310

Chapter 3 Applications of Differentiation

3.8

RATES OF CHANGE IN APPLICATIONS

In this section, we round out our exposition of the derivative by presenting a collection of
applications selected from a variety of fields. It has often been said that mathematics is the
language of nature. Today, the concepts of calculus are being applied in virtually every
field of human endeavor. The applications in this section represent but a small sampling of
some elementary uses of the derivative. These are not all of the uses of the derivative nor
are they necessarily the most important uses. Our intent is simply to present some interesting applications in a variety of settings.
Recall that the derivative of a function gives the instantaneous rate of change of that
function. So, when you see the word rate, you should be thinking of a derivative. You can
hardly pick up a newspaper without finding reference to some rates (e.g., inflation rate or
interest rate). These can be thought of as derivatives. There are also many quantities with
which you are familiar, but that you might not recognize as rates of change. Our first
example is of this type.
Suppose that Q(t) represents the electrical charge in a wire at time t. Then, the derivative Q  (t) gives the current flowing through the wire. To see this, consider the cross section of a wire as shown in Figure 3.86. Between times t1 and t2 , the net charge passing
through such a cross section is Q(t2 ) Q(t1 ). The average current (charge per unit time)
over this time interval is then defined as
Q(t2 ) Q(t1 )
.
t2 t1
Figure 3.86
An electrical wire.

The instantaneous current I (t) at any time t1 can then be found by computing the limit
I (t1 ) = lim

tt1

Q(t) Q(t1 )
.
t t1

(8.1)

Notice that (8.1) is simply the alternative definition of derivative discussed in section 2.2.
Thus, we have that I (t) = Q  (t).
Example 8.1
5 ohms 10 farads

12 henry

~
8 volts

Figure 3.87
A simple electrical circuit.

Modeling Electrical Current in a Wire

The electrical circuit shown in Figure 3.87 includes a 5-ohm resistor, a 12-henry inductor, a 10-farad capacitor and a battery supplying 8 volts of AC current modeled by the
oscillating function 8 sin 2t , where t is measured in seconds. Find the current in the
circuit at any time t.
Solution
It can be shown (using the elementary laws of electricity) that the charge
in this circuit is given by
Q(t) = 12 sin(4t /3) + 4 sin 2t coulombs.
The current is then
Q  (t) = 48 cos(4t /3) + 8 cos 2t amps (coulombs per second).

The next example we offer comes from chemistry. It is very important for chemists to
have a handle on the rate at which a given reaction proceeds. Reaction rates give chemists
information about the nature of the chemical bonds being formed and broken, as well as information about the type and quantity of product to expect. A simple situation is depicted

Section 3.8

Rates of Change in Applications

311

in the schematic
A + B C,
which indicates that chemicals A and B (the reactants) combine to form chemical C (the
product). Let [C](t) denote the concentration (in moles per liter) of the product. The average reaction rate between times t1 and t2 is
[C](t2 ) [C](t1 )
.
t2 t1
The instantaneous reaction rate at any given time t1 is then given by
[C](t) [C](t1 )
d[C]
lim
=
(t1 ).
tt1
t t1
dt
Depending on the details of the reaction, it is often possible to write down an equation
d[C]
relating the reaction rate
to the concentrations of the reactants, [A] and [B].
dt
Example 8.2

Modeling the Rate of a Chemical Reaction

In an autocatalytic chemical reaction, the reactant and the product are the same. The reaction continues until some saturation level is reached. From experimental evidence,
chemists know that the reaction rate is jointly proportional to the amount of the product
present and the difference between the saturation level and the amount of the product. If
the initial concentration of the chemical is 0 and the saturation level is 1 (corresponding
to 100%), then the concentration x(t) of the chemical satisfies the equation
x  (t) = rx(t)[1 x(t)],
where r > 0 is a constant.
Find the concentration of chemical for which the reaction rate x  (t) is a maximum.
Solution

To clarify the problem, we write the reaction rate as

f (x) = rx(1 x).

Our aim is then to find x 0 that maximizes f (x). We have

f  (x) = r(1)(1 x) + rx(1)
= r(1 2x)
and so, the only critical number is x = 12 . Notice that the graph of y = f (x) is a
parabola opening downward and hence, the critical number must correspond to the absolute maximum. (Draw your own graph to see for yourself.) Although the mathematical problem here was easy to solve, the result gives a chemist some precise information.
At the time the reaction rate reaches a maximum, the concentration of chemical equals
exactly half of the saturation level.

Calculus and elementary physics are quite closely connected historically. It should
come as no surprise, then, that physics provides us with such a large number of important
applications of the calculus. We have already explored the concepts of velocity and acceleration. Another important application in physics where the derivative plays a role involves
density. There are many different kinds of densities that we could consider. For example,
we could study population density (number of people per unit area) or color density (depth
of color per unit area) used in the study of radiographs. However, the most common type

312

Chapter 3 Applications of Differentiation

x1
x

Figure 3.88
A thin rod.

of density discussed is mass density (mass per unit volume). You probably already have
some idea of what we mean by mass density, but how would you define it? If the object of
interest is made of some homogeneous material (i.e., the mass of any portion of the object
of a given volume is the same), then the mass density is simply
mass density =

mass
volume

and this quantity is constant throughout the object. However, if the mass of a given volume
varies in different parts of the object, then this formula only represents the average density
of the object. In the next example we find a means of computing the mass density at a specific point in a nonhomogeneous object.
Suppose that the function f (x) gives us the mass (in kilograms) of the first x meters of
a thin rod (see Figure 3.88).
The total mass between marks x and x1 (x > x1 ) is given by f (x) f (x1 ) kg. The
average linear density (i.e., mass per unit length) between x and x1 is then defined as
f (x) f (x1 )
.
x x1
Finally, the linear density at x = x1 is defined as
(x1 ) = lim

xx1

f (x) f (x1 )
= f  (x1 ),
x x1

(8.2)

where we have again recognized the alternative definition of derivative.

Example 8.3

Density of a Thin Rod

Suppose that the mass in a thin rod is given by f(x) = 2x . Compute the linear density
at x = 2 and at x = 8 and compare the densities at the two points.
Solution

From (8.2), we have

1
1
(x) = f  (x) = (2) = .
2 2x
2x

Thus, (2) = 1/ 4 = 1/2 and (8) = 1/ 16 = 1/4. Notice that this says that the rod
is inhomogeneous (i.e., the mass density in the rod is not constant). Specifically, we
have that the rod is less dense at x = 8 than at x = 2.

Our next example comes from medicine (cardiology, to be precise). The rate of change in
this application is not actually a derivative. However, heart rate is one of the most familiar and

Section 3.8

Rates of Change in Applications

313

Figure 3.89
A normal electrocardiogram (EKG).

important rates of change in our lives. The language and concepts of calculus enable us to understand some of the exciting research currently being conducted in this field.
Cardiologists have long used graphs to help them identify heart problems. You are
probably familiar with the term electrocardiogram (ECG or more commonly, EKG), a
graph depicting some of the electrical activity of the heart. A small section of a normal
EKG is shown in Figure 3.89.
A well-trained cardiologist can determine an amazing amount of information from an
EKG strip. Each section of the EKG (containing one peak) corresponds to a primary phase
of heart activity, and the particular shape of the curve can indicate various problems with
the heart. You should notice that the plot appears to be periodic (that is, the same shape is
repeated over and over). It is a simple matter to measure the length of the period from peak
to peak (each peak corresponds to a contraction of the ventricles of the heart, and the
distance between peaks gives the heart rate). Specifically, if successive peaks occur at
times t1 and t2 seconds, the heart rate equals
1 beat
60
beats per minute.
=
t2 t1 seconds
t2 t1
Note that heart rate is not a derivative. To get a derivative, we would need to have a function representing heart beats as a function of time and would then take the limit as t2 t1 .
From an EKG, all we have is a number of observations at specific times and therefore cannot compute a limit. Having defined heart rate, what can we learn about the human body?
Example 8.4

Heart Rates and Electrocardiograms

70
60
50
40
30
20
10

Heart rate (beats per minute)

Heart rate (beats per minute)

Heart rate (beats per minute)

The following graphs depict heart rate over time. Which one would you say corresponds to a healthy individual at rest?

70
60
50
40
30
20
10

Time
10

20
30
Constant rate

70
60
50
40
30
20
10

Time
10

20
30
Smooth change

Time
10

20
Erratic rate

30

314

Chapter 3 Applications of Differentiation

Solution
Before answering, we should clarify a couple of points. First, since heart
rate is only measured at discrete time intervals (the intervals separating the peaks of the
EKG), a graph of heart rate over time is like a computer graph: a finite set of points. If
the points are connected and there are many points, the plot may appear to be smooth
like the idealized plots we show. Second, be sure to read the graphs carefully! Each
graph shows a person resting with a heart rate of about 60 beats per minute (bpm). The
difference is in the variation from this average. The first person has a constant heart rate
of exactly 60 bpm. The graph depicts a steady 1 beat per second, and is not the flatline of a person whose heart has stopped beating. The second person shows a smooth
rise and fall in heart rate, and the third person has a somewhat erratic heart rate.
Which heart rate do you think is healthy? Research by cardiologists in the 1980s
and 1990s indicate that healthy hearts generally have a very erratic heart rate, even at
rest (Youre right if you picked plot 3!). Highly periodic heart rates like those in plot 2
have been observed in patients who were resting comfortably prior to experiencing a
heart attack. The constant heart rate in plot 1 is typical of a person experiencing a heart
attack. We should point out that the variations in the healthy third plot are too small to
discern without sophisticated equipment. These observations have led to the design of a
new type of heart monitor that can be worn by at-risk patients. Current research is investigating the possibility of designing smart heart pacemakers that can identify and
avoid an impending crisis.

The following example comes from economics. Much as with heart rate, the rate of
change discussed here is not precisely a derivative, but the derivative has proved to be a
useful tool in economic modeling. In economics, the term marginal is used to indicate a
rate. Thus, marginal cost is the derivative of the cost function, marginal profit is the derivative of the profit function, and so on.

Example 8.5

Analyzing the Marginal Cost of Producing

a Commercial Product

Suppose that
C(x) = 0.02x 2 + 2x + 4000
is the total cost (in dollars) for a company to produce x units of a certain product. Compute the marginal cost at x = 100 and compare this to the actual cost of producing the
100th unit.
Solution
You might think that it is an unfair assumption to start with a function
that purports to represent cost. After all, cost is determined by accountants after a product is produced. Thats true, but in order to project what your cost would be for quantities you havent actually produced, it is helpful to develop a mathematical model of
cost. In practice, this means that you make some observations of the cost of producing
a number of different quantities and then try to fit that data to the graph of a known
function, which you can then analyze using the tools of calculus. (This is one way in
which the calculus is brought to bear on real-world problems.) The marginal cost
function is
C  (x) = 0.04x + 2
and so, the marginal cost at x = 100 is C  (100) = 4 + 2 = 6 dollars per unit. On the
other hand, the actual cost of producing item number 100 would be C(100) C(99).

Section 3.8

Rates of Change in Applications

315

(Why?) We have
C(100) C(99) = 200 + 200 + 4000 (196.02 + 198 + 4000)
= 4400 4394.02 = 5.98 dollars.
Note that this is very close to the marginal cost of $6.

Our final example comes from psychology. You have probably heard references to
the learning curve for a piece of computer software or other technical product. The
phrase comes from research attempting to quantify and understand the dynamics of learning something new.
Example 8.6

Analyzing a Learning Curve

Suppose that the percentage of problems a person works correctly on a test is

approximately
80
f (t) =
1 + 3e0.4t
after t hours of training. Graph the learning curve y = f (t) and compute and interpret
f  (2) and f  (10).

y
100
80
60
40
20
x
2

Figure 3.90
y = 80/(1 + 3e0.4t ).

Solution
The graph of y = f (t) looks like the one shown in Figure 3.90.
Note that with no training the person gets 20% correct and after some steady
improvement seems to have trouble getting beyond the 80% mark. To compute the derivative, note that we do not need to use the quotient rule, since the numerator is simply
a constant. We rewrite the function as

10

f (t) = 80(1 + 3e0.4t )1

and obtain
f  (t) = 80(1 + 3e0.4t )2 (1.2e0.4t )
= 96e0.4t (1 + 3e0.4t )2 .
After 2 hours, we then have
f  (2) = 96e0.8 (1 + 3e0.8 )2 7.8243 percent per hour.
That is, after 2 hours, a person can expect to add about 7.8 points to the test score by
training for an extra hour. At the 10-hour mark,
f  (10) = 96e4 (1 + 3e4 )2 1.5799 percent per hour.
Thus, the next hour of training only increases the test score by about 1.6 points.

We have now discussed examples of six rates of change drawn from engineering and
the sciences. Add these to the examples we have seen in previous sections (velocity, population growth, etc.) and we have an impressive list of applications of the derivative. Even
so, we have barely begun to scratch the surface. In any field where it is possible to quantify
and analyze the properties of a function, calculus and the derivative are powerful tools.
This list includes at least some portion of every major field of study. The continued study
of calculus will give you the ability to read (and understand) technical studies in a wide variety of fields and to see (as we have in this section) the underlying unity that mathematics
brings to human endeavors.

316

Chapter 3 Applications of Differentiation

EXERCISES 3.8
1.

2.

3.

A variable that is defined only at isolated points is called

a discrete variable. For example, the size of a calculus
class is discrete, since it must be an integer. A variable that can
assume an interval of values is called continuous. In the six
applications discussed in the text, in which cases is time a discrete variable and in which cases is it continuous? Which functions are discrete and which are continuous? Among the six examples given in the text, in which cases were the rates of
change truly derivatives? Explain this answer in terms of discrete and continuous variables.

7.

Suppose that the charge at a particular location in an electrical

circuit is Q(t) = e3t cos 2t + 4 sin 3t coulombs. What happens to this function as t ? Explain why the term
e3t cos 2t is called a transient term and 4 sin 3t is known as the
steady-state or asymptotic value of the charge function. Find
the transient and steady-state values of the current function.

8.

As in exercise 7, find the steady-state and transient values of

the current function if the charge function is given by Q(t) =
e2t (cos t 2 sin t) + te3t + 2 cos 4t.

An important model of population growth is the socalled logistic equation x  (t) = x(t)[1 x(t)]. Here,
x(t) represents not the actual population size but the proportion
of sustainable capacity: for instance, x(t) = 0.5 means that the
population is half of the total number of organisms that the environment can support and x(t) = 1.1 means that there are
10% more organisms than the available resources can support.
Note that the differential equation here is the same as was used
to describe an autocatalytic chemical reaction. The equation
has two competing contributions to the rate of change x  (t).
The term x(t) by itself would mean that the larger x(t) is, the
faster the population (or concentration of chemical) grows.
This is balanced by the term 1 x(t), which indicates that
the closer x(t) gets to 1, the slower the population growth is.
With these two terms together, the model has the property that
for small x(t), slightly larger x(t) means greater growth, but as
x(t) approaches 1, the growth tails off. Explain in terms of
population growth and the concentration of chemical why the
model is reasonable.

9.

If the concentration of a chemical changes according to the

equation x  (t) = 2x(t)[4 x(t)], find the concentration x(t)
for which the reaction rate is a maximum.

10.

If the concentration of a chemical changes according to the

equation x  (t) = 0.5x(t)[5 x(t)], find the concentration x(t)
for which the reaction rate is a maximum.

11.

Show that in exercise 9, the maximum concentration is 4 if

0 < x(0) < 4. Find the maximum concentration in exercise 10.

12.

Find the equation for an autocatalytic reaction in which the

maximum concentration is x(t) = 16 and the reaction rate
equals 12 when x(t) = 8.

13.

Mathematicians often study equations of the form x  (t) =

rx(t)[1 x(t)] instead of the more complicated x  (t) =
cx(t)[K x(t)], justifying the simplification with the statement
that the second equation reduces to the first equation. Starting
with y  (t) = cy(t)[K y(t)], substitute y(t) = Kx(t) and show
that the equation reduces to the form x  (t) = rx(t)[1 x(t)].
How does the constant r relate to the constants c and K?

14.

Suppose a chemical reaction follows the equation x  (t) =

cx(t)[K x(t)]. Suppose that at time t = 4 the concentration is
x(4) = 2 and the reaction rate is x  (4) = 3. At time t = 6, suppose that the concentration is x(6) = 4 and the reaction rate is
x  (6) = 4. Find the values of c and K for this chemical reaction.

15.

In a general second-order chemical reaction, chemicals A and

B (the reactants) combine to form chemical C (the product). If
the initial concentrations of the reactants A and B are a and b,
respectively, then the concentration x(t) of the product satisfies
the equation x  (t) = [a x(t)][b x(t)]. What is the rate of
change of the product when x(t) = a? At this value, is the concentration of product increasing, decreasing or staying the
same? Assuming that a < b and there is no product present
when the reaction starts, explain why the maximum concentration of product is x(t) = a.

16.

For the second-order reaction defined in exercise 15, find the

(mathematical) value of x(t) that minimizes the reaction rate.

Corporate deficits and debt are frequently in the news,

but the terms are often confused with each other. To take
an example, suppose a company finishes a fiscal year owing
$5,000. That is their debt. Suppose that in the following year
the company has revenues of $106,000 and expenses of
$109,000. The companys deficit for the year is $3,000 and the
companys debt has increased to $8,000. Briefly explain why
deficit is the derivative of debt.

4.

Many people find the healthy heart example (example 8.4) surprising. To make it more believable, explain
how your level of activity affects your heart rate. Also, explain
how breathing rate and emotional status can affect heart rate.
Given these and many other factors, discuss whether you
should expect your heart rate to be exactly constant.

5.

Suppose that the charge in an electrical circuit is Q(t) =

e2t (cos 3t 2 sin 3t) coulombs. Find the current.

6.

Suppose that the charge in an electrical circuit is Q(t) =

et (3 cos 2t + sin 2t) coulombs. Find the current.

Section 3.8

Show that the reaction rate for this value of x(t) is negative.
Explain why the concentration x(t) would never get this large,
so that this mathematical solution is not physically relevant.
Explain why x(t) must be between 0 and a, and find the maximum and minimum reaction rates on this closed interval.
17.

It can be shown that a solution of the equation x  (t) =

[a x(t)][b x(t)] is given by
a[1 e(ba)t ]
x(t) =
.
1 (a/b)e(ba)t
Find x(0), the initial concentration of chemical, and lim x(t),
t
the limiting concentration of chemical (assume a < b). Graph
x(t) on the interval [0, ) and describe in words how the concentration of chemical changes over time.

18.

For the solution in exercise 17, find and graph x  (t). Compute
lim x  (t) and describe in words how the reaction rate changes

over time.

m(x) = 4x sin x grams for 0 x 6

20.

m(x) = (x 1)3 + 6x grams for 0 x 2

21.

m(x) = 4x grams for 0 x 2

22.

m(x) = 4x 2 grams for 0 x 2

In exercises 2326, you are given times (in seconds) at which

peaks in a heart patients EKG occurred. Analyze the heart rate
and diagnose the patient as resting comfortably or in danger of a
heart attack.
23.

0.0, 0.99, 2.12, 3.19, 4.12, 5.17, 6.08, 6.95, 8.16, 9.26, 10.24,
11.28, 12.22, 13.14, 14.21, 15.18

24.

0.0, 0.99, 2.02, 3.01, 4.02, 5.03, 6.01, 6.99, 8.01, 9.02, 10.02,
11.02, 12.02, 13.01, 14.01, 15.02

25.

0.0, 0.98, 1.96, 2.90, 3.80, 4.74, 5.72, 6.70, 7.70, 8.72, 9.74,
10.80, 11.90, 13.06, 14.08, 15.10

26.

0.0, 0.98, 1.96, 3.02, 3.98, 5.04, 6.22, 7.17, 8.15, 9.17, 10.13,
11.08, 12.11, 13.06, 14.08, 15.10

27.

If the cost of manufacturing x items is C(x) = x 3 + 20x 2 +

90x + 15, find the marginal cost function and compare the
marginal cost at x = 50 and the actual cost of manufacturing
the 50th item.

28.

If the cost of manufacturing x items is C(x) = x 4 + 14x 2 +

60x + 35, find the marginal cost function and compare the

317

marginal cost at x = 50 and the actual cost of manufacturing

the 50th item.
29.

If the cost of manufacturing x items is C(x) = x 3 +

21x 2 + 110x + 20, find the marginal cost function and compare the marginal cost at x = 100 and the actual cost of manufacturing the 100th item.

30.

If the cost of manufacturing x items is C(x) = x 3 + 11x 2 +

40x + 10, find the marginal cost function and compare the
marginal cost at x = 100 and the actual cost of manufacturing
the 100th item.

31.

Suppose the cost of manufacturing x items is C(x) =

x 3 30x 2 + 300x + 100 dollars. Find the inflection point and
discuss the significance of this value in terms of the cost of
manufacturing.

32.

A baseball team owner has determined that if tickets are priced

at $10, the average attendance at a game will be 27,000 and if
tickets are priced at $8, the average attendance will be 33,000.
Using a linear model, we would then estimate that tickets
priced $9 would produce an average attendance of 30,000. Discuss whether or not you think the use of a linear model here is
reasonable. Then, using the linear model, determine the price at
which the revenue is maximized.

33.

The learning curve f (t) = 80/(1 + 3e0.4t ) was discussed in

example 8.6. Show that lim f (t) = 80 and show that f (t) is

In exercises 1922, the mass of the first x meters of a thin rod is

given by the function m(x) on the indicated interval. Find the linear density function for the rod. Based on what you find, briefly
describe the composition of the rod.
19.

Rates of Change in Applications

an increasing function for t > 0. Discuss whether you think

both results are reasonable properties for a learning curve.
34.

Suppose the maximum score that a person will ever score

on a test is 90. Further suppose that the person scores 60 after
2 hours of studying. Find a learning curve of the form f (t) =
a/(1 + 3ebt ) for this person on this test.

35.

Suppose that a person scores f (t) = 180/(2 + 4e0.2t ) after

t hours of studying. What is the persons score after 3 hours of
studying? Find f  (3) and estimate how many additional points
the person would earn by studying a fourth hour. Find f  (10)
and estimate how many additional points the person would
earn by studying an eleventh hour.

36.

Suppose that a person scores f (t) = 10/(1 + 4e0.5t ) after

t hours of studying. What is the persons score after 4 hours of
studying? Find f  (4) and estimate how many additional points
the person would earn by studying a fifth hour. Find f  (10) and
estimate how many additional points the person would earn by
studying an eleventh hour.

37.

The function f (t) = a/(1 + 3ebt ) has also been used to

model the spread of a rumor. Suppose that a = 70 and b = 0.2.
Compute f (2), the percentage of the population that has heard
the rumor after 2 hours. Compute f  (2) and describe what it
represents. Compute lim f (t) and describe what it represents.
t

318

Chapter 3 Applications of Differentiation

38.

After an injection, the concentration of drug in a muscle varies

according to a function of time, f (t). Suppose that t is measured in hours and f (t) = e0.02t e0.42t. Determine the time
when the maximum concentration of drug occurs.

39.

Suppose that the size of the pupil of an animal is given by

f (x) (mm), where x is the intensity of the light on the pupil. If
f (x) =

titration curve below, identify on the graph both inflection

points and briefly explain why chemists prefer to measure the
equivalence point and not the inflection point of exercise 47.
(Note: the horizontal axis of a titration curve indicates the
amount of base added to the mixture. This is directly proportional to the amount of converted acid in the region where
0 < f < 1.)

160x 0.4 + 90
,
4x 0.4 + 15

show that f (x) is a decreasing function. Interpret this result in

terms of the response of the pupil to light.
40.

Suppose that the body temperature 1 hour after receiving x mg

of a drug is given by T (x) = 102 16 x 2 (1 x/9) for
0 x 6. The absolute value of the derivative, |T  (x)|, is defined as the sensitivity of the body to the drug dosage. Find the
dosage which maximizes sensitivity.

41.

Let C(x) be the cost of manufacturing x items and define

C(x)
= C(x)/x as the average cost function. Suppose that

C(x) = 0.01x 2 + 40x + 3600. Show that C  (100) < C(100)

and show that increasing the production (x) by 1 will decrease
the average cost.

42.

For the cost function in exercise 41, show that C  (1000) >

C(1000)
and show that increasing the production (x) by 1 will
increase the average cost.

43.

For the cost function in exercise 41, prove that average cost is

minimized at the x-value where C  (x) = C(x).

44.

If the cost function is linear, C(x) = a + bx with a and b positive, show that there is no minimum average cost and that

C  (x) = C(x)
for all x.

45.

Let R(x) be the revenue and C(x) be the cost from manufacturing x items. Profit is defined as P(x) = R(x) C(x). Show
that at the value of x that maximizes profit, marginal revenue
equals marginal cost.

46.

Find the maximum profit if R(x) = 10x 0.001x 2 dollars and

C(x) = 2x + 5000 dollars.

47.

In the titration of a weak acid and strong base, the pH is given

f
where c is a constant (closely related to the
by c + ln
1 f
acid dissociation constant) and f is the fraction (0 < f < 1)
of converted acid (see Harris Quantitative Chemical Analysis
for more details). Find the value of f at which the rate of change
of pH is the smallest. What happens as f approaches 1?

48.

In exercise 47, you found the significance of one inflection

point of a titration curve. A second inflection point, called the
equivalence point, corresponds to f = 1. In the generalized

pH

ml of base added

49.

Epidemiology is the study of the spread of infectious

diseases. A simple model for the spread of fatal diseases
such as AIDS divides people into the categories of susceptible
(but not exposed), exposed (but not infected) and infected. The
proportions of people in each category at time t are denoted
S(t), E(t) and I (t), respectively. The general equations for this
model are
S  (t) = m I (t) bS(t)I (t),
E  (t) = bS(t)I (t) a E(t),
I  (t) = a E(t) m I (t),
where m, b and a are positive constants. Notice that each equation gives the rate of change of one of the categories. Each rate
of change has both a positive and negative term. Explain why
the positive term represents people who are entering the category and the negative term represents people who are leaving
the category. In the first equation, the term m I (t) represents
people who have died from the disease (the constant m is the
reciprocal of the life expectancy of someone with the disease).
This term is slightly artificial: the assumption is that the population is constant, so that when one person dies, a baby is born
who is not exposed or infected. The dynamics of the disease is
that susceptible (healthy) people get infected by contact with
infected people. Explain why the number of contacts between
susceptible people and infected people is proportional to S(t)
and I (t). The term bS(t)I (t), then, represents susceptible people who have been exposed by contact with infected people.
Explain why this same term shows up as a positive in the second equation. Explain the rest of the remaining two equations
in this fashion. (Hint: The constant a represents the reciprocal
of the average latency period. In the case of AIDS, this would
be how long it takes an HIV-positive person to actually develop
AIDS.)

50.

Without knowing how to solve differential equations

(we hope you will go far enough in your study of mathematics to learn to do so!), we can nonetheless deduce some

Chapter Review Exercises

319

x(t) decreases and we could again conjecture that

lim x(t) = 1. Changing equations, suppose that x  (t) =

important properties of the solutions of differential equations.

For example, consider the equation for an autocatalytic reaction x  (t) = x(t)[1 x(t)]. Suppose x(0) lies between
0 and 1. Show that x  (0) is positive by determining the possible
values of x(0)[1 x(0)]. Explain why this indicates that the
value of x(t) will increase from x(0), and will continue to increase as long as 0 < x(t) < 1. Explain why if x(0) < 1 and
x(t) > 1 for some t > 0, then it must be true that x(t) = 1 for
some t > 0. However, if x(t) = 1, then x  (t) = 0 and the solution x(t) stays constant (equal to 1). Therefore, we can conjecture that lim x(t) = 1. Similarly, show that if x(0) > 1, then

0.05x(t) + 2. This is a model of an experiment in which a radioactive substance is decaying at the rate of 5% but the substance is being replenished at the constant rate of 2. Find the
value of x(t) for which x  (t) = 0. Pick various starting values
of x(0) less than and greater than the constant solution and determine if the solution x(t) will increase or decrease. Based on
these conclusions, conjecture the value of lim x(t), the limitt
ing amount of radioactive substance in the experiment.

CHAPTER REVIEW EXERCISES

In exercises 1 and 2, find the linear approximation to f(x) at x0.

1. f (x) = e3x , x0 = 0
2. f (x) = x 2 + 3, x0 = 1
In exercises 3 and 4, use a linear approximation to estimate the
quantity.
3.

3
7.96

4. sin 3

In exercises 5 and 6, use Newtons method to find an approximate

root.
x 3 = ex

5.

x 3 + 5x 1 = 0

7.

Explain why Newtons method fails on x 3 3x + 2 = 0 with

x0 = 1.

8.

6.

1
1 + x is valid for
Show that the approximation
(1 x)
small x.

In exercises 918, do the following by hand. (a) Find all critical

numbers, (b) identify all intervals of increase and decrease,
(c) determine whether each critical number represents a local
maximum, local minimum or neither, (d) determine all intervals
of concavity and (e) find all inflection points.
9.

f (x) = x + 3x 9x

11.
13.

10.

f (x) = x 4x + 1

f (x) = x 4x + 2

12.

f (x) = x 3x 24x

14.

f (x) = x 2 ln x

15.

f (x) = xe4x

f (x) = x x 2 4

16.

f (x) = (x 1)

17.

f (x) =

18.

f (x) =

x2

x
+4

In exercises 1922, find the absolute extrema of the function on

the interval.
19.

f (x) = x 3 + 3x 2 9x on [0, 4]

20.

f (x) = x 3 + 3x 2 9x on [4, 0]

21.

f (x) = x 4/5 on [2, 3]

22.

f (x) = x 2 ex on [1, 4]

In exercises 2326, find the x-coordinates of all local extrema.

23.

f (x) = x 3 + 4x 2 + 2x

24.

f (x) = x 4 3x 2 + 2x

25.

f (x) = x 5 2x 2 + x

26.

f (x) = x 5 + 4x 2 4x

27.

Sketch a graph of a function with f (1) = 2, f (1) = 2,

f  (x) < 0 for 2 < x < 2, f  (x) > 0 for x < 2 and x > 2.

28.

Sketch a graph of a function with f  (x) > 0 for x = 0, f  (0)

undefined, f  (x) > 0 for x < 0 and f  (x) < 0 for x > 0.

In exercises 2938, sketch a graph of the function and completely

discuss the graph.
29.

f (x) = x 4 + 4x 3

30.

f (x) = x 4 + 4x 2

31.

f (x) = x 4 + 4x

32.

f (x) = x 4 4x 2

33.

f (x) =

x
x2 + 1

34.

f (x) =

35.

f (x) =

x2
+1

36.

f (x) =

37.

f (x) =

x3
1

38.

f (x) =

2/3

x
x2 + 2

x2
x2

x
x2 1
x2
1

x2

4
x2 1

320

Chapter 3 Applications of Differentiation

39.

Find the point on the graph of y = 2x 2 that is closest to (2, 1).

40.

Show that the line through the two points of exercise 39 is perpendicular to the tangent line to y = 2x 2 at (2, 1).

41.

A city is building a highway from point A to point B, which is

4 miles east and 6 miles south of point A. The first 4 miles south
of point A is swamp land, where the cost of building the highway is $6 million per mile. On dry land, the cost is $2 million
per mile. Find the point on the boundary of swamp land and dry
land to which the highway should be built to minimize the
total cost.

42.

Repeat exercise 41 with a cost of $16 million per mile on

swamp land. Explain why the optimal point in this exercise is
west of the optimal point found in exercise 41.

43.

A soda can in the shape of a cylinder is to hold 16 ounces. Find

the dimensions of the can that minimizes the surface area of
the can.

44.

Suppose that C(x) = 0.02x 2 + 4x + 1200 is the cost of manufacturing x items. Show that C  (x) > 0 and explain in business
terms why this has to be true. Show that C  (x) > 0 and explain
why this indicates that the manufacturing process is not very
efficient.

45.

The charge in an electrical circuit at time t is given by

Q(t) = e3t sin 2t coulombs. Find the current.

46.

If the concentration x(t) of a chemical in a reaction changes

according to the equation x  (t) = 0.3x(t)[4 x(t)], find the
concentration at which the reaction rate is a maximum.

47.

Suppose that the mass of the first x meters of a thin rod is given
by m(x) = 20 + x 2 for 0 x 4. Find the density of the rod
and briefly describe the composition of the rod.

48.

A person scores f (t) = 90/(1 + 4e0.4t ) points on a test after t

hours of studying. What does the person score without studying
at all? Compute f  (0) and estimate how many points one hour
of studying will add to the score.

49.

The cost of manufacturing x items is given by C(x) =

0.02x 2 + 20x + 1800. Find the marginal cost function. Compare the marginal cost at x = 20 to the actual cost of producing
the 20th item.

50.

For the cost function in exercise 49, find the value of x which

= C(x)/x .
minimizes the average cost C(x)

51.

Let n(t) be the number of photons in a laser field. One

model of the laser action is n  (t) = an(t) b[n(t)]2 ,
where a and b are positive constants. If n(0) = a/b, what is
n  (0)? Based on this calculation, would n(t) increase, decrease
or neither? If n(0) > a/b, is n  (0) positive or negative? Based
on this calculation, would n(t) increase, decrease or neither? If
n(0) < a/b, is n  (0) positive or negative? Based on this calculation, would n(t) increase, decrease or neither? Putting this information together, conjecture the limit of n(t) as t . Repeat this analysis under the assumption that a < 0. [Hint:
Because of its definition, n(t) is positive, so ignore any negative values of n(t).]

52.

One way of numerically approximating a derivative is

by computing the slope of a secant line. For example,
f (b) f (a)

f (a)
, if b is close enough to a. In this exercise,
ba
we will develop an analogous approximation to the second derivative. Graphically, we can think of the secant line as an approximation of the tangent line. Similarly, we can match the
second derivative behavior (concavity) with a parabola. Instead
of finding the secant line through two points on the curve, we
find the parabola through three points on the curve. The second
derivative of this approximating parabola will serve as an approximation of the second derivative of the curve. The first step
is messy, so we recommend using a CAS if one is available.
Find a function of the form g(x) = ax 2 + bx + c such that
g(x1 ) = y1 , g(x2 ) = y2 and g(x3 ) = y3 . Since g  (x) = 2a,
you actually only need to find the constant a. The so-called second difference approximation to f  (x) is the value of
g  (x) = 2a using the three points x1 = x x [y1 = f (x1 )],
x2 = x [y2 = f (x2 )] and x3 =x + x [y3 = f (x3 )]. Find the
second difference for f (x) = x + 4 at x = 0 with x = 0.5,
x = 0.1 and x = 0.01. Compare to the exact value of the
second derivative, f  (0).