you the camera angle is changing at a rate
of /3 radians per second. How fast is the car going? How fast will the camera
angle be changing a half second later? (The camera angle - race car problem is
like the lighthouse problem in that both use the formula x = b tan but in the
former problem dx/dt is constant while in the latter d/dt is constant.)
22 Some Theorems about Derivatives
Theorem 22.1 (First Derivative Test). Suppose that a function
f(x) dened on an interval a x b attains its minimum (or its
maximum) at a point c in the interval. Then either
(1) c is an endpoint, i.e. c = a or c = b; or
(2) f is not dierentiable at c; or
(3) c is a critical point of f, i.e. f
(c) = 0.
Proof. Assume that (1) and (2) fail, i.e. that a < c < b and that f
(c) exists.
We will prove (3). Since f
(c) exists we have that

f(x) f(c)
x c
f
(c) ()
for x c. If the ratio on the left is positive, then the numerator f(x) f(c) and
the denominator x c have the same sign; if the ratio on the left is negative,
then f(x)f(c) and xc have opposite signs. Since c is an interior point, there
are numbers in the interval to the right of c and close to c and other numbers
in the interval to the left of c and close to c. Thus
If f
(c) is positive, then f(x) f(c) > 0 for x near and to the right of c,
so c is not a maximum.
If f
(c) is positive, then f(x) f(c) < 0 for x near and to the left of c, so
c is not a minimum.
If f
(c) is negative, then f(x) f(c) < 0 x near and to the right of c, so c
is not a minimum.
If f
(c) is negative, then f(x) f(c) > 0 for x near and to the left of c, so
c is not a maximum.
74
Thus the only possibility is f
(c) = 0.
Remark 22.2. A function f(x) is said to have a
- global minimum at c if f(c) f(x) for all x in the domain of f;
- global maximum at c if f(c) f(x) for all x in the domain of f;
- local minimum at c if f(c) f(x) for all x some open interval about c;
- local maximum at c if f(c) f(x) for all x some open interval about c.
(Some books use the words absolute and relative for global and local.)
The rst derivative test says that the derivative of a dierentiable function
vanishes at a local minimum or maximum.
22.3. Examples. (1) The function L(x) = 2x+3 satises L
(x) = 2 and so is
dierentiable and has no critical point. On any interval it attains its minimum
at the left endpoint and its maximum at the right endpoint.
(2) The absolute value function g(x) = [x[ satises
g
(x) =
x
[x[
for x ,= 0 but g
(0) does not exist. On any interval containing 0 it attains its

minimum at 0 and its maximum at one of the two endpoints.
(3) The derivative of the function f(x) = x
3
3x is f
(x) = 3(x
2
1) and
f
(x) = 0 for x = 1. The point x = 1 does not lie in the interval 0 x 2

and f(0) = 0, f(1) = 2, and f(2) = 1 so on the interval 0 x 2 the
function attains its minimum value at 1 and its maximum value at 0.
Theorem 22.4 (Mean Value Theorem). Assume f(x) is continuous
on the closed interval a x b and dierentiable on a < x < b. Then
there is a point c with a < c < b and
f
(c) =
f(b) f(a)
b a
,
i.e. the tangent line to the graph at (c, f(c) is parallel to the secant
line joining (a, f(a)) and (b, f(b)).
75
(a,f(a))
y=W(x)
(b,f(b))
y=f(x)
y=L(x)
(c,f(c))
Proof: Consider the linear function
W(x) = f(a) +
f(b) f(a)
b a
(x a).
The graph y = W(x) is the line joining (a, f(a)) and (b, f(b)), i.e.
W(a) = f(a), W(b) = f(b).
The function
g(x) = f(x) W(x)
satises
g
(x) = f
(x)
f(b) f(a)
b a
, g(a) = g(b) = 0.
By the Extreme Value Theorem9.11 the function g attains its maximum and its
minimum. Since g(a) = g(b) = 0 at least one of the maximum or the minimum
must occur at an interior point c. By the First Derivative Test 22.1 we have
g
(c) = 0
as required. (The special case of the Mean Value Theorem where the values at
the endpoints are the same is called Rolles Theorem.)
Denition 22.5. A function y = f(x) is said to be increasing on an interval
i
x
1
< x
2
= f(x
1
) < f(x
2
)
for any two points x
1
, x
2
of the interval. (The symbol = means implies.)
Similarly f is said to be decreasing on an interval i
x
1
< x
2
= f(x
1
) > f(x
2
).
A function is monotonic on an interval i either it is increasing on that interval
or else it is decreasing on that interval.
76
Theorem 22.6 (Monotonicity Theorem). If
_
_
_
f
(x) > 0
f
(x) < 0
f
(x) = 0
_
_
_
for all
x in an interval I, then f is
_
_
_
increasing
decreasing
constant
_
_
_
on that interval.
Proof. Choose x
1
and x
2
in the interval I with x
1
< x
2
. By the Mean Value
Theorem there is a c with x
1
< c < x
2
and
f(x
2
) f(x
1
)
x
2
x
1
= f
(c). (#)
Since c is between x
1
and x
2
it lie in the interval I and hence f
(c) has the

sign (positive, negative, or zero) of the hypothesis. Hence the ratio on the left
in equation (#) has this same sign. Since x
1
< x
2
the denominator x
2
x
1
is
positive and hence the numerator f(x
2
) f(x
1
) has this same sign. If the sign
is positive, then f
(c) > 0 so f(x

2
) f(x
1
) > 0 so f(x
1
) < f(x
2
). If the sign is
negative, then f
(c) < 0 so f(x

2
)f(x
1
) < 0 so f(x
1
) > f(x
2
). If the derivative
is identically zero, then f
(c) = 0 so f(x
2
) f(x
1
) = 0 so f(x
1
) = f(x
2
).
Theorem 22.7 (Second Derivative Test). If f
(x) > 0 for all x in

some open interval I, and f
(c) = 0 at some c in I, then f(x) assumes

its minimum on I at c. Similarly, if f
(x) < 0 for all x in some open

interval I, and f
(c) = 0 at some c in I, then f(x) assumes its maximum

on I at c.
Proof: Let a and b be the endpoints of the interval so a < c < b. Assume
f
(x) > 0 for a < x < b. Then f
(x) is increasing on the interval. But f
(c) = 0
so f
(x) < 0 for a < x < c and f
(x) > 0 for c < x < b, i.e. f(x) is decreasing

for a < x < c and increasing for c < c < a. Hence f(x) is smallest when x = c.
Remark 22.8. The proof shows more generally that if f
(x) < 0 for a < x < c

and f
(x) > 0 for c < x < b then f(x) has a minimum at c on the interval
a < x < b.
77
Denition 22.9. A function y = f(x) is concave up on an interval i the sec-
ond derivative f
(x) is positive at every point x in the interval. The derivative

of f
is f
so, by the Monotonicity Theorem, the derivative f
of a concave up
function is increasing. Similarly, the function is concave down i its second
derivative is negative. For example, the function f(x) = x
2
is concave up on
any interval and the function g(x) = x
2
is concave down.
Theorem 22.10 (Secant Concavity Theorem). Suppose that f(x)
is concave up on the interval a x b and dene
W(x) = f(a) +
_
f(b) f(a)
b a
_
(x a)
so that the graph y = W(x) is the secant line joining the points (a, f(a))
and (b, f(b)). Then the graph y = f(x) lies below the graph of the secant
line, i.e.
f(x) < W(x).
for a < x < b. Similarly, if f(x) is concave down, the graph of the
function lies above the graph of the secant line.
Proof: Dene g(x) = f(x) W(x) as in the proof of the Mean Value Theorem.
Since W(x) is a linear function, its second derivative is zero so g
(x) = f
(x)
and g is also concave up. By the Mean Value Theorem there is a point c with
a < c < b and g
(c) = 0. Since g
is increasing this means that g
(x) < g
(c) = 0
for a < x < c and g
(x) > 0 for c < x < b. Hence g is decreasing on the interval

a < x < c and increasing on the interval c < x < b. Hence g(a) > g(x) for
a < x < c and g(x) < g(b) for c < x < b, i.e. g(x) < 0 for a < x < b. As
g = f W we get f(x) < W(x) for a < x < b as required. If f is concave down,
then f is concave up, so f < W, and hence W < f.
78
(a,f(a))
y=W(x)
(b,f(b))
y=L(x)
(c,f(c))
Figure 1: The Secant and Tangent Concavity Theorems
Theorem 22.11 (Tangent Concavity Theorem). Suppose that f(x)
is concave up on the interval a x b and let c be an interior point of
that interval, i.e. a < c < b. Dene
L(x) = f(c) +f
(c)(x c)
so that the graph y = L(x) is the tangent line to the graph y = f(x) at
the point (c, f(c)). Then the graph y = f(x) lies above the graph of the
tangent line, i.e.
L(x) f(x)
for a x b. Similarly if f(x) is concave down, the graph of the
function lies below the graph of the tangent line.
Proof: Consider the function g(x) = f(x) L(x). As the function L(x) is
linear, its second derivative is zero, so g
(x) = f
(x) and g is also concave up.

Moreover f
(c) = L
(c) = the slope of the tangent line at c, so g
(c) = 0. Since
g
is increasing this means that g
(x) < g
(c) = 0 for a < x < c and g
(x) > 0
for c < x < b. Hence g is decreasing on the interval a < x < c and increasing on
the interval c < x < b. Hence g(x) > g(c) = 0 for a < x < c and 0 = g(c) < g(x)
for c < x < b, i.e. g(x) > 0 for a < x < b. As g = f L we get L(x) < f(x)
for a < x < b as required. If f is concave down, then f is concave up, so
L < f, and hence f < L.
Remark 22.12. When I learned calculus the terms convex and concave were
used; I could never remember which was which. Small wonder. In common
parlance the concave side of a curve is a convex set. The present terminology is
better because of the following dumb poem:
79
Concave up is like a cup , Concave down is like a frown .
Example 22.13. If y = x
3
, then d
2
y/dx
2
> 0 for x > 0 so by the Secant and
Tangent Concavity Theorems it satises the following inequalities:
x
3
< a
3
+
b
3
a
3
b a
(x a) for 0 < a < x < b <
and
c
3
+ 3c
2
(x c) < x
3
for 0 < c, x < , x ,= c.
Similarly, the function d
2
y/dx
2
< 0 for x < 0 so
a
3
+
b
3
a
3
b a
(x a) < x
3
for < a < x < b < 0
and
x
3
< c
3
+ 3c
2
(x c) for < c, x < 0, x ,= c.
To make sure you appreciate this, plug in a few particular values for a, b, c and
x, say a = 1, b = 4, c = 2, x = 3, and evaluate both sides of the inequality.
Exercises
3
+ 3x
2
+ 3x + 5. Does f have an inverse? How
do you know? If f has an inverse, determine f
1
(5),
_
f
1
_
(5), f
1
(12), and
_
f
1
_
(12). Hint: For what values of x is f
(x) < 0?
Exercise 22.15. True or false?
T F. An increasing function has an inverse.
12
T F. A decreasing function has an inverse.
T F. If a continuous function has an inverse it must be monotonic. (Hint:
Intermediate Value Theorem.)
T F. If a function has an inverse it must be monotonic.
Exercise 22.16. On what intervals is the function y = x
3
x increasing?
Decreasing? Concave up? Concave down?
Exercise 22.17. Repeat the previous exercise for the following functions:
(a) y = x
4
x
2
(b) y = 1/(1 +x
2
) (c) y = (1 +x)/(1 x)
(d) y = e
x
(e) y = ln(x) (f) y = 2
x
(g) y = sin(x) (h) y = cos(x) (i) y = tan(x)
12
i.e. satises the horizontal line test.
80
Exercise 22.18. Write out the inequalities asserted by the Secant and Tangent
Concavity Theorems for each of the functions y = f(x) of the previous exer-
cise and each interval on which the function does not change concavity. (See
Example 22.13.)
Exercise 22.19. Find a point on the curve y = x
3
where the tangent is parallel
to the chord joining (1, 1) and (3, 27). What theorem does this illustrate?
23 Curve Plotting
Denition 23.1. A point on a curve where the concavity changes is called a
point of inection. On one side of a point of inection the tangent line is
below the curve, on the other side it is above. At the point of inection the
tangent line crosses the curve.
Theorem 23.2. If the curve is the graph y = f(x) of a twice dierentiable
function, the second derivative f
(x) vanishes at a point of inection.

Denition 23.3. The line y = b is a horizontal asymptote of the function
y = f(x) i either lim
x
f(x) = b or lim
x
f(x) = b or both. The line x = a
is a vertical asymptote of the function y = f(x) i either lim
xa+
f(x) =
or lim
xa
f(x) = or both.
Example 23.4. (i) The line y = 0 is a horizontal asymptote for the function
f(x) = 1/(1 +x
2
).
(ii) The lines y = /2 is are horizontal asymptotes for the function f(x) =
tan
1
(x).
(iii) The line x = 3 is a vertical asymptote for the function f(x) = 1/(x 3)
2
.
(iv) The line x = 2 is a vertical asymptote for the function f(x) = 1/(x 3).
(v) The lines x = (n +

2
) are the vertical asymptotes for the function f(x) =
tanx.
23.5. To graph y = f(x) on a < x < b proceed as follows:
(1) nd the interesting values of x: critical points, inection points, vertical
asymptotes.
(2) nd the value of y at each interesting point:
At a critical point or inection point x, nd f(x).
At a vertical asymptote x = c nd both one sided limits
f(c) = lim
xc
f(x), and f(c+) = lim
xc+
f(x).
(These values are .)
81
At the endpoints a and b nd
f(a+) = lim
xa+
f(x), and f(b) = lim
xb
f(x).
(Usually a = and b = +.)
(3) Make a table. On each interval bounded by a pair of adjacent interesting
points, nd the sign of f
(x) and the sign of f
(x). (Now you know on

which intervals f is increasing and on which intervals f is concave up.)
(4) Draw the graph.
Example 23.6. we graph y = x
3
3x. The formulas for the rst derivative
y
= dy/dx and y
= d
2
y/dx
2
are
y
= 3x
2
3, y
= 6x.
Here is a table showing the intervals on which f is increasing, decreasing, concave
up, concave down:
x 1 0 1
y 2 3 2
y
+ + + 0 0 + + +
y
0 + + + + + + +
Figure 23 shows the graph and the tangent line at the point of inection.
Example 23.7. We graph y = x/(1x
2
). The formulas for the rst derivative
y
= dy/dx and y
= d
2
y/dx
2
are
y
=
1 +x
2
(1 x
2
)
2
, y
=
2x(3 +x
2
)
(1 x
2
)
3
.
The x-axis is a horizontal asymptote at both ends as
lim
x
x
1 x
2
= 0, lim
x
x
1 x
2
= 0,
and the lines x = 1 and x = 1 are vertical asymptotes with
lim
x1
x
1 x
2
= , lim
x1+
x
1 x
2
= ,
lim
x1
x
1 x
2
= , lim
x1+
x
1 x
2
= .
Here is a table showing the intervals on which f is increasing, decreasing, concave
up, concave down:
x 1 1+ 0 1 1+
y 0 0 0
y
+ + + + + + + + + + + + +
y
+ + + 0 + + +
Figure 23 shows the graph.
82
y = x
3
x y = 1/(1 x
2
)
Figure 2: Two Graphs
Exercises
Exercise 23.8. Draw the graph of a function y = f(x) having all the following
properties:
a) horizontal asymptote y = 1/2 (as x ).
b) horizontal asymptote y = 1 (as x +),
c) vertical asymptotes x = 1 and x = 1/2.
d) continuous and decreasing on the interval (, 1).
e) continuous and increasing on the interval (1, 1/2).
f ) continuous and increasing on the interval (1/2, ).
Exercise 23.9. Determine where the curve y = x
4
4x
3
+ 1 is increasing
or decreasing, concave up or concave down. Where are its critical points and
inection points? Draw the graph.
Exercise 23.10. Describe the horizontal and vertical asymptotes (if any) of
y =
x
2
1x
2
.
Exercise 23.11. Find the intervals on which the curve y = x
3
3x is increasing,
decreasing, concave up, or concave down. Sketch the curve showing inection
points and local maxima and minima.
Exercise 23.12. Graph y = x + 1/x. The graph approaches what line as
x ?
83
Exercise 23.13. Graph y =
1
x
2
+ 1
.
x
x
2
+ 1
.
1
x
2
1
.
x
x
2
1
.
24 Max Min Word Problems
Example 24.1. We nd the points on the curve y = x
2
which are closest to
the point (0, b) for various values of b. The distance is
r =
_
(x 0)
2
+ (y b)
2
=
_
x
2
+ (x
2
b)
2
=
_
y + (y b)
2
.
If r attains its minimum at some point, then r
2
also attains its minimum at
that same point so we minimize r
2
. The algebra is slightly simpler if we use y
rather than x as a parameter but we must remember that, by symmetry, every
positive value of y determines two points (
y, y) and (
y, y) on the curve and

that the domain of the function r
2
as a function of y has an endpoint at y = 0.
The derivative of r
2
is
d(r
2
)
dy
= 1 + 2(y b)
which vanishes when y = b 1/2. If b < 1/2, then b 1/2 < 0 so the derivative
never vanishes and the minimum must occur at the endpoint y = 0 so the closest
point is (0, 0) Since the second derivative is positive there is a unique minimum
if b > 1/2. In this case there are two closest points: (
_
b 1/2, b 1/2).
Example 24.2. Suppose the river is bounded by the horizontal lines y = 0 (the
x-axis) and y = a. A man wants to go from A(0, a) to B(b, 0) as fast as possible
by rowing to P(x, 0) and running along the shore to B. He rows at speed w and
runs at speed v so the total time is
T(x) =
x
2
+a
2
w
+
b x
v
.
s
A
s
X
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H s
B
r
b
a
x
84
The function T(x) is dened for all x but we seek its minimum on 0 x b.
(Obviously T(x) < T(0) for x < 0, and for x > b the formula is not correct: one
should replace b x by [b x[.) The derivative is
T
(x) =
x
w
x
2
+a
2
1
v
which vanishes at
x =
wa
v
2
w
2
.
(This holds if v w; otherwise T
(x) doesnt vanish anywhere.) Now

T
(x) =
a
2
w(x
2
+a
2
)
3/2
> 0
so T(x) is concave up on < x < so has a unique minimum at x = x. If
x b and the minimum occurs at the end point x = b; otherwise the minimum
occurs at x = x. The inequality x < b is x
2
< b
2
. From the formula for x this is
w
2
(a
2
+b
2
) < v
2
b
2
.
If a = 3, b = 8, w = 6, v = 8, then w
2
(a
2
+ b
2
) = 36 73 < 64 64 = v
2
b
2
and
the min occurs at x = x. If a = 5, b = 5, w = 6, v = 8, then w
2
(a
2
+ b
2
) =
36 50 > 64 25 = v
2
b
2
and the min occurs at x = b.
Example 24.3. Here is a similar problem. A woman wants to cross a lake
from point A to point C directly opposite. She rows in a straight line to a point
B and then runs along the shore. Let = CAB so 2 = COB where O
is the center. Now ABC is a right triangle (it is inscribed in a semicircle) so
(assuming the radius of the lake is a) the time is
T() =
2a cos
w
+
2a
v
.
for /2 0. (Again w is the rowing speed and v is the running speed.)
Now
T
() =
2a sin
w
+
2a
v
which vanishes at = sin
1
(v/2w). But T
() = (2a/w) cos < 0 so the

interior critical point is a maximum, not a minimum. The minimum occurs at
one of the two end points. At the end points
T(/2) =
2a
v
, T(0) =
2a
w
,
and the minimum is the smaller of these two. Thus, to minimize the time, she
should run around the lake if w < v/, and row straight across if w > v/.
85
Exercises
Exercise 24.4. Find the local maxima and minima of f(x) = (x1)
3
(x+1)
2
.
Exercise 24.5. If we use x rather than y as the parameter in Example 24.1 we
have to minimize the function
r
2
= f(x) = x
2
+ (x
2
b)
2
over the interval < x < . Draw the graph of the function f(x) for b = 1/4
and for b = 1.
Exercise 24.6. Show that f(x) = x + 1/x has a local maximum and a local
minimum, but the value at the local maximum is less than the value at the local
minimum.
Exercise 24.7. A train is moving along the curve y = x
2
+ 2. A girl is at the
point (3, 2) At what point will the train be at when the girl and the train are
closest? Hint: You will have to solve a cubic equation, but the numbers have
been chosen so there is an obvious root.
Exercise 24.8. Find the local maxima and minima of f(x) = x + 2 sinx in
[0, 2].
Exercise 24.9. A 12 12 piece of sheet metal is to be cut along the solid lines
to form a closed rectangular box. (The bends are indicated by the dashed lines.)
Find the dimensions of the box of largest volume which can be constructed in
this manner. (See Thomas page 278.)
Exercise 24.10. A cylindrical can is to be made to hold 1 liter of oil. Find the
dimensions of the can which will minimize the cost of the metal to make the
can.
Exercise 24.11. Find the dimensions of the rectangle of area 96 cm
2
which
has minimum perimeter. What is this minimum perimeter?
Exercise 24.12. Find the largest possible volume of a right circular cylinder
inscribed in a hemisphere of radius r.
Exercise 24.13. Find the dimensions of the rectangle of area 96 cm
2
which
has minimum perimeter. What is this minimum perimeter?
Exercise 24.14. Given ABC and let AH be an altitude to side BC. If AB
and CH both have length 1, nd the length x of BH which will maximize the
area of ABC.
Exercise 24.15. The following questions all refer to the curve x
2
xy+y
2
= 9.
1. Find a formula for
dy
dx
(in terms of x and y).
86
2. Find all points (x, y) on the curve where x = 0 and nd
dy
dx
at each of
them.
3. Find all points (x, y) on the curve x
2
xy + y
2
= 9 where x = 3 and
nd
dy
dx
at each of them.
4. Find all points on the curve x
2
xy + y
2
= 9 where the tangent is hori-
zontal; then nd all the points where it is vertical.
5. Find all points of the curve that are closest to and farthest from the
origin. Hint: The square of the distance to the origin is r
2
= x
2
+y
2
and
r
2
= xy + 9 on the curve. (Finding where dr/dx = 0 requires solving two
quadratic equations in two unknowns, but the algebra is not dicult if
you dont make a mistake.)
6. Sketch the curve. Hint: The curve is an ellipse centered at the origin.
25 Exponential Growth
25.1. Suppose that a quantity N grows exponentially in time. This means
that the value of N at time t is given by a formula of form
N = N
0
a
t
where a and N
0
are constants. Notice that N = N
0
when t = 0 so that N
0
is
the initial value of N. The derivative is
dN
dt
= N
0
a
t
lna = kN, k = lna.
Consider a tiny time interval from t to t+t. The change in N over this interval
is
N = N(t + t) N(t).
The average growth rate on this interval is N/t and the instantaneous growth
rate at time t is the derivative dN/dt. The average percentage change in this
interval is the change N divided by the amount N present.
13
The average
percentage rate in this interval is the percentage change divided by the change
in time t. The instantaneous percentage rate is the limit as t 0. Thus
13
Per means divide and cent means 100. Thus % means 1/100. For example, 0.04=4%.
87
When a quantity N grows exponentially, its percentage growth rate
k =
dN/dt
N
is a positive constant, i.e. the growth rate is proportional to the amount
present:
dN
dt
= kN
The constant of proportionality is k = lna.
(When a < 1 the value of N decreases as t increases and we say that N decays
exponentially.)
25.2. The doubling time of a quantity N = N
0
e
kt
which is increasing expo-
nentialy is the time t such that N = 2N
0
. Since
2N
0
= N
0
e
kt
= 2 = e
kt
= ln2 = kt
the doubling time is t = (ln2)/k. Similarly, the half life of a quantity N =
N
0
e
kt
which is decreasing exponentialy is the time t such that N = N
0
/2, i.e.
t = (ln2)/k.
25.3. Many phenomena are governed by exponential growth laws, To name a
few:
1. Money invested in a bank account grows exponentially at 6% per year (or
whatever the interest rate is).
2. The population of the world grows exponentially at 1.5% per year. (This
will not continue forever.)
3. The amount of radioactivity in a radioactive material decays exponentially
(at a rate that depends on the material). For example carbon-14 decays
at 0.012% per year and polonium-210 decays at 0.495% per day.
4. Some chemical reactions occur at a rate proportional to the amount of the
chemical present.
25.4. Here is how a few dierent banks compound interest.
(i) The Alaska Bank pays interest as follows. Every year it looks at the balance
in an account and adds 6%. Thus an initial deposit of B
0
dollars grows to
B = B
0
(1.06)
t
dollars after t years (assuming no other deposits are made during this period.)
88
(ii) The Montana Bank pays interest as follows. Every month it looks at the
balance in an account and adds 6/12% = 0.5%. Thus an initial deposit of B
0
dollars grows to
B = B
0
(1.005)
12t
dollars after 12t months = t years (assuming no other deposits are made during
this period.)
(ii) The Delaware Bank pays interest as follows. Every day it looks at the
balance in an account and adds 6/365%. Thus an initial deposit of B
0
dollars
grows to
B = B
0
_
1 +
0.06
365
_
365t
dollars after 365t days = t years (assuming no other deposits are made during
this period.)
Theorem 25.5. lim
m
_
1 +
r
m
_
mt
= e
rt
.
Proof. Using the law that ln(b
c
) = c lnb and changing the dummy variable
shows that
lim
m
ln
_
1 +
r
m
_
mt
= lim
m
mt ln
_
1 +
r
m
_
= lim
h0+
t
h
ln(1 +rh)
so it is enough to prove that
lim
h0+
t
h
ln(1 +rh) = rt.
Now let f(x) = ln(1 +rx) and note that the last limit is the dierence quotient
(f(h) f(0))/h. Thus the limit is tf
(0) By the Chain Rule, f
(x) = r/(1 +rx)

so f
(0) = r so so the limit is rt. Now exponentiate and use the continuity of
the exponential.
25.6. The general formula for the balance B in a bank account after t years
if the balance is initially B
0
, the interest rate is r per year, and the interest is
compounded m times per year is
B = B
0
_
1 +
r
m
_
mt
.
(The formula is used even if t is not an integer.) When m becomes innite we
say that the interest is compounded continuously and the theorem says that
B = B
0
e
rt
.
In all cases the formula is of the form
B = B
0
a
t
89
where a = (1 + r/m)
m
if the compounding period is 1/m years and a = e
r
if the compounding is continuous. The following table shows the values of
a
t
_
1 +
r
m
_
mt
for t = 2, r = 0.05, n = mt, and various values of the number m
of compounding periods per year.
m
_
1 +
r
m
_
mt
1 1.102500000000000
12 1.104941335558328
52 1.105117820169223
365 1.105163349128883
1.105170918075648
Exercises
Exercise 25.7. Find k such that dB/dt = kB if B = B
0
_
1 +
r
m
_
mt
and if
B = B
0
e
rt
.
Exercise 25.8. Polonium-210 has a half life of 140 days. (a) If a sample has a
mass of 200 mg nd a formula for the mass that remains after t days. (b) Find
the mass after 100 days. (c) When will the mass be reduced to 10 mg? (d) Sketch
the graph of the mass as a function of time.
Exercise 25.9. The number N of bacteria in laboratory container t hours after
the start of an experiment is
N = 1000 2
t
where t is the number of hours since the beginning of the experiment. (a) How
many bacteria are present at time t = 0? (b) When will the number N be twice
that initial amount? Four times? Eight times? (c) At what (instantaneous)
rate is N increasing at time t? (d) At what (instantaneous) percentage rate is
N increasing at time t?
Exercise 25.10. After 3 days a sample of radon-222 decayed to 58% of its
original amount. (a) What is the half life of radon-222? (b) How long would it
take the sample to decay to 10% of its original amount?
Exercise 25.11. Radiocarbon dating works on the principle that
14
C decays
according to radioactive decay with a half life of 5730 years. A parchment
fragment was discovered that had about 74% as much
14
C as does plant material
on earth today. Estimate the age of the parchment.
90
Exercise 25.12. The population of the country of Slobia grows exponentially.
(a) If its population in the year 1990 was 1,990,000 and its population in the
year 2000 was 2,000,000 what will be its population in the year 2010? (b) How
long will it take the population to double?
Exercise 25.13. (Archer Daniels Midland) According to a recent TV com-
mercial from the Archer Daniels Midland Corporation, in fty years the world
will have to set ten billion places at the table. Another ADM commercial
says when this baby is old enough to vote, the world will have a billion more
mouths to feed. What is the present population of the world and how fast is
it increasing?
14
26 Indeterminate Forms (lHopitals Rule)
26.1. Often when we try to evaluate a derivative by just plugging in we get
nonsense like
0
0
,

, 0
0
, 1
,
0
.
these are called indeterminate forms. A special technique called lHopitals
Rule can be used to evaluate the limit in this case. We will de-emphasize this
technique because it encourages blind calculation and leads students to forget
what a limit is. However, we will touch on it briey. Here is the simplest case.
lHopitals Rule. Suppose that the functions f(x) and g(x) are dierentiable
and that f(a) = g(a) = 0. Then
lim
xa
f(x)
g(x)
= lim
xa
f
(x)
g
(x)
if the limit on the right exists.
We wont give a careful proof, but roughly speaking, the reason why the rule is
true is that
f(x)
g(x)
=
f(x) f(a)
x a
g(x) g(a)
x a
(a)
g
(a)
when x a.
26.2. lHopitals Rule also works for the indeterminate form /: If
lim
xa
f(x) = lim
xa
g(x) =
14
This problem cannot be solved algebraically. You can however nd two equations relating
the present world population and the rate of increase, eliminate the latter to get a single
equation in the former, and draw a graph with a computer to estimate the solution.
91
lim
xa
f(x)
g(x)
= lim
xa
f
(x)
g
(x)
if the limit on the right exists. Both forms 0/0 and / also work when
a = . Moreover, lHopitals Rule holds even when f(x) and g(x) are not
dened at x = a: it is enough to assume lim
xa
f(x) = lim
xa
g(x) = 0.
26.3. Warning. lHopitals Rule doesnt work unless plugging in gives an
indeterminate form. For example, with f(x) = x
3
, g(x) = x
2
, and a = 1 we
have
lim
xa
f(x)
g(x)
= lim
x1
x
3
x
2
= 1, but lim
xa
f
(x)
g
(x)
= lim
x1
3x
2
2x
=
3
2
.
Example 26.4. Exponentials defeat powers. Setting x = in x/e
x
gives
the indeterminate form / so by lHopitals Rule
lim
x
x
e
x
= lim
x
1
e
x
= 0.
Applying this argument twice gives
lim
x
x
2
e
x
= lim
x
2x
e
x
= lim
x
2
e
x
= 0.
Applying this argument n times gives
lim
x
x
n
e
x
= lim
x
nx
n1
e
x
= = lim
x
n!
e
x
= 0.
Example 26.5. Powers defeat logarithms. Setting x = in (lnx)/x gives
the indeterminate form / so by lHopitals Rule
lim
x
lnx
x
= lim
x
1/x
1
= 0.
Example 26.6. Other indeterminate forms can be treated by performing some
algebra rst. For the indeterminate form 0 try dividing by the second factor
to get the indeterminate form /. For example,
lim
x0+
xlnx = lim
x0+
lnx
1/x
= lim
x0+
1/x
1/x
2
= lim
x0+
x
2
x
= 0.
Example 26.7. For the indeterminate forms 0
0
or 1
try taking the logarithm

to get 0 as in Example 26.6. Then exponentiate the answer to solve the
original problem. For example,
ln lim
x0+
x
x
= lim
x0+
ln(x
x
) = lim
x0+
xlnx = 0
so
lim
x0+
x
x
= exp
_
ln lim
x0+
x
x
_
= exp(0) = 1.
92
Exercises
Exercise 26.8. Illustrate the indeterminate form / by nding four large
numbers A, B, C, D, all larger than one million, with A/B = 0.37 and C/D =
458.
Exercise 26.9. (i) Evaluate A/B when A = 0.01 and B = 0.00001. (ii) Evalu-
ate A/B when A = 0.00001 and B = 0.01. (iii) Which indeterminate form does
this illustrate?
Exercise 26.10. (i) Evaluate A
B
when A = 1.01 and B = 1000. (ii) Evaluate
A
B
when A = 1.00001 and B = 1000. (iii) Which indeterminate form does this
illustrate? (This problem requires a calculator.)
Exercise 26.11. (i) Evaluate A
B
when A = 10
1000
and B = 0.001. (ii) Eval-
uate A
B
when A = 10
1000
and B = 0.01. (iii) Which indeterminate form does
this illustrate? (This problem does not require a calculator.)
Exercise 26.12. Evaluate each of the following.
(i) lim
x
x3
x
(ii) lim
x
x3
x
(iii) lim
x0
x3
x
(iv) lim
x
3
x
x
(v) lim
x
3
x
x
(vi) lim
x
x
3
x
(vii) lim
x
x
3
x
(viii) lim
x
lnx
x
27 Antiderivatives
Denition 27.1. An antiderivative of a function f is a function F such that
F
= f. Note that
If F is an antidericative of f so is F + c for any constant c. (Proof: the
derivative of a constant is zero.)
Any two antiderivatives of f dier by a constant. (Proof: If the derivative
of a function is zero then that function is constant. See Theorem 22.6)
Remark 27.2. Antidierentiation suces to solve dierential equation where
only the derivative of the unknown function appears (e.g. constant gravity),
but not other dierential equations where both the unknown function and its
derivative appear. Later you will learn that antiderivatives can be used to
compute areas.
Example 27.3. If the position of a particle at time t is denoted by y then the
quantities
v =
dy
dt
, a =
d
2
y
dt
2
,
93
are called respectively the velocity and acceleration of the particle. According
to Newton, if a ball is thrown into the air its height y at time t satises
d
2
y
dt
2
= g
where g = 32ft/sec
2
. Hence there is a constant v
0
(the initial velocity) such
that
dy
dt
= gt +v
0
, v
0
=
dy
dt
t=0
and there is a constant y
0
(the initial height) such that
y =
gt
2
2
+v
0
t +y
0
, y
0
= y
t=0
.
Exercises
Exercise 27.4. Find the most general antiderivative of each of the following
functions. Check your answer by dierentiating.
(a) f(x) = x
3
+ 3x
2
+ 7 +x
2
(b) f(t) = sin(t)
(c) f(x) =
x
4
+ 3x
3
+ 7x +x
1
x
(d) f() = sec
2
()
(e) f(x) = e
x
(f) f(x) = x + 1/x.
Exercise 27.5. Find F(x) if F
(x) = x
3
+ 3x
2
+ 7 +x
2
and F(1) = 12.
Exercise 27.6. Find F(t) if F
(t) = sin(t) and F(0) = 7.

Exercise 27.7. Find F() for /2 < < /2 if F
() = sec
2
() and F(0) = 7.
Exercise 27.8. A ball is thrown upward with an initial speed of 48 ft/sec from
a roof which is 432 feet above the ground. Find its height t seconds later. When
does it reach its maximum height. What is its maximum height? When does it
hit the ground? How fast is it going when it hits the ground?
Exercise 27.9. One second after the ball in the previous problem is thrown
another ball is thrown upward with an initial speed of 24 ft/sec. Are the balls
ever at the same height? (Before they hit the ground of course.)
Exercise 27.10. A graph y = f(x) passes through the point (1, 6) and its slope
at the point (x, f(x)) is 2x + 1. What is f(2)?
Exercise 27.11. Find a function f(x) such that f
(x) = x
3
and the line x+y =
0 is tangent to the graph of f.
Exercise 27.12. A car traveling 50 miles per hour when the brakes are applied
producing a a constant deceleration of 40 ft/sec
2
. What is the distance covered
before the car comes to a stop. Hint: One mile= 5280 feet.
Exercise 27.13. A stone is dropped from a roof and hits the ground with a
velocity of 120 feet per second. What is the height of the roof?
94
28 Additional Problems
Exercise 28.1. For the function y = xe
x
make a table showing all horizontal
asymptotes, all vertical asymptotes, all local extrema, all points of inection,
intervals on which the function is increasing, intervals on which the function
is decreasing, intervals on which the function is concave up, and intervals on
which the function is concave down. Draw a graph of the function and on the
is graph draw the tangent line at each point of inection.
95
Chapter V
Integration
29 The Denite Integral
The denite integral
_
b
a
f(x) dx of a nonnegative function f(x) is the area
of the region a x b and 0 y f(x). The precise denition involves
approximating this region by skinny rectangles. The sum of the areas of these
rectangles is called a Riemman
15
sum and the denite integral is the limit of
the Riemann sums as the rectangles become skinnier and skinnier. To say this
precisely we rst need some terminology.
29.1. Sigma Notation. The notation

n
j=m
a
j
is short for the sum of the
numbers a
m
, a
m+1
, . . . , a
n
, i.e.
n
j=m
a
j
= a
m
+a
m+1
+ +a
n
.
For example,

5
j=2
j
2
= 2
2
+ 3
2
+ 4
2
+ 5
2
= 54. In the notation

n
j=m
a
j
the integers m and n are called the limits of summation, the expression a
j
is
called the jth summand, and the variable j is called the index of summation.
The index of summation is a dummy variable (see 7.7), i.e.
n
j=m
a
j
=
n
i=m
a
i
but the limits of summation are free variables (see 7.8), i.e.
n
j=m
a
j
,=
q
j=p
a
j
(usually).
Here are some obvious laws.
(Constants Law)
n
j=m
1 = n m+ 1.
(Linearity Law I)
n
j=m
(a
j
b
j
) =
_
_
n
j=m
a
j
_
_
_
_
n
j=m
b
j
_
_
.
(Linearity Law II)
n
j=m
ka
j
= k
n
j=m
a
j
.
15
pronounced Reemann
96
(Additivity Law)
n
j=m
a
j
=
_
_
p
j=m
a
j
_
_
+
_
_
n
j=p+1
a
j
_
_
for m p n.
(Order Law)
n
j=m
a
j

n
j=m
b
j
if a
j
b
j
for j = m, m+ 1, . . . , n.
29.2. Recall that [a, b] denotes the closed interval a x b with endpoints
a and b. A partition of the interval [a, b] is a nite sequence
P = (x
0
, x
1
, . . . , x
n
)
such that
a = x
0
< x
1
< x
2
< < x
n1
< x
n
= b.
The partition divides the interval [a, b] into subintervals
[x
0
, x
1
], [x
1
, x
2
], . . . , [x
n1
, x
n
].
The length of the kth interval is
x
j
= x
j
x
j1
.
The mesh |P| of the partition P is the length of the largest subiniterval:
|P| = maxx
1
, x
2
, . . . , x
n
.
When |P| 0 all of the subintervals are small and there are a large number of
them (i.e. n ). A Riemann partition of the interval [a, b] is a partition
P equipped with additional numbers
C = (c
1
, c
2
, . . . , c
n
)
such that c
j
lies in the jth interval, i.e.
x
j1
c
j
x
j
for j 1, 2, . . . , n.
Denition 29.3. Let f(x) be a function which is continuous on the interval
[a, b] and (P, C) be a Riemann partition of [a, b] as in 29.2. The number
S(f, P, C) =
n
j=1
f(c
j
)(x
j
x
j1
)
is called the Riemann sum of f for (P, C). In the notation of 29.2 this can
be written
S(f, P, C) =
n
j=1
f(c
j
)x
j
.
97
Denition 29.4. The denite integral of f(x) from x = a to x = b
is the limit
_
b
a
f(x) dx = lim
P0
S(f, P, C)
of the Riemann sums as the mesh |P| of the partition goes to 0. In
other words
n
j=1
f(c
j
)x
j

_
b
a
f(x) dx
when |P| 0. The function f is called integrable on the interval [a, b]
i this limit exists.
29.5. In the notation
_
b
a
f(x) dx the numbers a and b are called the limits of
integration, the function f is called the integrand, and variable x is called
the variable of integration. The variable of integration is a dummy variable:
_
b
a
f(x) dx =
_
b
a
f(t) dt
but the limits of summation are free variables:
_
b
a
f(x) dx ,=
_
q
p
f(x) dx (usually).
29.6. Integration Laws. Here are some important laws which which inte-
grable functions f and g satisfy. They are all proved the same way, namely by
showing that an analogous law holds for Riemann sums and then passing to the
limit.
Constants Law. The integral of a constant function is
_
b
a
k dx = k(b a).
This is because for any partition as in 29.2 we have
n
j=0
kx
j
= k
_
(x
1
x
0
) + (x
2
x
1
) + + (x
n
x
n1
)
_
= c(b a).
98
(This is a collapsing sum: the x
j
in each summand cancels the x
j
in the next
one. The two terms which dont cancel are x
0
= a from the rst summand
x
n
= b from the last.)
Linearity Law. The denite integral is linear in the integrand. This
means that
_
b
a
_
f(x) +g(x)
_
dx =
_
b
a
f(x) dx +
_
b
a
g(x) dx,
and, for any constant k,
_
b
a
kf(x) dx = k
_
b
a
f(x) dx.
This is because for any Riemann partition as in 29.2 we have
n
j=1
_
f(c
j
) +g(c
j
)
_
x
j
=
_
_
n
j=1
f(c
j
)x
j
_
_
+
_
_
n
j=1
g(c
j
)x
j
_
_
and
n
j=1
kf(c
j
)x
j
= k
n
j=1
f(c
j
)x
j
.
Additivity Law. If a < c < b, then
_
b
a
f(x) dx =
_
c
a
f(x) dx +
_
b
c
f(x) dx.
This is because any Riemann partition of [a, b] as in 29.2 with x
m
= c we have
n
j=1
f(c
j
)x
j
=
_
_
m
j=1
f(c
j
)x
j
_
_
+
_
_
n
j=m+1
f(c
j
)x
j
_
_
Remark 29.7. It is convenient to dene
_
a
b
f(x) dx =
_
b
a
f(x) dx.
Then the Additivity Law holds for all a, b, c not just a < c < b.
99
Order Law. The denite integral preserves order, i.e. if f(x) g(x)
for all x, then
_
b
a
f(x) dx
_
b
a
g(x) dx
This is because any Riemann partition as in 29.2 we have f(c
j
) g(c
j
) so
n
j=1
f(c
j
)x
j

n
j=1
g(c
j
)x
j
.
Key Estimate.
_
min
axb
f(x)
_
(b a)
_
b
a
f(x) dx
_
max
axb
f(x)
_
(b a)
Since min
axb
f(x) f(x) max
axb
f(x) the Key Estimate follows imme-
diately the Constants Law and the Order Law.
Theorem 29.8. A continuous function is integrable.
29.9. This theorem is normally proved in more advanced courses like Math 521,
but we can indicate the idea. Suppose that f is continuous on [a, b]. A partition
P determines two Riemann sums as follows. On each interval x
j1
x x
j
the function f assumes its maximum at some point c
j
and its minimum at some
other point c
j
. (Usually c
j
will be one of the endpoints and c
j
will be the other.)
The Riemann sum
U(f, P) =
n
j=1
f(c
j
)x
j
is called the upper sum and the Riemann sum
L(f, P) =
n
j=1
f(c
j
)x
j
100
is called the lower sum. (See Thomas page 345; there are pictures of upper
sums and lower sums in Exercise 77-80 in Thomas pages 354-5.) By the choice
of c
j
and c
j
we have f(c
j
) f(x) f(c
j
) for x
j1
x x
j
. In particular,
f(c
j
) f(c
j
) f(c
j
)
for any Riemann partition (P, C) and hence L(f, P) S(f, P, C) U(f, P),
i.e.
n
j=1
f(c
j
)x
j

n
j=1
f(c
j
)x
j

n
j=1
f(c
j
)x
j
.
The proof of Theorem 29.8 rests on the following fact:
Theorem 29.10. The denite integral of a continuous function is the
unique number which lies between every lower sum and every upper sum,
i.e.
n
j=1
f(c
j
)x
j

_
b
a
f(x) dx
n
j=1
f(c
j
)x
j
This is true because when the mesh is small the lower sum and the upper sum
are approximately the same. See the pictures in Thomas pages 354-355.
Example 29.11. Figure 3 shows an upper sum and a lower sum approximating
the integral
_
3
1
x
1
dx. The partition P is
a = x
0
= 1 < x
1
= 1.5 < x
2
= 2 < x
3
= 2.5 < x
4
= 3 = b
so n = 4 and all the intervals [x
j1
, x
j
] all have the same length:
x
1
= x
2
= x
3
= x
4
= 0.5.
Since the function f(x) = x
1
is decreasing, its minimum on any interval is
assumed at the right endpoint and its maximum is assumed at the left endpoint,
i.e. c
j
= x
j
and c
j
= x
j1
. Hence the lower sum is
L =
4
j=1
x
c
j
=
0.5
1.5
+
0.5
2.0
+
0.5
2.5
+
0.5
3.0
= 0.95
and the upper sum is
U =
4
j=1
x
c
j
=
0.5
1.0
+
0.5
1.5
+
0.5
2.0
+
0.5
2.5
= 1.2833 . . .
101
Figure 3: A lower sum and an upper sum
so Theorem 29.10 tells us that
0.95
_
3
1
dx
x
1.2833 . . . .
The exact value of the integral is ln3 which, according to my calculator, is
1.098612289 . . ..
Example 29.12. The picture at the
right shows shows a lower Riemann sum
for a function which has an interior
maximum. In the rst two intervals of
the partition the minimum occurs at the
left endpoint and in the other two the
minimum occurs at the right endpoint.
Exercises
Exercise 29.13. Use Riemann sums with four intervals of length one to nd
positive numbers L and U with
3 < L
_
5
1
_
3 +
1
x
_
dx U.
Exercise 29.14. Let P denote the partition
P = (x
0
, x
1
, x
2
, x
3
, x
4
) = (0, /6, /4, /3, /2)
of the interval [0,

2
]. Find the lower sum L and the upper sum U for the function
f(x) = sin(x) on this interval and show (using a calculator) that L < 1 < U.
102
Exercise 29.15. Let f(x) = 3x + 4, a = 2, b = 5. The region a x b,
0 y f(x) is a trapezoid.
1. Draw the trapezoid and nd the area
_
b
a
f(x) dx of this trapezoid by by
elementary geometry.
2. Evaluate the approximation

6
j=1
f(c
j
)x
j
and the mesh |P| when
x
0
= 2, c
1
= 2.2, x
1
= 2.5, c
2
= 2.8, x
2
= 3.0
c
3
= 3.2, x
3
= 3.5, c
4
= 3.9, x
4
= 4.1
c
5
= 4.2, x
5
= 4.4, c
6
= 4.8, x
6
= 5.0.
3. Draw a graph illustrating the trapezoid and the six rectangles x
j1
x
x
j
, 0 y f(c
j
) whose areas f(c
j
) x
j
sum to the approximation just
computed.
4. Evaluate the approximation

8
j=1
f(c
j
)x
j
when
x
j
= a +
k
8
(b a), c
j
=
x
j1
+x
j
2
.
If you like, you may use the formula
n
j=1
j =
n(n + 1)
2
.
2
, a = 1, b = 3, and
c
j
= x
j
= a +
k
8
(b a)
for k = 0, 1, . . . , 8.
1. Draw a graph showing the graph y = f(x) for a x b and also showing
the the eight rectangles x
j1
x x
j
, 0 y f(c
j
).
2. Calculate the approximation
8
j=1
f(c
j
)x
j
.
If you like, you may use the formula
n
j=1
j
2
=
n(n + 1)(2n + 1)
6
.
3. What is 1
2
+ 2
2
+ 3
2
+ + 20
2
?
103
Exercise 29.17. Here is a question from an old 221 exam. Find a number
smaller than
_
3
1
e
x
2
dx. The answer 10
100
is correct if you can prove it.
Exercise 29.18. Redraw the graph in Example 29.12 showing the upper sum
corresponding to the partition rather than the lower sum.
Exercise 29.19. The continuous function f takes the values
f(1) = 3, f(1.6) = 6.7, f(2.8) = 11.2, f(3.3) = 9.9, f(4) = 3,
and is increasing for 1 x 2.8 and decreasing for 2.8 x 4. Find the lower
Riemann sum L for the partition
x
0
= 1 < x
1
= 1.6 < x
2
= 2.8 < x
3
= 3.3 < x
4
= 4
of the interval [1, 4]. Sketch a possible graph and also draw the area represented
by the Riemann sum. (On an exam you would be instructed to leave the addition
and multiplication undone so as to make your work easier to grade.)
Exercise 29.20. Repeat the previous exercise with the upper Riemann sum
rather than the lower.
30 The Fundamental Theorem of Calculus
Theorem 30.1 (The Fundamental Theorem). Suppose that the
function f(x) is continuous on the closed interval [a, b]. Dene a func-
tion I(x) on [a, b] by
I(x) =
_
x
a
f(t) dt.
Let F be any antiderivative of f on a x b, i.e.
F
(x) = f(x)
Then
(I) I
(x) = f(x);
and
(II)
_
b
a
f(t) dt = F(b) F(a).
104
Proof: For part (I) we compute the dierence quotient:
I(x +h) I(x)
h
=
1
h
_
_
x+h
a
f(t) dt
_
x
a
f(t) dt
_
=
1
h
_
x+h
x
f(t) dt.
By the Key Estimate in 29.6 the right hand side satises
min
xtx+h
f(t)
1
h
_
x+h
x
f(t) dt max
xtx+h
f(t)
if h > 0. The inequality also holds for h < 0 if we replace x t x + h by
x +h t x. Because the function f is continuous, the max and the min are
both close to f(x) when h is small. Hence
1
h
_
x+h
x
f(t) dt f(x)
for h 0. This proves (I).
Part (II) is an easy consequence of part (I) as follows. If I
(x) = F
(x) then
F(x) = I(x) +C for some constant C so
F(b) F(a) = I(b) I(a) =
_
b
a
f(t) dt
_
a
a
f(t) dt =
_
b
a
f(t) dt.
30.2. We often write
_
f(x) dx = F(x) +C
to mean that F(x) is an antiderivative of f(x), i.e.
_
f(x) dx = F(x) +C F
(x) = f(x).
For example,
_
x
2
dx =
x
3
3
+C.
The function
_
f(x) dx is called the indenite integral of f(x). One should
think of it as a convenient notation for antiderivatives. The +C indicates that
it should be viewed as a set of functions: one for each choice of the constant
C. When evaluating the denite integral the choice of the constant C doesnt
matter since it cancels:
_
b
a
f(x) dx = (F(b) +C) (F(a) +C) = F(b) F(a).
30.3. Another handy notation is F(x)
b
a
dened by
F(x)
b
a
= F(b) F(a).
105
With this notation the Fundamental Theorem takes the form
_
b
a
F
(x) dx = F(x)
b
a
.
For example,
_
b
a
x
3
dx =
x
4
4
b
a
=
b
4
4

a
4
4
.
Exercises
Exercise 30.4. Evaluate the following:
(i)
_
2
1
x
2
dx (ii)
_
1
0
(1 2x 3x
2
) dx
(iii)
_
2
x
dx (iv)
_
4
0
xdx
(v)
_
1
0
x
3/7
dx (vi)
_
3
1
_
1
t
2

1
t
4
_
dt
(xii)
_
2
1
x
2
+ 1
x
dx (viii)
_
4
1
(
t 2/
t) dt
(ix)
_
2
0
(x
3
1)
2
dx (x)
_
1
0
u(
u +
3
u) du
(xi)
_
1
1
(x 1)(3x + 2) dx (xii)
_
2
1
(x + 1/x)
2
dx
Exercise 30.5. (1) Water ows into a container at a rate of three gallons per
minute for two minutes, ve gallons per minute for seven minutes and eleven gal-
lons per minute for two minutes. How much water is in the container? (2) Water
ows into a container at a rate of t
2
gallons per minute for 2 3 5. How
much water is in the container?
_
x
3
+ 3x
2
+ 7 +x
2
dx.
Exercise 30.7. True or false?
_
1
1
3
t
4
dt =
1
t
3
1
1
= 1 + 1 = 0.
d
dx
_
x
0
e
t
2
dt. Hint: Do not try to evaluate the
integral.
(i)
_
3
3
(x
4
+x
2
+ 1) dx (ii)
d
dx
_
x
3
(u
4
+u
2
+ 1) du
(iii)
_
3
3
_
x
5
+ 2 dx (iv)
d
dx
_
x
3
_
u
5
+ 2 du
106
(v)
d
dx
_
sin x
3
_
u
5
+ 2 du (vi)
d
dx
_

x
5
+2
3
sinudu
(vii)
d
dx
_
3
x
_
u
5
+ 2 du (viii)
d
dx
_
3
sin x
_
u
5
+ 2 du
_
5
0
f(x) dx where f(x) is dened by
f(x) =
_
x + 2 for x < 1
3
x
for x 1
Hint: Additivity Law.
31 Averages
See Thomas page 351.
Denition 31.1. The average value of the function f on the interval
[a, b] is
f
av
=
1
b a
_
b
a
f(x) dx.
31.2. The average velocity is the average of the velocity. Remember
the trip to Milwaukee? (See 6.3.) The position (=reading on the mile post) at
time t was s = f(t). The distance
s = f(t + t) f(t)
travelled over the time interval from t to t + t is called the displacement.
The ratio
v
av
=
s
t
is called the average velocity over this time interval, while the derivative
v =
ds
dt
= lim
t0
s
t
= f
(t)
is called the instantaneous velocity at time t. Now by the Fundamental
Theorem
s = f(t + t) f(t) =
_
t+t
t
f
() d.
107
But v = f
() is the instantaneous velocity at time so the average velocity

(from 6.3) is the average of the velocity function (which we just learned), i.e.
v
av
=
s
t
=
1
t
_
t+t
t
f
() d =
1
t
_
t+t
t
v() d.
Exercises
Exercise 31.3. Find the average value of f(x) = x
p
over the interval [0, 1] for
p = 1/2, 1, 2.
Exercise 31.4. Repeat the previous exercise for the interval [0,
1
5
] and the
interval [0, 5].
Exercise 31.5. Prove that if the function f(x) is continuous on the interval
[a, b] the there is a point c in the interval [a, b] such that f(c) = the average
value of f(x) on the interval, i.e.
_
b
a
f(x) dx = f(c)(b a).
This is called the Mean Value Theorem for Integrals. Hint: Apply the
Mean Value Theorem 22.4 to F(x) =
_
x
a
f(t) dt.
32 Change of Variables
Theorem 32.1 (Change of Variables Formula). Suppose that the
function u = g(x) is continuously dierentiable on the closed interval
[a, b] and that the function f(x) is continuous of the range of g. Then
_
b
a
f(g(x))g
(x) dx =
_
g(b)
g(a)
f(u) du. ()
32.2. The Change of Variables Formula is an easy consequence of the Funda-
mental Theorem. If F(u) is an antiderivative of f(u), then the right hand side
of () is
_
g(b)
g(a)
f(u) du = F(g(b)) F(g(a)).
108
By the Chain Rule
(F g)
(x) = F
(g(x))g
(x) = f(g(x))g
(x),
i.e. (F g)(x) is an antiderivative of the integrand in (). Hence by the Funda-
mental Theorem again
_
b
a
f(g(x))g
(x) dx = (F g)(x)(b) (F g)(a) = F(g(b)) F(g(a)).

Remark 32.3. The Change of Variables Formula is the Chain Rule back-
wards.
32.4. Here is another proof of the Change of Variables Formula which is more
like the proof of the analogous formula for multiple integrals which we will
study in Math 234. For this proof we must assume that g is increasing. Choose
a partition P = (x
0
, x
1
, . . . , x
n
) of the interval [x
0
, x
n
] = [a, b]. By the Mean
Value Theorem 22.4 each interval [x
k1
, x
k
] in the partition contains a point c
k
such that
g(x
k
) g(x
k1
) = g
(c
k
)(x
k
x
k1
). (i)
The partition P together with the points c
k
determine a Riemann partition of
the interval [a, b] so by the denition of the indenite integral we have
_
b
a
f(g(x))g
(x) dx
n
k=1
f(g(c
k
))g
(c
k
)(x
k
x
k1
) (ii)
when the mesh |P| of the partition is small. Now Q =
_
g(x
0
), g(x
1
), . . . , g(x
n
)
_
is a partition of the interval [g(x
0
), g(x
n
)] = [g(a), g(b)] and g(c
k
) lies in the
interval [g(x
k1
), g(x
k
)] so
_
g(b)
g(a)
f(u) du
n
k=1
f(g(c
k
))(g(x
k
) g(x
k1
)) (iii)
By (i) the right hand sides of (ii) and (ii) are equal, and the approximate
equalities are arbitrarily accurate so the left hand sides of (ii) and (ii) are also
equal.
32.5. The dierential notation from 11.4 is very handy when computing
with the Change of Variables Formula. I like to write the formula with the
dummy variable appearing in the superscript and subscript of the integral sign
to remind myself to change the limits of integration. For example, lets evaluate
_
x=2
x=0
1 +x
2
2xdx. We use the change of variables u = 1 +x
2
. Then
du = 2xdx, x = 0 = u = 1, x = 2 = u = 5
so
_
x=2
x=0
_
1 +x
2
2xdx =
_
u=5
u=1
udu =
2u
3/2
3
u=5
u=1
=
2(5
3/2
)
3

2(1
3/2
)
3
.
109
Example 32.6. When you use the Change of Variables Formula to compute an
indenite integral you must substitute back so that the answer is expressed in
the same variables as the original problem. For example, using the substitution
u = 3x, du = 3 dx we get
_
cos(3x) dx =
_
cos u
du
3
=
sinu
3
+C =
sin3x
3
+C.
32.7. Let s be the position of a particle at time t as in 6.3 and 31.2. and
suppose s = s
0
at some time t = t
0
and s = s
1
at some time t = t
1
. Let
v = ds/dt be the velocity at time t. Then by the Change of Variables formula
_
t1
t0
v dt =
_
t1
t0
ds
dt
dt =
_
s1
s0
ds = s
1
s
0
.
This says that the total displacement s
1
s
0
is the sum of all the innitesimal
displacements ds. On the other hand the total distance travelled by the particle
on this time interval is
_
t1
t0
[v[ dt =
_
t1
t0
ds
dt
dt =
_
s1
s0
[ds[.
This will be dierent from the absolute value [s
1
s
0
[ of the total displacement
s
1
s
0
if the particle changes direction (i.e. if v changes sign) on the time
interval.
Exercises
Exercise 32.8. Evaluate the indenite integral by making the indicated sub-
stitution.
(i)
_
x(x
2
1)
99
dx, u = x
2
1. (ii)
_
x
2
5 +x
3
dx, u = 5 +x
3
.
(iii)
_
(lnx)
2
x
dx, u = lnx. (iv)
_
xdx
1 +x
2
, u = x
2
.
(v)
_
x
x 1 dx, u = x 1. (vi)
_
e
x
(1 +e
x
)
5
dx, u = 1 +e
x
.
_
2
1
(x
5
+1)
2
5x
4
dx in two ways, rst by expanding and
evaluating the integral directly and then via the change of variables u = x
5
+1.
Exercise 32.10. Evaluate the denite integral.
(i)
_
1
0
x(x
2
1)
99
dx. (ii)
_
2
0
x
2
5 +x
3
dx.
(iii)
_
3
1
(lnx)
2
x
dx. (iv)
_
3
2
xdx
1 +x
2
.
110
(v)
_
5
3
x
x 1 dx. (vi)
_
1
0
e
x
(1 +e
x
)
5
dx.
(vii)
_
2
0
1 x
1 +x
dx. (viii)
_
2
0
1 x
1 +x
2
dx.
Exercise 32.11. A particle moves along a straight line. Its velocity at time t
is v = t
2
t 6. What is the distance between its position at time t = 0 and
its position at time t = 4? How far does it travel between time t = 0 and time
t = 4? Hint: v = ds/dt. The second question is dierent from the rst since
the particle turns around at time t = 3.
111
Chapter VI
Applications of Denite Integrals
The basic principle in this chapter is that the
whole is equal to the sum of its parts.
Every application rests on a formula of the form
Q =
_
dQ.
Riemann sums play a key role in the reasoning. In each application we reason
with Riemann sums to achieve an approximate equality
Q
Q
and then pass to the limit to nd Q exactly.
33 Plane Area
To nd the area of a plane gure by calculus we use the formula
A =
_
dA.
Example 33.1. We nd the area of the area bounded by the two curves y = x
3
and x = y
2
. The curves intersect at (x, y) = (0, 0) and (x, y) = (1, 1). The region
between the two curves is dened by the inequalities
0 x 1, x
3
y
x.
Break the area into strips parallel to the y-axis of width dx. The strip corre-
sponding to a given value of x has area
dA = (
x x
3
) dx
and the total area is
A =
_
dA =
_
1
0
(
x x
3
) dx =
_
2x
3/2
3

x
4
4
_
1
0
=
5
12
.
112
Remark 33.2. The reasoning in the previous example really involves Riemann
sums. The area was approximated by the Riemann sum
n
j=1
A
j
=
n
j=1
(
c
j
c
3
j
)x
j
where 0 = x
0
c
1
x
1
c
2
x
2
x
n1
c
n
x
n
= 1. and
x
j
= x
j
x
j1
. The limit of the Riemann sums for max
j

j
0 is the the
desired area.
Example 33.3. We evaluate the area bounded by the curve y = x
2
and the
line y = 2x in two ways.
(i) Using vertical strips: dA = (2x x
2
) dx, 0 x 2, so
A =
_
dA =
_
x=2
x=0
(2x x
2
) dx =
_
x
2
x
3
3
_
2
0
= 4
8
3
=
4
3
.
(ii) Using horizontal strips: dA = (
y y/2) dy, 0 y 4,
A =
_
dA =
_
y=4
y=0
_
y
y
2
_
dy =
_
2y
3/2
3

y
2
4
_
4
0
=
16
3
=
16
4
=
4
3
Remark 33.4. The curve y = x
2
is concave up. The Secant Concavity Theo-
rem 22.10 tells us that the line y = 2x is above the curve y = x
2
for 0 x 2.
Example 33.5. Exercise 33.6 asks us to nd the area bounded by the curve
y = x(4 x) and the line y = x in two ways: using vertical strips and using
horizontal strips. Here is how to get started on the latter method. The curve
and line intersect in the two points (x, y) = (0, 0) and (x, y) = (3, 3). The curve
y = x(4 x) can also be written as 4 y = (x 2)
2
. We have two kinds of
horizontal strips:
dA = (y 2 +
_
4 y) dy for 0 y 3
dA = 2
_
4 y dy for 3 y 4.
By the Additivity Rule
_
b
a
=
_
c
a
+
_
b
c
we have
A =
_
dA =
_
3
0
_
y 2 +
_
4 y
_
dy +
_
4
3
2
_
4 y dy.
Exercises
Exercise 33.6. Find the area bounded by the curve y = x(4 x) and the line
y = x in two ways, rst by using horizontal strips and the Additivity Rule (see
Example 33.5) and then using vertical strips.
113
Exercise 33.7. Find the area enclosed by the parabola y = 2 x
2
and the line
y = x. (Thomas Example 4 Page 380.)
Exercise 33.8. Find the area of the region in the rst quadrant bounded above
by y =

x and below by the x-axis and the line y = x2. (Thomas Example 5
Page 381.)
Exercise 33.9. Find the area in the previous exercise by integrating with re-
spect to y. (Thomas Example 6 Page 382.)
Exercise 33.10. Find the area of the region in the rst quadrant above the
curve y = x
2
and below the curve y = x
3
Exercise 33.11. Find the area of the region in the rst quadrant above the
curve x = y
2
and below the curve x = y
3
.
Exercise 33.12. Find a so that the curves y = x
2
and y = a cos x intersect at
the points (x, y) = (
4
,

2
16
). Then nd the area between these curves.
34 Volumes
To nd the volume V of a body by calculus we use the formula
V =
_
dV.
34.1. Suppose the body results by revolving the region
0 y f(x), a x b
about the x-axis. Then the line segment from (x, 0) to (x, f(x)) is the radius of
a disk whose area is A(x) = f(x)
2
so the volume is
V =
_
dV =
_
A(x) dx =
_
b
a
f(x)
2
dx.
The term dV = f(x)
2
dx represents the volume of an innitely thin circular
disk of width dx and radius f(x).
Example 34.2. If we revolve the half disk
0 y
_
a
2
x
2
, a x a
about the x-axis we get a sphere of radius a. Its volume is
V =
_
a
a
(
_
a
2
x
2
)
2
dx =
_
a
2
x
x
3
3
_
a
a
=
_
2a
3
2a
3
3
_
=
4a
3
3
.
114
Example 34.3. Consider the triangle bounded by the x-axis, the y-axis, and
the line ay +hx = ah, Revolve this triangle about the y-axis. (Note: y-axis not
x-axis.) The result is a right circular cone whose apex is at the point (0, h) and
whose base is a disk of radius a. The disk perpendicular to the y-axis centered at
the point (0, y) has radius x = a(hy)/h and hence area A(y) = a
2
(hy)
2
/h
2
.
Hence the volume is
V =
_
h
0
a
2
(h y)
2
h
2
dy =
a
2
h
2
_
(h y)
3
3
_
h
0
=
a
2
h
3
.
Thus the volume of a right circular cone is one third the area of its base time
its altitude.
34.4. The Slice Principle. Suppose that we want to nd the volume of
some solid body B. We choose a line L and for each point x on L we nd nd
the area A(x) of the intersection of B with the plane through x perpendicular
to L. Then
V =
_
dV where dV = A(x) dx.
The disk method for the volume of a surface of revolution is an instance of the
Slice Principle with the line L being the axis of rotation and A(x) = f(x)
2
.
Example 34.5. Here is another instance of the Slice Principle. A plane passes
through a diameter of the base of a right circular cylinder of radius a and
makes an angle of with the base. It cuts a wedge out of the cylinder: we
will nd the volume of this wedge. Represent the base of the cylinder as the
disk x
2
+ y
2
a
2
and let the diameter through which the plane passes be the
x-axis. A plane perpendicular to this diameter and passing through the point
(x, 0) cuts the wedge in a right triangle of base b =

a
2
x
2
and altitude
h =

a
2
x
2
tan so the area of the triangle is A(x)
1
2
bh =
1
2
(a
2
x
2
) tan.
The volume of the wedge is
V =
_
dV =
_
a
a
A(x) dx =
tan
2
_
a
a
(a
2
x
2
) dx =
2a
3
tan
3
34.6. Cylinders. The body formed by passing parallel lines through a plane
gure is called an (innite) cylinder. The parallel lines are called the gener-
ators of the cylinder. We do not assume that the generators are perpendicular
to the plane of the original gure. A second plane parallel to the rst and at
a distance h from it cuts o a cylinder of altitude h. By the Slice Principle
the volume of this cylinder is F = Ah where A is the area of the base. This is
because every plane parallel to the original plane cuts cylinder in a gure of the
same area A.
Remark 34.7. The diameter of a quarter dollar coin is about one inch so its
area is about (/4)in
2
. A stack of 80 quarters forms a cylinder about ve inches
tall. This right circular cylinder has volume 20in
3
. If we place a ruler along
the side of the stack and then tilt the ruler (keeping the quarters touching it)
115
we get a cylinder which is not a right cylinder. Its altitude is still ve inches
and its volume is still 20in
3
, the total volume of the 80 quarters. In fact, even
if the edges of the coins do not align along a straight line the volume of the
quarters is unchanged. (This illustrates the Slice Principle.)
34.8. Cones. Consider a plane gure of area A and a point P not on the
plane. The body formed by drawing line segments from the points of the gure
to the point P is called a cone. The point P is called the apex of the cone and
the distance from P to the plane is called the altitude of the cone. Let L be
the line though P perpendicular to the plane and x be the distance from P to
a variable point X on L. By similarity the plane through X and perpendicular
L cuts the cone in a gure of area
A(x) =
x
2
h
2
A
so the volume of the cone is
V =
_
dV = A
_
h
0
x
2
h
2
dx =
Ah
3
.
For example, the volume of a pyramid of altitude h whose base is a square of
side a is V = a
2
h/3. It does not matter if the apex of the pyramid is directly
over the center of the square. See Exercises 34.21-34.23.
34.9. Shells. Consider a body formed by rotating around the y-axis a gure
lying in the right half plane. Suppose that a line parallel to the y-axis and at
a distance x from it cuts the gure in a segment of length f(x). When this
line segment is rotated around the y-axis, it sweeps out a right circular cylinder
of radius x and height f(x). The area of this cylinder is A(x) = 2xf(x). A
second line segment at a distance of dx from the rst sweeps out a slightly
larger cylinder of about the same area and the cylindrical shell between these
two cylinders has volume
dV = 2xf(x) dx.
Example 34.10. We recalculate the volume of the cone from Example 34.3
using shells. This cone was constructed by rotating the triangle bounded by the
x-axis, the y-axis, and the line y = h(a x)/a about the y-axis. The volume of
the shell with coordinate x and width dx is
dV = 2x
h(a x)
a
dx
so the volume of the cone is
V =
_
a
0
2x
h(a x)
a
dx =
h
a
_
ax
2
2x
3
3
_
a
0
=
a
2
h
3
in agreement with the answer we found in Example 34.3.
116
Example 34.11. We calculate the volume of a sphere of radius a using shells.
The sphere is obtained by rotating the half disk
0 y
_
a
2
x
2
, a x a
about the x-axis. A line parallel to the x-axis at height y intersects the half
disk in a segment of length 2
_
a
2
y
2
and sweeps out a cylinder of area 2y
2
_
a
2
y
2
. Hence the volume is
V =
_
a
0
4y
_
a
2
y
2
dy = 2
_
0
a
2
udu = 2
2u
3/2
3
0
a
2
=
4a
3
3
To evaluate the integral we used the substitution u = a
2
y
2
, so du = 2y dy,
and x = a = u = 0 and x = 0 = u = a
2
.
Exercises
Exercise 34.12. Find the volume generated of a pyramid that has an altitude
of h and a base that is a square of side a. Hint: This is an example of a cone as
dened in 34.8. A plane parallel to the base and at a distance x from the apex
cuts the pyramid in a square. Determine the side length of the square using
similar triangles. (See Thomas Example 1 Page 398.)
Exercise 34.13. Find the volume of the solid generated by revolving the region
between the y-axis and the curve x = 2/y, 1 y 4, about the y-axis. (See
Thomas Example 7 Page 401.)
bounded by y = 1
x, the x-axis, and the line x = 4 about the x-axis.

bounded by y =

x and the lines y = 1, x = 4 about the line y = 1. Hint: The
radius of the circle swept out by by the point (x,
x) is

x 1. (See Thomas
Example 6 Page 401.)
between the parabola x = 2 y
2
and the y-axis about the x-axis.
between the parabola x = y
2
+1 and the line x = 3 about the line x = 3. Hint:
The radius of the circle swept out by by the point (y
2
+ 1, y) is 2 y
2
. (See
Thomas Example 8 Page 402.)
Exercise 34.18. Find the volume of the solid generated by revolving the solid
between the parabola y = x
2
and the line y = 2x in the rst quadrant about
the y-axis. Hint: A horizontal slice intersects the solid in a washer, i.e. the
area between two concentric circles. (See Thomas Example 10 Page 404.)
117
Exercise 34.19. A vase is constructed
by rotating the curve y =

x 1/100
for 0 y 6 around the y axis. It is
lled with water to a height y = h where
h < 6. (a) Find the volume of the water
in terms of h. (b) If the vase is lling
with water at the rate of 2 cubic units
per second, how fast is the height of the
water increasing when this height is 5
units?
Exercise 34.20. A triangle is formed by drawing lines from the two endpoints
of a line segment of length b to a vertex V which is at a height h above the line
of the line segment. Its area is then A =
_
h
y=0
dA where dA is the area of the
strip cut out by two parallel lines separated by a distance of dz and at a height
of z above the line containing the line segment. Find a formula for dA in terms
of b, z, and dz and evaluate the denite integral.
Exercise 34.21. A pyramid is formed by drawing lines from the four vertices
of a rectangle of area A to a apex P which is at a height h above the plane of
the rectangle. (The apex need not be above the center of the rectangle.) Its
volume is then V =
_
h
z=0
dV where dV is the volume of the slice cut out by two
planes parallel to the plane of the rectangle and separated by a distance of dz
and at a height of z above the plane of the rectangle. Find a formula for dV in
terms of A, z, and dz and evaluate the denite integral.
Exercise 34.22. A tetrahedron is formed by drawing lines from the three ver-
tices of a triangle of area A to a apex P which is at a height h above the plane
of the triangle. Its volume is then V =
_
h
z=0
dV where dV is the volume of the
slice cut out by two planes parallel to the plane of the triangle and separated
by a distance of dz and at a height of z above the plane of the rectangle. Find
a formula for dV in terms of A, z, and dz and evaluate the denite integral.
Exercise 34.23. A skew cone is formed by drawing lines from the perimeter
of a circle of area A to an apex P which is at a height h above the plane of the
circle. (The apex need not be above the center of the circle.) Its volume is then
V =
_
h
z=0
dV where dV is the volume of the slice cut out by two planes parallel
to the plane of the circle and separated by a distance of dz and at a height of z
above the plane of the rectangle. Find a formula for dV in terms of A, z, and
dz and evaluate the denite integral.
35 Arc Length
118
To nd the arc length s of a curve by calculus we use the formula
s =
_
ds.
35.1. Consider a parametric curve dened by the equations
x = f(t), y = g(t), a t b,
where the functions f and g are continuously dierentiable. The endpoints of
the curve are the points
A = (f(a), g(a)), B = (f(b), g(b)).
A partition a = t
0
< t
1
< t
2
< < t
n
= b of the interval [a, b] determines a
polygonal arc L with vertices
P
j
= (x
j
, y
j
) = (f(t
j
), g(t
j
)).
The polygonal arc L has the same endpoints A = P
0
and B = P
n
as the original
curve. Bt the Pythogorean Theorem the length L of the polygonal arc is
L =
n
j=1
s
j
where
s
j
=
_
(x
j
)
2
+ (y
j
)
2
, x
j
= x
j
x
j1
, y
j
= y
j
y
j1
. ()
Note that
s
j
=
s
j
t
j
t
j
=
_
x
j
t
j
_
2
+
_
y
j
t
j
_
2
t
j
and by the denition of the derivative
_
x
j
t
j
_
2
+
_
y
j
t
j
_
2
_
dx
dt
_
2
+
_
dy
dt
_
2
when t
j
0 and the right hand side is evaluated at any value of t in the
interval [t
j1
, t
j
]. The integral
s =
_
b
a
_
dx
dt
_
2
+
_
dy
dt
_
2
dt
is dened to be the arclength of the original parameterized curve
119
Theorem 35.2. (i) The length L of the polygonal arc is approximately equal
to the arclength when mesh is small:
max
j
t
j
0 = L s.
(ii) The value of the arclength s is independent of the parametrization of the
curve.
Part (i) requires proof because the formula for L is not obviously a Riemann
sum. The proof is not dicult, but we leave it for a more advanced course.
Part (ii) is an easy consequence of the Change of Variables Theorem 32.1 as
follows. Any other parameterization can be obtained from the original param-
eterizarion via a substitution t = T() where T is an increasing function whose
domain is an interval [, ] and whose range is the domain [a, b] of the original
parameterization. Thus
dt = T
(), d = T
() =
dt
d
d
and
= = t = a, = = t = b
so by the Chain Rule and the Change of Variables Theorem
_

_
dx
d
_
2
+
_
dy
d
_
2
d =
_

_
dx
dt

dt
d
_
2
+
_
dy
dt

dt
d
_
2
d
=
_

_
dx
dt
_
2
+
_
dy
dt
_
2
dt
d
d
=
_
b
a
_
dx
dt
_
2
+
_
dy
dt
_
2
dt
as required.
Remark 35.3. If we imagine the parameterization of the curve as describing
the position of a moving point at time t, then
s = s(t) =
_
t
a
_
dx
dt
_
2
+
_
dy
dt
_
2
dt
represents the distance travelled by the particle during the time interval [a, t].
By the Fundamental Theorem the derivative of s is
s
(t) =
ds
dt
=
_
dx
dt
_
2
+
_
dy
dt
_
2
.
This is often written
ds =
_
(dx)
2
+ (dy)
2
in analogy with the notation () used above in the formula for the length of the
polygonal arc.
120
Remark 35.4. In the special case where the curve is the graph y = f(x) of a
function we may take the parameterization
x = t, y = f(t)
and the formula for the arclength reduces to
s =
_
b
a
1 +
_
dy
dx
_
2
dx.
Example 35.5. For the curve y = (4
2/3)x
3/2
1, 0 x 1, we have
dy/dx = 2
2x
1/2
so
ds =
1 +
_
dy
dx
_
2
dx =

1 + 8xdx
and
s =
_
ds =
_
1
0
1 + 8xdx =
2
3

1
8
(1 + 8x)
3/2
1
0
=
13
6
Example 35.6. We nd the arclength of the arc of the circle of radius a sub-
tended by the lines y = (tan)x and y = (tan)x where 0 < < < /2.
This curve is given partameterically by the equations
x = a cos , y = a sin, .
We compute
dx
d
= a sin,
dy
d
= a cos , ds =
_
dx
d
_
2
+
_
dy
d
_
2
d = a d
so the arclength is s =
_

d = a( ). Using the formula

y =
_
a
2
x
2
, a cos x a cos
we get
ds =
1 +
_
dy
dx
_
2
dx =
a
a
2
x
2
and hence
s =
_
a cos
a cos
a dx
a
2
x
2
= a cos
1
_
x
a
_
a cos
a cos
= a +a
as before.
121
Exercises
Exercise 35.7. Find the length of each of the following curves. Warning: The
functions in these problems have been carefully chosen so that the resulting
integral can be done. The slightest error in algebra may lead to an integral
which cannot be done.
(i) x = 1 t, y = 32 + 3t, 2/3 t 1.
(ii) x = t
3
, y = 3t
2
/2, 0 t
3.
(iii) x = 8 cos t + 8t sint, y = 8 sint 8t cos t, 0 t /2.
(iv) y = x
3/2
, 0 x 4.
(v) y = (x
3/2
/3) x
1/2
, 1 x 3.
(vi) x = (y
3
/6) + 1/(2y), 2 y 3.
(vii) y =
_
x
0
_
u
2
+ 2u du, 1 x 4.
Exercise 35.8. Find a denite integral whose value is the length of each of
the following curves. Do not evaluate the integral. Do specify the limits of
integration.
(i) x = a cos , y = b sin, 0 2.
(ii) x = t
2
, y = t
3
, 0 t 3.
(iii) y = x
4
, 2 x 3.
(iv) y =
_
x
0
_
u
2
+ 1u du, 1 x 4.
36 Surface Area
See Thomas pages 436-446. This topic was not covered in Fall 2006.
To nd the area S of a surface by calculus we use the formula
S =
_
dS.
122
36.1. The region cut out of a right circular cone by two planes parallel to the
base is called a frustrum. Let P
1
and P
2
be the points in which a generator
of the cone intersects the two planes: the frustrum is swept out by rotating the
line segment P
1
P
2
about the axis of the cone. Let P be the midpoint of P
1
P
2
and r
1
, r
2
, and r = (r
1
+r
2
)/2, be the radii of the circles swept out by P
1
, P
2
,
and P respectively. If we cut the cone and lay it at we see that the area A of
the frustrum is the dierence of the areas of two sectors with a common central
angle and radii r
1
and r
2
, i.e.
A =
_
r
2
2
2

r
2
1
2
_
=
_
r
1
+r
2
2
_
(r
2
r
1
) = r (r
2
r
1
).
The second factor (r
2
r
1
)on the right is length of the line segment P
1
P
2
and
the rst factor r is the length of the circular are swept out by the midpoint P.
But this circular arc has the same length as the circle swept out by the midpoint
P in the original cone, i.e. r = 2x where x is the radius of this circle, i.e. the
distance from P to the axis of the cone. Hence
A = 2x, = r
2
r
1
. ()
36.2. Now we compute the surface area swept out when a curve in the right
half plane is rotated around the y-axis. As in 35.1 we consider a polygonal
arc L with vertices P
0
, P
1
, . . . , P
n
on and having the same endpoints A = P
0
and B = P
n
as . By Equation () the polygonal arc L sweeps out a surface (a
union of frustrums) with area
S(L) =
n
j=1
x
j
s
j
where x
j
is the x-coordinate of the midpoint of the line segment P
j1
P
j
and
s
j
is the length of this segment. If we choose a parameterization of the original
curve the sum S(L) becomes a Riemann sum approximating the area S()
swept out by . Hence
S() =
_
dS, dS = 2xds
where ds is the arclength element of the original curve .
Example 36.3. The surface area of of the sphere of radius a swept out when
the circular arc x =
_
a
2
y
2
of radius a is rotated about the y-axis is
S =
_
a
a
2x
1 +
_
dx
dy
_
2
dy.
The quantity under the radical sign is
1 +
_
dx
dy
_
2
= 1 +
_
y
_
a
2
y
2
_
2
=
a
2
a
2
y
2
123
so the integral evaluates to
S =
_
a
a
2ax
_
a
2
y
2
dy =
_
a
a
2a dy = 4a
2
where we used x =
_
a
2
y
2
in the penultimate step.
Remark 36.4. We can also nd the area of a sphere of radius a using the
standard parameterization
x = a cos , y = a sin,
2

2
.
For this parameterization ds = a d so
S =
_
2xds =
_
/2
/2
(2a cos )a d = 2a
2
sin
/2
/2
= 4a
2
.
Exercises
Exercise 36.5. Find the area of the surface generated by rotating each of the
following curves about the indicated axis. Warning: The functions in these
problems have been carefully chosen so that the resulting integral can be done.
The slightest error in algebra will lead to an integral which cannot be done.
(i) y = x
3
/9, 0 x 2, x-axis.
(ii) y =

x + 1, 1 x 5, x-axis.
(iii) x = (1/3)y
3/2
y
1/2
, 1 y 3, y-axis.
(iv) x
2/3
+y
2/3
= 1, 1 x 1, y 0, x-axis.
(v) y =
_
x
0
_
u
2
1 du, 1 x 2, y-axis.
Exercise 36.6. Find a denite integral whose value is the area of the surface
generated by rotating each of the following curves about the indicated axis. Do
not evaluate the integral. Do specify the limits of integration.
(i) y = x
3
/9, 0 x 2, y-axis.
(ii) y =

x
2
+ 1, 1 x 5, x-axis.
(iii) x = y
3/2
y
1/2
, 1 y 3, y-axis.
(iv) x
1/3
+y
1/3
= 1, 1 x 1, y 0, x-axis.
(v) y =
_
x
0
_
u
3
1 du, 1 x 2, y-axis.
Exercise 36.7. Find the surface area generated by revolving the circular arc
y =

1 x
2
, a x b about the x-axis. (The result depends only on b a,
i.e. all slices of the same width cut from a spherical loaf of bread have the same
amount of crust.)
124
37 Center of Mass
37.1. Every physical body has a center of mass. Every geometrical gure
has a centroid. The centroid of a gure is (by denition) the center of mass of
the body which results on giving the gure a uniform mass density. If a body
B with center of mass

P is suspended from a point Q on its boundary, it comes
to rest with the line Q
P vertical. If the body is a bar (lies in a line) or a plate

(lies in a plane) and is supported from any point other than the center of mass,
it will not balance.
37.2. Consider the case where the body consists of n collinear mass points
P
1
, P
2
, . . . , P
n
with masses m
1
, m
2
, . . . , m
n
. (Imagine that the straight line is a
teeter-totter and the mass points are children.) Let x be a coordinate on the
line, i.e. [x[ is the distance from a point O (the origin) on the line with x > 0
for points to the right of O and x < 0 for points to the left. If the coordinate
of the mass point P
j
is x
j
, then the coordinate x of the center of mass

P is the
weighted average
x =
1
M
j
m
j
x
j
, M =
j
m
j
. ()
The denominator M in the weighted average is the total mass and each mass
point receives a weight m
j
/M which represents the fraction of the total mass of
the mass point P
j
. (These weights sum to one.)
37.3. Here is an argument which explains why the bar doesnt balance if it is
supported at some point other than the center of mass. Imagine that the bar is
supported at the origin O and let the bar make an angle of with the horizontal
axis. Then the potential energy of the jth mass point is its mass m
j
times its
height x
j
sin. (When 0 /2 the points to the right of the origin have
positive height and those to the left have negative height. Remember that we
have dened x
j
to be the distance from P
j
to O along the bar, not the horizontal
coordinate of P
j
.) Hence the total potential energy is
U =
_
_
j
m
j
x
j
_
_
sin.
Physicists tell us that at equilibrium the potential energy is a local minimum
which implies that
0 =
dU
d
=
_
_
j
m
j
x
j
_
_
cos
which means that either cos = 0 (the bar is vertical) or else that
j
m
j
x
j
= 0
(the center of mass is at the origin).
125
37.4. Decomposition Principle. In many situations a set of mass points
behaves as if it is a single mass point of mass M concentrated at the center
of mass of the set. For example, imagine that the mass points are distributed
among k sets B
1
, B
2
, . . . , B
k
. Let the total mass of the ith set be M
i
and the
center of mass of the points in B
i
have coordinate x
i
. Then M =
i
M
i
is the
total mass and the center of mass

P is
x =
1
M
i
M
i
x
i
M =
i
M
i
.
For example imagine two boys with masses m
1
, m
2
at x
1
, x
2
> 0 on one side
of the teeter-totter and two girls with masses m
3
, m
4
at x
3
, x
4
< 0 on the
other side. The boys have total mass M
1
= m
1
+ m
2
and center of mass
x
1
= (m
1
x
1
+ m
2
x
2
)/M
1
and the girls have total mass M
2
= m
3
+ m
4
and
center of mass x
2
= (m
3
x
3
+m
4
x
4
)/M
2
. If a man of weight M
1
sits at x
1
and
a woman of weight M
2
sits at x
2
the center of mass is at
x =
M
1
x
1
+M
2
x
2
M
1
+M
2
=
(m
1
x
2
+m
2
x
2
) + (m
3
x
3
+m
4
x
4
)
(m
1
+m
2
) + (m
3
+m
4
)
which is the same as the center of mass of the boys and girls. I like to remember
the Decomposition Principle with the following dumb slogan:
The center of mass of the centers of mass is the center of mass.
37.5. Next we treat a solid bar. We assume that the bar has a mass density
which means that the amount dm mass in an innitely small interval of width
dx about the point with coordinate x is
dm = (x) dx.
(The mass density at a point of the bar might depend on position of the point
if (say) the bar would made of an alloy of two metals with more of one metal
at one end than at the other.) Suppose that the coordinate of a point on the
bar lies between the two values a and b. In analogy with Equation () above we
have
The total mass of the bar with coordinate x in [a, b] and mass density
(x) is
M =
_
dm =
_
b
a
(x) dx
and the coordinate of the center of mass

P is
x =
1
M
_
xdm =
1
M
_
b
a
x(x) dx.
126
Example 37.6. Suppose that the bar has a uniform mass density, i.e. the
amount of mass in a segment of the bar is proportional to the width of the
segment. Then the mass density is constant and the amount of mass in a tiny
segment of width dx is dm = dx. The center of mass is
x =
_
xdm
_
dm
=
_
b
a
xdx
_
b
a
dx
=
x
2
2
b
a
x
b
a
=
(b
2
a
2
)
2(b a)
=
(b a)(b +a)
2(b a)
=
b +a
2
.
Thus the center of mass is at the midpoint of the bar. Since the mass distribution
is uniform, the center of mass and centroid coincide.
37.7. Similar considerations hold for systems which are not conned to one
dimension. To specify a point in space it takes three coordinates (x, y, z) . For
a system of mass points P
j
= (x
j
, y
j
, z
j
) of mass m
j
connected by weightless
rigid rods the center of mass is the point

P = ( x, y, z) whose coordinates are
x =
1
M
j
m
j
x
j
, y =
1
M
j
m
j
y
j
, z =
1
M
j
m
j
z
j
,
where M =
j
m
j
is the total mass as before. If the system of mass points lies
in the (x, y)-plane, the z-coordinates of all the points (and hence of the center
of mass) are zero. For a body with a mass distribution the center of mass has
coordinates
x =
1
M
_
xdm, y =
1
M
_
y dm, z =
1
M
_
z dm,
where M =
_
dm is the total mass. For a wire the innitesimal mass dm in an
innitesimal piece of arc of length ds at the point P is given by
dm = ds,
where = (P) is the mass density at the point P. For a plate the innitesimal
mass dm in an innitesimal piece of the plate of area dA at the point P is given
by
dm = dA,
where = (P) is the mass density at the point P. For a body the innitesimal
mass dm in an innitesimal piece of the body of volume dV at the point P is
given by
dm = dV,
where = (P) is the mass density at the point P. Appropriate units for the
density are grams per centimeter for a wire, grams per square centimeter for
a plate, and grams per cubic centimeter for a body. Using Calculus 221 we
can nd the center of mass of a wire and, with the aid of the Decomposition
Principle (see below), some plates and bodies. Multiple integrals (taught in
Calculus 234) are required for the general case.
127
Example 37.8. (The Centroid of a Semicircle.) We nd the center of
mass of the semicircle
x = a cos , y = a sin, 0
with a uniform mass distribution . The element of arclength is
ds =
_
dx
d
_
2
+
_
dy
d
_
2
d = a d.
The amount of mass in it is dm = ds = a d. The coordinates of the center
of mass are
x =
_
xds
_
ds
=
_

0
cos d
_

0
ds
=
sin
= 0
and
y =
_
y ds
_
ds
=
_

0
sin d
_

0
ds
=
cos
=
2
.
Note that the constant appears in the numerator and denominator and thus
cancels. Thus when we want to nd the the center of mass of a uniform mass
distribution (i.e. constant mass density), we may as well assume that = 1.
37.9. Integral Decomposition Principle. The Decomposition Principle
holds quite generally, whenever a body is decomposed into subsets in almost
any way imaginable. Suppose that a plate in the (x, y)-plane is decomposed
into innitely many innitely thin strips parameterized by a variable u. The
mass of the strip is dm (it may depend on the parameter u) and its center of
mass is ( x, y) which also depends on u. An integral version of the Decomposition
Principle applies and the center of mass of the plate has coordinates
x =
1
M
_
xdm, y =
1
M
_
y dm, M =
_
dm.
A similar principle holds for a three dimensional solid body.
Example 37.10. (The Centroid of a Half Disk.) We nd the center of
mass of the half disk
0 y
_
a
2
x
2
, a x a
of radius a with a uniform mass distribution . The mass density is constant
so, by the same reasoning as in Example 37.8 , we might as well take = 1.
Then the total mass is the area of the half disk, i.e.
M =
_
dA =
_
a
a
_
a
2
x
2
dx = a
2
/2
128
(half the area of the disk of radius a) and the center of mass of an innitely thin
strip with end points (x, 0) and (x,
a
x
x
2
) is its midpoint ( x, y):
x = x, y =
a
x
x
2
2
.
By the Integral Decomposition Principle 37.9, the center of mass ( x, y) is given
by
x =
_
xdm
_
dm
=
_
a
a
x
_
_
a
2
x
2
_
dx
a
2
/2
=
1
a
2
_
0
0
udu = 0
(where we we made the substitution u = a
2
x
2
, du = 2xdx) and
y =
_
y dm
_
dm
=
_
a
a
_
a
2
x
2
2
_
_
_
a
2
x
2
_
dx
a
2
/2
=
1
a
2
_
a
a
(a
2
x
2
) dx =
1
a
2
_
a
2
x
x
3
3
_
a
a
=
4a
3
.
Note that the center of mass is on the y-axis as could be predicted by symmetry.
37.11. Here is a more general formulation of the argument used in Exam-
ple 37.13. Consider a thin plate with constant mass density bounded by the
curves y = f(x) and y = g(x) and between the lines x = a, and x = b. Assume
g(x) < f(x) for a < x < b. Then the center of mass of an innitely thin strip
with end points (x, g(x)) and (x, f(x)) is its midpoint ( x, y):
x = x, y =
f(x) +g(x)
2
.
The mass density is constant and thus cancels in the expression for the center
of mass so we take = 1. Hence the height of the strip is f(x) g(x) so, if the
strip has with dx, its mass (area) is
dm =
_
f(x) g(x)
_
dx.
By the Integral Decomposition Principle 37.9, the center of mass

P = ( x, y) is
given by
x =
_
xdm
_
dm
=
_
b
a
x
_
f(x) g(x)
_
dx
_
b
a
_
f(x) g(x)
_
dx
129
and
y =
_
y dm
_
dm
=
_
b
a
f(x) +g(x)
2

_
f(x) g(x)
_
dx
_
b
a
_
f(x) g(x)
_
dx
.
Theorem 37.12. The medians
16
of a triangle are concurrent: they intersect in
the centroid.
Proof: First we give a physics argument and then we give a calculus proof.
Decompose the triangle into innitely many innitely thin rectangular strips
parallel to one of the sides. Each strip has its center of mass at the midpoint.
By similar triangles each of these midpoints lies on the median. The mass of
the strip is proportional to its length and we can nd the coordinates of the
center of mass using the Integral Decomposition Principle. This is equivalent to
nding the center of mass of a variable mass distribution on the median so the
centroid lies on the median. (See Exercise 37.21.) Since the argument works for
each of the medians the centroid lies on all three medians.
Example 37.13. (The Centroid of a Triangle.) We prove Theorem 37.12
using calculus. Let the vertices be (a, h), (b, h), and (0, 0) with a < b and h > 0.
The sides of the triangle lie on the three lines x = ay/h, x = by/h, and y = h.
We view the horizontal line y = h as the base of the triangle so h is the altitude
and the area is h(b a)/2. Since the mass density is constant so we might as
well take take = 1. The base of the triangle has length b a so, by similar
triangles, the length of the horizontal line segment at height y is (b a)y)/h so
the the mass (i.e. the area) of the thin horizontal rectangular strip of width dy
and height y is
dm =
(b a)y
h
dy
The total mass is
M =
_
dm =
(b a)y
h
dy =
(b a)y
2
2h
h
0
=
(b a)h
2
.
(No surprise: the base of the triangle is b a and the height is h.) The center of
mass of the horizontal strip at height y is the midpoint ( x, y) of the horizontal
line segment so
x =
(a +b)y
2h
, y = y.
By the Integral Decomposition Principle 37.9,
y =
1
M
_
y dm =
2
(b a)h
_
h
0
y
(b a)y
h
dy =
2
h
2
_
h
0
y
2
dy =
2h
3
16
A median of a triangle is a line joining a vertex to the midpoint of the opposite side.
130
and
x =
1
M
_
xdm =
2
(b a)h
_
h
0
(a +b)y
2h

(b a)y
h
dy =
a +b
3
.
The horizontal line through the centroid is y = 2h/3 and it intersects the sides
x = ay/h and x = by/h in the points (2a/3, 2h/3) and (2b/3, 2h/3). The
midpoint of the line segment joining these two points is the centroid.
Exercises
Exercise 37.14. Two rods of lengths a and b and the same constant mass
density are welded together to form a right angle. Assume that the endpoints
of the rst rod are O = (0, 0) and A = (a, 0) and the endpoints of the second
rod are O and B = (0, b). Find the center of mass

P = ( x, y).
Exercise 37.15. Find the center of mass of a bar located on the x-axis with
end points x = 0 and x = 2 and mass density (x) = x
3
.
Exercise 37.16. Find the center of mass of a thin plate with constant mass
density bounded by the parabola y = x
2
and the line y = 3.
Exercise 37.17. Find the center of mass of a thin plate (a half disk) with
constant mass density mass bounded by the x-axis and the circle x
2
+y
2
= a
2
.
Is the answer the same as for the semicircle of Example 37.8?
Exercise 37.18. Find the center of mass of a thin plate (a quarter disk) with
constant mass density mass bounded by the x-axis, the y-axis, and the circle
x
2
+y
2
= a
2
.
density mass in the rst quadrant bounded by the lines y = 3, x = 3, and the
circle x
2
+y
2
= 9. (You can avoid some calculation by using geometry and the
previous problem to nd the area.)
density bounded by the curves y = x + 1, y = x
2
between the lines x = 0, and
x = 1 .
Exercise 37.21. Show that if a system of mass points in space all lie on the
same line, then the center of mass also lies on that line. (Hint: Choose coor-
dinates so that the line is the x-axis. Then the y
j
and and z
j
all vanish: you
must show that y and z also vanish.)
Exercise 37.22. Find the center of mass of a system of three mass points of
the same mass at the points A = (a, h), B = (b, h), C = (0, 0). Is the center of
mass of the three mass points the same as the centroid of the triangle ABC?
Exercise 37.23. Find the center of mass of a system of three bars with the same
uniform mass density connecting the vertices A = (a, h), B = (b, h), C = (0, 0).
Hint: By the Decomposition Principle this is the same as the center of mass of
a system of three possibly unequal mass points located at the midpoints of the
sides. Is the center of mass of the three bars the same as the centroid of the
triangle ABC?
131
Chapter VII
Loose Ends
This chapter contains material which we probably wont have enough
time to cover in the course.
38 The Natural Log Again*
In this section we prove Theorem 16.1 which dened the exponential function
exp
a
(x) = a
x
and also prove that this function is dierentiable.
38.1. It is easy to nd the antiderivative of a power, at least when the exponent
is not 1, namely
_
x
n
dx =
x
n+1
n + 1
+C
To handle the exceptional case we make a
Denition 38.2. The natural logarithm function is dened by
lnx =
_
x
1
dt
t
.
Its domain is the set of all positive numbers x.
Theorem 38.3. There is a number e such that
lnx = log
e
x
for all positive x.
Proof:
38.4. Now we can prove the Power Law
d
dx
x
p
= px
p1
for any exponent p not just rational numbers.
132
39 Taylor Approximation*
This is a warmup for innite series which we study in Math 222.
39.1. When f is a function and k 0 is an integer the notation f
(k)
denotes
kth derivative of f. Thus
f
(0)
(x) = f(x), f
(1)
(x) = f
(x), f
(2)
(x) = f
(x),
and so on. Given a number a in the domain of f and an integer n 0, the
polynomial
P
n
(x) =
n
k=0
f
(k)
(a)(x a)
k
k!
(#)
is called the degree n Taylor polynomial of f centered at a. The Taylor
polynomial P
n
(x) is the unique polynomial of degree n which has the same
derivatives as f at a up to order n:
P
(k)
n
(a) = f
(k)
(a) for k = 0, 1, 2, . . . , n.
39.2. The letter

is the Greek S (for sum) and is pronounced sigma so the
notation used in (#) is called sigma notation. It is a handy notation but if
you dont like it you can indicate the summation with dots:
n
k=m
a
k
= a
m
+a
m+1
+ +a
n1
+a
n
.
Hence the rst few Taylor polynomials are
P
0
(x) = f(a),
P
1
(x) = f(a) +f
(a)(x a),
P
2
(x) = f(a) +f
(a)(x a) +
f
(a)(x a)
2
2
,
P
3
(x) = f(a) +f
(a)(x a) +
f
(a)(x a)
2
2
+
f
(a)(x a)
3
6
.
39.3. The Taylor polynomial P
n
(x) for f(x) centered at a is the polynomial
of degree n which best approximates f(x) for x near a. To make this precise let
R
n
(x) = f(x) P
n
(x)
denote the nth Taylor Error of f at a. When R
n
(x) is small P
n
(x) is a good
approximation for f(x). How small is small? The answer is given by
133
Theorem 39.4 (Taylors Formula). Suppose that f is n + 1 times dier-
entiable and that f
(n+1)
is continuous. Let a be a point in the domain of f.
Then
lim
xa
R
n
(x)
(x a)
n
= 0.
39.5. Theorem 39.4follows from Theorem 22 on page 812 of Thomas; the
proof is on pages 818-819. The Theorem tells us that not only is the error
R
n
(x) small when x is close to a, it is so small that it is still small after being
divied by the small number (x a)
n
.
In order to use Taylors formula approximate a function f we pick a point
a where the value of f and of its derivatives is known exactly. Then the Taylor
polynomial P
n
(x) can be evaluated exactly for any x. We then need to estimate
the error R
n
(x) = f(x) P
n
(x), i.e. to nd an inequality
[R
n
(x)[ M[x a[
n+1
which tells us how small the error R
n
(x) is, i.e. how close P
n
(x) is to f(x).
Theorem 23 on page 813 of Thomas tells us how to nd M. Well study this on
Math 222.
Theorem 39.6 (Extended Mean Value Theorem). Let f, R, and a be as
in Theorem 39.4. Then for each b there is a number c
n+1
between a and b such
that
f(b) P
n
(b) =
f
(n+1)
(c
n+1
)(b a)
n+1
(n + 1)!
.
39.7. Note that the formula for the error f(b) P
n
(b) is the same as the next
term in the series (#) except that the n + 1st derivative f
(n+1)
is evaluated at
the unknown point c
n+1
instead of a. The Extended Mean Value Theorem is
proved in problem 74 on page 174 of the text. Equation () is an immediate
consequence.
Exercises

5
k=3
1
k
.

x. Find the polynomial P(x) of degree three such
that P
(k)
(4) = f
(k)
(4) for k = 0, 1, 2, 3.
1/3
. Find the polynomial P(x) of degree two
which best approximates f(x) near x = 8.
Exercise 39.11. Let f(x) and P(x) be as in 39.10. Evaluate P(10) and use
the Extended Mean Value Theorem to prove that
[10
1/3
P(10)[
80
27 256
.
Hint: The function g(x) = x
8/3
is decreasing so g(10) g(8).
134
Exercise 39.12. Find a polynomial P(x) of degree three such that
lim
x0
sin(x) P(x)
x
3
= 0.
Use the Extended Mean Value Theorem to show that
[ sin(x) P(x)[
[x[
4
24
.
40 Newtons Method*
135
Chapter VIII
More Problems
For graphing problems you may be asked to determine (a) where f(x) is dened,
(b) where f(x) is continuous, (c) where f(x) is dierentiable, (d) where f(x) is
increasing and where it is decreasing, (e) where f(x) is concave up and where
it is concave down, (f) what the critical points of f(x) are, (g) where the points
of inection are, (h) what (if any) the horizontal asymptotes to f(x) are, and
(i) what (if any) the vertical asymptotes to f(x) are. (A horizontal line y = b is
called a horizontal asymptote if lim
x
f(x) = b or lim
x
f(x) = b. A vertical line
x = a is called a vertical asymptote if lim
xa+
f(x) = or lim
xa
f(x) = .)
For proofs the question will be carefully worded to indicate what you may
assume in your proof. (See Problem 10 for example.) In this document you
may use without proof any previously asserted fact. For example, you may use
the fact that sin
() = cos() to prove that cos
() = sin() since the former

question precedes the latter below. (See Problems 12 and 13.) You may always
use high school algebra (like cos() = sin(/2 )) in your proofs.
1. State and prove the Sum Rule for derivatives. You may use (without proof)
the Limit Laws.
2. State and prove the Product Rule for derivatives. You may use (without
proof) the Limit Laws.
3. State and prove the Quotient Rule for derivatives. You may use (without
proof) the Limit Laws.
4. State and prove the Chain Rule for derivatives. You may use (without proof)
the Limit Laws. You may assume (as the proof in the Stewart text does) that
the inner function has a nonzero derivative.
5. State the Sandwich
17
Theorem.
6. Prove that
dx
n
dx
= nx
n1
, for all positive integers n.
7. Prove that
dx
n
dx
= nx
n1
, for n = 0.
8. Prove that
dx
n
dx
= nx
n1
, for all negative integers n.
9. Prove that
de
x
dx
= e
x
.
10. Prove that
lim
0
sin()
= 1.
17
Also called the Squeeze Theorem
136
You may assume without proof the Sandwich Theorem, the Limit Laws, and
that the sin and cos are continuous. Hint: See Problem 98.
11. Prove that
lim
0
1 cos()
= 0.
12. Prove that
d sinx
dx
= cos x.
13. Prove that
d cos x
dx
= sinx.
14. Prove that
d tanx
dx
= sec
2
x.
15. Prove that
d cot x
dx
= csc
2
x.
16. Prove that
d lnx
dx
=
1
x
.
17. Prove that
d sin
1
x
dx
=
1
1 x
2
.
18. Prove that
d cos
1
x
dx
=
1
1 x
2
.
19. Prove that
d tan
1
x
dx
=
1
1 +x
2
.
20. True or false? A dierentiable function must be continuous. If true, give a
proof; if false, illustrate with an example.
21. True or false? A continuous function must be dierentiable. If true, give a
proof; if false, illustrate with an example.
22. Explain why lim
x0
1/x does not exist.
23. Explain why lim
/2
tan does not exist.
24. Explain why lim
/2
sec does not exist.
25. Explain why lim
0
csc does not exist.
26. Explain why lim
x0
sin(1/x) does not exist.
27. Explain why lim
cos does not exist.

28. Let sgn(x) be the sign function. This function is given by
sgn(x) =
_
_
1, if x > 0,
0, if x = 0,
1, if x < 0.
137
Explain why lim
x0
sgn(x) does not exist.
29. Explain why lim
y0
2
1/y
does not exist.
30. Explain why lim
x1
2
1/(x1)
does not exist.
31. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = sin 2x.
32. Calculate lim
h0
f(x +h) f(x)
(x +h) x
when f(x) = cos 2x.
33. Calculate lim
xa
f(x) f(a)
x a
when f(x) = sin(x
2
).
34. Calculate lim
xx0
f(x) f(x
0
)
x x
0
when f(x) = cos(x
2
).
35. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) =

sinx.
36. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = xsinx.
37. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = e
x
.
38. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = e
sin x
.
39. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = ln(ax +b).
40. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = e
cos x
.
41. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = x
x
.
42. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) =
sinx
x
.
43. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) =

ax +b.
44. Calculate lim
x0
f(x + x) f(x)
(x + x) x
when f(x) = (mx +c)
n
.
45. Use dierentiation to estimate the number
127
4/3
125
4/3
2
approximately
without a calculator. Your answer should have the form p/q where p and q are
integers. Hint: 5
3
= 125.
138
46. What is the derivative of the area of a circle with respect to its radius?
47. What is the derivative of the volume of a sphere with respect to its radius?
48. Find the slope of the tangent to the curve y = x
3
x at x = 2.
49. Find the equations of the tangent and normal to the curve y = x
3
2x +7
at the point (1, 6).
50. Find the equation of the tangent line to the curve 3xy
2
2x
2
y = 1 at the
point (1, 1). Find d
2
y/dx
2
at this point.
51. Find the equations of the tangent and normal to the curve
x
2
a
2
+
y
2
b
2
= 1 at
the point (a cos , b sin).
52. Find the equations of the tangent and normal to the curve
x
2
a
2

y
2
b
2
= 1 at
the point (a sec , b tan).
53. Find the equations of the tangent and normal to the curve c
2
(x
2
+y
2
) = x
2
y
2
at the point (c/ cos , c/ sin).
54. Find the equations of the tangent and normal to the parabola y
2
= 4ax at
the point (at
2
, 2at).
55. Show that the equation of the tangent to the hyperbola
x
2
a
2

y
2
b
2
= 1 at the
point (p, q) is
xp
a
2

yq
b
2
= 1
56. Find the equations of the tangent and normal to the curve y = x
4
6x
3
+
13x
2
10x + 5 at the point where x = 1.
57. Find the linear and quadratic approximations to f(x) =
1
4 +x
at x = 0.

1 +x at x = 0.
1
(1 + 2x)
4
at x = 0.
60. Find the linear and quadratic approximations to f(x) = (1 +x)
3
at x = 0.
61. Find the linear and quadratic approximations to f(x) = sec x at x = 0.
62. Find the linear and quadratic approximations to f(x) = xsinx at x = 0.
63. Find the linear and quadratic approximations to f(x) = x
3
at x = 1.
1/3
at x = 8.
65. Find the linear and quadratic approximations to f() = sin at = /6.
1
at x = 4.
3
x at x = 1.
139

x at x = 4.

x
2
+ 9 at x = 4.
70. Use quadratic approximation to nd the approximate value of
401 without
a calculator. Hint:

400 = 20.
71. Use quadratic approximation to nd the approximate value of (255)
1/4
without a calculator. Hint: 256
1/4
= 4.
72. Use quadratic approximation to nd the approximate value of
1
(2.002)
2
without a calculator.
73. Approximate (1.97)
6
without a calculator. (Leave arithmetic undone.)
74. Let f be a function such that f(1) = 2 and whose derivative is known to be
f
(x) =

x
3
+ 1. Use a linear approximation to estimate the value of f(1.1).
Use a quadratic approximation to estimate the value of f(1.1).
75. Find the second derivative of x
7
with respect to x.
76. Find the second derivative of lnx with respect to x.
77. Find the second derivative of 5
x
with respect to x.
78. Find the second derivative of tan with respect to .
79. Find the second derivative of x
2
e
3x
with respect to x.
80. Find the second derivative of sin3xcos 5x with respect to x.
81. Find the third derivative of u
4
with respect to u.
82. Find the third derivative of lnx with respect to x.
83. Find the second derivative of tanx with respect to x.
84. If = sin
1
y show that
d
2
dy
2
=
y
(1 y
2
)
3/2
.
85. If y = e
t
cos t show that
d
2
y
dt
2
= 2e
t
sint.
86. If u = t + cot t show that sin
2
t
d
2
u
dt
2
2u + 2t = 0.
87. If y = e
tan x
show that cos
2
x
d
2
y
dx
2
(1 + sin2x)
dy
dx
= 0.
88. State LHopitals rule and give an example which illustrates how it is used.
89. Explain why LHopitals rule works. Hint: Expand the numerator and the
denominator in terms of x.
90. Give three examples to illustrate that a limit problem that looks like it is
coming out to 0/0 could be really getting closer and closer to almost anything
and must be looked at a dierent way.
140
coming out to 1
could be really getting closer and closer to almost anything

coming out to 0
0
could be really getting closer and closer to almost anything
93. Give three examples to illustrate that a limit problem that looks like it
is coming out to could be really getting closer and closer to almost
anything and must be looked at a dierent way.
94. Explain how limit problems that come out to /can always be converted
into limit problems that come out to 0/0 and why doing such a conversion is
useful.
95. Explain how limit problems that come out to can be converted into
limit problems that come out to 0/0 and why doing such a conversion is useful.
96. Explain how limit problems that come out to 0
0
can be converted into limit
problems that come out to 0/0 and why doing such a conversion is useful.
97. Explain how limit problems that come out to 1
can be converted into

limit problems that come out to 0/0 and why doing such a conversion is useful.
98. Use calculus to show that the area A of a sector of a circle with central
angle is A = (/2)R
2
where R is the radius and is measured in radians.
Hint: Divide the sector into n equal sectors of central angle = /n and area
A. As in the proof (see Problem 10) that
lim
0
sin()
= 1,
the area A lies between the areas of two right triangles whose areas can be
expressed in terms of R and trig functions of . Apply the Sandwich Theorem
to A = nA and use lHopitals rule or Problem 10.
99. Use calculus to show that the area of a circle of radius R is R
2
. Hint: The
area of a sector is a more general problem. (See problem 98.)
100. For which values of x is the function f(x) = x
2
+ 3x + 4 continuous?
Justify your answer with limits if necessary and draw a graph of the function
to illustrate your answer.
101. For which values of x is the function f(x) =
_
x
2
x6
x3
, if x ,= 3,
5, if x = 3,
con-
tinuous? Justify your answer with limits if necessary and draw a graph of the
function to illustrate your answer.
102. For which values of x is the function f(x) =
_
sin 3x
x
, if x ,= 0,
1, if x = 0,
contin-
uous? Justify your answer with limits if necessary and draw a graph of the
function to illustrate your answer.
141
103. Determine the value of k for which the function
f(x) =
_
_
_
sin2x
5x
, if x ,= 0,
k, if x = 0,
is continuous at x = 0. Justify your answer with limits if necessary and draw a
graph of the function to illustrate your answer.
104. What does it mean for a function f(x) to be continuous at x = a?
105. What does it mean for a function f(x) to be dierentiable at x = a?
106. What does f
(a) indicate you about the graph of y = f(x)? Explain why

this is true.
107. What does it mean for a function to be increasing? Explain how to use
calculus to tell if a function is increasing. Explain why this works.
108. What does it mean for a function to be concave up? Explain how to use
calculus to tell if a function is concave up. Explain why this works.
109. What is a horizontal asymptote of a function f(x)? Explain how to justify
that a given line y = b is a horizontal asymptote of f(x).
110. What is a vertical asymptote of a function f(x)? Explain how to justify
that a given line x = a is an vertical asymptote of f(x).
111. If f(x) = [x[, what is f
(2)?
112. Find the values of a and b so that the function
f(x) =
_
x
2
+ 3x +a, if x 1,
bx + 2, if x > 1,
is dierentiable for all values of x.
113. Graph f(x) =
_
2 x, if x 1,
x, if 0 x 1.
114. Graph f(x) =
_
2 +x, if x 0,
2 x, if x < 0.
115. Graph f(x) =
_
1 x, if x < 1,
x
2
1, if x 1.
116. Graph f(x) = x + 1/x.
117. Graph f(x) =
x
2
+ 2x 20
x 4
for 5 < x < 9.
118. Graph f(x) =
1
x
2
+ 1
.
142
119. Graph f(x) = xe
x
.
120. State Rolles theorem and draw a picture which illustrates the statement
of the theorem.
121. State the Mean Value Theorem and draw a picture which illustrates the
statement of the theorem.
122. Explain why Rolles theorem is a special case of the Mean Value Theorem.
123. Let f(x) = 1x
2/3
. Show that f(1) = f(1) but that there is no number
c in the interval (1, 1) such that f
(c) = 0. Why does this not contradict

Rolles theorem?
124. Let f(x) = (x1)
2
. Show that f(0) = f(2) but that there is no number
c in the interval (0, 2) such that f
(c) = 0. Why does this not contradict Rolles

theorem?
125. Show that the Mean Value Theorem is not applicable to the function
f(x) = [x[ in the interval [1, 1].
126. Show that the Mean Value Theorem is not applicable to the function
f(x) = 1/x in the interval [1, 1].
127. Find a point on the curve y = x
3
where the tangent is parallel to the
chord joining (1, 1) and (3, 27).
128. Show that the equation x
5
+ 10x + 3 = 0 has exactly one real root.
129. Find the local maxima and minima of f(x) = (5x 1)
2
+4 without using
derivatives.
130. Find the local maxima and minima of f(x) = (x3)
2
+9 without using
derivatives.
131. Find the local maxima and minima of f(x) = [x + 4[ + 6 without using
derivatives.
132. Find the local maxima and minima of f(x) = sin2x + 5 without using
derivatives.
133. Find the local maxima and minima of f(x) = [ sin4x + 3[ without using
derivatives.
134. Find the local maxima and minima of f(x) = x
4
62x
2
+ 120x + 9.
135. Find the local maxima and minima of f(x) = (x 1)(x + 2)
2
.
136. Find the local maxima and minima of f(x) = (x 1)
3
(x + 1)
2
.
137. Find the local maxima and minima of f(x) = x/2 + 2/x for x > 0.
138. Find the local maxima and minima of f(x) = 2x
3
24x + 107 in the
interval [1, 3].
139. Find the local maxima and minima of f(x) = sinx + (1/2) cos x in 0
x /2.
143
140. Show that the maximum value of
_
1
x
_
x
is e
1/e
.
141. Show that f(x) = x+1/x has a local maximum and a local minimum, but
the value at the local maximum is less than the value at the local minimum.
142. Find the maximum prot that a company can make if the prot function
is given by p(x) = 41 + 24x 18x
2
.
143. A train is moving along the curve y = x
2
+2. A girl is at the point (3, 2).
At what point will the train be at when the girl and the train are closest? Hint:
You will have to solve a cubic equation, but the numbers have been chosen so
there is an obvious root.
144. Find the local maxima and minima of f(x) = x + 2 sinx in [0, 2].
145. Divide 15 into two parts such that the square of one times the cube of the
other is maximum.
146. Suppose the sum of two numbers is xed. Show that their product is
maximum exactly when each one of them is half of the total sum.
147. Divide a into two parts such that the pth power of one times the qth power
of the other is maximum.
148. Which number between 0 and 1 exceeds its pth power by the maximum
amount?
149. Find the dimensions of the rectangle of area 96 cm
2
which has minimum
perimeter. What is this minimum perimeter?
150. Show that the right circular cone with a given volume and minimum
surface area has altitude equal to

2 times the radius of the base.
151. Show that the altitude of the right circular cone with maximum volume
that can be inscribed in a sphere of radius R is 4R/3.
152. Show that the height of a right circular cylinder with maximum volume
that can be inscribed in a given right circular cone of height h is h/3.
153. A cylindrical can is to be made to hold 1 liter of oil. Find the dimensions
of the can which will minimize the cost of the metal to make the can.
154. An open box is to be made out of a given quantity of cardboard of area
p
2
. Find the maximum volume of the box if its base is square.
155. Find the dimensions of the maximum rectangular area that can be fenced
with a fence 300 yards long.
156. Show that the triangle of the greatest area with given base and vertical
angle is isosceles.
157. Show that a right triangle with a given perimeter has greatest area when
it is isosceles.
144
158. What do distance, speed and acceleration have to do with calculus? Ex-
plain thoroughly.
159. A particle, starting from a xed point P, moves in a straight line. Its
position relative to P after t seconds is s = 11 + 5t + t
3
meters. Find the
distance, velocity and acceleration of the particle after 4 seconds, and nd the
distance it travels during the 4th second.
160. The displacement of a particle at time t is given by x = 2t
3
5t
2
+4t +3.
Find (i) the time when the acceleration is 8cm/s
2
, and (ii) the velocity and
displacement at that instant.
161. A particle moves along a straight line according to the law s = t
3
6t
2
+
19t 4. Find (i) its displacement and acceleration when its velocity is 7m/s,
and (ii) its displacement and velocity when its acceleration is 6m/s
2
.
162. A particle moves along a straight line so that after t seconds its position
relative to a xed point P on the line is s meters, where s = t
3
4t
2
+3t. Find
(i) when the particle is at P, and (ii) its velocity and acceleration at these times
t.
163. A particle moves along a straight line according to the law s = at
2
2bt+c,
where a, b, c are constants. Prove that the acceleration of the particle is constant.
164. The displacement of a particle moving in a straight line is x = 2t
3
9t
2
+
12t +1 meters at time t. Find (i) the velocity and acceleration at t = 1 second,
(ii) the time when the particle stops momentarily, and (iii) the distance between
two stops.
165. The distance s in meters travelled by a particle in t seconds is given by
s = ae
t
+ be
t
. Show that the acceleration of the particle at time t is equal to
the distance the particle travels in t seconds.
166. A ladder 10 feet long rests against a vertical wall. If the bottom of the
ladder slides away from the wall at a speed of 2 ft/s, how fast is the angle
between the top of the ladder and the wall changing when the angle is /4
radians?
167. A ladder 13 meters long is leaning against a wall. The bottom of the
ladder is pulled along the ground away from the wall at the rate of 2 m/s. How
fast is its height on the wall decreasing when the foot of the ladder is 5 m away
from the wall?
168. A television camera is positioned 4000 ft from the base of a rocket launch-
ing pad. A rocket rises vertically and its speed is 600 ft/s when it has risen 3000
feet. (a) How fast is the distance from the television camera to the rocket chang-
ing at that moment? (b) How fast is the cameras angle of elevation changing
at that same moment? (Assume that the telivision camera points toward the
rocket.)
169. Explain why exponential functions arise in computing radioactive decay.
145
170. Explain why exponential functions are used as models for population
growth.
171. Radiocarbon dating works on the principle that
14
C decays according to
radioactive decay with a half life of 5730 years. A parchment fragment was
discovered that had about 74% as much
14
C as does plant material on earth
today. Estimate the age of the parchment.
172. After 3 days a sample of radon-222 decayed to 58% of its original amount.
(a) What is the half life of radon-222? (b) How long would it take the sample
to decay to 10% of its original amount?
173. Polonium-210 has a half life of 140 days. (a) If a sample has a mass of 200
mg nd a formula for the mass that remains after t days. (b) Find the mass
after 100 days. (c) When will the mass be reduced to 10 mg? (d) Sketch the
graph of the mass as a function of time.
174. If the bacteria in a culture increase continuously at a rate proportional to
the number present, and the initial number is N
0
, nd the number at time t.
175. If a radioactive substance disintegrates at a rate proportional to the
amount present how much of the substance remains at time t if the initial
amount is Q
0
?
176. Current agricultural experts believe that the worlds farms can feed about
10 billion people. The 1950 world population was 2.517 billion and the 1992
world population was 5.4 billion. When can we expect to run out of food?
177. The Archer Daniel Midlands company runs two ads on Sunday mornings.
One says that when this baby is old enough to vote, the world will have one
billion new mouths to feed and the other says in thirty six years, the world
will have to set eight billion places at the table. What does ADM think the
population of the world is at present? How fast does ADM think the population
is increasing? Use units of billions of people so you can write 8 instead of
8, 000, 000, 000. (Hint: 36 = 2 18.)
178. The population of California grows exponentially at an instantaneous rate
of 2% per year. The population of California on January 1, 2000 was 20,000,000.
(a) Write a formula for the population N(t) of California t years after January
1, 2000. (b) Each Californian consumes pizzas at the rate of 70 pizzas per year.
At what rate is California consuming pizzas t years after 1990? (c) How many
pizzas were consumed in California from January 1, 2005 to January 1, 2009?
179. The population of the country of Slobia grows exponentially. (a) If its
population in the year 1980 was 1,980,000 and its population in the year 1990
was 1,990,000, what is its population in the year 2000? (b) How long will it
take the population to double? (Your answer may be expressed in terms of
exponentials and natural logarithms.)
180. If f
(x) = x 1/x
2
and f(1) = 1/2 nd f(x).
146
181.
_
(6x
5
2x
4
7x + 3/x 5 + 4e
x
+ 7
x
) dx
182.
_
(x/a +a/x +x
a
+a
x
+ax) dx
183.
_ _
x
3
x
4
+
7
3
x
2
6e
x
+ 1
_
dx
184.
_
2
x
dx 185.
_
4
2
(3x 5) dx
186.
_
2
1
x
2
dx 187.
_
1
0
(1 2x 3x
2
) dx
188.
_
2
1
(5x
2
4x + 3) dx 189.
_
0
3
(5y
4
6y
2
+ 14) dy
190.
_
1
0
(y
9
2y
5
+ 3y) dy 191.
_
4
0
xdx
192.
_
1
0
x
3/7
dx 193.
_
3
1
_
1
t
2

1
t
4
_
dt
194.
_
2
1
t
6
t
2
t
4
dt 195.
_
2
1
x
2
+ 1
x
dx
196.
_
2
0
(x
3
1)
2
dx 197.
_
1
0
u(
u +
3
u) du
198.
_
2
1
(x + 1/x)
2
dx 199.
_
3
3
_
x
5
+ 2 dx
200.
_
1
1
(x 1)(3x + 2) dx 201.
_
4
1
(
t 2/
t) dt
202.
_
8
1
_
3
r +
1
3
r
_
dr 203.
_
0
1
(x + 1)
3
dx
204.
_
2
5
x
4
1
x
2
+ 1
dx 205.
_
e
1
x
2
+x + 1
x
dx
206.
_
9
4
_
x +
1
x
_
2
dx 207.
_
1
0
_
4
x
5
+
5
x
4
_
dx
208.
_
8
1
x 1
3
x
2
dx 209.
_
/3
/4
sint dt
210.
_
/2
0
(cos + 2 sin) d 211.
_
/2
0
(cos + sin2) d
212.
_

2/3
sec xtanxdx 213.
_
/2
/3
csc xcot xdx
214.
_
/3
/6
csc
2
d 215.
_
/3
/4
sec
2
d
216.
_

3
1
6
1 +x
2
dx 217.
_
0.5
0
dx
1 x
2
147
218.
_
8
4
(1/x) dx 219.
_
ln 6
ln 3
8e
x
dx
220.
_
9
8
2
t
dt 221.
_
e
e
2
3
x
dx
222.
_
3
2
[x
2
1[ dx 223.
_
2
1
[x x
2
[ dx
224.
_
2
1
(x 2[x[) dx 225.
_
2
0
(x
2
[x 1[) dx
226.
_
2
0
f(x) dx where f(x) =
_
x
4
, if 0 x < 1,
x
5
, if 1 x 2
.
227.
_

f(x) dx where f(x) =

_
x, if x 0,
sinx, if 0 < x .
228. True or false?
_
1
1
3
t
4
dt =
1
t
3
1
1
= 1 + 1 = 0.
229. Explain what a Riemann sum is and write the denition of
_
b
a
f(x)dx as
a limit of Riemann sums.
230. State the Fundamental Theorem of Calculus.
231. (1) Water ows into a container at a rate of three gallons per minute for
two minutes, ve gallons per minute for seven minutes and eleven gallons per
minute for two minutes. How much water is in the container? (2) Water ows
into a container at a rate of t
2
gallons per minute for 0 t 5. How much
water is in the container?
232. Let f(x) be a function which is continuous and let A(x) be the area under
f(x) from a to x. Compute the derivative of A(x) by using limits.
233. Why is the Fundamental Theorem of Calculus true? Explain carefully
and thoroughly.
234. Give an example which illustrates the Fundamental Theorem of Calculus.
In order to do this compute an area by summing up the areas of narrow rectan-
gles and then show that applying the Fundamental Theorem of Calculus gives
the same answer.
235. Sketch the graph of the curve y =

x + 1 and determine the area of the
region enclosed by the curve, the x-axis and the lines x = 0, x = 4.
236. Make a sketch of the graph of the function y = 4 x
2
and determine the
area enclosed by the curve, the x-axis and the lines x = 0, x = 2.
237. Find the area under the curve y =

6x + 4 and above the x-axis between
x = 0 and x = 2. Draw a sketch of the curve.
148
238. Graph the curve y = x
3
and determine the area enclosed by the curve and
the lines y = 0, x = 2 and x = 4.
239. Graph the function f(x) = 9 x
2
, 0 x 3, and determine the area
enclosed between the curve and the x-axis.
240. Graph the curve y = 2
1 x
2
, x [0, 1], and nd the area enclosed
between the curve and the x-axis.
241. Determine the area under the curve y =

a
2
x
2
and between the lines
x = 0 and x = a.
242. Graph the curve y = 2
9 x
2
and determine the area enclosed between
the curve and the x-axis.
243. Graph the area between the curve y
2
= 4x and the line x = 3. Find the
area of this region.
244. Find the area bounded by the curve y = 4 x
2
and the lines y = 0 and
y = 3.
245. Find the area bounded by the curve y = x(x 3)(x 5), the x-axis and
the lines x = 0 and x = 5.
246. Find the area enclosed between the curve y = sin2x, 0 x /4 and the
axes.
247. Find the area enclosed between the curve y = cos 2x, 0 x /4 and
the axes.
248. Find the area enclosed between the curve y = 3 cos x, 0 x /2 and
the axes.
249. Find the area enclosed between the curve y = cos 3x, 0 x /6 and
the axes.
250. Find the area enclosed between the curve y = tan
2
x, 0 x /4 and
the axes.
251. Find the area enclosed between the curve y = csc
2
x, /4 x /2 and
the axes.
252. Find the area of the region bounded by y = 1, y = 2, x = y
3
, and x = 0.
253. Find the area of the region bounded by the parabola y = 4x
2
, x 0, the
y-axis, and the lines y = 1 and y = 4.
254. Find the area of the region bounded by the curve y = 4 x
2
and the lines
y = 0 and y = 3.
255. Graph y
2
+ 1 = x, x 2 and nd the area enclosed by the curve and the
line x = 2.
256. Graph the curve y = x/ + 2 sin
2
x and write a denite integral whose
value is the area between the x-axis, the curve and the lines x = 0 and x = .
Do not evaluate the integral. Do specify the limits of integration.
149
257. Find the area bounded by y = sinx and the x-axis between x = 0 and
x = 2. Hint: Make a careful drawing to decide what area is intended.
258. Find the area bounded by the curve y = cos x and the x-axis between
x = 0 and x = 2.
259. Give an example which shows that
_
b
a
f(x) dx is not always the true area
bounded by the curves y = f(x), y = 0, x = a, and x = b even though f(x) is
continuous between a and b.
260. Find the area of the region bounded by the parabola y
2
= 4x and the line
y = 2x.
261. Find the area bounded by the curve y = x(2 x) and the line x = 2y.
262. Find the area bounded by the curve x
2
= 4y and the line x = 4y 2.
263. Calculate the area of the region bounded by the parabolas y = x
2
and
x = y
2
.
264. Find the area of the region included between the parabola y
2
= x and the
line x +y = 2.
265. Find the area of the region bounded by the curves y =

x and y = x.
266. Use integration to nd the area of the triangular region bounded by the
lines y = 2x + 1, y = 3x + 1 and x = 4.
267. Find the area bounded by the parabola x
2
2 = y and the line x +y = 0.
268. Graph the curve y = (1/2)x
2
+ 1 and the straight line y = x + 1 and nd
the area between the curve and the line.
269. Find the area of the region between the parabolas y
2
= x and x
2
= 16y.
270. Find the area of the region enclosed by the parabola y
2
= 4ax and the
line y = mx.
271. Find a so that the curves y = x
2
and y = a cos x intersect at the points
(x, y) = (
4
,

2
16
). Then nd the area between these curves.
272. Write a denite integral whose value is the area of the region between the
two circles x
2
+ y
2
= 1 and (x 1)
2
+ y
2
= 1. Find this area. If you cannot
evaluate the integral by calculus you may use geometry to nd the area. Hint:
The part of a circle cut o by a line is a circular sector with a triangle removed.
circles x
2
+ y
2
= 4 and (x 2)
2
+ y
2
= 4. Do not evaluate the integral. Do
specify the limits of integration.
curves x
2
+ y
2
= 2 and x = y
2
Do not evaluate the integral. Do specify the
limits of integration.
150
curves x
2
+ y
2
= 2 and x = y
2
. Find this area. If you cannot evaluate the
integral by calculus you may use geometry to nd the area. Hint: Divide the
region into two parts.
276. Write a denite integral whose value is the area of the part of the rst
quadrant which is between the parabola y
2
= x and the circle x
2
+y
2
2x = 0.
Find this area. If you cannot evaluate the integral by calculus you may use
geometry to nd the area. Hint: Draw a careful graph. Divide a semicircle in
two.
277. Find the area bounded by the curves y = x and y = x
3
.
278. Graph y = sinx and y = cos x for 0 x /2 and nd the area enclosed
by them and the x-axis.
279. Write a denite integral whose value is the area inside the ellipse
x
2
a
2
+
y
2
b
2
=
1. Evaluate this area. Hint: After a suitable change of variable, the denite
integral becomes the denite integral whose value is the area of a circle.
280. Using integration nd the area of the triangle with vertices (1, 1), (0, 5)
and (3, 2).
281. Find the volume that results by rotating the triangle 1 x 2, 0 y
3x 3 around the x axis.
3x 3 around the y axis.
3x 3 around the line x = 1.
3x 3 around the line y = 1.
285. Find the volume that results by rotating the semicircle y =

R
2
x
2
about the x-axis.
286. A triangle is formed by drawing lines from the two endpoints of a line
segment of length b to a vertex V which is at a height h above the line of the
line segment. Its area is then A =
_
h
y=0
dA where dA is the area of the strip
cut out by two parallel lines separated by a distance of dz and at a height of z
above the line containing the line segment. Find a formula for dA in terms of
b, z, and dz and evaluate the denite integral.
287. A pyramid is formed by drawing lines from the four vertices of a rectangle
of area A to an apex P which is at a height h above the plane of the rectangle.
Its volume is then V =
_
h
z=0
dV where dV is the volume of the slice cut out by
two planes parallel to the plane of the rectangle and separated by a distance of
dz and at a height of z above the plane of the rectangle. Find a formula for dV
in terms of A, z, and dz and evaluate the denite integral.
151
288. A tetrahedron is formed by drawing lines from the three vertices of a
triangle of area A to an apex P which is at a height h above the plane of the
triangle. Its volume is then V =
_
h
z=0
dV where dV is the volume of the slice
cut out by two planes parallel to the plane of the triangle and separated by a
distance of dz and at a height of z above the plane of the rectangle. Find a
formula for dV in terms of A, z, and dz and evaluate the denite integral.
289. A cone is formed by drawing lines from the perimeter of a circle of area
A to an apex V which is at a height h above the plane of the circle. Its volume
is then V =
_
h
z=0
dV where dV is the volume of the slice cut out by two planes
parallel to the plane of the circle and separated by a distance of dz and at a
height of z above the plane of the rectangle. Find a formula for dV in terms of
A, z, and dz and evaluate the denite integral.
290. (a) A hemispherical bowl of radius a contains water to a depth h. Find
the volume of the water in the bowl. (b) Water runs into a hemispherical bowl
of radius 5 ft at the rate of 0.2 ft
3
/sec. How fast is the water level rising when
the water is 4 ft deep?
291. (Alternate wording for previous problem.) A hemispherical bowl is ob-
tained by rotating the semicircle x
2
+ (y a)
2
= a
2
, y a about the y-axis.
It is lled with water to a depth of h, i.e. the water level is the line y = h.
(a) Find the volume of the water in the bowl as a function of h. (b) Water runs
into a hemispherical bowl of radius 5 ft at the rate of 0.2 ft
3
/sec. How fast is
the water level rising when the water is 4 ft deep? (Hint: Use the method of
related rates and the Fundamental Theorem.)
292. A vase is constructed by rotating the curve x
3
y
3
= 1 for 0 y 8
around the y axis. It is lled with water to a height y = h where h < 8.
(a) Find the volume of the water in terms of h. (Express your answer as a
denite integral. Do not try to evaluate the integral.) (b) If the vase is lling
with water at the rate of 2 cubic units per second, how fast is the height of the
water increasing when this height is 2 units?
152
Chapter IX
Notes for TAs
40.1. My main objective is to get the students to use the notation correctly.
When grading, point out incorrect syntax, both mathematical and grammatical.
40.2. Read your email daily. After reading a message from me reply to me
(OK suces) so I know that you got it. I would be very happy if we all discussed
what is going on in the course via email. Back in the stone age we used to have
weekly meetings, but we can accomplish the same thing via email.
40.3. Learning occurs only when students are active, not passive. For this
reason I encourage students to interrupt the lecture with questions. I would be
delighted if a TA would interrupt once in a while, even if that TA does not need
to ask a question. It might encourage the students to speak up.
40.4. The students should do the exercises from the notes as soon as I cover
the corresponding material in lecture. If there are not enough problems, you
can assign more (either from Chapter VIII or from the text) or you can ask me
to assign more.
40.5. Grade Exercises 1.4 and 1.5 at the end of Section 1 of these notes.
The Math Department has an Early Warning System which aims to detect
students who may get into trouble as early as possible so as to get them into
the Tutorial Program. Weak algebra skills increase the likelihood that a student
will have diculty. Exercise 5.8 is also good preparation for word problems later
in the course. Also use this exercise to look for at risk students.
40.6. We dont allow calculators on exams. Encourage the students to leave
arithmetic undone; it makes grading easier. Tell them that an answer like
x = 1 + 3 is acceptable (if correct) but an answer like x = 1 + 3 = 5 will be
penalized. On the other hand, sometimes a little algebraic simplication at the
beginning of a problem makes the rest of the problem less error prone. For
example, I would replace 3x + 1 + 5x by 8x + 1 early in a calculation to avoid
copying errors.
40.7. I like to assign complicated computations as homework, especially when
there is another easier way to do the problem which I also assign. However, I
think the algebra we ask students to do on tests should not be so complicated.
40.8. Inverse functions are one of the trickiest things in the course. I like to
hammer on the fact that y = f(x) and x = f
1
(y) have the same graph, i.e.
y = f(x) x = f
1
(y).
Rather than interchanging x and y to to draw the graph of the inverse function,
I prefer to label the horizontal axis as y and the vertical axis as x. (The eect
is still reection in the diagonal.) Do Exercise 2.10 to reinforce this. I like to do
the inverse trig functions right after the trig functions to accustom the students
to inverse functions.
153
40.9. The change of variables formula in 7.9 means that you dont need to
do long division of polynomials when nding the limit of a rational function.
Instead of dividing by x a you replace x by a +h and then factor out h.
40.10. There are several good reasons for the Early Transcendentals ap-
proach:
1. By introducing the most important functions early on we assure that the
students get experience with them and are not short changed because we
run out of time at the end of the semester.
2. We make the course more challenging in the rst few weeks so that stu-
dents are not lulled into overcondence. (In the standard approach stu-
dents usually nd the rst exam easy.)
3. We can do more interesting problems earlier.
154

Mate Matic A

Uploaded by

Copyright:

Available Formats

Mate Matic A

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mate Matic A

Uploaded by

Copyright:

Available Formats

Math 221 1st Semester Calculus

Lecture Notes for Fall 2006.

always means the positive square root) these formulas

) > f(c) when c

whose value at the point a is given by

contains the interval I; f is said to

(4) we might take x = 4 and h = 0.001 0,

(4) we might take a = 4 and x = 4.001 4,

= 0 and the Quotient Rule reduces to

. Taking u = v in the Product Rule gives

which is the Power Rule with n = 2. Using this and taking v = u

, the derivative of the

(x) and with u = 1 +x

(x) in two ways, rst by

(x)? What is the relation

(x)? Hint: f(x) = g(x)

. The third derivative of a function f is the derivative of the

. The nth derivative of f is sometimes

(x) of the function

(1) we can dierentiate () and solve:

(x) by dierentiating ():

) for dy/dx in Exam-

(a) = 0 since (g(x) g(a))/(x a) g

(a) for x a. As before the

(a) = 0 for x a and

u, the direct method doesnt

u and then compute

(0). If there is insu-

y. Thus f(3) = 9 so by (#) f

(34). Warning: Dont try to nd a

) its derivative is given by

lnx using the Chain Rule as follows:

(a)(x a), M(y) = g(b) +g

(c) exists we have that

(0) does not exist. On any interval containing 0 it attains its

(x) = 0 for x = 1. The point x = 1 does not lie in the interval 0 x 2

(c) has the

(c) > 0 so f(x

(c) < 0 so f(x

(x) > 0 for all x in

(c) = 0 at some c in I, then f(x) assumes

(x) < 0 for all x in some open

(c) = 0 at some c in I, then f(x) assumes its maximum

(x) > 0 for a < x < b. Then f

(x) is increasing on the interval. But f

(x) < 0 for a < x < c and f

(x) > 0 for c < x < b, i.e. f(x) is decreasing

(x) < 0 for a < x < c

(x) is positive at every point x in the interval. The derivative

so, by the Monotonicity Theorem, the derivative f

is increasing this means that g

(x) > 0 for c < x < b. Hence g is decreasing on the interval

(x) and g is also concave up.

(c) = the slope of the tangent line at c, so g

is increasing this means that g

(c) = 0 for a < x < c and g

(12). Hint: For what values of x is f

(x) vanishes at a point of inection.

(x) and the sign of f

(x). (Now you know on