Section 8: The Rule of Chains

~
w
8. The Rule of Chains toc
Addition, subtraction, multiplication, and division are some of the
ways we use to build up new functions from old, and we have seen
how the operation of differentiation interacts with these basic arith-
metic processes. There is, as you know, one other way of combining
functions: composition.
A thorough study of composition of functions has already been taken
up; included in that discussion were the topics of composition, the me-
chanics of composition, and the very important notion of uncomposing
or decomposing a function.
In this section, we study how differentiation interacts with composi-
tion. The formal statement of this relationship is the Chain Rule, and
is stated immediately below.
As it turns out, the most important skill associated with a consistently
correct use of the Chain Rule is the ability to realize that a given func-
tion is, in fact, the composition of other functions, and the ability to
Section 8: The Rule of Chains

identify these functions. This realiaztion and identification is roughly

the process of uncomposing mentioned and referenced above.

Theorem 8.1. (The Chain Rule) Let y = f (u) and u = g(x) be

functions such that f is compatable for composition with g. Suppose
g is differentiable at x and f is differentiable at u = g(x), then the
composite function f ◦ g is differentiable at x, and
(f ◦ g)0 (x) = f 0 (g(x))g 0 (x). (1)

Proof.
Theorem Notes: The Chain Rule can be very mystifying when you see
it and use it the first time. Hopefully, this article will clear this up for
you.
The Chain Rule allows us to differentiate a more complicated
function by multiplying together the derivatives of the functions used
to compose the parent function.
In the theorem, I have conveniently labeled the functions and
their variables in such a way as to suggest the composition: y = f (u)
Section 8: The Rule of Chains

and u = g(x). Needless to say, generally, functions are not labeled in

the most pleasing manner.
Theorems are often stated in a very clean, precise and peda-
gogical manner, or, as often happens, not stated in a way they are
actually used in practice. In the case of this theorem, usually, you are
given a function, say F (x) = (x2 + 1)10 , to differentiate, the problem
becomes to uncompose the function. You must realize that F is the
composition of two functions: f (x) = x10 and g(x) = x2 + 1. (Check
for yourself that F (x) = f (g(x)).) Then you apply the theorem as
stated. What is not mentioned in the theorem is that you must un-
compose your function to apply the theorem. These decomposition
methods were covered in a paragraph on uncomposing functions.
Example 8.1. Find the derivative of F (x) = (x2 + 1)10 using the
Chain Rule.
Listed below is a chain of thoughts that are necessary to be successful
at applying the Chain Rule.
Chain Rule: The Procedure.
Section 8: The Rule of Chains

Step: 1 Recognize and realize that the given function is a compo-

sition of two (or more!) other functions.
Step: 2 Decompose the given function.
Step: 3 Apply the Chain Rule
Here’s another example, this one numerical.
Example 8.2. Calculate F 0 (2), where F (x) = (3x3 − x)7 .
Actually, both of the past examples followed same pattern. Presently,
we will identify this pattern; in this way, we don’t have to go through
such painful analysis every time.
Review the method of solution of Example 8.1 and Example 8.2,
then solve the following exercise. Solve it completely before looking
at the solution.
Exercise 8.1. Find the derivative of F (x) = (3x4 + 5x)1/2 using the
Chain Rule.
The Chain Rule can be applied in many situations. In the above ex-
amples, the Chain Rule is utilized to evaluate the derivative of specific
Section 8: The Rule of Chains

functions. In the next example, the Chain Rule is used to differenti-

ate the composition of an abstract function with a specific function.
Confused? Read on.
Example 8.3. Let f be a differentiable function, and define a new
function by
F (x) = f (x3 ).
Calculate F 0 (x) using the chain rule.
These kinds of problem types are encountered in differential equations
and other higher mathematical disciplines.
Study the reasoning of the previous example, then tackle the following
problem.
Exercise 8.2. Let f be a differentiable function, and define a new
function by
F (x) = f (1/x4 ).
Calculate F 0 (x) using the chain rule.
Section 8: The Rule of Chains

Exercise 8.3. Let g be a differentiable function, and define

F (x) = [g(x)]6 .
Calculate F 0 (x) using the Chain Rule.

8.1. Chaining with Leibniz

Let’s look at the Chain Rule from the point view of the Leibniz no-
tation. (Can notation have a point of view?)
Let y = f (u) and u = g(x) be compatible for composition. When we
compose, this establishes the y as a function of the x variable. If y is
considered a function of x, then the functional relationship is given
by y = (f ◦ g)(x) = f (g(x)). The Leibniz notation for that is

dy the derivative of the y-variable when it is

(f ◦ g)0 (x) = = considered a function of x, or with respect
dx
to x.
Section 8: The Rule of Chains

But y is naturally a function of u since y = f (u); thus

dy the derivative of the y-variable when it is

0
f (u) = = considered a function of u, or with respect
du
to u
But u is naturally a function of x since u = g(x); so,

du the derivative of the u-variable when it is

g 0 (x) = = considered a function of x, or with respect
dx
to x
See the discussion on the Leibniz Notation for a refresher course.
The Chain Rule can now be translated. The Chain Rule

(f ◦ g)0 (x) = f 0 (g(x))g 0 (x)

becomes
dy dy du
= . (2)
dx du dx
Let’s make this formula into a big deal:
Section 8: The Rule of Chains

The Chain Rule: Let y = f (u) and u = g(x) be differentiable

and compatible for composition, then
dy dy du
= .
dx du dx

The next example illustrates how the Leibniz form of the chain rule
is used.
d
Example 8.4. Calculate: (5x4 − 12x2 )3 .
dx
Use the Leibniz Notation to solve the following problem.
d
Exercise 8.4. Calculate: (3x3 − 6x)1/2 .
dx
Section 8: The Rule of Chains

8.2. The Power Rule Revisited

As was advertised earlier, we don’t to go through this painful pulling
of teeth. All example thus far followed the same pattern. Let us state
the Generalized Chain Rule!
Let u = f (x) be a differentiable function of x and let r ∈ Q, the set
of all rational numbers. Consider the problem of differentiating the
function [f (x)]r . We use chain rule techniques:
y = ur where u = f (x).
We have our decomposed function setup, and we want to calculate
dy/dx. This is a job for the Chain Rule!
dy dy du
= chain it
dx du dx
d ur du
=
du dx
du
= rur−1 Power Rule (3)
dx
Section 8: The Rule of Chains

Now, substituting for u = f (x) we obtain the formula,

d
[f (x)]r = r[f (x)]r−1 f 0 (x).
dx
This is a very nice formula, but I prefer the more laconic version (3)

Generalized Power Rule:

Let u be a function of x and r ∈ Q, the set of rational
numbers, then
d ur du
= rur−1 .
dx dx

Utilizing the Generalized Power Rule, we can differentiate complex

functions with great ease.
d
Example 8.5. (Skill Level 1) Calculate (1 − 3x3 )10 .
dx
Section 9: The Trigonometric Functions

d 1
Example 8.6. Calculate .
ds (s − s + 1)3/4
4

d
Exercise 8.5. Calculate (4x2 + 1)23 .
dx
 5
d x
Exercise 8.6. Calculate .
dx 1 + x
d 3√
Example 8.7. Calculate x 1 + 3x2 .
dx
d 4
Exercise 8.7. Calculate x (2x + 1)3/2 .
dx
Let’s combine the Power Rule with the quotient rule.
√
d w w
Example 8.8. Calculate .
dw (3w3 + 1)6
d (s + 1)3
Exercise 8.8. Calculate .
ds (2s + 1)5
Section 9: The Trigonometric Functions

9. The Trigonometric Functions

In this section, we tackle the problem of differentiating the Trigono-
metric functions.

9.1. Development of Trig Formulas

Let f (x) = sin(x). If we wanted to calculate the derivative of f we
first setup the Difference Quotient:
f (x + h) − f (x) sin(x + h) − sin(x)
=
h h
sin(x) cos(h) + cos(x) sin(h) − sin(x)
= (1)
h
sin(x)(cos(h) − 1) + cos(x) sin(h)
=
h
cos(h) − 1 sin(h)
= sin(x) + cos(x) (2)
h h
Section 9: The Trigonometric Functions

Where we have used the additive formula for sin(x) in (1) in the above
manipulations. Recall,
sin(x + h) = sin(x) cos(h) + cos(x) sin(h).
Now, we take the limit of (2) as h goes to 0.
 
0 cos(h) − 1 sin(h)
f (x) = lim sin(x) + cos(x)
h→0 h h
cos(h) − 1 sin(h)
= sin(x) lim + cos(x) lim (3)
h→0 h h→0 h
The equation (3) now makes is it clear the nub of the problem: We
need to calculate two limits
sin(h) cos(h) − 1
lim lim (4)
h→0 h h→0 h
But these two limits have been already been studied. We have shown
that
sin(h) cos(h) − 1
lim =1 lim =0 (5)
h→0 h h→0 h
Section 9: The Trigonometric Functions

Now we are ready to calculate the derivative of the function sin(x).

It’s been awhile, but let’s continue our derivative calculations. Recall,
f (x) = sin(x),

cos(h) − 1 sin(h)
f 0 (x) = sin(x) lim + cos(x) lim from (3)
h→0 h h→0 h
= sin(x)(0) + cos(x)(1) from (5)
= cos(x)
Thus, we have shown that if f (x) = sin(x), then f 0 (x) = cos(x). Using
the Leibniz notation we obtain
d
sin(x) = cos(x) x ∈ R. (6)
dx

Let’s declare victory over the problem of differentiating the trigono-

metric functions by stating complete results.
Section 9: The Trigonometric Functions

Theorem 9.1. The derivatives of the trigonometric functions are

given by the following sets of formulas:
d d
(1) sin(x) = cos(x) (2) cos(x) = − sin(x)
dx dx
d d
(3) tan(x) = sec2 (x) (4) cot(x) = − csc2 (x)
dx dx
d d
(5) sec(x) = sec(x) tan(x) (6) csc(x) = − csc(x) cot(x)
dx dx
Proof.
Theorem Notes: I have arranged the trigonometric functions into three
sets of two formulas each. If you stare at the three sets, you might see
a pattern. Observing this pattern enables you to reduce from 6 the
number of formulas to 3 — that’s a 50% decrease in the number of
allocated brain cells!
Each trig function has a “co”-function. The cofunction of sin(x)
is the cos(x). (The cofunction of cos(x) is sin(x).) The cofunction of
tan(x) is cot(x), and the cofunction of sec(x) is csc(x).
Section 9: The Trigonometric Functions

Formula (2) is obtained by “co-ing” formula (1), and prefixing

a negative sign. Formula (4) is obtained by “co-ing” formula (3), and
prefixing a negative sign. Formula (6) is obtained by “co-ing” formula
(5) and prefixing a negative sign. That’s the pattern.
Therefore, it suffices to know formulas (1), (3), and (5), and
know the “co-ing” trick.
d d
Exercise 9.1. Calculate [3 cos(x)] and [−5 sec(t)]. You may
dx dt
want to review the trigonometric formulas
Example 9.1. Find the slope of the line tangent to the graph of
y = sin(x) at x = π/3. Find the equation of the line tangent to the
graph at x = π/3.
Now, here’s one for you.
Exercise 9.2. Find the slope of the line tangent to the graph of
y = tan(x) at x = π/3. Find the equation of the line tangent to the
graph at x = π/3.
Section 9: The Trigonometric Functions

d d
Exercise 9.3. Using sin(x) = cos(x) and cos(x) = − cos(x)
dx dx
d
only, calculate tan(x).
dx
Now that we have the trig functions, we can combine them with other
functions through addition, multiplication, and division.
d 3
Example 9.2. Calculate x sin(x).
dx
To test your understanding of the Chain Rule, consider the following
exercise.
Exercise 9.4. Find the derivative of F (x) = sin(6x3 ) using the
Chain Rule.

9.2. The Chain Rule Revisited

Let’s apply the Chain Rule to a very common situation.
Section 9: The Trigonometric Functions

Suppose we have a function y = sin(x2 ), and we want to calculate

dy
. Set up the composition: Let y = sin(u), and u = x2 , then by the
dx
Chain Rule,
dy dy du
=
dx du dx
d d x2
= sin(u)
du dx
= cos(u)(2x)
= cos(x2 )(2x)
Thus, we have shown
d
sin(x2 ) = cos(x2 )(2x) = 2x cos(x2 ).
dx
Using this same reasoning process, we can take the basic trig formulas,
and generalize them to allow arbitrary differentiable arguments.
Section 9: The Trigonometric Functions

Let u be a differentiable function of x, then

d du d du
(1) sin(u) = cos(u) (2) cos(u) = − sin(u)
dx dx dx dx
d du d du
(3) tan(u) = sec2 (u) (4) cot(u) = − csc2 (u)
dx dx dx dx
d du d du
(5) sec(u) = sec(x) tan(u) (6) csc(u) = − csc(x) cot(u)
dx dx dx dx
d
Example 9.3. (Skill Level 0) Calculate cos(3x4 ).
dx
A piece of good advice: Do what a formula says to do, no more, no
less, and you can’t go wrong. Have the courage to do the obvious.
d
Exercise 9.5. Calculate tan(sin(x)). (Hint: Follow the advice
dx
above)
Having gained experience, try this one on for size.
d
Exercise 9.6. (Chaining Observed) Calculate sin(tan2 (3x3 )).
dx
Section 10: Higher Order Derivatives

Summary: Let’s catalog our new formulas one more time!

Trigonometric Differentiation Formulas:

d du d du
(1) sin(u) = cos(u) (3) tan(u) = sec2 (u)
dx dx dx dx
d du d du
(2) cos(u) = − sin(u) (4) cot(u) = − csc2 (u)
dx dx dx dx
d du
(5) sec(u) = sec(x) tan(u)
dx dx
d du
(6) csc(u) = − csc(x) cot(u)
dx dx
Section 10: Higher Order Derivatives

10. Higher Order Derivatives

Given a differentiable function f , we have seen that f 0 is also a func-
tion. We have also seen the Prime Notation convention: if a function
has a name, the name of its derivative is obtained by taking the parent
function and post-fixing a prime (0 ).
Because f 0 is itself a function, we can attempt differentiating it. The
derivative of f 0 is denoted by f 00 and is called the second derivative
of f .
Example 10.1. For f (x) = x5 calculate the first and second deriva-
tives of f .
Similarly, f 00 is a function as well, so we may differentiate it. The
derivative of f 00 is denoted by f 000 and is called the third derivative of
f . . . and so on.
Notation. The prime notation has a weakness. Write symbolically,
the tenth derivative of f . Answer : f 0000000000 . Hopefully, I counted my
primes correctly. At some point, the prime notation becomes a burden,
Section 10: Higher Order Derivatives

so we switch over to another. The fourth derivative of f is denoted by

f (4) , the fifth derivative is denoted by f (5) . In general, if n ∈ N, then
the nth derivative is denoted
f (n) = the nth derivative of f .
The integer n in the derivative f (n) is called theorder of the derivative.
Example 10.2. (Example 10.1Continued) For f (x) = x5 , calculate
all derivatives of f .
When the function has an anonymous name, we use the prime notation
Section 10: Higher Order Derivatives

as well.
y = sin(2x) (1)
0
y = 2 cos(2x)
y 00 = −4 sin(2x)
y 000 = −8 cos(2x)
y (4) = 16 sin(2x)
y (5) = 32 cos(2x)
y (6) = −64 sin(2x).
In the case of the sine function, we get nonzero derivatives of all orders.
Exercise 10.1. Verify that for y = sin(2x), a general formula for
y (n) is
y (n) = 2n sin(2x + nπ/2) n = 1, 2, 3, 4, . . . .
Section 10: Higher Order Derivatives

The Leibniz notation for higher order derivatives may seem a little
strange at first. If y = f (x), then
dy
y0 =
dx  
d dy d2 y
y 00 = = 2
dx dx dx
 2 
d d y d3 y
y 000 = =
dx dx2 dx3
 3 
d d y d4 y
y (4) = = .
dx dx3 dx4
In general, for n ∈ N, the nth derivative with respect to x is given by
dn y d dn−1 y
= ,
dxn dx dxn−1
and is calculated by differentiating the (n − 1)st -derivative.
Section 11: Implicit Differentiation

For example, if y = x4 , then

dy
= 4x3
dx
d2 y
= 12x2
dx2
d3 y
= 24x
dx3
d4 y
= 24
dx4
d5 y
=0
dx5
Note that the derivative of each line is the next line.
x
Exercise 10.2. Let y = , calculate y 0 ,y 00 , and y 000 .
2x + 1
Section 11: Implicit Differentiation

11. Implicit Differentiation

The title of this section is implicit differentiation. What is explicit
differentiation? Read on.

11.1. Statement of the Problem

Consider the following equation: x2 + y 2 = 1. As you know, the graph
of this equation is a circle of radius r = 1 with center at the origin.
This curve does not define y as a function of x as it fails the vertical
line test.
Despite not being a function, one feels that x2 + y 2 = 1 has tangent
lines to its graph. How do we calculate dy/dx? The techniques de-
veloped up to this point requires that y be an explicit function of x.
Then we can differentiate — maybe.
Solving for y in x2 + y 2 = 1, we get
p
y = ± 1 − x2 |x| ≤ 1. (1)
Section 11: Implicit Differentiation

The presence of the ± again suggests that y is not a function of x:

Each x has 2 associated y-values.
Another point of view of (1) is that the circle can be described by a
single equation, x2 + y 2 = 1, but requires two functions to describe
its curve; namely
p
y = 1 − x2 |x| ≤ 1 (2)
and p
y = − 1 − x2 |x| ≤ 1 (3)
The first function describes the upper semi-circle and the second func-
tion describes the lower semi-circle.
Therefore, if we wanted the slope of a line that is tangent to the upper
semi-circle, we would differentiate (2) to get
dy x
= −√ |x| < 1 (4)
dx 1 − x2
Section 11: Implicit Differentiation

Similarly, if we wanted the slope of a line that is tangent to the lower

semi-circle, we would differentiate (3) to get
dy x
=√ |x| < 1 (5)
dx 1 − x2
√
For example, the point ( x, y ) = ( 1/2, 3/2 ) lies on the unit circle
x2 + y 2 = 1. Find the slope of the tangent line to this point. Well, the
given point lies on the upper semi-circle, so we would use the formula
(4):
dy 1/2
= −p
dx √
( x,y )=( 1/2, 3/2 ) 1 − (1/2)2
1
= −√ . (6)
3
Section 11: Implicit Differentiation

Or, we may
√ be interested in a point on the lower semi-circle: ( x, y ) =
( 1/2, − 3/2 ). In this case, we use the formula (5),
dy 1/2
=p
dx √
( x,y )=( 1/2, 3/2 ) 1 − (1/2)2
1
=√ . (7)
3
In this case, all calculations are the same except for the sign.
As you can see from this extended example, this is the kind of pro-
cedure you would have to do if the curve is given in equational form:
(1) solve for y in terms of x – you may get several solutions, each
defining y as a function of x; (2) differentiate the functional relation
of interest in order to obtain dy/dx.
This method may be rather drawn out, or even impossible. Some
equations are so complicated that you cannot algebraically solve of y
in terms of x. In the equation,
x10 − 2xy − y 10 = 0.
Section 11: Implicit Differentiation

it is not possible to algebraically solve for y in terms of x. Yet, the

equation defines some curve and this curve has tangent lines. If our
goal was to calculate dy/dx we could not do so because we cannot
write y in the form of a function of x!
Therefore, another more powerful, more flexible method is needed.
This is the method of implicit differentiation. It is explained in the
next section.

11.2. The Technique Explained

Let’s state, in clear terms, the problem.
The Given. Given an equation
F (x, y) = c, (8)
that is, we are given some equation (F (x, y) = c) involving the vari-
ables x and y.
dy
The Problem. Find .
dx
Section 11: Implicit Differentiation

The Procedure.
1. Treat y as if it were an explicit function of x.
2. Differentiate both sides of (8) with respect to x. Use the various
differentiation formulas wherever applicable. When finished, we
will have an equation involving x, y, and dy/dx. Symbolically,
dy
G(x, y,
)=0 (9)
dx
3. Solve the equation (9) for dy/dx.
Let’s implement this procedure in a series of examples.
dy
Example 11.1. Find , for x2 + y 2 = 1 using implicit differentia-
dx
tion.
Next up is an example in which it is impossible to solve for y; conse-
quently, implicit differentiation is the only way we have to differentiate
this curve.
dy
Example 11.2. Calculate , where x10 − 2xy − y 10 = 0.
dx
Section 11: Implicit Differentiation

dy
Example 11.3. Consider the equation x sin(xy) = 1, find .
dx
Exercise 11.1. Consider the ellipse
x2 y2
+ = 1.
4 9
Find dy/dx using the technique of implicit differentiation.
Exercise 11.2. Consider the curve defined by the equation x2 y 7 −
x3 y 2 = 1. Calculate dy/dx.

11.3. Higher-Order Derivatives

Now let’s address the problem of calculating higher-order derivatives
using implicit differentiation.
dy d2 y d3 y
Example 11.4. Consider x2 + y 2 = 1. Calculate , , and
dx dx2 dx3
using implicit differentiation.
Section 11: Implicit Differentiation

Summary of the Technique. Here is the synthesis of the method for

calculating higher-order derivatives using implicit differentiation.
1. Given an equation: F (x, y) = c. Differentiate implicitly; you will
obtain another equation F0 (x, y, y 0 ) = 0. Solve this equation
for y 0 to obtain the solution:
y 0 = F1 (x, y). (10)
This means that y 0 is represented in terms of x and y.
2. Take equation (10) and differentiate it with respect to x. You
will obtain an equation of the general form:
y 00 = F2 (x, y, y 0 ). (11)
Take your answer for y 0 in (10) and substitute it into (11) to
obtain something like:
y 00 = F3 (x, y), (12)
that is, work to write your answer in terms of the basic variables
x and y.
Section 12: The Mean and the Extreme

3. To calculate the third derivative, y 000 , with respect to x, take

(12) and differentiate both sides with respect to x to obtain an
expression:
y 000 = F4 (x, y, y 0 ).
As before, take (10) and substitute for y 0 into this equation. We
will arrive at the equation:
y 000 = F5 (x, y). (13)
This represents y 000 in terms of x and y.
4. For higher derivatives, you would continue in the same manner.
Exercise 11.3. In the spirit of the above outline, write out the steps
for calculating y 0000 , also known as y (4) .
dy d2 y d3 y
Exercise 11.4. Let sin(y) = x. Calculate , , and using
dx dx2 dx3
the method of implicit differentiation. In all cases, leave the answer
in terms of x and y.
Section 12: The Mean and the Extreme

12. The Mean and the Extreme

This section still under construction. It is my intention to develop
a series of topics covering theory and techniques of finding local
and absolute extrema.
In this section we discuss the Extreme Value Theorem, which is
a fundamental tool in optimization problems, and the Mean Value
Theorem, which has many theoretical consequences and many useful
applications.

12.1. Extrema: Introductory Concepts

You know, as someone who has seen many graphs, that some curves
have high points and/or low points. In this section we give a formal
definition these terms and present a famous theorem that speaks to-
wards the existence of these high and low points.
Actually, there are two types of extreme points on a graph: local ex-
treme points and absolute extreme points. The term ‘extreme point’
Section 12: The Mean and the Extreme

refers to high or low points on a graph. Let’s have the formal defini-
tions.

Definition 12.1. Let f be a function defined on an interval I, and

let c ∈ I.
(1) We say that f has a local or relative minimum at c ∈ I pro-
vided there is a open interval J containing c such that for all
x ∈ J ∩ I, f (x) ≥ f (c).
(2) We say that f has a local or relative maximum at c ∈ I pro-
vided there is a open interval J containing c such that for all
x ∈ J ∩ I, f (x) ≤ f (c).

Definition Notes: The interval J in these definitions is used to de-

scribe a meaning of “closeness” to the number c. Basically, (1) in
Definition 12.1 states that c is a local minimum provided there is
a closeness, J, to c, such that if x is that close (x ∈ J) and if x is
within the domain of definition of f (x ∈ I), then f (x) ≤ f (c). That
is, everywhere “close” to c, f (c) is the smallest value of the function.
Section 12: The Mean and the Extreme

The condition x ∈ J ∩ I simply requires x to be “close” to c

(x ∈ J) and x be in the domain of f (x ∈ I).
Exercise 12.1. Why do you suppose in Definition 12.1 we required
J to be an open interval?
Of considerable more applied interest is the notion of absolute extrema,
the definitions of which follow.

Definition 12.2. Let f be a function having domain Dom(f ), and

let c ∈ Dom(f ).
(1) We say that c ∈ Dom(f ) is an absolute minimum of f over
Dom(f ) if for all x ∈ Dom(f ), f (x) ≥ f (c).
(2) We say that c ∈ Dom(f ) is an absolute maximum of f over
Dom(f ) if for all x ∈ Dom(f ), f (x) ≤ f (c).

Definition Notes: In the definition, Dom(f ) is not necessarily the nat-

ural domain of the function; it may be a subset of the natural domain
of the function f . For example, the function might be f (x) = x2 − x,
whose natural domain of definition is R, but may be interested in the
Section 12: The Mean and the Extreme

absolute maximum (or minimum) of this function over the interval

Dom(f ) = [ 0, 1 ].
It is important to remember that the notion of absolute maxi-
mum or minimum is itself relative concept — relative to the current
domain of definition of f . As a trivial and simple example of this re-
mark, consider the function f (x) = x and Dom(f ) = [ 0, 1 ]. In this
case, the absolute minimum is at c = 0 and the actual minimum value
is f (0) = 0. If we take the same function and change the domain to
Dom(f ) = [ 1, 2 ], then the absolute minimum occurs at c = 1 and the
minimum value is f (1) = 1.

Theorem 12.3. (Fermat’s Theorem) Let f be a function defined on

an interval I. Suppose c ∈ I is a local extremum such that c is not an
endpoint of I, and f 0 (c) exits, then f 0 (c) = 0.

Proof.
Theorem Notes: The proof is easy to understand and it is strongly
recommended that the student study it.
Section 12: The Mean and the Extreme

Now we turn to the larger question of locating local extrema.

Let c ∈ I be any local extrema of f , then either c is an endpoint,
or not; further, in the case c is not an endpoint, either f 0 (c) exists,
or f 0 (c) does not exist. In the former case, we know from Fermat’s
Theorem, f 0 (c) = 0.
To summarize, then, if c ∈ I is any local extrema of f , then
either c is an endpoint, f 0 (c) = 0, or f 0 (c) does not exist. Should you
be looking for local extrema, these would be the three places to look.

Before we continue, let’s define a useful terminology.

Definition 12.4. Let f be a function and let c ∈ Dom(f ), then c is

called a critical point of f if either f 0 (c) does not exist, or f 0 (c) = 0.
Definition Notes: A given critical point can be classified into any one
of three categories: local maximum, local minimum, or neither. In the
latter case, that is, if c is neither a local maximum nor minimum, very
often, with the kind of functions we deal with, this means that c is a
saddle point. Examples are listed below.
Section 12: The Mean and the Extreme

Here is a set of questions that occurs to me. Given that we have a

critical point c in hand:
1. How can we tell, in a sightless way whether the critical point is
a maximum, minimum, or neither.
2. How can we tell, in a sightless way whether the critical point is
an absolute extrema or a local extrema?
Notice that an emphasis is placed on sightless techniques. In the age
of the modern graphing calculator these questions can be answered by
simply graphing the function. These kinds of techniques are certainly
important at our level of play, but I am anticipating a more general
setting. In many applied problems it is desired to find maximums
and minimums of functions of many variables. In this case, there is
no geometry — one cannot graph the function and visually see the
maximums or minimums. We need to start developing techniques,
therefore, that will generalized to multidimensions where there are
not visualizations.
Section 12: The Mean and the Extreme

12.2. The Extreme Value Theorem

In this section we introduce the title theorem and give series of pro-
cedural steps for locating absolute extrema of a function over closed
and bounded intervals.
We want to hunt for the absolute extrema of a function over a specific
interval. It would be nice to know that the absolute extrema exist
before we start to look for them — this would save a lot of time,
pain, and mental anguish. The next theorem describes under what
conditions we will know that absolute maximums and minimums exist.

Theorem 12.5. (The Extreme Value Theorem) Let f be a continu-

ous function defined over a closed interval I = [ a, b ]. Then
(1) there exists (at least one) point xmin ∈ I such that f has an
absolute minimum at xmin ;
(2) there exists (at least one) point xmax ∈ I such that f has an
absolute maximum at xmax .
Section 12: The Mean and the Extreme

Proof. Beyond the scope of these notes.

Theorem Notes: If xmin is a point at which f has an absolute mini-
mum, then the minimum value is f (xmin ). In this case, sometimes we
write,
min f (x) = f (xmin );
a≤x≤b

the meaning of the notation is self-evident.

Similarly, if xmax is a point at which f has an absolute maxi-
mum, then the maximum value is f (xmax ). In this case, we write,
max f (x) = f (xmax );
a≤x≤b

As you know from your own graphing experiences, an absolute

maximum (minimum) may occur at several points (or at infinitely
many points). Think of the function f (x) = cos(x), I = [ 0, 4π ]. It
should be clear to you that
max cos(x) = 1,
0≤x≤4π
Section 12: The Mean and the Extreme

and that the maximum value is attained at x = 0, 2π, and 4π. In the
same way,
min cos(x) = −1,
0≤x≤4π

and that the minimum value is attained at x = π and 3π.

The Extreme Value Theorem and Fermat’s Theorem together
enable us to setup a definite procedure for finding the absolute ex-
trema of a continuous function f defined over a closed and bounded
interval I = [ a, b ].
Problem. Given a continuous function f defined on a closed and
bounded interval [ a, b ], we want to find the absolute maximum and
absolute minimum of the function f over the interval [ a, b ].
The Method. We proceed as follows:
1. Find the critical points of f over [ a, b ]: these are the numbers,
x, at which f 0 (x) = 0 or f 0 (x) does not exist. Let me represent
these numbers symbolically:
x1 , x2 , x3 , . . . , xn .
Section 12: The Mean and the Extreme

2. Include the endpoints a and b is the above list:

a, x1 , x2 , x3 , . . . , xn , b. (1)
3. Calculate the value of f at each of the numbers in the list (1).
x f (x)
a f (a)
x1 f (x1 )
x2 f (x2 )
x3 f (x3 )
.. ..
. .
xn f (xn )
b f (b)
4. The absolute maximum is the largest number in the right col-
umn and the absolute minimum is the smallest number in the
right column of the above table.
Justification of this Method. Here is a detailed explanation of the
reasoning behind this method.
Section 12: The Mean and the Extreme

Example 12.1. Consider the function f (x) = 3 + 4x − 3x3 restricted

to the interval [ −1, 2 ]. Find the absolute maximum and absolute min-
imum of f over the interval f and state at what values of x these
extrema are attained.

12.3. The Mean Value Theorem

In the previous sections we introduced definitions of local and absolute
extrema as well as some supporting theory that enable us to develop
a method of locating absolute extrema. In this section, we continue to
develop some underlying theory that will enable us to develop analytic
and sightless methods of classifying critical points.
We begin by presenting a theorem that is later used to prove the
Mean Value Theory, but is itself of considerable interest.

Theorem 12.6. (Rolle’s Theorem) Let f be a function be continuous

on the interval [ a, b ] and differentiable on the interval ( a, b ) such that
f (a) = f (b)
Section 12: The Mean and the Extreme

Then there exists a number c ∈ ( a, b ) such that f 0 (c) = 0.

Proof.
Let f be a differentiable function. Then
f (x + h) − f (x)
f 0 (x) := lim . (2)
h→0 h
This equation establishes a relationship between the values of the
function f near x and the value of the derivative f 0 at that same
x. The relationship is a complex one though: The two are connected
by the limit process. In this section we explore the Mean Value
Theorem, a theorem which establishes another relationship between
f and f 0 .
The Mean Value Theorem is an “obvious” truth, but like many
truths in mathematics, it is more difficult to prove than you would
suppose. Let me illustrate this obvious fact through an example.
Example 12.2. This example illustrates the Mean Value Theo-
rem using a skier and a ski slope.
Section 12: The Mean and the Extreme

Now let me formally state the Mean Value Theorem. See if you
can recognize the skier analogy in the theorem.

Theorem 12.7. (The Mean Value Theorem) Let f be a function be

continuous on the interval [ a, b ] and differentiable on the open interval
( a, b ). Then there exists a number c ∈ ( a, b ) such that

f (b) − f (a)
f 0 (c) = (3)
b−a
or,
f (b) − f (a) = f 0 (c)(b − a) (4)
or even,
f (b) = f (a) + f 0 (c)(b − a) (5)

Proof.
Theorem Notes: It is equation (3) that yields the skiing interpretation
of Example 12.2, do you see it?
The point T ( a, f (a) ) is the left-most point on the graph of f ,
the point B( b, f (b) ) is the right-most point on the graph of f . What
Section 12: The Mean and the Extreme

is the interpretation of right-hand side of (3)? It is the slope of the line

that passes through T and B — the gradient of the curve. According
to the Mean Value Theorem, there is some c ∈ ( a, b ) such that
(3) is true. Recall that f 0 (c) is the slope of the line tangent to the
graph at the point x = c. Thus, (3) states that there is some point on
the graph (x = c) at which the tangent is parallel to the gradient of
the graph. This is the content of Example 12.2.
The equation (4) is an algebraic variation of (3). It basically
gives a direct relationship between the values of the function f , and
the values of the derivative of the function f 0 . In subsequent studies,
we will take advantage of (4).
When the Mean Value Theorem is used in the form of (5)
it usually means that we are interested in relating the value f (b) to
f (a).
The phrasing of Theorem 12.7 is a little strange, “Suppose f
is . . . continuous on [ a, b ] and differentiable on ( a, b ).” why do you
suppose that is?
Section 12: The Mean and the Extreme

The next theorem is a simple, yet important, application of the Mean

Value Theorem.

Theorem 12.8. Suppose f is a function that is continuous on the

interval [ a, b ] and differentiable on the open interval ( a, b ) such that
f 0 (x) = 0 for all x ∈ ( a, b ),
then f is constant over the interval [ a, b ]; i.e., there is some constant,
C, such that
f (x) = C for all x ∈ [ a, b ].

Proof. The proof is simple enough to be presented “in-line.”

Suppose f is a function such that f 0 (x) = 0 for all x ∈ ( a, b ). Define
C := f (a). Claim f (x) = C for all x ∈ [ a, b ].
To that stated end, choose any x ∈ ( a, b ] and apply the Mean Value
Theorem to our function f but over the interval [ a, x ]. By the Mean
Value Theorem, there is some c ∈ ( a, x ) such that
f (x) − f (a) = f 0 (c)(x − a). (6)
Section 12: The Mean and the Extreme

But we are assuming that f 0 is identically zero; therefore, f 0 (c) = 0.

Updating (6) using this information we get

f (x) − f (a) = (0)(x − a) = 0

or,
f (x) = f (a)

But, f (a) = C, so

f (x) = C
which is what we wanted to prove. Hurray! 

Corollary 12.9. Suppose f and g are continuous over the interval

[ a, b ] and differentiable over the open interval ( a, b ) such that
f 0 (x) = g 0 (x) for all x ∈ ( a, b ),
then there is a constant, C, such that
f (x) = g(x) + C for all x ∈ ( a, b, ).
Section 12: The Mean and the Extreme

Proof. Define a new function by

H(x) = f (x) − g(x) x ∈ [ a, b ].

Then,
H 0 (x) = f 0 (x) − g 0 (x) = 0 x ∈ [ a, b ].
Now, by Theorem 12.8, there is a constant, C, such that H(x) = C.
This means

C = H(x) = f (x) − g(x) x ∈ [ a, b ]

or,
f (x) = g(x) + C x ∈ [ a, b ]
which is the advertised equation. 
Theorem Notes: Corollary 12.9 states that the only way two functions
can have exactly the same derivative is for the graph of one of the
functions to be a vertical translation of the graph of the other.
This corollary is a foundation stone of Indefinite Integration.
Watch of reference to this corollary when you study integration.
Section 12: The Mean and the Extreme

12.4. Criteria for Monotonicity

Let’s begin with a phalanx of definitions.

Definition 12.10. Let f be defined on an interval I.

(1) The function f is strictly increasing over the interval I pro-
vided
x1 , x2 ∈ I and x1 < x2 =⇒ f (x1 ) < f (x2 ).
(2) The function f is strictly decreasing over the interval I pro-
vided
x1 , x2 ∈ I and x1 < x2 =⇒ f (x1 ) > f (x2 ).

Definition Notes: A monotone function is one that is strictly increas-

ing or decreasing. Thus, if I say that “f is monotone over the interval
( a, b ), I mean that f is either strictly increasing or strictly decreasing
over that interval.
Section 12: The Mean and the Extreme

A function may be strictly increasing (decreasing) over its entire

domain, or a function may decrease over portions of the domain and
increase over other portions of the domain.
Example 12.3. The function f (x) = x is strictly increasing over its
entire natural domain, R.
Example 12.4. The function f (x) = x2 is strictly increasing over the
interval ( 0, +∞ ) and is strictly decreasing over the interval ( −∞, 0 ).
As was seen by these two illustrative examples, the increasing or de-
creasing nature of a function can be proven using algebra—but alge-
braic methods quickly become too difficult to use even for “moderately
complicated” functions.
Typically, there are two approaches of verifying that a function in-
creases or decreases over a specific interval:
1. Visual or graphical Methods. (If we graph the function we can
“see” where the function increases or decreases.)
2. Analytical Methods.
Section 12: The Mean and the Extreme

a. Algebraic Methods (as illustrated in the last two exam-

ples).
b. Calculus Methods.
In this section, of course, we are primarily interested in analytic meth-
ods using Calculus!
The next theorem states the relevant theory that leads to our analytic
methods.

Theorem 12.11. (The Monotonicity Theorem) Let f be continuous

on [ a, b ] and differentiable over the open interval ( a, b ).
(1) Test for Increasing. If f 0 (x) > 0 over the interval ( a, b ), then
f is strictly increasing over the interval [ a, b ].
(2) Test for Decreasing. If f 0 (x) < 0 over the interval ( a, b ), then
f is strictly decreasing over the interval [ a, b ].
Section 12: The Mean and the Extreme

Proof. The proof is a simple consequence of the much celebrated

Mean Value Theorem and can be easily understood by the reader.
You are invited to read it.
Theorem Notes: Theorem 12.11 essentially gives a way of detecting
the monotone nature of a differentiable function through the exam-
ination of its first derivative. We simply attempt to find everywhere
the first derivative of the function is positive—that will generally over
an interval of numbers—and by (a) above, the parent function f will
be increasing over that same interval.
Let’s begin with some simple examples . . . then it’s your turn.
Example 12.5. Find all intervals of increase and decrease of the
function f (x) = x4 − 8x2 + 2.
This example illustrates the way in which Theorem 12.11 is applied.
To find the intervals of increase and decrease of a function f , calculate
f 0 and do a sign analysis on the function f 0 — typically this is done
using the Sign Chart Method.
Section 12: The Mean and the Extreme

Example 12.6. Find the intervals of increase and decrease of the

function f (x) = x3 (x − 2)4 .

12.5. Classifying Critical Points: The First Derivative Test

The first derivative, in addition to containing monotone information
of a function, also contains information leading us to classify critical
points. In fact it is the monotone information that makes the classifi-
cation.

Theorem 12.12. (The First Derivative Test) Let f be a function

and c a critical point of f .
(1) Test for Local Maximum. Suppose f 0 is positive to the left of
c and negative to the right of c, then c is a local maximum.
(2) Test for Local Minimum. Suppose f 0 is negative to the left of
c and positive to the right of c, then c is a local minimum.
(3) Test for a Saddle Point. Suppose f 0 does not change signs at
c, then c is a saddle point.
Proof.
Let’s analyze the function in Example 12.5 and classify all critical
points.
Example 12.7. Find and classify all critical points of the function
f (x) = x4 − 8x2 + 2.
The function we examined in Example 12.6 is interesting as it pro-
vides an example of a saddle point. The next example examines this
function once again.
Example 12.8. Find and classify all critical points of the function
f (x) = x3 (x − 2)4 .
 Solutions to Exercises
8.1. Step 1: recognize and realize that the given function is a compo-
sition of two other functions. Step 2: decompose. Step 3: apply chain
rule.
The function F (x) = (3x4 +5x)1/2 is the composition of two functions:
the outer function: f (x) = x1/2 ; and the inner function: g(x) = 3x4 +
5x.
Preliminary Calculations:

f (x) = x1/2 g(x) = 3x4 + 5x

1
f 0 (x) = x−1/2 g 0 (x) = 12x3 + 5.
2
Solutions to Exercises (continued)

Therefore, by the Chain Rule we obtain,

F 0 (x) = (f ◦ g)0 (x) = f 0 (g(x))g 0 (x)
= f 0 (3x4 + 5x)(12x3 + 5)
1
= (3x4 + 5x)−1/2 (12x3 + 5).
2
Thus,
1
F 0 (x) = (12x3 + 5)(3x4 + 5x)−1/2 .
2

Did you get it?

Exercise 8.1.
Solutions to Exercises (continued)

8.2. The function F is obviously the composition of two functions:

y = f (x) g(x) = 1/x4 = x−4 .
Make the standard calculations.

f (x) g(x) = x−4

f 0 (x) g 0 (x) = −4x−5 .
Therefore, by the Chain Rule we obtain,
F 0 (x) = (f ◦ g)0 (x) = f 0 (g(x))g 0 (x)
= f 0 (1/x4 )(−4x−5 )
f 0 (1/x3 )
= −4
x5
Thus,
f 0 (1/x3 )
F 0 (x) = −4 .
x5
Exercise 8.2.
Solutions to Exercises (continued)

8.3. The function F is obviously the composition of two functions:

one abstract and the other specific. Define f (x) = x6 , then
F (x) = [g(x)]6 = f (g(x)).
Of course, f 0 (x) = 6x5 and so, by the Chain Rule, we have
F 0 (x) = f 0 (g(x))g 0 (x) = 6[g(x)]5 g 0 (x)
Thus,
F 0 (x) = 6[g(x)]5 g 0 (x).

Exercise Notes: This formula gives us a way of differentiating any

base function g raised to the sixth power. A useful formula indeed, if
you have the sixth power!
The formula is only valid when we have a base function raised
to the sixth power. What if we wanted to differentiate a function of
the form F (x) = [g(x)]7 ?
In the section The Power Rule Revisited, a general formula is
obtained for differentiating any function of the form F (x) = [g(x)]r ,
where r is any exponent.
Solutions to Exercises (continued)

Maybe you can develop the formula yourself before getting this
section. (Hint: Use this exercise as a template; change the exponent of
6 to an abstract exponent of r and follow the development line-by-line,
making all appropriate changes.)
Exercise 8.3.
Solutions to Exercises (continued)

8.4. The function is y = (3x3 − 6x)1/2 . Decompose using a substi-

tution technique. Let u = 3x3 − 6x, then y = u1/2 . Thus

y = u1/2 u = 3x3 − 6x
dy 1 du
= u−1/2 = 9x2 − 6.
du 2 dx
By the Chain Rule,
dy dy du
=
dx du dx
1
= u−1/2 (9x2 − 6). (A-1)
2
Again, typical of this technique, the answer comes out in terms of
x and u. A re-substitution will solve this problem: u = 3x3 − 6x.
Substituting this into (A-1) we obtain
dy 1
= (3x3 − 6x)−1/2 (9x2 − 6)
dx 2
Solutions to Exercises (continued)

The answer could be cleaned up a little.

dy 3(3x2 − 2)
= √
dx 2 3x3 − 6x
Exercise 8.4.
Solutions to Exercises (continued)

8.5. We want to differentiate a function raised to a power.

d
(4x2 + 1)23 = 23(4x2 + 1)22 (8x) = 184x(4x2 + 1)22 .
dx
Exercise 8.5.
Solutions to Exercises (continued)

8.6. We want to differentiate a function raised to a power.

 5  4
d x x d x
=5 Power
dx 1 + x 1+x dx 1 + x
 4
x (1 + x)(1) − x(1)
=5 Quot.
1+x (1 + x)2
 4
x 1
=5
1+x (1 + x)2
5x4
=
(1 + x)6
Thus,
 5
d x 5x4
= .
dx 1+x (1 + x)6
Exercise 8.6.
Solutions to Exercises (continued)

8.7. This is the derivative of a product of two functions. Use the

Product Rule.
d 4 d
x (2x + 1)3/2 = x4 (2x + 1)3/2 + 4x3 (2x + 1)3/2
dx dx
= x (3/2)(2x + 1)1/2 (2) + 4x3 (2x + 1)3/2
4

= x3 (2x + 1)1/2 (3x + 4(2x + 1))

= x3 (2x + 1)1/2 (11x + 4)
The ultimate in answers is
d 4
x (2x + 1)3/2 = x3 (11x + 4)(2x + 1)1/2 .
dx
Notes: If you have been seriously reading and working along with
these notes, you will, hopefully, begin to see a style of solution to
these differentiation problems. All of them are pretty much the same.

The secret? Use good notation, classify the function types (product,
quotient, sum, etc.), know the rules (the formulas), use good algebra.
Solutions to Exercises (continued)

That’s all. Exercise 8.7.

Solutions to Exercises (continued)

8.8. We are asked to differentiate a quotient of two functions of s.

You should have used the Quotient Rule.
Using the Quotient Rule and the Power Rule all in one magnificent
step we get,
d (s + 1)3 (2s + 1)5 (3)(s + 1)2 − (s + 1)3 (5)(2s + 1)4 (2)
= (A-2)
ds (2s + 1)5 (2s + 1)10
This is the initial macro expansion of the quotient and product rules.
This answer represents a minimal response to the question. However,
you should have the drive, the power, the ability to improve on this
answer. I hope this answer is unacceptable to you.
Assignment: Invoke your algebra module of knowledge and simplify.
Below is a multiple choice quiz, after you have finished simplifying,
make your choice. Do not look at the choices before your have finished
your simplifications. Simplify Now !
After you simplify, go to the next page for a little quiz!
Solutions to Exercises (continued)

Which of the following is a simplification of (A-2)?

(s + 1)2 3(s + 1)3 (4s + 7)
(a) (b)
(2s + 1)10 (2s + 1)10
(4s + 7)(s + 1)2 (4s + 7)(s + 1)2
(c) − (d) −
(2s + 1)6 (2s + 1)10
Exercise 8.8.
Solutions to Exercises (continued)

9.1. Let’s have a quiz.

d
Which of the following is the 3 cos(x)?
dx
(a) 3 cos(x) (b) −3 sin(x) (c) 3 sin(x) (d) −3 cos(x)
d
Which of the following is the − 5 sec(t)?
dt
(a) −5 sec2 (x) (b) −5 csc(x) cot(x)
(c) 5 cot2 (x) (d) −5 sec(x) tan(x)
Hopefully, you were two for two. Exercise 9.1.
Solutions to Exercises (continued)

9.2. Proceed as follows,

d
tan(x) = sec2 (x) x=π/3
from (3)
dx x=π/3

= sec2 (π/3)
=4
Recall
1 1 1
sec2 (π/3) = = = = 4.
cos2 (π/3)
(1/2)2 1/4
Now let’s turn to the problem of calculating the equation of the line
tangent to the graph at x = π/3.
In order to find any equation of a line we need two pieces of informa-
tion: A point the line passes through; the slope of the target line.
Work through the rest of the problem through a series of steps.
The Point: Which of the following can be taken to be our point?
p √
(a) ( π/3, 3/2 )(b) ( π/3, 2 ) (c) ( π/3, 4 ) (d) ( π/3, 3 )
Solutions to Exercises (continued)

The Slope: Which of the following is the slope of the line tangent to
the graph of y = tan(x).
(a) 1/2 (b) 2 (c) sec2 (x) (d) 4
The Equation: Which of the following is the equation of the line tan-
gent to the graph at x = π/3?
√
3 √ √
(a) y − = 3(x − π3 ) (b) y − 3 = 4(x − π3 )
√2 √
(c) y − 3 = sec2 (x)(x − π3 ) (d) y − π3 = 4(x − 3)
Exercise 9.2.
Solutions to Exercises (continued)

9.3. We first recall that

sin(x)
tan(x) = .
cos(x)
Therefore, we use the Quotient Rule.
d d sin(x)
tan(x) =
dx dx cos(x)
d d
cos(x) sin(x) − sin(x) cos(x)
= dx dx
cos2 (x)
cos(x) cos(x) − sin(x)(− sin(x))
=
cos2 (x)
cos2 (x) + sin2 (x)
=
cos2 (x)
1
=
cos2 (x)
= sec2 (x)
Solutions to Exercises (continued)

As you can see this is a straight forward enough calculation. So it

goes with all the trig functions tan(x), cot(x), sec(x), and csc(x). The
difficult task was to calculate the derivatives of sin(x) and cos(x), the
other four trig functions are ratios of these two. Exercise 9.3.
Solutions to Exercises (continued)

9.4. The function is F (x) = sin(6x3 ). You should have realized that
F is the composition of two more elementary functions: the outer
function is f (x) = sin(x) and the inner function is g(x) = 6x3 .
Preliminary Calculations:

f (x) = sin(x) g(x) = 6x3

f 0 (x) = cos(x) g 0 (x) = 18x2 .
Therefore, by the Chain Rule we obtain,
F 0 (x) = (f ◦ g)0 (x) = f 0 (g(x))g 0 (x)
= f 0 (6x3 )(18x2 )
= cos(6x3 )(18x2 ).
Thus,
F 0 (x) = 18x2 cos(6x3 ).
Exercise 9.4.
Solutions to Exercises (continued)

d
9.5. The problem tan(sin(x)) is nothing more than the derivative
dx
of the tangent function of some function of x. Use Trig formula (3),
d d
tan(sin(x)) = sec2 (sin(x)) sin(x)
dx dx
2
= sec (sin(x)) cos(x)
You simply have to learn to look at a function and classify it. In this
case, we wanted to differentiate the tangent of some function of x.
Keep it simple. Do the obvious.
The derivative of the tangent of some function of x is the secant
squared of that same function of x, times the derivative of that func-
tion of x. This would be the verbalization of the tangent formula.
Exercise 9.5.
Solutions to Exercises (continued)

d
9.6. Problem: sin(tan4 (3x3 )). Initially, we want to differentiate
dx
the sine of some function of x — that’s all. Apply (1),
d d
sin(tan4 (3x3 )) = cos(tan4 (3x3 )) tan4 (3x3 )
dx dx
Now to continue, we must analyze our next differentiation problem.
Are we to differentiate the tangent of some function of x, or are we
to differentiate a function raised to a power? The notation tan4 (3x3 ),
is short-hand for (tan(3x3 ))4 . This is a composition of two functions
with the 4th -power function the “outer” function; consequently, we
see the expression tan4 (3x3 ) as a function raised to a power — hence,
apply the Power Rule. Continuing the calculation now,
d d
sin(tan4 (3x3 )) = cos(tan4 (3x3 )) tan4 (3x3 )
dx dx
d
= cos(tan4 (3x3 ))(4) tan3 (3x3 ) tan(3x3 )
dx
Solutions to Exercises (continued)

We have another differentiation problem. This one is the derivative of

the tangent of some function. Use (3) of the trig function set.
d d
cos(tan4 (3x3 )) = cos(tan4 (3x3 ))(4) tan3 (3x3 ) tan(3x3 )
dx dx
d
= cos(tan4 (3x3 ))(4) tan3 (3x3 ) sec2 (3x3 ) (3x3 )
dx
= cos(tan (3x ))(4) tan (3x ) sec (3x )(9x2 )
4 3 3 3 2 3

The final answer made to look cute is

d
sec(tan4 (3x3 )) = 36x2 sec2 (3x3 ) tan(3x3 ) cos(tan4 (3x3 )).
dx
That’s quite a mouthful!
Exercise Notes: Actually, these differentiation problems are not that
hard, IF you follow the rules exactly as they are laid out. Analyze
the function type, apply the rules, slowly, methodically, one step at a
time just as I have done. You’ll come through o.k.
Solutions to Exercises (continued)

Perhaps now you can see from where the term “Chain Rule”
comes. When you differentiate a function that is the composition of
a large number of other functions, as was the case in this exercise,
we differentiate the “outer” function times the derivative of the “in-
ner” function. As a result, we generate a series of derivatives that are
chained together.
Sometimes I say that the process of differentiation is the process
of moving the d/dx symbol from left to right. Can you see why?
Exercise 9.6.
Solutions to Exercises (continued)

10.1. The first thing to know, is how to read the formula

nπ
y (n) = 2n sin(2x + ) n = 1, 2, 3, 4, . . . . (A-3)
2
Write (A-3) for the cases of n = 1, 2, 3, 4, dotsc.
π
n=1 y 0 = 2 sin(2x + ) (A-4)
2
n=2 y 00 = 4 sin(2x + π)
3π
n=3 y 000 = 8 sin(2x + )
2
n=4 y (4) = 16 sin(2x + 2π)
Compare these first derivatives with the results of (1). Except for the
powers of 2, they don’t compare well do they. However, if you apply
the additive rule for the sine function
sin(A + B) = sin(A) cos(B) + cos(A) sin(A)
to the equations in (A-4), I hope you’ll change your mind.
Exercise 10.1.
Solutions to Exercises (continued)

x
10.2. The function y is simply the quotient of two polynomi-
2x + 1
als; it should be the height of triviality to differentiate.
The first derivative:
d x
y0 =
dx 2x + 1
(2x + 1)(1) − x(2)
= Quot. Rule
(2x + 1)2
1
=
(2x + 1)2
= (2x + 1)−2
The second derivative:
d
y 00 = (2x + 1)−2
dx
= (−2)(2x + 1)−3 (2) Power Rule
= −4(2x + 1)−3
Solutions to Exercises (continued)

The third derivative:

d
y 000 = −4 (2x + 1)−3
dx
= (−4)(−3)(2x + 1)−4 (2) Power Rule
−4
= 24(2x + 1)
The ambitious student might calculate another half-dozen derivatives
for practice. Exercise 10.2.
Solutions to Exercises (continued)

11.1. We follow standard procedures.

Treat y as it were a function of x, and differentiate both sides with
respect to x:
d x2 y2 d1
( + )= .
dx 4 9 dx
Apply the differentiation rules, treating y as a function of x:
1 1 dy
(2x) + (2y) = 0.
4 9 dx
Solve the equation for dy/dx:
2 dy 1
y =− x
9 dx 2
dy 9x
=− .
dx 4y
Thus,
dy 9x
=− .
dx 4y
Solutions to Exercises (continued)

You’ll notice that dy/dx becomes infinite when y = 0. This usually

indicates vertical tangent lines. The points on the ellipse correspond-
ing to y = 0 are ( 2, 0 ) and ( −2, 0 ). These two points are the vertices
on the x-axis and we see that at these two points, we indeed have
vertical tangents.
Maple Plot. A plot of this equation can be obtained from Maple using
the commands:
> with(plots);
> implicitplot( x^2/4 + y^2/9 = 1, x=-2..2, y=-3..3);
Exercise 11.1.
Solutions to Exercises (continued)

11.2. The given equation x2 y 7 − x3 y 2 = 1 is an equation involving

the variables x and y. We cannot explicitly solve for y in terms of x;
therefore, we use the method of implicit differentiation.
Differentiate both sides with respect to x:
d 2 7 d1
(x y − x3 y 2 ) = .
dx dx
Apply the differentiation rules wherever applicable, Step 2, all the
while treating y as if it were a function of x, Step 1:
d 2 7 d 3 2
(x y ) − (x y ) = 0
dx dx
d y7 d x2 d y2 d x3
(x2 + y7 ) − (x3 + y2 )=0
dx dx dx dx
dy dy
(x2 (7y 6 ) + y 7 (2x)) − (x3 (2y ) + y 2 (3x2 ) = 0
dx dx
dy dy
7x2 y 6 + 2xy 7 − 2x3 y − 3x2 y 2 = 0.
dx dx
Solutions to Exercises (continued)

dy
Now, solve algebraically for , Step 3:
dx
dy dy
7x2 y 6 + 2xy 7 − 2x3 y − 3x2 y 2 = 0
dx dx
dy
(7x2 y 6 − 2x3 y) = 3x2 y 2 − 2xy 7
dx
dy
x2 y(7y 5 − 2x) = xy 2 (3x − 2y 5 )
dx
Thus,

dy x2 y(7y 5 − 2x)
= 2 ,
dx xy (3x − 2y 5 )
or,

dy x(7y 5 − 2x)
= .
dx y(3x − 2y 5 )
This then is the derivative of y with respect to x.
Solutions to Exercises (continued)

Question: Can you find a point, ( x, y ) that satisfies the equation

x2 y 7 − x3 y 2 = 1? What the graph look like? This is a job for Maple!
These questions are answered in the example contained in our discus-
sion of Differentiation using Maple. Exercise 11.2.
Solutions to Exercises (continued)

11.3. In the spirit of the outline, you should have said to take equa-
tion (13) and differentiate it with respect to x, the result of which
would be
y (4) = F6 (x, y, y 0 ).
Now, take (10), and substitute for y 0 in the above equation to obtain
the desired derivative in its most fundamental form:
y (5) = F7 (x, y).
Exercise 11.3.
Solutions to Exercises (continued)

11.4. You should have proceeded according to the outline above.

Indeed,
First Derivative:
d dx
1. sin(y) =
dx dx
dy
2. cos(y) =1
dx
dy
3. y 0 = = sec(y)
dx

Second Derivative:
Solutions to Exercises (continued)

dy 0 d
1. y 00 = = sec(y)
dx dx
dy
2. y 00 = sec(y) tan(y)
dx
00
3. y = sec(y) tan(y) sec(y)

4. y 00 = sec2 (y) tan(y)

The third derivative is done similarly. I calculated it to be

y 000 = sec2 (y)(3 sec2 (y) − 2),
expressed entirely in terms of the secant function. Verify please!
Exercise 11.4.
Solutions to Exercises (continued)

12.1. Assume the same notation as in Definition 12.1 and we assume

for the sake of discussion that c is a local minimum.
If J is required to be an open interval and c ∈ J, then c is not one of
the endpoints, since the endpoints do not belong to the interval.
Why is this important in the description of the concepts of local max-
imum or minimum? To properly answer that question, let’s consider
two cases: c is an endpoint of the interval I and c is not an endpoint
of I.
c is not an endpoint of I: In this case, when we compute J ∩ I we
will get an interval whose left-hand endpoint is strictly less than c
and whose right-hand endpoint is strictly greater than c. (Do you see
why?) That is, J ∩ I will include numbers on both sides of c. This is
the important point. The condition
f (x) ≥ f (c) x ∈ J ∩ I,
Solutions to Exercises (continued)

states that the graph of f is higher (than f (c)) to the left of c and
higher to the right of c; this is descriptive of the concept of local
minimum.
If J was not required to be open, then J could contain one of its
endpoints. In this case the requirement that c ∈ J could mean that
c is just an endpoint of J; therefore, c would be an endpoint of the
interval J ∩ I. The condition
f (x) ≥ f (c) x ∈ J ∩ I,
may be only saying that f is higher on one side of a. Think about it.
This would be counter to the concept of a low point on the graph. In
fact, if we don’t require J to be an open interval, then most any point
on the graph is both a maximum and minimum! Can you see why?
(Think of the function f (x) = x2 and take c = 1, try to argue that
c = 1 is both a local maximum and a local minimum by dropping the
word “open” from Definition 12.1.)
c is an endpoint of I: Assume for the purpose of argument that c is
the left-hand endpoint of the interval I. In this case, J ∩ I will be an
interval whose left-hand endpoint is c as well. The condition
f (x) ≥ f (c) x ∈ J ∩ I,
says that f is higher to the right of c and says nothing of what goes
on to the left of c. But if c is the left-hand endpoint of the domain of
definition I, we don’t care what goes on to the left of c because that
is outside the domain of the function f .
In this case, it really doesn’t matter whether J is open or not.
Exercise 12.1.
 Solutions to Examples
8.1. This is a continuation of the Theorem Notes. It was seen there
that F = f ◦ g, where f (x) = x10 and g(x) = x2 + 1. Note that we
have not relabeled the variables to suggest composition — sorry! The
Chain Rule formula states that at any x
F 0 (x) = (f ◦ g)0 (x) = f 0 (g(x))g 0 (x).
Let’s make the necessary calculations:

f (x) = x10 g(x) = x2 + 1

f 0 (x) = 10x9 g 0 (x) = 2x
therefore,
f 0 (g(x)) = f 0 (x2 + 1) = 10(x2 + 1)9
Finally,
F 0 (x) = (f ◦ g)0 (x) = f 0 (g(x))g 0 (x)
= 10(x2 + 1)9 2x = 20x(x2 + 1)9 .
Solutions to Examples (continued)

Thus,
d(x2 + 1)10
= 10x(x2 + 1)9 .
dx
Example Notes: That was relatively painless — I had no problems
at all with it. The key point is the ability of the student to decom-
pose functions (uncompose?), then differentiate each separately, and
multiply them together.
Example 8.1.
Solutions to Examples (continued)

8.2. Step 1: recognize and realize that the given function is a compo-
sition of two other functions. Step 2: decompose. Step 3: apply chain
rule.
The function F (x) = (3x3 − x)7 is the composition of two functions:
the outer function: f (x) = x7 ; and the inner function: g(x) = 3x3 −x.
Preliminary Calculations:

f (x) = x7 g(x) = 3x3 − x

f 0 (x) = 7x6 g 0 (x) = 9x2 − 1.
Therefore,
F 0 (x) = (f ◦ g)0 (x) = f 0 (g(x))g 0 (x)
= f 0 (3x3 − x)(9x2 − 1)
= 7(3x3 − x)6 (9x2 − 1).
Thus,
F 0 (x) = 7(9x2 − 1)(3x3 − x)6 .
Solutions to Examples (continued)

But wait, we wanted F 0 (2) — mere calculation.

F 0 (2) = 7(35)(22)6 .
I leave the punching of calculator buttons to you, but the derivative
is large. Example 8.2.
Solutions to Examples (continued)

8.3. In some sense, this is an easier problem than working with a

specific function. The function F is obviously the composition of two
functions:
y = f (x) g(x) = x3 .
It should be clear to you that F = f ◦ g.
Now, proceeding along standard lines of inquiry we make some pre-
liminary calculations:

f (x) g(x) = x3
f 0 (x) g 0 (x) = 3x2 .
Notice that because the function f is an abstract differentiable func-
tion, it is the height triviality to differentiate it: the derivative of f is
f 0 . Isn’t that easy?
Solutions to Examples (continued)

Therefore, by the Chain Rule we obtain,

F 0 (x) = (f ◦ g)0 (x) = f 0 (g(x))g 0 (x)
= f 0 (x3 )(3x2 )
= 3x2 f 0 (x3 ).
Thus,
F 0 (x) = 3x2 f 0 (x3 ).
Example 8.3.
Solutions to Examples (continued)

8.4. The function is y = (5x4 − 12x2 )3 . Decompose using a substi-

tution technique. Let u = 5x4 − 12x2 , then y = u3 . Thus

y = u3 u = 5x4 − 12x2
d d
y = 3u2 u = 20x3 − 24x.
du dx
By the Chain Rule,
dy dy du
=
dx du dx
= (3u2 )(20x3 − 24x)
= 6xu2 (5x2 − 6). (S-1)
The trouble with (S-1) is that the answer is in terms of x and u.
We want the derivative of y with respect to x; usually, we want the
answer entirely in terms of the independent variable, and that’s x. It
is simple enough to convert the answer in (S-1) to x — just replace u
with what it is equal to in terms of x: u = 5x4 − 12x2 .
Solutions to Examples (continued)

Substitute u = 5x4 − 12x2 into (S-1) to obtain

dy
= 6x(5x4 − 12x2 )2 (5x2 − 6).
dx
Thus, since y = (5x4 − 12x2 )3 , we have shown that
d
(5x4 − 12x2 )3 = 6x(5x4 − 12x2 )2 (5x2 − 6).
dx

Example Notes: Actually, the derivative can be simplified a bit more.

If we wanted to be true to our algebraic roots, we would continue.
d
(5x4 − 12x2 )3 = 6x5 (5x2 − 6)(5x2 − 12)2 .
dx
Verify please!
Example 8.4.
Solutions to Examples (continued)

d
8.5. The problem is to calculate (1−3x3 )10 . The first observation
dx
you should make is that we are required to differentiate a function
raised to a power. This is the first and most critical observation.
The second mental connection you need to make is this: Since we want
to differentiate a function raised to a power, the Power Rule needs to
be used first.
d d
(1 − 3x3 )10 = 10(1 − 3x3 )9 (1 − 3x3 ) Power Rule
dx dx
= 10(1 − 3x ) (−9x2 )
3 9
ditto
= −90x2 (1 − 3x3 )9
Thus,
d
(1 − 3x3 )10 = −90x2 (1 − 3x3 )9
dx
Example 8.5.
Solutions to Examples (continued)

8.6. The first thing we need to do is to rewrite the function to be

differentiated:
1
= (s4 − s + 1)−3/4 .
(s − s + 1)3/4
4

Now we can apply the power rule.

d 4 3 d
(s − s + 1)−3/4 = − (s4 − s + 1)−7/4 (s4 − s + 1)
ds 4 ds
(S-2)
3
= − (s4 − s + 1)−7/4 (4s3 − 1)
4
Thus,
d 1 3
= − (4s3 − 1)(s4 − s + 1)−7/4 .
ds (s − s + 1)
4 3/4 4

Example Notes: As you read the prepared examples, and do the ex-
ercises, be sure to verbalize the formulas as you use them. Usually, I
reference the verbal versions of the formulas. The Generalized Power
Solutions to Examples (continued)

Rule can be spoken as “the derivative of a function to a power is that

power times the base function raised to one less power, times the de-
rivative of the base function.” This is what I recited to myself as I
typed in equation (S-2).
Example 8.6.
Solutions to Examples (continued)

d 3√
8.7. The problem: x 1 + 3x2 . This is the derivative of a product
dx
of two functions. I’ll use the Product Rule.
d 3p 2 3 d 2 1/2 2 1/2 d x
3
x 1 + 3x = x (1 + 3x ) + (1 + 3x )
dx dx dx
1 3 d
= x (1 + 3x2 )−1/2 (1 + 3x2 ) + (1 + 3x2 )1/2 (3x2 )
2 dx
1 3
= x (1 + 3x2 )−1/2 (6x) + (1 + 3x2 )1/2 (3x2 )
2
This finishes the calculus part of the problem. Above, I have applied
the power rule (verbalize!) several times. We now start the algebra
part of the problem.
d 3p d 3
x 1 + 3x2 = x (1 + 3x2 )1/2
dx dx
1
= x3 ( )(1 + 3x2 )−1/2 (6x) + (1 + 3x2 )1/2 (3x2 )
2
= 3x (1 + 3x2 )−1/2 (x2 + (1 + 3x2 ))
2

= 3x2 (1 + 3x2 )−1/2 (4x2 + 1)

Solutions to Examples (continued)

Study the algebra steps — make sure these are familiar to you.
Thus,
d 3p 3x2 (4x2 + 1)
x 1 + 3x2 = √ .
dx 1 + 3x2
Example 8.7.
Solutions to Examples (continued)

8.8. We have initially the derivative of a quotient. Use the Quotient

Rule.
√
d w w d w3/2
=
dw (3w3 + 1)6 dw (3w3 + 1)6
(3w3 + 1)6 (3/2)w1/2 − w3/2 (6)(3w3 + 1)5 (9w2 )
=
(3w3 + 1)12
(S-3)
This finishes the calculus step. In the above calculation, I used the
Power Rule to calculate
d d
(3w3 + 1)6 = 6(3w3 + 1)5 (3w3 + 1) = 6(3w3 + 1)5 (9w2 )
dw dw
I’ll leave the algebra for you to verify. My simplified answer is
√ √
d w w 3 w(1 − 33w3 )
= (S-4)
dw (3w3 + 1)6 2(3w3 + 1)7

Keep working on (S-3) until you can obtain (S-4). Example 8.8.
Solutions to Examples (continued)

9.1. We are simply asked to calculate

d
sin(x) .
dx x=π/3

From formula (1) of Theorem 9.1,

d
sin(x) = cos(x).
dx
Therefore,
d
sin(x) = cos(x)|x=π/3
dx x=π/3
= cos(π/3)
1
= . (S-5)
2
Now turn to the problem of finding the equation of the tangent line
at x = π/3. This is just straight-line knowledge. We need two pieces
of information: A point on the target line, and the slope of the target
line.
Solutions to Examples (continued)
√
The Point: y = sin(x), when x = π/3, y = sin(x) = sin(π/3) = 3/2.
The given point then is √
π 3
P( , ).
3 2
The Slope: Our target line is tangent to the graph of y = sin(x) at
the point x = π/3. This is the interpretation of the derivative. Thus
dy 1
m tan = = ,
dx x=π/3 2
the last equality due to our earlier computation (S-5).
Equation of Tangent Line: y − y0 = m(x − x0 ) is the point-slope form
of the equation of a line that passes through P ( x0 , y0 ) having slope
m. Now we plug in the data,
√
3 1 π
y− = (x − ).
2 2 3
Example 9.1.
Solutions to Examples (continued)

d 3
9.2. The problem: Calculate x sin(x).
dx
We are asked to differentiate
(a) a sum (b) a product (c) a quotient (d) a sine function
We use standard techniques.
d 3 d d x3
x sin(x) = x3 sin(x) + sin(x)
dx dx dx
3 2
= x cos(x) + sin(x)(3x )
= x2 (x cos(x) + 3 sin(x))
That seemed easy. Example 9.2.
Solutions to Examples (continued)

9.3. We want to calculate the derivative of the cosine of a function

of x; use the (2) Trigonometric Derivative Formulas.
d d 3x4
cos(3x4 ) = − sin(3x4 )
dx dx
= − sin(3x )(12x3 )
4

= −12x3 sin(3x4 )
Example 9.3.
Solutions to Examples (continued)

10.1. We use the Power Rule to differentiate.

f (x) = x5
f 0 (x) = 5x4
f 00 (x) = 20x3
Example 10.1.
Solutions to Examples (continued)

10.2. We proceed as follows.

f (x) = x5
f 0 (x) = 5x4
f 00 (x) = 20x3
f 000 (x) = 60x2
f (4) (x) = 120x
f (5) (x) = 120
f (6) (x) = 0
For this function, beginning with the 6th -derivative, all derivatives are
identically 0. (This is characteristic of polynomials.)
f (n) (x) = 0 n ≥ 6.
Example 10.2.
Solutions to Examples (continued)

11.1. We follow the steps of implict differentiation.

Step 1: Think of y as an explicit function of x — but don’t bother to
try to find this explicit function.
Step 2: Take the equation
x2 + y 2 = 1,
and differentiate both sides with respect to x,
d 2 d1
(x + y 2 ) = ,
dx dx
and utilize the properties of differentiation

d x2 d y2
+ =0 / Additive Prop.
dx dx
dy
2x + 2y = 0. / Power Rule (S-6)
dx
Solutions to Examples (continued)

Thus,
dy
2x + 2y = 0. (S-7)
dx
In line (S-6), we have utilized Step 1. We treated y as if it were
a function of x. The derivative of y 2 , where y is a function of x, is,
according to the Power Rule, the exponent (2) times the base function
raised to one less power (y), times the derivative of the base function
(that’s dy/dx).
This completes Step 2. You’ll note the last equation generated, equa-
tion (S-7), is an equation involving the variables x, y, and dy/dx, as
was predicted by (9).
Solutions to Examples (continued)

Step 3: Solve (S-6) for dy/dx.

dy
2x + 2y =0
dx
dy
2y = −2x
dx
dy x
=− .
dx y
Thus,
dy x
=− . (S-8)
dx y
What does this representation of dy/dx mean? Notice that the answer
is in terms of both x and y. If the curve under consideration defined y
as a function of x, then knowledge of the value of x characterizes the
point on the curve under consideration. If the curve does not define
y as a function of x, as is the case for the circle, then knowledge of x
does not uniquely characterize the point on the curve. Corresponding
Solutions to Examples (continued)

to a particular x there is, if fact, two points on the curve — each

having a different tangent line.
Therefore, it should not be surprising that the representation of dy/dx
depends on both x and y. We need to know the complete set of coor-
dinates in order to characterize the point on the curve.
Thus,
dy 1/2
= −√
dx √
( x,y )=( 1/2, 3/2 ) 3/2
1
=√ .
3
This jives nicely with the much longer calculation (6).
Solutions to Examples (continued)

Similarly,
dy 1/2
=− √
dx √
( x,y )=( 1/2,− 3/2 ) − 3/2
1
=√ .
3
This corresponds to (7) made earlier. Example 11.1.
Solutions to Examples (continued)

11.2. Be simple minded, follow the procedure.

x10 − 2xy − y 10 = 0 (S-9)

d 10 d0
(x − 2xy − y 10 ) =
dx dx
10
d xy d y
10x9 − 2 − =0
dx dx
Now the second term is a product of two functions of x (Step 1, remem-
ber?), and the third term is a function of x raised to a power (Step 1,
remember?). For the second term we apply the Product Rule, and for
the third term, we apply the Power Rule. Continuing now,

d xy d y 10
10x9 − 2 − =0
dx dx
dy dx dy
10x9 − 2(x + y ) − 10y 9 =0
dx dx dx
dy dy
10x9 − 2x − 2y − 10y 9 = 0.
dx dx
Solutions to Examples (continued)

This is the end of Step 2. We now go into Step 3: Solve for dy/dx
dy dy
10x9 − 2x − 2y − 10y 9 =0
dx dx
dy
(10y 9 + 2x) = 10x9 − 2y
dx
dy
(5y 9 + x) = 5x9 − y.
dx
Finally,
dy 5x9 − y
= 9 .
dx 5y + x

For a particular numerical calculation, notice that the point

( x0 , y0 ) = ( 39/80 , 31/80 ) (S-10)
Solutions to Examples (continued)

satisfies the defining equation (S-9) (it took me awhile to find this
point, by the way!). The slope of the line tangent to the graph of
(S-9) at the point (S-10) is
dy 5x90 − y0
=
dx (x0 ,y0 ) 5y09 + x0
5 381/80 − 31/80
=
5 39/80 + 39/80
7 1 7
= = 39/10
33 1/10 9
≈ 2.090569740
Please verify the algebraic trench warfare above. Example 11.2.
Solutions to Examples (continued)

11.3. Given are given an equation x sin(xy) = 1 involving x and y.

It would be difficult, but not impossible, to solve for y; it is entirely
too much trouble. I’ll use implicit differentiation.
We begin by thinking (Step 1) of the symbol y as an anonymous
function of x, then we differentiation both sides of the equation as
per Step 2

x sin(xy) = 1
d d1
x sin(xy) =
dx dx
d
x sin(xy) + sin(xy) = 0.
dx
Now, in the first term, we have the derivative of the sine of some
function of x we’ll use Trig. (1),
d
x sin(xy) + sin(xy) = 0
dx
d
x(cos(xy) xy) + sin(xy) = 0.
dx
Solutions to Examples (continued)

Now we have a product of two functions of x; use the Product Rule,

d
x(cos(xy) xy) + sin(xy) = 0
dx
dy
x cos(xy)(x + y) + sin(xy) = 0.
dx
This finishes Step 2. Now we solve for dy/dx.
dy
x cos(xy)(x + y) + sin(xy) = 0
dx
dy
x2 cos(xy) + xy cos(xy) = − sin(xy)
dx
dy
x2 cos(xy) = − sin(xy) − xy cos(xy)
dx
dy sin(xy) + xy cos(xy)
=− .
dx x2 cos(xy)
Solutions to Examples (continued)

Thus,
dy sin(xy) + xy cos(xy)
=−
dx x2 cos(xy)
That’s the end of Step 3 and of this example. Example 11.3.
Solutions to Examples (continued)

11.4. We have calculate dy/dx already, Example 11.1, so I’ll skip

over that part:
dy x
=− . (S-11)
dx y
Second Derivative: The second derivative with respect to x is the
derivative of dy/dx with respect to x, i.e.,
d2 y d dy
2
= .
dx dx dx
We need only differentiation both sides of (S-11) with respect to x:
d dy d x
=− . (S-12)
dx dx dx y
The left-hand side of (S-12) is the second derivative, the right-hand
side is the derivative of the quotient of two functions of x — yes, we
take the implicit function attitude that y is an (implicit) function of
x.
Solutions to Examples (continued)

Let y 0 = dy/dx for simplicity of notation (and typing). Continuing

the calculation from (S-12),
d2 y d dy d x
= = −
dx2 dx dx dx y
y(1) − xy 0
=−
y2
y − xy 0
=− .
y2
This then is the desired second derivative. Notice that
y − xy 0
y 00 = −
y2
Solutions to Examples (continued)

is represented in terms of x, y, and y 0 . The answer can be improved

on by substituting our earlier calculation for y 0 .
y − xy 0
y 00 = −
y2
 
y − x − xy
=− / from (S-11)
y2
y 2 + x2
=− / algebra!
y3
1
= − 3, (S-13)
y
where, in the last step, we simplified y 2 + x2 down to 1, since it
is understood that the ( x, y ) pair satisfies the underlying equation
x2 + y 2 = 1. Thus,
1
y 00 = − 3 . (S-14)
y
Solutions to Examples (continued)

Third Derivative: To calculate the third derivative, we take the deriv-

ative of the second derivative, (S-14):

dy 00 d
y 000 = = − y −3 / from (S-14)
dx dx
dy
= 3y −4 / Power Rule
dx
 
−4 x
= 3y − / from (S-11)
y
3x
=− 5
y
Thus,
3x
y 000 = −
y5
Example 11.4.
Solutions to Examples (continued)

12.1. We simply follow the procedure for finding the absolute ex-
trema. First note that f is continuous over the interval [ −1, 2 ] so the
procedure applies.
Calculate Critical Points: Since f is a polynomial, it is everywhere
differentiable. This means the only critical points are the ones where
f 0 = 0.
f (x) = 3 + 4x − 3x3
f 0 (x) = 4 − 9x2
Now, set f 0 (x) = 0 and solve for x:
4 − 9x2 = 0
9x2 = 4
4
x2 =
9
2
x=±
3
Solutions to Examples (continued)

These are the critical points:

2 2
x1 = − , x2 = .
3 3
Note that both of these points belong to the target interval [ −1, 2 ].
We are not interested in anything outside this interval — had one or
more of the numbers been outside the interval, we would have deleted
them from our list.
Now we include the endpoints of the interval −1 and 2, and create a
table of calculations.
x f (x)
−1 2
2 11
−3 9
2 43
3 9 ⇐= max−1≤x≤2 f (x)
2 −13 ⇐= min−1≤x≤2 f (x)

Example 12.1.
Solutions to Examples (continued)

12.2. Suppose we are at the top of a hill in winter time wearing

our skis. The hill is a gentle one, flowing gently downward (as most
hills do) towards the bottom. Suddenly and gracefully, we push off
and ski down the hill to the bottom. As we slide down, our skis are
always tangent to the hill. In the course of our mad dash down the
hill, we can observe a mathematical truth. (Did you pick up on the
keyword tangent? I’m sure you did.) Let me try to describe to you
this mathematical truth to which I just referred.
Look at the hill from the side view. The hill then looks like a curve in
the xy-plane. Let T denote the point at the top of the hill the skier
started at, and let B denote the point at the bottom of the hill the
skier finished. Imagine drawing a line through these two points; call it
the line segment, T B, the gradient of the hill. Now here’s the trivial
observation that is stated formally by the Mean Value Theorem:
As the skier slides down the hill (remember, the skies are always
tangent to the hill), at some point on the hill, the skiier’s skis will be
parallel to the gradient of the hill.
Solutions to Examples (continued)

That is the observation: The skiers skies will be parallel to the gradient
of the hill at some time in the course of the ski trip. Example 12.2.
Solutions to Examples (continued)

12.3. Let x1 , x2 ∈ R such that x1 < x2 . Then

f (x1 ) = x1 < x2 = f (x2 ).
This means that f is strictly increasing. Example 12.3.
Solutions to Examples (continued)

12.4. Let’s look at this function over separate intervals.

On the interval ( 0, ∞ ): Let 0 < x1 < x2 . We want to argue that
f (x1 ) < f (x2 ). To that end
f (x2 ) − f (x1 ) = x22 − x21
= (x2 − x1 )(x2 + x1 )
>0 (S-15)
Since x1 < x2 we know x2 − x1 > 0. Since x1 > 0 and x2 > 0 we see
that x1 + x2 > 0. These observations justify (S-15).
But f (x2 )−f (x1 ) > 0 means f (x1 ) < f (x2 ), which is what we wanted
to prove. Thus, f is strictly increasing over the interval ( 0, ∞ ).
On the interval ( −∞, 0 ): Let x1 < x2 < 0. Prove that f (x1 ) > f (x1 ).
Indeed, as before,
f (x2 ) − f (x1 ) = x22 − x21
= (x2 − x1 )(x2 + x1 )
<0 (S-16)
Solutions to Examples (continued)

Since x1 < x2 we know x2 − x1 > 0. Since x1 < 0 and x2 < 0 we

see that x1 + x2 < 0. These observations justify (S-16). (Recall: the
product of a positive number and a negative number is a negative
number!)
But f (x2 )−f (x1 ) < 0 means f (x1 ) > f (x2 ), which is what we wanted
to prove. Thus, f is strictly decreasing over the interval ( −∞, 0 ).
Example 12.4.
Solutions to Examples (continued)

12.5. We simple take the first derivative of f

f 0 (x) = 4x3 − 16x. (S-17)

We ask ourselves the question: Where is f 0 positive and where is f 0

negative? To answer that question, we simply use standard techniques.
Begin by completely factoring the expression (S-17):
f 0 (x) = 4x(x2 − 4) = 4x(x − 2)(x + 2), (S-18)
where you can see that I have taken it easy on myself in the choice of
the function.
Next use the Sign Chart Method for analyzing when f 0 is positive and
negative.
The Sign Chart of f 0 (x) = 4x(x − 2)(x + 2)
x
0
x−2
2
Solutions to Examples (continued)

x+2
-2
f 0 (x)
-2 0 2
legend : • negative (−) • positive (+)
Intervals of increase are indicated Table of
in blue while intervals of decrease Increase/Decrease of f
are represented in red. The table f is decreasing on: ( −∞, −2 )
of summary results appears to the f is increasing on: ( −2, 0 )
right. f is decreasing on: ( 0, 2 )
f is increasing on: ( 2, +∞ )
Example 12.5.
Solutions to Examples (continued)

12.6. First calculate the first derivative and completely factor the
result.

f 0 (x) = 4x3 (x − 2)3 + 3x2 (x − 2)4 / Power Rule

= x2 (x − 2)3 [4x + 3(x − 2)] / factor it

= x2 (x − 2)3 (7x − 6) / combine similar terms

Thus,
f 0 (x) = x2 (x − 2)3 (7x − 6).
The Sign Chart of f 0 (x) = x2 (x − 2)3 (7x − 6)
x2
0
x−2
2
7x − 6
6/7
f 0 (x)
0 6/7 2
legend : • negative (−) • positive (+)
Solutions to Examples (continued)

Based on the Sign Chart results, The Table of

it is now easy to locate the in- Increase/Decrease of f
tervals of increase and decrease. f is increaseing on: ( −∞, 0 )
The blue indicates intervals of in- f is increasing on: ( 0, 6/7 )
crease and the red represents in- f is decreasing on: ( 6/7, 2 )
tervals of decrease. A table of re- f is increasing on: ( 2, +∞ )
sults appears to the right.
Example Notes: The sign of the factor x2 did not change at x = 0;
consequently, the monotone behavior did not change at x = 0. Actu-
ally, in the above table we could have combined the first two intervals
and simply stated that “f is increasing on the interval ( −∞, 6/7 ).”
What goes on at x = 0? This point is called a saddle point is a
the topic of discussion later in these notes.
The second factor of f 0 is (x − 2)3 ; however in the sign chart
I did not include the exponent. This was only for convenience. Since
the exponent is an odd integer the sign of (x − 2)3 is always the same
as the sign of x − 2, in the spirit of simplification, I analyzed the sign
of x − 2.
Solutions to Examples (continued)

For the first factor, I did include the exponent. This is because
the fact that the exponent is even effects the sign of the factor—
causing it to be always a nonnegative factor.
Example 12.6.
Solutions to Examples (continued)

12.7. The first derivative of f is given by

f 0 (x) = 4x(x2 − 4) = 4x(x − 2)(x + 2). (S-19)
To find the critical points, we look for points at which f 0 does not
exist (there are none) and points at which f 0 (x) = 0. No fancy solving
methods are necessary in (S-19) because the expression is completely
factored. It is clear form (S-19) that f 0 (x) = 0 with x = 0, −2, 2:
Critical Points: 0, −2, 2.
Now, we must examine the first derivative to the left and right of
these numbers. There is a couple of ways you can do this Examine
the Sign Chart of f 0 , or examine test points.
Sign Chart Method : From Example 12.5 the Sign Chart of f 0 was
computed to be
The Sign Chart of f 0 (x) = 4x(x − 2)(x + 2)
f 0 (x)
-2 0 2
legend : • negative (−) • positive (+)
Solutions to Examples (continued)

In the case of x = −2, the Sign Chart indicates that f 0 is negative to

the left and positive to the right of x = −2. By Theorem 12.12, f has
a local minimum there. For x = 0, the Sign Chart shows that f 0 is
positive to the left of x = 0 and negative to the right; this means that
x = 0 is a local maximum. Finally, to the left of x = 2, f 0 is negative
and to the right f 0 is positive.
The Sign Chart of f 0 (x) = 4x(x − 2)(x + 2)
f 0 (x)
-2 0 2
local min local max local min
legend : • negative (−) • positive (+)

Test Points: The creation of a Sign Chart does take time. Another
quicker method is to try some test points. Let me illustrated this
method for the critical point x = −2.
Take a number slightly to the left and slightly to the right of x = −2.
f 0 (−2.1) < 0
f 0 (−1.9) > 0
Solutions to Examples (continued)

We see from these calculations that f 0 is negative slightly to the left

of −2 and positive slightly to the right of −2. This indicates, by The-
orem 12.12, that f has a local minimum at x = −2.
Caveat Care must be taken not to choose a test point too far form
the critical point. You can see from the Sign Chart that the sign of f 0
changes several times. Should you choose a test point far enough away
from the critical point, the sign of f 0 may change and consequently,
leading you to a “false” conclusion.
Example 12.7.
Solutions to Examples (continued)

12.8. The first derivative is given by

f 0 (x) = x2 (x − 2)3 (7x − 6).
It is clear from this completely factored form of the first derivative
that the critical points are
6
Critical Points: 0, , 2.
7
0
The Sign Chart of f , as computed in Example 12.6 is
The Sign Chart of f 0 (x) = x2 (x − 2)3 (7x − 6)
f 0 (x)
0 6/7 2
legend : • negative (−) • positive (+)

Notice that f 0 is positive to the left of x = 0 and positive to the right

of x = 0. This means that f is increases up to the critical point, levels
off to a horizontal tangent at x = 0 (since f 0 (0) = 0) then continues
to increase to the right of x = 0. Thus, f 0 does not change sign at
x = 0; this means, from Theorem 12.12(3), f has a (increasing) saddle
point at x = 0.
The other two critical points are x = Classification of
6/7 and x = 2. The function f has a Critical Points
local maximum at x = 6/7 and has a 0 saddle point
local mimimum at x = 2. A table of 6/7 local maximum
summary results is given to the right. 2 local mimimum
Example 12.8.
Important Points
Important Points (continued)

To begin with, by the Extreme Value Theorem, what we are look

for exists — we just have to find them.
Let xmin denote a value of x at which f attains it’s absolute minimum,
i.e.
f (xmin ) = min f (x).
a≤x≤b

We need to “find” xmin . Where can xmin be? Logically speaking, xmin
is either an endpoint, a point at which f 0 does not exist, or a point at
which f 0 does exist. This represents an exhaustive analysis of xmin .
Now by Fermat’s Theorem, if f 0 exists at xmin , then f 0 (xmin ) = 0
— this is because xmin is a local extrema (since it is an absolute
extrema).
To update our analysis: Where can xmin be? It can be an endpoint,
or a point where f 0 does not exist, or at a point at which f 0 = 0. Or,
in other words, xmin is either an endpoint or a critical point.
Therefore, if we list all the endpoints and critical points, xmin must be
listed among them. Which one is it? Since f has an absolute minimum
Important Points (continued)

at xmin , then the smallest of the numbers in the right-hand column

of the table must be the absolute minimum.
A similar reasoning for the absolute maximum. Important Point
Important Points (continued)

A critical point is either a local minimum, a local maximum, or a

saddle point. The phrase classifying critical points refers to, in this
context, the problem of determining whether a given critical point is
either a local minimum, local maximum, or a saddle point.
Important Point
Important Points (continued)

Theorem 12.7 would not be true otherwise!

Example. Define a function f by

1 if x = 0
f (x) =
x otherwise
Think of the interval [ a, b ] as [ 0, 1 ]. The above function is differen-
tiable on the open interval ( 0, 1 ), in fact, f 0 (x) = 1 there, but not
continuous at x = 0.
Let’s calculate the right-hand side of equation (3):
f (b) − f (a) f (1) − f (0)
= =1−1=0
b−a 1−0
Yet, according to the Mean Value Theorem, there should be some
number c, 0 < c < 1, such that f 0 (c) = 0. (Recall, 0 is the value of
the right-hand side of equation (3).) But f 0 is never equal to 0 since
f 0 (x) = 1 for all x ∈ ( 0, 1 ). Thus the conclusion of the Mean Value
Theorem is not valid for this function.
Important Points (continued)

If the function f is not required to be continuous at the endpoints, one

could assign any arbitrary values to f at the endpoint; consequently,
there would be no reason to expect equation (3) to be true.
For those who want to know more. One could ask, “Why split hairs?
Why not just require f to be differentiable over [ a, b ]?” We could
certainly do that, but that would reduce the number of functions to
which the theorem can be applied. For example, the function

0 if x = 0
f (x) = (I-1)
x sin(1/x) if x 6= 0.
Think of the interval [ a, b ] as [ 0, 1 ] again. This function is continuous
on [ 0, 1 ] and differentiable on ( 0, 1 ), but f is not differentiable at
x = 0. (By not differentiable at the endpoint x = 0, I mean the
right-hand derivative of f at 0 does not exist.)
The Mean Value Theorem is applicable to this function, but if
we substitute the above revised wording, the “new” Mean Value
Theorem does not apply . . . because this f is not differentiable on
[ 0, 1 ].
Important Points (continued)

Exercise Verify the properties of the function f defined in (I-1): Ver-

ify that f is continuous at x = 0 but not differentiable at x = 0.
(Away from zero, the function f looks like x sin(1/x)—this function
is differentiable and the usual rules of differentiation can be applied
to compute its derivative.
Important Point