Ubc Diffcal PDF

“ T H E R E I S N O K N O W L E D G E T H AT I S N O T P O W E R .
R A L P H WA L D O E M E R S O N , ( 1 8 0 3 - 1 8 8 2 )
LEAH EDELSTEIN-KESHET
DIFFERENTIAL CALCU-
LUS FOR THE LIFE SCI-
ENCES
T H E A U T H O R AT T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A
Copyright © 2020 Leah Edelstein-Keshet
P U B L I S H E D B Y T H E AU T H O R AT T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A
http://www.math.ubc.ca/~keshet/OpenBook.pdf
License information: Differential Calculus for the Life Sciences by Leah Edelstein-Keshet is licensed under a Creative
Commons Attribution-NonCommercial-ShareAlike 4.0 International License. To view a copy of this license, visit
http://creativecommons.org/licenses/by-nc-sa/4.0/
Printing as of January 2020

Contents
1 Power functions as building blocks 25

1.1 Power functions 25
1.2 How big can a cell be? A model for nutrient balance 27
1.3 Sustainability and energy balance on Earth 30
1.4 First steps in graph sketching 31
1.5 Rate of an enzyme-catalyzed reaction 36
1.6 Predator Response 39
1.6.1 A ladybug eating aphids 39
1.7 Summary 41
Exercises 43
2 Average rates of change, average velocity and the secant line 51

2.1 Time-dependent data and rates of change 52
2.2 The slope of a straight line is a rate of change 54
2.3 The slope of a secant line is the average rate of change 55
2.4 From average to instantaneous rate of change 58
2.5 Introduction to the derivative 61
2.6 Summary 63
Exercises 65
3 Three faces of the derivative: geometric, analytic, and computational 69

3.1 The geometric view: zooming into the graph of a function 69
6
3.2 The analytic view: calculating the derivative 75

3.3 The computational view: software to the rescue! 80
3.4 Summary 82
Exercises 84
4 Differentiation rules, simple antiderivatives and applications 89

4.1 Rules of differentiation 89
4.2 Application of the second derivative to acceleration 97
4.3 Sketching the first and second derivative and the anti-derivatives 100
4.4 Summary 102
Exercises 104
5 Tangent lines, linear approximation, and Newton’s method 109

5.1 The equation of a tangent line 109
5.2 Generic tangent line equation and properties 112
5.3 Approximating a function by its tangent line 113
5.4 Tangent lines for finding zeros of a function: Newton’s method 116
5.5 Aphids and Ladybugs, revisited 119
5.6 Harder problems: finding the point of tangency 120
5.7 Summary 122
Exercises 124
6 Sketching the graph of a function using calculus tools 127

6.1 Overall shape of the graph of a function 127
6.2 Special points on the graph of a function 131
6.3 Sketching the graph of a function 134
6.4 Summary 138
Exercises 140
7
7 Optimization 145
7.1 Simple biological optimization problems 145
7.2 Optimization with a constraint 148
7.3 Checking endpoints 152
7.4 Optimal foraging 153
7.5 Summary 157
Exercises 160
8 Introducing the chain rule 167

8.1 The chain rule 167
8.2 The chain rule applied to optimization problems 172
8.3 Summary 179
Exercises 181
9 Chain rule applied to related rates and implicit differentiation 183

9.1 Applications of the chain rule to “related rates” 183
9.2 Implicit differentiation 187
9.3 The power rule for fractional powers 191
9.4 Summary 194
Exercises 196
10 Exponential functions 203

10.1 Unlimited growth and doubling 203
10.2 Derivatives of exponential functions and the function ex 206
10.3 Inverse functions and logarithms 211
10.4 Applications of the logarithm 214
10.5 Summary 217
Exercises 219
8
11 Differential equations for exponential growth and decay 227

11.1 Introducing a new kind of equation 227
11.2 Differential equation for unlimited population growth 231
11.3 Radioactive decay 236
11.4 Deriving a differential equation for the growth of cell mass 240
11.5 Summary 242
Exercises 244
12 Solving differential equations 249

12.1 Verifying that a function is a solution 249
12.2 Equations of the form y0 (t ) = a − by 252
12.3 Euler’s Method and numerical solutions 260
12.4 Summary 265
Exercises 267
13 Qualitative methods for differential equations 275

13.1 Linear and nonlinear differential equations 275
13.2 The geometry of change 280
13.3 Applying qualitative analysis to biological models 287
13.3.1 The radius of a growing cell 290
13.4 Summary 295
Exercises 297
14 Periodic and trigonometric functions 303

14.1 Basic trigonometry 304
14.2 Periodic functions 308
14.3 Inverse trigonometric functions 315
14.4 Summary 319
Exercises 321
9
15 Cycles, periods, and rates of change 325

15.1 Derivatives of trigonometric functions 325
15.2 Changing angles and related rates 328
15.3 The Zebra danio’s escape responses 332
15.4 Summary 338
Exercises 340
16 Additional exercises 347

Exercises 347
Appendices 361
A A review of Straight Lines 363

A.1 Geometric ideas: lines, slopes, equations 363
B A precalculus review 365

B.1 Manipulating exponents 365
B.2 Manipulating logarithms 365
C A Review of Simple Functions 367

C.1 What is a function? 367
C.2 Geometric transformations 368
C.3 Classifying 369
C.4 Power functions and symmetry 369
C.5 Inverse functions and fractional powers 372
C.6 Polynomials 375
D Limits 377
D.1 Limits for continuous functions 377
D.2 Properties of limits 378
10
D.3 Limits of rational functions 379

D.4 Right and left sided limits 381
D.5 Limits at infinity 382
D.6 Summary of special limits 383
E Proofs 385
E.1 Proof of the power rule 385
E.2 Proof of the product rule 386
E.3 Proof of the quotient rule 387
E.4 Proof of the chain rule 388
F Trigonometry review 391

F.1 Summary of the inverse trigonometric functions 393
G For further study 395

G.1 Michaelis-Menten transformed to a linear relationship 395
G.2 Spacing of fish in a school 396
G.3 A biological speed machine 397
G.4 Additional examples of geometric optimization 399
G.5 Optimal foraging with other patch functions 402
G.6 Trigonometric functions and differential equations 404
H Short Answers to Problems 407

H.0.1 Answers to Chapter 16 Problems 435
Bibliography 439
List of Figures
1.1 Graphs of power functions 26

1.2 Intersection point of two power functions 26
1.3 A spherical cell 28
1.4 Even and odd power functions 32
1.5 Sketching a simple polynomial 34
1.6 Sketching rational functions 36
1.7 Enzyme binds a substrate 37
1.8 Michaelis-Menten kinetics 37
1.9 Hill function kinetics 38
1.10 Holling 39
1.11 Ladybugs eating aphids 40
1.12 Rate of a chemical reaction 48
1.13 Speed of polymerization 48
1.14 Lineweaver-Burk plots 50
2.1 Kymograph used to track motion of a labeled protein 51

2.2 Milk temperature plots 52
2.3 Bluefin Tuna migration 53
2.4 The distance of a falling object 54
2.5 Velocity of labeled microtubule tips, estimated with a kymograph 55
2.6 Secant line 56
2.7 Secant line of arbitrary function 57
2.8 Secant line and height of a falling object 58
2.9 Refining data for cooling milk 59
2.10 The height of a falling object 60
3.1 Zooming in on the graph of a function 70

3.2 Zooming in on the sine graph 71
3.3 A function with a cusp 71
3.4 Zooming in on a discontinuity 71
3.5 The graph of a function 72
3.6 Function tangent lines 72
3.7 Function derivative sketch 72
3.8 Function whose derivative we sketch 72
12
3.9 Function derivative sketching 73

3.10 Molecular motors 73
3.11 Displacement and velocity of molecular motors 74
3.12 Continuous and discontinuous functions 77
3.13 Computational software plot of the derivative 81
3.14 Graphs of four functions 84
3.15 Two similar functions increasing at distinct rates 84
3.16 Motion of a particle 85
3.17 Functions and derivatives to sketch 85
3.18 Molecular motors 86
3.19 Concentration profile of a substance 87
4.1 Position, velocity, and acceleration for falling object 99

4.2 Two graphs: one smooth, one with cusps 100
4.3 Sketching antiderivatives 101
4.4 Two derivative plots 101
4.5 Sketching antiderivatives 102
4.6 Sketching an antiderivative 107
4.7 Sketching a derivative 107
4.8 Sketching derivatives and antiderivatives 107
5.1 A parabola and two tangent lines 110

5.2 A function with its tangent line 110
5.3 Shortest ladder 111
5.4 Generic tangent line 112
5.5 Over and underestimates using linear approximation 115
√
5.6 Linear approximation of x 115
5.7 Idea behind Newton’s method 117
5.8 Applying Newton’s method 118
5.9 Predation rate of a ladybug eating aphids 119
5.10 Finding point of tangency 120
5.11 Two points of tangency 121
5.12 Harder tangent line problem 122
5.13 Point of tangency for y = x2 124
5.14 Approximation with tangent line for f (x) = 1/x4 124
5.15 Linear approximation for f (x) = x3 124
5.16 Generic function and tangent line 125
5.17 Intersecting tangent lines for y = x2 125
5.18 Approximating from a graph 125
6.1 Increasing and decreasing derivatives and concavity 128

6.2 Inflection point 129
6.3 f 00 (0) = 0 does not imply an inflection point 129
6.4 Critical point 132
6.5 Local maximum and local minimum 133
13
6.6 B(x) = C (x2 − x3 ): which power dominates? 135

6.7 Shape of B(x) = C (x2 − x3 ) 136
6.8 Which powers dominate y = f (x) = 8 x5 + 5 x4 − 20 x3 ? 136
6.9 A function and its derivatives 137
6.10 Finding global maximum, minimum 138
7.1 Growth rate for Logistic growth 146

7.2 Properties of a cylinder 148
7.3 Wine barrels for Kepler’s wedding 150
7.4 A cylindrical wine barrel simplifies the problem 151
7.5 Global maxima or minima can occur at endpoints 153
7.6 Optimal foraging 154
7.7 Energy gain functions for time spent foraging 155
7.8 A sketch of the average rate of energy gain 156
7.9 Box with a square base 161
7.10 Growth rate for Logistic growth 165
8.1 Function composition 167

8.2 Population of carnivores, prey, and vegetation 171
8.3 Shortest path from nest to two food sources 173
8.4 Solution to shortest path from nest to food sources 174
8.5 Most economical wooden beam 174
8.6 Paying attention to find food 175
8.7 Probabilities of finding food types 176
8.8 Probabilities of finding food types 177
9.1 Growth of a spherical tumor 183

9.2 Convergent extension of tissue in embryonic development 184
9.3 Length of spider’s thread 185
9.4 Conical cup geometry 186
9.5 Living part of a tree trunk 187
9.6 Tree trunk geometry 187
9.7 Explicit and implicit functions 188
9.8 Tangent to an implicitly defined curve 189
9.9 Circle tangent line by implicit differentiation 189
9.10 The astroid 192
9.11 A rotated ellipse 193
9.12 Shadow cast by a burning candle 199
9.13 Elliptical cell 200
9.14 The Folium of Descartes 201
9.15 Circle inside a parabola 201
10.1 Powers of 2 204

10.2 The exponential function 2x 205
10.3 The function ax 206
14
10.4 Tangent line to ex 209

10.5 Ricker Equation model 210
x
10.6 e and its inverse function 212
10.7 Agroforestry optimization 213
10.8 Allometry data on a log scale 217
10.9 Determining constants for function y = Cekt 221
11.1 Exponential growth and decay 229

11.2 Flat age distribution assumption 233
11.3 Simple assumption about fertility 233
11.4 Simple assumption about mortality 233
11.5 Projected world population 234
11.6 Doubling time 236
11.7 Half-life 239
11.8 Where differential equations come from 242
12.1 Steady state 252

12.2 Deviation from Steady State 253
12.3 Deviation from Steady State 253
12.4 Solution to the original differential equation 253
12.5 Solutions to a differential equation 254
12.6 Temperature vs. time for cooling object 255
12.7 The carbon cycle 259
12.8 Dividing time 261
12.9 True and approximate solution to Newton’s law of cooling 264
13.1 Direction field and solution curves for dydt = 2y 282

13.2 Direction field and solution curves for dydt = y − y3 283
dT
13.3 Slope field and solution curves for dt = 0.2(10 − T ) 284
13.4 Slope field related to state space 285
13.5 Where y is increasing, decreasing, or static 285
13.6 Determining flow direction 286
13.7 Qualitative sketch for a DE 286
13.8 Plotting dy/dt vs. y for scaled logistic equation 288
13.9 Slope field and solution curves for the logistic equation 289
13.10 Plot of f (I ) versus I for a DE 294
14.1 Electrical activity of the heart: a repeating pattern 303

14.2 Circle circumference, diameter and radius 304
14.3 Definition of angles in radians 304
14.4 Definitions of trigonometric functions 306
14.5 Periodicity of sine and cosine 308
14.6 Amplitude, frequency and phase of the sine function 310
14.7 Periodicity of the ECG 311
14.8 Hormonal cycles 313
15
14.9 Periodic moon phases 314

14.10 Sine and its inverse, arcsine 316
14.11 Triangle with angle θ such that sin(θ ) = x 316
14.12 Cosine and its inverse, arccosine 317
14.13 Triangle with angle θ such that cos(θ ) is x 317
14.14 Tan and its inverse, arctangent 318
14.15 Triangle with angle θ such that tan(θ ) is x 318
15.1 Law of cosines 329

15.2 Angles made by clock hands 331
15.3 Visual angle changes 332
15.4 Escape response of Zebra danio 334
15.5 Escape response geometry 334
15.6 Rate of change of visual angle versus x 335
15.7 Rate of change of visual angle 336
15.8 Reaction distance for zebra danio escape response 337
15.9 Bird flock formation 341
15.10 Cannon ball trajectory 341
15.11 Cycloid curve 342
15.12 Rotating wheel 343
15.13 Ship sailing away 343
15.14 Window best view 344
15.15 Fireworks overhead 344
16.1 A function and its tangent line 349

16.2 Race track with circular ends 352
16.3 Shape of a cell 352
16.4 Growth of a vine 353
16.5 Newton’s law of cooling 353
16.6 Least pipe installing cost 355
16.7 Least heat lost in a building 355
16.8 Ducks in a row 356
16.9 Runners on a circular track 357
16.10 Tumor mass 357
16.11 Blood vessel branching 358
16.12 The surface sin(x) cos(y) = z 359
A.1 Slope of a straight line 363
C.1 An ellipse cannot be described by a single function 367

C.2 A variety of functions 368
C.3 Shifting a function 368
C.4 Function reflections 368
C.5 Classifying functions 369
C.6 Power functions and horizontal straight line 370
16
C.7 Graphs of inverse functions 374

C.8 An inverse of the power function y = x2 374
D.1 A function with a “hole” 381
F.1 Ratios of side lengths in right triangle 391

F.2 Law of cosines 392
F.3 Summary of the trigonometric functions and their inverses 393
G.1 Lysteria, a parasite that lives inside a cell 397

G.2 Spherical bead that can swim using an actin tail 398
G.3 Distance travelled by the bead as a function of time 398
G.4 Smoothed bead trajectory 399
G.5 Bead trajectory showing tangent lines to the graph 399
G.6 Bead trajectory and velocity 399
G.7 Wrapping a rectangular box 400
G.8 Surface area of a box 400
G.9 Cylinder inside a sphere 401
G.10Rooted tangent diagram for optimal foraging 403
List of Tables
Kx
1.1 Units for Michaelis-Menten kinetics, v = kn +x . 37
1.2 Chemical reaction speed data. 50
2.1 Heating milk temperature data. 52

2.2 Cooling milk temperature data. 52
2.3 Tuna swimming data 53
2.4 Refining the data for cooling 59
2.5 Human population data 66
2.6 Average rate of change from function data 66
2.7 Table of sine and cosine values 67
3.1 Some computed derivatives. 83

3.2 Derivative data for two functions 84
4.1 The Power Rule 90

√
5.1 Linear approximation to x 116
5.2 Data for Newton’s method 118
5.3 Newton’s method applied to the aphid-ladybug problem 119
6.1 Function behaviour of a polynomial 137
8.1 Data for Newton’s method 178
9.1 Geometric relationships for related rates 183
10.1 Constant in derivative of 2x 207

10.2 Constant in derivative of 10x 208
10.3 The natural base e for the exponential ex 208
10.4 Metabolic rates of animals of various sizes 216
10.5 Human population data 222
10.6 Fitting data to a function 223
12.1 Checking if a function satisfies a differential equation 250

12.2 Checking that a function satisfies a differential equation 250
12.3 Euler’s method 263
18
12.4 Cooling milk data 269
13.1 Table for the slope field diagram of dydt = 2y 281

13.2 Table for the slope field diagram of dydt = y − y3 283
dT
13.3 Table for a slope field diagram of dt = 0.2(10 − T ) 283
13.4 Slope field for the logistic equation 289
15.1 Trigonometric function derivatives 327

15.2 Inverse trigonometric function derivatives 328
D.1 Useful limits 383
F.1 Standard sine, cosine and tangent values 391

F.2 Standard inverse trigonometric values 393
19
This book is dedicated to students who have
a genuine desire to learn about the beauty
and usefulness of mathematics and to many
colleagues who have helped to shape my
interest and philosophy of teaching life-
science calculus.
Preface
This preface describes the main philosophy of the course, and serves as a
guide to the student and to the instructor. It outlines reasons for the way that
topics are organized, and how this organization is intended to help introduce
first year students to the major concepts and applications of differential
calculus. The material for this book was collected during two decades of
teaching calculus at the University of British Columbia, and benefitted greatly
from insights and ideas of colleagues, as well as questions, interest, and
enthusiasm of students and instructors.
Introduction to this book
Calculus arose as a tool for solving practical scientific problems through the
centuries. However, it is often taught as a technical subject with rules and for-
mulas (and occasionally theorems), devoid of its connection to applications.
In this course, the applications form an important focal point, with empha-
sis on life sciences. This places the techniques and concepts into practical
context, as well as motivating quantitative approaches to biology taught to
undergraduates. While many of the examples have a biological flavour, the
level of biology needed to understand those examples is kept at a minimum.
The problems are motivated with enough detail to follow the assumptions, but
are simplified for the purpose of pedagogy.
The mathematical philosophy is as follows: Power functions,
We start with basic observations about functions and graphs, with an Polynominals,
Rational functions
emphasis on power functions and polynomials. We use elementary properties
of a function to sketch its graph and to understand its shape, even before
discussing derivatives; later we refine such graph-sketching skills. We Graph Sketching
consider useful ideas with biological implications even in this basic context.
In fact, we discuss several examples in which two processes (such as growth Application: when
and predation or nutrient uptake and consumption) are at balance. We show does growth
balance mortality
how setting up the relevant algebraic problem revels when such a balance can
exist.
We introduce the derivative in three complementary ways: (1) As a rate of
change, (2) as the slope we see when we zoom into the graph of a function,
and (3) as a computational quantity that can be approximated by a finite
22
difference. We discuss (1) by first defining an average rate of change over

a finite time interval. We use actual data to do so, but then by refining the
time interval, we show how this average rate of change approaches the
instantaneous rate, i.e. the derivative. This helps to make the idea of the limit
more intuitive, and not simply a formal calculation. We illustrate (2) using
a sequence of graphs or interactive graphs with increasing magnification.
We illustrate (3) using simple computation that can be carried out on a
spreadsheet. The actual formal definition of the derivative (while presented
and used) takes a back-seat to this discussion. Derivatives of
The next philosophical aspect of the course is that we develop all the ideas Power functions,
Polynominals
and applications of calculus using simple functions (power and polynomials)
first, before introducing the more elaborate technical calculations. The aim
Refined Graph
is to highlight the usefulness of derivatives for understanding functions Sketching
(sketching and interpreting their behaviour), and for optimization problems,
before having to grapple with the chain rule and more intricate computation Optimization
of derivatives. This helps to illustrate what calculus can achieve, and decrease methods and
applications
the focus on rote mechanical calculations.
Once this entire “tour” of calculus is complete, we introduce the chain rule
and its applications, and then the transcendental functions (exponentials and
trigonometric). Both are used to illustrate biological phenomena (population
growth and decay, then, later on, cyclic processes). Both allow a repeated Chain Rule
exposure to the basic ideas of calculus - curve sketching, optimization, and
applications to related rates. This means that the important concepts picked
Derivatives of
up earlier in the context of simpler functions can be reinforced again. At this exponentials, etc.
point, it is time to practice and apply the chain rule, and to compute more
technically involved derivatives. But, even more than that, both these topics Exponential
allow us to informally introduce a powerful new idea, that of a differential population growth
equation.
By making the link between the exponential function and the differential Differential
equation dy/dx = ky, we open the door to a host of biological applications equations
where we seek to understand how a system changes: how a population size
grow? how does the mass of a cell change as nutrients are taken up and Application: What
if growth does not
consumed? By revisiting our initial discussions, we identify the “balance balance mortality?
points” as steady states, and we develop arguments to predict what changes
with time would be observed. The idea of a deviation away from steady state
also leads us to find the behaviour of solutions to the differential equation
dy/dt = a − by. This leads to many useful applications. including the
Slope field and
temperature of a cooling object, the level of drug in the bloodstream, simple
geometric analysis
chemical reactions, and many more.
Ultimately, a first semester calculus course is all about the applications Qualitative
of a derivative. We use this fact to explore nonlinear differential equations predictions of a
differential
of the first order, using qualitative sketches of the direction field and the equation
state space of the equation. Even though some of the (integration) methods
Applications to
for solving a differential equation are developed only in a second semester population growth
and disease spread
23
calculus course, we provide here the background to understanding what

such equations are saying, and what they imply. These simple yet powerful
qualitative methods allow us to get intuition to the behaviour of more realistic
biological models, including density-dependent (logistic) growth and even
the spread of disease. Many of the ideas here are geometric, and we return to
interpreting the meaning of graphs and slopes yet again in this context.
The idea of a computational approach is introduced and practiced in
several places, as appropriate. We use simple examples to motivate linear
approximation and Newton’s method for finding zeros of a function. Later,
we use Euler’s method to solve a simple differential equation computationally.
All these methods are based on the derivative, and most introduce the idea
of an iterated (repeated) process that is ideally handled by computer or
calculator. The exposure to these computational methods, while novel and
sometimes daunting, provides an important set of examples of how properly
understanding the mathematics can be used directly for effective design of
computational algorithms.
Acknowledgements
I am particularly grateful to Prof. William Casselman (emeritus, UBC) and

Prof David Austin (Grand Valley State University, and formerly UBC) for
the initial years of development of the approach and emphasis, as well as
overall philosophy of this calculus course. Much of the plan described above
had its beginnings while we were co-instructors in the 1990’s at UBC. I am
also grateful to recent colleagues, who have provided inspiration, input, and
advice, and who have helped to shape the biological applications in the book.
These include, most especially, Prof. Eric Cytrynbaum (UBC). Other notable
influences came from Wes Maciejewski (San Jose State University) who
first suggested the idea of videos. I wish to thank Dr. Sophie Burrill for the
formatting and improved figures. Many students contributed to proofreading,
finding typos, or adding to the collection of problems. A partial list includes
Glen Pugh, Carol Shao-Shan Huang and Lu Fan. I am also grateful to Will
Engle at the Center for Teaching and Learning Technology (CTLT) at UBC
for encouragement to make this work an open resource and to the UBC
Studios for assistance and support during preparation of videos.
1
Power functions as building blocks
Like tall architectural marvels that are made of simple units (beams, bricks,
and tiles), many interesting functions can be constructed from simpler
building blocks. In this chapter, we study a family of simple functions, the
power functions - those of the form f (x) = xn . Mastered Material Check
Our first task is to understand properties of the members of this “family”. 1. Can you define function?
We will see that basic observations of power functions such as x2 , x3 leads to 2. Give an example of a polynomial
insights into a biological problem of why the size of living cells is limited. function; a rational function.
Later, we use power functions as “ building blocks” to construct polynomials,

and rational functions . We then develop important approaches to sketch the
shapes of the resulting graphs.
1.1 Power functions
Section 1.1 Learning goals
1. Interpret the shapes of power functions relative to one another.
2. Justify that power functions with low powers dominate near the origin, and
power functions with high powers dominate far away from the origin.
3. Identify the points of intersection of two power functions.
Let us consider the power functions, that is functions of the form ` Click on this link and then adjust the
slider on this interactive desmos graph
y = f ( x ) = xn , to see how the power n affects the shape
of a power function in the first quadrant.
where n is a positive integer. Power functions are among the most elementary
and “elegant” functions - we only need multiplications to compute their value
at any point. They are thus easy to calculate, very predictable and smooth,
and, from the point of view of calculus, very easy to handle.
From Figure 1.1, we see that the power functions (y = xn for powers n =
2, . . . , 5) intersect at x = 0 and x = 1. This is true for all positive integer powers.
The same figure also demonstrates another fact helpful for curve-sketching:
the greater the power n, the flatter the graph near the origin and the steeper
26 DIFFERENTIAL CALCULUS FOR THE LIFE SCIENCES
4 y
the graph beyond x > 1. This can be restated in terms of the relative size of
the power functions. We say that close to the origin, the functions with lower x5 x4
3
powers dominate, while far from the origin, the higher powers dominate.
x3
More generally, a power function has the form 2
x2
n
y = f (x ) = K · x
1
where n is a positive integer and K, sometimes called the coefficient, is x

a constant. So far, we have compared power functions whose coefficient 0.2 0.4 0.6 0.8 1 1.2 1.4
is K = 1. We can extend our discussion to a more general case as well. Figure 1.1: Graphs of a few power func-
tions y = xn . All intersect at x = 0, 1. As
Example 1.1 Find points of intersection and compare the sizes of the two the power n increases, the graphs become
flatter close to the origin, (0, 0), and steeper
power functions at large x-values.
y1 = axn , and y2 = bxm .
Mastered Material Check
where a and b are constants. You may assume that both a and b are positive.
3. Use Figure 1.1 to approximate
when x5 = 2.
Solution. This comparison is a slight generalization of the previous discus-
4. What is the first quadrant?
sion. First, we note that the coefficients a and b merely scale the vertical
behaviour (i.e. stretch the graph along the y axis). It is still true that the two
functions intersect at x = 0; further, as before, the higher the power, the flatter 100 y
the graph close to x = 0, and the steeper for large positive or negative values 2x3
80
of x. However, now another point of intersection of the graphs occur when
60 5x2
axn = bxm ⇒ xn−m = (b/a).
40
We can solve this further to obtain a solution in the first quadrant 20
x
x = (b/a)1/(n−m) . (1.1) 1 2 3 4
Figure 1.2: Graphs of two power functions,
This is shown in Figure 1.2 for the specific example of y1 = 5x2 , y2 = 2x3 . y = 5x2 and y = 2x3 .
Close to the origin, the quadratic power function has a larger value, whereas
for large x, the cubic function has larger values. The functions intersect when
5x2 = 2x3 , which holds for x = 0 or x = 52 = 2.5. ♦
If b/a is positive, then in general the value given in (1.1) is a real number.
Example 1.2 Determine points of intersection for the following pairs of

functions:
(a) y1 = 3x4 and y2 = 27x2 ,
(b) y1 = 34 πx3 and y2 = 4πx2 .

Solution.
√
(a) Intersections occur at x = 0 and at ±(27/3)1/(4−2) = ± 9 = ±3.
(b) These functions intersect at x = 0, 3 but there are no other intersections at

negative values of x. ♦
P OW E R F U N C T I O N S A S B U I L D I N G B L O C K S 27
Note that in many cases, the points of intersection are irrational numbers Mastered Material Check
whose decimal approximations can only be obtained by a scientific calculator 5. What is an irrational number?
or by some approximation method (such as Newton’s Method, studied in
Section 5.4)
With only these observations we can examine a biological problem related
to the size of cells. By applying these ideas, we can gain insight into why
cells have a size limitation, as discussed in the next section.
1.2 How big can a cell be? A model for nutrient balance
i A summary of the cell size model.
We discuss what cell size is consistent
Section 1.2 Learning goals with a balance between nutrient
absorption and consumption in a cell.
1. Describe the derivation of a mathematical model for cell nutrient absorp-
tion and consumption.
2. Use parameters (k1 , k2 ) rather than specific numbers in mathematical

expressions.
3. Demonstrate the link between power functions in Section 1.1 and cell
nutrient balance in the model.
4. Interpret the results of the model.
Consider the following biologically motivated questions:

• What physical and biological constraints determine the size of a cell?
• Why do some size limitations exist?
• Why should animals be made of millions of tiny cells, instead of a just a
few large ones?
We already have enough mathematical prowess to address these questions -

particularly if we assume a cell is spherical. Of course, this is often not the
case. The shapes of living cells uniquely suit their functions. Many have
long appendages, cylindrical parts, or branch-like structures. But here, we
neglect all these beautiful complexities and look at a simple spherical cell
because it suffices to answer our questions. Such mathematical simplifica-
tions can be very illuminating: they allow us to form a mathematical model.
A mathematical model is just a representation of a real situation which sim-
plifies things by representing the most important aspects, and neglecting or
idealizing complicating details.
In this section, we follow a reasonable set of assumptions and mathemati-
cal facts to explore how nutrient balance can affect and limit cell size.
Building the model

In order to build the model we make some simplifying assumptions and
then restate them mathematically. We base the model on the following
assumptions:
1. The cell is roughly spherical (See Figure 1.3).

2. The cell absorbs oxygen and nutrients through its surface. The larger the
r
surface area, S, the faster the total rate of absorption. We assume that the
rate at which nutrients (or oxygen) are absorbed is proportional to the
surface area of the cell.
3. The rate at which nutrients are consumed (i.e., used up) in metabolism is
proportional the volume, V , of the cell. The bigger the volume, the more
nutrients are needed to keep the cell alive. Figure 1.3: An assumed spherical cell
absorbs nutrients at a rate proportional to its
surface area S, but consumes nutrients at a
We define the following quantities for our model of a single cell: rate proportional to its volume V .

A = net rate of absorption of nutrients per unit time, 6. What does “A is proportional to B”
mean?
C = net rate of consumption of nutrients per unit time,
7. What might the units be for
V = cell volume, quantities A,C,V , S and r?
S = cell surface area, 8. Given your choices for 7., what are
the units associated with k1 , k2 ?
r = radius of the cell.
We now rephrase the assumptions mathematically. By Assumption 2, the

absorption rate, A, is proportional to S: this means that
A = k1 S,
where k1 is a constant of proportionality. Since absorption and surface area

are positive quantities, only positive values of the proportionality constant
make sense, so k1 must be positive. The value of this constant depends on
properties of the cell membrane such as its permeability or how many pores it
contains to permit passage of nutrients. By using a generic constant - called
a parameter - to represent this proportionality constant, we keep the model
general enough to apply to many different cell types.
By Assumption 3, the rate of nutrient consumption, C is proportional to V ,
so that
C = k2V ,
where k2 > 0 is a second (positive) proportionality constant. The value of k2

depends on the cell metabolism, i.e. how quickly it consumes nutrients in
carrying out its activities.
By Assumption 1, the cell is spherical, thus its surface area, S, and volume,
V , are:
4
S = 4πr2 , V = πr3 . (1.2)
3
Putting these facts together leads to the following relationships between

nutrient absorption A, consumption C, and cell radius r:
A = k1 (4πr2 ) = (4πk1 )r2 ,

4 3 4
C = k2 πr = πk2 r3 .
3 3
Rewriting this relationship as

2 4
A(r ) = (4πk1 )r , and C (r ) = πk2 r3 . (1.3)
3
we observe that A,C are simply power functions of the cell radius, r, that is Mastered Material Check
9. What are constants a and c in terms
A(r ) = ar2 , C (r ) = cr3 . of k1 and k2 ?
10. Why are we considering different
Note: the powers are n = 3 for consumption and n = 2 for absorption. values of r in Example 1.3?
The discussion of power functions in Section 1.1 now contributes to our

analysis of how nutrient balance depends on cell size.
Nutrient balance depends on cell size

In our discussion of cell size, we found two power functions that depend on
the cell radius r, namely the nutrient absorption A(r ) and consumption C (r )
given in Eqns. (1.3). We first ask whether absorption or consumption of
nutrients dominates for small, medium, or large cells.
Example 1.3 (A fine balance) For what cell size is the consumption rate
exactly balanced by the absorption rate? Which rate (consumption or absorp-
tion) dominates for small cells? For large cells?
Solution.
The two rates “balance” (and their graphs intersect) when

4
A(r ) = C (r ) ⇒ πk2 r3 = (4πk1 )r2 .
3
A trivial solution to this equation is r = 0.

Note: this solution is not interesting biologically, but we should not forget it
in mathematical analysis of such problems.
If r 6= 0, then, canceling a factor of r2 from both sides gives:
k1
r=3 .
k2
This means absorption and consumption rates are equal for cells of this
size. For small r, the power function with the smaller power of r (namely
A(r )) dominates, but for very large values of r, the power function with the
higher power of r (namely C (r )) dominates. It follows that for smaller cells,

absorption A ≈ r2 is the dominant process, while for larger cells, consumption
rate C ≈ r3 dominates. We conclude that cells larger than the critical size
r = 3k1 /k2 are unable to keep up with the nutrient demand, and cannot
survive since consumption overtakes absorption of nutrients. ♦
Using the above simple geometric argument, we deduced that cell size
has strong implications on its ability to absorb nutrients or oxygen quickly
enough to feed itself. For these reasons, cells larger than some maximal size
(roughly 1mm in diameter) rarely occur.
A similar strategy also allows us to consider the energy balance and
sustainability of life on Earth - as seen next, in Section 1.3.
1.3 Sustainability and energy balance on Earth
1. Justify the given mathematical model that describes the energy input and
output on Planet Earth.
2. Use the given model to determine the energy equlibrium of the planet.
The sustainability of life on Planet Earth depends on a fine balance be-

tween the temperature of its oceans and land masses and the ability of life
forms to tolerate climate change. As a follow-up to our model for nutrient
balance, we introduce a simple energy balance model to track incoming and
outgoing energy and determine a rough estimate for the Earth’s temperature.
We use the following basic assumptions:
1. Energy input from the sun, given the Earth’s radius r, can be approximated
as Mastered Material Check
Ein = (1 − a)Sπr2 , (1.4) 11. Do you think Ein is proportional to
Earth’s surface area or volume?
where S is incoming radiation energy per unit area (also called the solar
constant) and 0 ≤ a ≤ 1 is the fraction of that energy reflected; a is
also called the albedo, and depends on cloud cover, and other planet
characteristics (such as percent forest, snow, desert, and ocean).
2. Energy lost from Earth due to radiation into space depends on the current
temperature of the Earth T , and is approximated as
Eout = 4πr2 εσ T 4 , (1.5)
where ε is the emissivity of the Earth’s atmosphere, which represents the

Earth’s tendency to emit radiation energy. This constant depends on cloud
cover, water vapour, as well as on greenhouse gas concentration in the
atmosphere; σ is a physical constant (the Stephan-Bolzmann constant)
which is fixed for the purpose of our discussion.
Notice there are several different symbols in Eqns. (1.4) and (1.5). Being
clear about which are constants and which are variables is critical to using
any mathematical model. As the next example points out, sometimes you
have a choice to make.
Example 1.4 (Energy expressions are power functions) Explain in what

sense the two forms of energy above can be viewed as power functions, and
what types of power functions they represent.
Solution. Both Ein and Eout depend on Earth’s radius as the power ∼ r2 .
However, since this radius is a constant, it is not fruitful to consider it as
an interesting variable for this problem (unlike the cell size example in
Section 1.2). However, we note that Eout depends on temperatureas ∼ T 4 . (We
might also select the albedo as a variable and in that case, we note that Ein
depends linearly on the albedo a.) ♦
Example 1.5 (Energy equilibrium for the Earth) Explain how the assump-
tions above can be used to determine the equilibrium temperature of the
Earth, that is, the temperature at which the incoming and outgoing radiation
energies are balanced.
Solution. The Earth is at equilibrium when
Ein = Eout ⇒ (1 − a)Sπr2 = 4πr2 εσ T 4 .
♦
We observe that the factors πr2cancel, and we obtain an equation that
can be solved for the temperature T . It is instructive to examine how this
temperature depends on the constants in the problem, and how it is affected
by cloud cover and greenhouse gas level. This is also explored in Exercise 21
1.4 First steps in graph sketching
1. Identify even and odd functions based on their graph.
2. Determine algebraically whether a function is even or odd.
3. Sketch the graph of a simple polynomial of the form y = axn + bxm .
4. Sketch a rational function such as y = Axn /(b + xm ).
Even and odd power functions

` Adjust the slider to see how the even
So far, we have considered power functions y = xn with x > 0. But in mathe- and odd power functions behave as their
matical generality, there is no reason to restrict the independent variable x to power increases.
Even power functions Odd power functions Figure 1.4: Graphs of power functions.
2.5 (a) A few even power functions: y = x2 ,
y
2 y = x4 and y = x6 . (b) Some odd power
x6 x4 x5 x3
2 functions: y = x, y = x3 and y = x5 . Note the
1 x symmetry properties.
1.5 2
x
−1.5 −1 −0.5 0.5 1 1.5
1
−1
0.5
−2
x
−1.5 −1 −0.5 0.5 1 1.5
(a) (b)
positive values. Thus we expand the discussion to consider all real values of x.
We examine now some symmetry properties that arise. Mastered Material Check
In Figure 1.4 (a) we see that power functions with an even power, such 12. Highlight the y-axes and circle the
as y = x2 , y = x4 , and y = x6 , are symmetric about the y-axis. In Figure 1.4(b) origins in Fig 1.4.
we notice that power functions with an odd power, such as y = x, y = x3 13. Consider Figure 1.4: where do even
power functions intersect? Odd?
and y = x5 are symmetric when rotated 180◦ about the origin. We adopt the
14. Show that f (a) = a5 − 3a is an odd
term even function and odd function to describe such symmetry properties. function.
More formally, 15. Give an example of a function
which is bounded.
f (−x) = f (x) ⇒ f is an even function, 16. Verify y = x2 is not one-to-one.
f (−x) = − f (x) ⇒ f is an odd function 17. What graphical property do
one-to-one functions share?
Many functions are not symmetric at all, and are neither even nor odd. See
Appendix C for further details.
Example 1.6 Show that the function y = g(x) = x2 − 3x4 is an even function
Solution. For g to be an even function, it should satisfy g(−x) = g(x). Let us

calculate g(−x) and see if this requirement holds. We find that
g(−x) = (−x)2 − 3(−x)4 = x2 − 3x4 = g(x).
Here we have used the fact that (−x)n = (−1)n xn , and that when n is even,
(−1)n = 1. ♦
All power functions are continuous and unbounded: for x → ∞ both
even and odd power functions satisfy y = xn → ∞. For x → −∞, odd power
functions tend to −∞. Odd power functions are one-to-one: that is, each
value of y is obtained from a unique value of x and vice versa. This is not true
for even power functions. From Fig 1.4 we see that all power functions go
through the point (0, 0). Even power functions have a local minimum at the
origin whereas odd power functions do not.
Definition 1.1 (Local Minimum) A local minimum of a function f (x) is

a point xmin such that the value of f is larger at all sufficiently close points.
Formally, f (xmin ± ε ) > f (xmin ) for ε small enough.
Sketching a simple (two-term) polynomial

Based on our familiarity with power functions, we now discuss functions
made up of such components. In particular, we extend the discussion to
polynomials (sums of power functions) and rational functions (ratios
of such functions). We also develop skills in sketching graphs of these
functions.
Example 1.7 (Sketching a simple cubic polynomial) Sketch a graph of the

polynomial ` Adjust the slider to see how positive
y = p(x) = x3 + ax. (1.6) and negative values of the coefficient a
affect the shape of this simple
How would the sketch change if the constant a changes from positive to polynomial.
negative?
Solution. The polynomial in Eqn. (1.6) has two terms, each one a power
function. Let us consider their effects individually. Near the origin, for x ≈ 0
the term ax dominates so that, close to x = 0, the function behaves as Mastered Material Check
17. Justify why the linear term
y ≈ ax. dominates near the origin, while the
cubic term dominates further out.
This is a straight line with slope a. Hence, near the origin, if a > 0 we would 18. Sketch the graph of any function
see a line with positive slope, whereas if a < 0 the slope of the line should be with horizontal asymptote y = 2.
negative. Far away from the origin, the cubic term dominates, so
y ≈ x3
at large (positive or negative) x values. Figure 1.5 illustrates these ideas.

In the first row we see the behaviour of y = p(x) = x3 + ax for large x,
in the second for small x. The last row shows the graph for an intermediate
range. We might notice that for a < 0, the graph has a local minimum as well
as a local maximum. Such an argument already leads to a fairly reasonable
sketch of the function in Eqn. (1.6). We can add further details using algebra
to find zeros - that is where y = p(x) = 0. ♦
Example 1.8 (Zeros) Find the places at which the polynomial Eqn. (1.6)
crosses the x axis, that is, find the zeros of the function y = x3 + ax.
Solution. The zeros of the polynomial can be found by setting
y = p(x ) = 0 ⇒ x3 + ax = 0 ⇒ x3 = −ax.
The above equation always has a solution x = 0, but if x 6= 0, we can cancel

and obtain
x2 = −a.
y y y
x x x
y y y
x x x
y y y
x x x
a<0 a=0 a>0
This would have no solutions if a is a positive number, so that in that case, Figure 1.5: The graph of the polynomial
y = p(x) = x3 + ax can be obtained
the graph crosses the x axis only once, at x = 0, as shown in Figure 1.5. If a by combining its two power function
is negative, then the minus signs cancel, so the equation can be written in the components. The cubic “arms” y ≈ x3 (top
row) dominate for large x (far from the
form
origin), while the linear part y ≈ ax (middle
x2 = |a| row) dominates near the origin. When these
are smoothly connected (bottom row) we
and we would have two new zeros at obtain a sketch of the desired polynomial.
p Shown here are three possibilities, for
x = ± |a|. a < 0, a = 0, a > 0, left to right. The value
of a determines the slope of the curve near
For example, if a = −1 then the function y = x3 − x has zeros at x = 0, 1, −1. x = 0 and thus also affects presence of a local
maximum and minimum (for a < 0).
♦
Example 1.9 (A more general case) Explain how you would use the ideas 19. Find the zeros of y = x3 + 3x.
of Example 1.7 to sketch the polynomial y = p(x) = axn + bxm . Without loss of
generality, you may assume that n > m ≥ 1 are integers.
Solution. As in Example 1.7, this polynomial has two terms that dominate
at different ranges of the independent variable. Close to the origin, y ≈ bxm
(since m is the lower power) whereas for large x, y ≈ axn . The full behaviour is
obtained by smoothly connecting these pieces of the graph. Finding zeros can
refine the graph. ♦
A step back. The reasoning used here is an important first step in sketching
the graph of a polynomial. In the ensuing chapters, we develop specialized
methods to find zeros of more complicated functions (using an approximation
technique called Newton’s method). We also apply calculus tools to deter-
mine points at which the function attains local maxima or minima (called
critical points), and how it behaves for very large positive or negative values
of x. That said, the elementary steps described here remain useful as a quick
approach for visualizing the overall shape of a graph.
Sketching a simple rational function

We apply similar reasoning to consider the graphs of simple rational func-
tions. A rational function is a function that can be written as
p1 (x)
y= , where p1 (x) and p2 (x) are polynomials.
p2 (x)
Example 1.10 (A rational function) Sketch the graph of the rational func-
tion ` Adjust the sliders to see how the
Axn values of n, A, and a affect the shape of
y= n , x ≥ 0. (1.7)
a + xn the rational function in (1.7).
What properties of your sketch depend on the power n? What would the
graph look like for n = 1, 2, 3?
Solution. We can break up the process of sketching this function into the
following steps:
• The graph of the function in Eqn. (1.7) goes through the origin (at x = 0,
we see that y = 0).
• For very small x, (i.e., x << a) we can approximate the denominator by the
constant term an + xn ≈ an , since xn is negligible by comparison, so that
Axn Axn

A
y= n ≈ = xn for small x.
a + xn an an
This means that near the origin, the graph looks like a power function, Mastered Material Check
y ≈ Cxn (where C = A/an ). 20. Why is an a constant?

21. Sketch the graph of any function
• For large x, i.e. x >> a, we have an + xn ≈ xn since x overtakes and with horizontal asymptote y = 2.
dominates over the constant a, so that
Axn Axn
y= ≈ = A for large x.
an + x n xn
This reveals that the graph has a horizontal asymptote y = A at large values
of x.
• Since the function behaves like a simple power function close to the origin,
we conclude directly that the higher the value of n, the flatter is its graph
near 0. Further, large n means sharper rise to the eventual asymptote.
The results are displayed in Figure 1.6. ♦
Small x Large x Smoothly connected

y y y n=3
n=1 n=2
n=2
n=1
n=3
A
x x x
Figure 1.6: The rational functions Eqns.(1.7)

with n = 1, 2, 3 are compared on this graph.
Close to the origin, the function behaves
1.5 Rate of an enzyme-catalyzed reaction like a power function, whereas for large x
there is a horizontal asymptote at y = A. As
n increases, the graph becomes flatter close
to the origin, and steeper in its rise to the
asymptote.
1. Describe the connection between Michaelis-Menten kinetics in biochem-

istry and rational functions described in Section 1.4.
2. Interpret properties of a graph such as Figure 1.8 in terms of properties of

an enzyme-catalyzed reactions.
Rational functions introduced in Example 1.10 often play a role in biochem-

istry. Here we discuss two such examples and the contexts in which they
appear. In both cases, we consider the initial rise of the function as well as its
eventual saturation.
Saturation and Michaelis-Menten kinetics

Biochemical reactions are often based on the action of proteins known as
enzymes that catalyze reactions in living cells. Fig. 1.7 depicts an enzyme
E binding to its substrate S to form a complex C. The complex breaks apart
into a product, P, and the original enzyme that can act once more. Substrate
is usually plentiful relative to the enzyme.
In the context of this example, x represents the concentration of substrate
in the reaction mixture. The speed of the reaction, v, (namely the rate at
which product is formed) depends on x. When you actually graph the speed
of the reaction as a function of the concentration, you see that it is not lin-
ear: Figure 1.8 is typical. This relationships, known as Michaelis-Menten
Figure 1.7: An enzyme (catalytic protein)

is shown binding to a substrate molecule
(circular dot) and then processing it into a
k1 k2 product (star shaped molecule).
k−1
?
E S C E P
kinetics, has the mathematical form v Michaelis-Menten kinetics
Kx K saturation
speed of reaction = v = , (1.8)
kn + x
where K, kn > 0 are constants specific to the enzyme and the experimental K/2
conditions.
Equation (1.8) is a rational function. Since x is a concentration, it must
initial rise x
be a positive quantity, so we restrict attention to x ≥ 0. The expression in kn
Eqn. (1.8) is a special case of the rational functions explored in Example 1.10, Figure 1.8: The graph of reaction speed,
where n = 1, A = K, a = kn . In Figure 1.8, we used plot this function for v, versus substrate concentration, x in an
enzyme-catalyzed reaction, as in Eqn. 1.8.
specific values of K, kn . The following observations can be made
This behaviour is called Michaelis-Menten
kinetics. Note that the graph at first rises
1. The graph of Eqn. (1.8) goes through the origin. Indeed, when x = 0 we almost like a straight line, but then it curves
and approaches a horizontal asymptote. This
have v = 0. graph tells us that the speed of the enzyme
cannot exceed some fixed level, i.e. it cannot
2. Close to the origin, the initial rise of the graph “looks like” a straight line. be faster than K.
We can see this by considering values of x that are much smaller than kn .
Then the denominator (kn + x) is well approximated by the constant kn .
Thus, for small x, v ≈ (K/kn )x, so that the graph resembles a straight line
through the origin with slope (K/kn ).
3. For large x, there is a horizontal asymptote. A similar argument for x kn ,

verifies that v is approximately constant at large enough x.
Michaelis-Menten kinetics represents one relationship in which satura-

tion occurs: the speed of the reaction at first increases as substrate concentra-
tion x is raised, but the enzymes saturate and operate at a fixed constant speed
K as more and more substrate is added.
units example Table 1.1: Units for Michaelis-Menten

kinetics, v = knKx
+x .
x concentration “nano Molar” , nM ≡ 10−9 Moles per litre
v concentration over time nM min−1
kn
K

Units. It is worth considering the units in Eqn. (1.8). Given that only quanti-
22. Complete Table 1.1.
ties with identical units can be added or compared, and that the units of the
two sides of the relationship must balance, fill Table 1.1.
Featured Problem 1.1 (Fish population growth 1) The Beverton-Holt

model relates the number of salmon in a population this year N1 to the num-
ber of salmon that were present last year N0 , according to the relationship
N0
N1 = k1 , k1 , k2 > 0 (1.9)
(1 + k2 N0 )
Sketch N1 as a function of N0 and explain how the constants k1 and k2 affect
the shape of the graph you obtain. Is there a population level N0 that would
be exactly the same from one year to the next? Are there any restrictions on
k1 or k2 for this kind of static (“steady state”) population to be possible?
Hill functions
The Michaelis-Menten kinetics we discussed above fit into a broader class of
Hill functions, which are rational functions of the form shown in Eqn. (1.7)
with n > 1 and A, a > 0. This function is often referred to in the life sciences
as a Hill function with coefficient n, (although the “coefficient” is actually a
power in the terminology used in this chapter). Hill functions occur when
an enzyme-catalyzed reaction benefits from cooperativity of a multi-step
process. For example, the binding of the first substrate molecule may enhance
the binding of a second. Hill function kinetics
Michaelis-Menten kinetics coincides with a Hill function for n = 1. In n=3
biochemistry, expressions of the form of Eqn. (1.7) with n > 1 are often n=2
denoted “sigmoidal” kinetics. Several such functions are plotted in Figure 1.9. n=1
We examined the shapes of these functions in Example 1.10.

All Hill functions have a horizontal asymptote y = A at large values of x.
If y is the speed of a chemical reaction (analogous to the variable we called
v), then A is the “maximal rate” or “maximal speed” of the reaction. Since
the Hill function behaves like a simple power function close to the origin,
the higher the value of n, the flatter is its graph near 0, and the sharper the chemical concentration, x
rise to the eventual asymptote. Hill functions with large n are often used to Figure 1.9: Hill function kinetics, from
Eqn. (1.7), with A = 3, a = 1 and Hill
represent “switch-like” behaviour in genetic networks or biochemical signal coefficient n = 1, 2, 3. See also Fig 1.6 for an
transduction pathways. analysis of the shape of this graph.
The constant a is sometimes called the “half-maximal activation level” for
the following reason: when x = a then
Aan Aa2 A
v= = 2
= .
an + an 2a 2
This shows that the level x = a leads to a reaction speed of A/2 which is half
of the maximal possible rate.
Featured Problem 1.2 Lineweaver-Burk plots. Hill functions can be

transformed to a linear relationship through a change of variables. Consider
the Hill function
Ax3
y= 3 .
a + x3
define y = 1/Y , X = 1/x3 . Show that Y and X satisfy a linear relationship.
Because we take the reciprocals of x and y, X and Y are sometimes called
reciprocal coordinates.
1.6 Predator Response
Interactions of predators and prey are often studied in ecology. Professor Holling Predator response
C.S. (“Buzz”) Holling, (a former Director of the Institute of Animal Resource Type I
predation rate, P(x)

Ecology at the University of British Columbia) described three types of Type II
predators, termed “Type I”, “Type II” and “Type III”, according to their
ability to consume prey as the prey density increases. The three Holling Type III
“predator functional responses” are shown in Fig. 1.10.
Quick Concept Checks

prey density, x
1. Match the predator responses shown in Fig. 1.10 with the descriptions
Figure 1.10: Holling’s Type I, II, and III
given below predator response. The predation rate P(x)
is the number of prey eaten by a predator
1. As a predator, I get satiated and cannot keep eating more and more per unit time. Note that the predation rate
depends on the prey density x.
prey.
2. I can hardly find the prey when the prey density is low, but I also get
satiated at high prey density.
3. The more prey there is, the more I can eat.
Based on Fig. 1.10, match the predator responses to functions shown below.
Hint: One of the curves “looks
P1 (x) = kx, like a straight line" (so which function
x here is linear?). One of the choices is
P2 (x) = K , a power function. (Will it fit any of the
a+x other curves? why or why not?). Now
P3 (x) = Kxn , n ≥ 2 consider the saturating curves and use
our description of rational functions
xn in Section 1.5 to select appropriate
P4 (x) = K , n≥2
an + xn formulae for these functions.
The generality of mathematics allows us to adapt concepts we studied in

one setting (enzyme biochemistry) to an apparently new topic (behaviour of
predators). i See this short video explanation of
the ladybug Type III predator response
to its aphid prey.
1.6.1 A ladybug eating aphids
Here we use ideas developed so far to address a problem in population
growth and biological control.
Featured Problem 1.3 (A balance of predation and aphid population growth)

Ladybugs are predators that love to eat aphids (their prey). Fig. 1.11 provides
data 1 that supports the idea that ladybugs are type 3 predators. 1
MP Hassell, JH Lawton, and JR Bed-
Let x= the number of aphids in some unit area (i.e., the density of the dington. Sigmoid functional responses by
invertebrate predators and parasitoids. The
prey). Then the number of aphids eaten by a ladybug per unit time in that unit Journal of Animal Ecology, 46(1):249–262,
area will be called the predation rate and denoted P(x). The predation rate 1977
usually depends on the prey density, and we approximate that dependence by
30
xn
P(x ) = K , where K, a > 0. (1.10)
Predation rate, P(x)

a + xn
n
20
Here we consider the case that n = 2. The aphids reproduce at a rate pro-
portional to their number, so that the growth rate of the aphid population G
10
(number of new aphids per hour) is
0
G(x) = rx where r > 0. (1.11) 0 20 40 60 80 100
Aphid density, x
(a) For what aphid population density x does the predation rate exactly Figure 1.11: The predation rate of a ladybug
depends on its aphid (prey) density.
balance the aphid population growth rate?
(b) Are there situations where the predation rate cannot match the growth
rate? Explain your results in terms of the constants K, a, r.
Hints and partial solution

` Use the sliders to manipulate the
predation constants K, a and the aphid
(a) The wording “the predation rate exactly balances the reproduction rate” growth rate parameter r. How many
solutions are there to P(x) = G(x)?
means that the two functions P(x) and G(x) are exactly equal. Hence, to Show that for some parameter values,
solve this problem, equate P(x) = G(x) and determine the value of x (i.e., there is only a trivial solution at x = 0.
the number of aphids) at which this equality holds. You will find that one Make a connection between this
observation and part (b) of Example 1.3.
solution to this equation is x = 0. But if x 6= 0, you can cancel one factor of
x from both sides and rearrange the equation to obtain a quadratic equation
whose solution can be written down (in terms of the positive constants
K, r, a).
Hint: Recall that a quadratic
(b) The solution you find in (a) is only a real number (i.e. a real solution equation ax√2 + bx + c = 0 has roots
−b ± b2 − 4ac
exists) if the discriminant (quantity inside the square-root) is positive. x= .
2a
Determine when this situation can occur and interpret your answer in These roots are real provided
b2 − 4ac ≥ 0.
terms of the aphid and ladybugs.
The solution to this problem is based on solving a quadratic equation, and

so, relies on the fact that we chose the value n = 2 in the predation rate. What
happens if n > 2? How do we solve the same kind of problem if n = 3, 4 etc?
We return to this issue, and develop an approximate technique (Newton’s
method) in a later chapter.
1.7 Summary
1. Functions of the form f (x) = K · xn (n a positive integer) are called power

functions with coefficient K.
2. Power functions with larger powers of n form graphs that are flatter near
the origin and steeper for x > 1.
3. An even function satisfies f (−x) = f (x); an odd function satisfies f (−x) =
− f (x). Identifying even and odd functions can aid in graph sketching.
4. The zeros of a function f (x) are roots of the equation f (x) = 0. Identifying
the root(s) of a function helps in sketching its graph.
5. Polynomials are sums of power functions. Rational functions are ratios of
polynomials. By examining the behaviour of terms that dominate near and
far from the origin, we can obtain a rough sketch of such functions.
6. Mathematical models can be used to describe scientific phenomenon. Mak-
ing reasonable assumptions and observations are necessary for building
a successful model. Translating these assumptions and observations into
mathematics is the key.
7. Hill functions can be transformed into a linear relationship using a change
of variables; the plots that result are called Lineweaver-Burk plots.
8. The mathematical models explored in this chapter concerned:
(a) cell size, based on nutrient balance;

(b) energy balance on Earth;
(c) biochemical reactions and Michaelis-Menten kinetics; and
(d) enzyme-catalyzed reactions and Hill functions.
9. Units, while often suppressed in math texts, can be immensely useful in

solving application problems. Only quantities with identical units can be
added, or compared. Two sides of an equation must have identical units.
1. When is x2 > x10 ?

(a) never (b) always (c) for small x (d) for large x
2. Why do we make assumptions when we build mathematical models?
3. Complete the sketch of the following graph, given that it is

4 y
x
−4 −2 2 4
−2
−4
(a) an even function

(b) an odd function
4. What is the relationship between Michaelis-Menten kinetics and Hill functions?

Exercises
1.1. Power functions. Consider the power function
y = axn , −∞ < x < ∞.
Explain, possibly using a sketch, how the shape of the function

changes when the coefficient a increases or decreases (for fixed n).
How is this change in shape different from the shape change that
results from changing the power n?
1.2. Transformations. Consider the graphs of the simple functions y =
x, y = x2 , and y = x3 . Describe what happens to each of these graphs
when the functions are transformed as follows:
(a) y = Ax, y = Ax2 , and y = Ax3 where A > 1 is some constant.
(b) y = x + a, y = x2 + a, and y = x3 + a where a > 0 is some constant.
(c) y = (x − b)2 , and y = (x − b)3 where b > 0 is some constant.
1.3. Sketching transformations. Sketch the graphs of the following
functions:
(a) y = x2 ,
(b) y = (x + 4)2 ,
(c) y = a(x − b)2 + c for the case a > 0, b > 0, c > 0.
(d) Comment on the effects of the constants a, b, c on the properties of
the graph of y = a(x − b)2 + c.
1.4. Sketching polynomials. Use arguments from Section 1.4 to sketch
graphs of the following polynomials:
(a) y = 2x5 − 3x2 ,
(b) y = x3 − 4x5 .
1.5. Finding points of intersection.
(a) Consider the two functions f (x) = 3x2 and g(x) = 2x5 . Find all
points of intersection of these functions.
(b) Repeat for functions f (x) = x3 and g(x) = 4x5 .
Note: finding these points of intersection is equivalent to calculating
the zeros of the functions in Exercise 4.
1.6. Qualitative sketching skills.
(a) Sketch the graph of the function y = ρx − x5 for positive and
negative values of the constant ρ. Comment on behaviour close to
zero and far away from zero.
(b) What are the zeros of this function and how does this depend on ρ?
(c) For what values of ρ would you expect that this function would
have a local maximum (“peak”) and a local minimum (“valley”)?
1.7. Finding points of intersection. Consider functions f (x) = Axn
and g(x) = Bxm . Suppose m > n > 1 are integers, and A, B > 0.
Determine the values of x at which the the functions are the same - i.e.
they intersect. Are there two places of intersection or three? How does
this depend on the integer m − n?
Note: The point (0, 0) is always an intersection point. Thus, we are
asking: when is there only one more and when there are two more
intersection points? See Exercise 5 for an example of both types.
1.8. More intersection points. Find the intersection of each pair of func-
tions.
√
(a) y = x, y = x2 ,
√
(b) y = − x, y = x2 ,
x2
(c) y = x2 − 1, 4 + y2 = 1.
1.9. Crossing the x-axis. Answer the following by solving for x in each
case. Find all values of x for which the following functions cross
the x-axis (equivalently: the zeros of the function, or roots of the
equation f (x) = 0.)
(a) f (x) = I − γx, where I, γ are positive constants.
(b) f (x) = I − γx + εx2 , where I, γ, ε are positive constants. Are there
cases where this function does not cross the x axis?
(c) In the case where the root(s) exist in part (b), are they positive,
negative or of mixed signs?
1.10. Crossing the x-axis. Answer Exercise 9 by sketching a rough graph
of each of the functions in parts (a-b) and using these sketches to
determine how many real roots there are and where they are located
(positive vs. negative x-axis).
Note: this exercise provides qualitative analysis skills that are helpful
in later applications.
1.11. Power functions. Consider the functions y = xn , y = x1/n , y = x−n ,
where n is an integer n = 1, 2, . . .).
(a) Which of these functions increases most steeply for values of x
greater than 1?
(b) Which decreases for large values of x?
(c) Which functions are not defined for negative x values?
(d) Compare the values of these functions for 0 < x < 1.
(e) Which of these functions are not defined at x = 0?
1.12. Roots of a quadratic. Find the range m such that the equation x2 −
2x − m = 0 has two unequal roots.
1.13. Rational Functions. Describe the shape of the graph of the func-
tion y = Axn /(b + xm ) in two cases:
(a) n > m and
(b) m > n.
1.14. Power functions with negative powers. Consider the function
A
f (x ) =
xa
where A > 0, a > 1, with a an integer. This is the same as the func-
tion f (x) = Ax−a , which is a power function with a negative power.
(a) Sketch a rough graph of this function for x > 0.
(b) How does the function change if A is increased?
(c) How does the function change if a is increased?
1.15. Intersections of functions with negative powers. Consider two
functions of the form
A B
f (x ) = , g(x ) = b .
xa x
Suppose that A, B > 0, a, b > 1 and that A > B. Determine where these
functions intersect for positive x values.
1.16. Zeros of polynomials. Find all real zeros of the following polynomi-
als:
(a) x3 − 2x2 − 3x,
(b) x5 − 1,
(c) 3x2 + 5x − 2.
(d) Find the points of intersection of the functions y = x3 + x2 − 2x + 1
and y = x3 .
1.17. Inverse functions. The functions y = x3 and y = x1/3 are inverse
functions (see Section 10.3 for a discussion of inverse functions).
(a) Sketch both functions on the same graph for −2 < x < 2 showing
clearly where they intersect.
(b) The tangent line to the curve y = x3 at the point (1, 1) has slope m =
3, whereas the tangent line to y = x1/3 at the point (1, 1) has
slope m = 1/3. Explain the relationship of the two slopes.
1.18. Properties of a cube. The volume V and surface area S of a cube
whose sides have length a are given by the formulae
V = a3 , S = 6a2 .
Note that these relationships are expressed in terms of power functions.

The independent variable is a, not x. We say that “V is a function of a”
(and also “S is a function of a”).
(a) Sketch V as a function of a and S as a function of a on the same set

of axes. Which one grows faster as a increases?
(b) What is the ratio of the volume to the surface area; that is, what
is VS in terms of a? Sketch a graph of VS as a function of a.
(c) The formulae above tell us the volume and the area of a cube of a
given side length. Suppose we are given either the volume or the
surface area and asked to find the side.
(i) Find the length of the side as a function of the volume (i.e.
express a in terms of V ).
(ii) Find the side as a function of the surface area.
(iii) Use your results to find the side of a cubic tank whose volume
is 1 litre. Units.
Note that 1 litre = 103 cm3 .
(iv) Find the side of a cubic tank whose surface area is 10 cm2 .
1.19. Properties of a sphere. The volume V and surface area S of a sphere
of radius r are given by the formulae
4π 3
V= r , S = 4πr2 .
3
Note that these relationships are expressed in terms of power functions
with constant multiples such as 4π. The independent variable is r,
not x. We say that “V is a function of r” (and also “S is a function
of r”).
(a) Sketch V as a function of r and S as a function of r on the same set
of axes. Which one grows faster as r increases?
(b) What is the ratio of the volume to the surface area; that is, what
is VS in terms of r? Sketch a graph of VS as a function of r.
(c) The formulae above tell us the volume and the area of a sphere of
a given radius. But suppose we are given either the volume or the
surface area and asked to find the radius.
(i) Find the radius as a function of the volume (i.e. express r in
terms of V ).
(i) Find the radius as a function of the surface area.
(i) Use your results to find the radius of a balloon whose volume
is 1 litre.
(i) Find the radius of a balloon whose surface area is 10 cm2
1.20. The size of cell. Consider a cell in the shape of a thin cylinder
(length L and radius r). Assume that the cell absorbs nutrient through
its surface at rate k1 S and consumes nutrients at rate k2V where S,V
Units.
are the surface area and volume of the cylinder. Here we assume that
Note that µM is 10−6 moles and µm is
k1 = 12µM µm−2 per min and k2 = 2µM µm−3 per min. 10−6 meters.
(a) Use the fact that a cylinder (without end-caps) has surface area S =
2πrL and volume V = πr2 L to determine the cell radius such that
the rate of consumption exactly balances the rate of absorption.
(b) What do you expect happens to cells with a bigger or smaller
radius?
(c) How does the length of the cylinder affect this nutrient balance?
1.21. Energy equilibrium for Earth. This exercise focuses on Earth’s
temperature, climate change, and sustainability.
(a) Complete the calculation for Example 1.5 by solving for the tem-
perature T of the Earth at which incoming and outgoing radiation
energies balance.
(b) Assume that greenhouse gases decrease the emissivity ε of the
Earth’s atmosphere. Explain how this would affect the Earth’s
temperature.
(c) Explain how the size of the Earth affects its energy balance accord-
ing to the model.
(d) Explain how the albedo a affects the Earth’s temperature.
1.22. Allometric relationship. Properties of animals are often related to
their physical size or mass. For example, the metabolic rate of the
animal (R), and its pulse rate (P) may be related to its body mass m
by the approximate formulae R = Amb and P = Cmd , where A,C, b, d
are positive constants. Such relationships are known as allometric
relationships.
(a) Use these formulae to derive a relationship between the metabolic
rate and the pulse rate (hint: eliminate m).
(b) A similar process can be used to relate the Volume V = (4/3)πr3
and surface area S = 4πr2 of a sphere to one another. Eliminate r
to find the corresponding relationship between volume and surface
area for a sphere.
1.23. Rate of a very simple chemical reaction. We consider a chemical re-
action that does not saturate, and a simple linear relationship between
reaction speed and reactant concentration.
A chemical is being added to a mixture and is used up in a reaction.
The rate of change of the chemical, (also called “the rate of the reac-
tion”) v M/sec is observed to follow a relationship Units.
Note that M stands for Molar, which is
v = a − bc the number of moles per litre.
where c is the reactant concentration (in units of M) and a, b are

positive constants.
Note: v is considered to be a function of c, and moreover, the relation-
ship between v and c is assumed to be linear.
(a) What units should a and b have to make this equation consistent?
v
Note: in an equation such as v = a − bc, each of the three terms
must have the same units. Otherwise, the equation would not make Reaction rate
sense.
slope −0.2
(b) Use the information in the graph shown in Figure 1.12 (and assume
that the intercept on the c axis is at 0.001M) to find the values of a
and b (hint: find the equation of the line in the figure, and compare
it to the relationship v = a − bc). M
0
(c) What is the rate of the reaction when c = 0.005 M?
Figure 1.12: Figure for Exercise 23; rate of a
1.24. Michaelis-Menten kinetics. Consider the Michaelis-Menten kinetics chemical reaction. Assume that the intercept
of the line on the c axis is at 0.01M
where the speed of an enzyme-catalyzed reaction is given by v =
Kx/(kn + x).
(a) Explain the statement that “when x is large there is a horizon-
tal asymptote” and find the value of v to which that asymptote
approaches.
(b) Determine the reaction speed when x = kn and explain why the
constant kn is sometimes called the “half-max” concentration.
1.25. A polymerization reaction. Consider the speed of a polymerization
reaction shown in Figure 1.13. Here the rate of the reaction is plotted
as a function of the substrate concentration; this experiment concerned
the polymerization of actin, an important structural component of
cells; data from [Rohatgi et al., 2001]. The experimental points are
shown as dots, and a Michaelis-Menten curve has been drawn to best
fit these points. Use the data in the figure to determine approximate
values of K and kn in the two treatments shown.
·10−2 Figure 1.13: Figure for Exercise. 25; speed

Maximum Polymerization Rate
of polymerization.
3
(arbitrary units)
2
GST-NcK
GST-NcK+PIP2
1
0
0 0.2 0.4 0.6
Concentration (µM )
1.26. Hill functions. Hill functions are sometimes used to represent a

biochemical “switch,” that is a rapid transition from one state to
another. Consider the Hill functions

x2 x5
y1 = , y2 = ,
1 + x2 1 + x5
(a) Where do these functions intersect?
(b) What are the asymptotes of these functions?
(c) Which of these functions increases fastest near the origin?
(d) Which is the sharpest “switch” and why?
1.27. Transforming a Hill function to a linear relationship. A Hill func-
tion is a nonlinear function - but if we redefine variables, we can
transform it into a linear relationship. The process is analogous to
transforming Michaelis-Menten kinetics into a Lineweaver-Burk plot,
as discussed in Appendix G.1.
(a) Determine how to define appropriate variables X and Y (in terms
of the original variables x and y) so that the Hill function y =
Ax3 /(a3 + x3 ) is turned into a linear relationship between X and Y .
(b) Indicate how the slope and intercept of the line are related to the
original constants A, a in the Hill function.
1.28. Hill function and sigmoidal chemical kinetics. It is known that the
rate v at which a certain chemical reaction proceeds depends on the
concentration of the reactant c according to the formula
Kc2
v= ,
a2 + c2
where K, a are some constants. When the chemist plots the values of
the quantity 1/v (on the “y” axis) versus the values of 1/c2 (on the “x
axis”), she finds that the points are best described by a straight line
with y-intercept 2 and slope 8. Use this result to find the values of the
constants K and a.
1.29. Lineweaver-Burk plots. Shown in the Figure 1.14(a) and (b) are two
Lineweaver-Burk plots (see Appendix G.1). By noting properties of
these figures comment on the comparison between the following two
enzymes:
(a) Enzyme (1) and (2).
(b) Enzyme (1) and (3).
1.30. Michaelis-Menten enzyme kinetics. The rate of an enzymatic reac-
tion according to the Michaelis-Menten kinetics assumption is
Kc
v= ,
kn + c
where c is concentration of substrate (shown on the x-axis) and v is the
reaction speed (given on the y-axis). Consider the data points given in
Table 1.2.
1/v 1/v Figure 1.14: Figure for Exercise 29:

Lineweaver-Burk plots.
(1)
(1)
(2) (3)
1/c 1/c
(a) (b)
Substrate concentration nM c 5 10 20 40 50 100

Reaction speed nM/min v 0.068 0.126 0.218 0.345 0.39 0.529
(a) Convert this data to a Lineweaver-Burk (linear) relationship (see Table 1.2: Chemical reaction speed data.
Appendix G.1 for discussion).

(b) Plot the transformed data values on a graph or spreadsheet, and
estimate the slope and y-intercept of the line you get.
(c) Use these results to find the best estimates for K and kn .
1.31. Spacing in a school of fish According to the biologist Breder [Breder,

1951], two fish in a school prefer to stay some specific distance apart.
Breder suggested that the fish that are a distance x apart are attracted
to one another by a force FA (x) = A/xa and repelled by a second
force FR (x) = R/xr , to keep from getting too close. He found the
preferred spacing distance (also called the individual distance) by
determining the value of x at which the repulsion and the attraction
exactly balance.
Find the individual distance in terms of the quantities A, R, a, r (all
assumed to be positive constants.)
2
Average rates of change, average velocity and the secant
line
A physicist might study the motion of a falling ball by taking strobe images at
fixed time intervals, and gluing them side by side to get a record of position
of the ball over time. In a similar manner, cell biologists study the motion of
proteins inside living cells. First, the proteins are labeled by fluorescent “tags”
(this makes them visible in microscopic images). Then images of some thin
strip of the cell are made at fixed time intervals, in regions through which the
“glowing” (fluorescent) proteins move. Finally, those thin strips are “glued”
together to form a record of the protein position over time, as shown in each
panel of Fig 2.1. Biologists refer to such images as kymographs.
The “streaks of light” in such kymographs allow us to determine the
locations of the labeled proteins over time. as well as their velocity in the cell.
But how fast were these proteins moving? Why are there zigzags in the left
panel? And what happened in the treated cells (right panel) that made the
Figure 2.1: Cell biologists track the motion
streaks look different from those in the “normal cell” (left panel)1 ? of proteins inside cells on images like
In this chapter we develop the tools to address some of these questions, this kymograph. Data courtesy of Anna
Akhmanova’s lab shows the position of
and to characterize what we mean by velocity. As a first step, we introduce
proteins that track growing ends of long
average rate of change. To motivate the idea, we examine data for common biopolymers called microtubules as they get
processes: changes in temperature, and motion of a falling object. Simple longer or shrink inside cells. Displacement
is shown horizontally (scale bar 5 µm) and
experiments are described in each case, and some features of the data are time vertically (scale bar 20 sec).
discussed. Based on each example, we calculate net change over some time 1
Benjamin P Bouchet, Ivar Noordstra, Mi-
interval and then define the average rate of change and average velocity. randa van Amersfoort, Eugene A Katrukha,
York-Christoph Ammon, Natalie D Ter Ho-
This concept generalizes to functions of any variable (not only time). We eve, Louis Hodgson, Marileen Dogterom,
interpret this idea geometrically, in terms of the slope of a secant line. Patrick WB Derksen, and Anna Akhmanova.
Mesenchymal cell invasion requires coop-
In both cases, we ask how to use average rate of change (over a given
erative regulation of persistent microtubule
interval) to find better and better approximations of the rate of change at a growth by slain2 and clasp1. Developmental
single instant, (i.e. at a point). We find that one way to arrive at this abstract cell, 39(6):708–723, 2016
concept entails refining the dataset - collecting data at closer and closer time
points. A second - more abstract - way is to use a limit. Eventually, this
procedure allows us to arrive at the definition of the derivative, which is the
instantaneous rate of change.
2.1 Time-dependent data and rates of change
1. Use (your favorite) graphical software package (spreadsheet, graphics

calculator, online tools, etc.) to plot data points such as those in Table 2.1.
2. Describe the trends seen in such data using words such as ‘increasing’,
‘decreasing’, ‘linear’, ‘nonlinear’, ‘shallow’, ‘steep’, ‘changes’, etc.
In this section we consider time dependent processes and develop the idea
of rates of change. We also use graphical software to represent the data.
Milk temperature and a recipe for yoghurt
To make yoghurt, heat milk to 190◦ F to kill off bacteria, then cool to 110◦ F.
time (min) Temperature (◦ F)
Add a spoonful of “live” pre-made yoghurt and keep the mixture at 110◦ F for
0.0 44.3
7-8 hours. This promotes growth of the microorganism Lactobacillus, that 0.5 61
turns milk into yoghurt. 1.0 77
1.5 92
2.0 108
Example 2.1 (Heating and cooling milk) Shown in Tables 2.1 and 2.2 are 2.5 122
sets of temperature measurements over time. Use your favorite software to 3.0 135.3
3.5 149.2
plot the data and describe the trends you see in each graph. 4.0 161.9
4.5 174.2
Solution. The data is plotted in Figure 2.2, and points are connected with line 5.0 186
segments. The heating phase is shown on the left. (Temperature increases Table 2.1: Heating milk temperature data.
at a nearly linear rate.) On the right, the milk is cooling and the temperature
time (min) Temperature (◦ F)
decreases, but the slope of the graph becomes shallower with time. ♦ 0 190
2 176
4 164.6
200 200
Temperature (◦F) Temperature (◦F) 6 155.4
8 148
150 180
10 140.9
14 131
160 18 123
100
22 116
140 26 111.2
50
120 Table 2.2: Cooling milk temperature data.
1 2 3 4 5 5 10 15 20 25
time (min) time (min) Figure 2.2: Plot of temperatures of milk
being heated (left) and cooled (right).
In this chapter, we will be concerned with describing the rates of change

of similar processes, that is in quantifying what we mean by “how fast is
the temperature changing” in examples of this sort. Before answering, we
introduce two other examples of time dependent data. 1. Reproduce one of the graphs in
Figure 2.2 using Celsius.
AV E R AG E R AT E S O F C H A N G E , AV E R AG E V E L O C I T Y A N D T H E S E C A N T L I N E 53
Data for swimming tuna

Example 2.2 (Bluefin tuna swimming distances) The tuna fishing industry
is of great economic value, but the danger of overfishing is recognized. Prof
Molly Lutcavage2 studied the swimming behaviour of Atlantic bluefin tuna 2
ME Lutcavage, RW Brill, GB Skomal,
(Thunnus thynnus L.) in the Gulf of Maine. She recorded their position over a BC Chase, JL Goldstein, and J Tutein.
Tracking adult north atlantic bluefin tuna
period of 1-2 days. Some of her approximate data is given in Table 2.3. Plot (thunnus thynnus) in the northwestern
the data points and describe the trends these display. atlantic using ultrasonic telemetry. Marine
Biology, 137(2):347–358, 2000
Solution. As shown in Figure 2.3, distance traveled by Tuna 1 is roughly

proportional to time spent, since its graph is roughly linear (almost a straight
time distance distance
line). This linear relationship between distance travelled and time spent is (hr) Tuna 2 (km) Tuna 1 (km)
called uniform motion. 0 0 0
Tuna 2 started with similar uniform motion, but later sped up. During 5 29 32
10 51 55
15 ≤ t ≤ 20h, it was swimming much faster. ♦ 15 78 80
20 140 111
25 160 125
200 distance (km) 30 182 150
35 218 180
Tuna 2
150
Table 2.3: Data for tuna swimming distance
Tuna 1 collected by Prof. Molly Lutcavage in the
100 Gulf of Maine.
50
time (hrs)
5 10 15 20 25 30 35 Figure 2.3: Distance travelled by two bluefin
tuna over 35 hrs
Distance of a falling object

Long ago, Galileo devised some ingenious experiments to track the position
of a falling object. He used his measurements to quantify the relationship
between the total distance fallen over a given time. Although Galileo did not
have formulae nor graph-paper in his day - and was thus forced to express
this relationship in a cumbersome verbal way - what he had discovered was
quite remarkable. Mastered Material Check
2. Use Figure 2.3 to approximate how
Example 2.3 (Gallileo’s formula for height of a falling object) Galileo far each tuna travelled after 18
discovered that the distance fallen, y(t ), is proportional to the square of the hours.
time t, that is 3. Does an object dropped from a

height of 15m hit the ground in 2
y(t ) = ct 2 , (2.1) seconds? 1 second?
where c is a constant. When distance is measured in meters (m) and time in

seconds (s) the constant is found to be c = 4.9m/s2 . Using Eqn. (2.1), plot
a graph of the distance fallen y(t ) versus time t for 0 ≤ t ≤ 2 seconds at
intervals of 0.1s. Connect the data points and comment on the shape of the
graph.
Solution. The graph is shown in Figure 2.4. We recognize this as a parabola, 20

y(t)
resulting from the quadratic relationship of y and t. (In fact, the relationship
15
is that of a simple power function with a constant coefficient.) ♦
y(t) = 4.9t2
10
Having looked at three examples of data for time-dependent processes,
we now turn to quantifying the rate at which change occurs in each process. 5
We start with the notion of average rate of change, and eventually refine and
idealize this idea it to develop rates of change at an instant in time. t
0.5 1 1.5 2
Figure 2.4: The distance y (in meters) of
2.2 The slope of a straight line is a rate of change an object falling under the force of gravity,
versus time t (in seconds).
1. Define rate of change for a linear relationship.
2. Compute the rate of change for a linear relationship.

In the examples discussed so far, we have plotted data and used words to 4. What does it mean for two variables
describe trends. Our goal now is to formalize the idea of change and rate to have a linear relationship?
of change. Let us consider the simplest case where a variable of interest, y 5. Why are we ‘idealizing’?
depends linearly on time, t. This was approximately true in some examples 6. Compute the rate of change for a
linear relationship which goes
seen previously (Figure 2.2a, parts of Figure 2.3). We can describe this kind through points (1, 4) and (2, 2).
of relationship by the idealized equation
y(t ) = mt + b. (2.2)
A graph of y versus t is then a straight line with slope m and intercept b. ` The equation of a straight line (2.2)
specifies the slope m and the y intercept
Definition 2.1 (Rate of change for a linear relationship) For a straight b of the line, as shown by manipulating
the sliders on this interactive graph.
line, we define the rate of change of y with respect to time t as the ratio:
Change in y
.
Change in t
We now make an important observation.
Observation (“Theorem”): The slope m of the straight line in Eqn. (2.2)

corresponds to definition 2.1 of the rate of change of a linear relationship. Hint: A “Theorem” is just a
mathematical statement that can be
established rigorously by an argument
Proof: Taking any two points (t1 , y1 ) and (t2 , y2 ) on that line, and using called a “proof”. While we will not
the notation ∆y, ∆t to represent the change in y and t we compute the ratio use such terminology often here, it is a
∆y/∆tand simplify algebraically to find: staple of mathematics.
Change in y ∆y y2 − y1
= =
Change in t ∆t t2 − t1
(mt2 + b) − (mt1 + b) mt2 − mt1
= = = m.
t2 − t1 t2 − t1
Thus, the slope m corresponds exactly to the notion of change of y per unit
time which we call henceforth the rate of change of y with respect to time.
It is important to notice that this calculation leads to the same result no matter
which two points we pick on the graph of the straight line. ♦
Featured Problem 2.1 (Velocity of growing microtubule tips) Shown in

Fig. 2.5 is a part of the kymograph from Fig. 2.1, but with a more “conven-
tional” view of position y (in µm) on the vertical axis versus time t (in sec-
onds) on the horizontal axis. The bright streaks are the tips of microtubules
(MT) at various positions as they grow inside a cell.
1. Use the image and the superimposed grid to estimate the average veloc-
ities of microtubule tips (up to one significant digit). This is best done on
the lower panel, where the position versus time graphs can be most easily
viewed.
2. Based on the lines in the graph, explain whether the three microtubule tips
shown are moving at similar or at quite different speeds.
3. Compare the normal and treated cells shown in two panels of Fig. 2.1,
carefully noting the fact that the coordinate system differs from that of
Fig. 2.5. How does the treatment affect the speed of microtubule tips?
4. Explain what could account for the apparent “zigzags” and “curves” (not
straight lines) seen in both panels of Fig. 2.1.
2.3 The slope of a secant line is the average rate of change
Section 2.3 Learning goals Figure 2.5: Part of the image from Fig. 2.5
is shown (top panel) on a yt (distance in
1. Define average rate of change; explain its connection with the slope of a µm vs time in seconds) coordinate system,
with a grid superimposed. The red lines
secant line.
represent positions of proteins that track the
tips of growing microtubules over time. In
2. Compute the average rate of change using time-dependent data over a the lower panel, the same lines are shown on
given time interval. the grid alone; we can use this to estimate
average velocity.
3. Given two points on the graph of a function, or two discrete data points,
find both the slope and the equation of a secant line through those points.
We generalize the ideas in Section 2.2 to consider rates of change for

relationships other than linear. Let y = f (t ) describe some relationship
between time t and a variable of interest y. (This could be a set of discrete Mastered Material Check
data points as in Figure 2.2, or a formula, as in Eqn. (2.1).) 7. Sketch some nonlinear function and
two different secant lines.
Pick any two points (a, f (a)), and (b, f (b)) satisfying y = f (t ), and
8. Is there any reason why we must
connect the points with a straight line. We refer to this line as the secant line, draw secant lines between pairs of
and we call its slope an average rate of change over the interval a ≤ t ≤ b. successive data points?
Formally, we define
Definition 2.2 (Secant Line) A secant line is a straight line connecting any
two specific points on the graph of a function.
Definition 2.3 (Average rate of change) The average rate of change of

y = f (t ) over the time interval a ≤ t ≤ b is the slope of the secant line through
the two points (a, f (a)), and (b, f (b)).
Based on the above definition, we compute the average rate of change of f

over the time interval a ≤ t ≤ b as
Change in f ∆f f (b) − f (a)
Average rate of change = = = .
Change in t ∆t b−a
Observe that the average rate of changes in general depending on which two
points we select, in contrast to the linear case. (See left panel in Figure 2.6.)
The word “average” sometimes causes confusion. One often speaks in a
different context about the average value of a set of numbers (e.g. the average
of {7, 1, 3, 5} is (7 + 1 + 3 + 5)/4 = 4.) However the term average rate of
change always means the slope of the straight line joining a pair of points.
Figure 2.6: A set of time dependent data

y = f (t) points (black circles) or smooth function
f (b) (dashed curve) f (t ) showing a secant line
through the points (a, f (a)), and (b, f (b)).
Another secant line is drawn through
(a, f (a)) and a different point to show the
secant line dependence on the points we select.
f (a)
another
secant line t
a b
For example, the average rate of change of milk temperature discussed in

Example 2.1 is
Change in temperature ∆T
= .
Time taken ∆t
Featured Problem 2.2 (Average rate of change of milk temperature) 1.
Use the data in Tables 2.1 and 2.2 to show that the average rates of
change of the temperature over the time interval 2 ≤ t ≤ 4 min for the cool-
ing phase is −5.7◦ /min. Repeat the calculation over a similar interval for
the heating phase, and show that you get 26.95◦ /min.
2. Write the equation of the secant line through the data points t = 2 and
t = 4 for the heating phase.
Definition 2.4 (Average velocity) For a moving body, the average velocity
over a time interval a ≤ t ≤ b is the average rate of change of distance over
the given time interval.
Example 2.4 (Swimming velocity of Bluefin tuna) Use the tuna swimming
data in Figure 2.3 to answer the following questions:
(a) Determine the average velocity of each of these two fish over the 35h
shown in the figure.
(b) What is the fastest average velocity shown in this figure, and over what
time interval and for which fish did it occur?
Solution.
(a) We find that Tuna 1 swam 180 km, whereas Tuna 2 swam 218 km over the
course of 35 hr. Thus, the average velocity of Tuna 1 was v̄ = 180/35 ≈
5.14 km/h, whereas for Tuna 2 it was 6.23 km/h.
(b) The fastest average velocity corresponds to the segment of the graph that
9. What was the average velocity of
has the greatest slope. This occurs for Tuna 2 during the time interval Tuna 1 during 15 ≤ t ≤ 20?
15 ≤ t ≤ 20. Indeed, over that 5 hr interval the tuna has a displacement
(net distance covered) of 140-78=62km. Its average velocity over that time
interval was thus 62/5 = 12.4km/h.
Example 2.5 (Equation of secant line 2) Find the equation of the secant
line connecting the first and last data points for the Tuna 1 swimming dis-
tances in Figure 2.3.
Solution. We defined the distance as 0 at time t = 0, so that the y intercept

of the secant line is 0. We have already computed the slope of the secant line
(average rate of change) as 5.14 km/h. Hence the equation of the secant line ` A secant line between two points, x0
is and x0 + h on the graph of a function
yS = 5.14t. f (x) is shown in this link. You can
change the base point x0 , the distance
♦ between the x coordinates, h, or you can
input your own function for f (x). The
We can extend the definition of average rate of change to any function f (x). slope of the secant line is the average
rate of change of f over the interval
Definition 2.5 (Average rate of change of a function) Suppose y = f (x) x0 ≤ x ≤ x0 + h
is a function of some arbitrary variable x. The average rate of change of f

between two points x0 and x0 + h is given by
y = f (x)
f (x0 + h)
Change in y ∆y [ f (x0 + h) − f (x0 )] [ f (x0 + h) − f (x0 )]
= = = .
Change in x ∆x ( x0 + h ) − x0 h
secant line
Here h is the difference of the x coordinates. The above ratio is the slope of
the secant line shown in Figure 2.7. f (x0 )
x
x0 x0 + h
Example 2.6 (Average velocity of a falling object) Consider a falling Figure 2.7: The graph of some arbitrary
object. Suppose that the total distance fallen at time t is given by Eqn. (2.1), function f (x) (dashed line) with a secant
line through the points (x0 , f (x0 )) and
y(t ) = ct 2 . Find the average velocity v̄, of the object over the time interval (x0 + h, f (x0 + h)). The slope of the secant
t0 ≤ t ≤ t0 + h. line is the average rate of change of f over
the given interval.
Solution. In Figure 2.8, we reproduce the data for the falling object from y(t) y(t) = 4.9t2
secant line and
Figure 2.4 and superimpose a secant line connecting two points labeled t0 and
average veloc-
t0 + h. ity
We compute the average velocity as follows: secant
y(t0 + h) − y(t0 ) line
v̄ = ← (definition of average velocity)
h
c(t0 + h)2 − c(t0 )2 t
= ← the function of interest t0 t0 + h
h
2 Figure 2.8: A secant line through two points
(t0 + 2ht0 + h2 ) − (t02 )

= c some algebra on the graph of distance versus time for an
h object falling under the force of gravity.
2ht0 + h2

= c simplifying the expression
h
= c(2t0 + h). (2.3)
The average velocity over the time interval t0 < t < t0 + h is v̄ = c(2t0 + h). ♦
2.4 From average to instantaneous rate of change
1. Demonstrate that a data set with more frequent measurements corresponds

to smaller time intervals ∆t between data points.
2. Describe the connection between average rate of change over a very small
time interval and instantaneous rate of change at a single point.
This section could also be titled “Shrinking the time-steps between

measurements.” So far, the average rates of change were computed over
finite intervals. Our ultimate goal is to refine this idea and define a rate of
change at each point, i.e. an instantaneous rate of change. But to do so,
we first consider how a data set can be refined by making more frequent
measurements to improve the notion of a rate of change close to a given point.
We discuss two examples below.
Refined temperature data

We can refine the original data of temperature T (t ) for cooling milk from
Figure 2.2 by taking more closely spaced time points. Table 2.4 provides a
sample of the refined data.
The leftmost plot in Figure 2.9 shows the original data set with measure-
ments every ∆t = 2 min. The second and third plots have successively more
refined measurements with shorter intervals between time points (∆t = 1 min
and ∆t = 0.5 min). After 10 minutes, fewer points were collected in each case.
Example 2.7 (Refined average rate of change) Use the data in Table 2.4 to
compute the average rate of change of the temperature over the time intervals
time Temp time Temp time Temp Table 2.4: Partial data for temperature in ◦ F
for the three graphs shown in Figure 2.9.
0 190 0 190 0 190 The pairs of columns indicate that the
2 176 1 182 0.5 185.5 data has been collected at more and more
4 164.6 2 176 1 182 frequent intervals h = ∆t.
6 155.4 3 169.5 1.5 179.2

8 148 4 164.6 2 176
10 140.9 5 159.8 2.5 172.9
200 200 200

Temperature (◦F )
Temperature (◦F )
Temperature (◦F )
180 180 180
160 160 160
140 140 140
120 120 120
5 10 15 20 25 5 10 15 20 25 5 10 15 20 25
time (min) time (min) time (min)
Figure 2.9: Three graphs of temperature

versus time for cooling milk.
2 ≤ t ≤ 2 + h where h = ∆t = 2, 1, 0.5 min, respectively. Which calculation
most accurately describes the behaviour “close to” t = 2min?
Solution. Computing the ratio ∆T /∆t, we obtain, for ∆t = 2, 1, 0.5 the

following average rates of change (in ◦ F per min): Mastered Material Check
10. At what time is the temperature
∆T T (2 + 2) − T (2) (164.6 − 176) approximately 120◦ F?
∆t = 2 : = = = −5.7,
∆t 4−2 (4 − 2)
∆T T (2 + 1) − T (2) (169.5 − 176)
∆t = 1 : = = = −6.5,
∆t 3−2 (3 − 2)
∆T T (2 + 0.5) − T (2) (172.9 − 176)
∆t = 0.5 : = = = −6.2.
∆t 2.5 − 2 (2.5 − 2)
The last of these has been calculated over the smallest time interval, and most
closely represents the rate of change of temperature close to the time t = 2
min. Exercise 2(b) leads to a similar comparison of this sort close to t = 0, and
results in a similar set of finer values for the average rate of change “near” the
initial data point. ♦
Refined data for the height of a falling object

We examine increasingly refined data for the height of a falling object in
Figure 2.10.
Figure 2.10(a) shows three stroboscopic images, each giving successive
vertical positions of an object falling from a height of 20 m over a 2 s time
20m 20m Figure 2.10: Data for the height Y of

a falling object starting with t = 0 at
the top, and following the object until
Y Y t = 2 at its lowest point. The data set is
refined (∆t = 0.5, 0.2, 0.1) to get more
and more accurate tracking of the object.
(a) Stroboscopic images. (b) Graphs of Y
versus t.
0m t 0m t
(a) (b)
period. The location of the ball is given first at intervals of ∆t = 0.5 seconds,
then at intervals of ∆t = 0.1 and finally ∆t = 0.05 s. In Figure 2.10(b), we
11. From what height was the object
graph the height Y = Y0 − ct 2 against time t. (The distance fallen is still dropped in Figure 2.10?
described by the function y(t ) = ct 2 , as in Example 2.3.) 12. If you wanted 50 equally spaced
By collecting data at finer time points, we can determine the “velocity” data points over a 2 s interval, what
would ∆t be?
of the object with greater accuracy. Indeed, taking smaller and smaller time
steps leads us to define instantaneous velocity.
Instantaneous velocity
i A brief summary of average and
instantaneous velocity in the example of
We know that the velocity over an interval can be calculated by finding the
a falling ball.
slope of a secant line connecting the endpoints of that interval. The slopes of
the secant lines in Figure 2.10(b) are steeper at the end of the time interval
than at the beginning - lending justification to what we intuitively know: a
falling object’s velocity increases as time passes.
With this in mind, to define an instantaneous velocity at some time t0 , we
compute average velocities over decreasing time intervals t0 ≤ t ≤ t0 + h,
allowing h to get smaller.
Note: we use the notation h → 0 to denote the shrinking the time interval.
For example, we make the strobe flash faster so that ∆t = (t0 + h) − t0 =
h → 0. At each stage, we calculate an average velocity, v̄ for the interval
t0 ≤ t ≤ t0 + h. As we continue to refine the measurements in this way, we
arrive at a value for the velocity that we denote the instantaneous velocity.
This number represents “the velocity of the ball at the very instant t = t0 ".
Definition 2.6 (Instantaneous velocity) The instantaneous velocity at time

t0 , denoted v(t0 ) is defined as
v(t0 ) = lim v̄
h→0
where v̄ is the average velocity over the time interval t0 ≤ t ≤ t0 + h. In other

words,
y(t0 + h) − y(t0 )
v(t0 ) = lim .
h→0 h
We shall be more explicit about the meaning of the notation lim in the next
h→0
chapter.
Example 2.8 (Computing an instantaneous velocity) Use Gallileo’s

formula for the distance fallen, Eqn. (2.1), y(t ) = ct 2 , to compute the instanta-
neous velocity of a falling object at time t0 .
Solution. We have already found the average velocity of the falling object
over a time interval t0 ≤ t ≤ t0 + h in Example 2.6, obtaining Eqn. (2.3),
v̄ = c(2t0 + h).
Then, by Definition 2.6,
v(t0 ) = lim v̄ = lim c(2t0 + h) = 2ct0 .

h→0 h→0
Here we have used the fact that the expression c(2t0 + h) approaches 2ct0 as
h shrinks to zero. This result holds for any time t0 . More generally, we could
write that at time t, the instantaneous velocity is v(t ) = 2ct. For example, for
c = 4.9m/s2 , the velocity of an object at time t = 1 s after it is released is
v(1) = 9.8 m/s. ♦
2.5 Introduction to the derivative
1. Explain the first examples of calculation of the derivative.
2. Describe how the derivative is obtained from an average rate of change.
3. Compute the derivative of very simple functions such as y = x2 , and y =

Ax + B.
We are ready for the the definition of the derivative.

` As h → 0, the secant line approaches
Definition 2.7 (The derivative) The derivative of a function y = f (x) at a tangent line. Use the slider for h to
show this trend, and note that the slope
a point x0 is the same as the instantaneous rate of change of f at x0 . It is of the secant line (average velocity)
dy
or f 0 (x0 ) and defined as

denoted dx x approaches the slope of the tangent line
0
(instantaneous velocity) at the point x0 .

dy [ f (x0 + h) − f (x0 )]
= f 0 (x0 ) = lim .
dx x0 h→0 h
We can use this to update our definition of instantaneous velocity:
Definition 2.8 (Velocity) If y = f (t ) is the position of an object at time t then

the derivative f 0 (t ) at time t0 is the instantaneous velocity, also simply called
the velocity of the object at that time.
Example 2.9 (Formal calculation of velocity) Use Gallileo’s formula to set

up and calculate the derivative of Eqn. (2.1), and show that it corresponds to
the instantaneous velocity obtained in Example 2.8.
Solution. We set up the calculation using limit notation, recalling Gallileo’s

formula states y(t ) = ct 2 . We compute
y(t0 + h) − y(t0 )
v(t0 ) = lim
h→0 h
c(t0 + h)2 − c(t0 )2
= lim
h→0 h
2
(t0 + 2ht0 + h2 ) − (t02 )

= lim c
h→0 h
2ht0 + h 2

= lim c = lim c(2t0 + h) = 2ct0 .
h→0 h h→0
(2.4)
All steps but the last are similar to the calculation (and algebraic simplifica-
tion) of average velocity (compare with Example 2.6). In the last step, we
formally allow the time increment h to shrink, which is equivalent to taking
limh→0 . ♦
Example 2.10 (Calculating the derivative of a function) Compute the

derivative of the function f (x) = Cx2 at some point x = x0 .
Solution. In the previous example, we calculated the derivative of the func-

tion y = f (t ) = ct 2 with respect to t. Here we merely have a similar (quadratic)
function of x. Thus, we have already solved this problem. By switching no-
tation (t0 → x0 and c → C) we can write down the answer, 2Cx0 at once.
However, as practice, we rewrite the steps in the case of the general point x
For y = f (x) = Cx2 we have
dy f (x + h) − f (x )
= lim
dx h→0 h
C (x + h)2 −Cx2
= lim
h→0 h
(x2 + 2xh + h2 ) − x2
= lim C
h→0 h
2
(2xh + h )
= lim C = lim C (2x + h) = 2Cx.
h→0 h h→0
Evaluating this result for x = x0 we obtain the answer 2Cx0 . ♦
We recognize from this definition that the derivative is obtained by starting

with the slope of a secant line (average rate of change of f over the interval
x0 ≤ x ≤ x0 + h) and proceeds to shrink the interval (limh→0 ) so that it
approaches a single point (x0 ). In later chapters, the resultant line is called the
tangent line and the value obtained identified as the instantaneous rate of
change of the function with respect to the variable x at the point of interest,
x0 . We explore properties and meanings of this concept in the next chapter.
Note: we have used different notations to denote the derivative of f (x) = y.

Further others exist. Each of the following may be used interchangeably:
df d dy
f 0 (x ), , f (x ), , y0 , D f (x ), and Dy.
dx dx dx
These notations evolved for historical reasons and are used interchangeabley
in science.
2.6 Summary
1. Graphs of time-dependent data are helpful to visualize trends such as

increasing and decreasing values, steepness, linearity, and so on.
2. An average rate of change is the ratio of change in a dependent variable (y)
over a range of the independent variable (x), often denoted ∆y/∆x.
3. A secant line is a straight line through any two points on the graph of a
function. The slope of a secant line is the average rate of change of the
function over interval between the x coordinates of the two points.
4. The average rate of change of a function y = f (x) on x0 ≤ x ≤ x0 + h can
be computed by the ratio
∆y Change in y [ f (x0 + h) − f (x0 )]

= =
∆x Change in x h
5. The instantaneous rate of change of a function f (x) at x0 can be found by

taking the limit as h → 0 of the average rate of change on x0 ≤ x ≤ x0 + h.
6. The derivative of a function y = f (x) at a point x0 is the same as the
instantaneous rate of change of f at x0 .
7. In the case of time-dependent data, refining the data can lead to a better
and better approximation of instantaneous rates of change.
8. This chapter explored data related to the following time-dependent pro-
cesses:
(a) height of a falling object;

(b) temperature of heating/cooling milk; and
(c) swimming velocity of Bluefin tuna.
1. How do we calculate average rate of change of a time dependent process over a given interval?
2. Over what interval does the function depicted in the graph below have the greatest average rate of change? Smallest
average rate of change?
4
y
2
x
−4 −2 2 4
−2
−4
3. Given the function defined by {(1, 3), (2, 5), (3, 7)}, how many different secant lines can be formed?
4. Use the definition to calculate the derivative of f (x) = 4x2 + 3 at x0 = 1.

Exercises
2.1. Heating milk. Consider the data gathered for heating milk in Ta-
ble 2.1and Figure 2.2 (left)
(a) Estimate the slope and intercept of the straight line shown in the
figure and use to write the equation of this line. According to this
approximate straight line relationship, what is the average rate of
change of the temperature over the 5 min interval shown?
(b) Find a pair of points such that the average rate of change of the
temperature is smaller than your result in part (a).
(c) Find a pair of points such that the average rate of change of the
temperature is greater than your result in part (a).
(d) Milk boils at 212◦ F, and the recipe for yoghurt calls for avoiding a
temperature this high. Use your common knowledge to explain why
the data for heating milk is not actually linear.
2.2. Refining the data. Table 2.4 shows some of the data for cooling milk
that was collected and plotted in Figure 2.9. Answer the following
questions.
(a) Use the table to determine the average rate of change of the temper-
ature over the first 10 min.
(b) Compute the average rate of change of the temperature over the
intervals 0 ≤ t ≤ 2, 0 ≤ t ≤ 1 and 0 ≤ t ≤ 0.5.
(c) Which of your results in (b) would be closest to the “instantaneous"
rate of change of the temperature at t = 0?
2.3. Height and distance dropped. We have defined the variable Y (t ) =
height of the object at time t and the variable y(t ) as the distance
dropped by time t.
(a) State the connection between these two variables for a ball whose
initial height is Y0 .
(b) How is the displacement over some time interval a ≤ t ≤ b related
between these two ways of describing the motion? (Assume that
the ball is in the air throughout this time interval).
2.4. Falling ball. A ball is dropped from height Y0 = 490 meters above
the ground. Its height, Y , at time t is known to follow the relation-
ship Y (t ) = Y0 − 21 gt 2 where g = 9.8 m /s2 .
(a) Find the average velocity of the falling ball between t = 1 and t = 2
seconds.
(b) Find the average velocity between t sec and t + ε sec where 0 ≤ ε ≤
1 is some small time increment (assume that the ball is in the air
during this time interval).
(c) Determine the time at which the ball hits the ground.
2.5. Tuna average velocity. Find the average velocity of Tuna 1 over
each of the time intervals shown in Table 2.3, that is for 0 ≤ t ≤ 5 hr,
5 ≤ t ≤ 10 hr, etc. year human population
2.6. Average velocity and secant line. The two points on Figure 2.8 (billions)
1 0.2
through which the secant line is drawn are (1.3, 8.2810) and (1.4, 9.6040). 1000 0.275
Find the average velocity over this time interval and then give the 1500 0.45
1650 0.5
equation of the secant line. 1750 0.7
2.7. Human Population Growth. Table 2.5 gives data for the human 1804 1
1850 1.2
population (in billions) over recorded history (with some estimates 1900 1.6
where data was not available). 1927 2
1950 2.55
Note: human population growth is further studied in Chapter 11. 1960 3
1980 4.5
(a) Plot the human population (in billions) versus time (in years) using 1987 5
graphing software of your choice. 1999 6
2011 7
(b) Determine the average rate of change of the human population over 2020 7.7
the successive time intervals.
Table 2.5: The human population (billions)
(c) Plot the average rate of change versus time (in years) and determine over the years AD 1 to AD 2020.
over what time interval that average rate of change was greatest.
(d) Over what period (i.e. time interval) was this average rate of
change increasing most rapidly? (hint: you should be able to
answer this question either by looking at the graph you have drawn
or by calculation)
2.8. Average velocity at time t. A ball is thrown from the top of a building
of height Y0 . The height of the ball at time t is given by
1
Y (t ) = Y0 + v0t − gt 2
2
where h0 , v0 , g are positive constants. Find the average velocity of the
ball for the time interval 0 ≤ t ≤ 1 assuming that it is in the air during
this whole time interval. Express your answer in terms of the constants
given.
2.9. Average rate of change. A certain function takes values given in
Table 2.6. Find the average rate of change of the function over the t 0 0.5 1.0 1.5 2.0
f (t ) 0 1 0 -1 0
intervals.
Table 2.6: Function values for Exercise 9.
(a) 0 ≤ t ≤ 0.5,
(b) 0 ≤ t ≤ 1.0,
(c) 0.5 ≤ t ≤ 1.5,
(d) 1.0 ≤ t ≤ 2.0.
2.10. Average rate of change. Find the average rate of change for each of
the following functions over the given interval.
(a) y = f (x) = 3x − 2 from x = 3.3 to x = 3.5.

(b) y = f (x) = x2 + 4x over [0.7, 0.85].
(c) y = − 4x and x changes from 0.75 to 0.5.
2.11. Trig mini-review. Consider the table of values of the trigonometric x sin (x ) cos (x )
functions sin(x) and cos(x) found in Table 2.7. 0 0 √
1
π 1 3
For the following, express your answer in terms of square roots and π. 6
π
√2
2
√2
2
4 √2 2
Do not compute decimal expressions. π 3 1
3 2 2
π
(a) Find the average rate of change of sin(x) over 0 ≤ x ≤ π/4. 2 1 0
(b) Find the average rate of change of cos(x) over π/4 ≤ x ≤ π/3. Table 2.7: Table of sine and cosine values
(c) Is there an interval over which the functions sin(x) and cos(x) have
the same average rate of change? (hint: consider the graphs of
these functions over one whole cycle, e.g. for 0 ≤ x ≤ 2π. Where do
they intersect?)
Note: trigonometry is reviewed in Appendix F and studied further in

Chapters 14 and 15.
2.12. Secant and tangent lines. Let y = f (x) = 1 + x2 and consider the
point (1, 2) on its graph and some point nearby, for example
(1 + h, 1 + (1 + h)2 ).
(a) Find the slope of a secant line connecting these two points.
(b) The slope of a tangent line to y = f (x) is the derivative f 0 (x). Use
the slope you calculated in (a) to determine what the slope of the
tangent line to the curve at (1, 2) would be.
(c) Find the equation of the tangent line through the point (1, 2).
2.13. Secant and tangent lines. Given the function y = f (x) = 2x3 + x2 − 4,
(a) find the slope of the secant line joining the points (4, f (4)) and (4 +
h, f (4 + h)) on its graph, where h is a small positive number, then
(b) find the slope of the tangent line to the curve at (4, f (4)).
2.14. Average rate of change. Consider the function f (x) = x2 − 4x and the
point x0 = 1.
(a) Sketch the graph of the function.
(b) Find the average rate of change over the intervals [1, 3], [−1, 1],
[1, 1.1], [0.9, 1] and [1 − h, 1], where h is some small positive
number.
(c) Find f 0 (1).
2.15. Approximation using a tangent line. Let y = f (x) = x2 − 2x + 3.
(a) Find the average rate of change over the interval [2, 2 + h].
(b) Find f 0 (2).
(c) Using only the information from (a), (b) and f (2) = 3, approxi-
mate the value of y when x = 1.99, without substituting x = 1.99
into f (x).
2.16. Average rate of change. For the following, express your answer in
terms of square roots and π. Do not compute the decimal expressions.
Note that
sin(x) cos(x)
tan(x) = , cot(x) =
cos(x) sin(x)
(a) Find the average rate of change of tan(x) over 0 ≤ x ≤ π4 .
(b) Find the average rate of change of cot(x) over π
4 ≤ x ≤ π3 .
2.17. Secant and tangent lines.
(a) Find the slope of the secant line to the graph of y = 2/x between the
points x = 1 and x = 2.
(b) Find the average rate of change of y between x = 1 and x = 1 + ε
where ε > 0 is some positive constant.
(c) What happens to this slope as ε → 0 ?
(d) Find the equation of the tangent line to the curve y = 2/x at the
point x = 1.
2.18. Velocity and average velocity. For each of the following motions
where s is measured in meters and t is measured in seconds, find the
velocity at time t = 2 and the average velocity over the given interval.
(a) s = 3t 2 + 5 and t changes from 2 to 3s.
(b) s = t 3 − 3t 2 from t = 3s to t = 5s.
(c) s = 2t 2 + 5t − 3 on [1, 2].
2.19. Acceleration. The velocity v of an object attached to a spring is given
by v = −Aω sin(ωt + δ ), where A, ω and δ are constants. Find the
average change in velocity (“acceleration”) of the object for the time
interval 0 ≤ t ≤ 2π
ω .
Note: acceleration is further explored in Chapter 4.
2.20. Definition of the derivative. Use the definition of derivative to calcu-
late the derivative of the function
1
f (x ) = .
x+1
Note: intermediate steps are required.

3
Three faces of the derivative: geometric, analytic, and
computational
In Chapter 2 we bridged two concepts: the average rate of change (slope

of secant line) and the instantaneous rate of change (the derivative). We
arrived at a technique for calculating a derivative algebraically. As a result,
we introduced limits - a concept that merits further discussion. One goal of
this chapter is to consider the technical aspects of limits - a requirement if we
are to use the definition of the derivative to determine derivatives of common
functions.
But first, we consider a distinct approach which is geometric in flavour.
Namely, we show that the local behaviour of a continuous function is de-
scribed by a tangent line at a point on its graph: we can visualize the tangent
line by zooming into the graph of the function. This duality - the geometric
(graphical) and analytic (algebraic calculation) views - form themes through-
out the discussions to follow. They are two complementary, but closely
related approaches to calculus.
3.1 The geometric view: zooming into the graph of a function
1. Describe the link between the local behaviour of a function (seen by

zooming into the graph) at a point and the tangent line to the graph of the
function at that point.
2. Sketch a function’s derivative given its graph.
Locally, the graph of a function looks like a straight line

In this section we consider well-behaved functions whose graphs are
“smooth”, as opposed to the discrete data points of Chapter 2. We link the
derivative to the local shape of the graph of the function. By local behaviour
we mean the shape we see when we zoom into a point on the graph. Imagine
using a microscope where the center of the field of vision is some point of
interest. As we zoom in, the graph looks flatter, until we observe a straight
line, as shown in Figure 3.1. Mastered Material Check
1. Write down the equation of a
Definition 3.1 (Tangent line) The straight line that we see when we zoom generic straight line.
into the graph of a smooth function at some point x0 is called the tangent 2. Identify the slope of that straight
line at x0 . line.
3. How many zeros does the
Definition 3.2 (Geometric definition of the derivative) The slope of the function f (x) = x3 − x have?
tangent line at the point x is denoted as the derivative of the function at the
given point.
y y
y 4 2.1
3 2
5
2 1.9
x 1 1.8
−2 −1 1 2
x x
−5 1 1.5 2 1.45 1.5 1.55 1.6
Figure 3.1: Zooming in on the graph of

y = f (x) = x3 − x at the point x = 1.5 makes
Example 3.1 (Zooming into a polynomial) Consider the function shown in the graph “look like” a straight line - the
tangent line. The slope of that tangent line is
Figure 3.1, y = f (x) = x3 − x, and the point x = 1.5. Find the tangent line to
the derivative of the function at x = 1.5.
the graph of this function by zooming in at the given point.
Solution. The graph of the function is shown in each panel of Figure 3.1,
where we have indicated the point of interest with a red dot. Now zooming
in on the given point. Locally, the graph resembles a straight line. This is the
tangent line to f (x) at x = 1.5. ♦
` Zooming in to the graph of a
Example 3.2 (Zooming into the sine graph at the origin) Determine the function.
(1) Click on the + (“zoom in”) button
derivative of the function y = sin(x) at x = 0 by zooming into the origin on its on this graph of sin(x) to see how, close
graph. Write down the equation of the tangent line at that point. to the point (0, 0), the graph looks like a
straight line of slope 1.
Solution. In Figure 3.2 we zoom into x = 0 on the graph of the function (2) Now change the function to
f (x) = |x| and note that when zooming
y = sin(x). into the cusp at x = 0, we do not see a
(single) tangent line. In that case we say
The sequence of zooms leads to a straight line (far right panel) that we that the derivative does not exist at the
given point.
identify once more as the tangent line to the function at x = 0. From the graph,
the slope of this tangent line is 1. We say that the derivative of the function
y = f (x) = sin(x) at x = 0 is 1, and write f 0 (0) = 1 to denote this. As this
line goes through (0, 0) and has slope 1, its equation is y = x. We can also say
that close to x = 0 the graph of y = sin(x) looks a lot like the line y = x. ♦ Mastered Material Check
4. Using the far right panel of
Figure 3.2, perform the calculations
that verify the y = sin(x) looks a lot
like y = x near x = 0.
5. Give an example of a function with
a discontinuity.
T H R E E F AC E S O F T H E D E R I VAT I V E : G E O M E T R I C , A NA LY T I C , A N D C O M P U TAT I O NA L 71
y y y
1
0.2
0.5 0.5
x x x
−2 2 −0.5 0.5 −0.2 0.2
−0.5 −0.5
−0.2
−1
Figure 3.2: Zooming into the graph of the
function y = f (x) = sin(x) at the point x = 0.
Eventually, the graph resembles a line of
At a cusp or a discontinuity, the derivative is not defined slope 1. This is the tangent line at x = 0 and
its slope, the derivative of y = sin(x) at x = 0
If we zoom into a function at a cusp or a discontinuity, there is no single is 1.
straight line that describes the local behaviour. For example, in Figure 3.3, we
see two distinct lines meeting at a sharp “corner”.
y y y
x
x x
Figure 3.3: A function has no tangent line at

a a cusp and the derivative is not defined at
Zooming into a discontinuity presents another problem: there is no line that point.
at all, as in Figure 3.4. Finally, a function like 1/x has a singularity at x = 0
which shows up as a vertical line whose slope is infinite. In all such cases,
we say that the function has no tangent line its derivative is not defined at the
cusp, discontinuity, or singularity point.
y
y y
x x
Figure 3.4: Zooming into a function at a
discontinuity shows no single distinct line
that might be a tangent line.
6. Sketch the “zooming in” graph of
the function y = f (x) = sin(x) at
From a function to a sketch of its derivative x = 1.
7. How many local minima are
The tangent line to the graph of a function varies from point to point along depicted in Figure 3.5?
the graph of the function - what we see when zooming in depends on the
location of the zoom. This means that the derivative f 0 (x) is, itself, also
a function. Here we consider the connection between these two functions
by using the graph of one to sketch the graph of other. The hand-sketch is
approximate, but preserves important elements.
f (x)
Example 3.3 Consider the function in Figure 3.5. Reason about the tangent
x
lines at various points along to sketch the derivative f 0 (x).
Solution. In Figure 3.6 we first sketch a few tangent lines along the graph of
f (x). Focus on the slopes (rather than height, length, or other properties) of Figure 3.5: The graph of a function. We
sketch its derivative.
f (x)
Figure 3.6: A few tangent lines of a func-

x tion.
tangents
2 1 0 −1 −0.5 0 2 3 slopes
the dashes. Copying these lines in a row below the graph, we estimate their f (x)
slopes roughly (approximate numerical values shown).

By convention, we “read” a graph from left to right. Slopes in Fig. 3.6 x
are first positive, then zero, then negative, increase again through zero, and
then positive. There are two locations with zero slope (horizontal dashes). tangents
2 1 0 −1 −0.5 0 2 3 slopes
Next, in Figure 3.7, we these rough values for slopes. Only a few points have
been plotted for f 0 (x), but these trends are clear: the derivative function has f 0 (x)
two zeros, and it dips below the axis between these places. In Figure 3.7 we
emphasize how the original function lines up with its derivative f 0 (x).
x
We have aligned these graphs so that the slope of f (x) matches the value
of f 0 (x) shown directly below. ♦ Figure 3.7: Sketching the derivative of a
function.
Example 3.4 Sketch the derivative of the function shown in Figure 3.8.
f (x) Figure 3.8: In Example 3.4, we sketch the

derivative of this function.
x
Function
Solution. We can get a good rough sketch by simply noting where the slopes f (x)
are positive, negative, or zero. See Figure 3.9 for the entire process. The thin
x
vertical lines demonstrate that f 0 (x) = 0 coincides with tops of hills or bottom
of valleys on the graph of the function f (x).
♦ Tangent lines
f (x)
Constant and linear functions and their derivatives x
Slopes
Example 3.5 (Derivative of y = C) Use a geometric argument to determine
+ 0
–
– 0
–
+
–
0

– 0 +
–
the derivative of the function y = f (x) = C at any point x0 on its graph.
Solution. This function is a horizontal straight line, whose slope is zero Derivative
f 0(x)
everywhere. Thus “zooming in” at any point x, leads to the same result, so the
derivative is 0 everywhere. ♦ x
Example 3.6 (Derivative of y = Bx) Use a geometric argument to determine

the derivative of the function y = f (x) = Bx at any point x0 on its graph.
Figure 3.9: Sketching the derivative of a
function for Example 3.4
Solution. The function y = Bx is a straight line of slope B. At any point on its
graph, it has the same slope, B. Thus the derivative is equal to B at any point
on the graph of this function. ♦
Notice that in the above two examples we have found the derivative for the
two power functions, y = x0 and y = x1 . We summarize:
8. If the equation of the tangent line to
The derivative of any constant function is zero. The derivative of the function
f (x) at x = 1 is y = mx + b, what is
y = x is 1. The derivative of the function y = k · x is k. f 0 (1)?
9. Sketch the graph of the derivative of
the function y = C. (What slope do
Molecular motors you see at each point when you
“zoom in”?
Microtubules (MT) are long, rod-like cellular structures (introduced in
Chapter 2) with both structural and transportation roles in living cells. Human
nerve cells can be up to 1 meter in length, which makes for a challenge to
move material from the cell body - where it is made - to the cell ends where
it is needed for repair or metabolism. Microtubules act like highways for vesicle
kinesin k
“molecular motors”, proteins that “drive” along these routes, transporting the +
necessary cargo.
microtubules
Microtubules have distinct ends (called “plus” and “minus” ends). Some
k
motors specialize in moving towards the + end, while others move towards +
the - end. Figure 3.10 is a schematic diagram of kinesin (represented by the Figure 3.10: The molecular motor kinesin
walks towards a microtubule “plus” end.
letter k), a plus-end directed motor. As shown in the figure, kinesin can hop It can detach and reattach to another
off one MT and onto another MT pointing in the opposite direction. microtubule.
In Example 3.7, we study a sample vesicle track (displacement, y over
time t) and decipher the sequence of motor events that caused that motion.
Example 3.7 (Motion of molecular motors) The displacement y(t ) of a

vesicle is shown in Figure 3.10(a).
(a) Sketch the corresponding instantaneous velocity v(t ) for the vesicle.
(b) Use your sketch to explain what was happening to the kinesin carrying
that vesicle.
Figure 3.11: Molecular motors: (a) The

displacement, y
(a) displacement of the vesicle. (b) The velocity

of the vesicle.
0 t
(b)
velocity, v
t
0
Solution.
(a) The plot in Figure 3.11(a) consists of straight line segments with sharp
corners (cusps). Over each of these line segments, the slope dy/dt (which
corresponds to the instantaneous velocity, v(t )) is constant. Segments
with positive slope correspond to motion towards the right (as in the top
microtubule track in Figure 3.10. Over times where the slope is negative,
the motion is to the left. Where the slope is zero (flat graph), the vesicle
was stationary.
In Figure 3.11(b), we sketch the graph of the instantaneous velocity, v(t ).

Observe that v(t ), which is the derivative of y(t ), is not defined at the
points where y(t ) has “corners”.
(b) Based on Figure 3.11(b), the kinesin motor was moving on a right-facing
microtubule, then hopped onto a left-facing microtubule, and then hopped
back to a right -facing microtubule. For a brief time it was either stuck or
detached from the microtubule tracks (stationary part). Finally, it hopped
onto a left-moving microtubule.
♦
3.2 The analytic view: calculating the derivative
1. Explain the definition of a continuous function.
2. Identify functions with various types of discontinuities.
3. Evaluate simple limits of rational functions.
4. Calculate the derivative of a simple function using the definition of the

derivative, 2.7.
Technical matters: continuous functions and limits

In Chapter 2, we saw functions made up of discrete data points. In Chapter 1
we studied several continuous functions: power, polynomial or rational.
Intuitively, on the graph of a continuous function, every point is connected to
neighbouring points. For example, the power function Eqn. (2.1), y(t ) = ct 2 is
√
continuous for all values of t, whereas a function such as y = x is continuous
and defined as a real number only for x ≥ 0. The function y = 1/(x + 1) is
defined and continuous for x 6= −1 (since division by zero is undefined). Mastered Material Check
We now make the intuitive discussion more precise with a formal defi- 10. Use your favorite graphing software
nition, based on the concept of a limit. We first define what it means for a to verify these statements about the
√
continuity of y = x
function to be continuous, and then show how limits are computed to test that and y = 1/(x + 1).
definition.
Definition 3.3 (Continuous function) We say that y = f (x) is continuous at

a point x = a in its domain if
lim f (x) = f (a).

x→a
By this we mean that the function is defined at x = a, that the above limit exists,
and that it matches with the value that the function takes at the given point.
This definition has two important parts. First, the function should be
defined at the point of interest, and second, the value assigned by the function
has to “fit the local behaviour” in the sense of the limit. This rules out a
“jump” or “break” in the graph. When the above is not true at some point xs ,
we say that the function is discontinuous at xs . We give a few examples to
demonstrate some different types of discontinuities that exist. At the same
time we illustrate how limits are calculated.
Function with a hole in its graph. Consider a function of the form
(x − a)2
f (x ) = .
(x − a)
Then if x 6= a, we can cancel a common factor, and obtain (x − a). At x = a,

the function is not defined because ( 00 ) - which is undefined - results. In short,
we have (
(x − a)2 x−a x 6= a
f (x ) = =
(x − a) undefined x = a.
Even though the function is not defined at x = a, we can still evaluate the limit Mastered Material Check
of f as x approaches a. We write 11. Sketch the graph of the function
(x−1)2
f (x ) = (x−1)
.
(x − a)2
lim f (x) = lim = lim x − a = 0,
x→a x→a (x − a) x→a
and say that “the limit as x approaches a” exists and is equal to 0. We also
say that the function has a removable discontinuity. If we add the point
(a, 0) to the set of points at which the function is defined then we obtain a
continuous function identical to the function x − a. See also Appendix D.
Function with jump discontinuity. Consider the function

(
−1 x ≤ a,
f (x ) =
1 x > a.
We say that the function has a jump discontinuity at x = a. As we approach

the point of discontinuity we observe that the function has two distinct values,
depending on the direction of approach. We formally capture this observation
using right and left hand limits,
lim f (x) = −1, lim f (x) = 1.

x→a− x→a+
Notice we use limx→a− to denote approaching a from the left, and limx→a+
to denote approaching a from the right. Since the left and right limits are
unequal, we say that “the limit does not exist” (abbreviated DNE).
Function with blow up discontinuity. Consider the function
1
f (x ) = .
x−a
Then as x approaches a, the denominator approaches 0, and the value of the
12. What type of discontinuity does
function goes to ±∞. We say that the function “blows up” at x = a and that x2 +4x+4
have?
x+2
the limit, lim f (x), does not exist. 13. What type of discontinuity does
x→a
Figure 3.12 illustrates the differences between functions that are continu- x2 +4x+4
x+4 have?
ous everywhere, those that have a hole in their graph, and those that have a
jump discontinuity or a blow up at some point a.
Examples of limits. We now examine several examples of computations of

limits. More details about properties of limits are provided in Appendix D.
By Definition 3.3, to calculate the limit of any function at a point of
continuity, we simply evaluate the function at the given point.
y continuous y y y
hole jump
blow-up
a x
x x x
a a
Figure 3.12: Left to right: a continuous

function, a function with a removable
Example 3.8 (Simple limit of a continuous function) Find the following discontinuity, a jump discontinuity, and a
function with blow up discontinuity.
limits:
1 x
(a) lim x2 + 2 (b) lim (c) lim .
x→3 x→1 x + 1 x→10 1 + x
Solution. In each case, the function is continuous at the point of interest (at
x = 3, 1, 10, respectively). Thus, we simply “plug in” the values of x in each
case to obtain
1 1 x 10
(a) lim x2 + 2 = 32 + 2 = 11 (b) lim = (c) lim = .
x→3 x→1 x+1 2 x→10 1+x 11
♦
Example 3.9 (Hole in graph limits) Calculate the limits of the following
functions. Note that each has a removable discontinuity (“a hole in its
graph”).
x2 − 6x + 9 x2 + 3x + 2
(a) lim , (b) lim .
x→3 x−3 x→−1 x+1
Solution. We first simplify algebraically by factoring the numerator, and then
14. How do these examples change if
evaluate the limit. Note that the simplification is possible so long as we evalu- the limit approaches a different
ate the limit, rather than the actual function, at the point of discontinuity. value of x?
x2 − 6x + 9 (x − 3)2
(a) lim = lim = lim (x − 3) = 0.
x→3 x−3 x→3 (x − 3) x→3
x2 + 3x + 2 (x + 1)(x + 2)
(b) lim = lim = lim (x + 2) = 1.
x→−1 x+1 x→−1 (x + 1) x→−1
Example 3.10 (Limit involving sin(x)) Use the observation made in Exam-
sin(x)
ple 3.2 to arrive at the value of lim .
x→0 x
Solution. Example 3.2 illustrated the fact that close to x = 0 the function
sin(x) has the following behaviour:
sin(x)
sin(x) ≈ x, or ≈ 1.
x
This is equivalent to the result
sin(x)
lim = 1. (3.1)
x→0 x
We read this “as x approaches zero, the limit of sin(x)/x is 1." This limit is
used in later calculations involving derivatives of trigonometric functions. ♦
Computing the derivative

As discussed in Section 2.5, calculating a derivative requires the use of limits.
To summarize:
1. For the secant line connecting the points x and x + h on the graph of a
function, in the limit h → 0, those points get closer together, leading to a
tangent line.
2. The slope of a secant line is an average rate of change, but in the limit
(h → 0), we obtain the derivative, which is the slope of the tangent line.
We illustrate how to use the definition of the derivative to compute a few

derivatives.
Example 3.11 (Derivative of a linear function) Using Definition 2.7 of the

derivative, compute the derivative of the function y = f (x) = Bx + C.
Solution. We have already used a geometric approach to find the derivative

of related functions in Examples 3.5-3.6. Here we do the formal calculation
as follows: Mastered Material Check
15. What other notations express the
f (x + h) − f (x )
f 0 (x) = lim (start with the definition) derivative of f (x)?
h→0 h
[B(x + h) + C ] − [Bx + C ]
= lim (apply it to the function)
h→0 h
Bx + Bh + C − Bx −C
= lim (expand the numerator)
h→0 h
Bh
= lim (simplify)
h→0 h
= lim B (cancel a factor of h)

h→0
=B (evaluate the limit). (3.2)
Hence, we confirmed that the derivative of f (x) = Bx + C is f 0 (x) = B. This

agrees with the sum of the derivatives of the two parts, Bx and C found in
Examples 3.5-3.6. (Indeed, as we establish shortly, the derivative of the sum
of two functions is the same as the sum of their derivatives.) ♦
Example 3.12 (Derivative of the cubic power function) Compute the

derivative of the function y = f (x) = Kx3 .
i See the steps in a related calculation

Solution. For y = f (x) = Kx3 we have of the derivative of y = f (x) = Kxn .
dy f (x + h) − f (x )
= lim
dx h→0 h
K (x + h)3 − Kx3
= lim
h→0 h
(x3 + 3x2 h + 3xh2 + h3 ) − x3
= lim K
h→0 h
(3x2 h + 3xh2 + h3 )
= lim K
h→0 h
= lim K (3x + 3xh + h2 )
2
h→0
= K (3x2 ) = 3Kx2 . (3.3)
Thus the derivative of f (x) = Kx3 is f 0 (x) = 3Kx2 . ♦
Example 3.13 Use the definition of the derivative to compute f 0 (x) for the
function y = f (x) = 1/x at the point x = 1.
i See the calculation of the derivative
Solution. We write down the formula for this calculation at any point x of y = f (x) = 1/x.
and then simplify algebraically, using common denominators to combine
fractions, and then, in the final step, calculate the limit formally. Lastly we
substitute the value x = 1 to find f 0 (1).
f (x + h) − f (x )
f 0 (x) = lim (the definition)
h→0 h
1 1
(x +h)
− x
= lim (applied to the function)
h→0 h
[x−(x+h)]
x (x +h)
= lim (common denominator)
h→0 h
−h
= lim (algebraic simplification)
h→0 hx(x + h)
−1
= lim (cancel factor of h)
h→0 x(x + h)
1
=− 2 (limit evaluated) (3.4)
x
Thus, the derivative of f (x) = 1/x is f 0 (x) = −1/x2 and at the point x = 1 it
takes the value f 0 (1) = −1. ♦
In Exercise 3. 8 we apply similar techniques to the derivative of the square-
root function to show that
√ 1
y = f (x ) = x ⇒ f 0 (x ) = √ . (3.5)
2 x
In the next chapter, we formalize some observations about derivatives of
power functions and rules of differentiation. This allows us to simplify some
of the calculations involved in finding derivatives.
3.3 The computational view: software to the rescue!
1. Use software to numerically compute an approximation to the derivative.
2. Explain that the approximation replaces a (true) tangent line with an

(approximating) secant line.
3. Explain using words how the derivative shape is connected with the shape
of the original function.
16. Describe how to find the derivative
of a function f (x) at x = x0
4. Interpret the differences between two types of biochemical kinetics:
analytically.
Michaelis-Menten and Hill function.
17. Describe how to find the derivative
of a function f (x) at x = x1
We have explored geometric and analytic aspects of the derivative. Here geometrically.
we show a third aspect of the derivative: its numerical implementation using

a simple spreadsheet. The ideas introduced here reappear in a variety of
problems where repetitive calculations are needed to arrive at a solution.
Definition 3.4 (Numerical derivative) A numerical derivative is an approx-

imation to the value of the derivative, obtained by using a finite value of ∆x,
∆f ∆f
f 0 (x)numerical ≈ rather than the actual value f 0 (x)actual = lim .
∆x ∆x→0 ∆x
The numerical derivative would be a good approximation of the true

derivative provided that ∆x is “small enough”, that is, a step size for which
the function f (x) does not changes dramatically. Since ∆ f is the difference
of two values of f , (∆ f = f (x + ∆x) − f (x)) it follows that the numerical
derivative is the same as the slope of a secant line. This important realization,
associated with the second learning goal in this section, means that a secant
line is often used to approximate a tangent line, and the slope of a secant line
is used to approximate a derivative in numerical computations. We see this
idea again in several contexts.
Derivative of Michaelis-Menten and Hill functions

A spreadsheet can be used to numerically approximate derivatives. We
Hint: Why would we ever use
illustrate this using as examples the reaction speeds for Michaelis-Menten a spreadsheet, when there is other
Eqn. (1.8) and Hill function kinetics, Eqn. (1.7) (see Section 1.5), repeated software for graphing? Basically,
spreadsheets are powerful tools for
below: computation, for manipulating data and
Kc for eliminating repetitive calculations.
vMM = f1 (c) = , (3.6)
a+c We gradually pick up some important
skills in this course.
Kcn
vHill = f2 (c) = . (3.7)
an + cn
(Both f1 (c), f2 (c) are shown as functions of c in Figure 3.13a.) Our goal is to
use a spreadsheet to compute a numerical approximation of the derivative of
these two reaction speeds with respect to the chemical concentration c.
Featured Problem 3.1 (The derivative on a spreadsheet) Use a spread-

sheet (or your favorite software) to plot the derivatives of the functions
vMM = f1 (c), vHill = f2 (c).
6 6
Hill function Hill function
Reaction Rate, v
4 4
dv/dc
Michaelis-Menten Michaelis-Menten
2 2
1 2 3 4 5 1 2 3 4 5
chemical concentration, c chemical concentration, c
(a) (b)
Figure 3.13: (a) A plot of f1 (c), f2 (c)

produced on a spreadsheet. (b) A plot of
Solution. Figure 3.13 shows the results of the spreadsheet calculation, both the functions and their (approximate)
but here we use only the Hill function to demonstrate typical spreadsheet derivatives.
manipulations.
(A) Open your favorite spreadsheet and label columns where data will be kept.
In the link we display a spreadsheet with columns for the step size ∆c,
concentration c, and values of the function f2 (c), with K = 5, a = 1, n = 4.
The last column contains the approximation for the derivative ∆ f /∆c. P Link to Google Sheets. This
spreadsheet shows how to create an
(1) We input the desired step size ∆c, here set to 0.1 (cell A2) approximation for the derivative of the
Hill function in (3.7). Fig 3.13 was
(2) Input the value of c at which to start the calculations. Here we used produced by a similar set of
calculations, see Feature Problem 3.1.
c = 0 as the left endpoint (input 0 into cell B2). Then let the spreadsheet
You can view this sheet and copy it
create the entire set of c values by inputting = B2 + $A$2 in cell B3, elsewhere. You cannot edit it as is.
and dragging the “fill handle” (small square dot on bottom right hand
corner) down the column. Note that the symbols $ are universally used
in spreadsheets to denote an absolute reference to a particular cell,
whereas all other references are relative.
(3) In cell C2 type = 5 ∗ B2 ∧ 4/(1 + B2 ∧ 4) to create the formula for the
desired function. The symbol ∧ denotes a power. This will generate the
first point on the graph to be plotted. Similarly drag the fill handle to
generate the rest of the points f2 (c) corresponding to the all c values.
(B) Next, we compute the desired numerical approximations of the derivatives.
(1) Use column D for the numerical derivative of f2 (c). To do so, approxi-
mate the actual derivative with a finite difference,
∆ f2 d f2
≈ .
∆c dc
Note: importantly, the two expressions are not equal. However, for
sufficiently small ∆c, they approximate one another well.
(2) Dragging the fill handle down the column generates the desired list of
values of the numerical derivative.
(C) Create a chart and plot the results. The x axis is the set of values of c.
Results of the above process (but modified for the two functions and their two
derivatives) lead to the graphs shown on the right panel of Figure 3.13. ♦
Example 3.14 Interpret the graphs of the derivatives in Figure 3.13b in

terms of the way that reaction speed increases as the chemical concentration
is increased in each of Michaelis-Menten and Hill function kinetics
Solution. Both derivatives are positive everywhere, since both f1 (c) and 18. Given time dependent data, can an
f2 (c) are increasing functions. For Michaelis-Menten, the derivative is exact derivative ever be determined?
always decreasing. This agrees with the observation that f1 (c) (thin yellow
curve) gradually levels off and flattens as c increases. While the reaction rate
vMM increases with c, the rate of increase, dv/dc, slows due to saturation at
higher c values.
In contrast, the Hill function derivative starts at zero, increases sharply,
and only then decreases to zero. Correspondingly, the Hill function (thin red
curve) is flat at first, then becomes steeply increasing, and finally flattens to
an asymptote. We can summarize this biochemically by saying that the initial
reaction rate vHill is small and hardly changes near c ∼ 0. For intermediate
range of c, the reaction rate depends sensitively on c (evidenced by large
dv/dc). As c increases to higher values, saturation slows down the rate of
reaction, leading to the drop in dv/dc. ♦
3.4 Summary
1. If we “zoom in” enough on a point x0 on a graph of a function (with

“smooth” behaviour), we see a straight line. This straight line is the tan-
gent line at that point. The slope of this line is the derivative (instanta-
neous rate of change) at that point, x0 .
2. Given the graph of a function f (x), the derivative f 0 (x) can be sketched by
approximating the slopes of the tangent lines of f (x), and plotting those
slopes as points.
3. A function is continuous at x = a in its domain if limx→a f (x) = f (a).

Discontinuous functions might have a hole (removable discontinuity), a
jump, or blow up.
4. Computing derivatives requires the use of limits:
f (x + h) − f (x )
f 0 (x) = limh→0
h
Limits are detailed further in Appendix D. In the absence of analytical
methods, or in the presence of only data, a numerical derivative calculus
can be used to approximate:
f(x) f0 (x)
∆f C 0
f 0 (x)numerical ≈ .
∆x Bx B
Bx + C B
5. Derivatives that were computed in this chapters are summarized in Ta- Kx3 3Kx2
1
− x12
ble 3.1. √x 1
√
x 2 x
6. The applications we encountered in this chapter included:
(a) molecular motors and vesicle transport; and Table 3.1: Some computed derivatives.
(b) Michaelis-Menten and Hill function kinetics for reaction speeds.
1. What geometric characteristic might stop a function from having a derivative?
2. Find the derivative of y = 2x + 1 using a
(a) geometric argument, and

(b) algebraic argument.
3. Draw the derivative of the function f (x) at x = 1, depicted on the graph below.
2
y
1
x
−1 1 2 3
−1
−2
4. Define a continuous function.

Exercises
3.1. Sketching the derivative (geometric view). Shown in Figure 3.14 are
four functions. Sketch the derivative of each of these functions.
y y y y
x x
x x
(a) (b) (c) (d)
Figure 3.14: Figure for Exercise 1; graphs of

four different functions.
3.2. Sketching the function given its derivative. Given the information
in Table 3.2 about the the values of the derivative of a function, g(x), x g 0 (x ) f 0 (x )
sketch a (very rough) graph the function for −3 ≤ x ≤ 3. -3 -1 0
-2 0 +
3.3. What the sign of the derivative tells us. Given the information about -1 2 0
the signs of the derivative of a function, f (x) found in Table 3.2, 0 1 -
sketch a (very rough) graph of the function for −3 ≤ x ≤ 3. 1 0 0
2 -1 +
3.4. Shallower or steeper rise. Shown in Figure 3.15 are two similar 3 -2 0
functions, both increasing from 0 to 1 but at distinct rates. Sketch the
Table 3.2: Derivative data for two different
derivatives of each one. Then comment on what your sketch would functions: g0 (x) (Exercise 2) and f 0 (x)
look like for a discontinuous “step function”, defined as follows: (Exercise 3).
(
0 x<0
f (x ) =
1 x ≥ 0.
y y Figure 3.15: Figure for Exercise 4; two

1 1 similar functions increasing at distinct rates.
0.8 0.8
0.6 0.6
F (x) F (x)
0.4 0.4
0.2 0.2
x x
−4 −2 2 4 −4 −2 2 4
(a) (b)
v
3.5. Introduction to velocity and acceleration.The acceleration of a
particle is the derivative of its velocity. Shown in Figure 3.16 is the
graph of the velocity of a particle moving in one dimension.
Indicate directly on the graph any time(s) at which the particle’s
acceleration is zero.
3.6. Velocity, continued. The vertical height of a ball, d (in meters) t
at time t (seconds) after it was thrown upwards was found to sat-
isfy d (t ) = 14.7t − 4.9t 2 for the first 3 seconds of its motion.
(a) What is the initial velocity of the ball (i.e. the instantaneous veloc- Figure 3.16: Figure for Exercise 5; velocity
ity at t = 0)? of a particle moving in one direction.
(b) What is the instantaneous velocity of the ball at t = 2 seconds?

3.7. Geometric view, continued. Consider Figure 3.17.
(a) Given the function in Figure 3.17(a), graph its derivative.
(b) Given the function in Figure 3.17(b), graph its derivative
(c) Given the derivative f 0 (x) shown in Figure 3.17(c) graph the
function f (x).
(d) Given the derivative f 0 (x) shown in Figure 3.17(d) graph the
function f (x).
Figure 3.17: Figures of functions and

1 10 derivatives for Exercise 7.
y = f (x) y = f (x)
5
0.5
x
2 4 6 8 10
x
0.5 1 1.5 2 −5
−0.5 −10
(a) (b)
10 3
f 0 (x) f 0 (x)
2
5
1
x
2 (c) 4 6 8 10 (d)
x
−1 −0.5 0.5 1
−5
−1
−10 −2
3.8. Computing the derivative of square-root (from the definition).

Consider the function
√
y = f (x) = x.
(a) Use the definition of the derivative to calculate f 0 (x). Consider

using the following algebraic simplification:
√ √ √ √
( a − b)( a + b) a−b
√ √ = √ √ .
( a + b) ( a + b)
(b) Find the slope of the function at the point x = 4.

(c) Find the equation of the tangent line to the graph at this point.
3.9. Computing the derivative. Use the definition of the derivative to
compute the derivative of the function y = f (x) = C/(x + a) where C
and a are arbitrary constants. Show that your result is f 0 (x) = −C/(x +
a)2 .
3.10. Computing the derivative. Consider the function
x
y = f (x ) = .
(x + a)
a
(a) Show that this function can be written as f (x) = 1 − .
(x + a)
(b) Use the results of Exercise 9 to determine the derivative of this
function (note: you do not need to use the definition of the deriva-
tive to do this computation). Show that you get f 0 (x) = (x+aa)2 .
3.11. Molecular motors.
(a) Figure 3.18 (a) shows the displacement of a vesicle carried by
a molecular motor. The motor can either walk right (R), left (L)
along one of the microtubules or it can unbind (U) and be station-
ary, then rebind again to a microtubule. Sketch a rough graph of
the velocity of the vesicle v(t ) and explain the sequence of events
(using the letters R, L, U) that resulted in this motion.
(b) Figure 3.18 (b) shows the velocity v(t ) of another vesicle. Sketch a
rough graph of its displacement starting from y(0) = 0.
y v
t
0
t
0
Figure 3.18: Figure for Exercise 11,
molecular motors.
3.12. Concentration gradient. Certain types of tissues - epithelia - are
made up of thin sheets of cells. Substances are taken up on one side of
the sheet by some active transport mechanism, and then diffuse down
a concentration gradient by a mechanism called facilitated diffusion
c(x)
on the opposite side.
Shown in Figure 3.19 is the concentration profile c(x) of some sub- facilitated
diffusion
stance across the width of the sheet (x represents distance). Sketch the active
corresponding concentration gradient, i.e. sketch c0 (x), the derivative transport
of the concentration with respect to x.

3.13. Tangent line to a simple function. What is the slope of the tangent
line to the function y = f (x) = 5x + 2 when x = 2? when x = 4 ? How x
would this slope change if a negative value of x was used? Why? distance across the sheet
3.14. Slope of the tangent line. Use the definition of the derivative to Figure 3.19: Figure for Exercise 12:
compute the slope of the tangent line to the graph of the function y = concentration profile of a substance.
3t 2 − t + 2 at the point t = 1.
3.15. Tangent line. Find the equation of the tangent line to the graph
of y = f (x) = x3 − x at the point x = 1.5 shown in Figure 3.1. You may
use the fact that the tangent line goes through (1.7, 1.47) as well as the
point of tangency.
3.16. Numerically computed derivative. Consider the two Hill functions
x2 x4
H1 (x) = , H2 (x) =
0.01 + x2 0.01 + x4
(a) Sketch a rough graph of these two functions on the same plot
and/or describe in words what the two graphs would look like.
(b) On a second plot, sketch a rough graph of both derivatives of these
functions and/or describe in words what the two derivatives would
look like.
(c) Using a spreadsheet or your favourite software, plot the two func-
tions over the range 0 ≤ x ≤ 1.
(d) Use the spreadsheet to calculate an approximation for the deriva-
tives H10 (x), H20 (x) and plot these two functions together.
Note: in order to have a reasonably accurate set of graphs, you must
select a small step size of ∆x ≈ 0.01.
3.17. More numerically computed derivatives. As we see in Chapter 14,
trigonometric functions such as sin(t ) and cos(t ) can be used to
describe biorhythms of various types. Here we numerically compute
the first and second derivative of y = sin(t ) and show the relationships
between the trigonometric functions and their derivatives. We use only
numerical methods (e.g. a spreadsheet), but in Chapter 14, we also
study the analytical calculation of the same derivatives.
(a) Use a spreadsheet (or your favourite software) to plot, on the same
graph the two functions
y1 = sin(t ), y2 = cos(t ), 0 ≤ t ≤ 2π ≈ 6.28.

Note that you should use a fairly small step size, e.g. ∆t = 0.01 to
get a reasonably accurate approximation of the derivatives.
(b) Use the same spreadsheet to (numerically) calculate (an approxi-
mate) derivative y01 (t ) and add it to your graph.
(c) Now calculate y001 (t ), that is (an approximation to) the derivative of
the derivative of the sine function and add this to your graph.
4
Differentiation rules, simple antiderivatives and applica-
tions
In Chapter 2 we defined the derivative of a function, y = f (x) by
dy f (x + h) − f (x )
= f 0 (x) = lim .
dx h→0 h
Using this formula, we calculated derivatives of a few power functions. Here,
we gather results so far, and observe a pattern, the power rule. This rule
allows us to compute higher derivatives (e.g. second derivative etc.), to differ-
entiate polynomials, and even to find antiderivatives by applying the rule “in
reverse” (finding a function that has a given derivative). All these calculations
are useful in common applications, including accelerated motion. These
are investigated later in this chapter. We round out the technical material by
stating several other rules of differentiation (product and quotient), allowing
us to easily calculate derivatives of rational functions.
4.1 Rules of differentiation
Learning goals for Section 4.1
1. Express the power rule (Table 4.1) and be prepared to apply it to both
derivatives and antiderivatives of power functions and polynomials.
2. Explain what is meant by the statement that “the derivative is a linear

operation”.
3. Describe the concept of an antiderivative and why it is defined only up to

some constant.
4. Express the product and quotient rules and be able to apply these to
calculating derivatives of products and of rational functions.
The derivative of power functions: the power rule

Function Derivative
We have already computed the derivatives of several of the power functions. f (x ) f 0 (x )
See Example 3.5 for y = x0 = 1 and Example 3.6 for y = x1 . See also 1 0
Example 2.10 for y = x2 and Example 3.12 for y = x3 . We gather these x 1
results in Table 4.1. From the table, we observe that the derivative of a power x2 2x
function is also a power function: the original power becomes a coefficient x3 3x2
and the new power is reduced by 1. We refer to this pattern as the power rule .. ..
. .
of differentiation.
xn nxn−1
We can show that this rule applies for any power function of the form
xn/m (n/m)x(n/m)−1
y = f (x) = xn where n is an integer power. The calculation is essentially the
same as examples illustrated in a previous chapter, but the step of expanding Table 4.1: The Power Rule of differentia-
the binomial (x + h)n entails lengthier algebra. (Such expansion contains terms tion states that the derivative of the power
function y = xn is nxn−1 . For now, we have
of the form xn−k hk multiplied by binomial coefficients, and we include the
established this result for integer n. Later,
details in Appendix E.) From now on, we simply use this result for any power we find this result holds for other powers
function with integer powers (rather than calculating the derivative using its that are not integer.
definition).
Example 4.1 Find the equation of the tangent line to the graph of the power
function y = f (x) = 4x5 at x = 1, and determine the y-intercept of that tangent
line.
Solution. Using the power rule, the derivative of the function is
f 0 (x) = 20x4 .
At the point x = 1, we have dy/dx = f 0 (1) = 20 and y = f (1) = 4. This

means that the tangent line goes through the point (1, 4) and has slope 20.
Thus, its equation is Mastered Material Check
1. Verify that the point (1, 4) satisfies
y−4
= 20 ⇒ y = 4 + 20(x − 1) = 20x − 16. the equation y = 20x − 16.
x−1
Letting x = 0 in the equation of the tangent line, we find that the y-intercept of
line is y = −16. ♦
Example 4.2 (Energy loss and Earth’s temperature) In Section 1.3, we

studied the energy balance on Earth. According to Eqn. (1.5), the rate of loss
of energy from the surface of the Earth depends on its temperature according
to the rule
Eout (T ) = 4πr2 εσ T 4 .
Calculate the rate of change of this outgoing energy with respect to the
temperature T .
Solution. The quantities π, ε, r are constants for this problem. Hence the rate
0 (T ) is
of change (‘derivative’) of energy with respect to T , denoted Eout
0
Eout (T ) = (4πr2 εσ ) · 4T 3 = (16πr2 εσ )T 3 .
D I F F E R E N T I AT I O N RU L E S , S I M P L E A N T I D E R I VAT I V E S A N D A P P L I C AT I O N S 91
Next, we find that the result for derivatives of power functions can be ex-
tended to derivatives of polynomials, using simple properties of the deriva-
tive.
The derivative is a linear operation

The derivative satisfies several convenient properties, among them:
1. the derivative of a sum of two functions is the same as the sum of the
derivatives; and
2. a constant multiplying a function can be brought outside the derivative.
We summarize these rules:
The derivative is a linear operation, that is:
d d f dg
( f (x) + g(x)) = + , (4.1)
dx dx dx
d df Mastered Material Check

C f (x ) = C . (4.2)
dx dx 2. Verify Eqns. (4.1) and (4.2) hold
for f (x) = x3 , g(x) = x2 and C = 4.
3. Give an example which shows
In general, a linear operation L is a rule or process that satisfies two squaring is not a linear operation.
properties:
1. L[ f + g] = L[ f ] + L[g] and
2. L[c f ] = cL[ f ],
where f , g are objects (such as functions, vectors, etc.) on which L acts, and
c is a constant multiplier. We refer to Eqns. (4.1) and (4.2) as the linearity
properties of the derivative.
The derivative of a polynomial

Using the linearity of the derivative, we can extend our differentiation power
rule to compute the derivative of any polynomial. Recall that polynomials are
sums of power functions multiplied by constants. A polynomial of degree n
has the form
p(x) = an xn + an−1 xn−1 + . . . a1 x + a0 , (4.3)
where the coefficients, ai are constant and n is an integer. Hence, the deriva-
tive of a polynomial is just the sum of derivatives of power functions (multi-
plied by constants). Formally, the derivative of Eqn. (4.3) is
dy
p0 (x) = = an · nxn−1 + an−1 · (n − 1)xn−2 + · · · + a1 . (4.4)
dx
(Observe that each term consists of the coefficient times the derivative of a
power function. The constant term a0 has disappeared since the derivative of
any constant is zero.) The derivative, p0 (x), is a function in its own right, and
a polynomial as well. Its degree, n − 1, is one less than that of p(x). In view of
this observation, we could ask: what is the derivative of the derivative?
Notation: we henceforth refer to the “derivative of the derivative” as a
2
second derivative, written in the notation p00 (x) or, equivalently ddx2p .
Using the same rules, we can compute, obtaining 4. Are their other notations for the
second derivative of p(x) that you
d2 p might expect?
p00 (x) = = an n(n − 1)xn−2 + an−1 (n − 1)(n − 2)xn−3 + · · · + a2 . (4.5)
dx2 5. Check that Eqn. (4.7) is an
antiderivative of Eqn. (4.6) by
The following examples should be used for practice. differentiating Eqn. (4.7).
Example 4.3 Find the first and second derivatives of the function
(a) y = f (x) = 2x5 + 3x4 + x3 − 5x2 + x − 2 with respect to x and
(b) y = f (t ) = At 3 + Bt 2 + Ct + D with respect to t.
Solution. We obtain the results
(a) f 0 (x) = 10x4 + 12x3 + 3x2 − 10x + 1 and f 00 (x) = 40x3 + 36x2 + 6x − 10.
(b) f 0 (t ) = 3At 2 + 2Bt + C and f 00 (t ) = 6At + 2B.
In (b) the independent variable is t, but, of course, the rules of differentiation

are the same. ♦
Antiderivatives of power functions and polynomials

Given a derivative, we can ask “what function was differentiated to lead
to this result?” This reverse process is termed antidifferentiation, and the
function we seek is then called an antiderivative. Antidifferentiation reverses
the operation of differentiation. We ask, for example, which function has as
its derivative
y0 (t ) = At n ? (4.6)
The original function, y(t ), should have a power higher by 1 (of the form t n+1 ),
but the “guess” yguess = At n+1 is not quite right, since differentiation results in
A(n + 1)t n . To fix this, we revise the “guess” to
1
y(t ) = A t n+1 . (4.7)
(n + 1)
Question. Is this the only function that has the desired property? No, there
are other functions whose derivatives are the same. For example, consider
adding an arbitrary constant C to the function in Eqn. (4.7) and note that we
obtain the same derivative (since the derivative of the constant is zero). We
summarize our findings:
1
The antiderivative of y0 (t ) = At n is y(t ) = A t n+1 + C. (4.8)
(n + 1)
We also note a similar result that holds for functions in general:

Given a function, f (x) we can only determine its antiderivative up to some
6. What is an example of a constant
(additive) constant. that is not additive?
7. Verify the solution to Example 4.4
We can extend the same ideas to finding the antiderivative of a polyno- by differentiating.
mial. 8. In Example 4.5, identify all:
(a) constants,
Example 4.4 (Antiderivative of a polynomial) Find an antiderivative of the
(b) dependent variables, and
polynomial y0 (t ) = At 2 + Bt + C.
(c) independent variables.
Solution. Since differentiation is a linear operation, we can construct the

antiderivative by antidifferentiating each of the component power functions.
Applying Eqn. (4.8) to the components we get,
1 1
y(t ) = A t 3 + B t 2 + Ct + D,
3 2
where D is an arbitrary constant. We see that the antiderivative of a polyno-
mial is another polynomial whose degree is higher by 1. ♦
Example 4.5 The second derivative of some function is y00 (t ) = c1t + c2 . Find
a function y(t ) for which this is true
Solution. The above polynomial has degree 1. Evidently, this function

resulted by taking the derivative of y0 (t ), which had to be a polynomial of
c1 c1
degree 2. We can check that either y0 (t ) = t 2 + c2t, or y0 (t ) = t 2 + c2t +
2 2
c3 (for any constant c3 ) could work. In turn, the function y(t ) had to be a
polynomial of degree 3. One such function is
c1 3 c2 2
y(t ) = t + t + c3t + c4 ,
6 2
where c4 is any constant. In short, the relationship is:
for differentiation y(t ) → y0 (t ) → y00 (t ),
whereas
for antidifferentiation y00 (t ) → y0 (t ) → y(t ).
♦
These results are used in applications to acceleration, velocity, and dis-
placement of a moving object in Section 4.2.
Product and quotient rules for derivatives

So far, using the power rule, and linearity of the derivative (Section 4.1),
we calculated derivatives of polynomials. Here we state without proof (see
Appendix E), two other rules of differentiation.
The product rule: If f (x) and g(x) are two functions, each differentiable in
the domain of interest, then
d [ f (x)g(x)] d f (x ) dg(x)
= g(x ) + f (x ).
dx dx dx
Another notation for this rule is
[ f (x)g(x)]0 = f 0 (x)g(x) + g0 (x) f (x).

Example 4.6 Find the derivative of the product of the two functions f (x) = x 9. What does ‘domain of interest’
mean?
and g(x) = 1 + x.
10. What is the domain of the
Solution. Using the product rule leads to function f (x) = 1x ?
11. Verify Example 4.6 by
noting f (x)g(x) = x(1 + x) = x + x2
d [ f (x)g(x)] d [x(1 + x)] d [x] d [(1 + x)] and differentiating.
= = · (1 + x ) + ·x
dx dx dx dx
= 1 · (1 + x) + 1 · x = 2x + 1.
The quotient rule: If f (x) and g(x) are two functions, each differentiable in
d f (x ) dg(x)
d f (x ) dx g(x) − dx f (x )

= .
dx g(x) [g(x)]2
We can also write this in the form
f (x ) 0 f 0 (x)g(x) − g0 (x) f (x)

= .
g(x ) [g(x)]2
Example 4.7 Find the derivative of the function y = ax−n = a/xn where a is
a constant and n is a positive integer.
Solution. We can rewrite this as the quotient of the two functions f (x) = a
and g(x) = xn . Then y = f (x)/g(x) so, using the quotient rule leads to the
derivative
dy f 0 (x)g(x) − g0 (x) f (x) 0 · xn − (nxn−1 ) · a
= =
dx [g(x)]2 ( xn ) 2
a n xn−1
= − = a (−n) x−n−1 .
x2n
♦
This calculation shows that the power rule of differentiation holds for
negative integer powers.
Featured Problem 4.1 (Dynamics of actin in the cell) Actin is a structural

protein that forms long filaments and networks in living cells. The actin net-
work is continually assembling from small components (actin monomers) and
disassembling back again. To study this process, scientists attach fluorescent
markers to actin, and watch the fluorescence intensity change over time. In
one experiment, both red and green fluorescent labels were used. The green
label fluoresces only after it is activated by a pulse of light, whereas the red
fluorescent protein is continually active.
It was noted that the red and green fluorescence intensities (R, G) satisfied
the following relationships: Such relationships between a function of
time and its own derivative are examples of
dR dG differential equations, a topic we revisit in
= (a − b)R, = −bG Chapter 11.
dt dt
where a, b are constants that characterize the rate of assembly and disassem-
d (R/G)
bly (“breakup”) of actin. Show that the derivative of the ratio, , can
dt
be expressed in terms of the ratio R/G.
Featured Problem 4.2 (Derivative of the Beverton-Holt function) The

Beverton-Holt model for fish population growth was discussed in Featured
Problem 1.1. A function relating the population of fish this year, y to the
population of fish in the previous year x was
x
y = f (x) = k1
(1 + k2 x)
(where we have simplified the notation, x = N0 , y = N1 of Eqn. 1.9. How
sensitive is this year’s population to slight changes in last year’s population?
Compute the derivative dy/dx to answer this question.
A preview of the chain rule

We give a brief preview of the chain rule, an important rule discussed in
detail in Chapter 8. This rule extends the rules of differentiation to composite
functions, that is, functions made up of applying a sequence of operations
one after the other. For example, the two functions
f (u) = u10 , u = g(x) = 3x2 + 1, (4.9)
could be applied one after the other to lead to the new composite function
10
f (g(x)) = 3x2 + 1 .
The chain rule states that the derivative of this new function with respect to x
is the product of derivatives of the individual functions.
The chain rule

If y = f (z) and z = g(x) are two functions, then the derivative of the composite
function f (g(x)) is
dy d f dz
= .
dx dz dx
Example 4.8 Use the chain rule to differentiate the composite function
10
y = f (g(x)) = 3x2 + 1 .
Solution. The functions being composed are the same as in Eqn. (4.9).
Applying the chain rule gives:
dy d f du
= = 10u9 · 6x = 60(3x2 + 1)9 x.
dx du dx
♦
The details of how to use the chain rule, and many applications are post-
poned to Chapter 8.
The power rule for fractional powers

Using the definition of the derivative, we have already shown that the deriva-
√
tive of x is y0 (x) = 2√
1
x
(see Exercise 2.8). We restate this result using
√
fractional power notation. Recall that x = x1/2 .
√
The derivative of y = x is y0 (x) = 12 x−1/2 .
This idea can be generalized to any fractional power. Indeed, we state here a
result (to be demonstrated in Chapter 9).
Derivative of fractional-power function:

dy m m
The derivative of y = f (x) = xm/n is = x( n −1) .
dx n
Notice that this appears in the last row of Table 4.1.
Example 4.9 (Energy loss and Earth’s temperature, revisited) In Exam-

ple 4.2, we calculated the rate of change of energy lost per unit change in the
Earth’s temperature based on Eqn. (1.5). Find the rate of change of Earth’s
temperature per unit energy loss based on the same equation.
Solution. We are asked to find dT /dEout . We first rewrite the relationship

to express T as a function of Eout . To do so, we solve for T in Eqn. (1.5),
obtaining
1/4 1/4
Eout 1 1/4 1/4
T= = Eout = KEout .
4πr2 εσ 4πr2 εσ
The first term is a constant, and we use the rule for a derivative of a fractional
power to compute that
1/4
dT 1 1 −3/4
= · ·E .
dEout 4πr2 εσ 4 out
Note: in Chapter 9 we will show that implicit differentiation can be used to

find the desired derivative without first solving for the variable of interest. ♦
4.2 Application of the second derivative to acceleration
1. Recognize that velocity and acceleration are first and second derivatives of
position with respect to time (and that velocity and position are first and
second antiderivatives of acceleration.)
2. Given constant acceleration, be able to find the velocity and displacement

of the moving object.
We have already defined the “derivative of a derivative” as the second

derivative. Here we provide a natural example of a second derivative, the
acceleration of an object: the rate of change of the velocity (which is, as we
have seen, the rate of change of a displacement). We restate these relation-
ships in terms of antiderivatives below.
Position, velocity, and acceleration

Consider an object falling under the force of gravity. Let y(t ) denote the
position of the object at time t. From now on, we refer to the instantaneous
velocity at time t simply as the velocity, v(t ).
Definition 4.1 (The velocity) Given the position of some particle as a

function of time, y(t ), we define the velocity as the rate of change of the
position, i.e. the derivative of y(t ):
dy
v(t ) = = y0 (t ).
dt
In general, v may depend on time, which we indicated by writing v(t ).
12. Give three different examples of
possible units for velocity.
Definition 4.2 (The acceleration) We define the acceleration as the (instan-
13. Give three different examples of
taneous) rate of change of the velocity, i.e. as the derivative of v(t ). possible units for acceleration.
dv
a(t ) = = v0 (t )
dt
(acceleration could also depend on time, hence a(t )).
Since the acceleration is the derivative of the derivative of the original

function, we also use the notation
d2y

d dy
a(t ) = = 2 = y00 (t ).
dt dt dt
Note: we have used three equivalent ways of writing a second derivative.

This notation evolved for historical reasons, and is used interchangeably in
science. Acceleration is a second derivative of the position.
Given a(t ), the acceleration as a function of t, we can use antidifferenti-
ation to obtain the velocity v(t ). Similarly, we can use the velocity v(t ) to
determine the position y(t ) (up to some constant). The constants must be
obtained from other information, as the next examples illustrate.
Example 4.10 (Uniformly accelerated motion of a falling object) Suppose

that the acceleration of an object is constant in time, i.e. a(t ) = −g = constant.
Use antidifferentiation to determine the velocity and the position of the object
as functions of time.
i Video summary of how to compute
Note: here we have a coordinate system in which the positive direction is the velocity and displacement given
uniform acceleration.
“upwards”, and so the acceleration, which is in the opposite direction, is
negative. On Earth, g = 9.8 m /s2 .
Solution. First, what function of time v(t ) has the property that
a(t ) = v0 (t ) = −g = constant?
The function a(t ) = −g is a polynomial of degree 0 in the variable t, so the

velocity, which is its antiderivative, has to be a polynomial of degree 1, such
as v(t ) = −gt. This is one antiderivative of the acceleration. Other functions,
such as
v(t ) = −gt + c, (4.10)
would work for any constant c.
How would we pick a value for the constant c? We need additional infor-
mation. Suppose we are told that v(0) = v0 is the known initial velocity. Later
we refer to the statement v(0) = v0 as an “initial condition,” since it specifies
how fast the object was moving initially at t = 0. Then, substituting t = 0 into
Eqn. (4.10), we find that c = v0 . Thus in general,
v(t ) = −gt + v0
where v0 is the initial velocity of the object.

Next, let us determine the position of the object as a function of t, that
is, y(t ). Recall that v(t ) = y0 (t ). Thus, antidifferentiation leads to a polyno-
mial of degree 2,
1
y(t ) = − gt 2 + v0t + k, (4.11)
2
where, as before we allow for some additive constant k. Reasoning as before, Mastered Material Check
we can determine the value of the constant k from a known initial position of 14. Verify that the derivative
the object y(0) = y0 . As before (plugging t = 0 into Eqn. (4.12)), we find that of y(t ) = −12gt 2 + v0 t + k is the
given expression for v(t ).
k = y0 , so that
1
y(t ) = − gt 2 + v0t + y0 . (4.12)
2
Here we assumed that the acceleration is due to gravity, −g, but any other
motion with constant acceleration would be treated in the same way. ♦
Summary, uniformly accelerated motion: If an object moves with constant

acceleration −g, then given its initial velocity v0 and initial position y0 at
y
time t = 0, the position at any later time is given by:
1
y(t ) = − gt 2 + v0t + y0 .
2
This powerful conclusion is a direct result of
1. the assumption that the acceleration is constant, t

2. the elementary rules of calculus, and v
3. the definitions of velocity and acceleration as first and second derivatives

of the position. t
We further explore the relationships between position, velocity and accelera-
tion of a falling object in the examples below.
Example 4.11 Determine when the object reaches its highest point, and its
velocity at that time.
a
t
Solution. When the object reaches its highest point, its velocity has de-
creased to zero. From then on, the velocity becomes negative and the object
falls back down. We solve for t in the equation v(t ) = 0:
v0
v(t ) = v0 − gt = 0 ⇒ ttop = .
g
♦
Example 4.12 When does the object hit the ground and with what velocity?
Figure 4.1: The position, velocity, and
acceleration of an object that is thrown
Solution. Since y is height above ground, the object hits the ground when upwards and falls under the force of gravity.
y = 0. Then we must solve for t in the equation y(t ) = 0. This turns out to be a
quadratic equation: ` Use the sliders to see how the initial
q velocity v0 and initial height y0 affect
the time to reach the top and the time to
1 v0 ± v20 + 2gy0 reach the ground on this interactive
y(t ) = gt 2 − v0t − y0 = 0, ⇒ tground = .
2 g graph.
We are interested in a solution with t ≥ 0, so, rejecting the negative root, i Video summary of how we find the
time that the object reaches its highest
point and the time that it hits the
ground.
q
v0 v20 + 2gy0
tground = + .
g g
The velocity of the object when it hits the ground, v(tground ), is then Mastered Material Check
 q  15. Answer Examples 4.11 and 4.12

v20 + 2gy0 when g = 5 m/s2 , v0 = 20 m/s
v0
q
v(tground ) = v0 − gtground = v0 − g  +  = − v2 + 2gy0 . and y0 = 100 m.
0
g g 16. What is the dashed line in Figure 4.1
indicating?
We observe that this velocity is negative, indicating (as expected) that the
object is falling down. Figure 4.1 illustrates the relationship between the
displacement, velocity and acceleration as functions of time. ♦
i A cheetah chasing a gazelle. Who
Featured Problem 4.3 (How to outrun a cheetah) A cheetah can run at will win the race?
top speed vc but it has to slow down (decelerate) to keep from getting too hot.
Assume that the cheetah has a constant rate of (negative) acceleration a = −ac .
A gazelle runs at a slower speed vg but it can maintain that constant speed
for a long time. If the gazelle is initially a distance d from the cheetah, and
running away, when would it be caught? Is there a (large enough) distance d
such that it never gets caught?
4.3 Sketching the first and second derivative and the anti-derivatives
1. Given a sketch of a function, sketch its first and second derivatives.
2. Given the sketch of a function, sketch its first and second antiderivatives.
We continue to practice sketching derivatives (first and second) as well

as antiderivatives, given a sketch of the original function. Here, we aim for
qualitative features, where only the most important aspects of the graphs
(locations of key points such as peaks and troughs) are indicated.
Example 4.13 (Sketching the derivative from the original function)

Sketch the first and second derivatives of the functions in Fig. 4.2.
y y Figure 4.2: Function graphs for Exam-

ple 4.13; one smooth, the other with cusps.
t t
Solution. In Figure 4.3 we show the functions y(t ) (top row), their first
derivatives y0 (t ) (middle row), and the second derivatives y00 (t ) (bottom row).
In each case, we determined the slopes of tangent lines as a first step.
(a) (b) Figure 4.3: Solutions for Example 4.13: a

y y sketch of first and second derivatives.
t t
0 + ++ + 0 0 + 0
y0 y0
t t
0 + 0− − 0 0 0 0
y 00
00
y
t t
17. Identify any cusps in Figure 4.2.
18. What is each dotted line indicating
in Figure 4.3?
Along flat parts of the graph), the derivative is zero. This is indicated at
19. What is the arrow in the middle of
several places in Fig. 4.3. In (b), there are cusps at which derivatives are not Figure 4.3 indicating?
defined. ♦
Example 4.14 (Sketching a function from a sketch of its derivative)

Sketch the antiderivatives y(x) for each derivative y0 (x) shown in Figure 4.4.
y0 y0 Figure 4.4: Derivatives y0 (x) for Exam-

ple 4.14.
x x
Solution. An antiderivative is only defined up to some (additive) constant.

In the bottom panels of Figure 4.5 we show sample antiderivatives for each
case. If we are given additional information, for example that y(0) = 0, we
could then select one specific curve out of this family of solutions. A second
point to observe is that antidifferentiation smoothes a function. Even though
y0 (x) has cusps in (b), we find that y(x) is smooth. We later see that the points
at which y0 (x) has a cusp correspond to places where the concavity of y(x)
changes abruptly.
y0 y0 Figure 4.5: Using the sketch of two func-

tions y0 (x) to sketch their antiderivative,
y(x ).
x x
+ 0 − − − 0 + + 0 − − 0 +
y y
–– –– –– ––
x x
4.4 Summary
1. The derivative is a linear operation, which means that the derivative

can be distributed over a sum of functions or exchanged with a constant
multiplying a function. The derivative of a polynomial is thus a sum of
derivatives of power functions.
2. A "second derivatives" is the derivate of a derivative.
3. Antiderivatives reverse the process of differentiation. Antidifferentiation

is, however, not unique. Given a function f (x), we can only determine its
antiderivative up to some arbitrary constant.
4. The product, quotient and chain rules are,
[ f (x)g(x)]0 = f 0 (x)g(x) + g0 (x) f (x),

0
f (x ) f 0 (x)g(x) − g0 (x) f (x)

= ,
g(x ) [g(x)]2
[ f (g(x))]0 = f 0 (g(x)) · g0 (x).
These allow us to compute derivatives of functions made up of simpler

components for which we have already established differentiation rules.
5. Given the graph of a function, we can use qualitative features - signs,

zeros, peaks, valleys, and tangent lines slopes to sketch both its derivatives
and antiderivatives.
6. The applications examined in this chapter included:
(a) energy loss and Earth’s temperature; and

(b) position, velocity and acceleration of an object.
In particular, these are related via differentiation and antidifferentiation.

Given an object’s position y(t ), its velocity is v(t ) = y0 (t ) and its accelera-
tion is a(t ) = v0 (t ) = y00 (t ).
1. Which of the following operations are linear?
(a) division
(b) exponentiation
(c) composition
(d) squaring
2. Why are the product and chain rules useful?
3. Suppose an object has acceleration a(t ) = 10m/s2 . What can you say about its:
(a) accelleration at t = 5s?
(d) position at time t?
(b) velocity at time t?
(e) position at time t = 5s?
(c) velocity at time t = 5s?
4. Consider the following graph which describes the position y of an object at time t:
30 y
20
10
t
1 2 3 4 5
Where is the object’s velocity minimal? Maximal?

Exercises
4.1. First derivatives. Find the first derivative for each of the following
functions.
(a) f (x) = (2x2 − 3x)(6x + 5),
(b) f (x) = (x3 + 1)(1 − 3x),
(c) g(x) = (x − 8)(x2 + 1)(x + 2),
(d) f (x) = (x − 1)(x2 + x + 1),
x2 − 9
(e) f (x) = ,
x2 + 9
2 − x3
(f) f (x) = ,
1 − 3x
b3
(g) f (b) = 2 ,
2−b3
m2
(h) f (m) = − (m − 2)(2m − 1),
3m − 1
(x2 + 1)(x2 − 2)
(i) f (x) = .
3x + 2
4.2. Logistic growth rate. In logistic growth, the rate of growth of a
population, R depends on the population size N as follows:

N
R = rN 1 − ,
K
where r and K are positive constants. Find the rate of change of

the growth rate with respect to the population size, that is com-
pute dR/dN.
Note: Logistic growth rate is further explored in Chapter 7.
4.3. Michaelis-Menten and Hill kinetics. Compute the derivatives of the
following functions:
(a) The Michaelis Menten kinetics of Eqn. (1.8),
Kx
v= .
kn + x
(b) The Hill function of Eqn. (1.7), that is
Axn
y= .
an + x n
4.4. Volume, surface area and radius of a sphere. The volume and
surface area of a sphere both depend on its radius:
4
V = πr3 , S = 4πr2 .
3
(a) Find the rate of change of the volume with respect to the radius and
the rate of change of the surface area with respect to the radius.
(b) Find the rate of change of the surface area to volume ratio S/V
with respect to the radius.
4.5. Derivative of volume with respect to surface area. Consider the vol-
ume and surface area of a sphere. (See Exercise 4 for the formulae.)
(a) Eliminate the radius and express V as a function of S.
(b) Find the rate of change of the volume with respect to the surface
area.
4.6. Surface area and volume of a cylinder. The volume of a cylinder
and the surface area of a cylinder with two flat end-caps are
V = πr2 L, S = 2πrL + 2πr2
where L is the length and r the radius of the cylinder.

(a) Find the rate of change of the volume and surface area with respect
to the radius, assuming that the length L is held fixed.
(b) Find the rate of change of the surface area to volume ratio S/V
with respect to the radius assuming that the length L is held fixed.
4.7. Growing circular colony. A bacterial colony has the shape of a
circular disk with radius r (t ) = 2 + t/2 where t is time in hours and r
is in units of mm. Express the area of the colony as a function of time
and then determine the rate of change of area with respect to time
at t = 2hr.
4.8. Rate of change of energy during foraging. When a bee forages for
nectar in a patch of flowers, it gains energy. Suppose that the amount
of energy gained during a foraging time span t is
Et
f (t ) = , where E, k > 0 are constants.
k +t
(a) If the bee stays in the patch for a very long time, how much energy
can it gain?
(b) Use the quotient rule to calculate the rate of energy gain while
foraging in the flower patch.
Note: foraging is returned to in Chapter 7.
4.9. Ratio of two species. In a certain lake it is found that the rate of
change of the population size of each of two species (N1 (t ), N2 (t )) is
proportional to the given population size. That is
dN1 dN2
= k1 N1 , = k2 N2 ,
dt dt
where k1 and k2 are constants. Find the rate of change of the ratio
d (N1 /N2 )
of population sizes (N1 /N2 ) with respect to time . Your
dt
answer should be in terms of k1 , k2 and the ratio N1 /N2 .
4.10. Invasive species and sustainability. An invasive species is one that

can out-compete and grow faster than the native species, resulting
in takeover and displacement of the local ecosystem. Consider the
two-lake system of Exercise 9. Suppose that initially, the ratio of the
native species N1 to the invasive species N2 is very large. Under what
condition (on the constants k1 , k2 ) does that ratio decrease with time,
i.e. does the invasive species take over?
4.11. Numerical derivatives. Consider the function
y(x) = 5x3 , 0 ≤ x ≤ 1.
(a) Use a spreadsheet (or your favourite software) to compute an

approximation of the derivative of this function over the given
interval for ∆x = 0.25 and compare to the true derivative, using the
power rule. Comment on the comparison.
(b) Recompute the approximation to the derivative using ∆x = 0.05 and
comment on the results.
4.12. Antiderivatives. Find antiderivatives of the following functions, that
is find y(t ).
(a) y0 (t ) = t 4 + 3t 2 − t + 3.
√
(b) y0 (x) = −x + 2.
(c) y0 = |x|.
4.13. Motion of a particle. The velocity of a particle is known to depend on
time according to the relationship
v(t ) = A − Bt 2 , A, B > 0 constants
(a) Find the acceleration a(t ).

(b) Suppose that the initial position of the particle is y(0) = 0. Find the
position at time t.
(c) At what time does the particle return to the origin?
(d) When is the particle farthest away from the origin?
(e) What is the largest velocity of the particle?
4.14. Motion of a particle. The position of a particle is given by the func-
tion y = f (t ) = t 3 + 3t 2 .
(a) Find the velocity and acceleration of the particle.
(b) A second particle has position given by the function y = g(t ) =
at 4 + t 3 where a is some constant and a > 0. At what time(s) are the
particles in the same position?
(c) At what times do the particles have the same velocity?
(d) When do the particles have the same acceleration?
4.15. Ball thrown from a tower. A ball is thrown from a tower of height h0 .
The height of the ball at time t is
h(t ) = h0 + v0t − (1/2)gt 2

where h0 , v0 , g are positive constants.
(a) When does the ball reach its highest point?
(b) How high is it at that point?
(c) What is the instantaneous velocity of the ball at its highest point ? f 0(x)
4.16. Sketching a graph. Sketch the graph of a function f (x) whose deriva-
x
tive is shown in Figure 4.6. Is there only one way to draw this sketch?
What difference might occur between the sketches drawn by two
different students? Figure 4.6: Figure for Exercise 16; sketch-
ing an antidertivative.
4.17. Sketching the second derivative. Given the derivative f 0 (x ) shown in
Figure 4.7, graph the second derivative f 00 (x).
10 Figure 4.7: Figure for Exercise 17; sketch-

f 0 (x) ing a derivative
5
x
2 4 6 8 10
−5
−10
4.18. Sketching graphs. Each of the graphs in Figure 4.8 depict the deriva-
tive of a function. Use these to sketch the corresponding antideriva-
tives and derivatives.
f 0 (x) Figure 4.8: Derivative graphs for Exer-

0.5 cise 18
0.4
0.3
d
dx F (x)
x 0.2
0.1
x
−4 −2 2 4
(a) (b)
5
Tangent lines, linear approximation, and Newton’s method
In Chapter 3, we defined the tangent line as the line we see when we zoom
into the graph of a (continuous) function, y = f (x), at some point. In much the
same sense, the tangent line approximates the local behaviour of a function
near the point of tangency., x0 . Given x0 , y0 = f (x0 ), and the slope m =
f 0 (x0 ) (the derivative), we can find the equation of the tangent line Mastered Material Check
1. What are the slope and y-intercept
rise y − y0
= = m = f 0 (x0 ) of the generic tangent line (given in
run x − x0 Eqn. (5.1))?
⇒ y = f (x0 ) + f 0 (x0 )(x − x0 ). (5.1)
(See Appendix A for a review of straight lines.)

We use Eqn. (5.1) in several applications, including linear approximation,
a method for estimating the value of a function near the point of tangency.
A further application of the tangent line is Newton’s method which locates
zeros of a function (values of x for which f (x) = 0).
5.1 The equation of a tangent line
1. Given a simple function y = f (x) and a point x, be able to find the equation
of the tangent line to the graph at that point.
2. Graph both a function and its tangent line using a spreadsheet or your
favorite software.
In the following examples, the equation of the tangent line is easily found.
Example 5.1 (Tangent to a parabola) Find the equations of the tangent

lines to the parabola y = f (x) = x2 at the points:
a) x = 1 and x = 2 (“Line 1” and “Line 2”).
b) Determine whether these tangent lines intersect, and if so, where.

Solution. The slopes of a tangent line is a derivative, which in this case is

f 0 (x) = 2x.
a) This means m1 = f 0 (1) = 2 · 1 = 2 (for Line 1) and m2 = f 0 (2) = 2 · 2 = 4

(for Line 2). The points of tangency are on the curve (x, x2 ). Thus these
are (1, 1) for Line 1 and (2, 4) for Line 2. With the slope and a point for 6 y
y = x2
each line, we find that Point of
tangency 2
y−1 4
Line 1: = m1 = 2, ⇒ y = 1 + 2(x − 1) ⇒ y = 2x − 1, intersection point
x−1
y−4 Tangent
Line 2: = m2 = 4 ⇒ y = 4 + 4(x − 2) ⇒ y = 4x − 4. 2 line 1
x−2 Point of
tangency 1
Tangent line 2
x
b) Two lines intersect if their y values (and x values) are the same. Equating y 0.5 1 1.5 2 2.5
values and solving for x, we get
Figure 5.1: The graph of the parabola
3 y = f (x) = x2 and its tangent lines at x = 1
2x − 1 = 4x − 4 ⇒ −2x = −3 ⇒ x= .
2 and x = 2. See Example 5.1 for the equations
and point of intersection of these tangent
Hence, the two tangent lines intersect at x = 3/2 as shown in Fig 5.1. ♦ lines.
2. Under what circumstance do two
The next example illustrates how a tangent line can be used to approxi- lines not intersect?
mate the zero of a function. This idea is developed into a useful approxima-
tion method called Newton’s method in Section 5.4.
` Manipulate the slider to see the
Example 5.2 a) Draw the graph of the function y = f (x) = x3 − x together tangent line at various points on the
graph of this function. Here D(x)
with its tangent line at the point x = 1.5. represents the derivative, D(x) = d f /dx,
and x0 is the point of tangency.
b) Where does that tangent line intersect the x-axis? You can zoom in or out, change the
range of the slider, or try a different
function.
c) Compare that point of intersection with a zero of the function.
Solution.
a) The function is f (x) = x3 − x, its derivative is f 0 (x) = 3x2 − 1, and the 6 y

y = x3 − x
point of interest is (x, f (x)) = (1.5, 1.875). A tangent line at x = 1.5 has 4
slope f 0 (1.5) = 3(1.5)2 − 1 = 5.75, so its equation is Tangent line
2
x
−2 −1 1 2
y − 1.875
= 5.75 ⇒ y = 1.875 + 5.75(x − 1.5) −2
x − 1.5
⇒ y = 5.75x − 6.75. −4 intersection
point
−6
The function and this tangent line are shown in Figure 5.2.
Figure 5.2: The graph of the function y =
b) The tangent line intersects the x-axis when y = 0, which occurs at f (x) = x3 − x is shown in black, together
with its tangent line at the point x = 1.5. The
6.75 point at which the TL intersects the x-axis is
0 = 5.75x − 6.75 ⇒ x= = 1.174. a (rough) approximation of a nearby zero of
5.75
the function.
TA N G E N T L I N E S , L I N E A R A P P ROX I M AT I O N , A N D N E W T O N ’ S M E T H O D 111
c) A true zero of the function is a value of x for which x3 − x = 0, and the

one closest to the result in (b) is x = 1. Thus the point (1.174, 0) where the
tangent line intersects the x-axis is close but not quite equal to (1, 0). (But
Newton’s method will help us fix this gap.) ♦
In many cases, directly solving for roots of functions (as in Example 5.2c)
is not possible. In Section 5.4 we discuss how a repetition of this idea can be
used to refine the approximation of a zero using Newton’s method.
Example 5.3 a) Find the equation of the tangent line to y = f (x) = x3 − ρx

for ρ > 0 constant, at the point x = 1.
b) Find where that tangent line intersects the x-axis. Mastered Material Check
3. How many zeros does the function
Solution. This is the same type of calculation, but the constant, ρ - chosen depicted in Figure 5.2 have?
to develop familiarity with alternative constant choices - makes the example 4. Check that the tangent line goes
slightly less straightforward. through the desired point and has
the slope we found. One way to do
a) The derivative of f (x) = x3 − ρx is f 0 (x) = 3x2 − ρ so at x = 1, the slope is this is to pick a simple value for ρ,
e.g. ρ = 1 and do a quick check that
m = f 0 (1) = 3 − ρ. The point of tangency is (1, f (1)) = (1, 1 − ρ ). Then, the answer matches what we have
the equation of the tangent line is found.
y − (1 − ρ ) 5. The following graph

√
= 3 − ρ. ⇒ y = (3 − ρ )(x − 1) + (1 − ρ ) depicts f (x) = x on the
x−1 interval [0, 10]. Draw its tangent line
⇒ y = (3 − ρ )x − 2. at the point x = 4.
4
b) To find the point of intersection, set y = (3 − ρ )x − 2 = 0 and solve for x to

2
obtain x = 2/(3 − ρ ).
1
♦
2 4 6 8 10
Example 5.4 Find the equation of the tangent line to the function y = f (x) =
√
x at the point x = 4.
√
Solution. By Eqn. (3.5), the derivative of y = f (x) = x is
1
f 0 (x ) = √ .
(2 x )
√
At x = 4, the slope is f 0 (4) = 1/(2 4) = 1/4 and the point of tangency
√
is (4, 4) = (4, 2). Given this point and the slope, we calculate that the
ladd
tangent line is:

er
y−2
= 0.25 ⇒ y = 2 + 0.25(x − 4).
x−4
anthill
♦
Featured Problem 5.1 (Shortest ladder) In Fig. 5.3, what is the shortest ground
ladder that you can use to reach the window at 6 meters height in the tower
if there is an anthill in the way? Assume that the equation of the anthill is
y = f (x) = −x2 + 3x + 2. Figure 5.3: The ladder (red line) must
reach the window at height 6m. What is the
shortest ladder that avoids the anthill whose
equation is y = f (x) = −x2 + 3x + 2?
5.2 Generic tangent line equation and properties
1. Explain the generic form of the tangent line equation (5.1) and be able to
connect it to the geometry of the tangent line.
2. Find the coordinate of the point at which the tangent line intersects the
x-axis (important for Newton’s Method later on in Section 5.4).
y
tangent line
Generic tangent line equation f (x)
We can find the general equation of a tangent line to an arbitrary function

f (x) at a point of tangency x0 . (The result is Eqn. (5.1).)
Shown in Figure 5.4 is a continuous function y = f (x), assumed to be
differentiable at some point x0 where a tangent line is attached. We see:
x
x0
1. The line goes through the point (x0 , f (x0 )).
Figure 5.4: The graph of an arbitrary
2. The line has slope given by the derivative evaluated at x0 , that is, function y = f (x) and a tangent line at
x = x0 . The equation of this generic tangent
m = f 0 (x0 ). line is Eqn. (5.2).
Then from the slope-point form of the equation of a straight line, Mastered Material Check
6. Circle the point (x0 , f (x0 )) on
y − f ( x0 ) Figure 5.4.
= m = f 0 ( x0 ) .
x − x0 7. Circle where the tangent line
depicted in Figure 5.4 would cross
Rearranging and eliminating the notation m, we have the desired result. the x-axis.
Summary, Tangent Line equation: The equation of a tangent line at x = x0 i Quick video with a derivation of the
to the graph of the differentiable function y = f (x) is generic equation of a tangent line.
y = f (x0 ) + f 0 (x0 )(x − x0 ). (5.2)
Where a generic tangent line intersects the x-axis

From the generic tangent line equation (5.2) we can determine the (generic)
coordinate at which it intersects the x-axis. The result is key to Newton’s
method for approximating the zeros of a function, explored in Section 5.4.
i The calculations for Example 5.5.
Example 5.5 Let y = f (x) be a smooth function, differentiable at x0 , and We show how to find the coordinate x1
where a tangent line intersects the x
suppose that Eqn. (5.2) is the equation of the tangent line to the curve at x0 . axis.
Where does this tangent line intersect the x-axis?
Solution. At the intersection with the x-axis, we have y = 0. Plugging this into
Eqn. (5.1) leads to
f (x0 )
0 = f (x0 ) + f 0 (x0 )(x − x0 ) ⇒ (x − x0 ) = −
f 0 ( x0 )
f (x0 )
⇒ x = x0 − 0 .
f (x0 )
Thus the desired x coordinate, which we refer to as x1 is 8. Where is x1 (as described in

Eqn. (5.3)) on Figure 5.4.?
f ( x0 )
x1 = x0 − . (5.3)
f 0 (x0 )
5.3 Approximating a function by its tangent line
1. Describe why a tangent line approximates the behaviour of a function

close to the point of tangency.
2. Use a tangent line to find a linear approximation to a value of a given

function at some point.
3. Determine whether a linear approximation overestimates or underesti-

mates the value of the function.
We have seen that the tangent line approximates the local behaviour of
a function, at least close enough to the point of tangency. Here we use this
idea in a formal procedure called linear approximation. The idea is to chose
a point (often called the base point) where the value of the function and its
derivative are known, or are easy to calculate, and use the tangent line at that
point to estimate values of the function in the vicinity. Specifically,
1. The generic equation of the tangent line to y = f (x) at x0 is given by

Eqn. (5.2). That line approximates the behaviour of the function close
to x0 , and leads to the so-called linear approximation of the function:
y = f (x0 ) + f 0 (x0 )(x − x0 ) ≈ f (x)
⇒ f (x) ≈ f (x0 ) + f 0 (x0 )(x − x0 ).
2. The approximation is exact at x = x0 , and holds well provided x is close to

x0 . (The expression on the right hand side is precisely the value of y on the
tangent line at x = x0 ).
Example 5.6 Use the fact that the derivative of the function f (x) = x2 is
f 0 (x) = 2x to find a linear approximation for the value (10.03)2 .
Solution. Without the aid of calculator, we know that value of f (x) at the
nearby point x = 10 is 102 = 100. The derivative is f 0 (x) = 2x, so at x = 10
the slope of the tangent line is f 0 (10) = 20. The equation of the tangent line
directly provides the linear approximation of the function.
y − 100
= 20 ⇒ y = 100 + 20(x − 10),
x − 10
⇒ f (x) ≈ 100 + 20(x − 10).
On the tangent line, the value of y corresponding to x = 10.03 is Mastered Material Check
9. Use the linear approximation for x2
f (10.03) ≈ y = 100 + 20(10.03 − 10) = 100 + 20(0.03) = 100.6 found in Example 5.6 to
approximate (9.97)2 .
which is our approximation to the value of the original function. This com- 10. Use the linear approximation
pares well with the calculator value f (10.03) =100.6009. ♦ for sin(x) found in Example 5.7 to
approximate sin(−0.05).
Example 5.7 (Approximating the sine of a small angle) Use a linear 11. Can you think of an example of a
function whose linear approximation
approximation to find a rough value for sin(0.1). is exact for all values?
12. Do you know what concave means?
Solution. We have not yet discussed finding derivatives of trigonometric
functions, but recall from Example 3.2 that close to x = 0 the function y =
sin(x) is well approximated by its tangent line, y = x. Hence, the linear
approximation of y = sin(x) near x = 0 is sin(x) ≈ x (provided x is in
radians, to be discussed in Chapter 14). Thus, at x = 0.1 radians, we find
that sin(0.1) ≈ 0.1, close to its true value of sin(0.1) = 0.09983 (found using
a calculator). ♦
Accuracy of the linear approximation

Example 5.8 (Over or underestimate?) For each of Examples 5.6 and 5.7,
determine whether the linear approximation over or underestimates the true
value of the function.
Solution. In Figure 5.5(a,b), we show the functions and their linear approx-
imations. In (a) we see that the tangent line to y = x2 at x = 10 is always
underneath the graph of the function, so a linear approximation underesti-
mates the true value of the function.
In (b), we see that the tangent line to y = sin(x) at x = 0 is above the
graph for x > 0 and below the graph for x < 0. This shows that the linear
approximation is larger than (overestimates) the function for x > 0 and smaller
than (underestimates) the function for x < 0. ♦
In Chapter 6, we associate these properties with the concavity of the
function, that is, whether the graph is locally concave up or down.
y =x −x
400 y 2 y Figure 5.5: Functions (black curves)

and their linear approximations (red) for
y = x2 y = sin(x) Examples 5.6 and 5.7. Whenever the
300 1
tangent line is below (above) the curve, we
x say that the linear approximation under
200 (over)-estimates the value of the function.
−4 −2 2 4
100 tangent line −1

tangent line
x
−2
5 10 15 20
(a) (b)
√ i Calculations for Example 5.9.

Example 5.9 a) Use linear approximation to estimate the value of 6.
b) Determine whether the linear approximation underestimates or overesti-

mates the function.
√
Solution. The derivative of y = f (x) = x = x1/2 is
1
f 0 (x) = √ = (1/2)x−1/2 . Mastered Material Check
2 x
13. A perfect square is an integer m of
Both the function and its derivative require evaluation of a square root. the form m = n2 where n is also an
Some numbers (“perfect squares” ) have convenient square roots. One integer. List several different perfect
squares.
such number, x = 4, is nearby, so we use it as the “base point” for a linear
approximation.
3 y 3 y √ Figure 5.6: Linear approximationp based at

Approximating 6 x = 4 to the function y = f (x) = (x). The
linear approx value panel on the right is a magnified view.
approx 2.5
2 at x = 4 actual value
2
1
√ 1.5
y= x
x x
2 4 6 8 4 5 6 7
(a) (b)
√
a) The slope of the tangent line at x = 4 is f 0 (4) = 1/(2 4) = 1/4 = 0.25
√
so the linear approximation of 6 is obtained as
y = f (4) + f 0 (4)(x − 4) ⇒ y = 2 + 0.25(x − 4)

√
⇒ 6 ≈ 2 + 0.25(6 − 4) = 2.5.
b) A graph of the function and its tangent line in Figure 5.6(a) and a zoomed
√
portion in Figure 5.6(b) compares the true and approximated values of 6.
The tangent line is above the graph of the function, so the linear approxi-
mation overestimates values of the function. ♦
The discrepancy between true and approximated values is called the error.
The closer we are to the base point, the smaller the error in the approxima-
tion. This is demonstrated by comparing the values in Table 5.1, computed
using a spreadsheet with base point x0 = 4.
√
Table 5.1: Linear approximation to x at the
base point x = 4. The exact value is recorded
exact value approx. value
√ in column 2 and the linear approximation in
x f (x ) = x y = f (x0 ) + f 0 (x0 )(x − x0 ) column 3. The approximation is reasonably
0.0000 0.0000 1.0000 good close to the base point.
2.0000 1.4142 1.5000 Mastered Material Check

4.0000 2.0000 2.0000 14. Find the error (true value minus
6.0000 2.4495 2.5000 approximate value)√in the linear
approximation for 6 found in
8.0000 2.8284 3.0000
Example 5.9 .
10.0000 3.1623 3.5000
15. Determine the error√of the
12.0000 3.4641 4.0000 approximation for 8 as given in
14.0000 3.7417 4.5000 Table 5.1.
16.0000 4.0000 5.0000 16. Which approximation has larger

error? Was this to be expected or
not?
5.4 Tangent lines for finding zeros of a function: Newton’s method
1. Describe the geometry on which Newton’s method is based (Figure 5.7).
2. Given f (x) and initial guess x0 , use Newton’s method to find improved
values x1 , x2 , etc., for the zero of f (x) (value of x such that f (x) = 0).
3. For a given function f (x), select a suitable initial guess x0 for Newton’s
method from which to start the iteration.
Definition 5.1 (Zero) Given a function f (x), we say that x∗ is a zero of f

if f (x∗ ) = 0. We also say that “x∗ is a root of the equation f (x) = 0”.
In many cases, it is impossible to compute a value of a zero, x∗ analyt-
ically. Based on tangent line approximations, we now explore Newton’s
method, an approximation that does the job.
Newton’s method
Consider the function y = f (x) shown in Figure 5.7. We have already found
that a tangent line approximates the behaviour of a function close to a point
y
of tangency. It can also be used to build up and refine an approximation of the tangent line
zeros of the function.
Our goal is to find a decimal approximation for the value x such that
f (x) = 0. (In Fig. 5.7, this value is denoted by x∗ .) Newton’s method is an f (x0 )
iterated scheme (a procedure that gets repeated). It is applied several times,
to generate a decimal expansion of the desired zero to any level of accuracy.
A starting value, x0 , initiates the method. This can be a rough first guess
x1 x∗ x0 x
for the zero we seek, found graphically, for example. Newton’s method,
applied a number of times, will generate better and better approximations of
Figure 5.7: In Newton’s method, we seek
the true zero, x∗ . a decimal approximation for x∗ , a zero of
Gluing a tangent line at x0 , we follow it down to its x-axis intersection. In y = f (x). A rough initial guess, x0 , is refined
by “sliding down the tangent line” (glued
a previous section, we have already computed that intersection point in (5.3), to the curve at x0 ). This brings us to an
to be improved guess x1 . Repeating this again
f ( x0 ) and again allows us to find the root to any
x1 = x0 − 0 .
f (x0 ) desired accuracy.
Then usually x1 is closer to x∗ , improving upon the initial guess.

Now use x1 as the (improved) guess, and repeat the process. This gener-
ates values
f ( x1 )
x2 = x1 − 0 .
f (x1 )
f ( x2 )
x3 = x2 − .
f 0 (x2 )
..
.
The values x2 , x3 rapidly approach the desired root x∗ . The same ‘recipe’
is repeated. In practice, when it works, Newton’s method converges quite
rapidly, that is, it approaches the root with excellent accuracy, after very few
repetitions (iterations). To summarize,
Newton’s method: Given an approximation xk for the root of the equation

f (x) = 0, we can improve the accuracy of that approximation with another
iteration using
f ( xk )
xk+1 = xk − 0 .
f (xk )
` Newton’s Method applied to the

Example 5.10 Find zeros of the function y = f (x) = x3 − x − 3 starting with function f (x) = x3 − x − 3.
the initial guess x0 = 1. Move the slider or the initial point x0 to
see how the first approximation x1 and
the second approximation x2 change.
Solution. The derivative is f 0 (x) = 3x2 − 1, so Newton’s method for the Compute the approximation x3 on the
improved guess is same graph.
f ( x0 ) x03 − x0 − 3 13 − 1 − 3
x1 = x0 − = x0 − = 1 − .
f 0 (x0 ) 3x02 − 1 3 · 12 − 1
So, starting with x0 = 1, we obtain Mastered Material Check

15. Perform the calculations to verify
x1 = 2.5, x2 = 1.929577465, x3 = 1.7078664, x4 = 1.672558473, x5 = 1.671700382. the x1 , x2 , x3 and x4 found in
Example 5.10.
The iterates converge to the result x ≈ 1.6717. ♦
Example 5.11 Use Newton’s method to find a decimal approximation of the

square root of 6.
y y
Figure 5.8: Newton’s method applied to
5 Newton’s method 5 Newton’s method
solving y = f (x) = x2 − 6 = 0.
f (x) = x2 − 6 f (x) = x2 − 6
x0 x1 x x0 x3 x2 x1 x
1 2 3 4 1 2 3 4
−5 −5

Solution. Because Newton’s method finds zeros of a function, it is first
16. Note that other functions have this
necessary to restate the problem in the form “find a value of x such that a property. For example, verify that
√ a
certain function f (x) = 0.” Clearly, one function that would accomplish this is root of f (x) = x4 − 36 is also 6.
17. Give another example of a
f ( x ) = x2 − 6 function f (x√
) for which f (x) = 0
has the root 6.
√
since f (x) = 0 corresponds to x2 − 6 = 0, i.e. x = 6. Then the derivative of
f (x) is f 0 (x) = 2x, and the recipe to repeat is
f (x0 ) x2 − 6
x1 = x0 − 0
= x0 − 0 .
f (x0 ) 2x0
Starting with an initial guess x0 = 1 (not very close to the value of the
P Link to Google Sheets. This
root), we show in Figure 5.8 how Newton’s method applies a tangent line to spreadsheet implements Newton’s
determination x1 . In the right panel, we see how the value of x1 is then used method for Example 5.11. You can view
the formulae by clicking on a cell in the
to obtain x2 by repeating the calculation. sheet but you cannot edit the sheet here.
A spreadsheet is ideal for carrying out the repetitive calculations, as shown
in Table 5.2. For example, we compute the following set of values using
k xk f (x k ) f 0 (x k ) x k +1
our spreadsheet. Observe that the fourth column contains the computed 0 1.00 -5.00 2.00 3.5
(Newton’s method) values, x1 , x2 , etc. These values are then copied onto the 1 3.5 6.250 7.00 2.6071
first column to be used as new “initial guesses”. After several repetitions, 2 2.6071 0.7972 5.2143 2.4543
3 2.4543 0.0234 4.9085 2.4495
the numbers calculated converge to 2.4495, and no longer change to that 4 2.4495 0.000 4.8990 2.4495
level of accuracy displayed. This signals that we have obtained the root to 5
Table 5.2: Newton’s method applied to
significant figures of accuracy. ♦
Example 5.11. We start with x0 = 1 as
Note: it is possible that Newton’s method fails to find a root - something we our initial approximation and refine it four
times.
do not explore further here. This might happen if our initial guess is too poor.
5.5 Aphids and Ladybugs, revisited

i Why do we need to use Newton’s
In Example 1.3, we asked when predation (by a ladybug) and growth rate Method if we already solved the aphid
exactly match for an aphid population. We did so by solving an equation -ladybug predation problem in
Section 1.6.1?
of the form P(x) = G(x) for x the aphid density and G(x) = rx the aphid
growth rate (r > 0), and P(x) the rate of predation of aphids by a ladybug.
Our solution relied on the quadratic formula. Now consider the case that the
predation rate is
x3
Number eaten per hour, P(x)

P(x ) = K , where K, a > 0. (5.4)
a3 + x 3
In this case, steps shown in Example 1.3 lead to a cubic equation for x,
which is not easy to solve by pen and paper. This is a classic situation where
Newton’s method proves useful.
Example 5.12 (Using Newton’s Method to solve the aphid-ladybug problem)

Set up the problem for obtaining the number of aphids at which predation by
0 20 40 60 80 100
a ladybug and population growth of aphids balance. Convert the equation to
Number of aphids, x
a form for which Newton’s method is suitable. Then use Newton’s method to
solve your problem. Assume that K = 30 aphids eaten per hour, a = 20 aphids,
and r = 0.5 per hour. To get a reasonable initial guess, plot P(x) and G(x) on Figure 5.9: The function P(x) superimposed
the same graph and determine roughly where they intersect. on a graph of data from [Hassell et al.,
1977].
Solution. The problem to be solved (assuming x 6= 0 is
x3 x2
P(x ) = G(x ) ⇒ K = rx ⇒ K =r
a3 + x 3 a3 + x3
simplifying algebraically leads to the equation ` Examine this graph of the predation
rate P(x) and the population growth rate
Kx2 = r (a3 + x3 ) ⇒ rx3 − Kx2 + ra3 ≡ f (x) = 0. G(x) to find reasonable initial guess(es)
for points of intersection. (We look only
for positive values, since x represents
Having converted the problem into the form f (x) = 0, we can apply New- the number of aphids.) These will be
ton’s method. We need the function and its derivative for Newton’s method used as value(s) for x0 in Newton’s
formula, method.
f (x) = rx3 − Kx2 + ra3

f 0 (x) = 3rx2 − 2Kx
Using the numerical values for the constants, and examining the graph of
the two functions P(x) and G(x), we find intersections at x = 0 and x0 ≈ 10.
k xk f (x k ) f 0 (x k ) x k +1
There is another intersection at x0 ≈ 60. To implement the method, we apply 0 10 1500.00 -450.00 13.33333333
Newton’s formula, 1 13.33333333 -148.15 -533.33 13.05555556
2 13.05555556 -0.78 -527.66 13.05407294
f ( x0 ) (rx03 − Kx02 + ra3 ) 3 13.05407294 0.00 -527.63 13.05407294
x1 = x0 − = x0 − 4 13.05407289 0.00 -527.63 13.05407294
f 0 (x0 ) 3rx02 − 2Kx0
Table 5.3: Newton’s method applied to
with x0 = 10. Table 5.3 summarizes the convergence to the root x =
Example 5.12. We start with x0 = 10 as the
13.05407289 after four rounds of improvement using Newton’s method. initial approximation.
Since x measures a density of aphids (e.g. average number per cm2 ), it is

reasonable to find a real valued solution (rather than an integer “number of
aphids”). i A short video discussion of the
Aphid-ladybug problem and how to use
We conclude that at this aphid density, the aphid population would have a
Newton’s Method to solve it on a
growth and a predation rate that exactly match. (Hence, we also expect that spreadsheet, as discussed in
the aphid population would neither increase nor decrease.) What happens if Example 5.12.
growth and predation rates do not match? In such cases, we expect change to
take place. How to analyze such situations will be the topic of a later chapter.
5.6 Harder problems: finding the point of tangency
1. Find a tangent line to a function that goes through some point (not neces-
sarily on the graph of the function).
2. Determine tangent lines to functions that contain unspecified parameters.
In this section, we present a sample of problems in which the path to a

solution is more subtle. In some of these, finding the point of tangency is part
of the question. We must use clues about the function to solve for that point,
as well as construct the tangent line equation from information supplied. In
other cases, the problem involves a parameter whose value is not specified y y = ax
initially. Such examples are meant to hone problem solving skills.
Example 5.13 Find any value(s) of the constant a such that the line y = ax is
y = f (x )
tangent to the curve y = f (x) = −x2 + 3x − 2.
x0 x
Solution. We do not know the coordinate of any such point, but we label it x0
in Figure 5.10 to denote that it is a definite (as yet to be determined) value.
Finding x0 is part of the problem. We collect the information to be used:
Figure 5.10: Figure for Example 5.13:
• The tangent line y = ax intersects the graph of the function y = f (x) =
finding the point of tangency.
−x2 + 3x − 2 at (x0 , f (x0 )). Mastered Material Check
• Equating y values of the tangent line and the curve y = f (x) at x0 we get: 18. Factor f (x) = −x2 + 3x − 2. Does
the shape depicted in Figure 5.10
make sense?
f (x0 ) = −x02 + 3x0 − 2 = ax0 .
• The equation of the tangent line is y = ax. Its slope is a, which is also the
derivative of f (x) at x0 . Equating slopes gives:
f 0 (x0 ) = −2x0 + 3 = a.
We have two equations with two unknowns, (a and x0 ). We can solve this
system by substituting the value of a from the first equation into the second,
getting
−x02 + 3x0 − 2 = (−2x0 + 3)x0 .
Simplifying: y y = ax
√
−x02 + 3x0 − 2 = −2x02 + 3x0 ⇒ x02 − 2 = 0, x0 = ± 2.
Thus, there are two possible points of tangency, as shown in Figure 5.11. y = f (x)
Finally, we find a using a = −2x0 + 3. We get:
x0 x
√ √ √ √
x0 = 2 ⇒ a = −2 2 + 3, and x0 = − 2 ⇒ a = 2 2 + 3.
♦
The solution to Example 5.13 was set up by
• listing of information provided,

• deducing a set of equations based on that information, and
• following a chain of reasoning to arrive at the final solution.
Practicing such multi-step problems is a critical part of training for many Figure 5.11: Two points of tangency in the
solution to Example 5.13
fields, including science, medicine, engineering, etc.
Example 5.14 Find the equation of the tangent line to the curve y = f (x) =
1 − x2 that goes through the point (1, 1).
Solution. Finding the point of tangency x0 is part of the problem. We use the
following facts:
• The tangent line goes through the point (x0 , f (x0 )) on the graph of the
function and has slope f 0 (x0 ).
• Consequently, its equation has the form Eqn. (5.2): y = f (x0 ) = f 0 (x0 )(x −
x0 ).
For the given function and point of tangency x0 , we have
f (x0 ) = 1 − x02 , f 0 (x0 ) = −2x0 .
Hence the tangent line equation is Mastered Material Check

19. Identify the slope of the
y = f (x0 ) + f 0 (x0 )(x − x0 ) = (1 − x02 ) − 2x0 (x − x0 ). line y = (1 − x02 ) − 2x0 (x − x0 ).
20. Determine the y-intercept of the
We are told that this line goes through the point (x, y) = (1, 1) so that line y = (1 − x02 ) − 2x0 (x − x0 ).
21. Check that both lines found as
1 = (1 − x02 ) − 2x0 (1 − x0 ), ⇒ 0 = x02 − 2x0 , ⇒ x02 = 2x0 . solutions to Example 5.14 go
through (1, 1) as desired.
Thus, there are two possible points of tangency, x0 = 0, 2 and two tangent lines
that satisfy the given condition. Plugging in these two values of x0 into the
generic equation for y leads to the two tangent line equations
1. y = 1, and
2. y = (1 − 22 ) − 2 · 2(x − 2) = −3 − 4(x − 2). ♦
We can also find points of tangency for functions that contain general
constants, as the next example illustrates.
Example 5.15 Shown in Figure 5.12 is the function

x
f (x ) = C
x+a
together with one of its tangent lines. The tangent line goes through a point
(−d, 0). Find the equation of the tangent line.
Solution. Finding the point of tangency x0 is part of the problem in this y

case too. We use the same approach, and employ facts (1) and (2) from
Example 5.14. We also use, for the specific function in this example,
x0 a
f (x0 ) = C ⇒ f 0 (x0 ) = C .
x0 + a ( x0 + a ) 2
(See Exercise 10 in Chapter 3). Hence, the equation of the tangent line is
x0 a
y = f (x0 ) + f 0 (x0 )(x − x0 ) = C +C ( x − x0 ) . −d x0 x
x0 + a (x0 + a)2
We can simplify this equation by factoring to obtain: Figure 5.12: The graph of a function and its
tangent line for Example 5.15.
C C
x02 + ax .

y= 2
(x0 (x0 + a) + a(x − x0 )) = 2
(x0 + a) (x0 + a)
It is important to realize that in this equation, x0 ,C and a represent fixed
(known) constants, and only x, y are variables. This means that the equation
expresses a linear relationship between x and y, as appropriate for a straight
line.
We know that the point (−d, 0) is on this line, so that (plugging in x =
−d, y = 0), we obtain
C
x02 − ad .

0= 2
( x0 + a )
√
Solving for x0 leads to x0 = ad. Moreover, we can now find the equation of
the tangent line in terms of these parameters.
C C
y= √ (ad + ax) ⇒ y= p (d + x )
( ad + a)2 ( (d/a) + 1)2
where we have simplified by factoring a from numerator and denominator.
We can easily see that when x = −d, we get y = 0, as required. This forms one
check that our calculations are correct. ♦
5.7 Summary
1. The equation of a tangent line to f (x) at x0 is given by
y = f (x0 ) + f 0 (x0 )(x − x0 ).
2. If L(x) is the tangent line to a function f (x) at x0 , then L(x) forms a linear
approximation to f (x) near the point x0 .
3. In some circumstances, the zero of a tangent line to a function f (x) at a

point x0 can form an initial approximation to the zero of f (x).
4. Newton’s method is based on the property of tangent lines. Newton’s
method can solve a problem of the form f (x) = 0. Given an initial guess
x0 , the method generates successive decimal approximations to the zeros
of the function to any desired accuracy. The iteration scheme is:
f (xk )
xk+1 = xk − .
f 0 (xk )
1. Is it possible for two different tangent lines of the same function to be parallel?
2. When would a tangent line not intersect the x-axis?
3. Consider the graph of the following function, and its tangent line at x = 1.
15
10
1 2 3 4 5
(a) When would the linear (tangent line) approximation result in an overestimate? Under-estimate?
(b) What is a reasonable interval on which to use this tangent line for approximation?
4. Why might Newton’s method not work?

Exercises
5.1. Equation of tangent line. Find the equation of the tangent line to the
function y = f (x) = |x + 1| at:
(a) x = −1,
(b) x = −2,
(c) x = 0.
If there is a problem finding a tangent line at one of these points, y
indicate what the problem is.
5.2. Equation of tangent line. A function f (x) satisfies f (1) = −1
and f 0 (1) = 2. What is the equation of the tangent line of f (x) at x =
1?
5.3. Point of tangency. Shown in Figure 5.13 is the graph of y = x2 with
one of its tangent lines.
(a) Show that the slope of the tangent to the curve y = x2 at the x
point x = a is 2a. 1 a
(b) Suppose that the tangent line intersects the x axis at the point (1, 0).
Figure 5.13: Figure for Exercise 3; y = x2
Find the coordinate, a, of the point of tangency. and a point of tangency.
5.4. Approximation with a tangent line. Shown in Figure 5.14 is the y
function f (x) = 1/x4 together with its tangent line at x = 1.

f (x)
(a) Find the equation of the tangent line.
(b) Determine the points of intersection of the tangent line with the x
and the y-axes.
(c) Use the tangent line to obtain a linear approximation to the value
of f (1.1). Is this approximation larger or smaller than the actual
x
value of the function at x = 1.1? 1
5.5. Linear approximation. Shown in Figure 5.15 is the function f (x) =
1
x3 with a tangent line at the point (1, 1). Figure 5.14: Figure for Exercise 4; f (x) = x4
and a point of tangency.
(a) Find the equation of the tangent line. y
(b) Determine the point at which the tangent line intersects the x axis. f (x)
(c) Compute the value of the function at x = 1.1. Compare this with the
value of y on the tangent line at x = 1.1.
This latter value is the linear approximation of the function at the (1, 1)
desired point based on its known value and known derivative at the
nearby point x = 1.
5.6. Generic tangent line. Shown in Figure 5.16 is the graph of a function x
and its tangent line at the point x0 .
Figure 5.15: Figure for Exercise 5; f (x) = x3
(a) Find the equation of the tangent line expressed in terms of x0 , f (x0 )
and a point of tangency.
and f 0 (x0 ).
y
(b) Find the coordinate x1 at which the tangent line intersects the x-
tangent
axis.
line
5.7. Estimating a square root. Use Newton’s method to find an ap-
√
proximate value for 8 (hint: first think of a function, f (x), such
√
that f (x) = 0 has the solution x = 8).
5.8. Finding points of intersection. Find the point(s) of intersection
of: y1 = 8x3 − 10x2 + x + 2 and y2 = x3 + 15x2 − x − 4 (hint: an x
intersection point exists between x = 3 and x = 4). x1 x0
5.9. Roots of cubic equations. Find the roots for each of the following Figure 5.16: Figure for Exercise 6; generic
cubic equations using Newton’s method: function and tangent line.
(a) x3 + 3x − 1 = 0
(b) x3 + x2 + x − 2 = 0
(c) x3 + 5x2 − 2 = 0 (hint: find an approximation to a first root a using
Newton’s method, then divide the left hand side of the equation
by (x − a) to obtain a quadratic equation, which can be solved by the
quadratic formula). unknown
coordinates to
x2
5.10. Intersecting tangent lines. The parabola y = has two tangent lines find
that intersect at the point (2, 3). These are shown as the dark lines in
Figure 5.17. Find the coordinates of the two points at which the lines
are tangent to the parabola.
Note: note that the point (2, 3) is not on the parabola.
5.11. An approximation for the square root. Use a linear approximation
to find a rough estimate of the following functions at the indicated (2, 3)
points. x
√ √ y
(a) y = x at x = 10 (use the fact that 9 = 3).
Figure 5.17: Figure for Exercise 10; y = x2
(b) y = 5x − 2 at x = 1.
5.12. An approximation for the cube root. Use the method of linear
approximation to find the cube root of
√
(a) 0.065 (hint: 3 0.064 = 0.4)
√
(b) 215 (hint: 3 216 = 6)
5.13. Approximating from a graph. Use the data in the graph in Fig-
ure 5.18 to make the best approximation you can to f (2.01).
5.14. Linear approximation. Approximate the value of f (x) = x3 − 2x2 +
3x − 5 at x = 1.001 using the method of linear approximation.
5.15. Approximating cube volume. Approximate the volume of a cube
whose length of each side is 10.1 cm. Figure 5.18: Figure for Exercise 13; using
a graph to approximate the value of a
5.16. Using Newton’s method to find a critical point. Consider the func- function.
tion
g(x) = x5 − 4x4 + 3x3 + x2 − 3x.
Critical points of a function are defined as values of x for which

g0 (x) = 0. However, for this fifth order polynomial, it is not easy to find
such points analytically (i.e., using pencil and paper).
(a) Use Newton’s Method to find a critical point for positive values
of x. Find an initial approximation for the critical point by plotting
the function, but use a spreadsheet and explain how you set up the
calculations. Provide an answer accurate to 8 decimal points.
(b) Explain why a starting value of x0 = 1 for Newton’s Method
does not lead to the positive critical point. You may support your
argument with a graph.
Note: in Section 6.2 we study critical points in greater depth.
5.17. Newton’s Method Use Newton’s method to find a decimal approxima-
√
tion of the value of 105 correct to 16 decimal points.
6
Sketching the graph of a function using calculus tools
The derivative of a function contains important information about the original

function. In this chapter we focus on how properties of the first and second
derivative can be used to help up refine curve-sketching techniques. The
mathematics we develop in this chapter is used in a variety of applications,
many found in Chapter 7.
6.1 Overall shape of the graph of a function
1. Identify that the sign of the first derivative corresponds to the increasing or
decreasing trend of a function.
2. Recognize that the sign of the second derivative corresponds to the concav-
ity (curvature) of a function.
Increasing and decreasing functions

Consider a function given by y = f (x). We first make the following observa-
tions: Mastered Material Check
1. Draw the graphs of two different
1. If f 0 (x) > 0 then f (x) is increasing. functions that are increasing on the
interval [0, 1].
2. If f 0 (x) < 0 then f (x) is decreasing.
2. Without using the derivative, how do
By convention, we read graphs from left to right, i.e. in the direction of the you identify when a function is
decreasing?
positive x-axis, so when we say “increasing” we mean that, as we move from
left to right, the value of the function gets larger.
We can use the same ideas to relate the second derivative to the first
derivative.
1. If f 00 (x) > 0 then f 0 (x) is increasing. This means that the slope of the
original function is getting steeper (more positive, from left to right). The
function curves upwards: we say that it is concave up. See Figure 6.1(a).
2. If f 00 (x) < 0 then f 0 (x) is decreasing. This means that the slope of the
original function is getting shallower (more negative or less positive from
left to right). The function curves downwards: we say that it is concave
down. See Figure 6.1(b).
i See video summarizing the
connection between the shape of the
graph of y = f (x), and the derivatives of
Concavity and points of inflection the function, f 0 (x), f 00 (x) of the
function.
The second derivative of a function provides information about the curvature
of the graph of the function, also called the concavity of the function.
• In Figure 6.1(a), f (x) is concave up, so its second derivative is positive.
• In Figure 6.1(b), f (x) is concave down, so second derivative is negative.
(a) (b) Figure 6.1: In (a) the function is concave up,

and its derivative increases (in the positive
direction). In (b), the function is concave
f (x) down, and its derivative decreases.
f (x)
x x
f 0 (x) f 0 (x)
x x
f 00 (x) f 00 (x)
x
Definition 6.1 A point of inflection of a function f (x) is a point x at which

the concavity of the function changes (Figure 6.2).
SKETCHING THE GRAPH OF A FUNCTION USING CALCULUS TOOLS 129
We can deduce from the definition and previous remarks that at a point of inflection point f (x)
inflection the second derivative changes sign. This is illustrated in Figure 6.2. f 00 (x) = 0
Note: It is not enough to show that f 00 (x) = 0 to conclude that x is an f 00 (x) > 0
inflection point. We summarize the one-way nature of this relationship in the concave up
box and discuss it further in Example 6.1.

f 00 (x) < 0
Inflection points concave down
1. If the function y = f (x) has a point of inflection at x0 then f 00 (x0 ) = 0. Figure 6.2: An inflection point is a place
where the concavity of a function changes.
2. If the function y = f (x) satisfies f 00 (x0 ) = 0, we cannot conclude that it
has a point of inflection at x0 . We must actually check that f 00 (x) changes
sign at x0 .
Example 6.1 Consider the the functions (a) f1 (x) = x3 and (b) f2 (x) = x4 .
Show that for both functions, the second derivative is zero at the origin
( f 00 (0) = 0) but that only one of these functions actually has an inflection point
at x = 0.
i A summary of Example 6.1,
reinforcing the fact that f 00 (x) = 0 is not
Solution. The functions are
enough to guarantee an inflection point!
We have to check that f 00 (x) changes
(a) f1 (x) = x3 , (b) f2 (x) = x4 . sign.
The first derivatives are
(a) f10 (x) = 3x2 , (b) f20 (x) = 4x3 .
and the second derivatives are:
(a) f100 (x) = 6x, (b) f200 (x) = 12x2 .
Figure 6.3: The functions (a) f1 (x) = x3

and (b) f2 (x) = x4 both satisfy f 00 (0) = 0.
x3 x4 However, only x3 has an inflection point
at x = 0, whereas x4 has a local minimum at
that point. This is not a contradiction, since
f200 (x) does not change sign at x = 0.
(a) (b)

Thus, at x = 0 we have f100 (0) = 0, f200 (0) = 0. However, x = 0 is NOT an 3. Check that f (x) = x4 does not
inflection point of x4 . In fact, it is a local minimum, as shown in Figure 6.3. change sign at x = 0 by comparing
the signs of f (−0.1) and f (0.1).
♦
Determining whether f 00 (x) changes sign
We defined an inflection point as a point on the graph of a function where

the second derivative changes sign. But how do we detect if this sign change
occurs at a given point?
We first state the following result i How we check where f 00 (x) changes
sign (to identify inflection points).
Sign change in a product of factors:

If an expression is a product of factors, e.g. g(x) = (x−a1 )n1 (x−a2 )n2 . . . (x−
am )nm , then
1. The expression can be zero only at the points x = a1 , a2 , . . . , am .
2. The expression changes sign only at points x = ai for which ni is an odd in-
teger power.

4. How might you determine sign
changes for a function which is not
Example 6.2 Determine where g(x) = x2 (x + 2)(x − 3)4 changes sign. given as product of factors?
5. Circle the places where the function
depicted below changes sign.
5
Solution. The zeros of g(x) are x = 0, −2, 3. At x = 0 and at x = 3 there is no
sign change as the terms x2 and (x − 3)4 are always positive. The factor (x + 2)
−2 2 4
goes from negative through zero to positive when x goes from x < −2
to x = −2 to x > −2. Hence, g(x) only changes sign at x = −2. ♦ −5
−10
Example 6.3 Where does the function f (x) = (2/5)x6 − x4 + x have
inflection point(s)?
Solution. The derivatives of the function are
f 0 (x ) = (12/5)x5 − 4x3 + 1,
f 00 (x) = 12x4 − 14x2 = 12x2 (x2 − 1) = 12x2 (x + 1)(x − 1).
Here we have completely factored the second derivative. Sign changes can
only occur when there are factors with odd powers, such as (x + 1) and (x − 1).
These change sign at x = −1, 1, respectively - making both inflection points.
There is NO sign change at x = 0, since the factor x2 is always positive. ♦
6.2 Special points on the graph of a function
1. Define a zero of a function and be able to identify zeros for simple func-
tions (factorizable polynomials).
2. Explain that a function f (x) can have various types of critical points
(maxima, minima, and other types) at which f 0 (x) = 0.
3. Find critical points for a given function.
4. Using first or second derivative tests, classify a given critical point as a

maximum, minimum (or neither).
In this section we use tools of algebra and calculus to identify special

points on the graph of a function. We first consider the zeros of a function,
and then its critical points.
Zeros of a function
Example 6.4 (Factoring) For the function y = f (x) = x2 − 5x + 6, find zeros
by factoring.
Solution. This polynomial has factors f (x) = (x − 3)(x − 2). Zeros are values
of x satisfying 0 = (x − 3)(x − 2), so either (x − 3) = 0 or (x − 2) = 0. Hence,
there are two zeros, x = 2, 3. ♦ Mastered Material Check
5. What is another term used for zeros
Example 6.5 Find the zeros of the function y = f (x) = x3 − 3x2 + x. of a function?
Solution. We can factor this function into f (x) = x(x2 − 3x + 1). From this
we see that x = 0 is one of the desired zeros of f . To find the others, we apply
the quadratic formula to the second factor, obtaining
1 p 1 √
x1,2 = (3 ± 32 − 4) = (3 ± 5).
2 2
Thus, there are a total of three zeros in this case:
1 √ 1 √
x = 0, (3 + 5), (3 − 5).
2 2
♦
Example 6.6 Find zeros of the function y = f (x) = x3 − x − 3.
Solution. This polynomial does not factor into integers, nor is it easy to apply
a cubic formula (analogous to the quadratic formula). However, as we saw in
Example 5.10, Newton’s method leads to an accurate approximation for the
only zero of this function, (x ≈ 1.6717) ♦
Critical points
Definition 6.2 A critical point of the function f (x) is any point x at which
the first derivative is zero, i.e. f 0 (x) = 0.
Figure 6.4: A critical point (place where

f 0 (x) = 0) can be a local maximum, local
Clearly, a critical point occurs whenever the slope of the tangent line to minimum, or neither.
the graph of the function is zero, i.e. the tangent line is horizontal. Figure 6.4
shows several possible shapes of the graph of a function close to a critical
point. From left to right, these are a local maximum, a local minimum and
two inflection points. i A summary of types of critical points
and how to tell them apart.
Critical points play an important role in many scientific applications,
as described in Chapter 7. Hence, we seek criteria to determine whether a
critical point is a local maximum, minimum, or neither.
Example 6.7 Consider the function y = f (x) = x3 + 3x2 + ax + 1. For what

range of values of a are there no critical points for this function?
Solution. We compute the first derivative f 0 (x) = 3x2 + 6x + a. A critical

point would occur whenever 0 = f 0 (x), which implies 0 = 3x2 + 6x + a. This
is a quadratic equation whose solutions are
√
−6 ± 36 − 4a · 3
x1,2 = .
6
There two real solutions provided the discriminant is positive, 36 − 12a > 0. Mastered Material Check
However, when 36 − 12a < 0 (which corresponds to a > 3), there are no real 6. Why are there no critical points
solutions and consequently no critical points. ♦ when 36 − 12a < 0 for the
function f (x?) = x3 + 3x2 + ax + 1
What happens close to a critical point

In Figure 6.5 we contrast the behaviour of two functions (top row), each
with a different type of critical point. We compare their first and second
derivatives close to that point (second and third rows, respectively). In each
case, the first derivative f 0 (x) = 0 at the critical point.
Near the local maximum (moving from left to right), the slope of f (x)
transitions from positive to zero (at the critical point) to negative. This means
that f 0 (x) is a decreasing function, as indicated in Figure 6.5. Since changes
in the first derivative are measured by its derivative f 00 (x), we find that the
f (x) f (x) Figure 6.5: Close to a local maximum, f (x)

local max local min is concave down, f 0 (x) is decreasing, so that
f 00 (x) is negative. Close to a local minimum,
f (x) is concave up, f 0 (x) is increasing, so
that f 00 (x) is positive.
x x
f 0 (x) f 0 (x)
x x
f 00 (x) f 00 (x)
second derivative is negative at a local maximum. The converse is true near a

local minimum (right column in Figure 6.5).
We collect and summarize conclusions about derivatives below.
Summary: first derivative
f 0 (x ) < 0 f 0 (x0 ) = 0 f 0 (x ) > 0

decreasing function critical point increasing function
at x0
• The first derivative test: depends on changes in the sign of the first
derivative close to a critical point, x0 .
Near a local maximum, the sign pattern is:
1. x < x0 , f 0 (x) > 0;

2. x = x0 , f 0 (x0 ) = 0;
3. x > x0 , f 0 (x0 ) < 0.
Near a local minimum, the sign pattern is:
1. x < x0 , f 0 (x) < 0;

2. x = x0 , f 0 (x0 ) = 0;
3. x > x0 , f 0 (x) > 0.
Summary: second derivative 7. Given the following graph of a
function’s second derivative, can
f 00 (x) < 0 f 00 (x0 ) = 0 f 00 (x) > 0 you say anything about the
concavity of the original function?
curve concave down inflection point curve concave up 1
f 00 (x)
at x0 only if f 00
0.5
changes sign
0.5 1 1.5 2
• The second derivative test: is based on the sign of f 00 (x0 ) for x0 a critical −0.5
point (satisfying f 0 (x0 ) = 0). −1
If f 00 (x0 ) < 0, x0 is a local maximum.
If f 00 (x0 ) = 0, test is inconclusive about max/min at x0 . If f 00 also changes

sign at x0 , then x0 is an inflection point.
If f 00 (x0 ) > 0, x0 is a local minimum.
6.3 Sketching the graph of a function
1. Given a function (polynomial, rational, etc.) be able to find its zeros,

critical points, inflection points, and determine where it is increasing or
decreasing, concave up or down.
2. Using a combination of the above techniques, together with methods

of Section 1.4, assemble a reasonably accurate sketch of the graph of a
function.
3. Using these techniques, identify all local as well as global extrema

(minima and maxima) of a function f (x) on an interval a ≤ x ≤ b.
In Section 1.4, we used elementary reasoning about power functions to

sketch the graph of simple polynomials. Now that we have learned more
advanced calculus techniques, we can hone such methods to produce more
accurate sketches of the graph of a function. We devote this section to
illustrating some examples.
Example 6.8 Sketch the graph of the function B(x) = C (x2 − x3 ). Assume
that C > 0 is constant.
i Steps in the calculations of
Solution. To prepare, we compute the derivatives: Example 6.8.
B0 (x) = C (2x − 3x2 ), B00 (x) = C (2 − 6x). 8. What is the independent variable in
Example 6.8? The dependent
1. Zeros. We start by finding the zeros of the function. Factoring makes this variable?
easy. Solving B(x) = 0, we find 9. Given our discussion on
“considering the powers” for
Example 6.8, add a reasonable scale
0 = C (x2 − x3 ) = Cx2 (1 − x), ⇒ ⇒ x = 0, 1.
to each of x-axes for the graphs in
Figure 6.6.
2. Consider the powers. Reasoning about the powers as in Chapter 1, we
surmise that close to the origin, x2 dominates (producing a parabolic
shape) whereas, far away, −x3 dominates (producing the shape of an
inverted cubic). We show this in a preliminary sketch, Figure 6.6.
B(x) B(x) B(x)
x x x
close to 0 0 far from 0
Figure 6.6: Figure for the function B(x) =

C (x2 − x3 ) in Example 6.8 showing which
3. First derivative. To find critical points, set B0 (x) = 0, obtaining power dominates.
B0 (x) = C (2x − 3x2 ) = 0, ⇒ 0 = 2x − 3x2 = x(2 − 3x),

2
⇒ x = 0, .
3
By considering the sketch in Figure 6.6, we can see that x = 0 is a local

minimum, and x = 2/3 a local maximum. Confirmation of this comes from
the second derivative.
4. Second derivative. From the second derivative, B00 (0) = 2 > 0, confirming
that x = 0 is a local minimum. Further, B00 (2/3) = 2 − 6 · (2/3) = −2 < 0
so x = 2/3 is a local maximum, as expected.
5. Classifying the critical points. Now identifying where B00 (x) = 0, we find
that
2 1
B00 (x) = C (2 − 6x) = 0, when 2 − 6x = 0 ⇒ x= =
6 3
we also note that the second derivative changes sign here: i.e. for x < 1/3, B(x)
B00 (x) > 0 and for x > 1/3, B00 (x) < 0. We may conclude that there is
inflection local max
an inflection point at x = 1/3. The final sketch, now labeled, is given in
Figure 6.7.
♦ x
1/3 2/3 1
Example 6.9 Sketch the graph of the function y = f (x) = 8 x5 + 5 x4 − 20 x3 local
min
Solution. This example is more challenging, but similar ideas apply.
1. Zeros. Factoring the expression for y and then using the quadratic formula Figure 6.7: A sketch of B(x) = C (x2 − x3 ) in
leads to Example 6.8.
5 1 √
y = x3 (8x2 + 5x − 20). ⇒ x = 0, − ± 665.
16 16
y
In decimal form, these are approximately x = 0, 1.3, −1.92.
2. Consider the powers. The highest power is 8 x5 , so far from the origin
we expect typical positive odd function behaviour. The lowest power is
−20 x3 , which means that close to zero, we expect to see a negative cubic.
This implies that the function “turns around”, creating local maxima and x
minima. We draw a rough sketch in Figure 6.8.
3. First derivative. Calculating the derivative of f (x) and then factoring

leads to
dy Figure 6.8: y = f (x) = 8 x5 + 5 x4 − 20 x3 .
= f 0 (x) = 40 x4 + 20 x3 − 60 x2 = 20x2 (2x + 3)(x − 1) From Example 6.9, this function behaves
dx roughly like the negative cubic near the
so this derivative is zero at: x = 0, 1, −3/2. We expect critical points at origin, and like 8 x5 for large x.
these places.
4. Second derivative. We calculate the second derivative and factor to obtain

d2y Mastered Material Check
= f 00 (x) = 160 x3 + 60 x2 − 120 x = 20x(8x2 + 3x − 6).
dx2 (a) Verify the given zeros of the second
derivative of Example 6.9.
Thus, the second derivative is zero at
3 1 √ 3 1 √
x = 0, − + 201, − − 201.
16 16 16 16
The values of these roots can be approximated by: x = 0, 0.69, −1.07.
10. Classifying the critical points. To identify the types of critical points, we
use the second derivative test.
• f 00 (0) = 0 so the test is inconclusive at x = 0.

• f 00 (1) = 20(8 + 3 − 6) > 0 implies a local minimum x = 1, and
11. Use software to verify the plot of the
• f 00 (−3/2) = −225 < 0 implies a local maximum at x = −3/2. function y = f (x) in Example 6.9.
Then plot 2y. What changes?
We summarize the results in Table 6.1.
The shape of the function, its first and second derivatives are shown in
Figure 6.9. ♦
Table 6.1: y = x3 (8x2 + 5x2 − 20) and its

detailed behaviour.
x= -1.92 -1.5 -1.07 0 0.69 1 1.3
f (x ) = 0 32.0 0 -7 0
f 0 (x ) = 0 0 0
f 0 0 (x ) = <0 0 0 0 >0
characteristic: zero max inflection inflection min zero
y 200 y 400 y
30
y = f (x) 200
150
20 y = f 0 (x) x
100 −2 −1.5 −1 −0.5 0.5 1 1.5
−200
10
50
−400
x
x 00
−2 −1.5 −1 −0.5 0.5 1 1.5 y = f (x) −600
−2 −1.5 −1 −0.5 0.5 1 1.5
−10 −800
Figure 6.9: The function y = f (x) =

8 x5 + 5 x4 − 20 x3 , and its first and second
Global maxima and minima, endpoints of an interval derivatives, f 0 (x) and f 00 (x).
A global maximum (also denoted absolute maximum) of a function over

some interval is the largest value that the function attains on that interval.
Similarly a global minimum (or absolute minimum) is the smallest value.
Note. For a function defined on a closed interval, we must check critical

points and endpoints of the interval to determine where global maxima and
minima occur, as illustrated in Example 6.10. i Steps in Example 6.10: Finding the
absolute minimum and maximum of a
2 function on a given interval.
Example 6.10 Consider y = f (x) = + x2 on the interval 0.1 ≤ x ≤ 4. Find
x
the absolute maximum and minimum.
Solution. We first compute the derivatives:

1 1
f 0 (x) = −2 + 2x, f 00 (x) = 4 + 2.
x2 x3
Solving for critical points by setting f 0 (x) = 0, we find
1 1
−2 + 2x = 0, ⇒ −2 = 2x ⇒ x3 = 1
x2 x2
which implies a critical point at x = 1. The second derivative at this point is
1
f 00 (1) = 4 + 2 = 6 > 0,
13
so that x = 1 is a local minimum.
138 DIFFERENTIAL CALCULUS FOR THE LIFE SCIENCES 20
15
We now calculate the value of the function at the endpoints x = 0.1 and 10
x = 4 and at the critical point x = 1 to determine where global and local
minima and/or maxima occur: 5
• f (0.1) = 2
+ 0.12 = 20.01; 1 2 3 4
0.1
Figure 6.10: The function f (x) = 2x + x2

• f (1) = 21 + 12 = 3; on the interval 0.1 < x < 4 has no local
maximum.
• f (4) = 24 + 42 = 16.5.
Consequently, x = 1 is both a local minimum and the global minimum on

the given interval. There are no local maxima. The global maximum occurs
at the left endpoint, x = 0.1. Figure 6.10 confirms our conclusions. ♦
6.4 Summary
1. We can improve the accuracy of a sketch of a function f (x) by examining

its derivatives. For example, if f 0 (x) > 0 then f (x) is increasing, and
if f 0 (x) < 0, then f (x) is decreasing. Further, if f 00 (x) > 0 then this cor-
responds to f 0 (x) increasing, which means f (x) is concave up. Similarly,
f 00 (x) < 0 corresponds to f (x) being concave down.
2. If f 00 (x0 ) = 0 and f 00 (x) changes sign at x0 , then x0 is a point of inflection
of f (x): a point where its concavity changes.
3. If f 0 (x0 ) = 0, then x0 is a critical point of the function f (x). A critical point
needs to be further tested to determine whether it is a local minimum, local
maximum or neither.
4. Given a critical point x0 (where f 0 (x0 ) = 0), the first derivative test
examines the sign of f 0 (x) near x0 . If the sign pattern of f 0 (x) is +, 0, −
(from left to right as we cross x0 ), then x0 is a local maximum. If the sign
pattern is −, 0, +, then x0 is a local minimum.
5. Given a critical point x0 (where f 0 (x0 ) = 0), the second derivative test
examines the sign of f 00 (x0 ). If f 00 (x0 ) < 0, then x0 is a local maximum.
If f 00 (x0 ) > 0, then x0 is a local minimum. If f 00 (x0 ) = 0, the second
derivative test is inconclusive.
6. A global maximum and minimum can only be identified after comparing
the values of the function at all critical points and endpoints of the interval.
(Remark: For discontinuous or non-smooth functions, we must also
examine values of f at points of discontinuity and cusps, not discussed in
this chapter).
1. Consider the function depicted below:

15
10
1 2 3 4 5
(a) Circle all zeros.

(b) Where are the local maxima?
(c) How many local minima are there?
(d) Where does the function change sign?
2. Draw a graph to justify that if f 0 (x) > 0, then f (x) is increasing.
3. Where does the function f (a) = a3 (a − π )(a + ρ )5 (a − 2)2 change signs?
4. Suppose you are given that γ is a critical point of some function g. What would you ask to learn more about the
shape of g?
Exercises
6.1. Zeros, local minima and maxima. A zero of a function is a place

where f (x) = 0.
(a) Find the zeros, local maxima, and minima of the polynomial y =
f (x) = x3 − 3x
(b) Find the local minima and maxima of the polynomial y = f (x) =
(2/3)x3 − 3x2 + 4x.
(c) Determine whether each of the polynomials given in parts (a) and
(b) have an inflection point.
6.2. Classifying critical points. Find critical points, zeros, and inflection
points of the function y = f (x) = x3 − ax. Then classify the types of
critical points that you have found.
6.3. Sketching graphs. For each of the following functions, sketch the
graph for −1 < x < 1, find f 0 (0), f 0 (1), f 0 (−1) and identify any local
minima and maxima.
(a) y = x2 ,
(b) y = −x3 ,
(c) y = −x4
(d) Using your observations, when can you conclude that a function
whose derivative is zero at some point has a local maximum at that
point?
6.4. Sketching a graph. Sketch a graph of the function y = f (x) = x4 −2x3 ,
using both calculus and methods of Chapter 1.
6.5. Global maxima and minima. Find the global maxima and minima
for the function in Exercise 4 on the interval 0 ≤ x ≤ 3.
6.6. Absolute maximum and minimum. Find the absolute maximum and
minimum values on the given interval:
(a) y = 2x2 on −3 ≤ x ≤ 3
(b) y = (x − 5)2 on 0 ≤ x ≤ 6
(c) y = x2 − x − 6 on 1 ≤ x ≤ 3
1 1
(d) y = + x on −4 ≤ x ≤ − .
x 2
6.7. Local vs. absolute. A function f (x) has as its derivative f 0 (x) =
2x2 − 3x
(a) In what regions is f increasing or decreasing?
(b) Find any local maxima or minima.
(c) Is there an absolute maximum or minimum value for this function?
6.8. Minimum value. Sketch the graph of x4 − x2 + 1 in the range −3 to 3.

Find its minimum value.
6.9. Critical points. Identify all the critical points of the following func-
tion.
y = x3 − 27
6.10. Critical and inflection points. Consider the function g(x) = x4 −

2x3 + x2 . Determine locations of critical points and inflection points.
6.11. No critical points. Consider the polynomial y = x3 + 3x2 + ax + 1.
Show that when a > 3 this polynomial has no critical points.
6.12. Critical points and generic parabola. Find the values of a, b, and c
if the parabola y = ax2 + bx + c is tangent to the line y = −2x + 3
at (2, −1) and has a critical point when x = 3.
6.13. Double wells and physics. In physics, a function such as
f (x) = x4 − 2x2
is often called a double well potential. Physicists like to think of this

as a “landscape” with hills and valleys. They imagine a ball rolling
along such a landscape: with friction, the ball eventually comes to rest
at the bottom of one of the valleys in this potential. Sketch a picture of
this landscape and use information about the derivative of this function
to predict where the ball might be found, i.e. where the valley bottoms
are located.
6.14. Function concavity. Find the first and second derivatives of the
function
x3
y = f (x ) = .
1 − x2
Use information about the derivatives to determine any local maxima
and minima, regions where the curve is concave up or down, and any
inflection points.
6.15. Classifying critical points. Find all the critical points of the function
y = f (x) = 2x3 + 3ax2 − 12a2 x + 1
and determine what kind of critical point each one is. Your answer
should be given in terms of the constant a, and you may assume
that a > 0.
6.16. Describing a function. The function f (x) is given by
y = f (x) = x5 − 10kx4 + 25k2 x3
where k is a positive constant.

(a) Find all the intervals on which f is either increasing or decreasing.
Determine all local maxima and minima.
(b) Determine intervals on which the graph is either concave up or

concave down. What are the inflection points of f (x)?
6.17. Muscle shortening. In 1938 Av Hill proposed a mathematical model
for the rate of shortening of a muscle, v, (in cm/sec) when it is work-
ing against a load p (in gms). His so called force-velocity curve is
given by the relationship
( p + a)v = b( p0 − p)
where a, b, p0 are positive constants.
(a) Sketch the shortening velocity versus the load, i.e., v as a function
of p.
Note: the best way to do this is to find the intercepts of the two
axes, i.e. find the value of v corresponding to p = 0 and vice versa.
(b) Find the rate of change of the shortening velocity with respect to
the load, i.e. calculate dv/d p.
(c) What is the largest load for which the muscle contracts? (hint: a
contracting muscle has a positive shortening velocity, whereas a
muscle with a very heavy load stretches, rather than contracts, i.e.
has a negative value of v).
6.18. Reaction kinetics. Chemists often describe the rate of a saturating

chemical reaction using Michaelis-Menten (Rm ) or sigmoidal (Rs )
kinetics
Kc Kc2
Rm (c) = , Rs (c) = 2
kn + c kn + c2
where c is the concentration of the reactant, K > 0, kn > 0 are con-
stants. R(c) is the speed of the reaction as a function of the concentra-
tion of reactant.
(a) Sketch the two curves. To do this, you should analyze the be-
haviour for c = 0, for small c, and for very large c. You will find
a horizontal asymptote in both cases. We refer to that asymptote
as the “maximal reaction speed”. What is the “maximal reaction
speed” for each of the functions Rm , Rs ?
Note: express your answer in terms of the constants K, kn .

(b) Show that the value c = kn leads to a half-maximal reaction speed.
For the questions below, you may assume that K = 1 and kn = 1.
(c) Show that sigmoidal kinetics, but not Michaelis-Menten kinetics
has an inflection point.
(d) Explain how these curves would change if K is increased; if kn is
increased.
6.19. Checking the endpoints. Find the absolute maximum and minimum
values of the function
1
f ( x ) = x2 + 2
x
on the interval [ 21 , 2]. Be sure to verify if any critical points are max-
ima or minima and to check the endpoints of the interval.
7
Optimization
Calculus was developed to solve practical problems. In this chapter, we con-

centrate on optimization problems, where finding “the largest,” “the smallest,”
or “the best” answer is the goal. We apply some of the techniques developed
in earlier chapters to find local and global maxima and minima. A new Mastered Material Check
challenge in this chapter is translating a word-problem into a mathematical 1. How do you find the critical points
problem. We start with elementary examples, and work to more complex of a function f (x)?
situations with biological motivation.
7.1 Simple biological optimization problems
1. Given a function, find the derivative of that function and identify all
critical points.
2. Using a combination of sketching and tests for critical points developed in

Section 6.2, diagnose the type of critical point.
In the first examples, the function to optimize is specified, making the

problem simply one of carefully applying calculus methods.
Density dependent (logistic) growth in a population
Biologists often notice that the growth rate of a population depends not only
on the size of the population, but also on how crowded it is. Constant growth
is not sustainable. When individuals have to compete for resources, nesting
sites, mates, or food, they cannot invest time nor energy in reproduction,
leading to a decline in the rate of growth of the population. Such population
growth is called density dependent growth.
One common example of density dependent growth is called the logistic
growth law. Here it is assumed that the growth rate of the population, G
depends on the density of the population, N, as follows:

K −N

G(N ) = rN .
K
Here N is the independent variable, and G(N ) is the function of interest. Mastered Material Check
All other quantities are constant: 2. Give an example of units for N.
3. What units might G carry?
• r > 0 is a constant, called the intrinsic growth rate, and
G
• K > 0 is a constant, called the carrying capacity. It represents the popula-
tion density that a given environment can sustain.
Importantly, when differentiating G, we treat r and K as “numbers”. A

generic sketch of G as a function of N is shown in Figure 7.1.
Example 7.1 (Logistic growth rate) Answer the following questions:
a) Find the population density N that leads to the maximal growth rate G(N ). N
0 K/2 K
b) Find the value of the maximal growth in terms of r, K.
c) For what population size is the growth rate zero?

Figure 7.1: In logistic growth, the pop-
Solution. We can expand G(N ): ulation growth rate rate G depends on
population size N as shown here.
K −N

r
G(N ) = rN = rN − N 2 ,
K K
from which it is apparent that G(N ) is a polynomial in powers of N, with
constant coefficients r and r/K.
a) To find critical points of G(N ), we find N such that G0 (N ) = 0, and then

test for maxima:
r r K
G0 (N ) = r − 2 N = 0. ⇒ r=2 N ⇒ N= .
K K 2
Hence, N = K/2 is a critical point, but is it a maximum? We check this
in one of several ways. First, a sketch in Figure 7.1 reveals a downwards-
opening parabola. This confirms a local maximum. Alternately, we can
apply a tool from Section 6.2 such as the second derivative test:
r
G00 (N ) = −2 <0 ⇒ G(N ) concave down
K
K
⇒ N= is a local maximum
2
Thus, the population density with the greatest growth rate is K/2.
b) The maximal growth rate is found by evaluating the function G at the

critical point, N = K/2,
!
K − K2

K K K 1 rK
G =r =r · = .
2 2 K 2 2 4
O P T I M I Z AT I O N 147
c) To find the population size at which the growth rate is zero, we set G = 0
and solve for N:
K −N

G(N ) = rN = 0.
K
There are two solutions. One is trivial: N = 0. (This is biologically
interesting in the sense that it rules out the ancient idea of spontaneous
generation - a defunct theory that held that life can arise on its own,
from dust or air. If N = 0, the growth rate is also 0, so no population
spontaneously arises according to logistic growth.) The second solution,
N = K means that the population is at its “carrying capacity”.
♦
We return to this type of growth in Chapter 13.
Cell size for maximal nutrient accumulation rate

According to the model of Section 1.2, the nutrient absorption and consump-
tion rates, A(r ),C (r ), of a simple spherical cell of radius r are:
4
A(r ) = k1 S = 4k1 πr2 , C (r ) = k2V = πk2 r3 ,
3
for k1 , k2 > 0 constants. The net rate of increase of nutrients, which is the
difference of the two is: Mastered Material Check
4 4. Give example units for each
N (r ) = A(r ) −C (r ) = 4k1 πr2 − πk2 r3 . (7.1) of A(r ), C (r ) and N (r ). Are there
3
restrictions in place?
This quantity is a function of the radius r of the cell.
Example 7.2 Determine the radius of the cell for which the net rate of
increase of nutrients N (r ) is largest.
Solution. We are asked to maximize N (r ) with respect to r. We first find

critical points of N (r ), keeping in mind that 8k1 π and 4k2 π are constant for
the purpose of differentiation. Critical points occur when N 0 (r ) = 0, i.e.
N 0 (r ) = 8k1 πr − 4k2 πr2 = 0 ⇒ 4πr (2k1 − k2 r ) = 0

k1
⇒ r = 0, 2 .
k2
To identify the type of critical point, we use the second derivative test
N 00 (r ) = 8k1 π − 8k2 πr = 8π (k1 − k2 r ).
Substituting in r = 2k1 /k2 , we find that

00 k1 2k1
N 2 = 8π k1 − k2 = −8πk1 < 0.
k2 k2
Thus, the second derivative is negative at at r = 2k1 /k2 , verifying that this is a
local maximum. Hence the net rate of nutrient uptake is greatest for cells of
radius r = 2k1 /k2 . ♦
7.2 Optimization with a constraint
1. Set up an optimization word problem involving formulae for volume and

surface area of geometric solids.
2. Identify a constraint in an optimization problem.
3. Use the constraint to eliminate one of the independent variables, and find a
desired critical point. (As before, this includes classifying the critical point
as a local minimum, maximum or neither.)
In the next examples, identifying the function to optimize is part of the

challenge. We also consider cases with more than one independent variable,
where a constraint is used to eliminate all but one.
A cylindrical cell with minimal surface area

Not all cells are spherical. Some are cylindrical or sausage shaped. We
explore how minimization of surface area would determine the overall shape
of a cylindrical cell with a circular cross-section.
The volume of the cell is assumed to be fixed, because the cytoplasm in
its interior cannot be “compressed”. However, suppose that the cell has a
“rubbery” membrane that tends to take on the smallest surface area possible.
In physical language, the elastic energy stored in the membrane tends to a
minimum. We want to find the proportions of the cylinder (that is, the ratio of
length to radius) so that the cell has minimal surface area.
L
L
2rπ
Figure 7.2: Properties of a cylinder
Recall the following properties for a cylinder of length L and radius r:
• The volume of a cylinder is the product of its base area, A, and its height,
L. That is, V = AL. For a cylinder with circular cross-section, V = πr2 L.
• As in Figure 7.2, a cylinder can be “cut and unrolled” into a rectangle with
side lengths L and 2πr, where r is the radius of the circular cross-section.
The surface area is the product of these side lengths, Sside = 2πrL. Mastered Material Check
5. If a cylindrical cell has
• If the “ends” of the cylinder are two flat circular caps then the sum of the volume 100µm3 and length 10µm,
areas of these two ends is Sends = 2πr2 . While in a real cell, the end caps what is its radius?
would not be actually flat, for simplicity, we assume flat, circular ends. 6. What is the surface area of a
cylindrical cell with
volume 100µm3 and length 10µm?
• The total surface area of the cylinder with flat ends is then
S = 2πrL + 2πr2 .
Mathematically, our problem can be restated as follows
Example 7.3 Minimize the surface area S = 2πrL + 2πr2 of the cell, given
that its volume V = πr2 L = K is constant1 . 1
I would like to thank Prof Nima Geffen
(Tel Aviv University) for providing the
inspiration for this example.
Solution. The shape of the cell depends on both the length, L, and the radius,
r, of the cylinder. However, these are not independent. They are related to
one another because the volume of the cell has to be constant. This is an
example of an optimization problem with a constraint, i.e., a condition that
has to be satisfied. The constraint is “the volume is fixed”, i.e.,
V = πr2 L = K,
where K > 0 is a constant. This constraint allows us to eliminate one variables.

For example, solving for L, we have
K
L= . (7.2)
πr2
substituting this into the function S yields
K K
S = 2πrL + 2πr2 . ⇒ S(r ) = 2πr + 2πr2 ⇒ S (r ) = 2 + 2πr2 ,
πr2 r
where S is now a function of a single independent variable, r (K and π are
constants).
We have formulated the mathematical problem: find the minimum of S(r ).
We compute derivatives to find and classify the critical points:
K K
S0 (r ) = −2 + 4πr, S00 (r ) = 4 + 4π.
r2 r3
Since K, r > 0, the second derivative is always positive, so S(r ) is concave up.
Any critical point we find is thus automatically a minimum. (In Exercise 7 we
also consider the first derivative test as practice.) Setting S0 (r ) = 0:
K
S0 (r ) = −2 + 4πr = 0.
r2
Solving for r, we obtain

1/3
K 3 K K
2 2 = 4πr ⇒ r = ⇒ r= .
r 2π 2π
We can also find the length of this cell from Eqn. 7.2.
1/3
4K
L= .
π
(Details are left for Exercise 7).
We can finally characterize the shape of the cell. One way is to specify
the ratio of its radius to its length. Based on our previous results, we find that
ratio to be:
L
=2
r
(Exercise 7). This implies that L = 2r which coincides with the length of a
diameter of the same circle. Mastered Material Check
This means the “cylindrical cell” with rubbery membrane would be short 7. Redo Example 7.3 for a cell with
and fat - something almost like a sphere of radius r with flattened ends. Some fixed volume 100µm3 .
cells do grow as long cylindrical filaments, but such growth is inconsistent

with an elastic membrane or a minimal surface area. (In fact, filamentous
growth is one way for cells to maximize their surface area, and reduce the
challenge of absorbing nutrients.) ♦
Wine for Kepler’s wedding

In 1613, Kepler set out to purchase a few barrels of wine for his wedding
party. To compute the cost, the merchant would plunge a measuring rod
through the tap hole, as shown in Figure 7.3 and measure the length L of the
“wet” part of rod. The cost would be set at a value proportional to L.
Kepler noticed that barrels come in different shapes. Some are tall and
skinny, while others are squat and fat. He conjectured that some shapes
would contain larger volumes for a given length L, i.e. would contain more
wine for the same price. Knowing mathematics, he set out to determine
which barrel shape would be the best bargain for his wedding.
Figure 7.3: Barrels come in various shapes.

But the cost of a barrel of wine was deter-
mined by the length L (dashed blue line
segment) of the wet portion of the rod
inserted into the tap hole. Kepler figured out
L
which barrels contain the most wine for a
given price.
Kepler sought the wine barrel that contains the most wine for a given
cost. This is equivalent to asking which cylinder has the largest volume
for a fixed (constant) length L. Below, we solve this optimization problem.

An alternate approach is to seek the wine barrel that costs least for a given
volume (Exercise 14), which leads to the same result.
Example 7.4 Find the proportions (height:radius) of the cylinder with

largest volume for a fixed length L (dashed line segment in Figure 7.3).
Solution. We make the following assumptions:
1. the barrel is a simple cylinder, as shown in Figure 7.4,
2. the tap-hole (normally sealed to avoid leaks) is half-way up the height of

the barrel, and
3. the barrel is full to the top with wine.

Let r, h denote the radius and height of the barrel. These two variables 8. Give two different examples of
uniquely determine the shape as well as the volume of the barrel. Note that barrel dimensions which would both
yeild a volume of 160L.
because the barrel is assumed to be full, the volume of the cylinder is the
same as the volume of wine, namely
V = base area × height. ⇒ V = πr2 h. (7.3)
The rod used to “measure” the amount of wine (and hence determine the cost
of the barrel) is shown as the diagonal of length L in Figure 7.4. Because the
cylinder walls are perpendicular to its base, the length L is the hypotenuse
of a right-angle triangle whose other sides have lengths 2r and h/2. (This
follows from the assumption that the tap hole is half-way up the side.) Thus,
by the Pythagorean theorem,
2
2 2 h
L = (2r ) + . (7.4)
2
The problem can now be stated mathematically: maximize V in Eqn. (7.3)

subject to a fixed value of L in Eqn. (7.4). The fact that L is fixed means
that we have a constraint, as before, that we use to reduce the number of h
variables in the problem. L h
Expanding the squares in the constraint and solving for r2 leads to 2
h2 h2

1
L2 = 4r2 + ⇒ r2 = L2 − .
4 4 4
When we use this to eliminate r from the expression for V , we obtain

2r
h2

π π 1
V = πr2 h = L2 − h= L2 h − h3 . Figure 7.4: We simplify the problem to
4 4 4 4
a cylindrical barrel with diameter 2r and
The mathematical problem to solve is now: find h that maximizes height h. We assumed that the height of
the tap-hole is h/2. Length L denotes
the “wet” portion of the merchant’s rod,

π 2 1 3
V (h) = L h− h . used to determine the cost. We observe a
4 4 Pythagorean triangle formed by the dashed
line segments.
The function V (h) is positive for h in the range 0 ≤ h ≤ 2L, and V = 0 at the
two endpoints of the interval. We can restrict attention to this interval since
otherwise V < 0, which makes no physical sense. Since V (h) is a smooth
function, we anticipate that somewhere inside this range of values there
should be a maximal volume.
Computing first and second derivatives, we find

0 π 2 3 2 00 π 3 3
V (h) = L − h , V (h) = 0 − 2 · h = − πh < 0.
4 4 4 4 8
Setting V 0 (h) = 0 to find critical points, we then solve for h:
3
V 0 (h) = 0 ⇒ L2 − h2 = 0 ⇒ 3h2 = 4L2
4
L2 L
⇒ h2 = 4 ⇒ h = 2√ .
3 3
We verify that this solution is a local maximum by the following reasoning.
The second derivative V 00 (h) = − 38 πh < 0 is always negative for any
positive value of h, so V (h) is concave down for h > 0, which confirms a local
maximum. We also noted that V (r ) is smooth, positive within the range of
interest and zero at the endpoints. As there is only one critical point in that
range, it must be a local maximum.
Finally, we find the radius of the barrel by plugging the optimal h into the
constraint equation, i.e. using
h2 L2

2 1 2 1 2 1 2 2
r = L − = L − = L
4 4 4 3 4 3
1
⇒ r = √ √ L.
3 2
The shape of the optimal barrel can now be characterized. One way to do so
is to specify the ratio of its height to its radius. (Tall skinny barrels have a
largeh/r ratio, and squat fat ones have a low ratio.) By the above reasoning,
the ratio of h/r for the optimal barrel is Mastered Material Check
9. If all barrels had a radius of 25cm,
h 2 √L √ given the result Example 7.4, what
= 1 3 = 2 2. (7.5)
would be the best barrel height?
r √ √ L
3 2
10. What would the volume of such a
Hence, for greatest economy, Kepler would have purchased barrels with barrel be?
√
height to radius ratio of 2 2 = 2.82 ≈ 3. ♦ 11. Consider a barrel with radius 25cm
and height 100cm. What is this
barrel’s volume?
7.3 Checking endpoints
1. Recognize the distinction between local and global extrema.
2. Find the global minimum or maximum in a given word problem.

In some cases, the optimal value of a function does not occur at any of
its local maxima, but rather at one of the endpoints of an interval. Here we
consider such an example.
Example 7.5 (Maximal perimeter) The area of a rectangle with side

lengths x and y is A = xy. Suppose that the variable x is only allowed to take
on values in the range 0.5 ≤ x ≤ 4. Find the dimensions of the rectangle
having largest perimeter whose fixed area is A = 1. Mastered Material Check
12. How do you calculate the perimeter
Solution. The perimeter of a rectangle whose sides are length x, y is of a rectangle?
13. Why can a negative root
P = x + y + x + y = 2x + 2y. of P0 (x) = 0 in Example 7.5 be
rejected as irrelevant?
We are to maximize this quantity subject to the area of the rectangle being
fixed, A = xy = 1. This is the constraint. We use it to solve for and to eliminate
y from P.
1 2
y = , ⇒ P(x) = 2x + .
x x
We look for x that maximizes P(x). Computing the derivatives,

1 4
P0 (x) = 2 1 − 2 , P00 (x) = 3 > 0. perimeter
x x
10
P (x)
Setting P0 (x) = 0, we find critical points satisfying x2 = 1 or x = ±1. We
8
reject the negative root as irrelevant. We have found that P00 (x) > 0 for all
x > 0, so the critical point is a local minimum! This is clearly not the maximum 6
we were looking for. This example reinforces the importance of diagnostic 4

tests for the type of critical point.
2
Next, checking the endpoints of the interval, we evaluate P(4) = 8.5 and
P(0.5) = 5. The largest perimeter for the rectangle thus occurs when x = 4, at x
1 1.5 2 2.5 3 3.5 4
the right endpoint of the domain, as shown in Figure 7.5. ♦
In Appendix G.4, we provide further examples of optimization in the Figure 7.5: In Example 7.5, the critical point
context of geometric solids. is a local minimum. The maximum occurs at
the right end point of the interval 0.5 ≤ x ≤ 4.
7.4 Optimal foraging

Section 7.4 Learning goals 14. Use Figure 7.5 to estimate the side
length x when P(x) = 6.
1. Explain the development of a simple model for an animal foraging (col- 15. Verify your estimate algebraically.
lecting food to gain energy) in a food patch.
2. Interpret graphs of the rate of energy gain in various food patches, and
explain the distinctions between types of food patches.
3. Determine how long to spend foraging in a food patch in order to optimize

the average rate of energy gain per unit time.
Animals spend much of their time foraging - searching for food. Time
is limited, since when the sun goes down, the risk of becoming food (to a
predator) increases, and the likelihood of finding food decreases. Individuals
who are most successful at finding food over that limited time have the
greatest chance of surviving. It is argued by biologists that evolution tends to
optimize animal behaviour by selecting those that are faster, stronger, or more
fit, or - in this case - most efficient at finding food.
In this section, we investigate a model for optimal foraging. We follow the
basic principles of [Stephens and Krebs, 1986] and [Charnov, 1976].
Figure 7.6: A bird travels daily to forage in a

food patch. We want to determine how long
f
it should stay in the patch to optimize its
overall average energy gain per unit time.
food patch
travel time τ
time t
Notation. We define the following notation:
• τ= travel time between nest and food patch (this is considered to be time
that is unavoidably wasted).
• t = residence time in the patch (i.e. how long to spend foraging in one
patch), also called foraging time,
• f (t ) = total energy gained by foraging in a patch for time t.
Energy gain in food patches. In some patches, food is ample and found Mastered Material Check
quickly, while in others, it takes time and effort to obtain. The typical time 16. Which of the energy gain functions
needed to find food is reflected by various energy gain functions f (t ) shown in Figure 7.7 are strictly increasing?
in Figure 7.7.
Example 7.6 (Energy gain versus patch residence time) For each panel
in Figure 7.7, explain what the graph of the total energy gain f (t ) is saying
about the type of food patch: how easy or hard is it to find food?
Solution. The types of food patches are as follows:
1. The energy gain is linearly proportional to time spent in the patch. In this
case, the patch has so much food that it is never depleted. It would make
sense to stay in such a patch for as long as possible.
2. Energy gain is independent of time spent. The animal gets the full quantity
as soon as it gets to the patch.
f (t) f (t) f (t) Figure 7.7: Examples of various total energy

gain f (t ) for a given foraging time t. The
shapes of these functions determine how
hard or easy it is to extract food from a food
patch.
t t t
(a) (b) (c)
f (t) f (t) f (t)
t t t
(d) (e) (f)
3. Food is gradually depleted, (the total energy gain levels off to some
constant as t increases). There is “diminishing return” for staying longer,
suggesting that it is best not to stay too long. 17. Which model(s) can you
automatically dismiss as not very
biologically realistic?
4. The reward for staying longer in this patch increases: the net energy gain
is concave up ( f 00 (t ) > 0), so its slope is increasing.
5. It takes time to begin to gain energy. After some time, the gain increases,
but eventually, the patch is depleted.
6. Staying too long in such a patch is disadvantageous, resulting in net loss of

energy. It is important to leave this patch early enough to avoid that loss.
Example 7.7 Consider the hypothetical patch energy function

Emaxt
f (t ) = where Emax , k > 0, are constants. (7.6)
k +t
a) Match this function to one of the panels in Figure 7.7.
b) Interpret the meanings of the constants Emax , k.
Solution.
a) The function resembles Michaelis-Menten kinetics (Figure 1.8). In
Figure 7.7, Panel (3) is the closest match.
b) From Chapter 1, Emax is the horizontal asymptote, corresponding to an

upper bound for the total amount of energy that can be extracted from the
patch. The parameter k has units of time and controls the steepness of the
function. Foraging for a time t = k, leads the animal to obtain half of the
total available energy, since f (k) = Emax /2 (Exercise 27(a)). ♦
Example 7.8 (Currency to optimize) We can assume that animals try to

maximize the average energy gain per unit time, defined by the ratio:
Total energy gained

R(t ) = ,
total time spent
Write down R(t ) for the assumed patch energy function Eqn. 7.6.
Solution. The ‘total time spent’ is a sum of the fixed amount of time τ
traveling, and time t foraging. The ‘total energy gained’ is f (t ). Thus, for the
patch function f (t ) assumed in Eqn. (7.6), Mastered Material Check
18. What units might be used in the
f (t ) Emaxt function R(t )?
R(t ) = = . (7.7)
(τ + t ) (k + t )(τ + t )
♦
We can now state the mathematical problem:
Find the time t that maximizes R(t ).
In finding such a t we are determining the optimal residence time.
Example 7.9 Use tools of calculus and curve-sketching to find and classify
the critical points of R(t ) in Eqn. (7.7).
Solution. We first sketch R(t ), focusing on t > 0 for biological relevance.
• For t ≈ 0, we have R(t ) ≈ (Emax /kτ )t, which is a linearly increasing

function.
• As t → ∞, R(t ) → Emaxt/t 2 → 0, so the graph eventually decreases to zero.
These two conclusions are shown in Figure 7.8 (left panel), and strongly
suggest that there should be a local maximum in the range 0 < t < ∞, as shown
in the right panel of Fig 7.8. Since the function is continuous for t > 0, this
Figure 7.8: In Example 7.9 we first compose

a rough sketch of the average rate of energy
R(t) R(t)
gain R(t ) in Eqn. (7.7). The graph is linear
near the origin, and decays to zero at large t.
t t
sketch verifies that there is a local maximum for some positive t value.
To find a local maximum, we compute R0 (t ) using the quotient rule

(Exercise 27c), and set R0 (t ) = 0:
kτ − t 2
R0 (t ) = Emax = 0. (7.8)
(k + t )2 (τ + t )2
This can only be satisfied if the numerator is zero, that is
√
kτ − t 2 = 0 ⇒ t1,2 = ± kτ.
Rejecting the (irrelevant) negative root, we deduce that the critical point of
√
the function R(t ) is tcrit = kτ. The sketch in Figure 7.8, verifies that this
critical point is a local maximum. ♦
Example 7.10 For practice, use one of the calculus tests for critical points to
√
show that tcrit = kτ is a local maximum for the function R(t ) in Eqn. (7.7).
Solution. R(t ) is a rational function, so a second derivative is messy. Instead,

we apply the first derivative test (Section 6.2), that is, we check the sign of
R0 (t ) on both sides of the critical point.
• Eqn. (7.8) gives R0 (t ). Its denominator is positive, so the sign of R0 (t ) is

determined by its numerator, (kτ − t 2 ).
• Thus, R0 (t ) > 0 for t < tcrit , and R0 (t ) < 0 for t > tcrit .
This confirms that the function increases up to the critical point and decreases
afterwards, so the critical point is a local maximum, henceforth denoted tmax .
♦
To optimize the average rate of energy gain, R(t ), we found that the
√
animal should stay in the patch for a duration of t = tmax = kτ. Mastered Material Check
19. Given tmax is the duration of time an
Example 7.11 Determine the average rate of energy gain at this optimal animal should stay in a patch, and τ
patch residence time, i.e. find the maximal average rate of energy gain. is travelling time, explain why the
constant k is also in units of time.
√
Solution. Computing R(t ) for t = tmax = kτ, we find that
Emaxtmax Emax 1
R(tmax ) = = √ . (7.9)
(k + tmax )(τ + tmax ) τ (1 + k/τ )2
The reader is asked to fill in the steps for this calculation in Exercise 27(d). ♦
In Appendix G.5, we extend this example by considering a more general
problem. Geometric arguments play a key role in that discussion.
7.5 Summary
1. Optimization is a process of finding critical points, and identifying local

and global maxima/minima.
2. A scientific problem that address “biggest/smallest, best, most efficient” is
often reducible to an optimization problem.
3. As with all mathematical models, translating scientific observations and

reasonable assumptions into mathematical terms is an important first step.
4. The following applications were considered:
(a) Density dependent population growth. Using a given logistic growth

law, the following parameters were considered:
• population growth rate (to be maximized),
• population density,
• intrinsic growth rate (constant),
• carrying capacity (constant).
(b) Nutrient absorption in a cell. Using the model developed in Section 1.2
for a spherical cell, we considered:
• nutrient absorption rate,

• nutrient consumption rate,
• cell radius,
• proportionality constants (determined based on context).
We maximized the net rate of increase of nutrients - a difference

between absorption and consumption rates.
(c) Surface area of a cylindrical cell, which tends to be minimized do to
energy conditions. The parameters we used were:
• cell length,
• cell radius,
• cell volume (constant),
• cell surface area (to be minimized).
(d) Wine for Kepler’s wedding, seeking the largest barrel volume for a
fixed diagonal length. The following parameters were considered:
• barrel volume, (to be maximized)

• barrel height,
• barrel radius,
• length of the diagonal (constant).
(e) Foraging time for an animal collecting food. We considered:
• travel time between nest and food patch,

• foraging time in the patch,
• energy gained by foraging in a patch for various time durations.
1. If the growth rate of a population follows the following logistic equation:
50000 − N

G(N ) = 1.2N ,
50000
where N is the density of the population, under what circumstances is the population growing fastest?
2. When finding a global maximum, why is always imperative to check the endpoints?
3. Demonstrate the variability of barrel dimensions by giving two different height and radius pairs which lead to a
volume of 50L.
4. Would the answer to Kepler’s wine barrel problem have changed if we had solved for h2 instead of r2 ?
Exercises
7.1. Find the numbers. The sum of two positive number is 20. Find the
numbers
(a) if their product is a maximum,
(b) if the sum of their squares is a minimum,
(c) if the product of the square of one and the cube of the other is a
maximum.
7.2. Distance, velocity and acceleration. A tram ride departs from its
starting place at t = 0 and travels to the end of its route and back. Its
distance from the terminal at time t can be approximately described by
the expression
S(t ) = 4t 3 (10 − t )
where t is in minutes, 0 < t < 10, and S is distance in meters.

(a) Find the velocity as a function of time.
(b) When is the tram moving at the fastest rate?
(c) At what time does it get to the furthest point away from its starting
position?
(d) Sketch the acceleration, the velocity, and the position of the tram on
the same set of axes.
7.3. Distance of two cars. At 9A.M., car B is 25 km west of car A. Car A
then travels to the south at 30 km/h and car B travels east at 40 km/h.
When are they closest to each other and what is this distance?
7.4. Cannonball movement. A cannonball is shot vertically upwards from
the ground with initial velocity v0 = 15m/sec. The height of the ball, y
(in meters), as a function of the time, t (in sec) is given by
y = v0t − 4.9t 2
Determine the following:

(a) the time at which the cannonball reaches its highest point,
(b) the velocity and acceleration of the cannonball at t = 0.5 s, and t =
1.5 s, and
(c) the time at which the cannonball hits the ground.
7.5. Net nutrient increase rate. In Example 7.2, we considered the net
rate of increase of nutrients in a spherical cell of radius r. Here we
further explore this problem.
(a) Draw a sketch of N (r ) based on Eqn. (7.1). Use your sketch to
verify that this function has a local maximum.
(b) Use the first derivative test to show that the critical point r = 2k1 /k2
is a local maximum.
7.6. Nutrient increase in cylindrical cell. Consider a long skinny cell in
the shape of a cylinder with radius r and a fixed length L. The volume
and surface area of such a cell (neglecting endcaps) are V = πr2 L = K
and S = 2πrL.
(a) Adapt the formula for net rate of increase of nutrients N (t ) for a
spherical cell Eqn. (7.1) to the case of a cylindrical cell.
(b) Find the radius of the cylindrical cell that maximizes N (t ). Be sure
to verify that you have found a local maximum.
7.7. Cylinder of minimal surface area. In this exercise we continue to
explore Example 7.3.
(a) Reason that the surface area of the cylinder, S(r ) = 2 Kr + 2πr2 is a
function that has a local minimum using curve-sketching.
K 1/3

(b) Use the first derivative test to show that r = 2π is a local
minimum for S(r ).
(c) Show the algebra required to find the value of L corresponding to
this r value and show that L/r = 2.
7.8. Dimensions of a box. A closed 3-dimensional box is to be constructed
in such a way that its volume is 4500 cm3 . It is also specified that the
length of the base is 3 times the width of the base.
Determine the dimensions of the box which satisfy these conditions
and have the minimum possible surface area. Justify your answer.
7.9. Dimensions of a box. A box with a square base is to be made so that
its diagonal has length 1; see Figure 7.9.
(a) What height y would make the volume maximal?
(b) What is the maximal volume? (hint: a box having side lengths `, w, h
has diagonal length D where D2 = `2 + w2 + h2 and volume V =
`wh).
Figure 7.9: Figure for Exercise 9; box with a

square base.
y D
x
x
7.10. Minimum distance. Find the minimum distance from a point on the
positive x-axis (a, 0) to the parabola y2 = 8x.
7.11. The largest garden. You are building a fence to completely enclose
part of your backyard for a vegetable garden. You have already
purchased material for a fence of length 100 ft.
What is the largest rectangular area that this fence can enclose?
7.12. Two gardens. A fence of length 100 ft is to be used to enclose two
gardens. One garden is to have a circular shape, and the other to be
square.
Determine how the fence should be cut so that the sum of the areas
inside both gardens is as large as possible.
7.13. Dimensions of an open box. A rectangular piece of cardboard with
dimension 12 cm by 24 cm is to be made into an open box (i.e., no
lid) by cutting out squares from the corners and then turning up the
sides.
Find the size of the squares that should be cut out if the volume of the
box is to be a maximum.
7.14. Alternate solution to Kepler’s wine barrel. In this exercise we
follow an alternate approach to the most economical wine barrel
problem posed by Kepler (as in Example 7.4).
Through this approach, we find the proportions (height:radius) of the
cylinder that minimizes the length L of the wet rod in Figure 7.3 for a
fixed volume.
(a) Explain why minimizing L is equivalent to minimizing L2 in
Eqn. (7.4)
(b) Explain how Eqn. (7.3) can be used to specify a constraint for this
problem. (hint: consider the volume, V to be fixed and show that
you can solve for r2 ).
(c) Use your result in (c) to eliminate r from the formula for L2 .
Now L2 (h) depends only on the height of the cylindrical wine
barrel.
(d) Use calculus to find any local minima for L2 (h). Be sure to verify
that your result is a minimum.
(e) Find the corresponding value of r using your result in (b).
(f) Find the ratio h/r. You should obtain the same result as in
Eqn. (7.5).
7.15. Rectangle with largest area. Find the side lengths, x and y, of the
rectangle with largest area whose diameter L is given (hint: eliminate
one variable using the constraint. To simplify the derivative, consider
that critical points of A would also be critical points of A2 , where A =
xy is the area of the rectangle. If you have already learned the chain
rule, you can use it in the differentiation).
7.16. Shortest path. Find the shortest path that would take a milk-maid
from her house at (10, 10) to fetch water at the river located along
the x-axis and then to the thirsty cow at (3, 5).
7.17. Water and ice. Why does ice float on water? Because the density of
ice is lower! In fact, water is the only common liquid whose maximal
density occurs above its freezing temperature. This phenomenon
favours the survival of aquatic life by preventing ice from forming
at the bottoms of lakes. According to the Handbook of Chemistry
and Physics, a mass of water that occupies one liter at 0◦ C occupies a
volume (in liters) of
V = −aT 3 + bT 2 − cT + 1
at T ◦ C where 0 ≤ T ≤ 30 and where the coefficients are
a = 6.79 × 10−8 , b = 8.51 × 10−6 , c = 6.42 × 10−5 .
Find the temperature between 0◦ C and 30◦ C at which the density of

water is the greatest. (hint: maximizing the density is equivalent to
minimizing the volume. Why is this?).
7.18. Drug doses and sensitivity. The reaction R(x) of a patient to a drug
dose of size x depends on the type of drug. For a certain drug, it was
determined that a good description of the relationship is:
R(x) = Ax2 (B − x)
where A and B are positive constants. The sensitivity of the patient’s

body to the drug is defined to be R0 (x).
(a) For what value of x is the reaction a maximum, and what is that
maximum reaction value?
(b) For what value of x is the sensitivity a maximum? What is the
maximum sensitivity?
7.19. Thermoregulation in a swarm of bees. In the winter, honeybees
sometimes escape the hive and form a tight swarm in a tree, where,
by shivering, they can produce heat and keep the swarm temperature
elevated.
Heat energy is lost through the surface of the swarm at a rate propor-
tional to the surface area (k1 S where k1 > 0 is a constant). Heat energy
is produced inside the swarm at a rate proportional to the mass of the
swarm (which you may take to be a constant times the volume). We
assume that the heat production is k2V where k2 > 0 is constant.
Swarms that are not large enough may lose more heat than they can
produce, and then they die. The heat depletion rate is the loss rate
minus the production rate. Assume that the swarm is spherical.
Find the size of the swarm for which the rate of depletion of heat
energy is greatest.
7.20. Cylinder inside a sphere. Work through the steps for the calculations
and classification of critical point(s) in Example G.2, that is, find
the dimensions of the largest cylinder that would fit in a sphere of
radius R.
7.21. Circular cone circumscribed about a sphere. A right circular cone
is circumscribed about a sphere of radius 5. Find the dimension of this
cone if its volume is to be a minimum.
Note: this is a rather challenging geometric problem.
7.22. Optimal reproductive strategy. Animals that can produce many
healthy babies that survive to the next generation are at an evolution-
ary advantage over other, competing, species. However, too many
young produce a heavy burden on the parents (who must feed and care
for them). If this causes the parents to die, the advantage is lost. Fur-
ther, competition of the young with one another for food and parental
attention jeopardizes the survival of these babies.
Suppose that the evolutionary Advantage A to the parents of having
litter size x is
A(x) = ax − bx2 .
Suppose that the Cost C to the parents of having litter size x is
C (x) = mx + e.
The Net Reproductive Gain G is defined as
G = A −C.
(a) Explain the expressions for A,C and G.

(b) At what litter size is the advantage, A, greatest?
(c) At what litter size is there least cost to the parents?
(d) At what litter size is the Net Reproductive Gain greatest?.
7.23. Behavioural Ecology. Social animals that live in groups can spend
less time scanning for predators than solitary individuals. However,
they waste time fighting with the other group members over the avail-
able food. There is some group size at which the net benefit is greatest
because the animals spend the least time on these unproductive activi-
ties - and thus can spend time on feeding, mating, etc.
Assume that for a group of size x, the fraction of time spent scanning
for predators is
1
S (x ) = A
(x + 1)
and the fraction of time spent fighting with other animals over food is
F (x ) = B(x + 1)2
where A, B are constants.

Find the size of the group for which the time wasted on scanning and
fighting is smallest.
7.24. Logistic growth. Consider a fish population whose density (indi-
viduals per unit area) is N, and suppose this fish population grows
logistically, so that the rate of growth R satisfies G
R(N ) = rN (1 − N/K )
where r and K are positive constants.

(a) Sketch R as a function of N or explain Figure 7.10.
(b) Use a first derivative test to justify the claim that N = K/2 is a local
maximum for the function G(N ). N
7.25. Logistic growth with harvesting. Consider a fish population of 0 K/2 K

density N growing logistically, i.e. with rate of growth R(N ) =
rN (1 − N/K ) where r and K are positive constants. The rate of
Figure 7.10: In logistic growth, the popula-
harvesting (i.e. removal) of the population is
tion growth rate G depends on population
size N as shown here.
h(N ) = qEN
where E, the effort of the fishermen, and q, the catchability of this type
of fish, are positive constants.
At what density of fish does the growth rate exactly balance the
harvesting rate? This density is called the maximal sustainable yield:
MSY.
7.26. Conservation of a harvested population. Conservationists insist
that the density of fish should never be allowed to go below a level at
which growth rate of the fish exactly balances with the harvesting rate.
At this level, the harvesting is at its maximal sustainable yield. If more
fish are taken, the population keeps dropping and the fish eventually
go extinct.
What level of fishing effort should be used to lead to the greatest
harvest at this maximal sustainable yield?
Note: you should first complete the Exercise 25.
7.27. Optimal foraging. Consider Example 7.7 for the optimal foraging
model.
(a) Show that the parameter k in Eqn. (7.6) is the time at which f (t ) =
Emax /2.
(b) Consider panel (5) of Figure 7.7. Show that a function such as a
Hill function would have the shape shown in that sketch. Interpret
any parameters in that function.
(c) Use the quotient rule to calculate the derivative of the function R(t )
given by Eqn. (7.7) and show that you get Eqn. (7.8).
(d) Fill in the missing steps in the calculation in Eqn. (7.9) to find the
optimal value of R(t ).
7.28. Rate of net energy gain while foraging and traveling. Animals
spend energy in traveling and foraging. In some environments this
energy loss is a significant portion of the energy budget. In such cases,
it is customary to assume that to survive, an individual would optimize
the rate of net energy gain, defined as
Net energy gained Energy gained − Energy lost
Q(t ) = = (7.10)
total time spent total time spent
Assume that the animal spends p energy units per unit time in all
activities (including foraging and traveling). Assume that the energy
gain in the patch (“patch energy function”) is given by Eqn. (7.6).
Find the optimal patch time, that is the time at which Q(t ) is maxi-
mized in this scenario.
7.29. Maximizing net energy gain: Suppose that the situation requires an
animal to maximize its net energy gained E (t ) defined as
E (t ) = energy gained while foraging

−energy spent while foraging and traveling.
(This means that E (t ) = f (t ) − r (t + τ ) where r is the rate of energy

spent per unit time and τ is the fixed travel time).
Assume as before that the energy gained by foraging for a time t in the
food patch is f (t ) = Emaxt/(k + t ).
(a) Find the amount of time t spent foraging that maximizes E (t ).
(b) Indicate a condition of the form k < ? that is required for existence
of this critical point.
8
Introducing the chain rule

So far, examples were purposefully chosen to focus on power, polynomial,
1. Give an example each of:
and rational functions that are each relatively easy to differentiate. We
(a) a power function,
now introduce the differentiation rule that opens up our repertoire to more
(b) a polynomial function, and
elaborate examples involving composite functions. This allows us to model
(c) a rational function.
more biological processes. We dedicate this chapter to the chain rule and its
applications.
8.1 The chain rule
1. Summarize function composition and express a composite function in

terms of the underlying composed functions.
2. Produce the chain rule of differentiation and apply it to find the derivative
of a composite function.
Function composition
Consider Figure 8.1 which depicts a function composition: an independent
variable, x, is used to evaluate a function, and the result, u = f (x) then acts as
an input to a second function, g. The final value is y = g(u) = g( f (x)).
Figure 8.1: Function composition.

u
x f g y
We refer to this two-step function operation as function composition.

√
Example 8.1 Consider the functions f (x) = x and g(x) = x2 + 1. Deter-
mine the functions obtained by composing these:
a) h1 (x) = g( f (x))
b) h2 (x) = f (g(x))
Solution.
√
a) For h1 we apply f first, followed by g, so h1 (x) = ( x)2 + 1 = x + 1
(provided x ≥ 0.)
b) For h2 , the functions are applied in the reversed order so that h2 (x) =
√
x2 + 1 (for any real x).
We note that the domains of the two functions are slightly different: h1 is only
defined for x ≥ 0 since f (x) is not defined for negative x, whereas h2 is defined
for all x. ♦ Mastered Material Check
2. Can you define the domain of a
Example 8.2 Express the function h(x) = 5(x3 − x2 )10 as the composition of function?
two simpler functions. 3. What are the domains of f (x)
and g(x) found in Example 8.2?
Solution. We can write this in terms of the two functions f (x) = x3 − x2 and
g(x) = 5x10 . Then h(x) = g( f (x)). ♦
The chain rule of differentiation

Given a composite function y = f (g(x)), we require a rule for differentiating y
with respect to x.
If y = g(u) and u = f (x) are both differentiable functions and y = g( f (x)) is

the composite function, then the chain rule of differentiation states that
dy dy du
= .
dx du dx
Informally, the chain rule states that the change in y with respect to x is a
product of two rates of change:
1. the rate of change of y with respect to its immediate input u, and
2. the rate of change of u with respect to its input, x.
Why does it work this way? Although the derivative is not a simple quotient,
we gain an intuitive grasp of the chain rule by writing
dy ∆y ∆y ∆u
≈ =
dx ∆x ∆u ∆x
then it is apparent that the “cancellation” of terms ∆u in numerator and
denominator lead to the correct fraction on the left. The proof of the chain
rule uses this essential idea, but care is taken to ensure that the quantity
∆u is nonzero, to avoid the embarrassment of dealing with the nonsensical
ratio 0/0. The proof of the chain rule is found in Appendix E.4.
I N T RO D U C I N G T H E C H A I N RU L E 169
Example 8.3 Compute the derivative of the function h(x) = 5(x3 − x2 )10 .
Solution. We express the function as y = h(x) = 5u10 where u = (x3 − x2 )

and apply the chain rule. Then
d (5u10 ) d ( x3 − x2 )

dy dy du
= = = 50u9 (3x2 − 2x).
dx du dx du dx
Substituting for u leads to dy/dx = 50(x3 − x2 )9 (3x2 − 2x). ♦

p
Example 8.4 Compute the derivative of the function y = f (x) = x 2 + a2 ,
where a is some positive real number.
i A quick summary of the first and
√
Solution. This function can be considered as the composition of g(u) = u = second derivatives of the function in
Example 8.4 for the case a = 1.
u1/2 and u(x) = x2 + a2 , that is, we can write f (x) = g(h(x)). Then using the
chain rule, we obtain
dy 1 x x
= · (x2 + d 2 )−1/2 · 2x = 2 =√ .
dx 2 (x + d 2 )1/2 x + d2
2
Example 8.5 Compute the derivative of the function

x
y = f (x ) = √ ,
x + d2
2
where d is some positive real number.
Solution. We use both the quotient rule and the chain rule for this calcula-
tion. √ √
dy [ x ] 0 · x2 + d 2 − [ x2 + d 2 ] 0 · x
= √ .
dx ( x2 + d 2 ) 2
Here the [. . . ]0 denotes differentiation. Then
√
dy 1 · x2 + d 2 − [ 12 · 2x · (x2 + d 2 )−1/2 ] · x
= .
dx ( x2 + d 2 )
We simplify algebraically by multiplying top and bottom by (x2 + d 2 )1/2 and

cancelling factors of 2 to obtain
dy x2 + d 2 − x2 d2
= 2 = 2 .
dx (x + d ) (x + d ) (x + d 2 )3/2
2 1/2 2 2
Interpreting the chain rule

While the chain rule is not rigorously proved here (see Appendix E.4), we
hope to extend our intuition about where it comes from. The following
intuitive examples may help to motivate why the chain rule is based on a
product of two rates of change.
Example 8.6 (Pollution level in a lake) A species of fish is sensitive to

pollutants in a lake. As humans populate the area around the lake, the fish
population declines due to increased pollution levels. Quantify the rate at
which the pollution level changes with time based on the pollution produced
per human and the rate of increase of the human population.
Solution. The rate of change (decline in this case) of the fish population
depends on:
• the rate of change of the human population, and
• the rate of change in the pollution created per person.
If either increases, the effect on the fish population increases.

The chain rule implies that the net effect is a product of the two rates.
Formally, for t time in years, x = f (t ) the number of people at the lake in year
t, and p = g(x) the pollution created by x people, the rate of change of the
4. If pollution is measured in
pollution p over time is a product of g0 (x) and f 0 (t ): micrograms per cubic meter
(µg/m3 ), give units for p
dp d p dx
= = g0 (x) f 0 (t ). and d p/dt.
dt dx dt
♦
Example 8.7 (Population of carnivores, prey, and vegetation) The pop-

ulation of large carnivores, C, on the African savannah depends on the
population of gazelles that are their prey, P. The gazelle population, in turn,
depends on the abundance of vegetation V , which depends on the amount of
rain in a given year, r. Quantify the rate of change of the carnivore popula-
tion with respect to rainfall.
Solution. We can express these dependencies through hypothetical functions

such as V = g(r ), P = f (V ) and C = h(P), each depicting a relationship in
the food chain.
A drought that decreases rainfall also decreases the abundance of vege-
tation. This then decreases the gazelle population, and eventually affect the
population of carnivores. The rate of change in the carnivores population with
respect to the rainfall, dC/dr, according to the chain rule, would be
dC dC dP dV
= · · .
dr dP dV dr
♦
Example 8.8 (Population of carnivores) Consider Example 8.7, with the

specific example of the relationships of carnivores and prey, C = h(P) = P2 ,
prey on vegetation, P = f (V ) = 2V , and vegetation on rainfall, V = g(r ) =
r1/2 (Figure 8.2). Quantify the rate of change of the carnivore population
with respect to rainfall.
C P V
P V r
Figure 8.2: An example in which the

population of carnivores, depends on prey
Solution. Computing all necessary derivatives, P acording to C = h(P) = P2 , while the
prey depend on vegetation, P = f (V ) = 2V ,
dV 1 dP dC and the vegetation depends on rainfall
= r−1/2 , = 2, = 2P, V = g(r ) = r1/2 .
dr 2 dV dP
so that, from Example 8.7, the rate of change in the carnivores with respect to
the rainfall is
dC dC dP dV 1 2P
= = r−1/2 (2)(2P) = 1/2 . Mastered Material Check
dr dP dV dr 2 r
5. Take an alternative approach to
Using the fact that V = r1/2 and P = 2V , we obtain Example 8.8 by expressing the
number of carnivores C explicitly in
dC 2P 2(2V ) terms of rainfall r, and then
= = = 4.
dr V V differentiating. Verify that both
approaches yield the same solution.
♦
6. While both approaches work in this
case, why might they not in general?
Example 8.9 (Budget for coffee) Your budget for coffee depends on the
number of cups consumed per day and on the price per cup. The total budget
changes if either price or the consumption goes up. Define appropriate
variables and quantify the rate at which the coffee budget changes if both
consumption and price change.
Solution. The total rate of change of the coffee budget is a product of the
change in the price and the change in the consumption. For t time in days,
x = f (t ) the number of cups of coffee consumed, and y = g(x) the price for x
cups of coffee, we obtain
dy dy dx
= = g0 (x) f 0 (t ).
dt dx dt
♦
Example 8.10 (Earth’s temperature and greenhouse gases) In Exercise 21

of Chapter 1, we found that the temperature of the Earth depends on the
albedo a (fraction of incoming radiation energy reflected) according to the
formula
(1 − a)S 1/4

T= . (8.1)
εσ
Assume that the albedo a is the only quantity that depends on the level of
greenhouse gases G, and that da/dG is known.Determine how the tempera-
ture changes as the level of greenhouse gases G increases.
Solution. Temperature T depends on the level of greenhouse gases G through 7. List all constants in Example 8.10.
the albedo a, so we write T (a(G)). Here S, ε, σ are all constants, so it 8. List all variables in Example 8.10.
simplifies calculation to rewrite T as
1/4
S
T (a) = (1 − a)1/4 .
εσ
According to the chain rule,
dT dT da
= .
dG da dG
Then
1/4 1/4
dT S d h i da S 1 da
= (1 − a)1/4 = (1 − a)(1/4)−1 · (−1) .
dG εσ da dG εσ 4 dG
Rearranging leads to
1/4
dT 1 S da
=− (1 − a)−3/4 .
dG 4 εσ dG
In general, greenhouse gasses affect both the Earth’s albedo a and its emissiv-
ity ε. We generalize our results in Exercise 3. ♦
8.2 The chain rule applied to optimization problems
1. Read and interpret the derivation of each optimization model.
2. Carry out the calculations of derivatives appearing in the problems (using

the chain rule).
3. Using optimization, find each critical point and identify its type.
4. Explain the interpretation of the mathematical results.
Armed with the chain rule, we can now differentiate a wider variety of
functions, and address problems that were not tractable with the power,
product, or quotient rules alone. We return to optimization problems where
derivatives require use of the chain rule.
Shortest path from food to nest

Ants are good mathematicians! They can find the shortest route connecting
their nest to a food source. But how do they do it? Each ant secretes a
chemical pheromone that other ants tend to follow. This marks up the
trail that they use and recruits nest-mates to food sources. The pheromone
(chemical message for marking a route) evaporates after a while, so that, for
a fixed given number of foraging ants, a longer trail has a less concentrated
chemical marking than a shorter trail. This means that whenever a shorter
route is found, the ants favour it. After some time, this leads to selection of
the shortest possible trail.
Shown in Figure 8.3 is a common laboratory test scenario: ants in an
artificial nest are offered two equivalent food sources. We ask: what is the
shortest total path connecting nest and sources? This is a simplified version
of the problem that the ants are solving.
Example 8.11 (Minimizing the total path length) Use the diagram to
determine the length of the shortest path that connects the nest to both food
sources. Assume that d << D.
Solution. We first consider two possibile paths connecting nest to food:
1. a V-shaped path and
2. a T-shaped path.
√
The length of a V-shaped path is 2 D2 + d 2 , whereas the length of a T-
shaped path is D + 2d. Mastered Material Check
Now consider a third possibility: a Y-shaped path, where the ants first 9. If D = 2m and d = 20cm, how long
walk straight ahead and then veer off to the left and right. All three possibili- is the V -shaped path? The T -shaped
path?
ties are shown in Figure 8.3.
Food d Food Food d Food Food d Food
D D D
Nest Nest Nest

(a) (b) (c)
Figure 8.3: Three ways to connect the

ants’ nest to two food sources, showing
Calculations are easiest if we denote the distance from the nest to the (a) V-shaped, (b) T-shaped, and (c) Y-shaped
Y-junction as D − x, so that x is distance shown in the diagram. The length of paths.
the Y-shaped path is then

p
LY = L(x) = (D − x) + 2 d 2 + x2 . (8.2)
Observe that when x = 0, then LT = D + 2d, which corresponds to the
√
T-shaped path length, whereas when x = D, then LV = 2 d 2 + D2 which
V-shaped path length. Thus, 0 ≤ x ≤ D is the appropriate domain, and we have
determined the values of L at the two domain endpoints.
To find the minimal path length, we look for critical points of the function
L(x). Differentiating (see Examples 8.4 and 8.5), we find
x d2
L0 (x) = −1 + 2 √ , L00 (x) = 2 > 0.
x + d2
2 (x2 + d 2 )3/2
Then critical points occur when
x
L0 ( x ) = 0 ⇒ −1 + 2 √ = 0.
x2 + d 2
Simplifying leads to
p d
x2 + d 2 = 2x ⇒ x2 + d 2 = 4x2 ⇒ 3x2 = d 2 x= √ .
⇒
3
To determine the type of critical point, we note that the second derivative is
positive and so the critical point is a local minimum.
Figure 8.4: (a) In the configuration √

for the
shortest path we found that x = d/ 3.
d (b) The total length of the path L(x) as a
function of x for D = 2, d √
= 1. The minimal
L(x)
x= √d 4.4 path occurs when x = 1/ 3 ≈ 0.577.
3
The length
√ of the shortest
√ path is then
L = D + 3d = 2 + 3 ≈ 3.73.
4.2
D
4
3.8 10. For D = 2m and d = 20cm, what is
x
the shortest path length for the ants?
0.5 1 1.5 2
(a) (b)
√
To determine the actual length of the path, we substitute x = d/ 3 into the
function L(x) and obtain (after simplification, see Exercise 4)
√
L = L(x) = D + 3d.
The final result is summarized in Figure 8.4. The shortest path is Y-shaped,
√ √
with x = d/ 3. The ants march straight for a distance D − (d/ 3), and then
their trail branches to the right and left towards the food sources. Does this happen really?
♦
Featured Problem 8.1 (Most economical wooden beam) A cylindrical
tree trunk is to be cut into a rectangular wooden beam. What is the most
economical way to cut the beam so as to waste the least amount of material?
Figure 8.5: A cylindrical trunk can produce
a variety of wooden beams. Which one
uses the most wood and wastes as little as
possible?
Food choice and attention

The following example is based on research about animal behaviour. As
before, we consider how animals should spend their time, but now a choice is
provided between two types of food.
Paying attention. Many types of food are cryptic - hidden in the environ-
ment - and so, require time and attention to find. Some types of food are
more easily detected, but other foods might provide greater nourishment.
Behavioural ecologist Reuven Dukas (McMaster U) studied how blue jays
allocate attention to two food types [Dukas and Ellner, 1993, Dukas and
Kamil, 2000, 2001]. The goal in the following problem is to find the optimal
subdivision of time and attention between the food types so as to maximizes
the total energy gain. Mastered Material Check
11. If x = 1 and full attention is devoted
Setting up the model. Suppose that there are two available food types. to finding food type 1, is any
Define notation as follows: attention devoted to finding food
type 2?
x = attention devoted to finding food of type 1
12. If x = 0.5, how much attention is
P(x) = probability of finding the food given attention x being paid to finding food of type 1?
We assume that 0 ≤ x ≤ 1, with x = 0 representing no attention and x = 1 Mastered Material Check
meaning full attention is devoted to finding food type 1. 13. Use Figure 8.6 to estimate the
attention x needed to have a 50%
Moreover, P is a probability which mean 0 ≤ P ≤ 1. We assume that probability of finding food type 1.
P(0) = 0 which means that when no attention is paid (x = 0) the probability of That is, roughly estimate x such
finding food is zero (P = 0). We also assume for simplicity that P(1) = 1, so that P(x) = 0.5 for each of the
probability curves.
when full attention paid x = 1, there is always success (P = 1).
14. If we fully divide attention between
Figure 8.6 displays hypothetical examples of P(x). The horizontal axis food types 1 and 2, and we
is attention 0 ≤ x ≤ 1, and the vertical axis, is the probability of success, spend 0.25 of our attention on
finding food type 1, how much
0 ≤ P ≤ 1. All these curves share the assumed properties of full success with attention is given to food type 2?
full attention, and no success with no attention. However, the curves differ in Converting the probability to a
overall shape. percentage may help with
understanding
Questions.
1. What is the difference between foods of type 1 and 4?

1
Probability, P (x)
2. Which food is easier to find, type 3 or type 4?

4
3. Compare the above to the case that P(x) = x and explain what this new case
implies. 3
2
4. What role is played by the concavity of the curve? 1
Observe that concave down curves such as 3 and 4 rise rapidly at small x, 0 attention, x 1
indicating that the probability of finding food increases a lot just by increas-
ing the attention by a little: these represent foods that are relatively easy to Figure 8.6: The probability, P(x), of finding
a food depends on the level of attention x
find. Other curves (1 and 2) are concave up, indicting that much more atten-
devoted to finding that food. Here 0 ≤ x ≤ 1,
tion is needed to gain appreciable increase in the probability of success: these with x = 1 being “full attention”. We show
represent foods that are harder to find. The concavity of the curves carries possible curves for four types of foods, some
easier to find than others.
this important information about the relative ease or difficulty of finding a

given food type.
Let x and y denote the attention divided between food types 1 and 2, and
suppose that type 2 is N-times more nutritious than type 1. We use P1 (x) and
P2 (y) to denote the probabilities of finding food of type 1 and 2. Then the
total nutritional value gained by splitting the attention is: ` Adjust the slider to see how the
relative nutritional value N of food type
V (x) = P1 (x) + N P2 (y) = P1 (x) + N P2 (1 − x). 2 affects the total value V (x) as a
function of the attention x. What type of
critical point do we find?
The goal is to optimize nutritional value V (x).
Example 8.12 (P1 and P2 as power function with integer powers)

Consider the case that the probability of finding the food types is given
by the simple power functions,
P1 (x) = x2 , P2 (y) = y3 .
i See an explanation of Example 8.12.
(These functions satisfy P(0) = 0, P(1) = 1, in accordance with Figure 8.6.)
Further, suppose that both foods are equally nutritious, so N = 1. Find the
optimal V (x).
Solution. The total nutritional value in this case is
V (x) = P1 (x) + N P2 (1 − x) = x2 + (1 − x)3 .
We look for a maximum value of V . Using the chain rule to differentiate, we

find that
0 2 15. Within Example 8.12, calculate how
V (x ) = 2x + 3(1 − x) (−1),
much nutritional value is gained by
V 00 (x) = 2 − 3(2)(1 − x)(−1) = 2 + 6(1 − x). the animal devoting x = 0.4514 of its
attention to food type 1.
We observe that a negative factor (−1) comes from applying the chain rule to 16. Consult Figure 8.7 to verify your
(1 − x)3 . Setting V 0 (x) = 0 we get result.
1
Nutritional value, V (x)
2x + 3(1 − x)2 (−1) = 0. ⇒ −3x2 + 8x − 3 = 0

√ 0.8
4± 7
⇒ x= ≈ 0.4514, 2.21. 0.6
3
Since attention takes on values in 0 ≤ x ≤ 1, we reject the second root. The 0.4
first root suggests that the animal should spend ≈ 0.45% of its attention on 0.2
food type 1 and the rest on type 2. However, to confirm such speculation, we
must check whether the critical point is a maximum. 0.2 0.4 0.6 0.8 1
attention, x
The second derivative is positive for all values of x in 0 ≤ xle1, signifying a
local minimum! The animal gains least by splitting its attention between two Figure 8.7: Figure for Example 8.12. The
food typess in this case. Indeed, from Figure 8.7, we see that the most gain probabilities of finding foods of types 1
occurs at either x = 0 (only food type 2 sought) or x = 1 (only food type 1 and 2 are concave up power functions. There
is no local maximum.
sought).This example reemphasizes the importance of checking the type of
critical point before drawing hasty conclusions. ♦
Example 8.13 (Fractional-power functions for P1 , P2 ) Now consider the

case that the probability of finding two food types is given by the concave ` Change the powers in the interactive
down power functions, graph to conform to Example 8.13. You
can adjust the relative nutritional value,
N. What type of critical point do we
P1 (x) = x1/2 , P2 (y) = y1/3 find?
and both foods are equally nutritious (N = 1). Find the optimal food value i See a brief recap of Example 8.13,
and why we expect to find a local
V (x ).
maximum in this case.
Solution. These functions also satisfy P(0) = 0, P(1) = 1, in accordance with 1.6

the sketches shown in Figure 8.6. Then
1.4
1/2 (1/3)
V (x) = P1 (x) + P2 (1 − x) = x + (1 − x ) ,
1.2
0 1 1
V (x) = 1/2 − ,
2x 3 (1.0 − x)(2/3)
0.2 0.4 0.6 0.8 1
1 2 attention, x
V 00 (x) = − − .
4 x(3/2) 9 (1.0 − x)(5/3)
We must solve V 0 (x) = 0 to find the critical point. Unfortunately, this problem, Figure 8.8: Figure for Example 8.13. The
probabilities of finding foods of types 1
turns out to be algebraically nasty. However, we can look for an approximate and 2 are both function that are concave
solution to the problem, using Newton’s Method. down. As a result there is a local maximum
for the nutritional value.
A plotting the graph of V (x) in Figure 8.8 demonstrates that there is a
maximum inside the interval 0 ≤ x ≤ 1, i.e., for attention split between finding
both foods. We further see from V 00 (x) that the second derivative is negative
for all values of x in the interval, indicating a local maximum, as expected. ♦
Applying Newton’s method to find the critical point.
Example 8.14 Use Newton’s Method to find the critical point for the func-
tion V (x) in Example 8.13.
Solution. Finding the critical point of V (x) reduces to solving V 0 (x)
= 0. 17. Justify algebraically why V 00 (x) in
0 Example 8.13 is negative on the
Let f (x) = V (x) - we must solve f (x) = 0 using Newton’s Method (Recall
interval 0 ≤ x ≤ 1.
Section 5.4).
18. Using Figure 8.8, what is the largest
Since the interval of interest is 0 ≤ x ≤ 1, we start with an initial “guess” for possible nutritional value?
the critical point at x0 = 0.5, midway along this interval. Then, according to
Newton’s method, the improved guess would be
f ( x0 )
x1 = x0 − .
f 0 (x0 )
and, repeating this, at the k’th stage,
f ( xk )
xk+1 = xk − .
f 0 ( xk )
To use this method, carefully note that
1 1
f (x ) = V 0 (x ) = − ,
2 x1/2 3 (1.0 − x)(2/3)
1 2
f 0 (x) = V 00 (x) = − − .
4 x(3/2) 9 (1.0 − x)(5/3)
Thus, we might use a spreadsheet in which cells A1 stores our initial guess,
whereas B1, C1, and D1 store the values of f (x), f 0 (x) and x0 − f (x0 )/ f 0 (x0 ).
In the typical syntax of spreadsheets, this might read something like the
following: P Link to Google Sheets. This
spreadsheet implements Newton’s
method for Example 8.14. You can view
A1 0.5 the formulae by clicking on a cell in the
B1 =(1/(2*(A1)^(1/2))-1/(3*(1-A1)^(2/3))) sheet but you cannot edit the sheet here.
C1 =(-1/(4*A1^(3/2))-2/(9*(1-A1)^(5/2)))
D1 =A1-B1/C1
This idea is implemented on a spreadsheet resulting in values shown in

Table 8.1 starting from x0 = 0.5. We see that the values converge to the
location of the critical point, x = 0.61977 (and y = 1 − x = 0.38022) within
the interval of interest.
k xk f (x k ) f 0 (x k ) x k +1
Epilogue. While the conclusions drawn above were disappointing in one 0 0.50000 0.17797 -1.96419 0.59061
specific case, it is not always true that concentrating all one’s attention on one 1 0.59061 0.04603 -2.62304 0.60816
type is optimal. We can examine the problem in more generality to find when 2 0.60816 0.01866 -2.83923 0.61473
3 0.61473 0.00816 -2.93065 0.61751
the opposite conclusion might be satisfied. 4 0.61751 0.00367 -2.97130 0.61875
In general, the value gained is 5 0.61875 0.00167 -2.98970 0.61931
6 0.61931 0.00076 -2.99809 0.61956
V (x) = P1 (x) + N P2 (1 − x). Table 8.1: Newton’s method applied to

Example 8.14. We start with x0 = 0.5.
A critical point occurs when
d Mastered Material Check

V 0 (x ) = [P1 (x) + N P2 (1 − x)] = P10 (x) + NP20 (1 − x)(−1) = 0.
dx 19. Can you see where the (−1) comes
from in V 0 (x)?
Suppose we have found a value of x in 0 < x < 1 where this is satisfied. We

then examine the second derivative:
d 0 d
V 00 (x) = [V (x)] = [P10 (x) − NP20 (1 − x)]
dx dx
= P100 (x) − NP200 (1 − x)(−1) = P100 (x) + NP200 (1 − x).
The concavity of the function V is thus related to the concavity of the two
functions P1 (x) and P2 (1 − x). If these are concave down (e.g. as in food
types 3 or 4 in Figure 8.6), then V 00 (x) < 0 and a local maximum occurs at any
critical point found by our differentiation.
Another way of stating this observation is: if both food types are relatively
easy to find, one can gain most benefit by splitting up the attention between
the two. Otherwise, if both are hard to find, then it is best to look for only one
at a time.
8.3 Summary
1. In this chapter we reviewed function composition, in which one func-

tion acts as the input to another function. The order of composition is
important: f (g(x)) 6= g( f (x)).
2. The chain rule can be used to differentiate composite functions. If y = g(u)
and u = f (x) are both differentiable, then
dy dy du
= .
dx du dx
3. Applications seen in this chapter include:
(a) pollution levels in a lake (depending on human population, and pollu-

tion created per person);
(b) populations of carnivores (depending on populations of prey, vegeta-
tion, and rainfall);
(c) budget for coffee (depending on amount consumed, and price per cup);
(d) Earth’s temperature (depending on the albedo, and green house gases);
(e) ants’ path to food (depending on locations of nest and food sources);
and
(f) food choice and attention (depending on probabilities of success,
nutrition gained).

p
1. Write f (x) = x2 (3x2 − 1)3 as the composition of two different functions in two different ways. Differentiate using
the chain rule.
2. Write f (x) as a composition of three different functions. Differentiate using the chain rule. Verify your solution
matches that of 1.
3. If an animal needs to divide its attention x between 4 food types, and P1 (x) = 0.2, P2 (x) = 0.1 and P3 (x) = 0.5, what
is P4 (x)?
4. Consider the following graph depicting the nutritional value gained when x attention is applied to food type 1
1.8
1.6
1.4
1.2
0.2 0.4 0.6 0.8 1

attention, x
(a) What is the maximum nutritional value that can be attained?

(b) How much attention should be paid to food type 1 in order to achieve this?
(c) Assuming there are only two different food types, how much attention should be paid to food type 2?
Exercises
8.1. Practicing the chain rule. Use the chain rule to calculate the follow-
ing derivatives
(a) y = f (x) = (x + 5)5 ,
(b) y = f (x) = 4(x2 + 5x − 1)8 ,
√
(c) y = f (x) = ( x + 2x)3 .
8.2. Growth curve. An example of a growth curve in population biology is
called the Bertalanffy growth curve, after Canadian biologist Ludwig
von Bertalanffy. This curve is defined by the equation credit: De Sapio
N = (a − b2−kt )3 ,
where the constants a, b and k are positive and a > b; N denotes the
size of the population and t denotes elapsed time. Find the growth
rate dN/dt of the population.
Note: if f (x) = 2ax , then f 0 (x) = 0.6931 · a2ax . Derivatives of such
exponential functions are studied in Chapter 10.
8.3. Earth’s temperature. We expand and generalize the results of Ex-
ample 8.10. As before, let G denote the level of greenhouse gases on
Earth, and consider the relationship of temperature of the earth to the
albedo a and the emissivity ε given by Eqn. (8.1).
(a) Suppose that a is constant, but ε depends on G. Assume that dε/dG
is given. Determine the rate of change of temperature with respect
to the level of greenhouse gasses in this case.
(b) Suppose that both a and ε depend on G. Find dT /dG in this more
general case (hint: the quotient rule as well as the chain rule are
needed).
8.4. Shortest path from nest to food sources.
(a) Use the first derivative test to verify that the value x = √d is a local
3
minimum of the function L(x) given by Eqn (8.2)
√
(b) Show that the shortest path is L = D + 3d.
(c) In Section 8.2 we assumed that d << D, so that the food sources
were close together relative to the distance from the nest. Now
suppose that D = d/2. How would this change the solution?
8.5. Geometry of the shortest ants’ path. Use the results of Section 8.2
to show that in the shortest path, the angles between the branches of
the Y-shaped path are all 120◦ . Recall that sin(30) = 1/2, sin(60) =
√
3/2.
8.6. More about the ant trail. Consider the lengths of the V and T-shaped
paths in the ant trail example of Section 8.2. We refer to these as LV
and LT ; each depend on the distances d and D in Figure 8.3.
(a) Write down the expressions for each of these functions.
(b) Suppose the distance D is fixed. How do the two lengths LV , LT
depend on the distance d? Use your sketching skills to draw a
rough sketch of Lv (d ), LT (d ).
(c) Use you sketch to determine whether there is a value of d for which
the lengths LV and LT are the same.
8.7. Divided attention. A bird in its natural habitat feeds on two kinds of
seeds whose nutritional values are
• 5 calories per seed of type 1, and
• 3 calories per seed of type 2.
Both kinds of seeds are hidden among litter on the forest floor and
have to be found. If the bird splits its attention into x1 (a fraction of 1
- its whole attention) searching for seed type 1 and x2 (also a fraction
of 1) searching for seed type 2, then its probability of finding 100
seeds of the given type is
P1 (x1 ) = (x1 )3 , P2 (x2 ) = (x2 )5 .
Assume that the bird pays full attention to searching for seeds so
that x1 + x2 = 1 where 0 ≤ x1 ≤ 1 and 0 ≤ x2 ≤ 1.
(a) Give an expression for the total nutritional value V gained by the
bird when it splits its attention. Use the constraint on x1 , x2 to
eliminate one of these two variables (for example, let x = x1 and
write x2 in terms of x1 .)
(b) Find critical points of V (x) and classify those points.
(c) Find absolute minima and maxima of V (x) and use your results to
explain the bird’s optimal strategy for maximizing the nutritional
value of the seeds it can find.
9
Chain rule applied to related rates and implicit differenti-
ation
9.1 Applications of the chain rule to “related rates”

i See A brief introduction.
1. Given a geometric relationship and a rate of change of one of the variables,

use the chain rule to find the rate of change of a related variable.
Volume of sphere V = 43 πr3
Surface area of sphere S = 4πr2
2. Use descriptive information about rates of change to set up the required Area of circle A = πr2
relationships, and to solve a word problem involving an application of the Perimeter of circle P = 2πr
chain rule (“related rates problem”). Volume of cylinder V = πr2 h
Volume of cone V = 13 πr2 h
In many applications of the chain rule, we are interested in processes that Area of rectangle A = xy
Perimeter of rectangle P = 2x + 2y
take place over time. We ask how the relationships between certain geometric Volume of box V = xyz
(or physical) variables affects the rates at which they change over time. Sides of right triangle c2 = a2 + b2
Many of these examples are given as word problems, and we must assemble Table 9.1: Geometric relationships used in
the required relationships to solve the problem. Some useful geometric various related rates problems.
relationships are presented in Table 9.1.

i Tumor growth example: See the
Example 9.1 (Tumor growth) The radius of a solid tumor expands at a calculation in action.
constant rate, k. Determine the rate of growth of the volume of the tumor
when the radius is r = 1cm. Assume that the tumor is approximately spherical
as depicted in Figure 9.1.
Solution. The volume of a sphere of radius r, is V (r ) = (4/3)πr3 . Here, r

changes with time, so V changes with time. We indicate this chain of depen-
dencies with the notation r (t ) and V (r (t )). Then function composition is
apparent:
r
4
V (r (t )) = π [r (t )]3 .
3
Then, using the chain rule,

d dV dr d 4 3 dr 4 dr dr Figure 9.1: Growth of a spherical tumor.
V (r (t )) = = πr = π · 3r2 = 4πr2 . Since the radius changes with time, the
dt dr dt dr 3 dt 3 dt dt
volume, too, changes with time. We use the
chain rule to link dV /dt to dr/dt.
But we are told that the radius expands at a constant rate, k, so that Mastered Material Check
1. What is dV /dt when the radius is
dr dV r = 2cm?
= k. ⇒ = 4πr2 k.
dt dt 2. What are the units of dV /dt?
Hence, the rate of growth of the volume is proportional to the square of the
radius; in fact, it is proportional to the surface area of the sphere. At the
instant that r = 1 cm,
dV
= 4πk.
dt
Importantly, the numerical value r = 1 cm holds only at one instant and is used
at the end of the calculation, after the differentiation and simplification steps
are completed. ♦
Featured Problem 9.1 (Growth of a cell) The mass of a cell is m = ρV

where V is cell volume and ρ is cell density. (Usually, cell density is constant
and close to that of water, ρ ≈ 1g/cm3 .) Relate the cell rate of change of mass
to rate of change of volume and to rate of change of radius. Assume that the
cell is spherical.
Example 9.2 (Convergent extension) Most animals are longer head to tail
than side to side. To obtain relative elongation along one axis, an embryo
undergoes a process called convergent extension whereby a block of tissue
elongates (extends) along one axis and narrows (converges) along the other
axis as shown in Figure 9.2. Here we consider this process. Mastered Material Check
Suppose that a rectangular block of tissue, with dimensions L = w = 10mm 3. How wide is the tissue
when L = 20mm if thickness τ and
and thickness τ = 1mm, extends at the rate of 1mm per day, while the volume volume V remain fixed?
V and thickness τ remain fixed. At what rate is the width w changing when
the length is L = 20mm?
Original tissue Extended tissue
L
Figure 9.2: Convergent extension of tissue
in embryonic development. Cells elongate
Solution. We are told that the volume V and the thickness τ remain constant. along one axis (which increases L) while
We find, using the initial length, width and thickness, that the volume is contracting along the other axis (decreasing
V = 10 · 10 · 1mm3 . Further, at any given time t, the volume of the rectangular w). Since the volume and thickness remain
fixed, the changes in L can be related to
block is changes in w.
V = L(t ) · w(t ) · τ.
V depends on L and w, both of which depend on time. Hence, there is a chain
of dependencies t → L, w, → V , Differentiating both sides with respect to t
C H A I N RU L E A P P L I E D T O R E L AT E D R AT E S A N D I M P L I C I T D I F F E R E N T I AT I O N 185
leads to
dV d
= (L(t ) · w(t )τ ) ⇒ 0 = (L0 (t ) · w(t ) + L(t ) · w0 (t )) τ.
dt dt
(Here we have used the product rule to differentiate L(t ) · w(t ) with respect
to t. We also used the fact that V is constant so its derivative is zero, and τ is
constant, so it multiplies the derivative of L(t )w(t ) as would any multiplica-
tive constant.) Consequently, canceling the constant factor and solving for
w0 (t ) results in
L0 (t )w(t )
L0 (t )w(t ) + L(t )w0 (t ) = 0 ⇒ w0 (t ) = − .
L(t )
At the instant that L(t ) = 20, w(t ) = V /(L(t )τ ) = 100/20 = 5. Hence we

find that
L0 (t )w(t ) 1mm/day · 5mm
w0 (t ) = − =− = −0.25mm/day.
L(t ) 20mm
The negative sign indicates that w is decreasing while L is increasing. ♦
Example 9.3 (A spider’s thread) A spider moves horizontally across the

ground at a constant rate, k, pulling a thin silk thread with it. One end of the i Spider silk example: See the
calculation in action.
thread is tethered to a vertical wall at height h above ground and does not
move. The other end moves with the spider. Determine the rate of elongation
of the thread.
l
Solution. Figure 9.3 illustrates the geometry, where x is the distance of the h
spider from the wall. We use the Pythagorean Theorem to relate the height of
the tether point h, the spider’s location x, and the length of the thread `: x
`2 = h2 + x2 . Figure 9.3: The length of a spider’s thread.
Here, h is constant, while x, ` change with time, so that
[`(t )]2 = h2 + [x(t )]2 .
Differentiating with respect to t leads to

d d 2 Mastered Material Check
[`(t )]2 = h + [x(t )]2 ,

dt dt 4. Repeat Example 9.3 given that the
thread is tethered 0.5m above the
d` dx d` 2x dx ground and the spider is walking at a
= 0 + 2x
2` ⇒ = . constant rate of 30cm/min.
dt dt dt 2` dt
Simplifying and using the fact that
dx
= k,
dt
leads to
d` x x
= k = k√ .
dt ` h2 + x 2
♦
R R Figure 9.4: The geometry of a conical cup

of (constant) height H and (constant) radius
R. The surface of the water is indicated with
dashed line. The water occupies a conical
r volume of height h(t ) and radius r (t ), which
H both get smaller as waters leaks out of
the cup. The diagram on the left strongly
h suggests that similar triangles would be
helpful in finding a relationship between the
variables.
Example 9.4 (A conical cup) Water is leaking at a constant rate out of a

conical cup of height H and radius R. Find the rate of change of the height of
water in the cup at the instant that the cup is full, if the volume is decreasing
at a constant rate, k.
Solution. Let us define h and r as the height and radius of water inside the
cone. Then we know that the volume of this (conically shaped) water in the i Draining Cone Example: See the
calculation in action.
cone is
1
V = πr2 h,
3
or, in terms of functions of time,
1
V (t ) = π [r (t )]2 h(t ).
3
We are told that
dV
= −k,
dt
where the negative sign indicates that volume is decreasing. By similar
triangles, we note that Mastered Material Check
5. What volume of water can be
r R R
= ⇒ r= h, contained in a cone of height 5cm
h H H and radius 3cm?
so we use this substitution to write the volume in terms of the height alone: 6. What are the units of the constant k?
2 7. What is meant by similar triangles?
1 R
V (t ) = π [h(t )]3 .
3 H
Then the chain rule leads to
2
dV 1 R dh
= π · 3[h(t )]2 .
dt 3 H dt
Now using the fact that volume decreases at a constant rate, we get
2
R dh dh −kH 2
−k = π [h(t )]2 ⇒ = .
H dt dt πR2 h2
The rate computed above holds at any time as the water leaks out of the
container. At the instant that the cup is full, h(t ) = H and r (t ) = R, so that
dh −kH 2 −k
= = .
dt πR2 H 2 πR2
For example, for a cone of height H = 4 and radius R = 3,
dh −k
= .
dt 9π
It is important to use information about a specific instant only after deriva-
tives are computed. ♦
Featured Problem 9.2 (Growth of a Tree Trunk) Consider a cylindrical

tree trunk of radius R. Living cells occupy a thin shell (thickness d) just
inside the tree bark. The interior of the trunk consists of dead cells that have
turned into wood.
1. What fraction F of the trunk volume is living tissue?

Figure 9.5: A thin shell of living tissue
2. How does the fraction F change with time as the tree grows? Assume that
(phloem) surrounds the dead wooden part
the radius of the trunk grows at a constant rate, and that the thickness d of a tree trunk (xylem). The fraction F of
does not change. Compute the rate of change of F at the instant that the living tissue changes as the tree grows.
radius is 5 times the thickness d.
Hints and setting up the problem

i An explanation of how to set up the
• Assume that the wooden interior is cylindrical, as is the trunk. Find the tree trunk problem.
volume of the shell by subtracting the volumes of these two cylinders, and
now write down the fraction F.
You should get
πh R2 − (R − d )2

F= .
πhR2
Simplify this expression.
• Compute the derivative dF/dt, remembering to use the Chain Rule. You Figure 9.6: The radius of the trunk is R and
may want to use the quotient rule for practice, or to first simplify the the thickness of the phloem (living part of
the trunk) is d. We assume a cylindrical
expression as much as possible and then compute a derivative. geometry. F = fraction of volume in the thin
(blue) cylindrical shell.
9.2 Implicit differentiation i Some further steps, and an outline of

what we are asked to do.
1. Identify the distinction between a function that is defined explicitly and

one that is defined implicitly.
2. Describe implicit differentiation geometrically.
3. Compute the slope of a curve at a given point using implicit differentiation,

find tangent line equations, and solve problems based on such ideas.
Implicit and explicit definition of a function

A review of the definition of a function (e.g. Appendix C) reminds us that for
a given x value, only one y value is permitted. For example, for
y = x2
any value of x leads to a single y value (Figure 9.7a). Geometrically, this

means that the graph of this function satisfies the vertical line property: a
given x value can have at most one corresponding y value. Not all curves
satisfy this property. The elliptical curve in Figure 9.7b clearly fails this,
intersecting some vertical lines twice. This simply means that, while we can
write down an equation for such a curve, e.g.
(x − 1)2
+ (y − 1)2 = 1,
4
we cannot solve for a simple function that describes the entire curve. Never-
theless, the idea of a tangent line to such a curve - and consequently the slope
of such a tangent line - is perfectly reasonable.
In order to make sense of this idea, we restrict attention to a local part of
the curve, close to some point of interest (Figure 9.7c). Then near this point,
the equation of the curve defines an implicit function, that is, close enough
to the point of interest, a value of x leads to a unique value of y. We refer to
this value as y(x) to remind us of the relationship between the two variables.
y y Figure 9.7: (a) A function has to satisfy

the vertical line property. Hence, the curve
shown in (b) cannot be a function. We can
write down an equation for the curve, but we
cannot solve for y explicitly. (c) However,
close to a given point on the curve (dark
point), we can think of how changing the x
coordinate of the point (shaded interval on x
x x axis) leads to a change in the corresponding
y coordinate on the same curve (shaded
(a) (b)
interval on y axis). (d) We can also ask
y y what is the slope of the curve at the given
point. This corresponds to lim∆x→0 ∆y/∆x.
Implicit differentiation can be used to
∆y
compute that derivative.
∆x
x x
(c) (d)
How can we generalize the notion of a derivative to implicit functions?

We observe from (Figure 9.7d) that a small change in x leads to a small
change in y. Without writing down an explicit expression for y versus x, we
can still determine these small changes, and form a ratio ∆y/∆x which is a
(secant line) slope. Now let ∆x → 0 to arrive at the slope of a tangent line
as before, dy/dx. In the next section we show how to do this using implicit
differentiation, an application of the chain rule.
Slope of a tangent line at the point on a curve

i A brief introduction to implicit
We now compute the tangent line at a point in several examples where it is differentiation and slope of a tangent
inconvenient or impossible, to isolate y as a function of x (see Figure 9.8). line to a circle.
First, consider the simple example of a circle. We aim to find the slope of
the tangent line at some point. The equation of a circle of radius 1 and centre
at the origin (0, 0) is
x2 + y2 = 1.
y y Figure 9.8: The curve in (a) is not a function

(x, y)
and hence it can only be described implicitly.
tangent line However, if we zoom in to a point in (b),
we can define the derivative as the slope of
(x, y) the tangent line to the curve at the point of
x interest.
ZOOM
x
(a) (b)
Here, the two variables are linked in a symmetric relationship. We can

solve for y, obtaining not one but two functions.
√
top of circle: y = f 1 (x ) = 1 − x2 , Mastered Material Check
√
and bottom of circle: y = f 2 ( x ) = − 1 − x2 . 8. In Example 9.5, why do we need to
specify “in the first quadrant"? What
are other possible point with
x = 1/2 on the circle?
However, this makes the work of differentiation more complicated than
necessary. Instead, we use implicit differentiation.
y
Example 9.5 (Tangent to a circle) a) Use implicit differentiation to find the
slope of the tangent line to the point x = 1/2 in the first quadrant on a
circle of radius 1 and centre at (0, 0).
b) Find the second derivative d 2 y/dx2 at the same point. x

Solution.
p √
a) When x = 1/2 then y = ± 1 − (1/2)2 = ± 3/2. The point in the first
√
quadrant has y coordinate y = + 3/2.
Figure 9.9: Tangent line to a circle by
implicit differentiation.
Viewing x as the independent variable locally, (and y depending on x on

the curve) we write
x2 + [y(x)]2 = 1.
Differentiating each side with respect to x:
dx2

d 2 d d
x + [y(x)]2 = 1 = 0 + [y(x)]2

⇒ = 0.
dx dx dx dx
Notice that in the second term, the value of x determines y which in turn
determines y2 . Applying the chain rule, we obtain
2
dy2 dy

dx dy
+ = 0. ⇒ 2x + 2y = 0.
dx dy dx dx
Thus
dy dy 2x x
2y = −2x ⇒ =− =− .
dx dx 2y y
√
At the point of interest, x = 1/2, y = 3/2. Thus the slope of the tangent
line is √
0 dy x 1/2 −1 − 3
y = = − = −√ =√ = .
dx y 3/2 3 3
b) The second derivative can be computed using the quotient rule
d2y

0 dy x d x
y = =− ⇒ 2
= −
dx y dx dx y
d2y 1 · y − x · y0 y − x −x
y y2 + x2 1
2
= − 2
= − 2
= − 3 = − 3.
dx y y y y
√
Substituting y = 3/2 from part (a) yields
d2y 1 8 9. Verify the result of Example 9.5(a)
2
=− √ =− ( .
dx ( 3/2) 3 3 3/2) by differentiating the explicit
function for the top half of a circle
We used the equation of the circle, and our result for the first derivative to of radius√1, centered at the origin:
f ( x ) = 1 − x2 .
simplify the above. ♦
10. Similarly, verify the result of
Example 9.5(b).
Note: we can see from the last expression that the second derivative is
11. The second derivative is positive
negative for y > 0, i.e. for the top semi-circle, indicating that this part of for y < 0. What does this say about
the curve is concave down (as expected). Indeed, as in the case of simple the bottom part of the circle?
functions, the second derivative can help identify concavity of curves.
Example 9.6 (Energy loss and Earth’s temperature) Redo Example 4.9
using implicit differentiation, that is: find the rate of change of Earth’s
temperature per unit energy loss based on Eqn. (1.5): Eout = 4πr2 εσ T 4 .
Solution. We rewrite the equation in the form
Eout (T ) = (4πr2 εσ )T 4
and observe that the term in braces is constant. We then differentiate both
sides with respect to Eout . We find,
dEout dT 4 dT dT
= (4πr2 εσ ) ⇒ 1 = (4πr2 εσ ) · 4T 3 .
dEout dT dEout dEout Mastered Material Check
The calculation is completed by rearranging this result. Thus 12. Does this result agree with that of
Example 4.9,
dT 1 1 dT

1
1/4
= = −3/4
Eout ?
dEout 16πr εσ T 3
2 dEout 16πr2 εσ
Justify algebraically.
is the rate of change of the Earth’s temperature per unit energy loss. ♦
13. Can you define an inverse function?
9.3 The power rule for fractional powers
Implicit differentiation is a useful technique for finding derivatives of inverse

functions. Here we use the known power rule for y = x2 to find the derivative
√
of its inverse function, y = x = x1/2 . This general idea recurs in later
chapters when we introduce new functions and their inverses.
i Using implicit differentiation to
√ √ √
Example 9.7 (Derivative of x) Consider the function y = x = x1/2 . Use compute the derivative of y = x.
implicit differentiation to compute the derivative of this function.

√
Solution. Let us rewrite the relationship y = x in the form y2 = x, but
consider y as the dependent variable, i.e. when we differentiate, we remember
that y depends on x:
[y(x)]2 = x.
Taking derivatives of both sides leads to
d d dy dy 1
[y(x)]2 = (x)

⇒ 2[y(x)] =1 ⇒ = .
dx dx dx dx 2y
√
We eliminate y by substituting y = x. Then
dy 1 1
= √ = x−1/2 .
dx 2 x 2
This verifies the power law for the above example. ♦

A similar procedure can be applied to any power function with fractional
power. When we apply similar steps, we get the following rule: Mastered Material Check
14. Justify the derivative of the
Derivative of fractional-power function: The derivative of fractional power rule by actually
carrying out an implicit
y = f (x) = xm/n differentiation calculation.
is
dy m m
= x( n −1) .
dx n
i See Demo created by David Austin

Example 9.8 (The astroid) The curve of the astroid.
x2/3 + y2/3 = 22/3
has the shape of an astroid. It describes the shape (Figure 9.10) generated by
a the path of a point on the perimeter of a disk of radius 12 rolling inside the
perimeter of a circle of radius 2.
Find the slope of the tangent line to a point on the astroid.
Solution. Considering y as the dependent variable, we use implicit differenti-

ation as follows:
Figure 9.10: The astroid is generated by a
d 2/3 d 2 −1/3 d dy
x + [y(x)]2/3 = 22/3 ⇒ x + (y2/3 ) = 0 disk or radius 1/2 rolling inside a circle of
dx dx 3 dy dx radius 2.
2 −1/3 2 −1/3 dy
⇒ x + y =0
3 3 dx
dy
⇒ x−1/3 + y−1/3 =0
dx
dy x−1/3
⇒ = − −1/3 .
dx y
15. Point out places on Fig. 9.10 at
The derivative fails to exist at x = 0 (where x−1/3 is undefined) and at y = 0 which the derivative fails to exist,
(where y−1/3 is undefined). ♦ and explain the properties of those
specific points on the astroid.
Example 9.9 (Horizontal tangent and concavity on a rotated ellipse) Find
the highest point on the (rotated) ellipse x2 + 3y2 − xy = 1.
i Using implicit differentiation to find
Solution. The highest point on the ellipse has a horizontal tangent line, so we the points on the top and bottom of the
ellipse in Example 9.9.
should look for the points on this curve at which dy/dx = 0.
1. Finding the slope of the tangent line: By implicit differentiation,
d 2 d d (x2 ) d (3y2 ) d (xy)

[x + 3y2 − xy] = 1 ⇒ + − = 0.
dx dx dx dx dx
We must use the product rule to compute the derivative of the last term xy:

dy dy dx dy dy
2x + 6y − x + y = 0 ⇒ 2x + 6y − x − 1y = 0.
dx dx dx dx dx
Grouping terms, we have
dy dy (y − 2x)
(6y − x) + (2x − y) = 0 ⇒ = .
dx dx (6y − x)
Setting dy/dx = 0, we obtain y − 2x = 0 so that y = 2x at the point of
interest. Next, we find the coordinates of the point.
2. Determining the coordinates of the point we want: We look for a point

that satisfies the equation of the curve as well as the condition y = 2x. There
are two equations and two unknowns. Plugging y = 2x into the original
equation of the ellipse, we get:
x2 + 3y2 − xy = 1 ⇒ x2 + 3(2x)2 − x(2x) = 1.
After simplifying, this equation becomes 11x2 = 1, leading to the two rotated ellipse
possibilities 0.6
1 2 0.4
x = ±√ , y = ±√ .
11 11
0.2
Which of these two points is at the top? The rotated ellipse is depicted in
Figure 9.11 which gives strong indication it is the positive solution - but −1 −0.5 0.5 1
we can confirm this analytically. −0.2
−0.4
3. Identifying the point at the top: The top point on the ellipse is located
at a point where the curve is concave down. Concavity can be determined −0.6
using the second derivative, computed (from the first derivative) using the
Figure 9.11: A rotated ellipse. In Exam-
quotient rule: ple 9.9, we find the point at the top of the
ellipse using implicit differentiation.
[y − 2x]0 (6y − x) − [6y − x]0 (y − 2x)

y00 =
(6y − x)2
[y − 2](6y − x) − [6y0 − 1](y − 2x)
0
= .
(6y − x)2
4. Plugging in information about the point: Now that we have set down
the form of this derivative, we make some important observations about
the specific point of interest. (This is done as a final step, only after all
derivatives have been calculated)
• We are only concerned with the sign of the second derivative. The denom-
inator is always positive (since it is squared) and so does not affect the
sign.
• At the top of ellipse, y0 = 0, simplifying some of the terms above.
• At the top of ellipse, y = 2x so the term (y − 2x) = 0.

We can thus simplify the expression for the y00 to obtain 16. Verify by hand that x = ± √111 are
two possible solutions to
[−2](6y − x) − [−x](0) [−2](6y − x) −2
y00 (x) = = = . Example 9.9.
(6y − x)2 (6y − x)2 (6y − x) 17. Repeat Example 9.9 looking for the
Using the fact that y = 2x, we get the final form lowest point on the rotated ellipse.
−2 −2
y00 (x) = = .
(6(2x) − x) 11x
Consequently, the second derivative is negative (implying concave down
curve) whenever x is positive. This tells us that at the point with positive x
√
value (x = 1/ 11), we are at the top of the ellipse. A graph of this curve is
shown in Figure 9.11. ♦
Featured Problem 9.3 (Tangent to an ellipse) Find the equation of a line

through the origin that is tangent to the ellipse
y2
(x − a)2 + = 1.
s2
This ellipse has its center at (a, 0) and has axes 1 and s.
9.4 Summary
1. The chain rule can be used to relate the changes in variables that depend
on one other in some “chain of relationships”. We use the term “related
rates” to describe such problems.
2. Curves that fail the vertical line property cannot be describe by a single
function y = f (x), even if we can represent some of those curves by
equation(s).
3. Zooming in on such a curve, we can define an implicit function that
describes some local piece of the curve.
4. When we use implicit differentiation in two variables, we treat one vari-
able as independent and the other as dependent. This allows us to differen-
tiate the equation with respect to the independent variable using the chain
rule.
5. Through implicit differentiation we showed that the derivative of y = xm/n
dy m
is dx = mn x( n −1) .
6. Applications addressed in this chapter included:
(a) tumor growth (volume depends on radius which depends on a growth

rate);
(b) convergent extension in tissue of an embryo (relationship between the
length and width of the growing tissue);
(c) growth of a cell (the relationship between volume and radius).
(d) spider’s thread (length of thread depends on the spider’s position,
which depends on time);
(e) growth of a tree trunk (determining the fraction of the trunk that is
living tissue as the tree grows)
(f) conical cup leaking water (height of water depends on volume, which
depends on time); and
(g) rate of Earth’s temperature change per unit energy loss.

da
1. Let a = b2/9 . Determine db .
2. In Example 9.4, suppose that the cone does not leak, but that it is being filled with water at a constant rates. How
would your work change?
3. Use implicit differentiation to find the slope of the tangent line to the circle x2 + y2 = 1 at the point x = −1? How does
your result relate to the orientation of the tangent line to the circle at that point?
4. Consider the following curve.

1
y
0.8
0.6
0.4
0.2
x
0.2 0.4 0.6 0.8 1
−0.2
Draw both tangent lines to this curve at x = 0.5.

Exercises
9.1. Growth of a cell. Consider the growth of a cell, assumed spherical

in shape. Suppose that the radius of the cell increases at a constant
rate k > 0 per unit time.
(a) At what rate would the volume, V , increase ?
(b) At what rate would the surface area, S, increase ?
(c) At what rate would the ratio of surface area to volume S/V change?
Would this ratio increase or decrease as the cell grows?
Note: note that answers are expressed in terms of the radius of the
cell.
9.2. Growth of a circular fungal colony. A fungal colony grows on a
flat surface starting with a single spore. The shape of the colony edge
is circular with the initial site of the spore at the centre of the circle.
Suppose the radius of the colony increases at a constant rate C per unit
time.
(a) At what rate does the area covered by the colony change ?
(b) The biomass of the colony is proportional to the area it occu-
pies. Let α be the factor of proportionality. At what rate does the
biomass increase?
9.3. Limb development. During early development, the limb of a fetus
increases in size, but has constant proportions. Suppose that the limb
is roughly a circular cylinder with radius r and length l in proportion
l/r = C
where C is a positive constant. It is noted that during the initial phase

of growth, the radius increases at an approximately constant rate, i.e.
that
dr/dt = a.
At what rate does the mass of the limb change during this time?
Note: assume that the density of the limb is 1 gm/cm3 and recall that
the volume of a cylinder is
V = Al
where A is the base area (in this case of a circle) and l is length.
9.4. Pouring water in a trough. A rectangular trough is 2 meter long, 0.5
meter across the top and 1 meter deep. At what rate must water be
poured into the trough such that the depth of the water is increasing
at 1 m/min when the depth of the water is 0.7 m?
9.5. Spherical balloon. Gas is being pumped into a spherical balloon at
the rate of 3 cm3 /s.
(a) How fast is the radius increasing when the radius is 15 cm?
(b) Without using the result from (a), find the rate at which the surface
area of the balloon is increasing when the radius is 15 cm.
9.6. Ice melting. A spherical piece of ice melts so that its surface area
decreases at a rate of 1 cm2 /min. Find the rate that the diameter
decreases when the diameter is 5 cm.
9.7. Point moving on a parabola. A point moves along the parabola y =
1 2
x in such a way that at x = 2 the x-coordinate is increasing at the rate
4
of 5 cm/s. Find the rate of change of y at this instant.
9.8. Boyle’s Law. In chemistry, Boyle’s Law describes the behaviour of
an ideal gas: this law relates the volume V occupied by the gas to the
temperature T and the pressure P as follows:
PV = nRT
where n, R are positive constants.

(a) Suppose that pressure is kept fixed by allowing the gas to expand as
the temperature is increased. Relate the rate of change of volume to
the rate of change of temperature.
(b) Suppose that the temperature is held fixed and the pressure is
decreased gradually. Relate the rate of change of the volume to the
rate of change of pressure.
9.9. Spread of a population. In 1905 a Bohemian farmer accidentally
allowed several muskrats to escape an enclosure. Their population Cite
grew and spread, occupying increasingly larger areas throughout
Europe. In a classical paper in ecology, it was shown by the scientist
Skellam (1951) that the square root of the occupied area increased at a
constant rate, k.
Determine the rate of change of the distance (from the site of release)
that the muskrats had spread. Assume that the expanding area of
occupation is circular.
9.10. A convex lens. A particular convex lens has a focal length of f =
10 cm. Let p be the distance between an object and the lens, and q the
distance between its image and the lens. These distances are related to
the focal length f by the equation:
1 1 1
= + .
f p q
Consider an object which is 30 cm away from the lens and moving

away at 4 cm/sec.
How fast is its image moving and in which direction?
9.11. A conical cup. Water is leaking out of a small hole at the tip of a Formula.
Note that the volume of a cone
is V = (π/3)r2 h.
conical paper cup at the rate of 1 cm3 /min. The cup has height 8 cm
and radius 6 cm, and is initially full up to the top.
Find the rate of change of the height of water in the cup when the cup
just begins to leak.
9.12. Conical tank. Water is leaking out of the bottom of an inverted
1
conical tank at the rate of 10 m3 /min, and at the same time is be-
ing pumped in the top at a constant rate of k m3 /min. The tank has
height 6 m and the radius at the top is 2 m.
Determine the constant k if the water level is rising at the rate
of 15 m/min when the height of the water is 2 m.
9.13. The gravel pile. Gravel is being dumped from a conveyor belt at the
rate of 30 ft3 /min in such a way that the gravel forms a conical pile
whose base diameter and height are always equal.
How fast is the height of the pile increasing when the height is 10 ft?
9.14. The sand pile. Sand is piled onto a conical pile at the rate of 10 m3 /min.
The sand keeps spilling to the base of the cone so that the shape al-
ways has the same proportions: that is, the height of the cone is equal
to the radius of the base.
Find the rate at which the height of the sandpile increases when the
height is 5 m.
9.15. Conical water reservoir. Water is flowing into a conical reservoir at a
rate of 4 m3 /min. The reservoir is 3 m in radius and 12 m deep.
(a) How fast is the radius of the water surface increasing when the
depth of the water is 8 m?
(b) In (a), how fast is the surface rising?
9.16. Sliding ladder. A ladder 10 meters long leans against a vertical wall.
The foot of the ladder starts to slide away from the wall at a rate of 3
m/s.
(a) Find the rate at which the top of the ladder is moving downward
when its foot is 8 meters away from the wall.
(b) In (a), find the rate of change of the slope of the ladder.
9.17. Sliding ladder. A ladder 5 m long rests against a vertical wall. If the
bottom of the ladder slides away from the wall at the rate of 0.5 m/min
how fast is the top of the ladder sliding down the wall when the base
of the ladder is 1 m away from the wall?
9.18. Species diversity in an area. Ecologists are often interested in the
relationship between the area of a region (A) and the number of differ-
ent species S that can inhabit that region. Hopkins (1955)suggested a
relationship of the form [Hopkins, 1955]
S = a ln(1 + bA)
where a and b are positive constants.

Find the rate of change of the number of species with respect to the
area. Does this function have a maximum?
9.19. The burning candle. A candle is placed a distance l1 from a thin
block of wood of height H. The block is a distance l2 from a wall as
shown in Figure 9.12. The candle burns down so that the height of the
flame, h1 decreases at the rate of 3 cm/hr. Find the rate at which the
length of the shadow y cast by the block on the wall increases.
Note: your answer should be in terms of the constants l1 and l2 . This is
a challenging problem.
Figure 9.12: Figure for Exercise 19; shadow

cast by burning candle.
H h1
h1
l1 l2
9.20. Implicit differentiation. Use implicit differentiation to show that the

derivative of the function
y = x1/3
is
y0 = (1/3)x−2/3 .
First write the relationship in the form y3 = x, and then find dy/dx.
9.21. Generalizing the Power Law.
(a) Use implicit differentiation to calculate the derivative of the func-
tion
y = f (x) = xn/m
where m and n are integers (hint: rewrite the equation in the

form ym = xn first).
(b) Use your result to derive the formulas for the derivatives of the
√
functions y = x and y = x−1/3 .
9.22. Tangent lines to a circle. The equation of a circle with radius r and
centre at the origin is
x 2 + y2 = r 2
(a) Use implicit differentiation to find the slope of a tangent line to the
circle at some point (x, y).
(b) Use this result to find the equations of the tangent lines of the circle
√
at the points whose x coordinate is x = r/ 3.
(c) Use the same result to show that the tangent line at any point on the
circle is perpendicular to the radial line drawn from that point to the
centre of the circle
Note: Two lines are perpendicular if their slopes are negative
reciprocals.
9.23. Implicit differentiation. For each of the following, find the derivative
of y with respect to x.
(a) y6 + 3y − 2x − 7x3 = 0
√
(b) ey + 2xy = 3
9.24. Tangent line to a circle. The equation of a circle with radius 5 and
centre at (1, 1) is
(x − 1)2 + (y − 1)2 = 25
(a) Find the slope of the tangent line to this curve at the point (4, 5).
(b) Find the equation of the tangent line.
9.25. Tangent to a hyperbola. The curve
x2 − y2 = 1
is a hyperbola. Use implicit differentiation to show that for large x

and y values, the slope dy/dx of the curve is approximately 1.
9.26. An ellipse. Use implicit differentiation to find the points on the ellipse
x 2 y2
+ =1
4 9
(xp,yp)
at which the slope is −1/2.
9.27. Motion of a cell.
In the study of cell motility, biologists often investigate a type of cell
called a keratocyte, an epidermal cell that is found in the scales of
fish. This flat, elliptical cell crawls on a flat surface, and is known to
be important in healing wounds. The 2D outline of the cell can be
approximated by the ellipse Figure 9.13: An elliptical cell
x2 /100 + y2 /25 = 1
where x and y are distances in µm. When the motion of the cell is Units.
Note that 1µm, often called “1 micron”,
filmed, points on the “leading edge” (top arc of the ellipse) move in a
is 10−6 meters.
direction perpendicular to the edge.
Determine the direction of motion of the point (x p , y p ) on the leading
edge, i.e. find the slope of the dark arrow in Fig. 9.13.
(1.5a, 1.5a)
9.28. The Folium of Descartes. A famous curve (see Figure 9.14) that was
studied historically by many mathematicians (including Descartes) is
x3 + y3 = 3axy a
Assume that a is a positive constant.

−a
(a) Explain why this curve cannot be described by a function such
as y = f (x) over the domain −∞ < x < ∞.
(b) Use implicit differentiation to find the slope of this curve at a Figure 9.14: The Folium of Descartes in
Exercise 28
point (x, y).
(c) Determine whether the curve has a horizontal tangent line any-
where, and if so, find the x coordinate of the points at which this
occurs.
(d) Does implicit differentiation allow you to find the slope of this
curve at the point (0, 0) ?
9.29. Isotherms in the Van-der Waal’s equation. In thermodynamics, the
Van der Waal’s equation relates the mean pressure, p of a substance to
its molar volume v at some temperature T as follows:
a
(p + )(v − b) = RT ,
v2
where a, b, R are constants. Chemists are interested in the curves
described by this equation when the temperature is held fixed.
Note: these curves are called isotherms.
(a) Find the slope, d p/dv, of the isotherms at a given point (v, p).
(b) Determine where points occur on the isotherms at which the slope
is horizontal.
y
9.30. The circle and parabola: A circle of radius 1 is made to fit inside
the parabola y = x2 as shown in figure 9.15. Find the coordinates of
the centre of this circle, i.e. find the value of the unknown constant c
(hint: set up conditions on the points of intersection of the circle and
(0, c)
the parabola which are labeled (a, b) in the figure. What must be true
about the tangent lines at these points?). (a,b)
9.31. Equation of a tangent line. Consider the curve whose equation is x
x3 + y3 + 2xy = 4, y = 1 when x = 1.
(a) Find the equation of the tangent line to the curve when x = 1. Figure 9.15: Figure for Exercise 30; circle
inside a parabola.
(b) Find y00 at x = 1.
(c) Is the graph of y = f (x) concave up or concave down near x = 1?
(hint: differentiate the equation x3 + y3 + 2xy = 4 twice with respect
to x).
10
Exponential functions
“The mathematics of uncontrolled growth are frightening. A single cell of

the bacterium E. coli would, under ideal circumstances, divide every twenty
minutes. That is not particularly disturbing until you think about it, but the fact
is that bacteria multiply geometrically: one becomes two, two become four,
four become eight, and so on. In this way it can be shown that in a single day,
one cell of E. coli could produce a super-colony equal in size and weight to the
entire planet Earth."
Michael Crichton, The Andromeda Strain, p. 247 [Crichton, 1969]
In this chapter, we introduce the exponential functions. We first describe

the discrete process of population doubling, represented by powers of 2,
namely, 2n , where n is some integer. We generalize to a continuous function Mastered Material Check
2x where x is any real number. We can then attach meaning to the notion 1. If a population has size P, what do
we mean by a doubled poulation
of the derivative of an exponential function. In doing so, we encounter a size??
specially convenient base denote e, leading to the most useful member of this 2. How large would the population be
class of functions, y = ex . We discuss applications to unlimited growth in a it it doubled twice?
population.
10.1 Unlimited growth and doubling
1. Explain the link between population doubling and integer powers of the
base 2.
2. Given information about the doubling time of a population and its initial
size, determine the size of that population at some later time.
3. Appreciate the connection between 2n for integer values of n and 2x for a

real number x.
The Andromeda Strain
The Andromeda Strain scenario (described by Crichton in the opening quo-

i A screencast summary of population
tation) motivates our investigation of population doubling and uncontrolled doubling and the Andromeda Strain.
growth. Consider 2n where n = 1, 2 . . . is an integer. We will study values of Edu.Cr.
this discrete function as the “variable”, n in the exponent changes. We list

some values and display a graph of 2n versus n in Figure 10.1. Notice that
an initially “gentle” growth becomes extremely steep in just a few steps, as
shown in the accompanying graph.
3. Compare the function f (n) = 2n
Note: properties of 2n (and related expressions) are reviewed in Appendix B.1 and g(n) = n2 for n = 1, 2, . . . , 5.
where common manipulations are illustrated. We assume the reader is How do these differ?
familiar with this material.
y Figure 10.1: Powers of 2 including both

negative and positive integers: here we show
1,000 2n for −4 < n < 10.
n 2n
0 1 y = 2n
1 2 800
2 4
3 8 600
4 16
5 32
6 64 400
7 128
8 256
9 512 200
10 1024
n
−4 −2 2 4 6 8 10
The function 2n first grows slowly, but then grows faster and faster as n
increases. As a side remark, the fact that 210 ≈ 1000 = 103 , will prove useful
4. Why would the approximation
for simple approximations. With this preparation, we can now check the 210 ≈ 103 be helpful?
accuracy of Crichton’s statement about bacterial growth.
Example 10.1 (Growth of E. coli) Use the following facts to check the
assertion made by Crichton’s statement at the beginning of this chapter.
• Mass of 1 E. coli cell : 1 picogram = 10−12 gm = 10−15 kg.
• Mass of Planet Earth : 6 · 1024 kg.

Solution. Based on the above two facts, we surmise that the size of an E. coli 5. How many cells of E. coli would
colony (number of cells, m) that together form a mass equal to Planet Earth there be after 20 minutes? 1 hour? 2
hours?
would be
6 · 1024 kg
m= = 6 · 1039 .
10−15 kg
EXPONENTIAL FUNCTIONS 205
Each hour corresponds to 3 twenty-minute generations. In a period of 24

hours, there are 24 × 3 = 72 generations, each doubling the colony size.
After 1 day of uncontrolled growth, the number of cells would be 272 . We can
find a decimal approximation using the observation that 210 ≈ 103 :
272 = 22 · 270 = 4 · (210 )7 ≈ 4 · (103 )7 = 4 · 1021 .
Using a scientific calculator, the value is found to be 4.7 · 1021 , so the approxi-
mation is relatively good. ♦
Apparently, the estimate made by Crichton is not quite accurate. However
it can be shown that it takes less than 2 days to produce a number far in
6. Verify that it takes less than 2 days
excess of the “size of Planet Earth”. The exact number of generations is left to produce a number far in excess of
as an exercise for the reader and is discussed in Example 10.12. the size of Planet Earth.
The function 2x and its “relatives”

rigour?
We would like to generalize the function 2n to a continuous function, so that
the tools of calculus - such as derivatives - can be used. To this end, we start
with values that can be calculated based on previous mathematical experience,
and then “fill in gaps”.
y y Figure 10.2: (a) Values of the function 2x

1,000 1,000 for discrete value of x. We can compute
y = 2x y = 2x many values (e.g. for x = 0, ±1, ±2, by
800 (discrete) 800 (smooth) simple arithmetical operations, and for
x = ±1/2, ±3/2 by evaluating square roots).
600 600 (b) The function 2x is connected smoothly to
form a continuous curve.
400 400
200 200
x x
−4 −2 2 4 6 8 10 −4 −2 2 4 6 8 10
(a) (b)
From previous familiarity with power functions such as y = x2 (not to be

confused with 2x ), we know the value of
√
21/2 = 2 ≈ 1.41421 . . .
We can use this value to compute

√ √
23/2 = ( 2)3 , 25/2 = ( 2)5 ,
and all other fractional exponents that are multiples of 1/2. We can add these
to the graph of our previous powers of 2 to fill in additional points. This is
shown on Figure 10.2(a).
Similarly, we could also calculate exponents that are multiples of 1/4

since
√
q
1/4
2 = 2
is a value that we can obtain. Adding these values leads to an even finer Mastered Material Check
set of points. By continuing in the same way, we “fill in” the graph of the 7. Given 21/2 ≈ 1.41421, find 23/2
emerging function. Connecting the dots smoothly allows us to define a value and 25/2 without using fractional
powers.
for any real x, of a new continuous function,
8. What method might you use to
determine a decimal approximation
y = f ( x ) = 2x . of 21/4 without computing
fractional powers?
Here x is no longer restricted to an integer, as shown by the smooth curve in 9. Why do we need to assume that
Figure 10.2(b). a > 0 for the exponential function
y = ax ?
Example 10.2 (Generalization to other bases) Plot “relatives” of 2x that

have other bases, such as y = 3x , y = 4x and y = 10x and comment about the
y
function y = ax where a > 0 is a constant (called the base). 1,000
10x 4x 3x 2x
Solution. We first form the discrete function an for integer values of n, 800
simply by multiplying a by itself n times. This is analogous to Figure 10.1. 600

So long as a is positive, we can “fill in” values of ax when x is rational (in the
400
same way as we did for 2x ), and we can smoothly connect the points to lead
to the continuous function ax for any real x. Given some positive constant a, 200
we define the new function f (x) = ax as the exponential function with base a.
x
Shown in Figure 10.3 are the functions y = 2x , y = 3x , y = 4x and y = 10x . ♦ −4 −2 2 4 6 8 10
Figure 10.3: The function y = f (x) = ax is

shown here for a variety of bases, a = 2, 3, 4,
10.2 Derivatives of exponential functions and the function ex and 10.
1. Using the definition of the derivative, calculate the derivative of the

function y = ax for an arbitrary base a > 0.
2. Describe the significance of the special base e.
3. Summarize the properties of the function ex , its derivatives, and how to

manipulate it algebraically.
4. Recall the fact that the function y = ekx has a derivative that is proportional
to the same function (y = ekx ).
Calculating the derivative of ax

i A screencast with the calculations for
In this section we show how to compute the derivative of the exponential this section and motivation for the
function. Rather then restricting attention to the special case y = 2x , we natural base e. Edu.Cr.
consider an arbitrary positive constant a as the base. Note that the base has to
be positive to ensure that the function is defined for all real x. For a > 0 let
y = f ( x ) = ax .
Then, using the definition of the derivative, Mastered Material Check

10. Describe geometrically the
ax+h − ax

dax derivative of ax .
= lim
dx h→0 h
ax ah − ax

= lim
h→0 h
( ah − 1)
= lim ax
h→0 h
h
a −1

x
= a lim .
h→0 h
The variable x appears only in the common factor ax that can be factored out.
The limit applies to h, not x. The terms inside square brackets depend only on
the base a and on h, but once the limit is evaluated, that term is some constant
(independent of x) that we denote by Ca . To summarize, we have found that
The derivative of an exponential function ax is of the form Ca ax where Ca
is a constant that depends only on the base a.
We now examine this in more detail with bases 2 and 10.
Example 10.3 (Derivative of 2x ) Write down the derivative of y = 2x using

the above result.
Solution. For base a = 2, we have

d2x
= C2 · 2x ,
dx
P Link to Google Sheets. The constant
where Ca in the derivative of ax is calculated
2h − 1 2h − 1 on this spreadsheet for a = 2. You can
C2 (h) = lim ≈ for small h.
h→0 h h copy and paste this to our own
spreadsheet and experiment with the
The decimal expansion value of C2 is determined in the next example. ♦
value of the base a. Try to find a value
of a between 2 and 3 for which Ca is
Example 10.4 (The value of C2 ) Find an approximation for the value of the close to 1.0.
constant C2 in Example 10.3 by calculating the value of the ratio (2h − 1)/h
h C2
for small (finite) values of h, e.g., h = 0.1, 0.01, etc. Do these successive 0.1 0.717735
approximations for C2 value approach a fixed real number? 0.01 0.695555
0.001 0.693387
Solution. We take these successively smaller values of h and compute the 0.0001 0.693171
0.00001 0.693150
value of C2 = (2h − 1)/h on a spreadsheet. 0.000001 0.693147
The results are shown in Table 10.1, where we find that C2 ≈ 0.6931. (The 0.0000001 0.693147
actual value has an infinitely long decimal expansion that we here represent
Table 10.1: The constant C2 in Exam-
by its first few digits.) Thus, the derivative of 2x is ple 10.4 is found by letting h get smaller
d2x and smaller. The value converges to
= C2 · 2x ≈ (0.6931) · 2x . C2 = 0.693147 .
dx
♦
h C 10
0.1 2.589254
0.01 2.329299
0.001 2.305238
Example 10.5 (The base 10 and the derivative of 10x ) Determine the
0.0001 2.302850
derivative of y = f (x) = 10x . 0.00001 2.302612
0.000001 2.302588
Solution. For base 10 we have 0.0000001 2.302585
10h − 1 Table 10.2: As in Table 10.1 but for

C10 (h) ≈ for small h. the constant C10 in Example 10.5. (The
h
We find, by similar approximation (Table 10.2), that C10 ≈ 2.3026, so that advantage of using a spreadsheet is that we
only need to change one cell to obtain this
d10x new set of values.)
= C10 · 10x ≈ (2.3026) · 10x .
dx Mastered Material Check
♦
11. What does it mean for a
x
Thus, the derivative of y = a is proportional to itself, but the constant of function f (x) to be proportional to
proportionality (Ca ) depends on the base. itself?
The natural base e is convenient for calculus

In Examples 10.3-10.5, we found that the derivative of ax is Ca ax , where the
constant Ca depends on the base. These constants are somewhat inconvenient,
but unavoidable if we use an arbitrary base. Here we ask:
Does there exists a convenient base (to be called “e”) for which the constant
is particularly simple, namely such that Ce = 1?
This is the property of the natural base that we next identify.
We can determine such a hypothetical base using only the property that
eh − 1
Ce = lim = 1. P Link to Google Sheets. The
h→0 h
calculation of a decimal approximation
This means that for small h to base e as shown in Table 10.3 .
eh − 1
≈ 1,
h
so that h approximation to e
eh − 1 ≈ h ⇒ eh ≈ h + 1 ⇒ e ≈ (1 + h)1/h . 0.1 2.5937425
0.01 2.7048138
More formally, 0.001 2.7169239
0.0001 2.7181459
e = lim (1 + h)1/h . (10.1)
h→0 0.00001 2.7182682
We can find an approximate decimal expansion for e by calculating the ra-
Table 10.3: We can use a spreadsheet to find
tio in Eqn. (10.1) for some very small (but finite value) of h on a spreadsheet. a decimal approximation to the natural base
Results are shown in Table 10.3. We find (e.g. for h = 0.00001) that e using Eqn. (10.1) and letting h approach
zero.
e ≈ (1.00001)100000 ≈ 2.71826.
To summarize, we have found that for the special base, e, we have the Mastered Material Check
following property: 12. Why can’t we simply plug in h = 0
into Eqn. (10.1) evaluate the limit?
The derivative of the function ex is ex . 1
13. Let h = n and rewrite Eqn (10.1).
The value of base e is obtained from the limit in Eqn. (10.1). This can be
14. Explain why each of
written in either of two equivalent forms. Properties 1. → 8. hold for the
The base of the natural exponential function is the real number defined as function ex .
follows:
1 n

e = lim (1 + h)1/h = lim 1 + .
h→0 n→∞ n
Properties of the function ex

We list below some of the key features of the function y = ex . Note that all
stem from basic manipulations of exponents as reviewed in Appendix B.1. ` Use the slider to adjust the value of
the base a in the function y = ax ;
1. ea eb = ea+b as with all similar exponent manipulations. Compare your result with the function
y = ex . Explain what you see for a > 1,
2. (ea )b = eab also stems from simple rules for manipulating exponents. a = 1, 0 < a < 1 and a = 0.
3. ex is a function that is defined, continuous, and differentiable for all real

numbers x.
4. ex > 0 for all values of x.
5. e0 = 1, and e1 = e.
6. ex → 0 for increasing negative values of x.
7. ex → ∞ for increasing positive values of x.
8. The derivative of ex is ex (shown in this chapter).

` Review: On this graph of f (x) = ex
Example 10.6 a) Find the derivative of ex at x = 0. add a generic tangent line at any point
x0 . (See Sections 5.1-5.2). Adjust a
b) Show that the tangent line at that point is the line y = x + 1. slider for x0 to get the configuration
shown in Fig. 10.4.
y
Solution. 4
ex
a) The derivative of ex is ex . At x = 0, e0 = 1.
3
b) The slope of the tangent line at x = 0 is therefore 1. The tangent line goes
2
through (0, e0 ) = (0, 1) so it has a y-intercept of 1. Thus the tangent line at tangent line
x = 0 with slope 1 is y = x + 1. This is shown in Figure 10.4. ♦
1
x
Composite derivatives involving exponentials −4 −2 2 4
Using the derivative of ex and the chain rule, we can now differentiate com- Figure 10.4: The function y = ex has the
posite functions in which the exponential function appears. property that its tangent line at x = 0 has
slope 1.
Example 10.7 Find the derivative of y = ekx .
Solution. Letting u = kx gives y = eu . Applying the simple chain rule leads to,
dy dy du
=
dx du dx
but
du dy
=k so = eu k = kekx .
dx dx
♦
We highlight this result for future use:
The derivative of y = ekx is dy
dy 15. Let y = e5x . What is dx ?
= kekx . 16. Let y = eπx . What is dy
dx dx ?
17. List all constants in Example 10.8.
18. List all variables in Example 10.8.
Example 10.8 (Chemical reactions) According to the collision theory of

bimolecular gas reactions, a reaction between two molecules occurs when
the molecules collide with energy greater than some activation energy, Ea ,
referred to as the Arrhenius activation energy. Ea > 0 is constant for the
given substance. The fraction of bimolecular reactions in which this collision
energy is achieved is
F = e−(Ea /RT ) ,
where T is temperature (in degrees Kelvin) and R > 0 is the gas constant.
Suppose that the temperature T increases at some constant rate, C, per unit
time.
Determine the rate of change of the fraction F of collisions that result in a
successful reaction.
Solution. This is a related rates problem involving an exponential function

that depends on the temperature, which depends on time, F = e−(Ea /RT (t )) . We
are asked to find the derivative of F with respect to time when the tempera-
ture increases.
We are given that dT /dt = C. Let u = −Ea /RT . Then F = eu . Using the
chain rule,
dF dF du dT
= .
dt du dT dt
Further, we have Ea , R,C are all constants, so
dF du Ea
= eu and = .
du dT RT 2
Assembling these parts, we have
dF Ea Ea CEa −(Ea /RT )
= eu 2 C = C T −2 e−(Ea /RT ) = e .
dt RT R RT 2
Thus, the rate of change of the fraction F of collisions that result in a success-
ful reaction is given by the expression above. ♦
Featured Problem 10.1 (Ricker model for fish population growth)

Salmon are fish with non-overlapping generations. The adults lay eggs
that are fertilized by males before the entire population dies. The eggs hatch
to form a new generation. In Featured Problem 1.1, we considered one model
for fish populations. Here we discuss a second model, the Ricker Equation,
wherein the fish population this year, N1 , is related to the population last year,
N0 , by the rule

N
r 1− K0
N1 = N0 e , r, K > 0. (10.2)
Here r is called an intrinsic growth rate, and K is the carrying capacity of the Figure 10.5: The functional form of the
Ricker equation.
population. We investigate the following questions.
(a) Is there a population level N0 that would stay constant from one year to
the next?
` Adjust the sliders to observe how the
parameters K and r affect the Ricker
equation 10.2. What is special about the
intersection of the two curves shown?
(b) Simplify the notation by setting x = N0 , y = N1 . Compute the derivative

dy/dx and interpret its meaning.
(c) What population level this year would result in the greatest possible
population next year?
The function ex satisfies a new kind of equation

We divert our attention momentarily to an interesting observation. We have
seen that the function
y = f (x) = ex
satisfies the relationship
dy
= f 0 (x) = f (x) = y.
dx
In other words, when differentiating, we get the same function back again.
We summarize this observation:
The function y = f (x) = ex is equal to its own derivative. It hence satisfies the
equation
dy
= y.
dx
An equation linking a function and its derivative(s) is called a differential
equation.
This is a new type of equation, unlike others previously seen in this course.
In Chapters 11-13 we show that these differential equations have many
applications to biology, physics, chemistry, and science in general.
10.3 Inverse functions and logarithms
1. Explain the concept of inverse function from both algebraic and geometric
points of view: given a function, determine whether (and for what re-
stricted domain) an inverse function can be defined and sketch that inverse
function.
2. Describe the relationship between the domain and range of a function and
the range and domain of its inverse function. (Review Appendix C.5).
3. Apply these ideas to the logarithm, which is the inverse of an exponential

function.
4. Reproduce the calculation of the derivative of ln(x) using implicit differen-

tiation.
In this chapter we defined the new function ex and computed its derivative.
Paired with this newcomer is an inverse function, the natural logarithm, ln(x).
Recall the following key ideas:
19. Are f (x) = xn and g(x) = x1/n also
• Given a function y = f (x), its inverse function, denoted f −1 (x) satisfies inverses of one another for even
integer n? Is this true for all x?
f ( f −1 (x)) = x, and f −1 ( f (x)) = x. 20. What is the inverse function for

y = x? Over what range of values is
the inverse defined?
• The range of f (x) is the domain of f −1 (x) (and vice versa), which implies 21. What is the inverse function to
that in many cases, the relationship holds only on some subset of the y = x2/3 and over what domain are
the two functions inverses of one
original domains of the functions. another?
• The functions f (x) = xn and g(x) = x1/n are inverses of one another for all
x when n is odd. ` Note symmetry about the line y = x
for this graph of f (x) = xn and
• The domain of a function (such as y = x2 or other even powers) must be g(x) = x1/n . Adjust the slider for n to
√ see how even and odd powers behave.
restricted (e.g. to x ≥ 0) so that its inverse function (y = x) is defined. What do you notice about the domain
over which g(x) is defined? Adjust the
• On that restricted domain, the graphs of f and f −1 are mirror images of slider for a to observe “corresponding
one another about the line y = x. Essentially, this stems from the fact that points” on the two graphs.
the roles of x and y are interchanged.
The natural logarithm is an inverse function for ex y

4
ex
For y = f (x) = ex we define an inverse function, shown on Figure 10.6. We
call this function the logarithm (base e), and write it as 2
ln(x)
y = f −1 (x) = ln(x).
−4 −2 2 4
We have the following connection: y = ex implies x = ln(y). The fact that −2

the functions are inverses also implies that y=x
−4 x
ln(x) x
e = x and ln(e ) = x.
Figure 10.6: The function y = ex is shown
ex
The domain of is −∞ < x < ∞, and its range is x > 0. For the inverse with its inverse, y = ln x.
function, this domain and range are interchanged, meaning that ln(x) is only
defined for x > 0 (its domain) and returns values in −∞ < x < ∞ (its range). As
shown in Figure 10.6, the functions ex and ln(x) are reflections of one another
about the line y = x.
Properties of the logarithm stem directly from properties of the exponen-
tial function. A review of these is provided in Appendix B.2. Briefly, Mastered Material Check
22. Give algebraic justification of the
1. ln(ab) = ln(a) + ln(b), three properties of logarithms.
2. ln(ab ) = b ln(a),
3. ln(1/a) = ln(a−1 ) = − ln(a).
Featured Problem 10.2 (Agroforestry) In agroforestry, the farming of

crops is integrated with growing of trees to benefit productivity and maintain
the health of an ecosystem. A tree can provide advantage to nearby plants
by creating better soil permeability, higher water retention, and more stable
temperatures. At the same time, trees produce shade and increased competi-
tion for nutrients. Both the advantage A(x) and the shading S(x) depend on
distance from the tree, with shading a dominant negative effect right under
the tree. Suppose that at a distance x from a given tree species, the net benefit
B to a crop plant can be expressed as the difference
Figure 10.7: Too close to a tree, shading
(grey) S(x) interferes with crop growth. Just
beyond this region, the advantage A(x) to
crop growth outweighs any disadvantage
2 /a2 2 /b2 due to shading. We seek to find the optimal
B(x) = A(x) −S(x), where A(x) = αe−x , S(x) = β e−x , αβ , a, b > 0 distance x for planting the crops.
` The Advantage A(x), the shading

(a) How far away from the tree will the two influences break even? (b) Find effect S(x), and the net benefit B(x) for
the optimal distance to plan crops so that they derive maximal benefit from a crop as functions of distance x from a
the nearby tree. tree are shown here. Move the sliders to
see how the spatial range a and the
magnitude β affect the graphs.
Derivative of ln(x) by implicit differentiation
Implicit differentiation is helpful whenever an inverse function appears.

Knowing the derivative of the original function allows us to compute the
derivative of its inverse by using their relationship. We use implicit differenti-
ation to find the derivative of y = ln(x).
First, restate the relationship in the inverse form, but consider y as the
dependent variable - that is think of y as a quantity that depends on x:
d y(x ) d
y = ln(x) ⇒ ey = x ⇒ e = x.
dx dx
Applying the chain rule to the left hand side,
dey dy dy dy 1 1
=1 ⇒ ey =1 ⇒ = y= .
dy dx dx dx e x
We have thus shown the following:
The derivative of ln(x) is 1/x:

` Tangent lines to the graphs of y = ex
d ln(x) 1 and y = ln(x) at corresponding points
= . are mirror images about the line y = x.
dx x
Adjust the slider to see the tangent lines
Inverse functions are mirror images of one another about the line y = x, since at various points along the curves. What
the role of independent and dependent variables are switched. Their tangent do we mean by “corresponding points”?
lines are also mirror images about the same line.
10.4 Applications of the logarithm
1. Describe the relationships between properties of ex and properties of its

inverse ln(x), and master manipulations of expressions involving both.
2. Use logarithms for base conversions.
3. Use logarithms to solve equations involving the exponential function, i.e.

solve A = ebt for t.)
4. Given a relationship such as y = axb , show that ln(y) is related linearly

to ln(x), and use data points for (x, y) to determine the values of a and b.
Using the logarithm for base conversion

The logarithm is helpful in changing an exponential function from one base
to another. We give some examples here. Mastered Material Check
23. Why might one base be preferred
Example 10.9 Rewrite y = 2x in terms of base e. over another?
Solution. We apply ln and then exponentiate the result. Manipulations of

exponents and logarithms lead to the desired results as follows:
y = 2x ⇒ ln(y) = ln(2x ) = x ln(2).
eln(y) = ex ln(2) ⇒ y = ex ln(2) .
We find (using a calculator) that ln(2) = 0.6931 . . . This coincides with the
value we computed earlier for C2 in Example 10.4, so we have
y = ekx where k = ln(2) = 0.6931 . . .
Example 10.10 Find the derivative of y = 2x .
Solution. In Example 10.9 we expressed this function in the alternate form
y = 2x = ekx with k = ln(2).
From Example 10.7 we have
dy
= kekx = ln(2)eln(2)x = ln(2)2x .
dx
Through the above base conversion and chain rule, we relate the constant C2
in Example 10.4 to the natural logarithm of 2: C2 = ln(2). ♦
The logarithm helps to solve exponential equations

Equations involving the exponential function can sometimes be simplified
and solved using the logarithm. We provide a few examples of this kind.
2
Example 10.11 Find zeros of the function y = f (x) = e2x − e5x .
2
Solution. We seek values of x for which f (x) = e2x − e5x = 0. We write
2
2x 5x2 2x 5x2 e5x 2 −2x
e −e =0 ⇒ e =e ⇒ =1 ⇒ e5x = 1.
e2x
2 −2x
Taking logarithm of both sides, and using the facts that ln(e5x ) = 5x2 − 2x
and ln(1) = 0, we obtain
2 −2x 2
e5x =1 ⇒ 5x2 − 2x = 0 ⇒ x = 0, .
5
We see that the logarithm is useful in the last step of isolating x, after simpli-
fying the exponential expressions appearing in the equation. ♦
Andromeda Strain, revisited. In Section 10.1 we posed the question: how

long does it take for the Andromeda strain population to attain a size of 6 · 1039
cells, i.e. to grow to an Earth-sized colony? We now solve this problem using
the continuous exponential function and the logarithm. Mastered Material Check
Recall that the bacterial doubling time is 20 min. If time is measured in 24. Verify that B(t ) agrees with
minutes, the number, B(t ) of bacteria at time t could be described by the Figure 10.1 and give powers of 2
smooth function: at t = 20, 40, 60, 80, . . . minutes.
B(t ) = 2t/20 . 25. When, in general, will B(t ) give a

power of 2?
Example 10.12 (The Andromeda strain) Starting from a single cell, how
long does it take for an E. coli colony to reach size of 6 · 1039 cells by doubling
every 20 minutes?
Solution. We compute the time it takes by solving for t in B(t ) = 6 · 1039 , as

shown below.
6 · 1039 = 2t/20 ⇒ ln(6 · 1039 ) = ln(2t/20 )

t
ln(6) + 39 ln(10) = ln(2).
20
Solving for t,
ln(6) + 39 ln(10) 1.79 + 39(2.3) 2643.27

t = 20 = 20 = 2643.4 min = hr.
ln(2) 0.693 60
Hence, it takes 44 hours (but less than 2 days) for the colony to “grow to the
size of planet Earth” (assuming the implausible scenario of unlimited growth).
♦
Example 10.13 (Using base e) Express the number of bacteria in terms of

base e (for practice with base conversions).
Solution. Given B(t ) = 2t/20 is the number of bacteria at time t, we proceed

as follows:
t
B(t ) = 2t/20 ⇒ ln(B(t )) = ln(2),
20
t ln(2)
eln(B(t )) = e 20 ln(2) ⇒ B(t ) = ekt where k = per min.
20
♦
The constant k has units of 1/time. We refer to k as the growth rate of the
bacteria. We observe that this constant can be written as:
ln(2)
k= .
doubling time
As we see next, this approach is helpful in scientific applications.
Logarithms help plot data that varies on large scale

Living organisms come in a variety of sizes, from the tiniest cells to the
largest whales. Comparing attributes across species of vastly different sizes basal
animal body weight metabolic
poses a challenge, as visualizing such data on a simple graph obscures both M (gm) rate (BMR)
extremes. mouse 25 1580
rat 226 873
Suppose we wish to compare the physiology of organisms of various sizes,
rabbit 2200 466
from that of a mouse to that of an elephant. An example of such data for dog 11700 318
metabolic rate versus mass of the animal is shown in Table 10.4. man 70000 202
horse 700000 106
It would be hard to see all data points clearly on a regular graph. For
this reason, it is helpful to use logarithmic scaling for either or both vari- Table 10.4: Animals of various sizes (mass
ables. We show an example of this kind of log-log plot, where both axes use M in gm) have widely different basal
metabolic rates (BMR, generally measured
logarithmic scales, in Figure 10.8. in terms of oxygen consumption rate, i.e. ml
In allometry, it is conjectured that such data fits some power function of O2 consumed per hr). A log-log plot of this
data is shown in Figure 10.8.
the form
y ≈ axb , where a, b > 0. (10.3)
Note: this is not an exponential function, but a power function with power b
and coefficient a.
Finding the allometric constants a and b using the graph in Fig 10.8 is
now explained. Mastered Material Check
26. Use software to plot the data given
Example 10.14 (Log transformation) Define Y = ln(y) and X = ln(x). in Table 10.4. Why is it so hard to
Show that (10.3) can be rewritten as a linear relationship between Y and X. plot on a regular graph?
Solution. We have
Y = ln(y) = ln(axb ) = ln(a) + ln(xb ) = ln(a) + b ln(x) = A + bX,
where A = ln(a). Thus, we have shown that X and Y are related linearly:
Y = A + bX, where A = ln(a).
This is the equation of a straight line with slope b and Y intercept A. ♦

Figure 10.8: A log-log plot of the data in

8 mouse Table 10.4, showing ln(BMR) versus ln(M ).
rat
ln(BM R)
rabbit
6 dog
human
horse
4
5 10 15 20
ln(M )
Example 10.15 (Finding the constants) Use the straight line superimposed
on the data in Figure 10.8 to estimate the values of the constants a and b.
Solution. We use the straight line that has been fitted to the data in Fig-
ure 10.8. The Y intercept is roughly 8.2. The line goes approximately
through (20, 3) and (0, 8.2) (open dots on plot) so its slope is ≈ (3 −
8.2)/20 = −0.26. According to the relationship we found in Example 10.14,
8.2 = A = ln(a) ⇒ a = e8.2 = 3640, and b = −0.26.
Thus, reverting to the original allometric relationship leads to

3640
y = axb = 3640x−0.26 = .
x0.26
From this we see that the metabolic rate y decreases with the size x of the
animal, as indicated by the data in Table 10.4.
10.5 Summary
1. We reviewed exponential functions of the form y = ax , where a > 0, the

base, is constant.
dy
2. The function y = ex is its own derivative, that is dx = ex . This function
dy
satisfies dx = y, which is an example of a differential equation.
3. If y = f (x), its inverse function is denoted f −1 (x) and satisfies f ( f −1 (x)) =
x and f −1 ( f (x)) = x. The graph of f −1 is the same as the graph of f re-
flected across the line y = x. The domain of a function may have to be
restricted so that its inverse function exists.
4. Let f (x) = ex . The inverse of this function is f −1 (x) = ln(x). The
derivative of ln(x) is 1x .
5. We can transform exponential relationships into linear relationships using
logarithms. Such transformations allow for more meaningful plots, and
can aid us in finding unknown constants in exponential relationships.
6. The applications in this chapter included:
(a) the Andromeda strain of E. coli (a bacterium) and its doubling;

(b) the Ricker equation for fish population growth from one year to the
next;
(c) chemical reactions: the fraction which result in a successful reaction;
(d) how the advantage and disadvantage of plants growing near a tree
depend on distance from the tree; and
(e) allometry: the relationship between body weight and basal metaboloci
rate.
1. Instead of 1 E. coli cell, suppose we began with 2 which also doubled every 20 min. How long would it take for the
population to grow to the size of the earth?
√
2. Given 3 ≈ 1.74205, compute without taking square roots:
(a) 33/2 ,
(b) 35/2 ,
dx
3. Let x = eρa . Determine da .
4. Consider the following log-log plot

5
log(y)
4
log(x)
1 2 3 4 5
(a) Let Y = log(y) and X = log(x). Find constants A and B such that Y = AX + B.
(b) Determine constants a and b such that y = axb .
Exercises
10.1. Polymerase Chain reaction (PCR). The polymerase chain reaction

(PCR) was invented by Mullis in 1983 to amplify DNA. It is based
on the fact that each strand of (double-stranded) DNA can act as a
template for the synthesis of the second (“complementary”) strand.
The method consists of repeated cycles of heating (which separates the
DNA strands) and cooling (allowing for new DNA to be assembled
on each strand). The reaction mixture includes the original DNA to be
amplified, plus enzymes and nucleotides, the components needed to
form the new DNA). Each cycle doubles the amount of DNA.
A particular PCR experiment consisted of 35 cycles.
(a) By what factor was the original DNA amplified? Give your answer
both in terms of powers of 2 and in approximate decimal (powers
of ten) notations.
(b) Use the approximation in the caption of Table 10.1 (rather than a
scientific calculator) to find the decimal approximation.
10.2. Invention of the game of chess. According to some legends, the
inventor of the game of chess (who lived in India thousands of years
ago) so pleased his ruler, that he was asked to chose his reward.
“I would be content with grains of wheat. Let one grain be placed on
the first square of my chess board, and double that number on the
second, double that on the third, and so on,” said the inventor. The
ruler gladly agreed.
A chessboard has 8 × 8 squares. How many grains of wheat would
be required for the last square on that board? Give your answer in
decimal notation.
Note: in the original wheat and chessboard problem, we are asked
to find the total number of wheat grains on all squares. This requires
summing a geometric series, and is a problem ideal for early 2nd term
calculus.
10.3. Computing powers of 2. In order to produce the graph of the con-
tinuous function 2x in Figure 10.2, it was desirable to generate many
points on that graph using simple calculations. Suppose you have an
ordinary calculator with the operations +, −, ×, /. You also know
√
that 2 ≈ 1.414.
How would you compute 2x for the values x = 7/2, x = −1/2,
and x = −5?
10.4. Exponential base requirement. Explain the requirement that a must
be positive in the exponential function y = ax . What could go wrong
if a was a negative base?
10.5. Derivative of 3x . Find the derivative of y = 3x . What is the value of the

multiplicative constant C3 that shows up in your calculation?
10.6. Graphing functions. Graph the following functions:
(a) f (x) = x2 e−x ,
(b) f (x) = ln(e2x ).
10.7. Changing bases. Express the following in terms of base e:
(a) y = 3x ,
1
(b) y = 7x ,
2
(c) y = 15x +2 .
Express the following in terms of base 2:
(d) y = 9x ,
(e) y = 8x ,
2 +3
(f) y = −ex .
Express the following in terms of base 10:
(g) y = 21x ,
(h) y = 1000−10x ,
2 −1
(i) y = 50x .
10.8. Comparing numbers expressed using exponents. Compare the
values of each pair of numbers (i.e. indicate which is larger):
(a) 50.75 , 50.65
(b) 0.4−0.2 , 0.40.2
(c) 1.0012 , 1.0013
(d) 0.9991.5 , 0.9992.3
10.9. Logarithms. Rewrite each of the following equations in logarithmic
form:
(a) 34 = 81,
(b) 3−2 = 19 ,
1 1
(c) 27− 3 = .
3
10.10. Equations with logarithms. Solve the following equations for x:
(a) ln x = 2 ln a + 3 ln b,
(b) loga x = loga b − 32 loga c.
10.11. Reflections and transformations. What is the relationship between
the graph of y = 3x and the graph of each of the following functions?
(a) y = −3x , (d) y = 3|x| ,
(b) y = 3−x , (e) y = 2 · 3x ,
(c) y = 31−x , (f) y = log3 x.

10.12. Equations with exponents and logarithms. Solve the following

equations for x:
(a) e3−2x = 5,
(b) ln(3x − 1) = 4,
(c) ln(ln(x)) = 2,
(d) eax = Cebx , where a 6= b and C > 0.
10.13. Derivative of exponential and logarithmic functions. Find the first Formula.
d 1
derivative for each of the following functions: dx (loga x) = x ln a
(a) y = ln(2x + 3)3 , 1

(f) y = a− 2 x ,
(b) y = ln3 (2x + 3),
(g) y = x3 · 2x ,
(c) y = ln(cos 12 x), x
(h) y = ee ,
(d) y = loga (x3 − 2x) et − e−t
(i) y = .
2
(e) y = e3x , et + e−t
10.14. Maximum, minimum and inflection points. Find the maximum
and minimum points as well as all inflection points of the following
functions:
1
(a) f (x) = x(x2 − 4), (d) f (x) = 1−x + 1+1 x , −1 < x < 1,
√ (0, 4)
(b) f (x) = x3 − ln(x), x > 0, (e) f (x) = x − 3 3 x,
(c) f (x) = xe−x , (f) f (x) = e−2x − e−x .

10.15. Using graph information. Shown in Figure 15 is the graph of y = Cekt
y = Cekt
for some constants C, k, and a tangent line. Use data from the graph to
determine C and k.
10.16. Comparing exponential functions. Consider the two functions
1. y1 (t ) = 10e−0.1t ,
(2, 0)
2. y2 (t ) = 10e0.1t .
Answer the following: Figure 10.9: Figure for Exercise 15; y = Cekt
and a tangent line.
(a) Which one is decreasing and which one is increasing?
(b) In each case, find the value of the function at t = 0.
(c) Find the time at which the increasing function has doubled from
this initial value.
(d) Find the time at which the decreasing function has fallen to half of
its initial value.
Note: these values of t are called the doubling time, and half-life,
respectively
10.17. Invasive species. An ecosystem with mature trees has a relatively con-
stant population of beetles (species 1) - around 109 . At t = 0, a single
reproducing invasive beetle (species 2) is introduced accidentally.
If this population grows at the exponential rate
N2 (t ) = ert , where r = 0.5 per month,
how long does it take for species 2 to overtake the population of the
resident species 1? Assume exponential growth for the entire duration.
10.18. Human population growth. It is sometimes said that the population
of humans on Earth is growing exponentially. This means year human population
(billions)
P(t ) = Cert , where r > 0. 1 0.2
1000 0.275
We investigate this claim. To this end, we consider the human popula- 1500 0.45
1650 0.5
tion beginning in year 1800 (t = 0). Hence, we ask whether the data in
1750 0.7
Table 10.5 fits the relationship 1804 1
1850 1.2
P(t ) = Cer(t−1800) , where t is time in years and r > 0? 1900 1.6
1927 2
1950 2.55
(a) Show that the above relationship implies that ln(P) is a linear 1960 3
function of time, and that r is the slope of the linear relationship 1980 4.5
(hint: take the natural logarithm of both sides of the relationship 1987 5
1999 6
and simplify). 2011 7
2020 7.7
(b) Use the data from Table 10.5 for the years 1800 to 2020 to investi-
gate whether P(t ) fits an exponential relationship (hint: plot ln(P), Table 10.5: The human population (billions)
where P is human population (in billions) against time t in years - over the years AD 1 to AD 2020.
we refer to this process as “transforming the data”.
(c) A spreadsheet can be used to fit a straight line through the trans-
formed data you produced in (b).
(i) Find the best fit for the growth rate parameter r using that
option.
(ii) What are the units of r?
(iii) What is the best fit value of C?
(d) Based on your plot of ln(P) versus t and the best fit values of r
and C, over what time interval was the population growing more
slowly than the overall trend, and when was it growing more
rapidly than this same overall trend?
(e) Under what circumstances could an exponentially growing popula-
tion be sustainable?
10.19. A sum of exponentials. Researchers that investigated the molecu-
lar motor dynein found that the number of motors N (t ) remaining
attached to their microtubule tracks at time t (in sec) after a pulse of
activation was well described by a double exponential of the form can we do better than ‘researchers’?
N (t ) = C1 e−r1 t + C2 e−r2 t , t ≥ 0.
They found that r1 = 0.1, r2 = 0.01 per second, and C1 = 75,C2 = 25

percent.
(a) Plot this relationship for 0 < t < 8 min. Which of the two expo-
nential terms governs the behaviour over the first minute? Which
dominates in the later phase?
(b) Now consider a plot of ln(N (t )) versus t. Explain what you see and
what the slopes and other aspects of the graph represent.
10.20. Exponential Peeling. The data in Table 10.6 is claimed to have been What is exponential peeling?
generated by a double exponential function of the form time N (t )
0.0000 100.0000
N (t ) = C1 e−r1 t + C2 e−r2 t , t ≥ 0. 0.1000 57.6926
0.2000 42.5766
0.3000 35.8549
Use the data to determine the values of the constants r1 , r2 , C1 , and C2 . 0.4000 31.8481
10.21. Shannon Entropy. In a recent application of information theory to 0.5000 28.8296
2.5000 4.7430
the field of genomics, a function called the Shannon entropy, H, was 4.5000 0.7840
considered. In it, a given gene is represented as a binary device: it can 6.0000 0.2032
be either “on” or “off” (i.e. being expressed or not). 8.0000 0.0336
If x is the probability that the gene is “on” and y is the probability that Table 10.6: Table for Exercise 20; data
it is “off”, the Shannon entropy function for the gene is defined as to be fit to a function of the form N (t ) =
C1 e−r1 t + C2 e−r2 t , t ≥ 0.
H = −x log(x) − y log(y)
Note that
• x and y being probabilities just means that they satisfy 0 < x ≤ 1,
and 0 < y ≤ 1 and
• the gene can only be in one of these two states, so x + y = 1.
Use these facts to show that the Shannon entropy for the gene is
greatest when the two states are equally probable, i.e. for x = y = 0.5.
10.22. A threshold function. The response of a regulatory gene to inputs
that affect it is not simply linear. Often, the following so-called
“squashing function” or “threshold function” is used to link the input x
to the output y of the gene:
1
y = f (x ) = ,
1 + e(ax+b)
where a, b are constants.
(a) Show that 0 < y < 1.
(b) For b = 0 and‘a = 1 sketch the shape of this function.
(c) How does the shape of the graph change as a increases?
10.23. Graph sketching. Sketch the graph of the function y = e−t sin πt.
10.24. The Mexican Hat. Consider the function
2 2 /3
y = f (x) = 2e−x − e−x
(a) Find the critical points of f .

(b) Determine the value of f at those critical points.

(c) Use these results and the fact that for very large x, f → 0 to draw a
rough sketch of the graph of this function.
(d) Comment on why this function might be called “a Mexican Hat”.
Note: The second derivative is not very informative here, and we

do not ask you to use it for determining concavity in this example.
However, you may wish to calculate it for practice with the chain rule.
10.25. The Ricker Equation. In studying salmon populations, a model
often used is the Ricker equation which relates the size of a fish
population this year, x to the expected size next year y. The Ricker
equation is
y = αxe−β x
where α, β > 0.
(a) Find the value of the current population which maximizes the
salmon population next year according to this model.
(b) Find the value of the current population which would be exactly
maintained in the next generation.
(c) Explain why a very large population is not sustainable.
Note: these populations do not actually change continuously, since
all the parents die before the eggs are hatched.
10.26. Spacing in a fish school. Life in a social group has advantages and
disadvantages: protection from predators is one advantage. Dis-
advantages include competition for food or resources. Spacing of
individuals in a school of fish or a flock of birds is determined by the
mutual attraction and repulsion of neighbours from one another: each
individual does not want to stray too far from others, nor get too close.
Suppose that when two fish are at distance x > 0 from one another, they
are attracted with “force” Fa and repelled with “force” Fr given by:
Fa = Ae−x/a
Fr = Re−x/r
where A, R, a, r are positive constants.

Note: A, R are related to the magnitudes of the forces, while a, r to the
spatial range of these effects.
(a) Show that at distance x = a, the first function has fallen to (1/e)
times its value at the origin. (Recall e ≈ 2.7.)
(b) For what value of x does the second function fall to (1/e) times its
value at the origin? Note that this is the reason why a, r are called
spatial ranges of the forces.
(c) It is generally assumed that R > A and r < a. Interpret what this
mean about the comparative effects of the forces.
(d) Sketch a graph showing the two functions on the same set of axes.
(e) Find the distance at which the forces exactly balance. This is called
the comfortable distance for the two individuals.
(f) If either A or R changes so that the ratio R/A decreases, does the
comfortable distance increase or decrease? Justify your response.
(g) Similarly comment on what happens to the comfortable distance
if a increases or r decreases.
10.27. Seed distribution. The density of seeds at a distance x from a parent
tree is observed to be
2 /a2
D(x) = D0 e−x ,
where a > 0, D0 > 0 are positive constants. Insects that eat these seeds
tend to congregate near the tree so that the fraction of seeds that get
eaten is
2 2
F (x) = e−x /b
where b > 0.
Note: These functions are called Gaussian or Normal distributions.
The parameters a, b are related to the “width” of these bell-shaped
curves.
The number of seeds that survive (i.e. are produced and not eaten by
insects) is
S(x) = D(x)(1 − F (x))
Determine the distance x from the tree at which the greatest number of
seeds survive.
10.28. Euler’s ‘e’. In 1748, Euler wrote a classic book on calculus, “Intro-
ductio in Analysin Infinitorum” [Euler, 1748]in which he showed that
the function ex could be written in an expanded form similar to an
(infinitely long) polynomial:
x2 x3
ex = 1 + x + + + ...
1·2 1·2·3
Use as many terms as necessary to find an approximate value for the
number e and for 1/e to 5 decimal places.
Note: in other mathematics courses we see that such expansions,
called power series, are central to approximations of many functions.
11
Differential equations for exponential growth and decay
In Section 10.2 we made an observation about exponential functions and a

new kind of equation - a differential equation - that such functions satisfy.
In this chapter we explore this observation in more detail. At first, this link
is based on the simple relationship between an exponential function and its
derivatives. Later, this expands into a more encompassing discussion of
1. how differential equations arise in scientific problems,

2. how we study their predictions, and
3. what their solutions can tell us about the natural world.
We begin by reintroducing these equations.
11.1 Introducing a new kind of equation

i A screencast summary of the
introduction: differential equations for
Section 11.1 Learning goals exponential growth and decay. Edu.Cr.
1. Explain that the exponential function and its derivative are proportional to
one another, and thereby satisfy a relationship of the form dy/dx = ky.
2. Give the definitions of a differential equation and of a solution to a differ-

ential equation.
3. Explain that y = ekt is a solution to the differential equation dy/dt = ky.
Observations about the exponential function

In Chapter 10, we introduced the exponential function y = f (x) = ex , and
noted that it satisfies the relationship
dex dy
= ex , ⇒ = y.
dx dx
The equation on the right (linking a function to its own derivative) is a new
kind of equation called a differential equation (abbreviated DE). We say that
f (x) = ex is a function that “satisfies” the equation, and we call this a solution
to the differential equation.
Note: The solution to an algebraic equation is a number, whereas the solution
to a differential equation is a function.
We call this a differential equation because it connects (one or more)
derivatives of a function with the function itself. Mastered Material Check
1. For whar constant C does y = Cex
Definition 11.1 (Differential equation) A differential equation is a mathe- satisfy the differential equation
dy/dx = y?
matical equation that relates one or more derivatives of some function to the
2. What function satisfies the DE
function itself. Solving the differential equation is the process of identifying dy/dz = y?
the function(s) that satisfies the given relationship.
We will be interested in applications in which a system or process varies

over time. For this reason, we will henceforth use the independent variable t,
for time in place of the former generic “x”.
Observations.
1. Consider the function of time: y = f (t ) = et . Show that this function

satisfies the differential equation Hint: Notice that we merely
changed the notation very slightly. Now
dy the derivative is “with respect to” t
= y. rather than x.
dt
2. The functions y = ekt (for k constant) satisfy the differential equation
dy
= ky. (11.1)
dt
We can verify by differentiating y = ekt , using the chair rule. Setting u = kt,
and y = eu , we have
dy dy du dy
= = eu · k = kekt = ky ⇒ = ky
dt du dt dt
Hence, we have established that y = ekt satisfies the DE (11.1).
It is interesting to ask: Is this is the only function that satisfies the differential
equation 11.1? Are there other possible solutions? What about a function
such as y = 2ekt or y = 400ekt ?
The reader should show that for any constant C, the function y = Cekt
is a solution to the DE (11.1). To do so, differentiate the function and plug
into (11.1). Verifying that the two sides of the equation are then the same Hint: Notice that the constant C in
establishes the result. While we do not prove it here, it turns out that y = Cekt front will appear in both the derivative
and the function, and so will not change
are the only functions that satisfy Eqn. (11.1). the equation.
Let us summarize what we have found out so far:
D I F F E R E N T I A L E Q UAT I O N S F O R E X P O N E N T I A L G ROW T H A N D D E C AY 229
Solutions to the differential equation
dy
= ky (11.2)
dt
are the functions
y = Cekt (11.3)
for C an arbitrary constant.
A few comments are in order. First, unlike algebraic equations - whose

solutions are numbers - differential equations have solutions that are func-
tions. Second, the constant k that appears in Eqn. (11.2), is the same as the Mastered Material Check
constant k in ekt . Depending on the sign of k, we get either 3. Give an example of an algebraic
equation and its solution.
a) exponential growth for k > 0, as illustrated in Figure 11.1(a), or 4. Verify that y = 3e−t satisfies
differential equation dy
dt = −y.
b) exponential decay for k < 0, as illustrated in Figure 11.1(b).
5. Why is ekt always positive?
Third, since ekt
is always positive, the constant C determines the sign of the 6. Plot, using software, y = Cet for
function as a whole - whether its graph lies above or below the t axis. each of C = −4, −2, 2 and 4.
A few curves of each type (C > 0,C < 0) are shown in each panel of 7. Plot, using software, y = Ce−t for
each of C = −4, −2, 2, and 4.
Figure 11.1. The collection of curves in a panel is called a family of solution
curves. The family shares the same value of k, but each member has a distinct
value of C. Next, we ask how to specify a particular member of the family as
the solution.
Figure 11.1: (a) A family of solutions to the

y y differential equation (DE) (11.2). These are
functions of the form y = Cekt for k > 0 and
arbitrary constant C. (b) Ampther family of
solutions of a DE of the form (11.2), but for
k < 0.
t
t
k>0 k<0
(a) (b)
The solution to a differential equation

Definition 11.2 (Solution to a differential equation) By a solution to a
differential equation, we mean a function that satisfies that equation.
We often refer to “solution curves” - the graphs of the family of solutions

of a differential equation, as shown, for example in the panels of Figure 11.1.
So far, we found that “many” functions can be valid solutions of the
differential equation (11.2), since we can chose the constant C arbitrarily
in the family of solutions y = Cekt . Hence, in order to distinguish one

specific solution of interest, we need additional information. This additional
information is called an initial value, or initial condition, and it specifies
one point belonging to the solution curve of interest. A common way to
set an initial value is to specify a fixed value of the function (say y = y0 ) at
time t = 0.
Definition 11.3 (Initial value) An initial value for a differential equation is

a specified, known value of the solution at some specific time point (usually at
time t = 0).
Example 11.1 Given the differential Eqn. (11.2) and the initial value ` Adjust the sliders in this interactive
graph to see how the values of k and C
affect the shape of the graph of the
y(0) = y0 , function y = Cekt as well as its initial
value y(0) = y0 .
find the value of C for the solution in Eqn. (11.3). Note the transitions that take place
when k changes from positive to
negative.
Solution. We proceed as follows:
kt k·0 0
y(t ) = Ce , so y(0) = Ce = Ce = C · 1 = C. 8. Given differential Eqn. (11.2) and
the initial value y(0) = 1, find C for
the solution in Eqn (11.3).
But, by the initial condition, y(0) = y0 . So,
9. Repeat the above but for the initial
value y(0) = 10.
C = y0
10. Draw the ty-plane with the
points (0, y0 ) for y0 = 1, 10.
and we have established that
11. Use differentiation to verify that the
unction y = 3e−0.5t in Example 11.2
y(t ) = y0 ekt , where y0 is the initial value. is a solution to dy/dt = −0.5y with
initial condition y(0) = 3.
♦
For example, in Figure 11.1, the initial value specifies that the solution we
want passes through a specific point in the ty-plane - namely, the point (0, y0 ).
Only one curve in the family of curves has that property. Hence, the initial
value picks out a unique solution.
Example 11.2 Find the solution to the differential equation
dy
= −0.5y
dt
that satisfies the initial condition y(0) = 3. Describe the behaviour of the
solution you have found.
Solution. The DE indicates that k = −0.5, so solutions are exponential

functions y = Ce−0.5t . The initial condition sets the value of C. From previous
discussion, we know that C = y(0) = 3. Hence, the solution is y = 3e−0.5t .
This is a decaying exponential. ♦
11.2 Differential equation for unlimited population growth
1. Recall the derivation of a model for human population growth and de-
scribe how it leads to a differential equation.
2. Identify that the solution to that equation is an exponential function.
3. Define per capita birth rates and rates of mortality, and explain the process
of estimating their values from assumptions about the population.
4. Compute the doubling time of a population from its growth rate and vice
versa.
i A screencast summary of the model
Differential equations are important because they turn up in the study of for (unlimited) human population
many natural processes that vary continuously. In this section we examine growth.
the way that a simple differential equation arises when we study continuous
uncontrolled population growth.
Here we set up a mathematical model for population growth. Let N (t ) be
the number of individuals in a population at time t. The population changes
with time due to births and mortality. (Here we ignore migration). Consider
the changes that take place in the population size between time t and t + h,
where ∆t = h is a small time increment. Then Mastered Material Check
12. What is the dependent variable in
this model? The independent
Change Number of Number of variable?
N (t + h) − N (t ) = = − (11.4)
in N births deaths 13. What are the units associated with
each variable in this model?
Eqn. (11.4) is just a “book-keeping” equation that keeps track of people en-
14. What does “x is proportional to y”
tering and leaving the population. It is sometimes called a balance equation.
mean?
We use it to derive a differential equation linking the derivative of N to the
value of N at the given time.
Notice that dividing each term by the time interval h, we obtain
N (t + h) − N (t )

Number of births Number of deaths
= − .
h h h
The term on the left “looks familiar”. If we shrink the time interval, h → 0, this
term is a derivative dN/dt, so
Rate of
     
Number of Number of
dN 
= change of N  =  births per  −  deaths per 
dt
per unit time unit time unit time
For simplicity, we assume that all individuals are identical and that the
number of births per unit time is proportional to the population size. Denote
by r the constant of proportionality. Similarly, we assume that the number of
deaths per unit time is proportional to population size with m the constant of
proportionality.
Both r and m have meanings: r is the average per capita birth rate, and m
is the average per capita mortality rate . Here, both are assumed to be fixed
positive constants that carry units of 1/time. This is required to make the
units match for every term in Eqn. (11.4). Then
number births per unit time

r = per capita birth rate = ,
population size
number deaths per unit time
m = per capita mortality rate = . 15. If there are 10 births/year in a
population size population of size 1000, what is the
Consequently, we have birth rate r? Give units.
16. If there are 11 deaths/year in a
population of size 1000, what is the
Number of births per unit time = rN,
mortality rate m? Give units.
Number of deaths per unit time = mN. 17. Given the above conditions, what is
the net growth rate k for such a
We refer to constants such as r, m as parameters. In general, for a given population? Give units. Is the
population growing or shrinking?
population, these would have specific numerical values that could be found
through experiment, by collecting data, or by making simple assumptions. In
Section 11.2, we show how some elementary assumptions about birth and
mortality could help to estimate approximate values of r and m.
Taking the assumptions and the form of the balance equation (11.4)
together we have arrived at:
dN
= rN − mN = (r − m)N. (11.5)
dt
This is a differential equation: it links the derivative of N (t ) to the func-
tion N (t ). By solving the equation (i.e. identifying its solution), we are be
able to make a projection about how fast a population is growing.
Define the constant k = r − m. Then k is the net growth rate, of the
population, so
dN
= kN, for k = (r − m).
dt
Suppose we also know that at time t = 0, the population size is N0 . Then:
• The function that describes population over time is (by previous results),
N (t ) = N0 ekt = N0 e(r−m)t . (11.6)
(The result is identical to what we saw previously, but with N rather than y
as the time-dependent function. We can easily check by differentiation that
this function satisfies Eqn. (11.5).)
• Since N (t ) represents a population size, it has to be non-negative to have

biological relevance. This is true so long as N0 ≥ 0.
• The initial condition N (0) = N0 , allows us to specify the (otherwise

arbitrary) constant multiplying the exponential function.
• The population grows provided k > 0 which happens when r − m > 0 i.e.
when birth rate exceeds mortality rate.
• If k < 0, or equivalently, r < m then more people die on average than are
born, so that the population shrinks and (eventually) go extinct.
A simple model for human population growth

The differential equation (11.5) and its initial condition led us to predict that number
a population grows or decays exponentially in time, according to Eqn. (11.6). of people
We can make this prediction quantitative by estimating the values of parame-
ters r and m. To this end, let us consider the example of a human population
and make further simplifying assumptions. We measure time in years. age
0 80
Assumptions.
Figure 11.2: We assume a uniform age
• The age distribution of the population is “flat”, i.e. there are as many 10 distribution to to determine the fraction
year-olds as 70 year olds. Of course, this is quite inaccurate, but a good of people who are fertile (and can give
place to start since it is easy to estimate some of the quantities we need. birth) or who are old (and likely to die).
While slightly silly, this simplification helps
Figure 11.2 shows such a uniform age distribution. estimate the desired parameters.
• The sex ratio is roughly 50%. This means that half of the population is
female and half male. number of
people fertile
• Women are fertile and can have babies only during part of their lives: we
assume that the fertile years are between age 15 and age 55, as shown in
Figure 11.3. age
0 15 55 80
• A lifetime lasts 80 years. This means that for half of that time a given
(55−15) Figure 11.3: We assume that only women
woman can contribute to the birth rate, or that 80 = 50% of women
between the ages of 15 and 55 years old are
alive at any time are able to give birth. fertile and can give birth. Then, according
to our uniform age distribution assumption,
• During a woman’s fertile years, we assume that on average, she has one half of all women are between these ages
baby every 10 years. and hence fertile.
• We assume that deaths occur only from old age (i.e. we ignore disease,
number of
war, famine, and child mortality.) people mortality occurs
here
• We assume that everyone lives precisely to age 80, and then dies instantly.
age
Based on the above assumptions, we can estimate the birthrate parameter r as 0 80
follows:
Figure 11.4: We assume that the people in
the age bracket 79-80 years old all die each
year, and that those are the only deaths. This,
number women years fertile number babies per woman too, is a silly assumption, but makes it easy
r= · ·
population years of life number of years to estimate mortality in the population.
Thus we compute that
1 1 1
r= · · = 0.025 births per person per year.
2 2 10
Note that this value is now a rate per person per year, averaged over the Mastered Material Check
entire population (male and female, of all ages). We need such an average 18. Under these assumptions, for a
rate since our model of Eqn. (11.5) assumes that individuals “are identical”. population size of 800, how many
male 35 year-olds would you
We now have an approximate value for the average human per capita birth expect? Women in their 60’s?
rate, r ≈ 0.025 per year. 19. Is the fertility assumption
reasonable? Why or why not?
Next, using our assumptions, we estimate the mortality parameter, m. With
20. Explain the units attached to the
the flat age distribution shown in Figure 11.2, there would be a fraction of birthrate parameter r.
1/80 of the population who are precisely removed by mortality every year
(i.e. only those in their 80th year.) In this case, we can estimate that the per
capita mortality is:
1
m= = 0.0125 deaths per person per year.
80
The net per capita growth rate is k = r − m = 0.025 − 0.0125 = 0.0125 per
person per year. We often refer to the constant k as a growth rate constant
and we also say that the population grows at the rate of 1.25% per year.
Example 11.3 Using the results of this section, find a prediction for the
population size N (t ) as a function of time t.
Solution. We have found that our population satisfies the equation

Human population
dN 20
= (r − m)N = kN = 0.0125N,
dt
so that 15
0.0125t N (t)
N (t ) = N0 e , (11.7)
10
where N0 is the starting population size. Figure 11.5 illustrates how this
function behaves, using a starting value of N (0) = N0 = 7 billion. ♦
5
20 40 60 80 100
Example 11.4 (Human population in 100 years) Given the initial condi-
t (years)
tion N (0) = 7 billion, determine the size of the human population at t = 100
years predicted by the model. Figure 11.5: Projected world population (in
billions) over 100 years, based on the model
Solution. At time t = 0, the population is N (0) = N0 = 7 billion. Then in in Eqn. (11.7) and assuming that the initial
billions, population is ≈ 7 billion.
N (t ) = 7e0.0125t
so that when t = 100 we would have
N (100) = 7e0.0125·100 = 7e1.25 = 7 · 3.49 = 24.43. 21. Based on Figure 11.5, when would
we expect the human population to
Thus, with a starting population of 7 billion, there would be about 24.4 billion reach 15 billion?
after 100 years based on the uncontrolled continuous growth model. ♦
A critique. Before leaving our population model, we should remember that
our projections hold only so long as some rather restrictive assumptions are
made. We have made many simplifications, and ignored many features that
would seriously affect these results. These include (among others),
• variations in birth and mortality rates that stem from competition for
resources and,
• epidemics that take hold when crowding occurs, and
• uneven distributions of resources or space.
We have also assumed that the age distribution is uniform (flat), but that is not
accurate: the population grows only by adding new infants, and this would
skew the distribution even if it is initially uniform. All these factors suggest
that some “healthy skepticism” should be applied to any model predictions.
Predictions may cease to be valid if model assumptions are not satisfied. This
caveat will lead us to think about more realistic models for population growth.
Certainly, the uncontrolled exponential growth would not be sustainable
in the long run. That said, such a model is a good starting point for a first
description of population growth, later to be adjusted.
Growth and doubling

In Chapter 10, we used base 2 to launch our discussion of exponential growth
and population doublings. We later discovered that base e is more convenient
for calculus, having a more elegant derivative. We also saw in Chapter 10,
that bases of exponents can be inter-converted. These skills are helpful in our
discussion of doubling times below.
The doubling time. How long would it take a population to double, given
that it is growing exponentially with growth rate k? We seek a time t such
that N (t ) = 2N0 . Then
N (t ) = 2N0 and N (t ) = N0 ekt ,
implies that the population has doubled when t satisfies
2N0 = N0 ekt , ⇒ 2 = ekt ⇒ ln(2) = ln(ekt ) = kt.
We solve for t. Thus, the doubling time, denoted τ is:
ln(2)
τ= .
k
22. What are the units associated
Example 11.5 (Human population doubling time) Determine the doubling with τ?
time for the human population based on the results of our approximate 23. The human population hit 3 billion
in 1959. How does this fit with our
growth model.
(imperfect) model?
Solution. We have found a growth rate of roughly k = 0.0125 per year for the
human population. Based on this, it would take
ln(2)
τ= = 55.45 years
0.0125
for the population to double. Compare this with the graph of Fig 11.5, and
note that over this time span, the population increases from 6 to 12 billion. ♦
Note: the observant student may notice that we are simply converting back
from base e to base 2 when we compute the doubling time.
We summarize an important observation:
In general, an equation of the form
dy
= ky
dt
that represents an exponential growth has a doubling time of
ln(2) y
τ= .
k
This is shown in Figure 11.6. We have discovered that based on the
uncontrolled growth model, the population doubles every 55 years! After 110
years, for example, there have been two doublings, or a quadrupling of the 2y0
population.
y0
Example 11.6 (A ten year doubling time) Suppose we are told that some t
animal population doubles every 10 years. What growth rate would lead to τ
such a trend?
Figure 11.6: Doubling time for exponential
growth.
Solution. In this case, τ = 10 years. Rearranging
ln(2)
τ= ,
k
we obtain
ln(2) 0.6931
k= = ≈ 0.07 per year.
τ 10
Thus, a growth rate of 7% leads to doubling roughly every 10 years. ♦
11.3 Radioactive decay
1. Describe the model for the number of radioactive atoms and explain how
this leads to a differential equation.
2. Determine the solution of the resulting differential equation.
3. Given the initial amount, determine the amount of radioactivity remaining

at a future time.
4. Describe the link between half-life of the radioactive material and its
decay rate; given the value of one, be able to find the value of the other.
A radioactive material consists of atoms that undergo a spontaneous

change. Every so often, some radioactive atom emits a particle, and decays
into an inert form. We call this a process of radioactive decay. For any one
atom, it is impossible to predict when this event would occur exactly, but
based on the behaviour of a large number of atoms decaying spontaneously,
we can assign a probability k of decay per unit time. Can we give an “for example..” for this?
In this section, we use the same kind of book-keeping (keeping track of
the number of radioactive atoms remaining) as in the population growth
example, to arrive at a differential equation that describes the process. Once
we have the equation, we determine its solution and make a long-term
prediction about the amount of radioactivity remaining at a future time.
Deriving the model

We start by letting N (t ) be the number of radioactive atoms at time t. Gener-
ally, we would know N (0), the number present initially. Our goal is to make
simple assumptions about the process of decay that allows us to arrive at a
mathematical model to predict values of N (t ) at any later time t > 0.
Assumptions.
(1) The process of radioactive decay is random, but on average, the probabil-
ity of decay for a given radioactive atom is k per unit time where k > 0 is
some constant.
(2) During each (small) time interval of length ∆t = h, a radioactive atom has
probability kh of decaying. This is merely a restatement of (1).
Suppose that at some time t, there are N (t ) radioactive atoms. Then, 24. Suppose a given atom has a 1%
chance of decay per 24 hours. What
according to our assumptions, during the time period t ≤ t ≤ t + h, on average is this atom’s probability of decay
khN (t0 ) atoms would decay. How many are there at time t + h? We can write per week? Per hour?
the following balance-equation:
     
Amount left Amount present Amount decayed
at time = at time  −  during time interval 
     

t +h t t ≤ t ≤ t +h
or, restated:
N (t + h) = N (t ) − khN (t ). (11.8)
Here we have assumed that h is a small time period. Rearranging Eqn. (11.8)
leads to
N (t + h) − N (t )
= −kN (t ).
h
Considering the left hand side of this equation, we let h get smaller and
smaller (h → 0) and recall that
N (t + h) − N (t ) dN
lim = = N 0 (t )
h→0 h dt
where we have used the notation for a derivative of N with respect to t. We

have thus shown that a description of the population of radioactive atoms
reduces to
dN
= −kN. (11.9)
dt
We have, once more, arrived at a differential equation that provides a link
between a function of time N (t ) and its own rate of change dN/dt. Indeed,
this equation specifies that dN/dt is proportional to N, but with a negative
constant of proportionality which implies decay.
Above we formulated the entire model in terms of the number of radioac-
tive atoms. However, as shown below, the same equation holds regardless of
the system of units used measure the amount of radioactivity
Example 11.7 Define the number of moles of radioactive material by y(t ) =

N (t )/A where A is Avogadro’s number (the number of molecules in 1 mole:
≈ 6.022 × 1023 - a dimensionless quantity, i.e. just a number with no associ-
ated units). Determine the differential equation satisfied by y(t ).
Solution. We write y(t ) = N (t )/A in the form N (t ) = Ay(t ) and substitute

this expression for N (t ) in Eqn. (11.9). We use the fact that A is a constant to
simplify the derivative. Then
dN Ady(t ) dy(t )
= −kN ⇒ = −k(Ay(t )) ⇒ A = A(−ky(t ))
dt dt dt
cancelling the constant A from both sides of the equations leads to
dy(t ) dy
= −ky(t ), or simply = −ky. (11.10)
dt dt
Thus y(t ) satisfies the same kind of differential equation (with the same
negative proportionality constant) between the derivative and the original
function. We will refer to (11.10) as the decay equation. ♦
Solution to the decay equation (11.10)

Suppose that initially, there was an amount y0 . Then, together, the differential
equation and initial condition are
dy
= −ky, y ( 0 ) = y0 . (11.11)
dt
We often refer to this pairing between a differential equation and an initial
condition as an initial value problem. Next, we show that an exponential
function is an appropriate solution to this problem
Example 11.8 (Checking a solution) Show that the function
y(t ) = y0 e−kt . (11.12)
is a solution to initial value problem (11.11).

Solution. Ee compute the derivative of the candidate function (11.12), and

rearrange, obtaining
dy(t ) d de−kt
= [y0 e−kt ] = y0 = −ky0 e−kt = −ky(t ).
dt dt dt
This verifies that for the derivative of the function is −k times the original
function, so satisfies the DE in (11.11). We can also check that the initial
condition is satisfied:
y(0) = y0 e−k·0 = y0 e0 = y0 · 1 = y0 .
Hence, Eqn. (11.12) is the solution to the initial value problem for radioactive
decay. For k > 0 a constant, this is a decreasing function of time that we refer
to as exponential decay. ♦
The half life

Given a process of exponential decay, how long would it take for half of the
original amount to remain? Let us recall that the “original amount” (at time
t = 0) is y0 . Then we are looking for the time t such that y0 /2 remains. We
must solve for t in
y0
y(t ) = .
2
We refer to the value of t that satisfies this as the half life.
Example 11.9 (Half life) Determine the half life in the exponential decay
described by Eqn. (11.12).
Solution. We compute:
y0 1
= y0 e−kt ⇒ = e−kt .
2 2
Now taking reciprocals:
1 y
2= = ekt .
e−kt
Thus we find the same result as in our calculation for doubling times, namely, y0
ln(2) = ln(ekt ) = kt,

y0
2
so that the half life is
ln(2)
τ= .
k
t
This is shown in Figure 11.7. τ
Example 11.10 (Chernobyl: April 1986) In 1986 the Chernobyl nuclear Figure 11.7: Half-life in an exponentially
power plant exploded, and scattered radioactive material over Europe. The decreasing process.
radioactive element iodine-131 (I131 ) has half-life of 8 days whereas cesium-

137 (Cs137 ) has half life of 30 years. Use the model for radioactive decay to
predict how much of this material would remain over time.
Solution. We first determine the decay constants for each of these two
elements, by noting that
ln(2)
k= ,
τ
and recalling that ln(2) ≈ 0.693. Then for I131 we have
ln(2) ln(2)
k= = = 0.0866 per day.
τ 8
Then the amount of I131 left at time t (in days) would be
yI (t ) = y0 e−0.0866t .
For Cs137
ln(2)
k= = 0.023 per year.
30
so that for T in years,
yC (T ) = y0 e−0.023T .
Note: we have used T rather than t to emphasize that units are different in the
two calculations done in this example.
Example 11.11 (Decay to 0.1% of the initial level) How long it would take
for I131 to decay to 0.1 % of its initial level? Assume that the initial level
occurred just after the explosion at Chernobyl.
Solution. We must calculate the time t such that yI = 0.001y0 : 25. Repeat the calculation in
Example 11.11 for Cesium.
0.001y0 = y0 e−0.0866t ⇒ 0.001 = e−0.0866t ⇒ ln(0.001) = −0.0866t. 26. Convert the Cesium decay time units
to days and repeat the calculation of
Therefore, Example 11.10 with the new time
ln(0.001) −6.9 units.
t= = = 79.7 days.
−0.0866 −0.0866 27. If the decay rate of a substance is
10% per day, what is its half-life?
Thus it would take about 80 days for the level of Iodine-131 to decay to 0.1%
of its initial level. ♦
11.4 Deriving a differential equation for the growth of cell mass

i Derivation of a differential equation
In Section 1.2, we asked how the size of a living cell influences the balance that describes cell growth resulting from
between the rates of nutrient absorption (called A) and consumption (denoted absorption and consumption of
nutrients.
C). But what if the two processes do not balance? What happens to the cell if
the rates are unequal?
If a cell absorb nutrients faster than nutrients are consumed (A > C), some
of the excess nutrients accumulate, and this buildup of nutrient mass can be
converted into cell mass. This can result in growth (increase of cell mass).
Conversely, if the consumption rate exceeds the rate of absorption of nutrients
(C > A), the cell has a shortage of metabolic “fuel”, and needs to convert
some of its own mass into energy reserves that can power its metabolism -
this would lead to loss in cell mass.
We can keep track of such changes in cell mass by using a simple “balance
equation”. The balance equation states that “the rate of change of cell mass is
the difference between the rate of nutrient (mass) coming in (A) and the rate
of nutrient (mass) being consumed (C), i.e.
dm
= A −C. (11.13)
dt
Each term in this equation must have the same units, mass of nutrient per
unit time. A contributes positively to mass increase, whereas C is a rate of
depletion that makes a negative contribution (hence the signs associated
with terms in the equation). It also makes sense to adopt the assumptions
previously made in Section 1.2 (and Featured Problem 9.1) that
A = k1 S, C = k2V , m = ρV ,
where S,V , ρ are the surface area, volume, and density of the cell, and Mastered Material Check What are the
units of k1 , k2 , ρ?
k1 , k2 , ρ are positive constants. Then Eqn. (11.13) becomes
dm d (ρV )
= A −C ⇒ = k1 S − k2V . (11.14)
dt dt
The above equation is rather general, and does not depend on cell shape.
Now consider the special case of a spherical cell for which V = (4/3)πr3 ,
S = 4πr2 . This simplification will permit us to convert the balance equation
into a differential equation that describes changes in cell radius over time.
Now Eqn. (11.14) can be rewritten as
d [ρ · (4/3)πr3 ]
= k1 (4πr2 ) − k2 (4/3)πr3 . (11.15)
dt
We can simplify the derivative on the right hand side using the chain rule, as
done in Featured Problem 9.1, obtaining
4π dr
ρ π (3r2 ) = k1 (4πr2 ) − k2 (4/3)πr3 . (11.16)
3 dt
What does this tell us about cell radius?
One way to satisfy Eqn. (11.16) is to set r = 0 in each term. While this
is a “solution” to the equation, it is not biologically interesting. (It merely
describes a “cell” of zero radius that never changes.) Suppose r 6= 0. In that
case, we can cancel out a factor of r2 from both sides of the equation. (We
can also cancel out 4π.) After some simplification, we arrive at
Hint: If we use units of
dr k2 dr 1 k2 µm (=10−6 m) for cell radius,
ρ = k1 − r, ⇒ = k1 − r .
dt 3 dt ρ 3 pg (=10−12 gm) for mass, and measure
time in hours, then approximate values
With appropriate units and taking into account typical cell size and density, of the constants are
ρ = 1pg µm−3 ,
this equation might look something like
k1 = 1pg µm−2 hr−1 , and
k2 = 0.3pg µm−3 hr−1 .
dr
= (1 − 0.1 r ) . (11.17) In that case, the equation for cell radius
dt is dr/dt = (1 − 0.1 · r ).
From a statement about how cell mass changes, we have arrived at a resultant
prediction about the rate of change of the cell radius. The equation we
obtained is a differential equation that tells us something about a growing cell.
In an upcoming chapter, we will build tools to be able to understand what this
equation says, how to solve it for the cell radius r (t ) as a function of time t,
and what such analysis predicts about the dynamics of cells with different
initial sizes.
11.5 Summary
1. A differential equation is a statement linking the rate of change of some

state variable with current values of that variable. An example is the
simplest population growth model: if N (t ) is population size at time t:
dN
= kN.
dt
2. A solution to a differential equation is a function that satisfies the equation.

For instance, the function N (t ) = Cekt (for any constant C) is a solution to
the unlimited population growth model (we check this by the appropriate
differentiation). Graphs of such solutions (e.g. N versus t) are called
solution curves.
3. To select a specific solution, more information (an initial condition) is

needed. Given this information, e.g. N (0) = N0 , we can fully characterize
the desired solution.
Scientific problem
or system
4. The decay equation is one representative of the same class of problems,
and has an exponentially decaying solution.
dy
= −ky, y ( 0 ) = y0 ⇒ Solution: y(t ) = y0 e−kt . (11.18)
dt
Facts, observations “Laws of Nature”
5. So far, we have seen simple differential equations with simple (exponen- assumptions, or statements about
tial) functions for their solutions. In general, it may be quite challenging hypotheses rates of change
to make the connection between the differential equation (stemming from

some application or model) with the solution (which we want in order to
understand and predict the behaviour of the system.)
Predictions about
In this chapter, we saw examples in which a natural phenomenon (popula- the systme
Mathematical
model
tion growth, radioactive decay, cell growth) motivated a mathematical model behaviour
that led to a differential equation. In both cases, that equation was derived by
making a statement that tracked the amount or number or mass of a system
over time. Numerous simplifications were made to derive each differential
equation. For example, we assumed that the birth and mortality rates stay Differential
Solutions to the
fixed even as the population grows to huge sizes. equation(s)
differential
describing the
equations
system
Figure 11.8: A “flow chart” showing

how differential equations originate from
scientific problems.
With regard to a larger context.
• Our purpose was to illustrate how a simple model is created, and what
such models can predict.
• In general, differential equation models are often based on physical laws

(“F = ma”) or conservation statements (“rate in minus rate out equals net
rate of change”, or “total energy = constant”).
• In biology, where the laws governing biochemical events are less formal,
the models are often based on some mix of speculation and reasonable
assumptions.
• In Figure 11.8 we illustrate how the scientific method leads to a cycle

between the mathematical models and their test and validation using
observations about the natural world.
1. Identify each of the following with either exponential growth or exponential decay:
(a) y = 20e3t ;
(b) y = 5e−3t ;
dy
(c) dt = 3t;
dy
(d) dx = −5x.
2. Determine the doubling time of the exponential growth function N (t ) = 500e2t .
3. Determine the half life of the of the exponential decay function N (t ) = 500e−2t .
4. Consider the following figure depicting exponential growth:

20
15
y
10
20 40 60 80 100
t
What is the doubling time of this function?
Exercises
11.1. Checking solutions of differential equations. A differential equation

is an equation in which some function is related to its own deriva-
tive(s).
For each of the following functions, calculate the appropriate deriva-
tive, and show that the function satisfies the indicated differential
equation
(a) f (x) = 2e−3x , f 0 (x) = −3 f (x)
(b) f (t ) = Cekt , f 0 (t ) = k f (t )
(c) f (t ) = 1 − e−t , f 0 (t ) = 1 − f (t )
11.2. Linear differential equations. Consider the function y = f (t ) = Cekt
where C and k are constants. For what value(s) of these constants does
this function satisfy the equation
dy
(a) dt = −5y,
dy
(b) dt = 3y.
Note: an equation which involves a function and its derivative is called

a differential equation.
11.3. Checking initial value solution to a differential equation. Check
that the function (11.6) satisfies the differential equation (11.2) and the
initial condition N (0) = N0 .
11.4. Solving linear differential equations. Find a function that satisfies
each of the following differential equations.
Note: all your answers should be exponential functions, but they may
have different dependent and independent variables.
dy
(a) = −y,
dt
dc
(b) = −0.1c and c(0) = 20,
dx
dz
(c) = 3z and z(0) = 5.
dt
11.5. Andromeda strain, revisited. In Chapter 10 we discussed the growth
of bacteria, starting from a single cell. The doubling time of the
bacteria was given as 20 min.
Find the appropriate differential equation that describes this growth,
the appropriate initial condition, and the exponential function (with
base e) that is the solution to that differential equation. Use units of
hours for time t.
11.6. Population growth in developed and developing countries. In
Canada, women have only about 2 children during their 40 years
of fertility, and people live to age 80. In underdeveloped countries,

people on average live to age 60 and women have a child roughly
every 4 years between ages 13 and 45.
Compare the per capita birth and mortality rates and the predicted pop-
ulation growth or decay in each of these scenarios, using arguments
analogous to those of Section 11.2.
Find the growth rate k in percent per year and the doubling time for
the growing population.
11.7. Population growth and doubling. A population of animals has a
per-capita birth rate of b = 0.08 per year and a per-capita death rate
of m = 0.01 per year. The population density, P(t ) is found to satisfy
the differential equation
dP(t )
= bP(t ) − mP(t )
dt
(a) If the population is initially P(0) = 1000, find how big the popula-
tion is in 5 years.
(b) When does the population double?
11.8. Rodent population. The per capita birthrate of one species of rodent
is 0.05 newborns per day. This means that, on average, each member
of the population results in 5 newborn rodents every 100 days. Sup-
pose that over the period of 1000 days there are no deaths, and that the
initial population of rodents is 250.
(a) Write a differential equation for the population size N (t ) at time t

(in days).
(b) Write down the initial condition that N satisfies.
(c) Find the solution, i.e. express N as some function of time t that
satisfies your differential equation and initial condition.
(d) How many rodents are there after 1 year ?
11.9. Growth and extinction of microorganisms.

(a) The population y(t ) of a certain microorganism grows continuously
and follows an exponential behaviour over time. Its doubling time
is found to be 0.27 hours. What differential equation would you use
to describe its growth ?
Note: you must find the value of the rate constant, k, using the
doubling time.
(b) With exposure to ultra-violet radiation, the population ceases to
grow, and the microorganisms continuously die off. It is found that
the half-life is then 0.1 hours. What differential equation would
now describe the population?
11.10. A bacterial population. A bacterial population grows at a rate propor-

tional to the population size at time t. Let y(t ) be the population size at
time t. By experiment it is determined that the population at t = 10 min
is 15, 000 and at t = 30 min it is 20, 000.
(a) What was the initial population?
(b) What is the population at time t = 60min?
11.11. Antibiotic treatment. A colony of bacteria is treated with a mild
antibiotic agent so that the bacteria start to die. It is observed that
the density of bacteria as a function of time follows the approximate
relationship b(t ) = 85e−0.5t where t is time in hours.
Determine the time it takes for half of the bacteria to disappear; this is
called the half-life.
Find how long it takes for 99% of the bacteria to die.
11.12. Two populations. Two populations are studied. Population 1 is found
to obey the differential equation
dy1 /dt = 0.2y1
and population 2 obeys
dy2 /dt = −0.3y2
where t is time in years.

(a) Which population is growing and which is declining?
(b) Find the doubling time (respectively half-life) associated with the
given population.
(c) If the initial levels of the two populations were y1 (0) = 100
and y2 (0) = 10, 000, how big would each population be at time t?
(d) At what time would the two populations be exactly equal?
11.13. The human population. The human population on Earth doubles
roughly every 50 years. In October 2000 there were 6.1 billion hu-
mans on earth.
(a) Determine what the human population would be 500 years later
under the uncontrolled growth scenario.
(b) How many people would have to inhabit each square kilometer of
the planet for this population to fit on earth? (Take the circumfer-
ence of the earth to be 40, 000 km for the purpose of computing its
surface area and assume that the oceans have dried up.)
11.14. Fish in two lakes. Two lakes have populations of fish, but the con-
ditions are quite different in these lakes. In the first lake, the fish
population is growing and satisfies the differential equation
dy
= 0.2y
dt
where t is time in years. At time t = 0 there were 500 fish in this

lake. In the second lake, the population is dying due to pollution. Its
population satisfies the differential equation
dy
= −0.1y,
dt
and initially there were 4000 fish in this lake.
At what time are the fish populations in the two lakes identical?
11.15. First order chemical kinetics. When chemists say that a chemical
reaction follows “first order kinetics”, they mean that the concentration
of the reactant at time t, i.e. c(t ), satisfies an equation of the form dc
dt =
−rc where r is a rate constant, here assumed to be positive. Suppose
the reaction mixture initially has concentration 1M (“1 molar”) and
that after 1 hour there is half this amount.
(a) Find the “half life” of the reactant.
(b) Find the value of the rate constant r.
(c) Determine how much is left after 2 hours.
(d) When is only 10% of the initial amount be left?
11.16. Chemical breakdown. In a chemical reaction, a substance S is broken
down. The concentration of the substance is observed to change
at a rate proportional to the current concentration. It was observed
that 1 Mole/liter of S decreased to 0.5 Moles/liter in 10 minutes.
(a) How long does it take until only 0.25 Moles per liter remain?
(b) How long does it take until only 1% of the original concentration
remains?
11.17. Half-life. If 10% of a radioactive substance remains after one year,
find its half-life.
11.18. Carbon 14. Carbon 14, or 14C, has a half-life of 5730 years. This
means that after 5730 years, a sample of Carbon 14, which is a ra-
dioactive isotope of carbon, has lost one half of its original radioactiv-
ity.
(a) Estimate how long it takes for the sample to fall to roughly 0.001 of
its original level of radioactivity.
(b) Each gram of 14C has an activity given here in units of 12 decays
per minute. After some time, the amount of radioactivity decreases.
For example, a sample 5730 years old has only one half the original
activity level, i.e. 6 decays per minute. If a 1 gm sample of material
is found to have 45 decays per hour, approximately how old is it?
Note: 14C is used in radiocarbon dating, a process by which the age
of materials containing carbon can be estimated. W. Libby received
the Nobel prize in chemistry in 1960 for developing this technique.
11.19. Strontium-90. Strontium-90 is a radioactive isotope with a half-life

of 29 years. If you begin with a sample of 800 units, how long does
it take for the amount of radioactivity of the strontium sample to be
reduced to
(a) 400 units
(b) 200 units
(c) 1 unit
11.20. More radioactivity. The half-life of a radioactive material is 1620 years.
(a) What percentage of the radioactivity remains after 500 years?
(b) Cobalt 60 is a radioactive substance with half life 5.3 years. It is
used in medical application (radiology). How long does it take
for 80% of a sample of this substance to decay?
11.21. Salt in a barrel. A barrel initially contains 2 kg of salt dissolved
in 20 L of water. If water flows in the rate of 0.4 L per minute and the
well-mixed salt water solution flows out at the same rate, how much
salt is present after 8 minutes?
11.22. Atmospheric pressure. Assume the atmospheric pressure y at a
height x meters above the sea level satisfies the relation
dy
= ky.
dx
If one day at a certain location the atmospheric pressures are 760
and 675 torr (unit for pressure) at sea level and at 1000 meters above
sea level, respectively, find the value of the atmospheric pressure
at 600 meters above sea level.
12
Solving differential equations
In Chapter 11, we introduced differential equations to keep track of continu-

ous changes in the growth of a population or the decay of radioactivity. We
encountered a differential equation that tracks changes in cell mass due to
nutrient absorption and consumption. Finally, we learned that the solutions to
a differential equation is a function. In applications studied, that function can
be interpreted as predictions of the behaviour of the system or process over
time.
In this chapter, we further develop some of these ideas. We explore
several techniques for finding and verifying that a given function is a solution
to a differential equation. We then examine a simple class of differential
equations that have many applications to processes of production and decay,
and find their solutions. Finally, we show how an approximation method
provides for numerical solutions of such problems.
12.1 Verifying that a function is a solution
1. Given a function, check whether that function does or does not satisfy a
given differential equation.
2. Verify whether a given function does or does not satisfy an initial condi-
tion.
In this section we concentrate on analytic solutions to a differential

equation. By analytic solution, we mean a “formula” such as y = f (x) that
satisfies the given differential equation. We saw in Chapter 11 that we can
check whether a function satisfies a differential equation (e.g., Example 11.8)
by simple differentiation. In this section, we further demonstrate this process.
Example 12.1 Show that the function y(t ) = (2t + 1)1/2 is a solution to the
differential equation and initial condition

dy 1
= , y(0) = 1.
dt y
Solution. First, we check the derivative, obtaining
dy(t ) d (2t + 1)1/2 1

= = (2t + 1)−1/2 · 2
dt dt 2
1 1
= (2t + 1)−1/2 = = .
(2t + 1)1/2 y
Hence, the function satisfies the differential equation. We must also verify
the initial condition. We find that y(0) = (2 · 0 + 1)1/2 = 11/2 = 1. Thus the
initial condition is also satisfied, and y(t ) is indeed a solution. ♦
Example 12.2 Consider the differential equation and initial condition

dy
= 1 − y, y(0) = y0 . (12.1)
dt
a) Show that the function y(t ) = y0 e−t is not a solution to this differential
equation.
b) Show that the function y(t ) = 1 − (1 − y0 )e−t is a solution.

LHS RHS
dy
Solution. 1−y
dt
a) To check whether y(t ) = y0 e−t is a solution to the differential equa- d [y0 e−t ]
1 − y0 e−t
tion (12.1), we substitute the function into each side (“left hand side", dt
LHS; “right hand side”. RHS) of the equation. We show the results in
the columns of Table 12.1. After some steps in the simplification, we see −y0 e−t
Table 12.1: The function y(t ) = y0 e−t is not
that the two sides do not match, and conclude that the function is not a a solution to the differential equation (12.1).
solution, as it fails to satisfy the equation Plugging the function into each side of the
DE and simplifying (down the rows) leads to
b) Similarly, we check the second function. The calculations expressions that do not match.
are shown in columns of Table 12.2. We find that RHS=LHS, so the

differential equation is satisfied. Finally, let us show that the initial condi-
tion y(0) = y0 is also satisfied. Plugging in t = 0 we have LHS RHS
dy
1−y
y(0) = 1 − (1 − y0 )e0 = 1 − (1 − y0 ) · 1 = 1 − (1 − y0 ) = y0 . d
dt
[1 − (1 − y0 )e−t ] 1 − 1 − (1 − y0 )e−t

dt
Thus, both differential equation and initial condition are satisfied.
de−t
−(1 − y0 ) (1 − y0 )e−t
♦ dt
(1 − y0 )e−t X
Example 12.3 (Height of water draining out of a cylindrical container) A Table 12.2: (b) The function y(t ) = 1 −
cylindrical container with cross-sectional area A has a small hole of area a (1 − y0 )e−t is a solution to the differential
at its base, through which water leaks out. It can be shown that height of equation (12.1). The expressions we get
by evaluating each side of the differential
water h(t ) in the container satisfies the differential equation equation do match.
dh √
= −k h, (12.2)
dt
S O LV I N G D I F F E R E N T I A L E Q UAT I O N S 251
(where k is a constant that depends on the size and shape of the cylinder and
√
its hole: k = Aa 2g > 0 and g is acceleration due to gravity.) Show that the
function
p t 2
h(t ) = h0 − k (12.3)
2
is a solution to the differential equation (12.2) and initial condition h(0) = h0 . Mastered Material Check
1. Draw a diagram of the system
described in Example 12.3.
2. What set of units would be
Solution. We first easily verify that the initial condition is satisfied. Substi- reasonable for each of the
tute t = 0 into the function (12.3). Then we find h(0) = h0 , verifying the parameters in Example 12.3.
initial conditions. 3. Create a table to organize the

calculations for this example,
To show that the differential equation (12.2) is satisfied, we differentiate similar to Tables 12.1 and 12.2.
the function in Eqn. (12.3):
dh(t ) t −k

d p t 2 p
= h0 − k = 2 h0 − k ·
dt dt 2 2 2
t
p q
= −k h0 − k = −k h(t ).
2
Here we have used the power law and the chain rule, remembering that h0 , k
p
are constants. Now we notice that, using Eqn. (12.3), the expression for h(t )
exactly matches what we have computed for dh/dt. Thus, we have shown
that the function in Eqn. (12.3) satisfies both the initial condition and the
differential equation. ♦
As shown in Examples 12.1- 12.3, if we are told that a function is a

solution to a differential equation, we can check the assertion and verify that
it is correct or incorrect. A much more difficult task is to find the solution of a
new differential equation from first principles.
In some cases, integration, learned in second semester calculus, can be
used. In others, some transformation that changes the problem to a more
familiar one is helpful - an example of this type is presented in Section 12.2.
In many cases, particularly those of so-called non-linear differential equa-
tions, great expertise and familiarity with advanced mathematical methods
are required to find the solution to such problems in an analytic form, i.e. as
an explicit formula. In such cases, approximation and numerical methods are
helpful.
12.2 Equations of the form y0 (t ) = a − by
1. Define steady states of a differential equation, and be able to find such

special solutions.
2. Starting with the differential equation (12.4) y0 = a − by, find a new

differential equation for the deviation away from a steady state, z(t ) and
show that it is a simple decay equation.
3. Use the transformed (decay) equation to find the solution for z(t ) and, for
y(t ) in the original equation, (12.4).
4. Explain Newton’s Law of Cooling (NLC), and the differential equation of

the same type, y0 = a − by. Find its solution and explain what this solution
means.
5. Use the solution to NLC to predict the temperature of a cooling or heating

object over time.
6. Describe a variety of related examples, and use the same methods to solve
and interpret these (examples include chemical production and decay, the
velocity of a skydiver, the concentration of drug in the blood, and others).
In this section we introduce an important class of differential equations

that have many applications in physics, chemistry, biology, and other applica-
tions. All share a similar structure, namely all are of the form
dy
= a − by, y(0) = y0 . (12.4)
dt
First, we show how a solution to such equation can be found. Then, we
examine a number of applications.
Special solutions: steady states

We first ask about “special solutions” to the differential equation (12.4) in
which there is no change over time. That is, we ask whether there are values
of y for which dy/dt = 0. i An explanation of the way we find
solutions to equations of the form
From (12.4), we find that such solutions would satisfy dy
dt = a − by, with y(0) = y0 .
dy a
=0 ⇒ a − by = 0 ⇒ y= .
dt b
In other words, if we were to start with the initial value y(0) = a/b, then
that value would not change, since it satisfies dy/dt = 0, so that the solution
at all future times would be y(t ) = a/b. (Of course, this is a perfectly good
function; it is simply a function that is alway constant.)
We refer to such constant solutions as Steady States.
Figure 12.1: y = a/b is a constant solution to
the differential equation in (12.4). We call
this type of solution a steady state.
Other solutions: away from steady state

What happens if we start with a value of y that is not exactly at the “special”
steady state? Let us rewrite the DE in a more suggestive form,
dy dy a
= a − by ⇒ = −b y − ,
dt dt b
(having factored out −b). The advantage is that we recognize the expression
(y − ab ) as the difference, or deviation of y away from its steady state value.
(That deviation could be either positive or negative, depending on whether
y is larger or smaller than a/b.) We ask whether this deviation gets larger
or smaller as time goes by, i.e., whether y gets further away or closer to its
steady state value a/b.
Define z(t ) as that deviation, that is
a
z(t ) = y(t ) − , Figure 12.2: We define z(t ) as the deviation
b of y from its steady state value. Here we
Then, since a, b are constants, we recognize that show two typical initial values of z, where
z0 = y0 − ab .
dz dy
= .
dt dt
Second, the initial value of z follows simply from the initial value of y:
a a
z(0) = y(0) − = y0 − .
b b
Now we can transform the equation (12.4) into a new differential equation
for the variable z by using these two facts. We can replace the y derivative by
the z derivative, and also, using Eqn. (12.4), find that
dz dy a
= = −b y − = −bz.
dt dt b
Hence, we have transformed the original DE and IC into the new problem
Figure 12.3: The deviation away from steady
dz h ai state (blue, grey curves) is z(t ) = y(t ) − a/b.
= −bz, z(0) = z0 , where z0 = y0 − .
dt b We can solve the differential equation for
But this is the familiar decay initial value problem that we have already z(t ) because it is a simple exponential
decay equation. Here we show two typical
solved before. So solutions for z.
z(t ) = z0 e−bt .
We have arrived at the conclusion that the deviation from steady state decays
exponentially with time, provided that b > 0. Hence, we already know that y
should get closer to the constant value a/b as time goes by!
We can do even better than this, by transforming the solution we found
for z(t ) into an expression for y(t ). To do so, use the definition once more,
setting Figure 12.4: Finally, we can determine the
a a −bt solution y(t ).
z(t ) = z0 e−bt ⇒ y(t ) − = y0 − e .
b b
Solving for y(t ) then leads to
a a −bt ` Adjust the sliders to see how the
y(t ) = + y0 − e . (12.5)
b b parameters a and b and the initial value
y0 affect the shape of the function y(t )
in the formula (12.5).
y
Example 12.4 (a = b = 1) Suppose we are given the differential equation 3
and initial condition Solutions to
dy differential equation
= 1 − y, y(0) = y0 . (12.6) dy
dt 2 dt = 1 − y
Determine the solution to this differential equation.
Solution. 1
By substituting a = 1, b = 1 in the solution found above, we observe that
y(t ) = 1 − (1 − y0 )e−t .
0.5 1 1.5 2
time, t
Representative curves in this family of solutions are shown in Figure 12.5 for
various initial values y0 . ♦ Figure 12.5: Solutions to Eqn. (12.6) are
We now apply the methods to a number of examples. functions that approach y = 1.
Featured Problem 12.1 (Predicting the size of a growing cell) Find a 4. Find the steady state of Eqn. (12.6).
solution to the differential equation (11.17) for the radius of a growing cell 5. From Figure 12.5, determine what
r (t ) (in units of µm=10−6 m) as a function of time t (in hours), that is find were the four different initial
conditions used.
r (t ) assuming that at time t = 0 the cell is 2µm in radius.
6. Rewrite these four initial conditions
By solving the above problem, we get a detailed prediction of cell growth as the initial deviations away from
steady state, that is, give the initial
based on assumed rates of nutrient intake and consumption. values, z0 of the deviation.
Newton’s law of cooling

Consider an object at temperature T (t ) in an environment whose ambient
temperature is E. Depending on whether the object is cooler or warmer than
the environment, it heats up or cools down. From common experience we
know that, after a long time, the temperature of the object equilibrates with its
environment.
Isaac Newton formulated a hypothesis to describe the rate of change of
temperature of an object. He assumed that
The rate of change of temperature T of an object is proportional to the
difference between its temperature and the ambient temperature, E.
To rephrase this statement mathematically, we write
dT
is proportional to (T (t ) − E ).
dt
This implies that the derivative dT /dt is some constant multiple of the term
(T (t ) − E ). However, the sign of that constant requires some discussion.
Denote the constant of proportionality by α temporarily, and suppose α ≥ 0.
Let us check whether the differential equation
dT
= α (T (t ) − E ),
dt
makes physical sense. Mastered Material Check
Suppose the object is warmer than its environment (T (t ) > E). Then 7. What can we say about the units
T (t ) − E ) > 0 and α ≥ 0 implies that dT /dt > 0 which says that the of T and E?
temperature of the object should get warmer! But this does not agree with our
everyday experience: a hot cup of coffee cools off in a chilly room. Hence
α ≥ 0 cannot be correct. Based on this, we conclude that Newton’s Law of
Cooling, written in the form of a differential equation, should read:
dT
= k(E − T (t )), where k > 0. (12.7)
dt
Note: the sign of the term in braces has been switched.
Typically, given the temperature at some initial time T (0) = T0 , we want to
predict T (t ) for later time.
Example 12.5 Consider the temperature T (t ) as a function of time. Solve

the differential equation for Newton’s law of cooling
dT Mastered Material Check
= k (E − T ),
dt 8. Fill in the details for Example 12.5.
together with the initial condition T (0) = T0 . 9. In Figure 12.6, what are the five
different initial temperatures, T0
Solution. As before, we transform the variable to reduce the differential corresponding to each solution
curve?
equation to one that we know how to solve. This time, we select the new
10. In Figure 12.6, how many curves
variable to be z(t ) = E − T (t ). Then, by steps similar to previous examples, represent a heating object and how
we find that many a cooling object?
dz(t )
= −kz.
dt 20
We also rewrite the initial condition in terms of z, leading to z(0) = E −
temperature, T
15
T (0) = E − T0 . After carrying out Steps 1-3 as before, we find the solution for
T (t ), 10
T (t ) = E + (T0 − E )e−kt . (12.8)
5
In Figure 12.6 we show a family of curves of the form of Eqn. (12.8) for
t
five different initial temperature values (we have set E = 10 and k = 0.2 for all 0
0 2 4 6 8 10 12 14
these curves). ♦
Next, we interpret the behaviour of these solutions. Figure 12.6: Temperature versus time, T (t ),
for a cooling object.
Example 12.6 Explain (in words) what the form of the solution in Eqn. (12.8)
of Newton’s law of cooling implies about the temperature of an object as it
warms or cools.
Solution. We make the following remarks
• It is straightforward to verify that the initial temperature is T (0) = T0

(substitute t = 0 into the solution of Eqn. (12.8)). Now examine the time
dependence. Only one term, e−kt depends on time. Since k > 0, this is an
exponentially decaying function, whose magnitude shrinks with time. The
whole term that it multiplies, (T0 − E )e−kt , continually shrinks. Hence,
T (t ) = E + (T0 − E )e−kt ⇒ as t → ∞, e−kt → 0,

so T (t ) → E.
Thus the temperature of the object always approaches the ambient temper-
ature. This is evident in the solution curves shown in Figure 12.6.
• We also observe that the direction of approach (decreasing or increasing)

depends on the sign of the constant (T0 − E ). If T0 > E, the temperature
approaches E from above, whereas if T0 < E, the temperature approaches E
from below. Mastered Material Check
11. Consider three cups of coffee left in
• In the specific case that T0 = E, there is no change at all. T = E satisfies a 20◦ C room. If one is iced, another
dT /dt = 0, and corresponds to a steady state of the differential equation, is piping hot, and the third is room
temperature, which cup will not
as previously defined.
change temperature? Which, thus,
represents a steady state?
Steady states are studied in more detail in Chapter 13.
12. Convert the temperatures in
Example 12.7 to Fahrenheit and
Using Newton’s law of cooling to solve a mystery repeat.
Now that we have a detailed solution to the differential equation representing

Newton’s law of cooling, we can apply it to making exact determinations of
temperature over time, or of time at which a certain temperature was attained.
The following example illustrates an application of this idea.
Example 12.7 (Murder mystery) It is a dark clear night. The air temper-
ature is 10◦ C. A body is discovered at midnight. Its temperature is 27◦ C.
One hour later, the body has cooled to 24◦ C. Use Newton’s law of cooling to
determine the time of death.
i Details of the calculations for
Solution. We assume that body-temperature just before death was 37◦ C Example 12.7.
(normal human body temperature). Let t = 0 be the time of death. Then the
initial temperature is T (0) = T0 = 37◦ C. We want to find the time elapsed
until the body was found, i.e. time t at which the temperature of the body had
cooled down to 27◦ C. We assume that the ambient temperature, E = 10, was
constant. From Newton’s law of cooling, the body temperature satisfies
dT
= k(10 − T ).
dt
From previous work and Eqn. (12.8), the solution to this DE is
T (t ) = 10 + (37 − 10)e−kt .
We do not know the value of the constant k, but we have enough informa-
tion to find it. First, at discovery, the body’s temperature was 27◦ . Hence at
time t
27 = 10 + 27e−kt ⇒ 17 = 27e−kt .
Also at t + 1 (one hour after discovery), the temperature was 24◦ C, so
T (t + 1) = 10 + (37 − 10)e−k(t +1) = 24, ⇒ 24 = 10 + 27e−k(t +1) .
Thus,
14 = 27e−k(t +1) .
We have two equations for the two unknowns t and k. To solve for k, take
a ratio of the sides of the equations. Then
27e−k(t +1)

14 14
= = e−k ⇒ −k = ln = −0.194 .
17 27e−kt 17
This is the constant that describes the rate of cooling of the body. Mastered Material Check
To find the time of death, t, use 13. Give the concluding sentence for
Example 12.7. Be sure to include an

17 actual time of death, given that the
17 = 27e−kt ⇒ −kt = ln = −0.4626 body was discovered at midnight.
27
14. Use a plotting program to graph
finally, solving for t, we get T (t ) for Example 12.7.
15. Use your plot to estimate how long
0.4626 0.4626 it took for the body to cool off to
t= = = 2.384 hours. 33◦ C.
k 0.194
♦
Related applications and further examples

Having gained familiarity with specific examples, we now return to the
general case and summarize the results.
The differential equation and initial condition
dy
= a − by, y(0) = y0 (12.9)
dt
has the solution
a a
y(t ) = − − y0 e−bt . (12.10)
b b
Suppose that a, b > 0 in Eqn. (12.9). Then we can summarize the behaviour
of the solutions (12.10) as follows:
• The time dependence of Eqn. (12.10) is contained in the term e−bt , which
(for b > 0) is exponentially decreasing. As time increases, t → ∞, the
exponential term becomes negligibly small, so y → a/b.
• If initially y(0) = y0 > a/b, then y(t ) approaches a/b from above, whereas
if y0 < a/b, it approaches a/b from below.
• If initially y0 = a/b, there is no change at all (dy/dt = 0). Thus y = a/b is

a steady state of the DE in Eqn. (12.9).
Recognizing such general structure means that we can avoid repeating

similar calculations from scratch in related examples. Newton’s law of
cooling is one representative of the class of differential equations of the form
Eqn. (12.9). If we set a = kE, b = k and T = y in Eqn. (12.9), we get back to
Eqn. (12.7). As expected from the general case, T approaches a/b = E, the
ambient temperature, which corresponds to a steady state of NLC.
Next, we describe other examples that share this structure, and hence
similar dynamic behaviour.
Friction and terminal velocity A falling object accelerates under the force
of gravity, but friction slows down this acceleration. The differential equa- Note. Eqn. (12.11) comes from a simple
force balance:
tion satisfied by the velocity v(t ) of the falling object with friction is
ma = Fgravity − Fdrag ,
dv
= g − kv (12.11) and from the assumption that Fdrag = µv,
dt
where µ > 0 is the “drag coefficient”.
where g > 0 is acceleration due to gravity and k > 0 is a constant representing Dividing both sides by m and
replacing a by dv/dt leads to this
the effect of air resistance. (In contrast to the “upwards pointing” coordinate
equation, with k = µ/m.
system used in Example 4.10, here we focus on how the magnitude of the
velocity changes with time.) Usually, a frictional force is assumed to be
proportional to the velocity of the object, and to act in a direction that slows
it down. (This accounts for the negative sign in Eqn. (12.11).) Parachutes
operate on the principle of enhancing that frictional force to damp out the
acceleration of a skydiver. Hence, Eqn. (12.11) is often called the skydiver
equation.
Example 12.8 Use the general results for Eqn. (12.9) to write down the
solution to the differential equation (12.11) for the velocity of a skydiver
given the initial condition v(0) = v0 . Interpret your results in a simple
description of what happens over time.
Solution. Eqn. (12.11) is of the same form as Eqn. (12.9), and has the same 16. Assign appropriate units to each of
type of solutions. We merely have to adjust the notation, by identifying the parameters in Example 12.8.
17. When a sky-diver steps into the void,
v(t ) → y(t ), g → a, k → b, v0 → y0 . her initial vertical velocity is zero.
Write down her velocity v(t ) based
Hence, without further calculation, we can conclude that the solution of on results of Example 12.8 .
(12.11) together with its initial condition is:
g g
v(t ) = − − v0 e−kt . (12.12)
k k
The velocity is initially v0 , and eventually approaches g/k which is the
steady state or terminal velocity for the object. Depending on the initial
speed, the object either slows down (if v0 > g/k) or speed up (if v0 < g/k) as
it approaches the terminal velocity. ♦
Chemical production and decay. A chemical reaction inside a fixed reaction
volume produces a substance at a constant rate Kin . A second reaction results
in decay of that substance at a rate proportional to its concentration. Let c(t )
denote the time-dependent concentration of the substance, and assume that
time is measured in units of hours. Then, writing down a balance equation
leads to a differential equation of the form
dc
= Kin − γc. (12.13)
dt
Here, the first term is the rate of production and the second term is the
rate of decay. The net rate of change of the chemical concentration is then
the difference of the two. The constants Kin > 0, γ > 0 represent the rate
of production and decay - recall that the units of each term in any equation
have to match.For example, if the concentration c is measured in units of
milli-Molar (mM), then dc/dt has units of mM/hr, and hence Kin must have
units of mM/h and γ must have units of 1/hr.
Example 12.9 Write down the solution to the DE (12.13) given the initial
condition c(0) = c0 . Determine the steady state chemical concentration.
Solution. Translating notation from the general case to this example,
c(t ) → y(t ), Kin → a, γ → b.
Then we can immediately write down the solution:

Kin Kin
c(t ) = − − c0 e−γt . (12.14)
γ γ
Regardless of its initial condition, the chemical concentration will approach a
steady state concentration is c = Kin /γ. ♦
In this section we have seen that the behaviour found in the general case of
the differential equation (12.4), can be reinterpreted in each specific situation
of interest. This points to one powerful aspect of mathematics, namely the
ability to use results in abstract general cases to solve a variety of seemingly
unrelated scientific problems that share the same mathematical structure.
Featured Problem 12.2 (Greenhouse Gasses and atmospheric CO2 )

Climate change has been attributed partly to the accumulation of greenhouse
gasses (such as carbon dioxide and methane) in the atmosphere.
Here we consider a simplified illustrative model for the carbon cycle that
tracks the sources and sinks of CO2 in the atmosphere. Consider C (t ) as the
Figure 12.7: CO2 is produced by emissions
level of atmospheric carbon dioxide. Define the production rate of C02 due from burning fossil fuel and other human
to utilization of fossil fuel and other human activity to be EFF , and let the activities (orange arrow). The oceans and
rate of absorption of CO2 by the oceans be SOCEAN . We will also assume that plant biomass are both sinks that absorbs
CO2 (light green arrows).
living plants absorb CO2 at a rate proportional to their biomass and to the
CO2 level.
1. Explain the following differential equation for atmospheric CO2 :

Hint: CO2 is usually given in
dC units of “parts per million”,
= EFF − SOCEAN − γPC. (12.15) ppm (=10−6 ) ,1 ppm = 2.1 GtC.
dt
(1GtC= 1 gigaton carbon = 109 tons.)
2. Assuming that EFF , SOCEAN , γ, P are constants, find the steady state level Time is typically given in years, so rates
are “per year” (yr−1 ).
of CO2 in terms of these parameters. Approximate parameter values:
EFF ≈ 10 GtC yr−1 ,
3. Find C (t ), that is, predict the amount of CO2 over time, assuming that SOCEAN ≈ 3 GtC yr−1 ,
C (0) = C0 . P ≈ 560 Gt plant biomass,
γ ≈ 1.35 · 10−5 yr−1 Gt−1 .
4. Graph the function C (t ) for parameter values given in the problem,

assuming that C0 = 400ppm= 840 GtC.
5. How big an effect would be produced on the CO2 level in 50 years if 15%
of the plant biomass is removed to deforestation just prior to t = 0?
Note: Information for Problem 12.2 is adapted from [Le Quéré et al., 2016],
and may reflect many simplifications and approximations. In actual fact, most
“constants” in the problem are time-dependent, making the real problem of
predicting CO2 levels much more challenging.
12.3 Euler’s Method and numerical solutions
1. Explain the idea of a numerical solution to a differential equation and how

this compares with an exact or analytic solution.
2. Describe how Euler’s method is based on approximating the derivative by

the slope of a secant line.
3. Use Euler’s method to calculate a numerical solution (using a spreadsheet)

to a given initial value problem.
So far, we have explored ways of understanding the behaviour predicted

by a differential equation in the form of an analytic solution, namely an
explicit formula for the solution as a function of time. However, in reality this
is typically difficult without extensive training, and occasionally, impossible
even for experts. Even if we can find such a solution, it may be inconve-
nient to determine its numerical values at arbitrary times, or to interpret its
behaviour.
For this reason, we sometimes need a method for computing an approx-
imation for the desired solution. We refer to that approximation as a nu-
merical solution. The idea is to harness a computational device - computer,
laptop, or calculator - to find numerical values of points along the solution
curve, rather than attempting to determine the formula for the solution as a
function of time. We illustrate this process using a technique called Euler’s
method, which is based on an approximation of a derivative by the slope of a
secant line.
Below, we describe how Euler’s method is used to approximate the
solution to a general initial value problem (differential equation together with
initial condition) of the form
dy
= f (y), y(0) = y0 .
dt
Set up. We first must pick a “step size,” ∆t, and subdivide the t axis into
discrete steps of that size. We thus have a set of time points t1 ,t2 , . . . , spaced
∆t apart as shown in Figure 12.8. Our procedure starts with the known initial
value y(0) = y0 , and uses it to generate an approximate value at the next time
point (y1 ), then the next (y2 ), and so on. We denote by yk the value of the
independent variable generated at the k’th time step by Euler’s method as an
approximation to the (unknown) true solution y(tk ).
Figure 12.8: The time axis is subdivided into

steps of size ∆t.
∆t
time
t0 t1 t2 t3 t4 t5 Mastered Material Check
18. If ∆t = 0.1 and t0 = 0, what are t1 ,t2
and t3 ?
Method. We approximate the differential equation by a finite difference 19. Explain the difference between the
value y1 and the true solution y(t1 ).
equation
20. If ∆t is not sufficiently small, why
dy yk+1 − yk might Euler’s method give a bad
= f (y) approximated by = f (yk ). approximation to the solution?
dt ∆t
This approximation is reasonable only when ∆t, the time step size, is small.
Rearranging this equation leads to a process (also called recurrence relation)
for linking values of the solution at successive time points,
yk+1 − yk
= f (yk ), ⇒ yk+1 = yk + ∆t · f (yk ). (12.16)
∆t
Application. We start with the known initial value, y0 . Then (setting the
index to k = 0 in Eqn. (12.16)) we obtain
y1 = y0 + f (y0 )∆t.
The quantities on the right are known, so we can compute the value of y1 ,
which is the approximation to the solution y(t1 ) at the time point t1 . We can
then continue to generate the value at the next time point in the same way, by
approximating the derivative again as a secant slope. This leads to
y2 = y1 + f (y1 )∆t.
The approximation so generated, leading to values y1 , y2 , . . . is called Euler’s

method. Mastered Material Check
Applying this approximation repeatedly, leads to an iteration method, 21. In Euler’s method, can you
that is, the repeated computation determine t2 directly? That is,
without first computing t1 ?
y1 = y0 + f (y0 )∆t, 22. In Euler’s method, can you

determine y2 directly? That is,
y2 = y1 + f (y1 )∆t, without first computing y1 ?
..
.
yk+1 = yk + f (yk )∆t.
From this iteration, we obtain the approximate values of the function yk ≈ y(tk )
for as many time steps as desired starting from t = 0 in increments of ∆t up to
some final time T of interest.
It is customary to use the following notations:
• t0 : the initial time point, usually at t = 0.
• h = ∆t : common notations for the step size, i.e. the distance between the
points along the t axis.
• tk : the k’th time point. Note that since the points are at multiples of the step
size that we have picked, tk = k∆t = kh.
• y(t ) : the actual value of the solution to the differential equation at time t.
This is usually not known, but in the examples discussed in this chapter,
we can solve the differential equation exactly, so we have a formula for the
function y(t ). In most hard scientific problems, no such formula is known
in advance.
• y(tk ) : the actual value of the solution to the differential equation at one of
the discrete time points, tk (again, not usually known).
• yk : the approximate value of the solution obtained by Euler’s method. We

hope that this approximate value is fairly close to the true value, i.e. that
yk ≈ y(tk ), but there is always some error in the approximation. More
advanced methods that are specifically designed to reduce such errors are
discussed in courses on numerical analysis.
Euler’s method applied to population growth

We illustrate how Euler’s method is used in a familiar example, that of
unlimited population growth.
Example 12.10 Apply Euler’s method to approximating solutions for the

simple exponential growth model that was studied in Chapter 11,
dy
= ay, y ( 0 ) = y0
dt
where a is a constant (see Eqn 11.2).
Solution. Subdivide the t axis into steps of size ∆t, starting with t0 = 0, and 23. Carry our Example 12.10
t1 = ∆t,t2 = 2∆t, . . . The first value of y is known from the initial condition, with ∆t = 0.1, a = 1, and y0 = 1.
24. Plot the first 5 points you determine.
y0 = y(0) = y0 . Compare with the true solution.
25. Solve the initial value problem in
We replace the differential equation by the approximation Example 12.11 analytically.
Compare the points (0, 100),
yk+1 − yk (0.1, 95), (0.2, 90.25)
= ayk ⇒ yk+1 = yk + a∆tyk , k = 1, 2, . . . and (0.3, 85.7375) with the true
∆t solution at the corresponding t
values.
In particular,
y1 = y0 + a∆ty0 = y0 (1 + a∆t ),
y2 = y1 (1 + a∆t ),
y3 = y2 (1 + a∆t ),
and so on. At every stage, the quantity on the right hand side depends only on
value of yk that as already known from the step before. ♦
The next example demonstrates Euler’s method applied to a specific
differential equation.
Example 12.11 Use Euler’s method to find the solution to P Link to Google Sheets. This
spreadsheet implements Euler’s method
dy
= −0.5y, y(0) = 100. for Example 12.11. You can view the
dt formulae by clicking on a cell in the
Use step size ∆t = 0.1 to approximate the solution for the first two time steps. sheet but you cannot edit the sheet here.
k tk yk
Solution. Euler’s method applied to this example would lead to
0 0 100.00
1 0.1 95.00
2 0.2 90.25
y0 =100. 3 0.3 85.74
4 0.4 81.45
y1 =y0 (1 + a∆t ) = 100(1 + (−0.5)(0.1)) = 95, etc.
5 0.5 77.38
We show the first five values in Table 12.3. Clearly, these kinds of repeated
Table 12.3: Euler’s method applied to
calculations are best handled on a spreadsheet or similar computer software. Example 12.11.
Euler’s method applied to Newton’s law of cooling

We apply Euler’s method to Newton’s law of cooling. Upon completion,
we can directly compare the approximate numerical solution generated by
Euler’s method to the true (analytic) solution, (12.8), that we determined
earlier in this chapter.
Example 12.12 (Newton’s law of cooling) Consider the temperature of an

object T (t ) in an ambient temperature of E = 10◦ . Assume that k = 0.2/min.
Use the initial value problem
dT
= k(E − T ), T (0) = T0
dt
to write the exact solution to Eqn. (12.8) in terms of the initial value T0 .
Solution. In this case, the differential equation has the form

dT
= 0.2(10 − T ),
dt
and its analytic solution, from Eqn. (12.8), is
T (t ) = 10 + (T0 − 10)e−0.2t . (12.17)
♦
Below, we use Euler’s method to compute a solution from each of several
initial conditions, T (0) = 0, 5, 15, 20 degrees.
Example 12.13 (Euler’s method applied to Newton’s law of cooling)

Write the Euler’s method procedure for the approximate solution to the
problem in Example 12.12.
Solution. Euler’s method approximates the differential equation by

Tk+1 − Tk
= 0.2(10 − Tk ).
∆t
or, in simplified form,
Tk+1 = Tk + 0.2(10 − Tk )∆t.
y
20
Euler’s method
time approx solution exact soln
tk Tk T (t)
15 True solution 0.0000 0.0000 0.0000
1.0000 2.0000 1.8127
2.0000 3.6000 3.2968
10 3.0000 4.8800 4.5119
4.0000 5.9040 5.5067
5.0000 6.7232 6.3212
5 6.0000 7.3786 6.9881
7.0000 7.9028 7.5340
8.0000 8.3223 7.9810
∆t = 1.0
2 4 6 8 10
time, t
Figure 12.9: Euler’s method applied to

Newton’s law of cooling. The graph shows
the true solution (red) and the approximate
Example 12.14 Use Euler’s method from Example 12.13 and time steps of solution (black).
size ∆t = 1.0 to find a numerical solution to the the cooling problem. Use a
spreadsheet for the calculations. Note that ∆t = 1.0 is not a “small step;” we
use it here for illustration purposes.
Solution. The procedure to implement is
Tk+1 = Tk + 0.2(10 − Tk )∆t.

In Figure 12.9 we show a typical example of the method with initial value 26. What change would you make in the
process set up in Example 12.14 to
T (0) = T0 and with the time step size ∆t = 1.0. Black dots represent the dis- improve the approximation made by
crete values generated by the Euler method, starting from initial conditions, Euler’s method?
T0 = 0, 5, 15, 20. Notice that the black curve is simply made up of line seg-
ments linking points obtained by the numerical solution. On the same graph,
we also show the analytic solution (red curves) given by Eqn. (12.17) with the
same four initial temperatures. We see that the black and red curves start out
at the same points (since they both satisfy the same initial conditions). How-
ever, the approximate solution obtained with Euler’s method is not identical
to the true solution. The difference between the two (gap between the red and
black curves) is the numerical error in the approximation.
12.4 Summary
1. Given a function, we can check whether it is a solution to a differential

equation by performing the appropriate differentiation and algebraic
simplification.
2. Solutions to differential equations in which there is no change at all
(“constant solutions”) are referred to as steady states.
3. The differential equations
dy
= a − by, y(0) = y0
dt
has a steady state solution y = a/b.
4. If we define the deviation from steady state, z(t ) = y(t ) − ab , we get a decay
equation for z(t ) that has exponentially decreasing solutions provided b > 0.
This says that the eviation from steady state always decrease over time.
5. The resulting solution for y(t ) is
a a
y(t ) = − − y0 e−bt .
b b
6. For some differential equations, it is not always possible to determine an
analytic solution (explicit formula). Numerical solutions can be found
using Euler’s method, and serve as an approximate solution.
7. Euler’s method takes a known initial value y0 and uses the iteration
scheme:
yk+1 = yk + f (yk )∆t.
to generate successive values of yk that approximate the solution at time

points tk = k∆t
8. Applications considered in this chapter included:
(a) height of water draining out of a cylindrical container (verifying a
solution to a differential equation);
(b) Newton’s law of cooling (described by a linear differential equation);
(c) growth of the radius of a cell;
(d) the accumulation of greenhouse gasses in the atmosphere;
(e) friction and terminal velocity; and
(f) chemical production and decay.
1. Explain why an object at room temperature is at a steady state for Newton’s law of cooling.
2. The following graph depicts solution curves to a particular differential equation of the form dy/dt = a − by.
30
20
y
10
t
2 4 6 8 10 12 14
(a) Estimate the value that these solution curves are approaching.
(b) Which solutions are approaching from above? From below?
3. Consider the following initial value problem:
dy
= 2 − 4y, y(0) = 4,
dt
(a) What value does its solution curve approach?
(b) Does its solution approach from above or below?
4. Why is a large value of ∆t not a good idea when using Euler’s method?
Exercises
12.1. Water draining form a container. In Example 12.3, we verified

√ 2
that the function h(t ) = h0 − k 2t is a solution to the differential
equation (12.2). Based on the meaning of the problem, for how long
does this solution remain valid?
12.2. Verifying a solution. Verify that the function y(t ) = 1 − (1 − y0 )e−t
satisfies the initial value problem (differential equation and initial
condition) (12.6).
12.3. Linear differential equation. Consider the differential equation
dy
= a − by
dt
where a, b are constants.
(a) Show that the function
a
y(t ) = −Ce−bt
b
satisfies the above differential equation for any constant C.
(b) Show that by setting
a
C= − y0
b
we also satisfy the initial condition
y(0) = y0 .
Remark: you have shown that the function

a −bt a
y(t ) = y0 − e +
b b
is a solution to the initial value problem (i.e differential equation
plus initial condition)
dy
= a − by, y(0) = y0 .
dt
12.4. Complete the steps. Complete the algebraic steps to show that the
solution to Eqn. (12.4) can be obtained by the substitution z(t ) =
a − by(t ).
12.5. Verifying a solution. Show that the function
1
y(t ) =
1−t
is a solution to the differential equation and initial condition
dy
= y2 , y(0) = 1.
dt
Comment on what happens to this solution as t approaches 1.
12.6. Verifying solutions. For each of the following, show the given func-
tion y is a solution to the given differential equation.
dy
(a) t · = 3y, y = 2t 3 .
dt
d2y
(b) + y = 0, y = −2 sint + 3 cost.
dt 2
d2y dy
(c) 2
− 2 + y = 6et , y = 3t 2 et .
dt dt
12.7. Verifying a solution. Show the function determined by the equa-
tion 2x2 + xy − y2 = C, where C is a constant and 2y 6= x, is a solution
dy
to the differential equation (x − 2y) = −4x − y.
dx
12.8. Determining the constant. Find the constant C that satisfies the given
initial conditions.
(a) 2x2 − 3y2 = C, y|x=0 = 2.
dy
(b) y = C1 e5t + C2te5t , y|t =0 = 1 and dt |t =0 = 0.
(c) y = C1 cos(t −C2 ), y|t = π = 0 and dy
dt |t = 2 = 1.
π
2
12.9. Friction and terminal velocity. The velocity of a falling object

changes due to the acceleration of gravity, but friction has an effect of
slowing down this acceleration. The differential equation satisfied by
the velocity v(t ) of the falling object is
dv
= g − kv
dt
where g is acceleration due to gravity and k is a constant that rep-
resents the effect of friction. An object is dropped from rest from a
plane.
(a) Find the function v(t ) that represents its velocity over time.
(b) What happens to the velocity after the object has been falling for a
long time (but before it has hit the ground)?
12.10. Alcohol level. Alcohol enters the blood stream at a constant rate k gm
per unit time during a drinking session. The liver gradually converts
the alcohol to other, non-toxic byproducts. The rate of conversion per
unit time is proportional to the current blood alcohol level, so that the
differential equation satisfied by the blood alcohol level is
dc
= k − sc
dt
where k, s are positive constants. Suppose initially there is no alcohol
in the blood.
Find the blood alcohol level c(t ) as a function of time from t = 0, when
the drinking started.
12.11. Checking a solution. Check that the differential equation (12.7) has
the right sign, so that a hot object cools off in a colder environment.
12.12. Details of Newtons Law of Cooling. Fill in the missing steps in the
solution to Newton’s Law of Cooling in Example 12.5.
12.13. Newton’s Law of Cooling. Newton’s Law of Cooling states that the
rate of change of the temperature of an object is proportional to the
difference between the temperature of the object, T , and the ambient
(environmental) temperature, E. This leads to the differential equation
dT
= k (E − T )
dt
where k > 0 is a constant that represents the material properties and, E
is the ambient temperature. (We assume that E is also constant.)
T (t ) = E + (T0 − E )e−kt
which represents the temperature at time t satisfies this equation.

(b) The time of death of a murder victim can be estimated from the
temperature of the body if it is discovered early enough after the
crime has occurred.
Suppose that in a room whose ambient temperature is E = 20◦ C,
the temperature of the body upon discovery is T = 30◦ C, and that a
second measurement, one hour later is T = 25◦ C.
Determine the approximate time of death.
Remark: use the fact that just prior to death, the temperature of the
victim was 37◦ C.
12.14. A cup of coffee. The temperature of a cup of coffee is initially 100 de-
grees C. Five minutes later, (t = 5) it is 50 degrees C. If the ambient
temperature is A = 20 degrees C, determine how long it takes for the
temperature of the coffee to reach 30 degrees C.
12.15. Newton’s Law of Cooling applied to data. The data presented in
Table 12.4 was gathered in producing Figure 2.2 for cooling milk
during yoghurt production. According to Newton’s Law of Cooling,
this data can be described by the formula time (min) Temp
0.0 190.0
T = E + (T (0) − E ) e−kt . 0.5 185.5
1.0 182.0
where T (t ) is the temperature of the milk (in degrees Fahrenheit) at 1.5 179.2
2.0 176.0
time t (in min), E is the ambient temperature, and k is some constant 2.5 172.9
that we determine in this exercise. 3.0 169.5
3.5 167.0
(a) Rewrite this relationship in terms of the quantity Y (t ) = ln(T (t ) − 4.0 164.6
4.5 162.2
E ), and show that Y (t ) is related linearly to the time t.
5.0 159.8
(b) Explain how the constant k could be found from this converted
Table 12.4: Cooling milk data for Exer-
form of the relationship. cise 15.
(c) Use the data in the table and your favourite spreadsheet (or similar
software) to show that the data so transformed appears to be close
to linear. Assume that the ambient temperature was E = 20◦ F.
(d) Use the same software to determine the constant k by fitting a line
to the transformed data.
12.16. Infant weight gain. During the first year of its life, the weight of a
baby is given by
√
y(t ) = 3t + 64
where t is measured in some convenient unit.

(a) Show that y satisfies the differential equation
dy k
=
dt y
where k is some positive constant.

(b) What is the value for k?
(c) Suppose we adopt this differential equation as a model for human
growth. State concisely (that is, in one sentence) one feature about
this differential equation which makes it a reasonable model. State
one feature which makes it unreasonable.
12.17. Lake Fishing. Fish Unlimited is a company that manages the fish
population in a private lake. They restock the lake at constant rate (to
restock means to add fish to the lake): N fishers are allowed to fish in
the lake per day. The population of fish in the lake, F (t ) is found to
satisfy the differential equation
dF
= I − αNF (12.18)
dt
(a) At what rate are fish added per day according to Eqn. (12.18)? Give
both value and units.
(b) What is the average number of fish caught by one fisher? Give both
the value and units.
(c) What is being assumed about the fish birth and mortality rates
in Eqn. (12.18)?
(d) If the fish input and number of fishers are constant, what is the
steady state level of the fish population in the lake?
(e) At time t = 0 the company stops restocking the lake with fish. Give
the revised form of the differential equation (12.18) that takes this
into account, assuming the same level of fishing as before. How
long would it take for the fish to fall to 25% of their initial level?
(f) When the fish population drops to the level Flow , fishing is stopped
and the lake is restocked with fish at the same constant rate
(Eqn (12.18), with α = 0.) Write down the revised version of

Eqn. (12.18) that takes this into account. How long would it take
for the fish population to double?
12.18. Tissue culture. Cells in a tissue culture produce a cytokine (a chem-

ical that controls the growth of other cells) at a constant rate of 10
nano-Moles per hour (nM/h). The chemical has a half-life of 20 hours.
Give a differential equation (DE) that describes this chemical pro-

duction and decay. Solve this DE assuming that at t = 0 there is no
cytokine. [1nM=10−9 M].
12.19. Glucose solution in a tank. A tank that holds 1 liter is initially
full of plain water. A concentrated solution of glucose, contain-
ing 0.25 gm/cm3 is pumped into the tank continuously, at the rate 10 cm3 /min
and the mixture (which is continuously stirred to keep it uniform) is
pumped out at the same rate.
How much glucose is in the tank after 30 minutes? After a long time?
(hint: write a differential equation for c, the concentration of glucose
in the tank by considering the rate at which glucose enters and the rate
at which glucose leaves the tank.)
12.20. Pollutant in a lake. A lake of constant volume V gallons con-
tains Q(t ) pounds of pollutant at time t evenly distributed throughout
the lake. Water containing a concentration of k pounds per gallon
of pollutant enters the lake at a rate of r gallons per minute, and the
well-mixed solution leaves at the same rate.
(a) Set up a differential equation that describes the way that the amount
of pollutant in the lake changes.
(b) Determine what happens to the pollutant level after a long time if
this process continues.
(c) If k = 0 find the time T for the amount of pollutant to be reduced to
one half of its initial value.
12.21. A sugar solution. Sugar dissolves in water at a rate proportional to the
amount of sugar not yet in solution. Let Q(t ) be the amount of sugar
undissolved at time t. The initial amount is 100 kg and after 4 hours
the amount undissolved is 70 kg.
(a) Find a differential equation for Q(t ) and solve it.
(b) How long does it take for 50 kg to dissolve?
12.22. Leaking water tank. A cylindrical tank with cross-sectional area A
has a small hole through which water drains. The height of the water
in the tank y(t ) at time t is given by:
√ kt 2
y(t ) = ( y0 − )
2A
where k, y0 are constants.

(a) Show that the height of the water, y(t ), satisfies the differential
equation
dy k√
=− y.
dt A
(b) What is the initial height of the water in the tank at time t = 0 ?
(c) At what time is the tank be empty ?
(d) At what rate is the volume of the water in the tank changing
when t = 0?
12.23. Determining constants. Find those constants a, b so that y = ex and
y = e−x are both solutions of the differential equation
y00 + ay0 + by = 0.
12.24. Euler’s method. Solve the decay equation in Example (12.11) analyt-
ically, that is, find the formula for the solution in terms of a decaying
exponential, and then compare your values to the approximate solution
values y1 and, y2 computed with Euler’s method.
12.25. Comparing approximate and true solutions:
(a) Use Euler’s method to find an approximate solution to the differen-
tial equation
dy
=y
dx
with y(0) = 1. Use a step size h = 0.1 and find the values of y up
to x = 0.5. Compare the value you have calculated for y(0.5) using
Euler’s method with the true solution of this differential equation.
What is the error i.e. the difference between the true solution and
the approximation?
(b) Now use Euler’s method on the differential equation
dy
= −y
dx
with y(0) = 1. Use a step size h = 0.1 again and find the values of y
up to x = 0.5. Compare the value you have calculated for y(0.5)
using Euler’s method with the true solution of this differential
equation. What is the error this time?
12.26. Beginning Euler’s method. Give the first 3 steps of Euler’s method
for the problem in Example 12.13.
12.27. Euler’s method and a spreadsheet. Use the spreadsheet and Euler’s
method to solve the differential equation shown below:
dy/dt = 0.5y(2 − y)
Use a step size of h = 0.1 and show (on the same graph) solutions for
the following four initial values:
y(0) = 0.5, y(0) = 1, y(0) = 1.5, y(0) = 2.25
For full credit, include a short explanation your process (e.g. 1-2 sen-
tences and whatever equations you implemented on the spreadsheet.)
13
Qualitative methods for differential equations

Not all differential equations are easily solved analytically. Furthermore,
1. What is meant by an analytic
even when we find the analytic solution, it is not necessarily easy to interpret, solution to a differential equation?
graph, or understand. This situation motivates qualitative methods that 2. What other kind of solutions are
promote an overall understanding of behaviour - directly from information in possible?
the differential equation - without the challenge of finding a full functional 3. Give an example of a nonlinear
function f (y).
form of the solution.
In this chapter we expand our familiarity with differential equations and
assemble new, qualitative techniques for understanding them. We consider
differential equations in which the expression on one side, f (y), is nonlinear,
i.e. equations of the form
dy
= f (y)
dt
in which f is more complicated than the form a − by. Geometric techniques,
rather than algebraic calculations form the core of the concepts we discuss.
13.1 Linear and nonlinear differential equations
1. Identify the distinction between unlimited and density-dependent popula-

tion growth. Be able to explain terms in the logistic equation in its original
version, Eqn. (13.1), and its rescaled version, Eqn. (13.3).
2. State the definition of a linear differential equation.
3. Explain the law of mass action, and derive simple differential equations
for interacting species based on this law.
In the model for population growth in Chapter 11, we encountered the

differential equation
dN
= kN,
dt
where N (t ) is population size at time t and k is a constant per capita growth

rate. We showed that this differential equation has exponential solutions. It
means that two behaviours are generically obtained: explosive growth if k > 0
or extinction if k < 0. Mastered Material Check
The case of k > 0 is unrealistic, since real populations cannot keep growing 4. What happens in the case that k = 0?
indefinitely in an explosive, exponential way. Eventually running out of space Explain under what conditions this
might arise and what happens to the
or resources, the population growth dwindles, and the population attains some population N (t ) in this case.
static level rather than expanding forever. This motivates a revision of our
previous model to depict density-dependent growth.
The logistic equation for population growth

Let N (t ) represent the size of a population at time t, as before. Consider the
dN (K − N )
= rN . (13.1)
dt K
We call this differential equation the logistic equation.The logistic equation
has a long history in modelling population growth of humans, microorgan-
isms, and animals. Here the parameter r is the intrinsic growth rate and K is
the carrying capacity. Both r, K are assumed to be positive constants for a
given population in a given environment.
In the form written above, we could interpret the logistic equation as
(K − N )

dN
= R(N ) · N, where R(N ) = r .
dt K
The term R(N ) is a function of N that replaces the constant rate of growth k
(found in the unrealistic, unlimited population growth model). R is called the
density dependent growth rate.
Linear versus nonlinear

The logistic equation introduces the first example of a nonlinear differ-
ential equation. We explain the distinction between linear and nonlinear
differential equations and why it matters.
Definition 13.1 (Linear differential equation) A first order differential

equation is said to be linear if it is a linear combination of terms of the form
dy
, y, 1
dt Mastered Material Check
that is, it can be written in the form 5. Can the differential equation
dy
dt = a − by be written in the form
dy (13.2)? If so, what are the values of
α +βy+γ = 0 (13.2) α, β , γ?
dt
where α, β , γ do not depend on y. Note that “first order” means that only the
first derivative (or no derivative at all) may occur in the equation.
Q UA L I TAT I V E M E T H O D S F O R D I F F E R E N T I A L E Q UAT I O N S 277
So far, we have seen several examples of this type with constant coeffi-
cients α, β , γ. For example, α = 1, β = −k, and γ = 0 in Eqn. 11.2 whereas
α = 1, γ = −a, and β = b in Eqn. (12.4). A differential equation that is not of
this form is said to be nonlinear.
Example 13.1 (Linear versus nonlinear differential equations) Which of

the following differential equations are linear and which are nonlinear?
dy dy dy
(a) = y2 , (b) − y = 5, (c) y = −1.
dt dt dt
√
Solution. Any term of the form y2 , y, 1/y, etc. is nonlinear in y. A product
such as y dy
dt is also nonlinear in the independent variable. Hence equations (a),
(c) are nonlinear, while (b) is linear. ♦
The significance of the distinction between linear and nonlinear differ- 6. For what values of α, β and γ can
ential equations is that nonlinearities make it much harder to systematically Example 13.1(b) be put into the
form (13.2)?
find a solution to the given differential equation by “analytic” methods. Most
linear differential equations have solutions that are made of exponential func-
tions or expressions involving such functions. This is not true for nonlinear
equations.
However, as we see shortly, geometric methods are very helpful in under-
standing the behaviour of such nonlinear differential equations.
Law of Mass Action

Nonlinear terms in differential equations arise naturally in various ways. One
common source comes from describing interactions between individuals, as
the following example illustrates.
In a chemical reaction, molecules of types A and B bind and react to
form product P. Let a(t ), b(t ) denote the concentrations of A and B. These
concentrations depend on time because the chemical reaction uses up both
types in producing the product.
The reaction only occurs when A and B molecules “collide” and stick to
one another. Collisions occur randomly, but if concentrations are larger, more
collisions take place, and the reaction is faster. If either the concentration a or
b is doubled, then the reaction rate doubles. But if both a and b are doubled,
then the reaction rate should be four times faster, based on the higher chances
of collisions between A and B. The simplest assumption that captures this Mastered Material Check
dependence is 7. If the concentration of A is tripled,
and that of B is doubled, how much
faster would we expect the reaction
rate of reaction is proportional to a · b ⇒ rate of reaction = k · a · b rate to be?
8. Why does the product a · b, rather
where k is some constant that represents the reactivity of the molecules. than the sum a + b appear in the Law
We can formally state this result, known as the Law of Mass Action as of Mass Action ?
follows:
The Law of Mass Action: The rate of a chemical reaction involving an

interaction of two or more chemical species is proportional to the product of
the concentrations of the given species.
Example 13.2 (Differential equation for interacting chemicals) Substance A

is added at a constant rate of I moles per hour to a 1-litre vessel. Pairs of
molecules of A interact chemically to form a product P. Write down a differ-
ential equation that keeps track of the concentration of A, denoted y(t ).
Solution. First consider the case that there is no reaction. Then, the addition 9. In each of Examples 13.2 and 13.3,
of A to the reactor at a constant rate leads to changing y(t ), described by the clearly identify the constant
quantities.
dy
= I.
dt
When the chemical reaction takes place, the depletion of A depends on
interactions of pairs of molecules. By the law of mass action, the rate of
reaction is of the form k · y · y = ky2 , and as it reduces the concentration, it
appear with a minus sign in the DE. Hence
dy
= I − ky2 .
dt
This is a nonlinear differential equation - it contains a term of the form y2 . ♦
Example 13.3 (Logistic equation reinterpreted) Rewrite the logistic

equation in the form
dN
= rN − bN 2
dt
(where b = r/K is a positive quantity).
a) Interpret the meaning of this restated form of the equation by explaining

what each of the terms on the right hand side could represent.
b) Which of the two terms dominates for small versus large population
levels?
Solution.
a) This form of the equation has growth term rN proportional to population

size, as encountered previously in unlimited population growth. However,
there is also a quadratic (nonlinear) rate of loss (note the minus sign)
−bN 2 . This term could describe interactions between individuals that lead
to mortality, e.g. through fighting or competition.
b) From familiarity with power functions (in this case, the functions of N
that form the two terms, rN and bN 2 ) we can deduce that the second,
quadratic term dominates for larger values of N, and this means that when
the population is crowded, the loss of individuals is greater than the rate of
reproduction. ♦
Scaling the logistic equation

Consider units involved in the logistic equation (13.1):
dN (K − N )
= rN .
dt K
This equation has two parameters, r and K. Since units on each side of
an equation must balance, and must be the same for terms that are added
or subtracted, we can infer that K has the same units as N, and thus it is
a population density. When N = K, the population growth rate is zero
(dN/dt = 0).
It turns out that we can understand the behaviour of the logistic equation
by converting it to a “generic” form that does not depend on the constant K.
We do so by transforming variables, which amounts to choosing a convenient
way to measure the population size.
Example 13.4 (Rescaling) Define a new variable

N (t )
y(t ) = ,
K
with N (t ) and K as in the logistic equation. Then N (t ) = Ky(t ).
a) Interpret what the transformed variable y represents.
b) Rewrite the logistic equation in terms of this variable.

10. Suppose an environment can
Solution. sustain 2000 aphids per plant, and
the current population size on a
a) The variable, y(t ) represents a scaled version of the population density. In- given plant is 1700. What is K, N
stead of measuring the population in some arbitrary units - such as number and y based on this information?
of individuals per acre, or number of bacteria per ml - y(t ) measures the 11. This population is at what percent of
its carrying capacity?
population in “multiples of the carrying capacity.”
For example, if the environment can sustain 1000 aphids per plant (so
K = 1000 individuals per plant), and the current population size on a given
plant is N = 950 then the value of the scaled variable is y = 950/1000 =
0.95. We would say that “the aphid population is at 95% of its carrying
capacity on the plant.”
b) Since K is assumed constant, it follows that

dN dy
N (t ) = Ky(t ) ⇒ =K .
dt dt
Using this, we can simplify the logistic equation as follows:
dN (K − N ) dy (K − Ky)
= rN , ⇒ K = r (Ky) ,
dt K dt K (13.3)
dy
⇒ = ry(1 − y).
dt
♦
Eqn. (13.3) “looks simpler” than Eqn. (13.1) since it depends on only one Mastered Material Check
parameter, r. Moreover, by understanding this equation, and transforming 12. What are the units of the parameter
r?
back to the original logistic in terms of N (t ) = Ky(t ), we can interpret
13. How might we use the parameter r
results for the original model. While we do not go further with transforming to define a time-scale?
variables at present, it turns out that one can also further reduce the scaled
logistic to an equation in which r = 1 by “rescaling time units”.
13.2 The geometry of change
1. Explain what is a slope field of a differential equation. Given a differential

equation (linear or nonlinear), construct such a diagram and use it to
sketch solution curves.
2. Describe what a state-space diagram is; construct such a diagram and

use it to interpret the behaviour of solution curves to a given differential
equation.
3. Identify the relationships between a slope field, a state-space diagram, and

a family of solution curves to a given differential equation.
4. Identify steady states of a differential equation and determine whether they

are stable or unstable.
5. Given a differential equation and initial condition, predict the behaviour of

the solution for t > 0.
In this section, we introduce a new method for understanding differen-

tial equations using graphical and geometric arguments. Such methods
circumvent the solutions that we expressed in terms of analytic formulae.
We resort to concepts learned much earlier - for example, the derivative as
a slope of a tangent line - in order to use the differential equation itself to
assemble a sketch of the behaviour that it predicts. That is, rather than writing
down y = F (t ) as a solution to the differential equation (and then graphing that
function) we sketch the qualitative behaviour of such solution curves directly
from information contained in the differential equation.
Slope fields
Here we discuss a geometric way of understanding what a differential equa-
tion is saying using a slope field, also called a direction field. We have
already seen that solutions to a differential equation of the form
dy
= f (y)
dt
are curves in the (y,t )-plane that describe how y(t ) changes over time (thus,
these curves are graphs of functions of time). Each initial condition y(0) = y0
is associated with one of these curves, so that together, these curves form a
family of solutions.
What do these curves have in common, geometrically?
• the slope of the tangent line (dy/dt) at any point on any of the curves
is related to the value of the y-coordinate of that point - as stated in the
differential equation.
• at any point (t, y(t )) on a solution curve, the tangent line must have
slope f (y), which depends only on the y value, and not on the time t.
Note: in more general cases, the expression f (y) that appears in the differ-
ential equation might depend on t as well as y. For our purposes, we do not
consider such examples in detail.
By sketching slopes at various values of y, we obtain the slope field
through which we can get a reasonable idea of the behaviour of the solutions
to the differential equation.
Example 13.5 Consider the differential equation

dy
= 2y. (13.4)
dt
Compute some of the slopes for various values of y and use this to sketch a
slope field for this differential equation.
Solution. Equation (13.4) states that if a solution curve passes through a 14. Solve Differential Eqn. (13.4)
point (t, y), then its tangent line at that point has a slope 2y, regardless of the analytically.
value of t. This example is simple enough that we can state the following: for
positive values of y, the slope is positive; for negative values of y, the slope is
negative; and for y = 0, the slope is zero.
We provide some tabulated values of y indicating the values of the slope
f (y), its sign, and what this implies about the local behaviour of the solution
and its direction. Then, in Figure 13.1 we combine this information to
y f (y ) slope of behaviour of y direction of Table 13.1: Table for the slope field diagram
of differential equation (13.4), dy
dt = 2y,
tangent line arrow described in Example 13.5.
-2 -4 -ve decreasing &
-1 -2 -ve decreasing &
0 0 0 no change →
1 2 +ve increasing %
2 4 +ve increasing %
generate the direction field and the corresponding solution curves. Note that
the direction of the arrows (rather than their absolute magnitude) provides the
most important qualitative tendency for the slope field sketch. ♦
y
2 2 y Figure 13.1: Direction field and solution
curves for differential equation, dy
dt = 2y
1 1
0 0
−1 −1
−2 t −2 t
0 1 2 3 4 5 0 1 2 3 4 5
In constructing the slope field and solution curves, the following basic
rules should be followed:
1. By convention, time flows from left to right along the t axis in our graphs,
so the direction of all arrows (not usually indicated explicitly on the slope
field) is always from left to right.
2. According to the differential equation, for any given value of the vari-
able y, the slope is given by the expression f (y) in the differential equa-
tion. The sign of that quantity is particularly important in determining
whether the solution is locally increasing, decreasing, or neither. In the
tables, we indicate this in the last column with the notation %, &, or →.
3. There is a single arrow at any point in the ty-plane, and consequently so-
lution curves cannot intersect anywhere (although they can get arbitrarily
close to one another).
We see some implications of these rules in our examples.
Example 13.6 Consider the differential equation

dy
= f (y) = y − y3 . (13.5)
dt
Create a slope field diagram for this differential equation.
i A summary of steps in creating the
Solution. Based on the last example, we focus on the sign, rather than the slope field for Example 13.6.
value of the derivative f (y), since that sign determines whether the solutions
increase, decrease, or stay constant. Recall that factoring helps to find zeros,
and to identify where an expression changes sign. For example,
dy
= f (y) = y − y3 = y(1 − y2 ) = y(1 + y)(1 − y).
dt
The sign of f depends on the signs of the factors y, (1 + y), (1 − y). For Mastered Material Check
y < −1, two factors, y, (1 + y), are negative, whereas (1 − y) is positive, so 15. Graph the
that the product is positive overall. The sign of f (y) changes at each of the function f (y) = y(1 + y)(1 − y) and
indiate where it changes sign.
three points y = 0, ±1 where one or another of the three factors changes sign,
16. Repeat the process for the
function f (y) = y2 (1 + y)2 (1 − y).
as shown in Table 13.2. Eventually, to the right of all three (when y > 1), the
sign is negative. We summarize these observations in Table 13.2 and show
the slopes field and solution curves in Figure 13.2. ♦
y sign of f (y) behaviour of y direction of Table 13.2: Table for the slope field diagram
of the DE (13.5) described in Example 13.6.
arrow
y < −1 +ve increasing %
-1 0 no change →
-0.5 -ve decreasing &
0 0 no change →
0.5 +ve increasing %
1 0 no change →
y>1 -ve decreasing &
\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
1 1
Figure 13.2: Direction field and solution
/ / / / / / / / / / / / / / / / / / / / curves for differential equation (13.5)
0.5 / / / / / / / / / / 0.5 / / / / / / / / / /

0 0

−0.5 \ \ \ \ \ \ \ \ \ \ −0.5 \ \ \ \ \ \ \ \ \ \
\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
−1 −1
/ / / / / / / / / / / / / / / / / / / /
0 5 10 15 20 0 5 10 15 20
Example 13.7 Sketch a slope field and solution curves for the problem of a
cooling object, and specifically for
dT
= f (T ) = 0.2(10 − T ). (13.6)
dt
Solution. The family of curves shown in Figure 13.3 (also Figure 12.6) are
solutions to (13.6). The function f (T ) = 0.2(10 − T ) corresponds to the
slopes of tangent lines to these curves. We indicate the sign of f (T ) and
thereby the behaviour of T (t ) in Table 13.3. Note that there is only one
T sign of f (T ) behaviour of T direction of Table 13.3: Table for the slope field dia-
gram of dT
dt = 0.2(10 − T ) described in
arrow Example 13.7.
T < 10 +ve increasing %
T = 10 0 no change →
T > 10 -ve decreasing &
change of sign, at T = 10. For smaller T , the solution is always increasing and
for larger T , the solution is always decreasing. The slope field and solution
17. Indicate the regions Figure 13.3
curves are shown in Figure 13.3. In the slope field, one particular value of t is where T is increasing.
coloured to emphasize the associated changes in T , as in Table 13.3. ♦ 18. Where is T not changing in
Figure 13.3?
20 20 Figure 13.3: Slope field and solution

curves for a cooling object that satisfies the
temperature, T
temperature, T
15 15 differential equation (13.6) in Example 13.7.
10 10
5 5
0
t 0
t
0 5 10 15 20 0 5 10 15 20
We observe an agreement between the detailed solutions found analyt-

ically (Example 12.5), found using Euler’s method (Example 12.13), and
those sketched using the new qualitative arguments (Example 13.7).
State-space diagrams
In Examples 13.5-13.7, we saw that we can understand qualitative features of
solutions to the differential equation
dy
= f (y), (13.7)
dt
by examining the expression f (y). We used the sign of f (y) to assemble a
slope field diagram and sketch solution curves. The slope field informed us
about which initial values of y would increase, decrease or stay constant. We
next show another way of determining the same information.
First, let us define a state space, also called phase line, which is essen-
tially the y-axis with superimposed arrows representing the direction of flow.
Definition 13.2 (State space (or phase line)) A line representing the depen-
dent variable (y) together with arrows to describe the flow along that line
(increasing, decreasing, or stationary y) satisfying Eqn. (13.7) is called the
state space diagram or the phase line diagram for the differential equation.
Rather than tabulating signs for f (y), we can arrive at similar conclusions
by sketching f (y) and observing where this function is positive (implying
that y increases) or negative (y decreases). Places where f (y) = 0 (“zeros
of f ”) are important since these represent steady states (“static solutions”,
where there is no change in y). Along the y axis (which is now on the hori-
zontal axis of the sketch) increasing y means motion to the right, decreasing y
means motion to the left.
As we shall see, the information contained in this type of diagram pro-
vides a qualitative description of solutions to the differential equation, but
with the explicit time behaviour suppressed. This is illustrated by Figure 13.4,
where we show the connection between the slope field diagram and the state
space diagram for a typical differential equation.
y
y
% f (y)
&
t y
%
&
(a) (b) (c)
Figure 13.4: The relationship of the slope

field and state space diagrams. (a) A typical
Example 13.8 Consider the differential equation slope field. A few arrows have been added
to indicate the direction of time flow along
dy the tangent vectors. Now consider “looking
= f (y) = y − y3 . (13.8)
dt down the time axis” as shown by the “eye”
in this diagram. Then the t axis points
Sketch f (y) versus y and use your sketch to determine where y is static, and towards us, and we see only the y-axis as
where y increases or decreases. Then describe what this predicts starting in (b). Arrows on the y-axis indicate the
directions of flow for various values of y
from each of the three initial conditions: as determined in (a). Now “rotate” the y
axis so it is horizontal, as shown in (c). The
(i) y(0) = −0.5, direction of the arrows exactly correspond to
places where f (y), in (c), is positive (which
(ii) y(0) = 0.3, or implies increasing y, →), or negative (which
implies decreasing y, ←). The state space
(iii) y(0) = 2. diagram is the y-axis in (b) or (c).
Solution. From Example 13.6, we know that f (y) = 0 at y = −1, 0, 1. This i Video explanation of the steps in the
solution to Example 13.8.
means that y does not change at these steady state values, so, if we start a
system off with y(0) = 0, or y(0) = ±1, the value of y is static. The three
places at which this happens are marked by heavy dots in Figure 13.5(a).
f (y) f (y) Figure 13.5: Steady states (dots) and

intervals for which y increases or decreases
for the differential equation (13.8). See
Example 13.8.
y y
(a) (b)
We also see that f (y) < 0 for −1 < y < 0 and for y > 1. In these intervals,
y(t ) must be a decreasing function of time (dy/dt < 0). On the other hand, for
0 < y < 1 or for y < −1, we have f (y) > 0, so y(t ) is increasing. See arrows
on Figure 13.5(b). We see from this figure that there is a tendency for y to
move away from the steady state value y = 0 and to approach either of the
steady states at 1 or −1. Starting from the initial values given above, we have
(i) y(0) = −0.5 results in y → −1,
(ii) y(0) = 0.3 leads to y → 1, and
(iii) y(0) = 2 implies y → 1. ♦
Example 13.9 (A cooling object) Sketch the same type of diagram for the
problem of a cooling object and interpret its meaning.
f (T )
Solution. Here, the differential equation is
2
dT
= f (T ) = 0.2(10 − T ). (13.9)
dt
A sketch of the rate of change, f (T ) versus the temperature T is shown in T
Figure 13.6. We deduce the direction of the flow directly form this sketch. ♦ 10
Example 13.10 Create a similar qualitative sketch for the more general
form of linear differential equation
dy Figure 13.6: Figure for Example 13.9, the
= f (y) = a − by. (13.10)
dt differential equation (13.9).
For what values of y would there be no change?
19. In Figures 13.6 and 13.7, where is
Solution. The rate of change of y is given by the function f (y) = a − by. This the function positive?
is shown in Figure 13.7. The steady state at which f (y) = 0 is at y = a/b. 20. Consider Eqn. (13.10) analytically:
what value does y approach?
Starting from an initial condition y(0) = a/b, there would be no change. We
f (y)
also see from this figure that y approaches this value over time. After a long
time, the value of y will be approximately a/b. ♦
Steady states and stability

From the last few figures, we observe that wherever the function f on the
right hand side of the differential equations crosses the horizontal axis
(satisfies f = 0) there is a steady state. For example, in Figure 13.6 this takes
y
place at T = 10. At that temperature the differential equation specifies that a
dT /dt = 0 and so, T = 10 is a steady state, a concept we first encountered in b
Chapter 12.
Figure 13.7: Qualitative sketch for
Eqn. (13.10) in Example 13.10.
Definition 13.3 (Steady state) A steady state is a state in which a system is
not changing.
Example 13.11 Identify steady states of Eqn. (13.8),

dy
= y3 − y.
dt
Solution. Steady states are points that satisfy f (y) = 0. We already found
those to be y = 0 and y = ±1 in Example 13.8. ♦
From Figure 13.5, we see that solutions starting close to y = 1 tend to get
closer and closer to this value. We refer to this behaviour as stability of the
steady state.
Definition 13.4 (Stability) We say that a steady state is stable if states that
are initially close enough to that steady state will get closer to it with time.
We say that a steady state is unstable, if states that are initially very close to
it eventually move away from that steady state.
Example 13.12 Determine the stability of steady states of Eqn. (13.8):

dy
= y − y3 .
dt
Solution. From any starting value of y > 0 in this example, we see that after
21. In the state space diagram in
a long time, the solution curves tend to approach the value y = 1. States Figure 13.4, identify the stable
close to y = 1 get closer to it, so this is a stable steady state. For the steady steady states.
state y = 0, we see that initial conditions near y = 0 move away over time.
Thus, this steady state is unstable. Similarly, the steady state at y = −1 is
stable. In Figure 13.5 we show the stable steady states with black dots and the
unstable steady state with an open dot. ♦
13.3 Applying qualitative analysis to biological models
1. Practice the techniques of slope field, state-space diagram, and steady state
analysis on the logistic equation.
2. Explain the derivation of a model for interacting (healthy, infected)

individuals based on a set of assumptions.
3. Identify that the resulting set of two ODEs can be reduced to a single
ODE. Use qualitative methods to analyse the model behaviour and to
interpret the results.
The qualitative ideas developed so far will now be applied to to problems

from biology. In the following sections we first use these methods to obtain
a thorough understanding of logistic population growth. We then derive a
model for the spread of a disease, and use qualitative arguments to analyze
the predictions of that differential equation model.
Qualitative analysis of the logistic equation

We apply the new methods to the logistic equation.
Example 13.13 Find the steady states of the logistic equation, Eqn. (13.1): i The scaled logistic equation, its slope
field, and steady state values are
dN (K − N ) discussed here.
= rN .
dt K
Solution. To determine the steady states of Eqn. (13.1), i.e. the level of
population that would not change over time, we look for values of N such that
dN
= 0.
dt
This leads to
(K − N )
rN = 0,
K
which has solutions N = 0 (no population at all) or N = K (the population is at
its carrying capacity). ♦
We could similarly find steady states of the scaled form of the logistic i A second way to analyze the scaled
equation, Eqn. (13.3). Setting dy/dt = 0 leads to logistic equation, using the phase line
approach, and its connection to the
dy slope field method as described in
0= = ry(1 − y) ⇒ y = 0, or y = 1. Example 13.14.
dt
This comes as no surprise since these values of y correspond to the values
N = 0 and N = K.
Example 13.14 Draw a plot of the rate of change dy/dt versus the value of
y for the scaled logistic equation,Eqn. (13.3): Mastered Material Check
dy 22. Circle the steady states in
= ry(1 − y). Figure 13.8 and identify which one
dt
is stable.
Solution. In the plot of Figure 13.8 only y ≥ 0 is relevant. In the interval
23. Why is y < 0 not relevant in
0 < y < 1, the rate of change is positive, so that y increases, whereas for y > 1, Example 13.14?
the rate of change is negative, so y decreases. Since y refers to population
Rate of change
size, we need not concern ourselves with behaviour for y < 0. dy/dt
From Figure 13.8 we deduce that solutions that start with a positive y
value approach y = 1 with time. Solutions starting at either steady state y = 0
or y = 1 would not change. Restated in terms of the variable N (t ), any initial y
population should approach its carrying capacity K with time. ♦ 1
We now look at the same equation from the perspective of the slope field.
Example 13.15 Draw a slope field for the scaled logistic equation with
Figure 13.8: Plot of dy/dt versus y for the
r = 0.5, that is for
the scaled logistic equation (13.3).
dy
= f (y) = 0.5 · y(1 − y). (13.11)
dt
Solution. We generate slopes for various values of y in Table 13.4 and plot
the slope field in Figure 13.9(a). ♦
Finally, we practice Euler’s method to graph the numerical solution to
Eqn. (13.11) from several initial conditions.
Example 13.16 (Numerical solutions to the logistic equation) Use Euler’s

method to approximate the solutions to the logistic equation (13.11).
y sign of f (y) behaviour of direction of Table 13.4: Table for slope field for the
logistic equation (13.11). See Fig 13.9(a) for
y arrow the resulting diagram.
0 0 no change →
0<y<1 +ve increasing %
1 0 no change →
y>1 -ve decreasing &
1.2 Figure 13.9: (a) Slope field and (b) solution

1.2
1
curves for the logistic equation (13.11),
1 dy
dt = 0.5 · y(1 − y)
population
population
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 5 10 15 20 0 2 4 6 8 10
(a) time (b) time

Solution. In Figure 13.9(b) we show a set of solution curves, obtained by 24. What initial values y0 were used in
solving the equation numerically using Euler’s method. To obtain these drawing the different solution curves
depicted in Figure 13.9(b)?
solutions, a value of h = ∆t = 0.1 was used. The solution is plotted for various
initial conditions y(0) = y0 . The successive values of y were calculated
according to
yk+1 = yk + 0.5yk (1 − yk )h, k = 0, . . . 100.
From Figure 13.9(b), we see that solution curves approach the steady state y = P Link to Google Sheets. This
1, meaning that the population N (t ) approaches the carrying capacity K for spreadsheet implements Euler’s method
for Example 13.16. A chart showing
all positive starting values. A link to the spreadsheet that implements Euler’s solutions from four initial conditions is
method is included. ♦ included.
Example 13.17 (Inflection points) Some of the curves shown in Fig-

ure 13.9(b) have an inflection point, but others do not. Use the differential
equation to determine which of the solution curves have an inflection point.
Solution. We have already established that all initial values in the range 0 <
y0 < 1 are associated with increasing solutions y(t ). Now we consider the
concavity of those solutions. The logistic equation has the form Mastered Material Check
25. How do we know that initial
dy
= ry(1 − y) = ry − ry2 conditions in the range 0 < y0 < 1
dt lead to increasing solutions?
Differentiate both sides using the chain rule and factor, to get
d2y dy dy dy
= r − 2ry = r (1 − 2y).
dt 2 dt dt dt
An inflection point would occur at places where the second derivative

changes sign. This is possible for dy/dt = 0 or for (1 − 2y) = 0. We have
already dismissed the first possibility because we argued that the rate of
change is nonzero in the interval of interest. Thus we conclude that an
inflection point would occur whenever y = 1/2. Any initial condition satis-
fying 0 < y0 < 1/2 would eventually pass through y = 1/2 on its way to the
steady state level at y = 1, and in so doing, would have an inflection point. ♦
A changing aphid population

In Chapters 1 and 5, we investigated a situation when predation and growth
rates of an aphid population exactly balanced. But what happens if these two
rates do not balance? We are now ready to tackle this question.
Featured Problem 13.1 (aphids) Consider the aphid-ladybug problem Hint: Growth rate (number of
(Example 1.3) with aphid density x, growth rate G(x) = rx, and predation rate aphids born per unit time) contributes
positively, whereas predation rate
by a ladybug P(x) as in (1.10). (a) Write down a differential equation for the (number of aphids eaten per unit time)
aphid population. (b) Use your equation, and a sketch of the two functions contributes negatively to the rate of
to answer the following question: What happens to the aphid population change of aphids with respect to time
(dx/dt).
starting from various initial population sizes?
13.3.1 The radius of a growing cell

In Section 11.4 we examined a cell in which nutrient absorption and con-
sumption each contribute to changing the mass balance of the cell. We first
wrote down a differential equation of the form
dm
= A −C.
dt
Assuming the cell was spherical, we showed that this equation results in the
differential equation for the cell radius r (t ):

dr 1 k2
= k1 − r , k1 , k2 , ρ > 0 (13.12)
dt ρ 3
Using tools in this chapter, we can now understand what this implies about
cell size growth.
Featured Problem 13.2 (How cell radius changes) Apply qualitative

methods to Eqn. (13.12) so as to determine what happens to cells starting
from various initial sizes. Is there a steady state cell size? How do your
results compare to our findings in Section 1.2?
A model for the spread of a disease

In the era of human immunodeficiency virus (HIV), Severe Acute Respira-
tory Syndrome (SARS), Avian influenza (“bird flu”) and similar emerging
infectious diseases, it is prudent to consider how infection spreads, and how it

could be controlled or suppressed. This motivates the following example.
For a given disease, let us subdivide the population into two classes:
healthy individuals who are susceptible to catching the infection, and those
that are currently infected and able to transmit the infection to others. We
consider an infection that is mild enough that individuals recover at some
constant rate, and that they become susceptible once recovered.
Note: usually, recovery from an illness leads to partial temporary immunity.
While this, too, can be modelled, we restrict attention to the simpler case
which is tractable using mathematics we have just introduced.
The simplest case to understand is that of a fixed population (with no
birth, death or migration during the timescale of interest). A goal is to
predict whether the infection spreads and persists (becomes endemic) in the
population or whether it runs its course and disappear. We use the following i A video summary of the model for
the spread of a disease, together with its
notation:
analysis.
S(t ) = size of population of susceptible (healthy) individuals,

I (t ) = size of population of infected individuals,
N (t ) = S(t ) + I (t ) = total population size.
We add a few simplifying assumptions.
1. The population mixes very well, so each individual is equally likely to

contact and interact with any other individual. The contact is random.
2. Other than the state (S or I), individuals are “identical,” with the same
rates of recovery and infectivity.
3. On the timescale of interest, there is no birth, death or migration, only

exchange between S and I.
Example 13.18 Suppose that the process can be represented by the scheme
S + I → I + I,
I→S
The first part, transmission of disease from I to S involves interaction. The

second part is recovery. Use the assumptions above to track the two popula-
tions and to formulate a set of differential equations for I (t ) and S(t ).
Solution. The following balance equations keeps track of individuals

     
Rate of Rate of gain Rate of loss
 change of  =  due to disease  −  due to
     

I (t ) transmission recovery
According to our assumption, recovery takes place at a constant rate per unit
time, denoted by µ > 0 . By the law of mass action, the disease transmission
rate should be proportional to the product of the populations, (S · I ). Assigning

β > 0 to be the constant of proportionality leads to the following differential
equations for the infected population:
dI
= β SI − µI.
dt
Similarly, we can write a balance equation that tracks the population of
susceptible individuals:
     
Rate of Rate of Loss Rate of gain
 change of  = −  due to disease  +  due to
     

S(t ) transmission recovery
Observe that loss from one group leads to (exactly balanced) gain in the other
group. By similar logic, the differential equation for S(t ) is then
dS
= −β SI + µI.
dt
We have arrived at a system of equations that describe the changes in each of
the groups, Mastered Material Check
26. Identify any constants in
dI
= β SI − µI, (13.13a) Eqns. (13.13)(a) and (b).
dt
27. What are the units of those
dS
= −β SI + µI. (13.13b) constants?
dt
28. Why does the hint given in
♦ Example 13.19 help?
From Eqns. (13.13) it is clear that changes in one population depend on
both, which means that the differential equations are coupled (linked to one
another). Hence, we cannot “solve one” independently of the other. We must
treat them as a pair. However, as we observe in the next examples, we can
simplify this system of equations using the fact that the total population does
not change.
Example 13.19 Use Eqns.(13.13) to show that the total population does not
change (hint: show that the derivative of S(t ) + I (t ) is zero).
i Video showing that the population
Solution. Add the equations to one another. Then we obtain N (t ) = I (t ) + S(t ) is constant.
d dI dS
[I (t ) + S(t )] = + = β SI − µI − β SI + µI = 0.
dt dt dt
Hence
d dN
[I (t ) + S(t )] = = 0,
dt dt
which mean that N (t ) = [I (t ) + S(t )] = N=constant, so the total population
does not change. (In Eqn. (13.1), here N is a constant and I (t ), S(t ) are the
variables.) ♦
Example 13.20 Use the fact that N is constant to express S(t ) in terms
of I (t ) and N, and eliminate S(t ) from the differential equation for I (t ). Your
equation should only contain the constants N, β , µ.

Solution. Since N = S(t ) + I (t ) is constant, we can write S(t ) = N − I (t ). 29. Redo Example 13.20 but eliminate
Then, plugging this into the differential equation for I (t ) we obtain I (t ) instead of S(t ) .
30. Analyze the equation you get for
dI dI dS(t )/dt as done for dI/dt in
= β SI − µI, ⇒ = β (N − I )I − µI.
dt dt Example 13.21.
Example 13.21 a) Show that the above equation can be written in the form
dI
= β I (K − I ),
dt
where K is a constant.
b) Determine how this constant K depends on N, β , and µ.
c) Is the constant K positive or negative?
Solution.
a) We rewrite the differential equation for I (t ) as follows:

dI µ µ
= β (N − I )I − µI = β I (N − I ) − = βI N − −I .
dt β β
b) We identify the constant,

µ
K = N− .
β
c) Evidently, K could be either positive or negative, that is

(
N ≥ βµ ⇒ K ≥ 0,
µ
N< β ⇒ K < 0.
♦
Using the above process, we have reduced the system of two differential
equations for the two variables I (t ), S(t ) to a single differential equation
for I (t ), together with the statement S(t ) = N − I (t ). We now examine
implications of this result using the qualitative methods of this chapter.
Example 13.22 Consider the differential equation for I (t ) given by

dI µ
≡ f (I ) = β I (K − I ), where K = N − . (13.14)
dt β
Find the steady states of the differential equation (13.14) and draw a state
space diagram in each of the following cases:
(a) K ≥ 0,
(b) K < 0.
Use your diagram to determine which steady state(s) are stable or unstable.
Solution. Steady states of Eqn. (13.14) satisfy dI/dt = β I (K − I ) = 0. 31. What is the significance of the grey
Hence, these steady states are I = 0 (no infected individuals) and I = K. The shaded regions in Fig. 13.10.
latter only makes sense if K ≥ 0. We plot the function f (I ) = β I (K − I ) in 32. Draw Fig. 13.10 for K = 0.
Eqn. (13.14) against the state variable I in Figure 13.10 (a) for K ≥ 0 and (b) 33. Why is I = K not an admissible
steady state if K < 0?
for K < 0. Since f (I ) is quadratic in I, its graph is a parabola and it opens
downwards. We add arrows pointing right (→) in the regions where dI/dt > 0
and arrows pointing left (←) where dI/dt < 0.
f (I) f (I)
Figure 13.10: State-space diagrams for
differential equation (13.14). Plots of f (I )
as a function of I in the cases (a) K ≥ 0,
and (b) K < 0. The grey regions are not
biologically meaningful since I cannot be
negative.
I 0 I
0 K K
(a) (b)
In case (a), when K ≥ 0, we find that arrows point toward I = K, so this

steady state is stable. Arrows point away from I = 0, so this represents an
unstable steady state. In case (b), while we still have a parabolic graph with
two steady states, the state I = K is not admissible since K is negative. Hence
only one steady state, at I = 0 is relevant biologically, and all initial conditions
move towards this state. ♦
Example 13.23 Interpret the results of the model in terms of the disease,
assuming that initially most of the population is in the susceptible S group,
and a small number of infected individuals are present at t = 0.
Solution. In case (a), as long as the initial size of the infected group is
positive (I > 0), with time it approaches K, that is, I (t ) → K = N − µ/β . The
rest of the population is in the susceptible group, that is S(t ) → µ/β (so that
S(t ) + I (t ) = N is always constant.) This first scenario holds provided K > 0
which is equivalent to N > µ/β . There are then some infected and some
healthy individuals in the population indefinitely, according to the model. In
this case, we say that the disease becomes endemic.
In case (b), which corresponds to N < µ/β , we see that I (t ) → 0 regardless Mastered Material Check
of the initial size of the infected group. In that case, S(t ) → N so with time, 34. In the case that β = 0.001per person
the infected group shrinks and the healthy group grows so that the whole per day and µ = 0.1 per day, how
population becomes healthy. From these two results, we conclude that the large would the population have to
be for the disease to become
disease is wiped out in a small population, whereas in a sufficiently large endemic?
population, it can spread until a steady state is attained where some fraction 35. Frequent hand-washing can be a
of the population is always infected. In fact we have identified a threshold protective measure that decreases
the spread of disease. Which
that separates these two behaviours: parameter of the model would this
affect and in what way?
Nβ
>1 ⇒ disease becomes endemic,
µ
Nβ
<1 ⇒ disease is wiped out.
µ
The ratio of constants in these inequalities, R0 = Nβ /µ is called the basic i A video summarizing the
interpretation of the model and the
reproduction number for the disease. Many current and much more detailed
meaning of the constant R0 = Nβ /µ.
models for disease transmission also have such threshold behaviour, and
the ratio that determines whether the disease spreads or disappears, R0 is of
great interest in vaccination strategies. This ratio represents the number of
infections that arise when 1 infected individual interacts with a population
of N susceptible individuals.
13.4 Summary
1. A differential equation of the form α dydt + β y + γ = 0 is linear (and

“first order”). We encountered several examples of nonlinear DEs in this
chapter.
2. A (possibly nonlinear) differential equation dy dt = f (y) can be analyzed
qualitatively by observing where f (y) is positive, negative or zero.
3. A slope field (or “direction field”) is a collection of tangent vectors for
solutions to a differential equation. Slope fields can be sketched from f (y)
without the need to solve the differential equation.
4. A solution curve drawn in a slope field corresponds to a single solution to
a differential equation, with some initial y0 value given.
5. A state space (or “phase line” diagram) for the differential equation is a y
axis, together with arrows describing the flow (increasing/decreasing/stationary)
along that axis. It can be obtained from a sketch of f (y).
6. A steady state is stable if nearby states get closer. A steady state is unsta-
ble if nearby states get further away with time.
7. Creating/interpreting slope field and state space diagrams is helpful in
understanding the behavior of solutions to differential equations.
8. Applications considered in this chapter included:
(a) the logistic equations for population growth (a nonlinear differential

equation, scaling, steady state and slope field demonstration);
(b) the Law of Mass Action (a nonlinear differential equation);
(c) a cooling object (state space and phase line diagram demonstration);
and
(d) disease spread model (an extensive exposition on qualitative differential
equation methods).
1. Why is it helpful to rescale an equation?
2. Identify which of the following differential equations are linear:
(a) 5 dy
dt − y = −0.5 (c) dy
dx + πy + ρ = 3
2 dx
(b) dy
+y+1 = 0 (d) dt + x + 2 = −3x
dt
3. Consider the following slope field:

y
20
15
(a) Where is y decreasing?

10
(b) What is y approaching?
5
t
5 10 15 20
4. Circle the stable steady states in the following state space diagram
f (y)
y
Exercises
13.1. Explaining connections. Explain the connection between Eqn. (13.2)

and the equations Eqn. 11.2 and Eqn. (12.4).
13.2. Slope fields. Consider the differential equations given below. In
each case, draw a slope field, determine the values of y for which no
change takes place - such values are called steady states - and use
your slope field to predict what would happen starting from an initial
value y(0) = 1.
dy
(a) = −0.5y
dt
dy
(b) = 0.5y(2 − y)
dt
dy
(c) = y(2 − y)(3 − y)
dt
13.3. Drawing slope fields. Draw a slope field for each of the given differ-
ential equations:
dy
(a) dt = 2 + 3y
dy
(b) dt = −y(2 − y)
dy
(c) dt = 2 − 3y + y2
dy
(d) dt = −2(3 − y)2
dy
(e) dt = y2 − y + 1
dy
(f) dt = y3 − y
dy √
(g) dt = y(y − 2)(y − 3)2 , y ≥ 0.
13.4. Linear or Nonlinear. Identify which of the differential equations in

Exercise 2 and 3 is linear and which nonlinear.
13.5. Using slope fields. For each of the differential equations (a) to (g) in
Exercise 3, plot dy
dt as a function of y, draw the motion along the y-axis,
identify the steady state(s) and indicate if the motions are toward or
away from the steady state(s).
13.6. Direction field. The direction field shown in the figure below corre-
sponds to which differential equation?
dy
(A) = ry(y + 1)
dt
dy
(B) = r (y − 1)(y + 1)
dt
dy
(C) = −r (y − 1)(y + 1)
dt
dy
(D) = ry(y − 1)
dt
dy
(E) = −ry(y + 1)
dt
13.7. Differential equation. Given the differential equation and initial

condition
dy
= y2 (y − a), y(0) = 2a
dt
where a > 0 is a constant, the value of the function y(t ) would
(A) approach y = 0;
(B) grow larger with time;
(C) approach y = a;
(D) stay the same;
(E) none of the above.
13.8. There’s a hole in the bucket. Water flows into a bucket at constant
rate I. There is a hole in the container. Explain the model
dh √
= I − k h.
dt
Analyze the behaviour predicted. What would the height be after a
long time? Is this result always valid, or is an additional assumption
needed? (hint: recall Example 12.3.)
13.9. Cubical crystal. A crystal grows inside a medium in a cubical shape
with side length x and volume V . The rate of change of the volume is
given by
dV
= kx2 (V0 −V )
dt
where k and V0 are positive constants.
dx
(a) Rewrite this as a differential equation for dt .
(b) Suppose that the crystal grows from a very small “seed.” Show that
its growth rate continually decreases.
(c) What happens to the size of the crystal after a very long time?
(d) What is its size (that is, what is either x or V ) when it is growing at
half its initial rate?
13.10. The Law of Mass Action. The Law of Mass Action in Section 13.1
led to the assumption that the rate of a reaction involving two types of
molecules (A and B) is proportional to the product of their concentra-
tions, k · a · b.
Explain why the sum of the concentrations, k · (a + b) would not make
for a sensible assumption about the rate of the reaction.
13.11. Biochemical reaction. A biochemical reaction in which a substance S
is both produced and consumed is investigated. The concentration c(t )
of S changes during the reaction, and is seen to follow the differential
equation
dc c
= Kmax − rc
dt k+c
where Kmax , k, r are positive constants with certain convenient units.
The first term is a concentration-dependent production term and the
second term represents consumption of the substance.
(a) What is the maximal rate at which the substance is produced? At
what concentration is the production rate 50% of this maximal
value?
(b) If the production is turned off, the substance decays. How long
would it take for the concentration to drop by 50%?
(c) At what concentration does the production rate just balance the
consumption rate?
13.12. Logistic growth with proportional harvesting. Consider a fish
population of density N (t ) growing at rate g(N ), with harvesting, so
that the population satisfies the differential equation
dN
= g(N ) − h(N ).
dt
(K−N )
Now assume that the growth rate is logistic, so g(N ) = rN K
where r, K > 0 are constant. Assume that the rate of harvesting is
proportional to the population size, so that
h(N ) = qEN
where E, the effort of the fishermen, and q, the catchability of this type
of fish, are positive constants.
Use qualitative methods discussed in this chapter to analyze the
behaviour of this equation. Under what conditions does this lead to a
sustainable fishery?
13.13. Logistic growth with constant number harvesting. Consider the
same fish population as in Exercise 12, but this time assume that the
rate of harvesting is fixed, regardless of the population size, so that
h(N ) = H
where H is a constant number of fish being caught and removed per

unit time. Analyze this revised model and compare it to the previous
results.
13.14. Scaling time in the logistic equation. Consider the scaled logistic
equation (13.3). Recall that r has units of 1/time, so 1/r is a quantity
with units of time. Now consider scaling the time variable in (13.3) by
defining t = s/r. Then s carries no units (s is “dimensionless”).
Substitute this expression for t in (13.3) and find the differential

equation so obtained (for dy/ds).
13.15. Euler’s method applied to logistic growth. Consider the logistic
dy
= ry(1 − y).
dt
Let r = 1. Use Euler’s method to find a solution to this differential

equation starting with y(0) = 0.5, and step size h = 0.2. Find the values
of y up to time t = 1.0.
13.16. Spread of infection. In the model for the spread of a disease, we used
the fact that the total population is constant (S(t ) + I (t ) = N =constant)
to eliminate S(t ) and analyze a differential equation for I (t ) on its
own.
Carry out a similar analysis, but eliminate I (t ). Then analyze the

differential equation you get for S(t ) to find its steady states and
behaviour, practicing the qualitative analysis discussed in this chapter.
13.17. Vaccination strategy. When an individual is vaccinated, he or she is
“removed” from the susceptible population, effectively reducing the
size of the population that can participate in the disease transmission.
For example, if a fraction φ of the population is vaccinated, then
only the remaining (1 − φ )N individuals can be either susceptible or
infected, so S(t ) + I (t ) = (1 − φ )N. When smallpox was an endemic
disease, it had a basic reproductive number of R0 = 7.
What fraction of the population would have had to be vaccinated to

eradicate this disease?
13.18. Social media. Sally Sweetstone has invented a new social media
App called HeadSpace, which instantly matches compatible mates
according to their changing tastes and styles. Users hear about the App
from one another by word of mouth and sign up for an account. The
account expires randomly, with a half-life of 1 month. Suppose y1 (t )
are the number of individuals who are not subscribers and y2 (t ) are
the number of are subscribers at time t. The following model has been
suggested for the evolving subscriber population
dy1
= by2 − ay1 y2 ,
dt
dy2
= ay1 y2 − by2 .
dt
(a) Explain the terms in the equation. What is the value of the con-
stant b?
(b) Show that the total population P = y1 (t ) + y2 (t ) is constant.
Note: this is a conservation statement.
(c) Use the conservation statement to eliminate y1 . Then analyze the
differential equation you obtain for y2 .
(d) Use your model to determine whether this newly launched social
media will be successful or whether it will go extinct.
13.19. A bimolecular reaction. Two molecules of A can react to form a
new chemical, B. The reaction is reversible so that B also continually
decays back into 2 molecules of A. The differential equation model
proposed for this system is
da
= −µa2 + 2β b
dt
db µ
= a2 − β b,
dt 2
where a(t ), b(t ) > 0 are the concentrations of the two chemicals.
(a) Explain the factor 2 that appears in the differential equations
and the conservation statement. Show that the total mass M =
a(t ) + 2b(t ) is constant.
(b) Use the techniques in this chapter to investigate what happens in
this chemical reaction, to find any steady states, and to explain the
behaviour of the system
14
Periodic and trigonometric functions
Nature abounds with examples of cyclic processes. Perhaps most familiar

is the continually repeating heartbeat that accompanies us through life.
Electrically active muscles power the heart. That electrical activity can
be recorded on the surface of the body by an electrocardiogram (ECG),
as shown in Figure 14.1. In a normal healthy human at rest, the electrical
activity pattern associated with a complete cycle repeats itself with roughly
1 heartbeat per second. At exercise, the beating heart pumps faster, so the
pattern repeats more frequently.
To study the behaviour shown in Figure 14.1, we must first develop
language that describes such periodic phenomena. For example, we need
to quantify what is meant by “more frequent repetition" of a heart beat, Figure 14.1: An electrocardiogram (ECG)
“skipping a beat,” or other shifts in the pattern of this or of any other cycling shows the pattern of electrical activity of the
heart. Vertical axis: millivolts, horizontal
system. axis: time in seconds. At rest (top), the heart
Before trying to understand intricate examples such as ECG’s, we begin beats approximately 60 times per minute,
with simple prototypes of periodic functions: the trigonometric functions, but in exercise (lower trace), the heart rate
increases.
sine and cosine. Belonging to a wider class of periodic functions, these
cases illustrate ideas of amplitude, frequency, period, and phase. Many
cyclic phenomena can be described approximately by suitably adjusting such
basic functions. Our study in this elementary context then aids in the goal
of analyzing periodic functions in general - and periodicity of the ECG’s in
particular (we return to this in Example 14.2).
As a second theme, we return to inverse functions and show that restric-
tions must be applied to ensure the existence of an inverse, particularly for
trigonometric functions. Then, in Chapter 15, we calculate the derivatives
of trigonometric functions and explore applications to rates of change of
periodic phenomena or changing angles.
14.1 Basic trigonometry
1. Define radian as a measure for angles.
2. Describe the correspondence between a point moving on a unit circle and

the sine and cosine of the angle it forms at the origin.
3. Make correspondence between ratios of sides of a Pythagorean triangle

and the trigonometric functions of one of its angles.
4. Review properties of the functions sin(x) and cos(x) and other trigono-
metric functions. State and apply the connections between these functions
(“trigonometric identities”).
Trigonometric functions are closely associated with angles and ratios of

sides of a right-angle triangle. They are also connected to the motion of a
point moving around a unit circle. Before we articulate these connections, we
must agree on a universal way of measuring angles.
Angles and circles

Angles can be measured in a number of ways. One is to assign a value in Radius, R
degrees, with the convention that one complete revolution is represented by
360◦ . It turns out that this measure is not particularly convenient, and we
instead replace it with a more universal quantity. Diameter, D
Our definition of angles will be based on the fact that circles of all sizes
have one common geometric feature: they have the same ratio of circumfer-
ence, to diameter, no matter what their size (or where in the universe they
occur). We call that ratio π, that is
Circumference, C
Circumference of circle
π= .
Diameter of circle Figure 14.2: Circumference, diameter and
radius of a circle.
By construction, the diameter D of a circle is a distance that corresponds to
twice the radius R of that circle, so
D = 2R. S
This leads to the familiar relationship of circumference C, to radius R,
θ
C = 2πR. R
This statement is merely a definition of the constant π.
As shown in Figure 14.3, an angle θ can be put into correspondence with
an arc along the edge of a circle. For a circle of radius R and angle θ we
define the arclength, S by the relation S = Rθ where θ is measured in a Figure 14.3: The angle θ in radians is
related to the radius R of the circle, and
the length of the arc S by the simple
formula, S = Rθ .
PERIODIC AND TRIGONOMETRIC FUNCTIONS 305
convenient way that we now select. Both S and R carry units of “distance" or
“length". But their ratio is,
S
θ= ,
R
so the units in numerator and denominator cancel, and the angle θ is dimen-
sionless (carries no units). Mastered Material Check
Now consider a circle of radius R = 1 (called a unit circle) and denote by s 1. What angle θ corresponds to a 1/6
revolution around the perimeter of a
a length of arc around the entire perimeter of this unit circle. Then circle?
s 2. Sketch an angle of π/4 radians.
θ= .
1 3. If the radius of a circle is 2 and an
In particular, for one complete revolution around the circle, the arclength is arc on its perimeter has length 0.5,
what is the angle corresponding to
s = 2π · 1 = 2π, which is just the circumference of the unit circle. In that case, that arc (in radians)?
it makes sense to consider the angle corresponding to one revolution as
2π
θ= = 2π.
1
This leads naturally to the definition of the radian: we identify an angle of
2π radians with one complete revolution around the circle. Note that (like
degrees or other measures of angles), a radian is a number that carries no
“units”.
We can now use this measure for angles to assign values to any frac-
tion of a revolution, and thus, to any angle. For example, an angle of 90◦
corresponds to one quarter of a revolution around the perimeter of a unit
circle, so we identify the angle π/2 radians with it. One degree is 1/360 of a
revolution, corresponding to 2π/360 radians, and so on.
We summarize the properties of radians:
1. The length of an arc along the perimeter of a circle of radius R corre-

sponding to an angle θ between two radii is S = Rθ for θ in radians.
2. An angle in radians is the ratio of the arclength it subtends in a circle to

the radius of that circle (and hence, a radian carries no units).
3. One complete revolution, or one full cycle corresponds to an angle of 2π

radians.
We can convert between degrees and radians by remembering that 360◦

corresponds to 2π radians (180◦ then corresponds to π radians, 90◦ to π/2
radians, etc.) Mastered Material Check
4. How many radians does 270◦
correspond to?
Defining the trigonometric functions sin(t ) and cos(t )
5. Label cos(t ) and sin(t ) where
Consider a point (x, y) moving around the rim of a circle of radius 1, and appropriate on Figure 14.4.
let t be some angle (measured in radians) formed by the x-axis and the radius
vector to the point (x, y) as shown in Figure 14.4.
Figure 14.4: The equation of a circle

of radius 1, with center at the origin is
x2 + y2 = 1. The radius vector to the
point (x, y) forms an angle t (radians) with
, (x, y) 1 the x-axis. In the triangle, shown enlarged
on the right, lengths of sides are labeled.
y The trigonometric functions are just ratios of
two sides of this triangle.
t
t
x
We define the functions sine and cosine, both dependent on the angle t
(abbreviated sin(t ) and cos(t )) as follows:
y x
sin(t ) = = y, cos(t ) = = x
1 1
That is, the function sine tracks the y coordinate of the point as it moves
around the unit circle, and the function cosine tracks its x coordinate. ` A demonstration of the link between
motion on a circle and the function
Note: A review and definitions of trigonometric quantities is given in Fig- y = sin(x). Click on the arrow left of the
ure F.1, Appendix F as ratios of sides in a right angle triangle. The hy- parameter a or shift the slider on a to
potenuse in our diagram is simply the radius r = 1 of the circle. see the moving point.
Featured Problem 14.1 (Cosine as motion around circle) Adapt the inter-
active sine graph to represent the link between the graph of y = cos(x) and the
x coordinate of a point moving around a circle.
Properties of sin(t ) and cos(t )

We now explore the consequences of these definitions:
Values of sine and cosine.
• The radius of the unit circle is 1. This means that the x coordinate of any
point (x, y) on the unit circle cannot be larger than 1 or smaller than −1.
The same holds for the y coordinate. Thus, the functions sin(t ) and cos(t )
are always swinging between −1 and 1. (−1 ≤ sin(t ) ≤ 1 and −1 ≤
cos(t ) ≤ 1 for all angles t). The maximum value of each function is 1, the
minimum is −1, and the average is 0.
• We adopt the convention that when the radius vector points along the
x-axis, the angle is t = 0, and coordinates of the point are x = 1, y = 0. This
implies that cos(0) = 1, sin(0) = 0.
• When the radius vector points up the y-axis, the angle is π/2 (correspond-
ing to one quarter of a complete revolution), and coordinates of the point
are x = 0, y = 1 so that cos(π/2) = 0, sin(π/2) = 1.
• Through geometry, we can also determine the lengths of all sides - and
hence the ratios of the sides - of particular triangles, namely
– equilateral triangles (in which all angles are 60◦ ), and Mastered Material Check
– right triangles (two equal angles of 45◦ ). 6. Review Appendix F and then use
triangles to determine the x and y
These types of calculations (omitted here) lead to some easily determined coordinates of angles of 60◦
and 45◦ in the unit circle.
values for the sine and cosine of such special angles. These are shown in
7. Why does cosine have its largest
Table F.1 of the Appendix F. value when the angle t = 0, at the
beginning of the cycle?
Characteristics of sine and cosine.
• Both sin(t ) and cos(t ) go through the same values every time the angle t
completes another cycle around the circle. We refer to such functions as
periodic functions.
• The two functions, sine and cosine depict the same underlying motion,
viewed from two perspectives: cos(t ) represents the projection of the
circularly moving point onto the x-axis, while sin(t ) is the projection of
the same point onto the y-axis. In this sense, the functions are “twins”, and
as such we expect many relationships connecting them.
• The cosine has its largest value at the beginning of the cycle, when t = 0
(since cos(0) = 1), while sine has its peak value a little later (sin(π/2) =
1). Throughout their circular race, the sine function is π/2 radians ahead
of the cosine, that is, π
cos(t ) = sin t + .
2
See Figure 14.5 for graphs of both functions showing this shift by π/2.
• The period T of the sine function sin(t ) is defined as the value of t for
which one whole cycle (around the circle) has been completed. Ac-
cordingly, this period is T = 2π. Similarly the period of the cosine
function cos(t ) is also 2π. (See Figure 14.5.)
• The point (x, y) = (cos(t ), sin(t )) is on a circle of radius 1, and, thus, its
coordinates satisfy
x2 + y2 = 1.
This implies that
sin2 (t ) + cos2 (t ) = 1 (14.1)
for any angle t. This is an important relation, (also called a trigonometric
identity), and one that is frequently used. See Appendix F for a review of
other trigonometric identities.
• The sine and cosine functions have symmetries that we already encoun-
tered: sin(t ) is an odd function (symmetric about the origin) and the
cos(t ) is an even function (symmetric about the y-axis). These symme-
tries also imply that sin(−t ) = − sin(t ) and cos(−t ) = cos(t ).
Figure 14.5: Periodicity of the sine and

cosine. Note that the two curves are just
y = sin(t)
shifted versions of one another.
0 π 3π 5π t
2 π 2 2π 2 3π 4π
−1
period, T
y = cos(t)
0 π 3π 5π t
2 π 2 2π 2 3π
−1
period, T
Other trigonometric functions

` You can use this desmos graph to
Although we shall mostly be concerned with the two basic functions de- see all the trigonometric functions. Turn
scribed above, several others are historically important and are encountered the graphs on or off by clicking on the
(grey) circles to the left of the formulae.
frequently in integral calculus. These include the following: Notice the vertical asymptotes on some
of these functions and think about
sin(t ) 1 where these asymptotes occur.
tan(t ) = , cot(t ) = ,
cos(t ) tan(t )
1 1
sec(t ) = , csc(t ) = .
cos(t ) sin(t )
We review these and the identities that they satisfy in Appendix F. We also
include the Law of Cosines in Eqn. (F.2), and angle-sum identities in the
same appendix. Sine and cosine are the functions we focus on here.
14.2 Periodic functions
1. Define a periodic function.
2. Given a periodic function, determine its period, amplitude and phase.
3. Given a graph or description of a periodic or rhythmic process, “fit” an

approximate sine or cosine function with the correct period, amplitude and
phase.
In Section 14.1, we identified the period of sin(t ) and cos(t ) as the value
of t at which one full cycle is completed. Here we formalize the definition of
a periodic function, define its period, frequency, and other properties.
Definition 14.1 (Periodic function) A function is said to be periodic if
f (t ) = f (t + T ),
where T is a constant that we call the period of the function. Graphically,

this means that if we shift the function by a constant “distance” along the
horizontal axis, we see the same picture again. Make demo?
Example 14.1 Show that the trigonometric functions are indeed periodic.
Solution. The point (x, y) in Figure 14.4 repeats its trajectory every time 8. If cos(α ) = β , what
a revolution around the circle is complete. This happens when the angle t is cos(α + 2π )? cos(α − 6π )?
completes one full cycle of 2π radians. Thus, as expected, the trigonometric
functions are periodic, that is
sin(t ) = sin(t + 2π ), and cos(t ) = cos(t + 2π ).
Similarly
tan(t ) = tan(t + 2π ), and cot(t ) = cot(t + 2π ).

` Use the sliders on this desmos graph
to see how the amplitude, A, frequency
We say that the period of these functions is T = 2π radians. The same applies w, and phase φ affect the graph of the
to sec(t ) and csc(t ), that is, all six trigonometric functions are periodic. ♦ function y = M + A sin(w(t − φ )). You
can also vary the mean value M.
Phase, amplitude, and frequency Mastered Material Check
In Appendix C we review how the appearance of any function changes when 9. How many zeros are depicted in
each panel of Figure 14.6?
we transform variables. For example, replacing the independent variable x
10. How many local minima are
by x − a (or αx) shifts (or scales) the function horizontally, multiplying f by a depicted in each panel of
constant C scales the function vertically, etc. The same ideas apply to shapes Figure 14.6?
of a trigonometric function when similar transformations are applied. 11. In each panel of Figure 14.6,
A function of the form identify where y = 1.
12. Indicate a single period on each
panel of Figure 14.6.
y = f (t ) = A sin(ωt )
has both t and y-axes scaled, as shown in Figure 14.6(c). The the amplitude
of the graph, A scales the y axis so that the oscillation swings between a
minimum of −A and a maximum of A. The frequency ω, scales the t-axis.
This cycles are crowded together (if ω > 1) or stretched out (if ω < 1). One
full cycle is completed when
ωt = 2π,
and this occurs at time

2π
t= .
ω
Figure 14.6: The sine function y = sin(t )

(shown in (a)) is transformed in several
y y = sin(t) y y = A sin(t) ways. (b) Multiplying the function by
2 2
a constant (A = 2) stretches the graph
vertically. A is called the amplitude. (c)
1 1 Multiplying the independent variable by a
constant (ω = 3) increases the frequency,
i.e. the number of cycles per unit time.
π π t π π t (d) Subtracting a constant (a = 0.8) from
2
3π 2π 2
3π
2
2π
2 the independent variable shifts the graph
−1 −1 horizontally to the right.
−2 −2
(a) (b)
y y = A sin(ωt) y y = A sin(ω(t − a))
2 2
1 1
π π 3π
t
2π π π 3π
t
2π
2 2 2 2
−1 −1
−2 −2
(c) (d)

This time, called the period of the function is denoted by T . The connection
13. What is the period of a
between frequency and period is: trigonometric functions whose
frequency is 5 cycles per min?
2π 2π
ω= , ⇒ T= . 14. What is the frequency of a
T ω trigonometric functions whose
period is 1 hr?
If we examine a graph of the function
y = f (t ) = A sin(ω (t − a)),
we find that the basic sine graph has been shifted in the positive t direction by 15. Sketch a graph of y = 3 sin(t − π2 )
a, as in Figure 14.6(d). At time t = a, the value of the function is and y = 3 cos(t )
16. Sketch a graph of y = cos(4(t − π ))
y = f (t ) = A sin(ω (a − a)) = A sin(0) = 0,
so the cycle “starts” with a delay of t = a relative to the basic sine function.
Another common variant of the same function can be written in the form
y = f (t ) = A sin(ωt − φ ).
Here φ is called the phase shift of the oscillation. The above two forms are
the same if we identify φ with ωa. The phase shift, φ has no units, whereas a
has units of time, which is the same as the units of t. When φ = 2π, (which
happens when a = 2π/ω), the graph has been moved over to the right by one
full period, making it identical to the original periodic graph.
Some of the scaled, shifted, sine functions described here are shown in
Figure 14.6.
The periodic electrocardiogram

With the terminology of periodic function in place, we can now describe the
ECG pattern for both normal resting individuals and those at exercise.
Recall that at rest, the heart beats approximately once per second. Con-
sider the ECG trace on the left of Figure 14.7. This corresponds to a single
heartbeat, and so, takes 1 second from start to finish. Suppose t represents
time in seconds, and let y = f (t ) represent the electrical activity (in mV) Figure 14.7: One full ECG cycle (left) has
been “wrapped around a circle”.
at time t. Then, since this pattern repeats, the function f is periodic, with
period T = 1 second. We can write
f (t ) = f (t + 1), t in seconds.
However, suppose the individual starts running. Then this relationship

no longer holds, since heartbeats become more frequent, and the length of
their period, T , decreases. This suggests a more natural way to mark off time
- the amount of time it takes to complete a heartbeat cycle. Thus, rather than
seconds, we define a such a variable, denoting it “the cycle time” and use the
notation τ̄. Then the connection between clock time t and cycle time is
t = time in seconds = number of cycles · length of 1 cycle in sec = τ̄ · T
Restated, Mastered Material Check

t
τ̄ = . 17. If a runner’s heart beats every 0.6
T seconds, how many beat cycles
Note: the “number of cycles” need not be an integer - for example, τ̄ = 2.75 ellapse in 10 seconds?
means that we are 3/4 way into the third electrical activity cycle. 18. Suppose t = 4 sec and τ̄ = 5.
Since f is a periodic function, we can “join up” its two ends and “wrap it Determine T and interpret this
situation.
around a circle”, as shown in the schematic on Figure 14.7. Then successive
heartbeats are depicted by traversing the circle over and over again. This
suggests identifying the beginning of a cycle with 0 and the end of a cycle
with 2π. To do so, we revise our cycle time clock as follows. Define
τ = number of radians traversed since time 0.
Then every heartbeat corresponds to this clock adding an increment of 2π.

The connection between this cycle clock and time t in seconds is
t
τ = 2π τ̄ = 2π (14.2)
T
(We check that when 1 cycle is complete, t = T and τ = 2π, as desired.)
We can now describe the periodicity of the ECG in terms of the cycle
clock by the formula
f (τ ) = f (τ + 2π ). (14.3)
As a check, we show in the next example that the relationship in Eqn. (14.3)
reduces to the familiar period and frequency notation in terms of our original
time t in seconds.
Example 14.2 (Period and frequency of heartbeat) Use the formula we

arrived at for τ and its connection to clock time t to transform back to time t
in seconds. Express the periodicity of the function f both in terms of the
period T and the frequency ω of heartbeats.
Solution. Start with Eqn. (14.3),
f (τ ) = f (τ + 2π ).
Substitute τ = 2πt/T from Eqn. (14.3) and simplify by making a common

denominator. Then

2πt 2πt
f =f + 2π
T T

2π
=f (t + T ) .
T
We now rewrite this in terms of the frequency ω = 2π/T to arrive at
f (ωt ) = f (ω (t + T )) .
This relationship holds for any regular heartbeat, whether at rest or exercise
where the frequency of the heartbeat, ω, is related to the period (duration of 1
beat cycle) by ω = 2π/T . ♦
Rhythmic processes
Many natural phenomena are cyclic. It is sometimes convenient to represent
such phenomena with a simple periodic functions, such as sine or cosine.
Given some periodic process, we determine its frequency (or period), ampli-
tude, and phase shift. We create a trigonometric function (sine or cosine) that
approximates the desired behaviour.
To select a function, it helps to remember that (at t = 0) cosine starts at
its peak, while sine starts at its average value of 0. A function that starts at
the lowest point of the cycle is − cos(t ). In most cases, the choice of sine or
cosine to represent the cyclic phenomenon is arbitrary, they are related by a
simple phase shift.
Next, pick a constant ω such that the trigonometric function sin(ωt )
(or cos(ωt )) has the correct period using the relationship ω = 2π/T . We
then select the amplitude, and horizontal and vertical shifts to complete the
process. The examples below illustrate this process.
Example 14.3 (Daylight hours:) In Vancouver, the shortest day (8 hours of

light) occurs around December 22, and the longest day (16 hours of light)
is around June 21. Approximate the cyclic changes of daylight through the
season using the sine function.
Solution. On Sept 21 and March 21 the lengths of day and night are equal,
and then there are 12 hours of daylight (each of these days is called an
equinox). Suppose we identify March 21 as the beginning of a yearly day-
night length cycle. Let t be time in days beginning on March 21. One full
cycle takes a year, i.e. 365 days. The period of the function we want is thus
T = 365
and its frequency (in units of per day) is
ω = 2π/365.
Daylight shifts between the two extremes of 8 and 16 hours: i.e. 12 ± 4

hours. This means that the amplitude of the cycle is 4 hours. The oscillation
take place about the average value of 12 hours. We have decided to start a
cycle on a day for which the number of daylight hours is the average value
(12). This means that the sine would be most appropriate, so the function that
19. In August, the average number of
best describes the number of hours of daylight at different times of the year
daylight hours is 14. How does this
is: fit with our model?
2π
D(t ) = 12 + 4 sin t 20. Repeat Example 14.3 using a cosine
365 function.
where t is time in days and D the number of hours of light. ♦
Example 14.4 (Hormone levels) The level of a certain hormone in the

bloodstream fluctuates between undetectable concentration at t =7:00 and 100
ng/ml (nanograms per millilitre) at t =19:00 hours. Approximate the cyclic
variations in this hormone level with the appropriate periodic trigonometric
H(t) period: T = 24 hrs
function. Let t represent time in hours from 0:00 hrs through the day.
100
Solution. We first note that it takes one day (24 hours) to complete a cycle.
This means that the period of oscillation is 24 hours, so that the frequency is
50
2π 2π π
ω= = = .
T 24 12
The level of hormone varies between 0 and 100 ng/ml, which can be ex- 0 t
pressed as 50 ± 50 ng/ml. (The trigonometric functions are symmetric cycles, 1 7 13 19 24
and we are finding both the average value about which cycles occur and the 12 hrs
amplitude of the cycles.) We could consider the time midway between the 6 hrs
low and high points, namely 13:00 hours as the point corresponding to the
upswing at the start of a cycle of the sine function. (See Figure 14.8 for the Figure 14.8: Hormonal cycles. The full
cycle takes 24 hrs (hence the period is
sketch.) Thus, if we use a sine to represent the oscillation, we should shift it
T = 24h and the frequency is ω = 2π/24
by 13 hrs to the left. per hour). The level H (t ) swings between 0
and 100 ng/ml. From the given information,
we see that the average level is 50 ng/ml,
and that the origin of a sine curve should be
at t = 13 (i.e. 1/4 of the cycle which is 6 hrs
past the time point t = 7).
Assembling these observations, we obtain the level of hormone, H at time

t in hours: π
H (t ) = 50 + 50 sin (t − 13) .
12
In the expression above, the number 13 represents a shift along the time axis,
and carries units of time. We can express this same function in the form

πt 13π
H (t ) = 50 + 50 sin − .
12 12
In this version, the quantity
13π
φ=
12
is the phase shift.
In selecting the periodic function to use for this example, we could have
made other choices. For example, the same periodic can be represented by
any of the functions listed below:
π
H1 (t ) = 50 − 50 sin (t − 1) ,
12 Mastered Material Check
π
H2 (t ) = 50 + 50 cos (t − 19) , 21. Verify that H1 , H2 and H3 all have
12
the same periods.
π
H3 (t ) = 50 − 50 cos (t − 7) .
12
All these functions have the same values, the same amplitudes, and the same
periods. ♦
Example 14.5 (Phases of the moon) A cycle of waxing and waning moon
takes 29.5 days approximately. Construct a periodic function to describe the
changing phases, starting with a “new moon” (totally dark) and ending one
cycle later.
0 29.5
Solution. The period of the cycle is T = 29.5 days, so
2π 2π Figure 14.9: Periodic moon phases. The
ω= = .
T 29.5 horizontal axis is time in days, and the
vertical axis represents P(t ), the fraction of
Let P(t ) be the fraction of the moon face on day t in the cycle. Then we
the moon that is visible from Earth on day t.
should construct the function so that 0 < P < 1, with P = 1 in mid cycle (see
Figure 14.9). The cosine function swings between the values -1 and 1. To
obtain a positive function in the desired range for P(t ), we add a constant and
scale the cosine as follows:
1
[1 + cos(ωt )].
2
This is still not quite right, since at t = 0 this function takes the value 1, rather
than 0, as shown in Figure 14.9. To correct this we can either introduce a
phase shift, i.e. set
1 Mastered Material Check
P(t ) = [1 + cos(ωt + π )].
2 22. What fraction of the moon do you
Then when t = 0, we get P(t ) = 0.5[1 + cos π ] = 0.5[1 − 1] = 0. Or we can expect to be visible one week into
the cycle?
write
1
P(t ) = [1 − cos(ωt )],
2
which achieves the same result. ♦
14.3 Inverse trigonometric functions
Note: the material in this section can be omitted without loss of continuity in
the next chapter. If this is done, merely skip Sections 15.1 and 15.3.
1. Review inverse functions, and define the inverse trigonometric functions.
2. Explain why the domain of a periodic function must be restricted to define

its inverse. Given any of the trigonometric functions, identify the suitably
restricted domain on which an inverse function can be defined.
3. Simplify and/or interpret the meaning of expressions involving the trigono-

metric and inverse trigonometric functions.
Trigonometric functions provides another opportunity to illustrate the

roles and properties of inverse functions. The inverse of a trigonometric
function leads to exchange in the roles of the dependent and independent
variables, as well as the the roles of the domain and range. Recall that geo-
metrically, an inverse function is obtained by reflecting the original function
about the line y = x. However, we must take care that the resulting graph rep-
resents a true function, i.e. satisfies all the properties required of a function.
The domains of sin(x) and cos(x) are both −∞ < x < ∞ while their ranges 23. What property of a function might
are −1 ≤ y ≤ 1. In the case of the function tan(x), the domain excludes values fail when we define an inverse
function?
±π/2 as well as angles 2nπ ± π/2 at which the function is undefined. The
24. Give an example of two different
range of tan(x) is −∞ < y < ∞. functions which are not one-to-one.
There is one difficulty in defining inverses for trigonometric functions: the 25. Why is the range of tan(x)
fact that these functions repeat their values in a cyclic pattern means that a −∞ < y < ∞? Why are some points
not in the domain?
given y value is obtained from many possible values of x. For example, the
values x = π/2, 5π/2, 7π/2, etc. all satisfy sin(x) = 1. We say that these
functions are not one-to-one. Geometrically, this means the graphs of the
trigonometric functions intersect a horizontal line in numerous places.
When these graphs are reflected about the line y = x, they would intersect
a vertical line in many places, and would fail to be functions: the function
would have multiple y values corresponding to the same value of x, which
is not allowed. See Appendix C where the inverse function for y = x2 is
discussed. ` The trigonometric and inverse
To avoid this difficulty, we restrict the domains of a trigonometric func- trigonometric functions are shown. Turn
the graphs on or off by clicking on the
tions to a portion of their graphs that does not repeat. To do so, we select an (grey) circles to the left of the formulae.
interval over which the given trigonometric function is one-to-one, i.e. over Notice that each function and its inverse
are reflections of one another about the
which there is a unique correspondence between values of x and values of y. line y = x. Also observe that the domain
We then define the corresponding inverse function, as described below. of the inverse functions is restricted, to
avoid multiple y values for a given x
Arcsine is the inverse of sine. The function y = sin(x) is one-to-one on the value.
interval −π/2 ≤ x ≤ π/2. We define the associated function y = Sin(x)

(shown in red on Figures 14.10(a) and (b)) by restricting the domain of
the sine function to −π/2 ≤ x ≤ π/2. On the given interval, we have
−1 ≤ Sin(x) ≤ 1. We define the inverse function, called arcsine
y = arcsin(x) −1 ≤ x ≤ 1
in the usual way, by reflection of Sin(x) through the line y = x as shown in

Figure 14.10(a).
Figure 14.10: (a) The original trigonometric

y y function, sin(x), in black, as well as the
portion restricted to a smaller domain,
y = Sin(x) Sin(x), in red. The red curve is shown
1 1 y=x again in part b. (b) Relationship between
the functions Sin(x), defined on −π/2 ≤
0.5
x ≤ π/2 (in red) and arcsin(x) defined on
−1 ≤ x ≤ 1 (in orange). Note that one is the
−6 −4 −2 2 4 6x −1.5 −1 −0.5 0.5 1 1.5x reflection of the other about the line y = x.
−0.5 y = Sin(x) The graphs in parts (a) and (b) are not on the
same scale.
−1 −1
y = sin(x) y = arcsin(x)
(a) (b)
To interpret this function, we note that arcsin(x) is “the angle (in radians)
whose sine is x”. In Figure 14.11, we show a triangle in which θ = arcsin(x). 1
x
This follows from the observation that the sine of theta, opposite over hy-
potenuse (sin θ = opp
hyp as reviewed in Appendix F) is x/1 = x. The length of θ
√
the other side of the triangle is then 1 − x2 by the Pythagorean theorem. √
√ √ 1 − x2
For example arcsin( 2/2) is the angle whose sine is 2/2, namely π/4
(we can also see this by checking the values of trigonometric functions of Figure 14.11: This triangle has been
standard angles shown in Table F.1) constructed so that θ is an angle such
that sin(θ ) = sx/1 = x. This means that
The functions sin(x) and arcsin(x), reverse (or “invert”) each other’s θ = arcsin(x).
effect, that is:
arcsin(sin(x)) = x for − π/2 ≤ x ≤ π/2,
sin(arcsin(x)) = x for − 1 ≤ x ≤ 1.
Note: the allowable values of x that can be “substituted in” are not exactly Mastered Material Check
the same for these two cases. In the first case, x is an angle whose sine we 26. What is the value of arcsin(sin(x))
compute first, and then reverse the procedure. In the second case, x is a for x = π/2 and for 5π/2?
number whose arc-sine is an angle. 27. What is the value of sin(arcsin(x))

for x = 1?
While we can evaluate arcsin(sin(x)) for any value of x, the result may not
28. Is arcsin(x) defined for x = π/2?
agree with the original value of x - unless we restrict attention to the interval
−π/2 ≤ x ≤ π/2.
Example 14.6 Let x = π. Compare x and arcsin(sin(x)).

Solution. When x = π we get that sin(x) = 0. Thus arcsin(sin(x)) =

arcsin(0) = 0, which is not the same as x = π. ♦
The other case also requires careful attention:
Example 14.7 Let x = 2. Compare x and sin(arcsin(x)).
Solution. Notice that x = 2 is outside of the interval −1 ≤ x ≤ 1. Thus

arcsin(2) is simply not defined, and so neither is sin(arcsin(x)). ♦
Indeed, care must be taken in handling the inverse trigonometric functions.
Inverse cosine. We cannot use the same interval (as that for sine) to restrict
the cosine function because cos(x) is an even function, symmetric about
the y axis (and not one-to-one on the interval π/2 ≤ x ≤ π/2). Instead, the
appropriate interval is 0 ≤ x ≤ π. Let y = Cos(x) = cos(x) for 0 ≤ x ≤ π be
Figure 14.12: (a) The original function

y y cos(x), is shown in black; the restricted
3 domain version, Cos(x) is shown in red.
y = Cos(x) The same red curve appears in part (b) on
1
a slightly different scale. (b) Relationship
2 y = arccos(x) y=x
0.5 between the functions Cos(x) (in red) and
arccos(x) (in orange). Note that one is the
1 reflection of the other about the line y = x.
−6 −4 −2 2 4 6x
y = Cos(x)
−0.5
−1 1 2 3x
−1
y = cos(x)
−1
(a) (b)
the restricted-domain version of cosine (red curve in Figure 14.12). On the

interval 0 ≤ x ≤ π, we have 1 ≥ Cos(x) ≥ −1. We define arccosine
y = arccos(x) −1 ≤ x ≤ 1
as the inverse of Cos(x), as shown in orange in Figure 14.12(b).

The meaning of the expression y = arccos(x) is “the angle (in radians)
whose cosine is x.” For example, arccos(0.5) = π/3 because π/3 is an angle
whose cosine is 1/2 = 0.5. The triangle in Figure 14.13, is constructed so 1 √
1 − x2
that θ = arccos(x). (This follows from the fact that cos(θ ) is adjacent over
hypotenuse.) The length of the third side of the triangle is obtained using the θ
Pythagorian theorem. x
The inverse relationship between the functions mean that
Figure 14.13: This triangle has been
constructed so that θ is an angle such
arccos(cos(x)) = x for 0 ≤ x ≤ π, that cos(θ ) = x/1 = x implying that
θ = arccos(x).
cos(arccos(x)) = x for − 1 ≤ x ≤ 1.
The same subtleties apply as in the case of arcsine.

Inverse tangent. The function y = tan(x) is one-to-one on an interval

π/2 < x < π/2. Unlike the case for Sin(x), we must exclude the endpoints,
where the function tan(x) is undefined. We therefore restrict the domain to
π/2 < x < π/2, that is, we define,
y = Tan(x) = tan(x) π/2 < x < π/2.
Again in contrast to the sine function, as x approaches either endpoint of

this interval, the value of Tan(x) approaches ±∞, i.e. −∞ < Tan(x) < ∞. This
means that the domain of the inverse function is −∞ < x < ∞, i.e. the inverse
function is defined for all real values of x. We define the inverse tan function:
y = arctan(x) − ∞ < x < ∞.
y y Figure 14.14: (a) The function tan(x), is

10 6 shown in black, and Tan(x) in red. The
same red curve is repeated in part b. (b)
4
5 y=x Relationship between the functions Tan(x)
y = T an(x) y = tan(x) 2 (in red) and arctan(x) (in orange). Note that
y = arctan(x) one is the reflection of the other about the
x line y = x.
−4 −2 2 4 −6 −4 −2 2 4 6 t
−2
−5
−4
y = T an(x)
−10 −6
√
(a) (b) 1 + x2
As before, we can understand the meaning of the inverse tan function, by θ

constructing a triangle in which θ = arctan(x), shown in Figure 14.15. 1
The inverse tangent “inverts” the effect of the tangent on the relevant
interval: Figure 14.15: This triangle has been
arctan(tan(x)) = x for − π/2 < x < π/2 constructed so that θ is an angle such
that tan(θ ) = x/1 = x. This means that
θ = arctan(x).
tan(arctan(x)) = x for −∞ < x < ∞
The same comments hold in this case.

A summary of the above inverse trigonometric functions, showing their
graphs on a single page is provided in Figure F.3 in Appendix F. Some of the
standard angles allow us to define precise values for the inverse trig functions.
A table of such standard values is given in the same Appendix (See Table F.2).
For other values of x, one has to calculate the decimal approximation of the
function using a scientific calculator.
Example 14.8 Simplify the following expressions:
a) arcsin(sin(π/4)),
b) arccos(sin(−π/6))
Solution.
a) arcsin(sin(π/4)) = π/4 since the functions are simple inverses of one

another on the domain −π/2 ≤ x ≤ π/2.
b) We evaluate this expression piece by piece: First, note that sin(−π/6) =

−1/2. Then arccos(sin(−π/6)) = arccos(−1/2) = 2π/3. The last
equality is obtained from Table F.2.
Example 14.9 Simplify the expressions:
a) tan(arcsin(x)),
b) cos(arctan(x)).
Solution. 29. Redraw and further label triangles to
aid in the solutions to Example 14.9.
a) Consider first the expression arcsin(x), and note that this represents an
30. Evaluste the expression
angle (call it θ ) whose sine is x, i.e. sin(θ ) = x. Refer to Figure 14.11 for a arccos(sin(π/4)).
sketch of a triangle in which this relationship holds. Now note that tan(θ ) 31. Simplify the expression
in this same triangle is the ratio of the opposite side to the adjacent side, sin(arccos(x)).
i.e.
x
tan(arcsin(x)) = √ .
1 − x2
b) Figure 14.15 shows a triangle that captures the relationship tan(θ ) = x or
θ = arctan(x). The cosine of this angle is the ratio of the adjacent side to
the hypotenuse, so that
1
cos(arctan(x)) = √ .
2
x +1
14.4 Summary
1. This chapter introduced and reviewed angles, cyclic processes, trigonomet-

ric, and periodic functions
2. amplitude, period, frequency, and phase were defined and identified with
graphical properties
3. the functions cosine and sine correspond to (x, y) coordinates of a point
moving around a circle of radius 1. tan(x) = sin(x)/ cos(x) is their ratio.
4. rhythmic processes can be approximated by a sine or a cosine graph once
the period, amplitude, mean, and phase shift are identified.
5. to define an inverse trigonometric function, the domain of the original trig
function has to be restricted to make it one-to-one (no repeated y values).

(a) electrocardiograms detecting the electrical activity of the heart;
(b) daylight hours fluctuating with period of one year;
(c) hormone levels that change on a daily rhythm; and
(d) phases of the moon, with a 29.5 day period.
1. What is the range of the function y = 8 sin(2t )?
2. Does a phase shift change the period of a trigonometric function?
3. If, in a 1-minute interval, a heart beats 50 times, what is the length of a heart beat cycle?
4. Using the following right angle triangle, determine:

√ α
a (a) tan(α );
b
(b) arccos(sin(θ ));
θ π/2
(c) cos(arccos(sin(α ).
0.3
Exercises
14.1. Radians, degrees and right tirangles. Convert the following expres-
sions in radians to degrees:
(a) π,
(b) 5π/3,
(c) 21π/23,
(d) 24π.
Convert the following expressions in degrees to radians:
(e) 100o ,
(f) 8o ,
(g) 450o ,
(h) 90o .
Using a right triangle, evaluate each of the following:
(i) cos(π/3),
(j) sin(π/4),
(k) tan(π/6).
14.2. Graphing functions. Graph the following functions over the indicated
ranges:
(a) y = x sin(x) for −2π < x < 2π,
(b) y = ex cos(x) for 0 < x < 4π.
14.3. Sketching the graph. Sketch the graph for each of the following
functions:
1 π
(a) y = sin 3 x − ,
2 4
(b) y = 2 − sin x,
(c) y = 3 cos 2x,

1 π
(d) y = 2 cos x+ .
2 4
14.4. Converting angles. The radian is an important unit associated with
angles. One revolution about a circle is equivalent to 360 degrees
or 2π radians.
(a) Convert the following angles (in degrees) to angles in radians.
Express each as multiples of π, not as decimal expansions:
(i) 45 degrees
(ii) 30 degrees
(iii) 60 degrees
(iv) 270 degrees.

(b) Find the sine and the cosine of each of the above angles.
14.5. Trigonometric functions and rhythmic functions. Find the appro-
priate trigonometric function to describe the following rhythmic
processes:
(a) Daily variations in the body temperature T (t ) of an individual
over a single day, with the maximum of 37.5o C at 8:00 am and a
minimum of 36.7o C 12 hours later.
(b) Sleep-wake cycles with peak wakefulness (W = 1) at 8:00 am and
8:00pm and peak sleepiness (W = 0) at 2:00pm and 2:00 am.
In both cases, express t as time in hours with t = 0 taken at 0:00 am.
14.6. Trigonometric functions and rhythmic functions. Find the appro-
priate trigonometric function to describe the following rhythmic
processes:
(a) The displacement S cm of a block on a spring from its equilibrium
position, with a maximum displacement 3 cm and minimum
displacement −3 cm, a period of √2π and at t = 0, S = 3.
g/l
(b) The vertical displacement y of a boat that is rocking up and down

on a lake, with y measured relative to the bottom of the lake. It has
a maximum displacement of 12 meters and a minimum of 8 meters,
a period of 3 seconds, and an initial displacement of 11 meters
when measurement was first started (i.e., t = 0).
14.7. Sunspot cycles. The number of sunspots (solar storms on the sun)
fluctuates with roughly 11-year cycles with a high of 120 and a low
of 0 sunspots detected. A peak of 120 sunspots was detected in the
year 2000.
Which of the following trigonometric functions could be used to
approximate this cycle?
2π

(A) N = 60 + 120 sin 11 (t − 2000) + 2
π
11

(B) N = 60 + 60 sin 2π (t + 2000)
11

(C) N = 60 + 60 cos 2π (t + 2000)
N = 60 + 60 sin 2π

(D) 11 (t − 2000)
N = 60 + 60 cos 2π

(E) 11 (t − 2000)
14.8. Inverse trigonometric functions. As seen in Section 14.3, the inverse

trigonometric function arctan(x) (also written tan−1 (x)) means the
angle θ where −π/2 < θ < π/2 whose tan is x. Thus cos(arctan(x)
(or cos(tan−1 (x)) is the cosine of that same angle. By using a right
√
triangle whose sides have length 1, x and 1 + x2 we can verify that
p
cos(arctan(x)) = 1/ 1 + x2 .
Use a similar geometric argument to arrive at a simplification of the

following functions:
(a) sin(arcsin(x)),
(b) tan(arcsin(x),
(c) sin(arccos(x).
14.9. Inverse trigonometric functions. The value of tan(arccos(x)) is
which of the following?
(A) 1 − x2 ,
(B) x ,
(C) 1 + x2 ,
√
2
(D) 1−xx ,
√
2
(E) 1x+x ,
14.10. Inverse trigonometric functions. The function y = tan(arctan(x)) has
which of the following for its domain and range?
(A) Domain 0 ≤ x ≤ π; Range −∞ ≤ y ≤ ∞
(B) Domain −∞ ≤ x ≤ ∞; Range −∞ ≤ y ≤ ∞
(C) Domain −π ≤ x ≤ π; Range −π ≤ y ≤ π;
(D) Domain −π/2 ≤ x ≤ π/2; Range −π/2 ≤ y ≤ π/2;
(E) Domain−∞ ≤ x ≤ ∞; Range 0 ≤ y ≤ π
14.11. Simplify trigonometric identity.
(a) Use a double-angle trigonometric identity to simplify the following
expression as much as possible:
y = cos(2 arcsin(x)).
(b) For what values of x is this simplification possible?

15
Cycles, periods, and rates of change
Having acquainted ourselves with properties of the trigonometric functions

and their inverses in Chapter 14, we are ready to compute their derivatives
and apply our results to understanding rates of change of these periodic
functions. We compute derivatives, and then use these results in a medley of
problems on optimization, related rates, and differential equations.
15.1 Derivatives of trigonometric functions
1. Use the definition of the derivative to calculate the derivatives of sin(x)

and cos(x).
2. Using the quotient rule, compute derivatives of tan(x), sec(x), csc(x),

and cot(x).
3. Using properties of the inverse trigonometric functions and implicit

differentiation, calculate derivatives of arcsin(x), arccos(x), and arctan(x).
Limits of trigonometric functions

In Chapter 3, we zoomed in on the graph of the sine function close to the
origin (Figure 3.2). Through this, we reasoned that
sin(x) ≈ x, for small x.
` Observe the behaviour of the two

Restated, with h replacing the variable x
limits, (15.1) and (15.2) by zooming
into the graphs for h close to zero.
sin(h) (Click on the + tab to zoom in.)
sin(h) ≈ h, for small h ⇒ ≈ 1 for small h
h
The smaller is h, the better this “tangent line” approximation becomes. In
more formal limit notation, we say that
sin(h)
lim = 1. (15.1)
h→0 h
(See Eqn. (3.1).) This is a very important limit, and we apply it directly in
computing the derivative of trigonometric functions using the definition of the
derivative.
A similar analysis of the graph of the cosine function, leads to a second Mastered Material Check
important limit: 1. Perform the calculation to verify
d cos(x)
cos(h) − 1 that dx = − sin(x).
lim = 0. (15.2)
h→0 h 2. Based on properties of the sine
We can now apply these to computing derivatives of both the sine and the function, estimate the value of sin(x)
for x = 0.01 radians and for x = 0.01
cosine functions. degrees.
3. Repeat this for estimates of cos(x)
Derivatives of sine, cosine, and other trigonometric functions for these two values of x.
Let y = f (x) = sin(x) be the function to differentiate, where x is now the

independent variable (previously t). We use the definition of the derivative to
compute the derivative of this function.
Example 15.1 (Derivative of sin(x)) Compute the derivative of y = sin(x)

using the definition of the derivative.
i Explanation of the calculation of the
Solution. We apply the definition of the derivative as follows: derivative of sin(x) using the definition
of the derivative.
f (x + h) − f (x )
f 0 (x ) = lim
h→0 h
d sin(x) sin(x + h) − sin(x)
= lim
dx h→0 h
sin(x) cos(h) + sin(h) cos(x) − sin(x)
= lim
h→0 h
cos(h) − 1 sin(h)

= lim sin(x) + cos(x)
h→0 h h
cos(h) − 1 sin(h)

= sin(x) lim + cos(x) lim
h→0 h h→0 h
= cos(x).
Observe that a trigonometric identity (for the sum of angles - see Eqn. (F.3))
and Limits (15.1) and (15.2) were used to obtain the final result. ♦
A similar calculation using the function cos(x) leads to the result
d cos(x)
= − sin(x).
dx
Note: the same two limits appear in this calculation, as well as the trigono-
metric identity Eqn. (F.4).
We can now calculate the derivative of the any of the other trigonometric
functions using the quotient rule.
Example 15.2 (Derivative of the function tan(x)) Compute the derivative

of y = tan(x) using the quotient rule. i Using the quotient rule to compute
the derivative of y = tan(x).
C Y C L E S , P E R I O D S , A N D R AT E S O F C H A N G E 327
Solution. We apply the quotient rule:
d tan(x) [sin(x)]0 cos(x) − [cos(x)]0 sin(x)

= .
dx cos2 (x)
Using the recently found derivatives for the sine and cosine, we have
4. Verify one or more of the derivatives
d tan(x) sin2 (x) + cos2 (x) of csc(x), sec(x) or cot(x) using the
= . quotient rule.
dx cos2 (x)
5. For what ranges of values of x and y
But the numerator of the above can be simplified using the trigonometric are the two statements y = arctan(x)
as tan(y) = x equivalent?
identity Eqn. (14.1), leading to
6. For what range(s) of values of y are
these two functions not inverses of
d tan(x) 1
= = sec2 (x). one another?
dx cos2 (x)
♦
Derivatives of the six trigonometric functions are given in Table 15.1. y = f (x ) f 0 (x )
The first three are frequently encountered in practical applications and worth sin(x) cos(x)
committing to memory. cos(x) − sin(x)
tan(x) sec2 (x)
csc(x) − csc(x) cot(x)
Featured Problem 15.1 (Lung volume) Breathing is a rhythmic process. sec(x) sec(x) tan(x)
The volume of air in the lungs can be modelled by a function of the form cot(x) − csc2 (x)
V (t ) = C + A sin(ωt + φ ), Table 15.1: Derivatives of the trigonometric

functions.
where V is the volume in millilitres (ml) and t is time in seconds. Suppose

that the minimum and maximum volumes are 1400 and 3400 ml, respectively,
and that the maximum rate of change of V is 1200 ml/sec. What is the period
of V (t )?
Derivatives of the inverse trigonometric functions

Implicit differentiation - introduced in Chapter 9 - can be used to deter-
mine the derivatives of the inverse trigonometric functions, explored in
Section 14.3. As an example, we demonstrate how to compute the derivative
of arctan(x). To do so, we need to recall that the derivative of the func-
tion tan(x) is sec2 (x). We also use the identity tan2 (x) + 1 = sec2 (x). (See
Eqn. (F.1).)
Let y = arctan(x). Then on the appropriate interval, we can replace this
relationship with the equivalent one:
tan(y) = x.
Differentiating implicitly with respect to x on both sides, we obtain

dy
sec2 (y) = 1,
dx
dy 1 1
= 2
= 2
.
dx sec (y) tan (y) + 1
y = f (x ) f 0 (x )
Now using again the relationship tan(y) = x, we obtain
arcsin(x) √1
1−x2
d arctan(x) 1
= 2 . arccos(x) −√ 1
dx x +1 1−x2
This expression is used frequently in integral calculus. The derivatives of the arctan(x) 1
x2 + 1
.
important inverse trigonometric functions are given in Table 15.2.
Table 15.2: Derivatives of the inverse
trigonometric functions.
15.2 Changing angles and related rates
1. Explain how the chain rule is applied to geometric problems with angles
that are change over time (“related rates”).
2. Given a description of the geometry and/or rate of change of angle or side

of a triangle, set up the mathematical problem and solve it using geometry
and/or properties of the trigonometric functions.
The examples in this section provide practice with chain rule applica-
tions based on trigonometric functions. We discuss a number of problems,
and show how the basic properties of these functions, together with some
geometry, are used to arrive at desired results.
Example 15.3 (A point on a circle) A point moves around the rim of a

circle of radius 1 so that the angle θ between the radius vector and the x axis
changes at a constant rate,
θ = ωt,
where t is time. Determine the rate of change of the x and y coordinates of
that point.
Solution. We have θ (t ), x(t ), and y(t ) all functions of t. (The geometry is

captured by Figure 14.4, but the angle has been renamed θ , and we consider
it to be time-dependent.) The fact that θ is proportional to t means that
dθ
= ω.
dt
The x and y coordinates of the point are related to the angle by
x(t ) = cos(θ (t )) = cos(ωt ),
y(t ) = sin(θ (t )) = sin(ωt ).

Then, by the chain rule,
dx d cos(θ ) dθ
= ,
dt dθ dt
dy d sin(θ ) dθ
= .
dt dθ dt
Performing the required calculations, we have
dx
= − sin(θ )ω,
dt
dy
= cos(θ )ω. 7. Write a concluding sentence for
dt Example 15.3.
♦ 8. If ω = 2 per min, what is the rate of
change of the x and y coordinates
Example 15.4 (Runners on a circular track) Two runners start at the when θ = π/2?
same position on a circular race track of length 400 meters. Joe Runner
takes 50 sec, while Michael Johnson takes 43.18 sec to complete the 400 me-
ter race. Find the rate of change of the angle formed between the two runners
and the center of the track, assuming that they run at a constant rate.
Solution. We are told that the track is 400 meters in length (total). However,
this information does not actually enter into the solution. Joe completes one
cycle around the track (2π radians) in 50 sec, while Michael completes a
cycle in 43.18 sec. This means that
• Joe has a period of T1 = 50 sec, and a frequency of ω1 = 2π/T1 = 2π/50

radians per sec.
• Michael has period is T2 = 43.18 sec, and a frequency of ω2 = 2π/T2 =

2π/43.18 radians per sec.
Let θJ , θM be the angles subtended between one of the runners and the
starting line. We take the x-axis as that starting line, by convention, as in
Figure 14.4. From this, we find that
dθJ 2π
= = 0.125 radians per sec,
dt 50
dθM 2π
= = 0.145 radians per sec.
dt 43.18
Thus, the angle between the runners, θM − θJ changes at the rate
d (θM − θJ ) dθM dθJ

= − = 0.145 − 0.125 = 0.02 radians per sec.
dt dt dt
♦
b
Example 15.5 (Simple law of cosines) The law of cosines applies to an c
arbitrary triangle, as reviewed in Appendix F (see Eqn (F.2)). Consider the θ
triangle shown in Figure 15.1. Suppose that the angle θ increases at a
constant rate, dθ /dt = k. If the sides a = 3, b = 4, are of constant length,
a
determine the rate of change of the length c opposite this angle at the instant
that c = 5. Figure 15.1: Law of cosines states that
c2 = a2 + b2 − 2ab cos(θ ).
Definitely needs a demo
Solution. Let a, b, c be the lengths of the three sides, with c the length of the
side opposite angle θ . The law of cosines states that
c2 = a2 + b2 − 2ab cos(θ ).
We identify the changing quantities by writing this relation in the form
c2 (t ) = a2 + b2 − 2ab cos(θ (t ))
so it is evident that only c and θ vary with time, while a, b remain constant.
Differentiating with respect to t and using the chain rule leads to:
dc d cos(θ ) dθ
2c = −2ab .
dt dθ dt
But d cos(θ )/dθ = − sin(θ ) and dθ /dt = k, so that
dc ab dθ ab
= − (− sin(θ )) = k · sin(θ ).
dt c dt c
At the instant in question, a = 3, b = 4, and c = 5, forming a Pythagorean
triangle in which the angle opposite c is θ = π/2. We can see this fact using Mastered Material Check
the law of cosines, and noting that 9. Verify that a triangle formed by
sides of length 3, 4 and 5 is a
c2 = a2 + b2 − 2ab cos(θ ), 25 = 9 + 16 − 24 cos(θ ). Pythagorean triangle.
10. Redo Example 15.5 with a triangle
This implies that 0 = −24 cos(θ ), cos(θ ) = 0 so that θ = π/2. Substituting in which a = 5, b = 12 and c is the
third side, at the instant when the
these into our result for the rate of change of the length c leads to
tringle so formed is Pythagorean.
dc ab 3·4 11. Write a concluding sentence for
= k= k. Example 15.5.
dt c 5
♦
Example 15.6 (Clocks) Find the rate of change of the angle between the
minute hand and hour hand on a clock.
Solution. We call θ1 the angle that the minute hand subtends with the x-axis
(horizontal direction) and θ2 the angle that the hour hand makes with the
same axis as depicted in Figure 15.2(a).
If our clock is working properly, each hand moves around at a con-
stant rate. The hour hand traces one complete revolution (2π radians) ev-
ery 12 hours, while the minute hand completes a revolution every hour. Both
hands move in a clockwise direction, which (by convention) is towards
negative angles. This means that
dθ1
= −2π radians per hour,
dt
dθ2 2π
=− radians per hour.
dt 12
The angle between the two hands is the difference of the two angles, i.e.
θ = θ1 − θ2 .
Figure 15.2: Figure for Examples 15.6 and

15.7; θ1 and θ2 are the angles the minute
and hour hands, respectively, form with
θ1 θ1 − θ2 the x-axis.
x x
θ2
(a) (b)
Thus, Mastered Material Check

dθ d dθ1 dθ2 2π 12. Highlight θ1 − θ2 on Figure 15.2(a).
= ( θ1 − θ2 ) = − = −2π + .
dt dt dt dt 12 13. What property of the derivative
allows us to simplify the
We find that the rate of change of the angle between the hands is subtraction dθ dθ2
dt − dt ?
1
14. Explain the significance of the

dθ 11 minus sign in the solution to
= −π . Example 15.7.
dt 6
15. Redo Examples 15.6 and 15.7 for a
♦ clock whose two hands have lengths
12 and 5 cm. (The longer hand is the
minute hand.)
Example 15.7 (Clocks, continued) Suppose that the length of the minute
hand is 4 cm and the length of the hour hand is 3 cm. At what rate is the
distance between the ends of the hands changing when it is 3:00 o’clock?
Solution. We use the law of cosines on the triangle shown in Figure 15.2.
Side lengths are a = 3, b = 4, and c(t ) opposite the angle θ (t ). From the
previous example, we have
dc ab dθ
= sin(θ ) .
dt c dt
At precisely 3:00 o’clock, the angle in question is θ = π/2 and it can also be
seen that the Pythagorean triangle abc leads to
c2 = a2 + b2 = 32 + 42 = 9 + 16 = 25
11
so that c = 5. We found from our previous analysis that dθ /dt = 6 π. Using
this information leads to:
dc 3·4 π 11 22
= sin − π = − π cm/hr.
dt 5 2 6 5
♦
15.3 The Zebra danio’s escape responses
1. Describe the geometry of a visual angle, and determine how that angle
changes as the distance to the viewed object (or the size of the object)
changes (an application of “related rates”).
2. Determine how the rate of change of the visual angle of a prey fish (zebra
danio) changes as a predator of a given size approaches it at some speed.
3. Explain the link between the rate of change of the visual angle and the
triggering of an escape response.
4. Using the results of the analysis, explain in words under what circum-
stances the prey does (or does not) manage to escape from its predator.
We consider an example involving trigonometry and related rates with a

biological application. We first consider the geometry on its own, and then
link it to the biology of predator avoidance and escape responses.
Visual angles
Example 15.8 (Visual angle) In the triangle shown in Figure 15.3, an object
of height s is moving towards an observer. Its distance from the observer
at some instant is labeled x(t ) and it approaches at some constant speed, v.
Determine the rate of change of the angle θ (t ) and how it depends on speed,
size, and distance of the object. Often θ is called a visual angle, since it
represents the angle that an image subtends on the retina of the observer.
` Demo of the changing visual angle.
Solution. The object approaches at some constant speed, v so that You can see how the angle depends on
the size of the object s. Notice that the
dx angle hardly changes when the object is
= −v. far away, and changes dramatically as
dt
the object gets closer.
where the minus sign means that the distance x is decreasing. Using the
trigonometric relations, we see that
s
tan(θ ) = . s
x
If the size, s, of the object is constant, then the changes with time imply that θ(t)
s x(t)
tan(θ (t )) = .
x(t )
Figure 15.3: A visual angle θ would change
We differentiate both sides of this equation with respect to t, and obtain
as the distance x decreases. The size s is
assumed constant. See Example 15.8.
d tan(θ ) dθ

d s
= ,
dθ dt dt x(t )
dθ 1 dx
sec2 (θ ) = −s 2 ,
dt x dt
so that
dθ 1 1 dx
= −s 2 .
dt sec (θ ) x2 dt
We can use the trigonometric identity
sec2 (θ ) = 1 + tan2 (θ )
to express our answer in terms of the size, s, the distance of the object, x and
the speed v:
s 2 x2 + s2
sec2 (θ ) = 1 + =
x x2
so
x2

dθ 1 dx s
= −s 2 2 = 2 2 v.
dt x +s x2 dt x +s
Thus, the rate of change of the visual angle is Mastered Material Check
16. Where did the minus sign go in the
sv/(x2 + s2 ). final result of dθ
dt for Example 15.8?
The angle thus changes at a rate proportional to the speed of the object. ♦
The dependence on the size of the object is more involved. This is ex-
plored next.
The Zebra danio and a looming predator

Visual angles are important to predator avoidance. We use the ideas of
Example 15.8 to consider a problem in biology, studied by Larry Dill, a
biologist at Simon Fraser University in Burnaby, BC.
The Zebra danio is a small tropical fish, with many predators. In order
to survive, it must sense danger quickly enough to be able to escape from
a pair of hungry jaws. At the same time, over-reacting to every perceived
motion could be counter-productive, wasting energy and time better spent on
foraging. Here, we investigate the visual basis of an escape response, based
on a hypothesis by Dill [Dill, 1974a,b].
Figure 15.5 shows the relation between the angle subtended at the danio’s
eye and the profile size S of an approaching predator, currently located at
distance x away. Suppose that the predator approaches at constant speed, v so
that the distance x is decreasing. Then
dx
= −v.
dt
Using this information and geometry, we characterize the rate of change of
the angle α.
Figure 15.4: A cartoon showing the visual

angle, α (t ) and how it changes as a predator
α x S
approaches its prey, the Zebra danio.
prey predator
α x S
Example 15.9 (Danio’s visual angle) Use the above geometry to express the x
α S
rate of change of angle α in terms of the size S and speed v of the approach-
ing predator, and its distance away, x.
Solution. If we consider the top half of the triangle shown in Figure 15.5 Figure 15.5: The geometry of the escape
we find a Pythagorean triangle identical to that of Example 15.8, but with response problem.
θ = α/2 and s = S/2. The side labeled x is identical in both pictures. Thus,
the trigonometric relation that holds is:
α (S/2)
tan = . (15.3)
2 x
Furthermore, based on the results of Example 15.8, we know that dα/dt can 17. Verify the calculations for the
be written as solution of Example 15.9.
dα S/2 Sv 18. Write a concluding sentence for
=2 2 2
v= 2 . (15.4) Example 15.9; address units.
dt x + (S/2) x + (S2 /4)
♦
Featured Problem 15.2 (Demo of looming predator) Set up an interactive

(Desmos) graph that represents an elliptical “predator” centered at x = a
that approaches its prey along the x axis. Draw the visual angle that this
predator creates at (0, 0). Assume that the equation of the predator is given
by the ellipse (x − a)2 + y2 /s2 = 1. By animating the parameter a, you should
be able to make the predator move. (Refer to Featured Problem 9.3 for the
geometry.)
Example 15.10 (Distance-dependence) Use the relationship in Eqn. (15.4)

to sketch a rough graph of the rate of change of the visual angle α versus the
distance x of the predator.
Solution. We are asked to sketch dα/dt versus x. Let us denote by f (x) the
function of x that we want to graph. Then from Eqn. (15.4),
Sv
f (x ) ≡ .
x2 + (S2 /4)
We first make three observations about this function. Reference sketching skills developed
throughout course? Particular chapters?
• When x = 0, (and the predator has reached its prey),
Sv 4v
f (0) = = .
0 + (S2 /4) S
This provides the “y-intercept” of the graph.
• Since x appears in the denominator, the function f (x) is always decreas-

ing.
f (x)
• For x → ∞ (the predator is very far away), the denominator is very large, so
f (x) → 0.
4v
S
The sketch of f (x) (the rate of change of the visual angle, dα/dt) versus
distance x of the predator is shown in Figure 15.6.
♦
x
When to escape? What sort of visual input should the danio respond to, if
it is to be efficient at avoiding the predator? In principle, we would like to Figure 15.6: The function f (x) plotted
consider a response that has the following features: against x. This graph shows that the rate
of change of visual angle dα/dt ≡ f (x) is
small when the distance to the predator x is
• If the predator is too far away, or moving slowly, (or moving further away), large.
it is likely harmless and should not trigger an escape response.
• If the predator is moving quickly towards the danio, it is likely a threat,

and should trigger the escape response.
Accordingly, the hypothesis proposed by Dill is that:
The escape response is triggered when the predator approaches so quickly,

that the rate of change of the visual angle is greater than some critical value.
Then the critical value, denoted Kcrit > 0, is a constant that depends on
the “skittishness of the prey or level of perceived danger of its environment.
Hence, the danio’s escape response is triggered when
dα
= Kcrit .
dt
♦
Example 15.11 (Finding the predator’s distance) How far away is the
predator when the escape response is triggered?
Solution. Rewrite the above condition using the dependence of dα/dt on the
geometric quantities in the problem. Then, we must solve for x in
Sv
= Kcrit . (15.5)
x2 + (S2 /4)
Figure 15.7(a) illustrates a geometric solution, showing the line y = Kcrit and
the curve y = Sv/(x2 + (S2 /4)) superimposed on the same graph.
Figure 15.7: The rate of change of the

visual angle dα/dt in two cases, when the
dα/dt dα/dt quantity 4v/S is above (a) and below (b)
some critical value.
Kcrit
4v 4v
S S
Kcrit
x x
xreact
(a) (b)
The value x = xreact is the distance of the predator at the instant that the
escape response is triggered, termed the reaction distance. Solving for x,
(Exercise 11) leads to:
s s
S2

Sv v S
xreact = − = S − . (15.6)
Kcrit 4 Kcrit 4
Example 15.12 (Lunch) Interpret the reaction distance xreact . Are there ever
cases in which the prey does not notice a predator in time to escape?
Solution. Figure 15.7(b) illustrates a case where Kcrit = Sv/(x2 + (S2 /4))
is never satisfied. This could happen if either the danio has a very high
threshold of alert (large Kcrit ), or if 4v/S is too low. That happens either if S
is very large (big predator) or if v is small (predator slowly “stalks” its prey).
From this scenario, we find that in some situations, the fate of the danio
would be sealed in the jaws of its pursuer. ♦
Large slow predators beat danio’s escape response. Notice the reaction
distance, xreact , of the danio with reaction threshold Kcrit is largest for certain
sizes of predators. In Figure 15.8, we plot the reaction distance xreact versus
the predator size S. We see that for very small predators (S ≈ 0) or large
predators (S ≈ 4v/Kcrit ) the distance at which escape response is triggered is
very small. This means that the danio may miss noticing such predators until
they are too close for a comfortable escape, resulting in calamity.
On the other hand, some predators are detected when they are very far
away- they have a large xreact . This is explored in Example 15.13.
Figure 15.8: The reaction distance xreact is

10 8 shown as a function of the predator size S in
(a) and as a function of the predator speed v
8 in (b). A very small or very large predator
6 may fail to be detected until it is too close to
6 escape. A very slow predator might never be
xreact
xreact
4 detected.
4
2
2
19. Which panel of Figure 15.8 would
10 20 30 40 2 4 6 8 10
you use to determine the distance at
predator size S predator speed v which a danio would react to a
(a) (b) predator of fixed size?
20. According to Figure 15.8 what size
predator is easiest to detect farthest
away? what size is undetectable
until it is too close?
Example 15.13 (Bad design for a predator) Some predators are more
21. According to Figure 15.8 what is the
easily detected than others. Find the size of predator for which the reaction slowest moving predator that is just
distance is maximal, and interpret your finding. barely detected? How far away is a
predator detected if its speed is 10
distance units per unit time?
Solution. We solve this problem using differentiation (Exercise 12) and
find that xreact has a critical point at S = 2v/Kcrit . From Figure 15.8(a), we
see that this critical point is a local maximum. We can also reason based on
Eqn. (15.6): xreact cannot be negative. However, we see that xreact = 0 at S = 0
and at S = 4v/Kcrit . Hence, xreact has a local maximum for some predator
size between these two values. In short, a predator of size S = 2v/Kcrit would
be detected as far away as possible (largest possible xreact ), giving the prey a
good chance to escape. ♦
Observe that at sizes S > 4v/Kcrit , the reaction distance is not defined. We
also see this Figure 15.7(b): when Kcrit > 4v/S, the straight line and the curve
fail to intersect, and there is no solution.
Figure 15.8(b) illustrates the dependence of the reaction distance xreact on
the speed v of the predator. We find that for small values of v, xreact is not
defined: the danio would not notice the threat posed by predators that swim
very slowly. See Exercise 13 for the largest velocity that fails to trigger the
escape response.
Alternate approach involving inverse trig functions
Note: this section is optional and can be skipped or left for independent
study.
In Section 15.3, we studied the escape response of the Zebra danio and
showed that the connection between the visual angle and distance to predator
satisfies α (S/2)
tan = . (15.7)
2 x
We also computed the rate of change of the visual angle per unit time using
implicit differentiation and related rates. Here we illustrate an alternate
approach using inverse trigonometric functions. (This provides practice with
differentiation of inverse trigonometric functions.)
Example 15.14 Use the inverse function arctan to restate the angle α in
Eqn. 15.7 as a function of x. Then differentiate that function using the chain
rule to compute dα/dt.
Solution. Restate the relationship using the function arctan:

α S
= arctan .
2 2x
Both the angle α and the distance x change with time; we indicate this by
writing
S
α (t ) = 2 arctan .
2x(t )
Applying the chain rule, let u = S/2x. Recall that S is a constant, so α (u) =
2 arctan(u). Using the derivative of the inverse trigonometric function,
d arctan(u) 1
= 2
du u +1
and the chain rule lead to
dα (t ) d arctan(u) du dx

1 S
= = 2 − 2 (−v).
dt du dx dt u +1 2x (t ) Mastered Material Check
22. Verify the simplification to the final
Simplifying leads to the same result,
solution in Example 15.14.
dα Sv
= 2 .
dt x + (S2 /4)
This rate of change of the visual angle agrees with Example 15.8. ♦
15.4 Summary
1. We used two limits

sin(h) cos(h) − 1
lim = 1, lim =0
h→0 h h→0 h
to computed the derivatives of two trigonometric functions
d sin(x) d cos(x)
= cos(x), = − sin(x)
dx dx
2. Implicit differentiation was used to compute the derivatives of the inverse
3. Trigonometric related rates, and implicit differentiation were discussed.

(a) runners on a circular track (changing angle between them);
(b) clock hands (changing angle between hands); and
(c) Zebra danio’s escape response (changing visual angle).
1. What is the derivative of y = cos(x) + sin(x)?
2. Rewrite the reaction distance formula twice: first clearly identifying the size, and second the speed, of
the predator as the independent variable.
3. Consider the following graph in which xreact is the dependent variable:

6
4
xreact
2 4 6 8 10 12 14
Is size or speed the independent variable here? Why?

Exercises
15.1. First derivatives. Calculate the first derivative for the following
functions:
(a) y = sin x2 ,
(b) y = sin2 x,
√
(c) y = cot2 3 x,
(d) y = sec(x − 3x2 ),
(e) y = 2x3 tan x,
x
(f) y = cos x ,
(g) y = x cos x,
2 1
(h) y = e− sin x ,
(i) y = (2 tan 3x + 3 cos x)2 ,
(j) y = cos(sin x) + cos x sin x.
15.2. Derivatives. Take the derivative of the following functions.
(a) f (x) = cos(ln(x4 + 5x2 + 3)),
p
(b) f (x) = sin( cos2 (x) + x3 ),
(c) f (x) = 2x3 + log3 (x),
(d) f (x) = (x2 ex + tan(3x))4 ,
q
(e) f (x) = x2 sin3 (x) + cos3 (x).
15.3. Point moving on a circle. A point is moving on the perimeter of a
circle of radius 1 at the rate of 0.1 radians per second.
(a) How fast is its x coordinate changing when x = 0.5?
(b) How fast is its y coordinate changing at that time?
15.4. Graphing trigonometric functions. The derivatives of the two
important trig functions are [sin(x)]0 = cos(x) and [cos(x)]0 = − sin(x).
Use these derivatives to answer the following questions.
Let f (x) = sin(x) + cos(x), 0 ≤ x ≤ 2π
(a) Find all intervals where f (x) is increasing.
(b) Find all intervals where f (x) is concave up.
(c) Locate all inflection points.
(d) Graph f (x).
π
15.5. Tangent lines. Find all points on the graph of y = tan(2x), − < x <
4
π
, where the slope of the tangent line is 4.
4
15.6. Bird formation. A “V” shaped formation of birds forms a symmetric

structure in which the distance from the leader to the last birds in the
V is r = 10m, the distance between those trailing birds is D = 6m and
the angle formed by the V is θ , as shown in Figure 15.9.
Suppose that the shape is gradually changing: the trailing birds
start to get closer so that their distance apart shrinks at a constant
rate dD/dt = −0.2m/min while maintaining the same distance from
the leader. Assume that the structure is always in the shape of a V as
the other birds adjust their positions to stay aligned in the flock.
What is the rate of change of the angle θ ?
Figure 15.9: Figure for Exercise 6; bird

r flock formation.
θ D
15.7. Hot air balloon. A hot air balloon on the ground is 200 meters away
from an observer. It starts rising vertically at a rate of 50 meters
per minute. Find the rate of change of the angle of elevation of the
observer when the balloon is 200 meters above the ground.
15.8. Cannon-ball. A cannon-ball fired by a cannon at ground level at
angle θ to the horizon (0 ≤ θ ≤ π/2) travels a horizontal distance
(called the range, R) given by the formula below:
1 2
R= v sin θ cos θ .
16 0
Here v0 > 0, the initial velocity of the cannon-ball, is a fixed constant
and air resistance is neglected (see Figure 15.10).
What is the maximum possible range?
15.9. Leaning ladder. A ladder of length L is leaning against a wall so that θ
its point of contact with the ground is a distance x from the wall, and R
its point of contact with the wall is at height y. The ladder slips away
from the wall at a constant rate C. Figure 15.10: Figure for Exercise 8; cannon-
ball trajectory.
(a) Find an expression for the rate of change of the height y.
(b) Find an expression for the rate of change of the angle θ formed
between the ladder and the wall.
15.10. Cycloid curve. A wheel of radius 1 meter rolls on a flat surface with-
out slipping. The wheel moves from left to right, rotating clockwise at
a constant rate of 2 revolutions per second.
Stuck to the rim of the wheel is a piece of gum, (labeled G); as the
wheel rolls along, the gum follows a path shown by the wide arc
(called a “cycloid curve”) in Figure 15.11. The (x, y) coordinates
of the gum (G) are related to the wheel’s angle of rotation θ by the
formulae
x = θ − sin θ ,
y = 1 − cos θ ,
where 0 ≤ θ ≤ 2π.
(a) How fast is the gum moving horizontally at the instant that it
reaches its highest point?
(b) How fast is it moving vertically at that same instant?
Figure 15.11: Figure for Exercise 10; a

cycloid curve.
y
x = θ − sin θ
G y = 1 − cos θ
1
y θ
x
x θ 2π
15.11. Zebra danio’s reaction distance. Solve Eqn. (15.5) for x and show
that you get the reaction distance x ≡ xreact given in Eqn (15.6).
15.12. Bad design for a predator. Some predators are more easily detected
than others. Use Eqn. (15.6) to find the size of predator for which the
reaction distance is maximal.
15.13. Sneaking up on the prey.
(a) Use Eqn. (15.6) to show that a predator moving “slowly enough”
can sneak up on the prey without being detected.
(b) What is the largest velocity for which a predator of size S is not
detected by a prey that responds to a visual sighting when the rate
of change of the visual angle exceeds the threshold Kcrit ?
15.14. Inverse trigonometric derivatives. Find the first derivative of the

following functions.
1
(a) y = arcsin x 3 ,
1
(b) y = (arcsin x) 3 ,
(c) θ = arctan(2r + 1),
(d) y = x arcsec 1x ,
√
(e) y = ax a2 − x2 − arcsin ax , a > 0,
(f) y = arccos 1+2tt 2 .
15.15. Rotating wheel. In Figure 15.12, the point P is connected to the P
3 5
point O by a rod 3 cm long. The wheel rotates around O in the clock-
Q
wise direction at a constant speed, making 5 revolutions per second. O
The point Q, which is connected to the point P by a rod 5 cm long,
moves along the horizontal line through O.
How fast and in what direction is Q moving when P lies directly
Figure 15.12: Figure for Exercise 15;
above O? rotating wheel.
Note: recall the law of cosines: c2 = a2 + b2 − 2ab cos θ .
15.16. Sailing ship. A ship sails away from a harbor at a constant speed v.
The total height of the ship including its mast is h. See Figure 15.13.
(a) At what distance away does the ship disappear below the horizon? h
d
(b) At what rate does the top of the mast appear to drop toward the
horizon just before this?
Note: in ancient times this effect lead people to conjecture that
R θ
the earth is round (radius R), a fact which you need to take into
account.
dy
15.17. Implicit differentiation. Find dx using implicit differentiation.
(a) y = 2 tan(2x + y),
(b) sin y = −2 cos x,
(c) x sin y + y sin x = 1.
Figure 15.13: Figure for Exercise 16; ship
15.18. Equation of a tangent line. Use implicit differentiation to find the sailing away.
equation of the tangent line to the following curve at the point (1, 1):
x sin(xy − y2 ) = x2 − 1
15.19. Implicit differentiation and arcsin. The function y = arcsin(ax)

is a so-called inverse trigonometric function. It expresses the same
relationship as does the equation ax = sin(y).
Note: this function is defined only for values of x between 1/a
and −1/a.
Use implicit differentiation to find y0 .
Figure 15.14: Figure for Exercise 20;

distance from a window.
1.5 window
θ 0.1
α
0
x
15.20. Best view.Your room has a window whose height is 1.5m. The bottom
edge of the window is 10cm above your eye level, as depicted in
Figure 15.14. How far away from the window should you stand to get
the best view?
Note: “best view” means the largest visual angle, i.e. angle between
the lines of sight to the bottom and to the top of the window.
fireworks
15.21. Fireworks. You are directly below English Bay during a summer
fireworks event and looking straight up. A single fireworks explo-
sion occurs directly overhead at a height of 500m as depicted in
Figure 15.15. The rate of change of the radius of the flare is 100m/sec.
500
θ
Assuming that the flare is a circular disk parallel to the ground (with
its centre directly overhead), what is the rate of change of the visual
O
angle at the eye of an observer on the ground at the instant that the
radius of the disk is r = 100 meters? Figure 15.15: Figure for Exercise 21;
fireworks overhead.
Note: the visual angle is the angle between the vertical direction and
the line between the edge of the disk and the observer.
15.22. Differential equations and their solutions. Match the differential
equations given in parts (i-iv) with the functions in (a-f) which are
solutions for them.
Differential equations:
(i) d 2 y/dt 2 = 4y
(ii) d 2 y/dt 2 = −4y
(iii) dy/dt = 4y
(iv) dy/dt = −4y
Solutions:
(a) y(t ) = 4 cos(t )
(b) y(t ) = 2 cos(2t )
(c) y(t ) = 4e−2t

(d) y(t ) = 5e2t
(e) y(t ) = sin(2t ) − cos(2t ),
(f) y(t ) = 2e−4t .
Note: each differential equation may have more than one solution
15.23. Periodic motion.
(a) Show that the function y(t ) = A cos(wt ) satisfies the differential
equation
d2y
= −w2 y
dt 2
where w > 0 is a constant, and A is an arbitrary constant.
Note: w corresponds to the frequency and A to the amplitude of an
oscillation represented by the cosine function.
(b) It can be shown using Newton’s laws of motion that the motion of a
pendulum is governed by a differential equation of the form
d2y g
= − sin(y),
dt 2 L
where L is the length of the string, g is the acceleration due to
gravity (both positive constants), and y(t ) is displacement of the
pendulum from the vertical.
What property of the sine function is used when this equation is
approximated by the Linear Pendulum Equation:
d2y g
2
= − y.
dt L
(c) Based on this Linear Pendulum Equation, what function would
represent the oscillations? What would be the frequency of the
oscillations?
(d) What happens to the frequency of the oscillations if the length of
the string is doubled?
15.24. Jack and Jill. Jack and Jill have an on-again off-again love affair. The
sum of their love for one another is given by the function
y(t ) = sin(2t ) + cos(2t ).
(a) Find the times when their total love is at a maximum.

(b) Find the times when they dislike each other the most.
15.25. Differential equations and critical points. Let
y = f (t ) = e−t sint, − ∞ < t < ∞.
(a) Show that y satisfies the differential equation y00 + 2y0 + 2y = 0.

(b) Find all critical points of f (t ).
16
Additional exercises
Exercises
16.1. Multiple choice.
(1) The equation of the tangent line to the function y = f (x) at the
point x0 is
(a) y = f 0 (x0 ) + f (x0 )(x − x0 )
(b) y = x0 + f (x0 )/ f 0 (x0 )
(c) y = f (x) − f 0 (x)(x − x0 )
(d) y = f (x0 ) + f 0 (x0 )(x − x0 )
(e) y = f (x0 ) − f 0 (x0 )(x − x0 )
(2) The functions f (x) = x2 and g(x) = x3 are equal at x = 0 and
at x = 1. Between x = 0 and at x = 1, for what value of x are their
graphs furthest apart?
(a) x = 1/2
(b) x = 2/3
(c) x = 1/3
(d) x = 1/4
(e) x = 3/4
(3) Consider a point in the first quadrant on the hyperbola x2 − y2 = 1
with x = 2. The slope of the tangent line at that point is
√
(a) 2/ 3
√
(b) 2/ 5
√
(c) 1/ 3
√
(d) 5/2
(e) 2/3
(4) For a, b > 0, solving the equation ln(x) = 2 ln(a) − 3 ln(b) for x
leads to
(a) x = e2a−3b
(b) x = 2a − 3b
(c) x = a2 /b3
(d) x = a2 b3
(e) x = (a/b)6
(5) The function y = f (x) = arctan(x) − (x/2) has local maXima (LX),
local minima (LM) and inflection points(IP) as follows:
(a) LX: x = 1, LM: x = −1, IP: x = 0.
(b) LX: x = −1, LM: x = 1, IP: x = 0.
(c) LX: x = −1, LM: x = 1, IP: none.
√ √
(d) LX: x = 3, LM: x = − 3, IP: x = 0.
√ √
(e) LX: x = − 3, LM: x = 3, IP: x = 0.
(6) Consider the function y = f (x) = 3e−2x − 5e−4x .
(a) The function has a local maximum at x = (1/2) ln(10/3).
(b) The function has a local minimum at x = (1/2) ln(10/3).
(c) The function has a local maximum at x = (−1/2) ln(3/5).
(d) The function has a local minimum at x = (1/2) ln(3/5).
(e) The function has a local maximum at x = (−1/2) ln(3/20).
(7) Let m1 be the slope of the function y = 3x at the point x = 0 and
let m2 be the slope of the function y = log3 x at x = 1 Then
(a) m1 = ln(3)m2
(b) m1 = m2
(c) m1 = −m2
(d) m1 = 1/m2
(e) m1 = m2 / ln(3)
(8) Consider the curve whose equation is x4 + y4 + 3xy = 5. The slope
of the tangent line, dy/dx, at the point (1, 1) is
(a) 1
(b) -1
(c) 0
(d) -4/7
(e) 1/7
(9) Two kinds of bacteria are found in a sample of tainted food. It is
found that the population size of type 1, N1 and of type 2, N2 satisfy
the equations
dN1
dt = −0.2N1 , N1 (0) = 1000,
dN2
dt = 0.8N2 , N2 (0) = 10.
A D D I T I O NA L E X E R C I S E S 349
Then the population sizes are equal N1 = N2 at the following time:

(a) t = ln (40)
(b) t = ln (60)
(c) t = ln (80)
(d) t = ln (90)
(e) t = ln (100)
(10) In a conical pile of sand the ratio of the height to the base radius is
always r/h = 3. If the volume is increasing at rate 3 m3 /min, how Formula.
Note that the volume of a cone with
fast (in m/min) is the height changing when h = 2m?
height h and radius r is V = (π/3)r2 h.)
(a) 1/(12π )
(b) (1/π )1/3
(c) 27/(4π )
(d) 1/(4π )
(e) 1/(36π ) y
(11) Shown in Figure 16.1 is a function and its tangent line at x = x0 . The
y = f (x)
tangent line intersects the x axis at the point x = x1 . Based on this
figure, the coordinate of the point x1 is
f ( x0 )
(a) x1 = x0 + f 0 ( x0 )
,
(b) x1 = x0 − 0
f (x0 )(x − x0 ),
f 0 (x1 )
(c) x1 = x0 − f (x1 ) x
f 0 ( x1 ) x1 x0
(d) x1 = x0 + f ( x1 )
,
f (x0 )
(e) x1 = x0 − f 0 ( x0 ) Figure 16.1: Graph for Exercise 1; a
function and its tangent line.
(12) Consider Euler’s method. For the differential equation and initial
condition
dy
= (2 − y), y(0) = 1
dt
using one time step of size ∆t = 0.1 leads to which value of the
solution at time t = 0.1?
(a) y(0.1) = 2,
(b) y(0.1) = 2.1,
(c) y(0.1) = 2.2,
(d) y(0.1) = 1.2,
(e) y(0.1) = 1.1,
(13) Consider the function y = cos(x) and its tangent line to this function
at the point x = π/2. Using that tangent line as a linear approxima-
tion of the function would lead to
(a) Overestimating the value of the actual function for any nearby x.
(b) Underestimating the value of the actual function for any
nearby x.
(c) Overestimating the function when x > π/2 and underestimating

the function when x < π/2.
(d) Overestimating the function when x < π/2 and underestimating
the function when x > π/2.
(e) Overestimating the function when x < 0 and underestimating the
function when x > 0.
16.2. Related Rates. Two spherical balloons are connected so that one Formula.
Note that the volume of a sphere of
inflates as the other deflates, the sum of their volumes remaining
radius r is V = (4/3)πr3 .
constant. When the first balloon has radius 10 cm and its radius is
increasing at 3 cm/sec, the second balloon has radius 20 cm. What is
the rate of change of the radius of the second balloon?
16.3. Particle velocity. A particle is moving along the x axis so that its
distance from the origin at time t is given by
x(t ) = (t + 2)3 + λt
where λ is a constant
(a) Determine the velocity v(t ) and the acceleration a(t ).
(b) Determine the minimum velocity over all time.
16.4. Motion. A particle’s motion is described by y(t ) = t 3 − 6t 2 + 9t
where y(t ) is the displacement (in metres) t is time (in seconds)
and 0 ≤ t ≤ 4 seconds.
(a) During this time interval, when is the particle furthest from its
initial position ?
(b) During this time interval, what is the greatest speed of the particle?
(c) What is the total distance (including both forward and backward
directions) that the particle has travelled during this time interval?
16.5. Falling object. Consider an object thrown upwards with initial veloc-
ity v0 > 0 and initial height h0 > 0. Then the height of the object at
time t is given by
1
y = f (t ) = − gt 2 + v0t + h0 .
2
Find critical points of f (t ) and use both the second and first derivative
tests to establish that this is a local maximum.
16.6. Linear approximation. Find a linear approximation to the func-
tion y = x2 at the point whose x coordinate is x = 2. Use your result to
approximate the value of (2.0001)2 .
16.7. HIV virus. Initially, a patient has 1000 copies of the virus. How long
does it take until the HIV infection is detectable? Assume that the
number of virus particles y grows according to the equation
dy
= 0.05y
dt
where t is time in days, and that the smallest detectable viral load
is 350, 000 particles. Leave your answer in terms of logarithms.
16.8. Fish generations. In Fish River, the number of salmon (in thousands),
x, in a given year is linked to the number of salmon (in thousands), y,
in the following year by the function
y = Axe−bx
where A, b > 0 are constants.

(a) For what number of salmon is there no change in the number from
one year to the next?
(b) Find the number of salmon that would yield the largest number of
salmon in the following year.
16.9. Polynomial. Find a polynomial of third degree that has a local maxi-
mum a x = 1, a zero and an inflection point at x = 0, and goes through
the point (1, 2) (hint: assume p(x) = ax3 + bx2 + cx + d and find the
values of a, b, c, d).
16.10. Critical points.
(a) Find critical points for the function y = ex (1 − ln(x)) for 0.1 ≤ x ≤ 2
and classify their types.
(b) The function y = ln(x) − ex has a critical point in the interval 0.1 ≤
x ≤ 2. It is not possible to solve for the value of x at that point, but it
is possible to find out what kind of critical point that is. Determine
whether that point is a local maximum, minimum, or inflection
point.
16.11. Lennard-Jones potential. The Lennard-Jones potential, V (x) is the
potential energy associated with two uncharged molecules a distance x
apart, and is given by the formula
a b
V (x ) = −
x12 x6
where a, b > 0. Molecules would tend to adjust their separation
distance so as to minimize this potential. Find any local maxima or
minima of this potential. Find the distance between the molecules, x,
at which V (x) is minimized and use the second derivative test to verify
that this is a local minimum.
16.12. Rectangle inscribed in a circle. Find the dimensions of the largest
rectangle that can fit exactly into a circle whose radius is r.
16.13. Race track. Figure 16.2 shows a 1 km race track with circular ends.
Find the values of x and y that maximize the area of the rectangle.
16.14. Leaf shape. Now suppose that Figure 16.2 shows the shape of a
leaf of some plant. If the plant grows so that x increases at the
rate 2 cm/year and y increases at the rate 1 cm/year, at what rate is
the leaf’s entire area increasing?
Figure 16.2: This shape is investigated in

both Exercise 13 and 14; maximizing area or
a racetrack, rate of growth of a leaf.
y
16.15. Shape of E. coli. A cell of the bacterium E. coli has the shape of
a cylinder with two hemispherical caps, as shown in Figure 16.3.
Consider this shape, with h the height of the cylinder, and r the radius
of the cylinder and hemispheres.
Figure 16.3: Shape of the cells described in

Exercises 15 and 16
r
(a) Find the values of r and h that lead to the largest volume for a fixed Formula.
Note that a hemisphere of radius r has
constant surface area, S= constant.
volume V = (2/3)πr3 and surface
(b) Describe or sketch the shape you found in (a). area S = 2πr2 .
For a cylinder: V = πr2 h and S = 2πrh.
(c) A typical E. coli cell has h = 1µm and r = 0.5µm. Based on your
results in (a) and (b), would you agree that E. coli has a shape
that maximizes its volume for a fixed surface area? (Explain your
answer).
16.16. Changing cell shape. If the cell shown in Figure 16.3 is growing so
that the height increases twice as fast as the radius, and the radius
is growing at 1 µm per day, at what rate does the volume of the cell
increase? Leave your answer in terms of the height and radius of the
cell.
16.17. Minima and Maxima.
(a) Consider the polynomial y = 4x5 − 15x4 . Find all local minima
maxima, and inflection points for this function.
(b) Find the global minimum and maximum for this function on the
interval [−1, 1].
16.18. Minima and Maxima. Consider the polynomial y = −x5 − x4 + 3x3 .
Use calculus to find all local minima maxima, and inflection points for
this function.
16.19. Growth of vine. A vine grows up a tree in the form of a helix as
shown on the left in Figure 16.4. If the length of the vine increases at
a constant rate α cm/day, at what rate is the height of its growing tip
increasing? Assume that the radius of the tree is r and the pitch of the
helix (i.e. height increase for each complete turn of the helix) is p, a
positive constant. Note that the right panel in Figure 16.4 shows the
unwrapped cylinder, with the vine’s location along it.
Figure 16.4: Growth of a vine in the shape

of a spiral for Exercise 19.
16.20. Newton’s law of cooling. Newton’s Law of cooling leads to a differ-

ential equation that predicts the temperature T (t ) of an object whose
initial temperature is T0 in an environment whose temperature is E.
The predicted temperature is given by T (t ) = E + (T0 − E )e−kt where t
is time and k is a constant. Shown in Figure 16.5 is some data points
plotted as ln(T (t ) − E ) versus time in minutes. The ambient tempera-
ture was E = 22◦ C. Also shown on the graph is the line that best fits
those 11 points. Find the value of the constant k.
ln(T (t) − E Figure 16.5: Figure for Exercise 20;

Newton’s law of cooling.
4
3.5
Best fit line
5 10 15 20 25 30
time in minutes
16.21. Blood alcohol. Blood alcohol level (BAL), the amount of alcohol
in your blood stream (here represented by B(t ), is measured in mil-
ligrams of alcohol per 10 millilitres of blood. At the end of a party
(time t = 0), a drinker is found to have B(0) = 0.08 (the legal level for
driving impairment), and after that time, B(t ) satisfies the differential
equation
dB
= −kB, k > 0
dt
where k is a constant that represents the rate of removal of alcohol
form the blood stream by the liver.
(a) If the drinker had waited for 3 hrs before driving (until = 3), his
BAL would have dropped to 0.04. Determine the value of the rate
constant k (specifying appropriate units) for this drinker.
(b) According to the model, how much longer would it take for the
BAL to drop to 0.01?
16.22. Population with immigration. An island has a bird population of
density P(t ). New birds arrive continually with a constant colonization
rate C birds per day. Each bird also has a constant probability per day,
γ, of leaving the island. At time t = 0 the bird population is P(0) = P0
(a) Write a differential equation that describes the rate of change of the
bird population on the island.
(b) Find the steady state of that equation and interpret this in terms of
the bird population.
(c) Give the solution of the differential equation you found in (b) and
show that it satisfies the following two properties:
(i) the initial condition,
(ii) as t → ∞ it approaches the steady state you found in (b).
(d) If the island has no birds on it at time t = 0, how long would it take
for the bird population to grow to 80% of the steady state value?
16.23. Learning.
(a) It takes you 1 hrs (total) to travel to and from SFU every day to
study Philosophy 101. The amount of new learning (in arbitrary
units) that you can get by spending t hours at the university is given
approximately by
10t
LP (t ) = .
9+t
How long should you stay at SFU on a given day if you want to
maximize your learning per time spent?
Note: time spent includes travel time.
(b) If you take Math 10000 instead of Philosophy, your learning at
time t is
LM (t ) = t 2 .
How long should you stay at SFU to maximize your learning in that
case?
16.24. Learning and forgetting. Knowledge can be acquired by studying,
but it is forgotten over time A simple model for learning represents the
amount of knowledge, y(t ), that a person has at time t (in years) by a
dy
= S − f y,
dt
where S ≥ 0 is the rate of studying and f ≥ 0 is the rate of forgetting.
We assume that S and f are constants that are different for each
person.
Note: your answers to the following questions will contain constants
such as S or f .
(a) Mary never forgets anything. What does this imply about the
constants S and f ? Mary starts studying in school at time t = 0
with no knowledge at all. How much knowledge does she have
after 4 years (i.e. at t = 4)?
(b) Tom learned so much in preschool that his knowledge when en-
tering school at time t = 0 is y = 100. However, once Tom is in
school, he stops studying completely. What does this imply about
the constants S and f ? How long does it take him to forget 75 % of
what he knew?
(c) Jane studies at the rate of 10 units per year and forgets at rate
of 0.2 per year. Sketch a “direction field” (“slope field”) for the
differential equation describing Jane’s knowledge. Add a few
curves y(t ) to show how Jane’s knowledge changes with time.
16.25. Least cost. A rectangular plot of land has dimensions L by D as de- C
picted in Figure 16.6. A pipe is to be built joining points A and C. The y
pipe can be above ground along the border of the plot (Section AB),
L
but has to be buried underground along the segment BC. The cost per B
unit length of the underground portion is 3 times that of the cost of the
above ground portion. Determine the distance y so that the cost of the
pipe is as low as possible. D A
16.26. Least heat loss. In an effort to increase sustainability, the university

Figure 16.6: Figure for Exercise 25; least
is aiming to use the shortest length of pipes to connect the buildings cost for installing a pipe.
(labeled B) with the steam power plant (labeled P) in Figure 16.7 in
B L B
order reduce loss of heat from the pipes to the surroundings.
(a) Set up a model and describe how you would solve this problem.
Assume that the four sites are located at the corners of a square of
side length L.
L
(b) How long is the shortest total pipe length connecting P to the 3
other corners?
P B
Figure 16.7: Figure for Exercise 26; least

heat loss.
16.27. Logistic equation and its solution.

1
y(t ) =
1 + e−t
satisfies the differential equation
dy
= y(1 − y).
dt
(b) What is the initial value of y at t = 0?
(c) For what value of y is the growth rate largest?
(d) What happens to y after a very long time?
16.28. Human growth. Given a population of 6 billion people on Planet

Earth, and using the approximate growth rate of r = 0.0125 per year,
how long ago was this population only 1 million? Assume that the
growth has been the same throughout history (which is not actually
true).
16.29. Ducks in a row. Graduate student Ryan Lukeman studies behaviour
of duck flocks swimming near Canada Place in Vancouver, BC. This
figure from his PhD thesis shows his photography set-up. Here H = 10
meters is the height from sea level up to his camera aperture at the
observation point, D = 2 meters is the width of a pier (a stationary α
platform whose size is fixed), and x is the distance from the pier to the
leading duck in the flock (in meters). A visual angle subtended at the H
camera is shown as α.
If the visual angle is increasing at the rate of 1/100 radians per
second, at what rate is the distance x changing at the instant that x = 3 fffff
meters? D x
16.30. Circular race track. Two runners are running around a circular race
track whose length is 400 m, as shown in Figure 16.9(a). The first Figure 16.8: Figure for Exercise 29; ducks
in a row.
runner makes a full revolution every 100 s and the second runner
every 150 s. They start at the same time at the start position, and the
angles subtended by each runner with the radius of the start position
are θ1 (t ), θ2 (t ), respectively. As the runners go around the track
both θ1 (t ) and θ2 (t ) change with time.
(a) At what rate it the angle φ = θ1 − θ2 changing?

(b) What is the angle φ at t = 25s?
(c) What is the distance between the runners at t = 25s?
Note: here “distance” refers to the length of the straight line con-
necting the runners.
(d) At what rate is the distance between the runners changing at t = 25s?
(a) R1
Figure 16.9: Figure for Exercises 30 and 31.
R2 (b) The angles in (a) are θ1 (t ), θ2 (t ). In (b), the
angle between the runners is φ .
START START
16.31. Phase angle and synchrony. Suppose that the same two runners as
in Exercise 30 would speed up or slow down depending on the angle
between them, φ (see Figure 16.9). Then φ = φ (t ) changes with time.
We assume that the angle φ satisfies a differential equation of the form
dφ
= A − B sin(φ )
dt
where A, B > 0 are constants.
(a) What values of φ correspond to steady states (i.e. constant solu-
tions) of this differential equation?
(b) What restriction should be placed on the constants A, B for these
steady states to exist?
(c) Suppose A = 1, B = 2. Sketch the graph of f (φ ) = A − B sin(φ )
fo −π ≤ φ ≤ π and use it to determine what happens if the two
runners start at the same point, (φ = 0) at time t = 0.
16.32. Tumor mass. Figure 16.10 (not drawn to scale) shows a tumor mass
containing a necrotic (dead) core (radius r2 ), surrounded by a layer of
actively dividing tumor cells. The entire tumor can be assumed to be
spherical, and the core is also spherical.
Figure 16.10: Figure for Exercise 32; tumor

mass.
necrotic r2
r1
core
active cells
(a) If the necrotic core increases at the rate 3 cm3 /year and the volume Formula.
of the active cells increases by 4 cm3 /year, at what rate is the outer Note that the volume and surface area of
a sphere are V = (4/3)πr3 , S = 4πr2 .
radius of the tumor (r1 ) changing when r1 = 1 cm. (Leave your
answer as a fraction in terms of π; indicate units with your answer.)
(b) At what rate (in cm2 /yr) does the outer surface area of the tumor
increase when r1 = 1cm?
16.33. Blood vessel branching. Shown in Figure 16.11 is a major artery,

(radius R) and one of its branches (radius r). A labeled schematic
diagram is also shown (right). The length 0A is L, and the distance
between 0 and P is d, where 0P is perpendicular to 0A. The location
of the branch point (B) is to be determined so that the total resistance
to blood flow in the path ABP is as small as possible. (R, r, d, L are
positive constants, and R > r.)
L
R
0 B A
r P
Figure 16.11: Figure for Exercise 33; blood
vessel branching.
(a) Let the distance between 0 and B be x. What is the length of the
segment BA and what is the length of the segment BP?
(b) The resistance of any blood vessel is proportional to its length and Formula.
Note that “z is inversely proportional
inversely proportional to its radius to the fourth power Based on
to y” means that z = k/y for some
this fact, what is the resistance, T1 , of segment BA and what is the constant k.
resistance, T2 , of the segment BP?
(c) Find the value of the variable x for which the total resistance, T (x) =
T1 + T2 is a minimum.
16.34. Implicit differentiation. A surface that looks like an “egg carton” as

depicted in Figure 16.12(a) can be described by the function
z = sin(x) cos(y).
The intersection of this surface with the plane z = 1/2 is a curve

(also called a level curve). One such level curve is shown in Fig-
ure 16.12(b).
Figure 16.12: (a) The surface sin(x) cos(y) =

z. (b) One level curve for this surface is
y shown on its own, and then enlarged, with
π (b)
some tangent lines. The tangent line to
3
this level curve at the point (π/2, π/3) is
horizontal.
x
π
2
(a) Use implicit differentiation to find the slope of the tangent line to a
point on such a curve.
(b) Find the slope of the tangent line to the same level curve at the
point x = π2 .
(c) Find the slope of the tangent line to the same level curve at the
point x = π4 .
Appendices
A
A review of Straight Lines
A.1 Geometric ideas: lines, slopes, equations
Straight lines have some important geometric properties, namely:

The slope of a straight line is the same everywhere along its length.
Slope of a straight line. We define the slope of a straight line as follows:

y
∆y
Slope =
∆x ∆y
where ∆y means “change in the y value” and ∆x means “change in the x ∆x
x
value” between two points. See Figure A.1 for what this notation represents.
Equation of a straight line. Using this basic geometric property, we can find
the equation of a straight line given any of the following information about
the line: Figure A.1: The slope of a line (usually
given the symbol m) is the ratio of the
change in the y value, ∆y to the change in
• The y-intercept, b, and the slope, m: the x value, ∆x.
y = mx + b.
• A point (x0 , y0 ) on the line, and the slope, m, of the line:
y − y0
=m
x − x0
• Two points on the line, say (x1 , y1 ) and (x2 , y2 ):
y − y1 y2 − y1
=
x − x1 x2 − x1
Note: any of these can be rearranged or simplified to produce the standard

form y = mx + b, as discussed in the problem set.
The following examples serve as a refresher on finding the equation of the
line that satisfies each of the given conditions.
Example A.1 In each case write the equation of the straight line that satis-
fies the given statements.
Note: you should also be able to easily sketch the line in each case.
(a) The line has slope 2 and y-intercept 4.
(b) The line goes through the points (1, 1) and (3, −2).
(c) The line has y-intercept -1 and x-intercept 3.
(d) The line has slope −1 and goes through the point (−2, −5).
Solution.
(a) We can use the standard form of the equation of a straight line, y = mx + b
where m is the slope and b is the y-intercept to obtain the equation: y =
2x + 4.
(b) The line goes through the points (1, 1) and (3, −2). We use the fact that
the slope is the same all along the line. Thus,
(y − y0 ) (y1 − y0 )
= = m.
(x − x0 ) (x1 − x0 )
Substituting in the values (x0 , y0 ) = (1, 1) and (x1 , y1 ) = (3, −2),
(y − 1) (1 + 2) 3
= =− .
(x − 1) (1 − 3) 2
This tells us that the slope is m = −3/2. We find that

3 3 3 3 5
y − 1 = − (x − 1) = − x + , ⇒ y = − x+ .
2 2 2 2 2
(c) The line has y-intercept −1 and x-intercept 3, i.e. goes through the
points (0, −1) and (3, 0). We can use the method in (b) to get
1
y = x−1
3
Alternately, as a shortcut, we could find the slope,
∆y 1
m= = .
∆x 3
Note: ∆ means “change in the value”, i.e. ∆y = y1 − y0 .
Thus m = 1/3 and b = −1 (y-intercept), leading to the same result.
(d) The line has slope −1 and goes through the point (−2, −5). Then,
(y + 5)
= −1, ⇒ y + 5 = −1(x + 2) = −x − 2, ⇒ y = −x − 7.
(x + 2)
B
A precalculus review
B.1 Manipulating exponents
Recall: 2n = 2 · 2 . . . 2 (with n factors of 2). This means:
. . . · · · 2} = 2n+m .
2n · 2m = (2 · 2 . . . 2) · (2 · 2 . . . 2) = |2 · 2 · {z
| {z } | {z }
n factors m factors n + m factors
Similarly, we can derive many properties of manipulations of exponents. A
list of these appears below, and holds for any positive base a.
1. 2a 2b = 2a+b as with all similar exponent manipulations.
2. (2a )b = 2ab also stems from simple rules for manipulating exponents.
3. 2x is a function that is defined, continuous, and differentiable for all real

values x.
4. 2x > 0 for all values of x.
5. We define 20 = 1, and we also have that 21 = 2.
6. 2x → 0 for increasingly negative values of x.
7. 2x → ∞ for increasing positive values of x.
B.2 Manipulating logarithms
The following properties hold for logarithms of any base (we used base 2
in our previous section and keep the same base here). Properties of the
logarithm stem directly from properties of the exponential function, and
include the following:
1. log2 (ab) = log2 (a) + log2 (b).
2. log2 (ab ) = b log2 (a).
3. log2 (1/a) = log2 (a−1 ) = − log2 (a).

C
A Review of Simple Functions
We review a few basic concepts related to functions.
C.1 What is a function?
A function is just a way of expressing a special relationship between a value

we consider as the input (“x”) value and an associated output (“y”) value. We
write this relationship in the form
y = f (x )
to indicate that y depends on x. The only constraint on this relationship is that,

for every value of x we can get at most one value of y. This is equivalent to
the “vertical line property”: the graph of a function can intersect a vertical
line at most at one point. The set of all allowable x values is called the
domain of the function, and the set of all resulting values of y are the range.
y
Naturally, we do not always use the symbols x and y to represent indepen-
dent and dependent variables. For example, the relationship
4
V = πr3 x
3
expresses a functional connection between the radius, r, and the volume, V ,

of a sphere. We say in such a case that “V is a function of r”.
Figure C.1: The above elliptical curve
All the sketches shown in Figure C.2 are valid functions. The first is cannot be the graph of a function. The
merely a collection of points, x values and associated y values. The second dashed vertical line intersects the graph at
a histogram. The third sketch is meant to represent the collection of smooth more than one point: this means that a given
value of x corresponds to "too many" values
continuous functions - those of most interest to usnin the study of calculus. of y. If we restrict ourselves to the top part
On the other hand, the example shown in Figure C.1 is not the graph of a of the ellipse only (or the bottom part only),
then we can create a function which has the
function. We see that a vertical line intersects this curve at more than one corresponding graph.
point. This is not permitted since a given value of x may have at most one
corresponding value of y.
y y y
x x
Figure C.2: All the examples above repre-

sent functions.
C.2 Geometric transformations
It is helpful to recognize what happens to the graph of a function when

the relationship between the variables is changed slightly. Often this is
called applying a transformation. In Figure C.3 we illustrate (a) an original
function f (x), (b) the function f (x − a) which shifts f to the right along the
positive x-axis by a distance a, and (c) the function f (x) + b shifting f up
the y-axis by height b.
y y y
b
a
x x x
(a) y = f (x) (b) y = f (x − a) (c) y = f (x) + b
Figure C.3: Shifting the graph of a function

horizontally and vertically
Figure C.4 illustrates what happens to a function when shifts, scaling, or y

reflections occur: a function y = f (x) is shown with a black solid line. On y = f (−x) y = f (x)
the same graph are superimposed the reflections of this graph about the x
axis, y = − f (x) (dashed black), about the y axis y = f (−x) (red), and about
the y and the x-axis, y = − f (−x) (red dashed). The latter is equivalent to a x
rotation of the original graph about the origin.
y = −f (−x) y = −f (x)
Figure C.4: A function and its reflections

about x and y-axes.
A REVIEW OF SIMPLE FUNCTIONS 369
C.3 Classifying
While life offers amazing complexity, one way to study living things is to
classify them into related groups. A biologist looking at animals might group
them according to certain functional properties - being warm blooded, being
mammals, having fur or claws, or having some other interesting characteristic.
In the same way, mathematicians often classify the objects they study (e.g.
functions) into related groups.
An example of way to group functions into very broad classes is also
shown in Figure C.5.
Figure C.5: Classifying functions according

to their properties.
constant linear power smooth wild
constant slope
easily computed
has a derivative
unpredictable
From left to right, the complexity of behaviour in this chart grows: at left,
we see constant and linear functions; these are “most convenient” or simplest
to describe: one or two parameters suffice (e.g. intercept or slope). Further to
the right are smooth and continuous functions, while rightmost some more
irregular, discontinuous function represents those that are outside the group of
the “well-behaved”.
In Section C.4 we study examples along this spectrum. Towards this end,
we describe properties they share, properties they inherit form their “cousins,”
and new characteristics that appear at distinct branches.
C.4 Power functions and symmetry
In this section we list some features of each family of power functions.

Even integer powers. For n = 2, 4, 6, 8 . . . the shape of the graph of y = xn is
as shown in Figure 1.4(a). Note the following characteristics of these graphs:
1. The graphs of all the even power functions intersect at x = 0 and at
at x = ±1. The value of y corresponding to both of these is y = +1. Mastered Material Check
1. What are the coordinates of the
three intersection points of even
power functions?
2. All graphs have a lowest point - a minimum value - at x = 0.
3. As x → ±∞, y → ∞. We equivalently say that these functions are “un-

bounded from above.”
4. The graphs are all symmetric about the y-axis. This special type of symme-
try is of interest in other types of functions (not just power functions). A
function with this property is called an even function.
Odd integer powers. For n = 1, 3, 5, 7, . . . and other odd powers, the graphs
have shapes shown in Figure 1.4(b) and the following characteristics: Mastered Material Check
2. What are the coordinates of the
1. The graphs of the odd power functions intersect at x = 0 and at x = ±1. three intersection points of odd
power functions?
2. None of the odd power functions have a minimum value.
3. As x → +∞, y → +∞. As x → −∞, y → −∞. The functions are “unbounded

from above and below.”
4. The graphs are all symmetric about the origin. This special type of symme-
try is of interest in other types of functions (not just power functions). A
function with this type of symmetry is called an odd function.
Further properties of intersections

Consider the even and odd functions graphed in Figure C.6. Notice that a
horizontal line intersects the graph only once for the odd power but possibly
twice for the even power.
Note: we must allow for the case when the line does not intersect at all, or
that it intersects precisely at the minimum point.
These observations holds in the case of general even and odd power functions
as well.
Figure C.6: The even power functions

y intersect a horizontal line in up to two
places, while the odd power functions
y 2 y=C intersect such a line in only one place.
2
1
1.5
y=C y = x3
1 −1.5 −1 −0.5 0.5 1 1.5x
y = x2
−1
0.5
−2
x
−1.5 −1 −0.5 0.5 1 1.5
(a) (b)
A horizontal line has an equation of the form y = C where C is some

constant. To find where it intersects the graph of a power function y = xn , we
solve an equation of the form
xn = C. (C.1)
To do so, take nth root of both sides:
(xn )1/n = C1/n .
Simplifying, using algebraic operations on powers leads to
(xn )1/n = xn/n = x1 = x = C1/n .
However, we must allow that there may be more than one solution to
Eqn. (C.1), as shown for some C > 0 in Figure C.6. This demonstrates a
distinction between odd and even power functions. If n is even then the
solutions to Eqn. (C.1) are
x = ±C1/n ,
whereas if n is odd, there is a single solution,
x = C1/n .
In general, we can define even and odd functions.
Definition C.1 (Even and odd functions) A function that is symmetric

about the y-axis is said to be an even function. A function that is symmetric
about the origin is said to be an odd function.
Even functions satisfy the relationship
f (x) = f (−x).
Odd functions satisfy the relationship
f (x) = − f (−x).
Examples of even functions include y = cos(x), y = −x8 , y = |x| which are

all their own mirror images when reflected about the y-axis. Examples of odd
functions are y = sin(x), y = −x3 , y = x. Each of these functions is its own
double-reflection (about y and then x-axes).
In a later calculus course, when we compute integrals, taking these sym-
metries into account can help to simplify (or even avoid) calculations.
Optional: Combining even and odd functions

Not every function is either odd or even. However, if we start with symmetric
functions, some manipulations can either preserve or reverse the symmetry.
Example C.1 Show that the product of an even and an odd function is an
odd function.
Solution. Let f (x) be even. Then
f (x) = f (−x).
Let g(x) be an odd function. Then g(x) = −g(−x). We define h(x) to be the
product of these two functions,
h(x ) = f (x )g(x ).
Using the properties of f and g,
f (x)g(x) = f (−x)[−g(−x)],
so, rearranging, we get
h(x) = f (x)g(x) = f (−x)[−g(−x)] = −[ f (−x)g(−x)].
but this is just the same as −h(−x). We have established that
h(x) = −h(−x),
so the new function is odd. ♦

A function is not always even or odd. Many functions are neither even nor
odd. However, it is possible to show that given any function, y = f (x), we can
write it as a sum of an even and an odd function. This is left as a challenge
for the reader, with the following hint:
Hint: suppose f (x) is not an even nor an odd function. Consider defining the
two associated functions:
1
fe (x) = ( f (x) + f (−x)),
2
and
1
f0 (x) = ( f (x) − f (−x)).
2
Can you draw a sketch of what these would look like for the function given in
Figure C.3(a)? Show that fe (x) is even and that f0 (x) is odd. Now show that
f (x ) = f e (x ) + f 0 (x ).
C.5 Inverse functions and fractional powers
Suppose we are given a function expressed in the form
y = f (x ).
This implies that x is the independent variable, and y is obtained from it

by evaluating a function, i.e. by using the “rule” or operation specified by
that function. This mathematical statement expresses a certain relationship
between the two variables, x and y, in which the roles are distinct: x is a value
we pick, and y is calculated from it.
However, sometimes we can express a relationship in more than one way:
as an example, if the connection between x and y is simple squaring, then
provided x > 0, we might write either
√
y = x2 or x = y1/2 = y
to express the same relationship. In other words

√
y = x2 ⇔ x = y.
We have used two distinct functions to describe the relationship from two
points of view: one function involves squaring and the other takes a square
root. We may also notice that for x > 0,
√
f (g(x)) = ( x)2 = x,
q
g( f (x)) = (x2 ) = x,
i.e. that these two functions invert each other’s effect.
Functions that satisfy
y = f (x ) ⇔ x = g(y)
are said to be inverse functions. We often use the notation f −1 (x) to denote
the function that acts as an inverse function to f (x).
Graphical property of inverse functions

The graph of an inverse function y = f −1 (x) is geometrically related to the
graph of the original function: it is a reflection of y = f (x) about the 45◦
line, y = x. This relationship is shown in Figure C.7(a) for a pair of functions f
and f −1 .
But why should this be true? The idea is as follows: suppose that (a, b) is
any point on the graph of y = f (x). This means that b = f (a). That, in turn,
implies that a = f −1 (b), which then tells us that (b, a) must be a point on
the graph of f −1 (x). But the points (a, b) and (b, a) are related by reflection
about the line y = x. This is true for any arbitrary point, and so must be true for
all points on the graphs of the two functions.
Restricting the domain

The above argument establishes that, given the graph of a function, its inverse
is obtained by reflecting the graph in an imaginary mirror placed along a
line y = x.
However, a difficulty could arise. In particular, for the function
y = f ( x ) = x2 ,
Figure C.7: The point (a, b) is on the graph

y y of y = f (x). If the roles of x and y are
y = f −1(x) interchanged, this point becomes (b, a).
Geometrically, this point is the reflection
(b, a) y = f (x)
of (a, b) about the line y = x. Thus, the
(a, b) graph of the inverse function y = f −1 (x) is
y = f (x) related to the graph of the original function
by reflection about the line y = x. In (b),
y = f −1(x) the inverse is not a function, as it does not
(a, b) satisfy the vertical line property. In (a),
both f and its reflection satisfy that property,
(b, a) and thus the inverse, f −1 is a true function.
x x
(b) (a)
a reflection of this type would lead to a curve that cannot be a function, as

shown in Figure C.8.
Note: the sideways parabola would not be a function if we included both its
branches, since a given value of x would have two associated y values.
To fix such problems, we simply restrict the domain to x > 0, i.e. to the
solid parts of the curves shown in Figure C.8. For this subset of the x-axis, we
have no problem defining the inverse function.
y Figure C.8: The graph of y = f (x) = x2

(black) and of its inverse function (red). We
cannot define the inverse for all x, because
1
y = x1/2 the red parabola does not satisfy the vertical
line property. However, if we restrict the
original function to positive x values, this
omit this y = x2 problem is circumvented.
branch x
−1.5 −1 −0.5 0.5 1 1.5
−1
blue curve is not a function
if this branch is included
Observe that the problem described above would be encountered for any
of the even power functions by virtue of their symmetry about the y-axis but
not by the odd power functions. For example,
y = f ( x ) = x3 , y = f −1 (x) = x1/3
are inverse functions for all x values: when we reflect the graph of x3 about
the line y = x we do not encounter the problem of multiple y values.
This follows directly from the horizontal line properties that we saw
in Figure C.6. When we reflect the graphs shown in Figure C.6 about the
line y = x, the horizontal lines are reflected onto vertical lines. Odd power
functions have inverses that intersect a vertical line exactly once, i.e. they
satisfy the “vertical line property” discussed earlier.
C.6 Polynomials
Recall that a polynomial is a function of the form
y = p(x) = an xn + an−1 xn−1 + · · · + a1 x + a0 .
This form is sometimes referred to as superposition (i.e. simple addition) of

the basic power functions with integer powers. The constants ak are called
coefficients. In practice, some of these may be zero. We restrict attention to
the case where all these coefficients are real numbers. The highest power n
(whose coefficient is not zero) is called the degree of the polynomial.
We are interested in these functions for several reasons. Primarily, we
find that computations involving polynomials are particularly easy, since
operations include only the basic addition and multiplication.
Features of polynomials
• Zeros of a polynomial: values of x such that
y = p(x) = 0.
If p(x) is quadratic (a polynomial of degree 2) then the quadratic formula

gives a simple way of finding roots of this equation/zeros of the polyno-
mial. Generally, for most polynomials of degree higher than 5, there is
no analytical recipe for finding zeros. Geometrically, zeros are places
where the graph of the function y = p(x) crosses the x-axis. This fact is ex-
ploited in Chapter 5 to approximate the values of the zeros using Newton’s
Method.
• Critical Points: places on the graph where the value of the function is
locally larger than those nearby (local maxima) or smaller than those
nearby (local minima) are of interest to us. Calculus is one of the main
tools for detecting and identifying such places.
• Behaviour for very large x: all polynomials are unbounded as x → ∞ and

as x → −∞. For large enough values of x, the power function y = f (x) = xn
with the largest power, n, dominates over other power functions with
smaller powers, as seen in Chapter 1. For
p(x) = an xn + an−1 xn−1 + · · · + a1 x + a0 ,

the first (highest power) term dominates for large x. Thus for large x
(whether positive or negative),
p(x) ≈ an xn for large x.
• Behaviour for small x: close to the origin, power functions with smallest
powers dominate (see Chapter 1). Thus, for x ≈ 0 the polynomial is
governed by the behaviour of the smallest (non-zero coefficient) power,
i.e.,
p(x) ≈ a1 x + a0 for small x.
D
Limits
We introduced notation involving limits without carefully defining what was

meant. Here, such technical matters are briefly discussed.
The concept of a limit helps us to describe the behaviour of a function
close to some point of interest. This is useful in the case of functions that are
either not continuous, or not defined somewhere. We use the notation
lim f (x)
x→a
to denote the value the function f approaches as x gets closer and closer to
the value a.
D.1 Limits for continuous functions
If x = a is a point at which the function is defined and continuous (informally:

has no “breaks in its graph”) the value of the limit and the value of the
function at a point are the same, i.e.
If f is continuous at x = a then
lim f (x) = f (a).

x→a
Example D.1 Find lim f (x) for the function y = f (x) = 10.
x→0
Solution. This function is continuous (and constant) everywhere. In fact, the

value of the function is independent of x. We conclude immediately that
lim f (x) = lim 10 = 10.

x→0 x→0
Example D.2 Find lim f (x) for the function y = f (x) = sin(x).
x→0
Solution. This function is a continuous trigonometric function, and has the

value sin(0) = 0 at the origin. Thus
lim f (x) = lim sin(x) = 0.

x→0 x→0
♦
Power functions are continuous everywhere. This motivates the next
example.
Example D.3 Compute the limit lim xn where n is a positive integer.

x→0
Solution. The function in question, f (x) = xn , is a simple power function

that is continuous everywhere. Further, f (0) = 0 for any n a positive integer.
Hence the limit as x → 0 coincides with the value of the function at that point,
so
lim xn = 0.
x→0
D.2 Properties of limits
Suppose we are given two functions, f (x) and g(x). we also assume that both
functions have (finite) limits at the point x = a. Then the following statements
follow.
1. lim ( f (x) + g(x)) = lim f (x) + lim g(x)

x→a x→a x→a
2. lim (c f (x)) = c lim f (x)

x→a x→a

3. lim ( f (x) · g(x)) = lim f (x) · lim g(x)
x→a x→a x→a
4. Provided that lim g(x) 6= 0, we also have that

x→a
lim f (x)
!
f (x )

x→a
lim = .
x→a g(x ) lim g(x)
x→a
The first two statements are equivalent to linearity of the process of

computing a limit.
Example D.4 Find lim f (x) for the function y = f (x) = 2x2 − x3 .
x→2
Solution. Since this function is a polynomial, and so continuous everywhere,

we can simply plug in the relevant value of x, i.e.
lim 2x2 − x3 = 2 · 22 − 23 = 0.

x→2
Thus when x gets closer to 2, the value of the function gets closer to 0. ♦
Note: when the function is continuous, the value of the limit is the same as
the value of the function at the given point.
LIMITS 379
D.3 Limits of rational functions
Case 1: Denominator nonzero

We first consider functions that are the quotient of two polynomials, y =
f (x)/g(x) at points were g(x) 6= 0. This allows us to apply Property 4 of
limits together with what we have learned about the properties of power func-
tions and polynomials. Much of this discussion is related to the properties
of power functions and dominance of lower (higher) powers at small (large)
values of x, as discussed in Chapter 1. In the examples below, we consider
both limits at the origin (at x = 0) and at infinity (for x → ∞). The latter means
“very large x”. See Section 1.4for examples of the informal version of the
same reasoning used to reach the same conclusions.
Example D.5 Find the limit as x → 0 and as x → ∞ of the quotients
Kx Axn
(a) , (b) .
kn + x an + xn
Solution. We recognize (a) as an example of the Michaelis-Menten kinetics,

found in (1.8) and (b) as a Hill function in (1.7) of Chapter 1. We now
compute, first for x → 0,
Kx Axn
(a) lim = 0, (b) lim = 0.
x→0 kn + x x→0 an + xn
This follows from the fact that, provided a, kn 6= 0, both functions are continu-
ous at x = 0, so that their limits are the same as the actual values attained by
the functions. Now for x → ∞
Kx Kx Axn Axn
(a) lim = lim = K, (b) lim = lim = A.
x→∞ kn + x x→∞ x x→∞ an + xn x→∞ xn
This follows from the fact that the constants kn , an are always “swapped out”
by the value of x as x → ∞, allowing us to obtain the result. Other than the
formal limit notation, there is nothing new here that we have not already
discussed in Sections 1.5. ♦
Below we apply similar reasoning to other examples of rational functions.
Example D.6 Find the limit as x → 0 and as x → ∞ of the quotients
3x2 1+x
(a) , (b) .
9 + x2 1 + x3
Solution. For part (a) we note that as x → ∞, the quotient approaches 3x2 /x2 =
3. As x → 0, both numerator and denominator are defined and the denominator
is nonzero, so we can use the 4th property of limits. We thus find that
3x2 3x2
(a) lim = 3, lim = 0,
x→∞ 9 + x2 x→0 9 + x2
For part (b), we use the fact that as x → ∞, the limit approaches x/x3 = x−2 →
0. As x → 0 we can apply property 4 yet again to compute the (finite) limit, so
that
1+x 1+x
(b) lim = 0, lim = 1.
x→∞ 1 + x3 x→0 1 + x3
Example D.7 Find the limits of the following function at 0 and ∞
x4 − 3x2 + x − 1
y= .
x5 + x
Solution. For x → ∞ powers with the largest power dominate, whereas
for x → 0, smaller powers dominate. Hence, we find
x4 − 3x2 + x − 1 x4 1
lim 5
= lim 5
= lim = 0.
x→∞ x +x x→∞ x x→∞ x
x4 − 3x2 + x − 1 −1 1
lim = lim = − lim = ∞
x→0 x5 + x x→0 x x→0 x
So in the latter case, the limit does not exist. ♦
Case 2: zero in the denominator and “holes” in a graph

In the previous examples, evaluating the limit, where it existed, was as simple
as plugging the appropriate value of x into the function itself. The next
example shows that this is not always possible.
Example D.8 Compute the limit as x → 4 of the function f (x) = 1/(x − 4)
Solution. This function has a vertical asymptote at x = 4. Indeed, the value of

the function shoots off to +∞ if we approach x = 4 from above, and −∞ if we
approach the same point from below. We say that the limit does not exist in
this case. ♦
Example D.9 Compute the limit as x → −1 of the function f (x) = x/(x2 − 1)
Solution. We compute
x x
lim = lim
x→−1 x2 − 1 x→−1 (x − 1)(x + 1)
It is evident (even before factoring as we have done) that this function has a
vertical asymptote at x = −1 where the denominator approaches zero. Hence,
the limit does not exist. ♦
Next, we describe an extremely important example where the function has
a “hole” in its graph, but where a finite limit exists. This kind of limit plays a
huge role in the definition of a derivative.
Example D.10 Find lim f (x) for the function y = (x − 2)/(x2 − 4).
x→2
LIMITS 381
Solution. This function is a quotient of two rational expressions f (x)/g(x)

but we note that limx→2 g(x) = limx→2 (x2 − 4) = 0. Thus we cannot use
property 4 directly. However, we can simplify the quotient by observing that
for x 6= 2 the function y = (x − 2)/(x2 − 4) = (x − 2)/(x − 2)(x + 2) takes on
the same values as the expression 1/(x + 2). At the point x = 2, the function
itself is not defined, since we are not allowed division by zero. However, the
limit of this function does exist:
(x − 2)
lim f (x) = lim .
x→2 x→2 ( x2 − 4 )
Provided x 6= 2 we can factor the denominator and cancel:
y
(x − 2) (x − 2) 1
lim = lim = lim
x→2 (x2 − 4) x→2 (x − 2)(x + 2) x→2 (x + 2)
Now we can substitute x = 2 to obtain y = f (x)

1 1
lim f (x) = =
x→2 (2 + 2) 4 1/4
♦
x
Example D.11 Compute the limit 2
K (x + h)2 − Kx2 (x−2)

Figure D.1: The function y = (x2 −4) has
lim .
h→0 h a “hole” in its graph at x = 2. The limit
of the function as x approaches 2 does
Solution. This is a calculation we would perform to compute the derivative exist, and “supplies the missing point”:
of the function y = Kx2 from the definition of the derivative. Details have limx→2 f (x) = 14 .
already been displayed in Example 2.10. The essential idea is that we expand
the numerator and simplify algebraically as follows:
(2xh + h2 )
lim K = lim K (2x + h) = 2Kx.
h→0 h h→0
Even though the quotient is not defined at the value h = 0 (as the denominator
is zero there), the limit exists, and hence the derivative can be defined. ♦
See also Example 3.12 for a similar calculation for the function Kx . 3
D.4 Right and left sided limits
Some functions are discontinuous at a point, but we may still be able to

define a limit that the function attains as we approach that point from the
right or from the left. (This is equivalent to gradually decreasing or gradually
increasing x as we get closer to the point of interest.
Consider the function
(
0 if x < 0;
f (x ) =
1 if x > 0.
This is a step function, whose values is 0 for negative real numbers, and 1 for
positive real numbers. The function is not even defined at the point x = 0 and
has a jump in its graph. However, we can still define a right and a left limit as
follows:
lim f (x) = 1, lim f (x) = 0.
x→+ 0 x→− 0
That is, the limit as we approach from the right is 0 whereas from the left it
is 1. We also state the following result:
If f (x) has a right and a left limit at a point x = a and if those limits are equal,
then we say that the limit at x = a exists, and we write
lim f (x) = lim f (x) = lim f (x)

x→+ a x→− a x→a
Example D.12 Find lim f (x) for the function y = f (x) = tan(x).
x→π/2
Solution. The function tan(x) = sin(x)/ cos(x) cannot be continuous at x =

π/2 because cos(x) in the denominator takes on the value of zero at the
point x = π/2. Moreover, the value of this function becomes unbounded
(grows without a limit) as x → π/2. We say in this case that “the limit does
not exist”. We sometimes use the notation
lim tan(x) = ±∞.

x→π/2
(We can distinguish the fact that the function approaches +∞ as x ap-
proaches π/2 from below, and −∞ as x approaches π/2 from higher values.)
♦
D.5 Limits at infinity
We can also describe the behaviour “at infinity” i.e. the trend displayed by a
function for very large (positive or negative) values of x. We consider a few
examples of this sort below.
Example D.13 Find lim f (x) for the function y = f (x) = x3 − x5 + x.

x→∞
Solution. All polynomials grow in an unbounded way as x tends to very

large values. We can determine whether the function approaches positive or
negative unbounded values by looking at the coefficient of the highest power
of x, since that power dominates at large x values. In this example, we find
that the term −x5 is that highest power. Since this has a negative coefficient,
the function approaches unbounded negative values as x gets larger in the
positive direction, i.e.
lim x3 − x5 + x = lim −x5 = −∞.

x→∞ x→∞
♦
LIMITS 383
Example D.14 Determine the following two limits:
(a) lim e−2x , (b) lim e5x ,

x→∞ x→−∞
Solution. The function y = e−2x becomes arbitrarily small as x → ∞. The

function y = e5x becomes arbitrarily small as x → −∞. Thus we have
(a) lim e−2x = 0, (b) lim e5x = 0.

x→∞ x→−∞
Example D.15 Find the limits below:

1
(a) lim x2 e−2x , (b) lim e−x ,
x→∞ x→0 x
Solution. For part (a) we state here the fact that as x → ∞, the exponential
function with negative exponent decays to zero faster than any power func-
tion increases. For part (b) we note that for the quotient e−x /x we have that
as x → 0 the top satisfies e−x → e0 = 1, while the denominator has x → 0.
Thus the limit at x → 0 cannot exist. We find that
1
(a) lim x2 e−2x = 0, (b) lim e−x = ∞,
x→∞ x→0 x
D.6 Summary of special limits
As a reference, in the table below, we collect some of the special limits that
are useful in a variety of situations.
Function x→ Limit notation value Table D.1: A collection of useful limits.

e−ax , a > 0 ∞ lim e−ax 0
x→∞
e−ax , a > 0 −∞ lim e−ax ∞
x→−∞
eax , a > 0 ∞ lim eax ∞
x→∞
kx
ekx 0 lim e 1
x→0
xn e−ax , a>0 ∞ lim xn e−ax 0
x→∞
ln(ax), a > 0 ∞ lim ln(ax) ∞
x→∞
ln(ax), a > 0 1 lim ln(ax) 0
x→1
ln(ax), a > 0 0 lim ln(ax) −∞
x→0
x ln(ax), a > 0 0 lim x ln(ax) 0
x→0
ln(ax) ln(ax)
,a>0 ∞ lim 0
x x→∞ x
sin(x) sin(x)
0 lim 1
x x→0 x
We can summarize the information in this table informally as follows:

1. The exponential function ex grows faster than any power function as x

increases, and conversely the function e−x = 1/ex decreases faster than any
power of (1/x) as x grows. The same is true for eax provided a > 0.
2. The logarithm ln(x) is an increasing function that keeps growing without

bound as x increases, but it does not grow as rapidly as the function y = x.
The same is true for ln(ax) provided a > 0. The logarithm is not defined
for negative values of its argument and as x approaches zero, this function
becomes unbounded and negative. However, it approaches −∞ more
slowly than x approaches 0. For this reason, the expression x ln(x) has a
limit of 0 as x → 0.
E
Proofs
This Appendix was written by Dr. Sophie Burrill.
E.1 Proof of the power rule
We present a proof for the power rule. Recall from Section 4.1:
The power rule states that the derivative of the power function f (x) = xn
is nxn−1 .
Proof. We begin with the definition of the derivative:
f (x + h) − f (x ) ( x + h ) n − xn
f 0 (x) = lim = lim .
h→0 h h→0 h
As mentioned in Section 4.1, the binomial (x + h)n entails lenthy algebra. We

employ the binomial theorem (which we do not prove):
Binomial theorem. If n is a positive integer, then
n(n − 1) n−2 2 n(n − 1)(n − 2) n−3 3

(x + y)n = xn + nxn−1 y + x y + x y
2·1 3·2·1
n(n − 1) . . . (n − k + 1) n−k k
+... + x y + . . . + nxyn−1 + yn
k (k − 1) . . . 2 · 1
Note that this means the expansion of (x + y)n is a sum of terms of the form
ck xn−k yk , k = 0, 1, . . . , n
where ck (called a “binomial coefficient”) is a coefficient that depends on

both n and k. The exact form of ck is not necessary for the proof of the power
rule - except for the terms c0 = 1 and c1 = n, the coefficients of xn and xn−1 y.
Let us use the binomial theorem and expand the numerator in the defini-
tion of the derivative:
f (x + h) − f (x ) = ( x + h ) n − xn
= (c0 xn + c1 xn−1 h + c2 xn−2 h2 + c3 xn−3 h3
+ . . . + cn−1 xhn−1 + cn hn ) − xn
We can rewrite using the fact that c0 = 1 and note that the terms xn cancel:
(x + h)n − xn = (xn + c1 xn−1 h + . . . + cn hn ) − xn = c1 xn−1 h + . . . + cn hn .
All of the remaining terms have h as a factor, which we can factor out:
(x + h)n − xn = h[c1 xn−1 + c2 xn−2 h + · · · + cn hn−1 ].
Substituting this into the definition of the derivative we achieve:
f (x + h) − f (x ) ( x + h ) n − xn
f 0 (x ) = lim = lim
h→0 h h→0 h
h[c1 xn−1 + c2 xn−2 h + · · · + cn hn−1 ]
= lim
h→0 h
= lim [c1 xn−1 + c2 xn−2 h + · · · + cn hn−1 ]
h→0
Note that all terms except for the first have h as a factor and so tend to 0
as h → 0. This gives that
f 0 (x) = c1 xn−1 ,
and we have already noted that c1 = n, so
f 0 (x) = nxn−1 ,
which proves the power rule.
E.2 Proof of the product rule
We proof the product rule. Recall from Section 4.1:
The product rule: If f (x) and g(x) are two functions, each differentiable in
d [ f (x)g(x)] d f (x ) dg(x)
= g(x ) + f (x ).
dx dx dx
Another notation for this rule is
[ f (x)g(x)]0 = f 0 (x)g(x) + g0 (x) f (x).

P RO O F S 387
Proof. Let k(x) = f (x)g(x), the product of the two functions. We use the
definition of the derivative:
k (x + h) − k (x ) f (x + h)g(x + h) − f (x )g(x )
k0 (x) = lim = lim
h→0 h h→0 h
Adding 0 = f (x)g(x + h) − f (x)g(x + h) allows us to perform some helpful
factoring:
f (x + h)g(x + h) − f (x )g(x + h) + f (x )g(x + h) − f (x )g(x )
k 0 (x ) = lim
h→0 h
[ f (x + h) − f (x)] · g(x + h) + f (x) · [g(x + h) − g(x)]
= lim
h→0 h
f (x + h) − f (x ) g(x + h) − g(x )
= lim g(x + h) + f (x )
h→0 h h
Due to properties of limits (see Appendix D.2) we can distribute the limit and
recognize familiar derivatives:
f (x + h) − f (x ) g(x + h) − g(x )
k 0 (x ) = lim lim g(x + h) + lim f (x) lim
h→0 h h→0 h→0 h→0 h
= f 0 ( x ) g ( x ) + f ( x ) g0 ( x ) .
Thus we have proved the power rule, that
[ f (x)g(x)]0 = f 0 (x)g(x) + f (x)g0 (x).
E.3 Proof of the quotient rule
We provide a proof the quotient rule. Recall from Section 4.1:
The quotient rule: If f (x) and g(x) are two functions, each differentiable in
d f (x ) dg(x)
d f (x ) dx g(x) − dx f (x )

= .
dx g(x) [g(x)]2
We can also write this in the form
f (x ) 0 f 0 (x)g(x) − g0 (x) f (x)

= .
g(x ) [g(x)]2
Proof. This proof also follows from the definition of the derivative; it con-
f (x )
tains some careful arithmetic. Let k(x) = g(x) . Using the definition of the
derivative we get:
f (x +h) f (x )
k (x + h) − k (x ) g(x +h)
− g(x ) 1 f (x + h) f (x )

0
k (x) = lim = lim = lim − .
h→0 h h→0 h h→0 h g(x + h) g(x )
Finding a common denominator and then adding 0 = g(x + h) f (x + h) − g(x +

h) f (x + h) in the numerator we proceed:
1 f (x + h)g(x ) − f (x )g(x + h)

k0 (x) = lim
h→0 h g(x + h)g(x )
1 f (x + h)g(x ) − f (x + h)g(x + h) + f (x + h)g(x + h) − f (x )g(x + h)

= lim
h→0 h g(x + h)g(x )
1 − f (x + h)[g(x + h) − g(x)] + g(x + h)[ f (x + h) − f (x)]

= lim
h→0 h g(x + h)g(x )
Using properties of limits and identifying the definition of the derivative

for g0 (x) and f 0 (x) leads us to:
− f (x + h)g0 (x) g(x + h) f 0 (x)

k0 (x) = lim +
h→0 g(x + h)g(x ) g(x + h)g(x )
− f (x)g0 (x) + g(x) f 0 (x) f 0 (x)g(x) − g0 (x) f (x)
= = .
[g(x)]2 [g(x)]2
We have thus proved the quotient rule. Despite the arithmetic required,
hopefully the fact that the definition of the derivative is all that is required
provides the reader with some comfort.
E.4 Proof of the chain rule
We present a plausibility argument for the chain rule. Recall from Section 8.1:
If y = g(u) and u = f (x) are both differentiable functions and y = g( f (x)) is

the composite function, then the chain rule of differentiation states that
dy dy du
= .
dx du dx
Proof. We first note that if a function is differentiable, it is also continuous.
Because of this continuity, when x changes a very little, u can change only by
a little - there are no abrupt jumps. Thus, using our notation, if ∆x → 0 then
∆u → 0.
Now consider the definition of the derivative dy/du:
dy ∆y
= lim .
du ∆u→0 ∆u
This means that for any (finite) ∆u,
∆y dy
= + ε,
∆u du
where ε → 0 as ∆u → 0. Then
dy
∆y = ∆u + ε∆u.
du
P RO O F S 389
Now divide both sides by some (nonzero) ∆x:
∆y dy ∆u ∆u
= +ε .
∆x du ∆x ∆x
Taking ∆x → 0 we get ∆u → 0, (by continuity) and hence also ε → 0 so that as
desired,
dy dy du
= .
dx du dx

F
Trigonometry review
The definition of trigonometric functions in terms of the angle θ in a right

triangle are reviewed in Figure. F.1.
Figure F.1: Review of the relation between

ratios of side lengths (in a right triangle) and
sin θ = opp/hyp
use
trigonometric functions of the associated
opposite
e n angle.
ot
hyp cos θ = adj/hyp
θ tan θ = opp/adj
adjacent
Based on these definitions, certain angles’ sine and cosine can be found
explicitly - and similarly tan(θ ) = sin(θ )/ cos(θ ). This is shown in Table F.1.
degrees radians sin (t ) cos (t ) tan (t ) Table F.1: Values of the sines, cosines, and
tangent for standard angles.
0 0 0 1
√
0
π 1 3 √1
30 6 √2 √2 3
π 2 2
45 4 √2 2 1
π 3 1
√
60 3 2 2 3
π
90 2 1 0 ∞
We also define the other trigonometric functions as follows:
sin(t ) 1
tan(t ) = , cot(t ) = ,
cos(t ) tan(t )
1 1
sec(t ) = , csc(t ) = .
cos(t ) sin(t )
Sine and cosine are related by the identity
π
cos(t ) = sin(t + ).
2
This identity then leads to two others of similar form. Dividing each side of
the above relation by cos2 (t ) yields
tan2 (t ) + 1 = sec2 (t ), (F.1)
whereas division by sin2 (t ) gives us
1 + cot2 (t ) = csc2 (t ).
These aid in simplifying expressions involving the trigonometric functions, as

we shall see.
Law of cosines. This law relates the cosine of an angle to the lengths of sides
formed in a triangle (see Figure F.2). b
c
c2 = a2 + b2 − 2ab cos(θ ) (F.2)
θ
where the side of length c is opposite the angle θ .
There are other important relations between the trigonometric functions
a
(called trigonometric identities). These should be remembered.
Angle sum identities. The trigonometric functions are nonlinear. This means Figure F.2: Law of cosines states that c2 =
a2 + b2 − 2ab cos(θ ).
that, for example, the sine of the sum of two angles is not just the sum of
the two sines. One can use the law of cosines and other geometric ideas to
establish the following two relationships:
sin(A + B) = sin(A) cos(B) + sin(B) cos(A) (F.3)
cos(A + B) = cos(A) cos(B) − sin(A) sin(B) (F.4)
These two identities appear in many calculations and aid in computing

derivatives of basic trigonometric formulae.
Related identities. The identities for the sum of angles can be used to derive
a number of related formulae. For example, by replacing B by −B we get the
angle difference identities:
sin(A − B) = sin(A) cos(B) − sin(B) cos(A)

cos(A − B) = cos(A) cos(B) + sin(A) sin(B)
By setting θ = A = B, we find the subsidiary double angle formulae:
sin(2θ ) = 2 sin(θ ) cos(θ )

cos(2θ ) = cos2 (θ ) − sin2 (θ )
and these can also be written in the form
2 cos2 (θ ) = 1 + cos(2θ )
2 sin2 (θ ) = 1 − cos(2θ ).
T R I G O N O M E T RY R E V I E W 393
F.1 Summary of the inverse trigonometric functions

x arcsin (x ) arccos (x )
In Table F.2 we show the table of standard values of functions arcsin(x) √−1 −π/2 π
−√3/2 −π/3 5π/6
and arccos(x). In Figure F.3 we summarize the the relationships between the − 2/2 −π/4 3π/4
original trigonometric functions and their inverses. −1/2 −π/6 2π/3
0 0 π/2
y y √1/2 π/6 π/3
√2/2 π/4 π/4
1
y = Sin(x) y=x
3/2 π/3 π/6
1 1 π/2 0
0.5 y = sin(x) Table F.2: Standard values of the inverse
x x
−6 −4 −2 2 4 6 −1.5 −1 −0.5 0.5 1 1.5
−0.5 y = Sin(x)
−1 −1
y = arcsin(x)
y y Figure F.3: A summary of the trigonometric

3
functions and their inverses. (a) Sin(x)
1 y = Cos(x) y = arccos(x) (b) arcsin(x), (b) Cos(x) (d) arccos(x), (e)
2 y=x Tan(x) (f) arctan(x). The red curves are the
0.5
restricted domain portions of the original
trig functions. The gold curves are the
x 1
−6 −4 −2 2 4 6 inverse functions.
y = Cos(x)
−0.5
x
−1 1 2 3
−1
y = cos(x)
−1
y y
10 6
4
5 y=x
y = T an(x) 2
y = arctan(x)
−6 −4 −2 2 4 6 x −6 −4 −2 2 4 6x
−2
−5
y = tan(x) −4
y = T an(x)
−10 −6
G
For further study
In Sections G.1 and G.2 we suggest topics that are related to the material
in Chapter 1. Section G.3 supplements biological examples in Chapters 3
and 4. The material in G.4 supplements examples in Chapter 7 and provides
additional practice with optimization. The optimal foraging time for a specific
patch function studied in Chapter 7 is generalized in Section G.5. Section G.6
extends the study of trigonometric functions and their derivatives as seen in
Chapter 15 to differential equations.
G.1 Michaelis-Menten transformed to a linear relationship
Michaelis-Menten kinetics explored in Eqn. (1.8) is a nonlinear saturating

function in which the concentration x is the independent variable on which
the reaction velocity, v depends:
Kx
v= .
kn + x
As discussed in Section 1.5, the constants K and kn depend on the enzyme
and are often quantified in a biochemical assay of enzyme action. Histori-
cally, a convenient way to estimate the values of K and kn was to measure v
for many different values of the initial substrate concentration. Without non-
linear fitting software widely available, Eqn. (1.8) was transformed (meaning
that it was rewritten) as a linear relationship.
We can do so as well with algebra. We begin by taking reciprocals of our
equation and expanding:
1 kn + x
= ,
v Kx
kn x
= +
Kx Kx

kn 1 1
= +
K x K
This suggests defining the two constants:
kn 1
m= , b= .
K K
In which case, the relationship between 1/v and 1/x becomes linear:

1 1
=m + b. (G.1)
v x
Both the slope, m and intercept b of the straight line provide information
about the parameters. The relationship in Eqn. (G.1), which is a disguised
variant of Michaelian kinetics, is called the Lineweaver-Burk relationship. In
Exercise 30 this is used to estimate the values of K and kn from biochemical
data about an enzyme.
G.2 Spacing of fish in a school
Many animals live or function best when they are in a group. Social groups
include herds of wildebeest, flocks of birds, and schools of fish, as well as
swarms of insects. Life in a group can affect the way that individuals forage
(search for food), their success at detecting or avoiding being eaten by a
predator, and other functions such as mating, protection of the young, etc.
Biologists are interested in the ecological implications of groups on their own
members or on other species with whom they interact, and how individual
behaviour, combined with environmental factors and random effects affect
the shape, spacing and function of the groups.
In many social groups, the spacing between individuals is relatively
constant from one part of the formation to another, because animals that
get too close start to move away from one another, whereas those that get
too far apart are attracted back. These spacing distances can be observed
in a variety of groups, and were described in many biological publications.
For example, Emlen [Emlen Jr, 1952] studied flocks of birds and found that
gulls are spaced at about one body length apart. Similarly, Conder [J., 1949]
observed a 2 − 3 body lengths spacing distance in tufted ducks while Miller
[R.S. and J.D., 1966] found that in the flock he observed, sandhill cranes try
to keep about 5.8 ft apart.
To explain why certain spacing is maintained in a group of animals, it
was proposed that there are mutual attraction and repulsion interactions
(effectively acting like simple forces) between individuals. Breder [Breder,
1954] followed a number of species of fish that school, and measured the
individual spacing in units of the fish body length, showing that individuals
are separated by 0.16 − 0.25 body length units. He suggested that the effective
forces between individuals were similar to inverse power laws for repulsion
and attraction. Breder considered a quantity he called cohesiveness, defined
as:
A R
c = m − n, (G.2)
x x
where A, R are magnitudes of attraction and repulsion, x is the distance
between individuals, and m, n are integer powers that govern how quickly
the interactions fall off with distance. We could re-express the formula in
F O R F U RT H E R S T U DY 397
Eqn. (G.2) as
c = Ax−m − Rx−n
Thus, the function shown in Breder’s cohesiveness formula is related to

our power functions, but the powers are negative integers. A specific case
considered by Breder was m = 0, n = 2, i.e. constant attraction and inverse
square law repulsion,
c = A − (R/x2 )
Breder specifically considered the “point of neutrality”, where c = 0. The

distance at which this occurs is:
x = (R/A)1/2
where attraction and repulsion are balanced. This is the distance at which two
fish would be most comfortable: neither tending to move apart, nor get closer
together.
G.3 A biological speed machine
Lysteria monocytogenes is a parasite that lives inside cells of the host, causing
a nasty infection. It has been studied by cellular biologists for its amazingly
fast propulsion, which uses the host’s actin filaments as “rocket fuel”. Actin
is part of the structural component of all animal cells, and is known to play
a major role in cell motility. Lysteria manages to “hijack” this cellular
mechanism, assembling it into its own comet tail, which it uses to propel
inside the cell and pass from one cell to the next. Figure G.1 illustrates part of
these curious traits.
Figure G.1: The parasite Lysteria lives

inside a host cell. It assembles a “rocket-
like” tail made up of actin, and uses this
assembly to move around the cell, and to
pass from one host cell to another. Figure
from [Tilney and Portnoy, 1989].
Researchers in cell biology use Lysteria to learn about motility at the

cellular level. It has been discovered that certain proteins on the external
surface of this parasite (ActA) are responsible for the ability of Lysteria
to assemble an actin filament tail. Surprisingly, even small plastic beads
artificially coated in Lysteria’s ActA proteins can perform the same “trick”:
they assemble an actin tail which pushes the bead like a tiny rocket.
Figure G.2: Small spherical beads coated

with part of Lysteria’s special actin-
assembly kit also gain the ability to swim
around. Figure from [Bernheim-Groswasser
et al., 2002].
In a recent papr, Bernheim-Groswasser et al. [Bernheim-Groswasser et al.,

2002] describe the motion of these beads, shown in Figure G.2. When the
position of the bead is plotted on a graph with time as the horizontal axis,
(see Figure G.3) we find that the trajectory is not a simple one: it appears that
the bead slows down periodically, and then accelerates.
Figure G.3: The distance traveled by a little

bead is shown as a function of time. The
arrows point to times when the particle
slowed down or stopped. We can use this
data to analyze the velocity of the particles.
Figure from [Bernheim-Groswasser et al.,
2002].
With the techniques developed in Chapters 3 and 4, we can analyze the

experimental data shown in Figure G.3 to determine both the average velocity
of the beads, and the instantaneous velocity over the course of the motion.
Average velocity of the bead. We can get a rough idea of how fast the micro-
beads are moving by computing an average velocity over the time interval
shown on the graph. We can use two (approximate) data points (t, D(t )),
at the beginning and end of the run, for example (45, 20) and (80, 35): the
average velocity is
∆D
v̄ =
∆t
35 − 20
v̄ = ≈ 0.43µ min−1
80 − 45
so the beads move with average velocity 0.43 microns per minute. Units.
One micron is 10−6 meters.
The changing instantaneous velocity. Because the actual data points are
taken at finite time increments, the curve shown in Figure G.3 is not smooth.
We smoothen it, as shown in Figure G.4 for a simpler treatment. In Fig-
ure G.5 we sketch this curve together with a collection of lines that represent
the slopes of tangents along the curve. A horizontal tangent has slope zero:
this means that at all such points (also indicated by the arrows for emphasis),
the velocity of the beads is zero. Between these spots, the bead has picked up
speed and moved forward until the next time in which it stops.
We show the velocity v(t ), which is the derivative of the original func-
tion D(t ) in Figure G.6. As shown here, the velocity has periodic increases
and decreases.
Figure G.4: The (slightly smoothened) bead

trajectory is shown here.
Figure G.5: We inserted a sketch of the

tangent line configurations along the
trajectory from beginning to end. We
observe that some of these tangent lines are
horizontal, implying a zero derivative, and,
thus, a zero instantaneous velocity at that
time.
Figure G.6: Here we sketched the velocity

on the same graph.
G.4 Additional examples of geometric optimization
Rectangular box with largest surface area

We consider two examples of optimization where volumes, lengths, and/or
surface areas are considered.
Example G.1 (Wrapping a rectangular box) A box with square base and
arbitrary height has string tied around each of its perimeter. The total length
of string so used is 10 inches. Find the dimensions of the box with largest
surface area, i.e. determine the largest amount of wrapping paper needed to
y
wrap this box.
Solution. The total length of string is shown in Figure G.7. It consists of

three perimeters of the box is as follows:
x
L = 2(x + x) + 2(x + y) + 2(x + y) = 8x + 4y = 10.
x
This total length is to be kept constant, so this equation is the constraint in
this problem. This means that x and y are related to one another. We use this Figure G.7: A rectangular box is to be
wrapped with paper.
fact to eliminate one of them from the formula for surface area.
The surface area of the box is
S = 4(xy) + 2x2
since there are two faces (top and bottom) which are squares (area x2 ) and
four rectangular faces with area xy. At the moment, the total surface area S
is expressed in terms of both variables. Suppose we eliminate y from S by
rewriting the constraint in the form:
5
y= − 2x.
2
S(x)
Then

5
S(x) = 4x − 2x + 2x2 = 10x − 8x2 + 2x2 = 10x − 6x2 .
2
We show the shape of this function in Figure G.8. Note that S(x) = 0 at x = 0
and at 10 − 6x = 0 (which occurs at x = 5/3).
Since S is now expressed as a function of one variable, we can find its
critical points by setting S0 (x) = 0, i.e., solving x
0 5/6 5/3
S0 (x) = 10 − 12x = 0
Figure G.8: Figure for Example G.1; surface
for x: we get x = 10/12 = 5/6. To find the corresponding value of y we area of a box.
substitute our result back into the constraint. This results in

15 − 10

5 5 5
y = −2 = = .
2 6 6 6
Thus the dimensions of the box of interest are all the same, i.e. it is a cube
with side length 5/6.
We can verify that
S00 (x) = −12 < 0,
(indeed this holds for all x), which means that x = 5/6 is a local maximum.
Further, we can find that

2
5 5 5 25
S=4 +2 =
6 6 6 6
square inches, the maximum surface area of a box with such a constraint.
Figure G.8 shows how the surface area varies as the dimension x of the box is
varied.
A cylinder in a sphere
Example G.2 (Fitting a cylinder inside a sphere) Find the cylinder of
maximal volume that would fit inside a sphere of radius R.
Solution. We sketch a cylinder inside a sphere as in Figure G.9.
Figure G.9: Definition of variables and

geometry to consider.
R
R h/2 h/2
r r
It is helpful to add the radius of the sphere and of the cylinder. We define
the following:
h = height of cylinder,
r = radius of cylinder,
R = radius of sphere.
Then R is assumed a given fixed positive constant, and r and h are dimensions
of the cylinder to be determined.
From Figure G.9 we see that the cylinder fits if the top and bottom rims
touch the circle. When this occurs, the dark line in Figure G.9 is a radius of
the sphere, and so has length R.
The connection between the variables (our constraint) is given from
Pythagoras’ theorem by:
2
2 2 h
R =r + .
2
We maximize the volume of the cylinder,
V = πr2 h
subject to the above constraint.

Eliminating r2 using the Pythagoras theorem leads to
h2

2
V (h) = π R − h.
4
We see that the problem is very similar to the previous discussion. The reader
can show by working out the steps that
V 0 (h) = 0
occurs at the critical point

2
h= √ R
3
and this is a local maximum.
G.5 Optimal foraging with other patch functions
In Section 7.4, we computed an optimal foraging time for a specific patch

function f (t ) given by Eqn. (7.6). However, we can gain insight and obtain
an interesting result without making this assumption. We now consider a
similar analysis with a more general example.
Example G.3 Carry out the calculations for the optimal value patch res-
idence time for a general patch energy function f (t ), without using the
formula Eqn. (7.6).
Solution. We use the expression for R(t ) given by Eqn. (7.7). Differentiating,
we find the first derivative,
f 0 (t )(τ + t ) − f (t ) G(t )
R0 (t ) = 2
=
(τ + t ) H (t )
where
G(t ) = f 0 (t )(τ + t ) − f (t ), H (t ) = (τ + t )2 .
(The calculation is easier with this notation.) To maximize R(t ) we set
R0 (t ) = 0
which can occur only when the numerator of the above equation is zero, i.e.
G(t ) = 0.
This means that

f 0 (t )(τ + t ) − f (t ) = 0
so that, after simplifying algebraically,
f (t )
f 0 (t ) = . (G.3)
τ +t
A geometric argument. In practice, we need to specify a function for f (t ) in

order to solve for the optimal time t. However, we can also solve this problem
using a geometric argument.
Eqn (G.3) equates two quantities that can be interpreted as slopes. On
the right is the slope of a tangent line. On the left is the slope (rise over run)
of some right triangle whose height is f (t ) and whose base length is τ + t.
In Figure G.10, we show each slope on its own: in the left panel, f 0 (t ) is
the slope of the tangent line to the graph of f (t ). In the central panel, we
have constructed some triangle with the property that its hypotenuse has
slope f (t )/[τ + t ]. On the right panel we have superimposed both, selecting
a value of t for which the slope of the triangle is the same as the slope of the
tangent line.
energy gain energy gain

f (t)
f 0(t)
f (t) f (t)
t
τ +t τ 0 t
Figure G.10: The solution to the optimal

foraging problem can be expressed geomet-
Notice that in order to fit the triangle on the same diagram, we had to place rically in the form shown in this figure. The
its tip at the point −τ along the horizontal axis. When these slopes coincide, tangent line at the (optimal) time t should
it means that we have satisfied Eqn. (G.3), and found the desired time t for have the same slope as the hypotenuse of the
right triangle shown above. The diagram on
optimal foraging. the far right is sometimes termed the “rooted
We can use this observation to come up with the following steps to solve tangent” diagram.
an optimal foraging problem in general:
1. A biologist conducts some field experiments to determine the mean travel

time from food to nest, τ, and the shape of the energy gain function f (t ).
Note: this may require capturing the animal and examining the contents
of its stomach. We leave this task to our biological colleagues.
2. We draw a sketch of f (t ) as shown in rightmost panel of Figure G.10 and

extend the t axis in the negative direction. At the point −τ we draw a line
that just touches the curve f (t ) at some point (i.e. a tangent line). The
slope of this line is f 0 (t ) for some value of t.
3. The value of t at the point of tangency is the optimal time to spend in the
patch!
The diagram drawn in our geometric solution (right panel in Figure G.10) is
often called a “rooted tangent”.
We have shown that the point labeled t indeed satisfies the condition that
we derived above for R0 (t ) = 0, and hence is a critical point.
Checking the type of critical point. We still need to show that this solution
leads to a maximum efficiency, (rather than, say a minimum or some other
critical point). We do this by examining R00 (t ).
Recall that
G(t )
R0 (t ) =
H (t )
in terms of the notation used above. Then
G0 (t )H (t ) − G(t )H 0 (t )
R00 (t ) = .
H 2 (t )
But, according to our remark above, at the patch time of interest (the candi-
date for optimal time),
G(t ) = 0
so that
G0 (t )H (t ) G0 (t )
R00 (t ) = = .
H 2 (t ) H (t )
We substitute the derivative of G0 (t ), H (t ) into this ratio:
G(t ) = f 0 (t )(τ + t ) − f (t ) ⇒ G0 (t ) = f 00 (t )(τ + t ) + f 0 (t ) − f 0 (t )

= f 00 (t )(τ + t )
We find that
f 00 (t )(τ + t ) f 00 (t )
R00 (t ) = 2
= .
(τ + t ) (τ + t )
The denominator of this expression is always positive, so the sign of R00 (t ) is
the same as the sign of f 00 (t ). But in order to have a maximum efficiency at
some residence time, we need R00 (t ) < 0. This tells us that the gain function
has to have the property that f 00 (t ) < 0, i.e. has to be concave down at the
optimal residence time.
Returning to some of the shapes of the function f (t ) that we saw in
Figure 7.7, we see that only some of these lead to an optimal solution. In
cases (1), (2), (4) the function f (t ) has no points of downwards concavity
on its graph. This means that in such cases there is no local maximum. The
optimal efficiency would then be attained by spending as much time as
possible in just one patch, or as little time as possible in any patch, i.e. it
would be attained at the endpoints.
G.6 Trigonometric functions and differential equations
As we saw in Chapter 15, the functions sin(t ) and cos(t ) are related to one
another via differentiation: one is the derivative of the other (with a multiple
of the factor (−1)):
d sin(t ) d cos(t )
= cos(t ), = − sin(t ).
dt dt
The connection becomes even clearer when we examine the second deriva-
tives of these functions:
d 2 sin(t ) d cos(t ) d 2 cos(t ) d sin(t )
2
= = − sin(t ), 2
=− = − cos(t ).
dt dt dt dt
Thus, for each of the functions y = sin(t ), y = cos(t ), we find that the function
and its second derivative are related to one another by the differential equa-
tion (DE) d 2 y/dt 2 = −y. Here the highest derivative is a second derivative,
and we denote this a second order DE.
More generally, we make the following observations:
The functions
x(t ) = cos(ωt ), y(t ) = sin(ωt )
satisfy a pair of differential equations,
dx dy
= −ωy, = ωx.
dt dt
The functions
x(t ) = cos(ωt ), y(t ) = sin(ωt )
also satisfy a related differential equation with a second derivative
d2x
= −ω 2 x.
dt 2
These follow by the same reasoning, where the chain rule is applied in
differentiation.
Students of physics may recognize the equation that governs the behaviour
of a harmonic oscillator, and see the connection between the circular motion
of our point on the circle, and the differential equation for periodic motion.
H
Short Answers to Problems
Answers to Chapter 1 Problems

• Problem 1.1:
• Problem 1.2:
(a) Stretched in y direction by factor A; (b) Shifted up by a; (c) Shifted in
positive x direction by b.
• Problem 1.3:
Not Provided
• Problem 1.4:
• Problem 1.5:
√
(a) x = 0, (3/2)1/3 ; (b) x = 0, x = ± 1/4.
• Problem 1.6:
(b) a < 0: x = 0; a ≥ 0: x = 0, ±a1/4 ; (c) a > 0.
• Problem 1.7:
A 1/(m−n) A 1/(m−n)

if m − n even: x = ± B ,x = 0; if m − n odd: x = B ,x =
0
• Problem 1.8:
√ √
7 3 7 3
(a) (0, 0) and (1, 1); (b) (0, 0); (c) ( 2 , 4 ), (− 2 , 4 ), and (0, −1).
• Problem 1.9:
p
γ± γ 2 − 4Iε
(a) x = I/γ, (b) x = .
2ε
• Problem 1.10:
• Problem 1.11:
y = xn ; y = x−n ; y = x1/n , n = 2, 4, 6, . . .; y = x−n , n = 1, 2, 3, . . .
• Problem 1.12:
m > −1
• Problem 1.13:
• Problem 1.14:
Not Provided
• Problem 1.15:
1
B b−a
x=
A
• Problem 1.16:
(a) x = 0, −1, 3; (b) x = 1; (c) x = −2, 1/3; (d) x = 1.
S H O RT A N S W E R S T O P RO B L E M S 409
• Problem 1.17:
(a) Intersections x = −1, 0, 1.
• Problem 1.18:
V 1 1 1
√
(a) V ; (b) = a, a > 0; (c) a = V 3 ; a = ( 16 S) 2 ; a = 10 cm; a = 315 cm.
S 6
• Problem 1.19:
r 3 1/3 1/3
1 1/2 1/2

(a) V ; (b) ; (c) r = 4π V ; r = 4π S ; r ≈ 6.2035 cm;
3
r ≈ 0.8921 cm.
• Problem 1.20:
r = 2k1 /k2 = 12µm.
• Problem 1.21:
(1 − a)S 1/4

(a) T = .
εσ
• Problem 1.22:
3V 2/3
d/b
(a) P = C AR

; (b) S = 4π 4π .
• Problem 1.23:
(a) a: Ms−1 , b: s−1 ; (b) b = 0.2, a = 0.002; (c) v = 0.001.
• Problem 1.24:
(a) v ≈ K, (b) v = K/2.
• Problem 1.25:
K ≈ 0.0048, kn ≈ 77 nM
• Problem 1.26:
(a) x = −1, 0, 1 (b) 1 (c) y1 (d) y2 .
• Problem 1.27:
Line of slope a3 /A and intercept 1/A
• Problem 1.28:
K = 0.5, a = 2
• Problem 1.29:
Not Provided
• Problem 1.30:
m ≈ 67, b ≈ 1.2, K ≈ 0.8, kn ≈ 56
• Problem 1.31:
1
x = AR r−a .

• Problem 2.1:
(a) m = 28◦ /min, b = 50.
• Problem 2.2:
(a) −4.91◦ F/min. (b) -7, -8, -9 ◦ F/min. (c) -9 ◦ F/min.
• Problem 2.3:
Displacements have same magnitude, opposite signs.
• Problem 2.4:
(a) −14.7 m/s; (b) −gt − gε
2 ; (c) t = 10 s.
• Problem 2.8:
v0 − g/2
• Problem 2.5:
5.8, 4.4, 5.4, 12.4, 4, 4.4, 7.2 (km/hr)
• Problem 2.6:
v̄ = 13.23 m/s.; secant line is y = 13.23x − 2.226
• Problem 2.7:
• Problem 2.9:
(a) 2; (b) 0; (c) −2; (d) 0.
• Problem 2.10:
32
(a) 3; (b) 5.55; (c) 3 .
• Problem 2.11:
√ √
2 2 6(1− 2)
(a) π ; (b) π ; (c) π/4 ≤ x ≤ 5π/4 (one solution).
• Problem 2.12:
(a) 2 + h; (b) 2; (c) y = 2x.
• Problem 2.13:
2h2 + 25h + 104; slope =104
• Problem 2.14:
(b) 0, −4, −1.9, −2.1, −2 − h; (c) −2.
• Problem 2.15:
(a) 2 + h; (b) 2; (c) 2.98.
• Problem 2.16:
√
4 3−12
(a) π4 ; (b) π .
• Problem 2.17:
−2
(a) −1; (b) 1+ε ; (c) Slope approaches -2; (d) y = −2x + 4.
• Problem 2.18:
(a) v(2) = 12 m/s; v̄ = 15 m/s; (b) v(2) = 0 m/s; v̄ = 25 m/s; (c)
v(2) = 13 m/s; v̄ = 11 m/s.
• Problem 2.19:
0
• Problem 2.20:
−1
(x +1)2

• Problem 3.1: – Problem 3.5:
Not Provided
• Problem 3.6:
(a) 14.7 m/s; (b) −4.9 m/s.
• Problem 3.7:
Not Provided
• Problem 3.8:
√
(a) f 0 (x) = 1/(2 x), (b) 0.25 (c) y = 2 + 0.25(x − 4).
• Problem 3.9:
1
f 0 (x) = −C
(x + a)2
• Problem 3.10:
Not Provided.
• Problem 3.11:
(a) R, U, R, L, U, R, L
• Problem 3.12:
Not Provided
• Problem 3.13:
5; 5; no change; linear function
• Problem 3.14:
5.
• Problem 3.15:
y = 5.8x − 6.825.
• Problem 3.16:
(a) Horizontal asymptotes at y = 1.
• Problem 3.17:
Not Provided.

• Problem 4.1:
(a) 36x2 − 16x − 15; (b) −12x3 + 3x2 − 3; (c) 4x3 − 18x2 − 30x − 6; (d) 3x2 ;
8
36x 6x3 − 3x2 + 6 18b2 − 7b 3 −36m3 + 72m2 − 36m + 5
(e) 2 ; (f) ; (g) 2 ; (h) ;
(x + 9)2 (1 − 3x)2 3(2 − b 3 )2 (3m − 1)2
9x4 + 8x3 − 3x2 − 4x + 6
(i) .
(3x + 2)2
• Problem 4.2:
dR r
= r − 2 N.
dN K
• Problem 4.3:
dv kn dy nxn−1 an
(a) =K , (b) = A .
dx (kn + x)2 dx (an + xn )2
• Problem 4.4:
dV dS
(a) = 4πr2 , and = 8πr.
dr dr
d (S/V ) 1
(b) = −3 2
dr r
• Problem 4.5:
S3/2 1
(a)V = , (b) dV /dS = · S1/2 .
6(π )1/2 4(π )1/2
• Problem 4.6:
(a) V 0 (r ) = 2πL + 4πr, S0 (r ) = 2πL + 4πr, (b) 2/r2 .
• Problem 4.7:
A0 (2) = 3πmm2 /hr
• Problem 4.8:
(a) E
• Problem 4.9:
d (N1 /N2 ) N1
= (k1 − k2 )
dt N2
• Problem 4.10:
k2 > k1 .
• Problem 4.11:
Not provided.
• Problem 4.12:
1 1 1 √
(a) y(t ) = t 5 + t 3 − t 2 + 3t + C, (b) y(x) = − x2 + 2x + C, (c)
5 2 2
1
y = | x2 | + C.
2
• Problem 4.13:
p p
(a) a(t ) = −2Bt, (b) y(t ) = At − (B/3)t 3 , (c) t = 3A/B, (d) t = A/B,
(e) v = A.
• Problem 4.14:
p p
(a) v = 3t 2 + 6t, a = 6t + 6; (b) t = 0, 3/a; (c) t = 0, 3/2a; (d)
√
t = 1/ 2a.
• Problem 4.15:
v20
(a) t = v0 /g; (b) h0 + ; (c) v = 0.
2g
• Problem 4.16:– Problem 4.18:
Not Provided

• Problem 5.1:
(a) no tangent line; (b) y = −(x + 1); (c) y = (x + 1).
• Problem 5.2:
y = 2x − 3
• Problem 5.3:
(b) a = 2.
• Problem 5.4:
(a) y = −4x + 5; (b) x = 5/4, y = 5; (c) y = 0.6, smaller.
• Problem 5.5:
(a) y = 3x − 2; (b) x = 2/3; (c) 1.331; 1.3.
• Problem 5.6:
f (x0 )
(a) y = f 0 (x0 )(x − x0 ) + f (x0 ); (b) x = x0 − .
f 0 (x0 )
• Problem 5.7:
2.83
• Problem 5.8:
(a) (3.41421, 207.237), (0.58580, 0.762), (−0.42858, −0.895);
• Problem 5.9:
(a) x = 0.32219; (b) x = 0.81054; (c) x = 0.59774, x = −0.68045, x =
−4.91729.
• Problem 5.10:
(3, 9), (1, 1)
• Problem 5.11:
19
(a) 6 ; (b) 3.
• Problem 5.12:
(a) 0.40208; (b) 5.99074.
• Problem 5.13:
0.99
• Problem 5.14:
−2.998
• Problem 5.15:
1030 cm3
• Problem 5.16:
Not provided.
• Problem 5.17:
10.2469507659596

• Problem 6.1:
√
(a) zeros: x = 0, x = ± 3; loc. max.: x = −1; loc. min.: x = 1; (b) loc.
min.: x = 2; loc. max.: x = 1; (c) (a): x = 0; (b): x = 3/2.
• Problem 6.2:
√ √
Zeros at 0, ± a; inflection point at 0, local maximum at − a/3, local
√
minimum at − a/3.
• Problem 6.3:
(a) f 0 (x) = 2x, f 0 (0) = 0, f 0 (1) = 2 > 0, f 0 (−1) = −2 < 0. Local
minimum at x = 0; (b) f 0 (x) = −3x2 , f 0 (0) = 0, f 0 (1) = −3 < 0,
f 0 (−1) = −3 < 0. No local maxima nor minima; (c) f 0 (x) = −4x3 ,
f 0 (0) = 0, f 0 (1) = −4 < 0, f 0 (−1) = 4 > 0. Local maximum at x = 0.
• Problem 6.4:
Zeros at 0, 2, Inflection point at 0, local min at 3/2.
• Problem 6.5:
Global maximum at x = 3, global minimum at x = 3/2.
• Problem 6.6:
(a) max.: 18; min.: 0; (b) max.: 25; min.: 0; (c) max.: 0; min.: −6; (d)
max.: −2; min.: −17/4.
• Problem 6.7:
(a) increasing: −∞ < x < 0, 1.5 < x < ∞; decreasing for 0 < x < 1.5; (b) 0,
local maximum; 1.5, local minimum; (c) No.
• Problem 6.8:
min.: 3/4
• Problem 6.9:
x=0
• Problem 6.10:
√
1 3
critical points: x = 0, 1, 1/2; inflection points: x = ±
2 6
• Problem 6.11:
Not Provided
• Problem 6.12:
a = 1, b = −6, c = 7
• Problem 6.13:
Not Provided
• Problem 6.14:
√ √
min. at x = − 3; max. at x = 3; c.u.: x < −1, 0 < x < 1; c.d.: −1 < x <
0,x > 1; infl.pt.: x = 0
• Problem 6.15:
loc. min.:x = a loc. max.: x = −2a
• Problem 6.16:
(a) increasing: x < 0, 0 < x < 3k, x > 5k; decreasing:
√ 3k < x < 5k;√loc. max.:
6 6
x = 3k; loc. min.:x = 5k; (b) c.u.: 0 < x < (3 − )k, x > (3 + )k; c.d.:
√ √ 2 √ 2
6 6 6
x < 0, (3 − )k < x < (3 + )k; infl.pts.: x = 0, (3 ± )k.
2 2 2
• Problem 6.17:
(a + p0 )
(b) dv/d p = −b ; (c) p = p0 .
( p + a)2
• Problem 6.18:
Not Provided
• Problem 6.19:
abs. max. of 4.25 at end points; abs. min. of 2 at x = 1

• Problem 7.1:
(a) 10, 10; (b) 10, 10; (c) 12, 8.
• Problem 7.2:
(a) v(t ) = 120t 2 − 16t 3 ; (b) t = 5; (c) t = 7.5.
• Problem 7.3:
9 : 24A.M., 15 km
• Problem 7.4:
(a) t ≈ 1.53 sec; (b) v(0.5) = 10.1 m/sec, v(1.5) = 0.3 m/sec, a(0.5) =
−9.8 m/sec2 , a(1.5) = −9.8 m/sec2 ; (c) t ≈ 3.06 sec.
• Problem 7.5:
See Example 7.2.
• Problem 7.6:
(a) N (r ) = 2k1 πrL − πk2 r2 L, (b) r = k1 /k2 .
• Problem 7.7:
Not provided.
• Problem 7.8:
30 × 10 × 15 cm
• Problem 7.9:
√ √
(a) y = (1/ 3); (b) 3/9.
• Problem 7.10:
√
|a| if a < 4; 2 2a − 4 if a ≥ 4
• Problem 7.11:
A = 625 ft2
• Problem 7.12:
All of the fencing used for a circular garden.
• Problem 7.13:
√
Squares of side 6 − 2 3 cm.
• Problem 7.14:
Not provided.
• Problem 7.15:
A square with A = L2 /2
• Problem 7.16:
16
Straight lines from (10, 10) to ( , 0) then to (3, 5).
3
• Problem 7.17:
4 ◦C
• Problem 7.18:
(a) x = 2B/3, R = (4/27)AB3 ; (b) x = B/3, S = AB2 /3.
• Problem 7.19:
r = 2k1 /k2
• Problem 7.20:
Not provided.
• Problem 7.21:
√
h = 20, r = 5 2
• Problem 7.22:
a a−m
(b) x = ; (c) x = 0; (d) x = .
2b 2b
• Problem 7.23:
x = (A/2B)1/3 − 1
• Problem 7.24:
Not provided.
• Problem 7.25:
NMSY = K (1 − qE/r )
• Problem 7.26:
E = r/2q
• Problem 7.27:
• Problem 7.28:
√
topt = kτ.
• Problem 7.29:
r
k
t = Emax − k.
r

• Problem 8.1:
(a) y0 (x) = 5(x + 5)4
• Problem 8.2:
dN
== 2.0793 b k(a − b2−kt )2 2−kt .
dt
• Problem 8.3:
1/4
(1 − a)S

dT 1 dε
(a) =− ε −5/4 .
dG 4 σ dG
• Problem 8.4:
(c) Global minimum occurs at an endpoint, rather than at a critical point.
• Problem 8.5:
Not Provided
• Problem 8.6:
(c) d = 3D/4.
• Problem 8.7:
√
3± 5
(a) V (x)= 500x3 + 300(1 − x)5 , (b) Critical point at x1 = is a local
2
minimum. (c) Best strategy is x1 = 1, x2 = 0.

• Problem 9.1:
3k
(a) 4πr2 k; (b) 8πrk; (c) − .
r2
• Problem 9.2:
(a) dA/dt = 2πrC; (b) dM/dt = α2πrC.
• Problem 9.3:
dM
= Cπ (3r2 )a
dt
• Problem 9.4:
dV
= 1 m3 /min
dt
• Problem 9.5:
1 2
(a) cm/s; (b) cm2 /s.
300π 5
• Problem 9.7:
5 cm/s
• Problem 9.8:
dV nR dT dV nRT dP
(a) = ; (b) =− 2 .
dt P dt dt P dt
• Problem 9.9:
π −1/2 k.
• Problem 9.6:
1
− cm/min
10π
• Problem 9.10:
1 cm/sec toward lens
• Problem 9.11:
dh −1
= cm/min
dt 36π
• Problem 9.12:
1 4π
k= +
10 45
• Problem 9.13:
dh 6
= ft/min
dt 5π
• Problem 9.14:
2
h0 (5) = m/min
5π
• Problem 9.15:
1 1
(a) 4π m/min; (b) π m/min.
• Problem 9.16:
25
(a) −4 m/s; (b) − per sec.
32
• Problem 9.17:
−1
√ m/min
4 6
• Problem 9.18:
dS ab
= ; no.
dA 1 + bA
• Problem 9.19:
dy l2
= 3 cm/hr
dt l1
• Problem 9.20-9.21:
Not Provided
• Problem 9.22:
dy 2x x 1 √ √ √
(a) = − = − ; (b) y = √ (−x + r 3); y = (1/ 2)(x − r 3).
dx 2y y 2
• Problem 9.23:
dy 21x2 + 2 dy 2y
(a) = 5 ; (b) =− y ;
dx 6y + 3 dx e + 2x
• Problem 9.24:
(a) −3/4; (b) y = −(3/4)x + 8.
• Problem 9.25:
Not Provided
• Problem 9.26:
2 9 2 9
( √ , √ ) and (− √ , − √ )
10 10 10 10
• Problem 9.27:
4y p
m=
xp
• Problem 9.28:
dy (ay − x2 )
(b) = 2 ; (c) x = 0, x = 21/3 a; (d) No.
dx (y − ax)
• Problem 9.29:
dp a a
(a) = (2 3 ) − ( p + 2 ) / (v − b).
dv v v
• Problem 9.30:
(0, 5/4)
• Problem 9.31:
4
(a) y − 1 = −1(x − 1); (b) y00 = ; (c) concave up.
5

• Problem 10.1:
235 ≈ 32 · 109 .
• Problem 10.2:
264 ≈ 1.6 · 1019 .
• Problem 10.3:
11.312, 0.7072, 0.03125
• Problem 10.4:
Not provided
• Problem 10.5:
dy
= C3 · 3x and C3 = ln(3) = 1.098.
dx
• Problem 10.6-10.7 :
Not Provided
• Problem 10.8:
(a) 50.75 > 50.65 ; (b) 0.4−0.2 > 0.40.2 ; (c) 1.0012 < 1.0013 ; (d) 0.9991.5 >
0.9992.3 .
• Problem 10.9:
Not Provided
• Problem 10.10:
b
(a) x = a2 b3 ; (b) x = 2 .
c3
• Problem 10.11:
Not Provided
• Problem 10.12:
3 − ln(5) e4 + 1 2 ln(C )
(a) x = ; (b) x = ; (c) x = e(e ) = ee·e ; (d) x = .
2 3 a−b
• Problem 10.13:
dy 6 dy 6[ln(2x + 3)]2 dy 1 1
(a) = ; (b) = ; (c) = − tan x; (d)
dx 2x + 3 dx 2x + 3 dx 2 2
dy 3x2 − 2 dy 2 dy 1 1 dy
= 3 ; (e) = 6xe3x ; (f) = − a− 2 x ln a; (g) =
dx (x − 2x) ln a dx dx 2 dx
dy x dy 4
x2 2x (3 + x ln 2); (h) = ee +x ; (i) = t .
dx dx (e + e−t )2
• Problem 10.14:
2 2 1
(a) min.: x = √ ; max.: x = − √ ; infl.pt.: x = 0; (b) min.: x = √ 3
; (c)
3 3 3
max.: x = 1; inf.pt.: x = 2; (d) min.: x = 0; (e) min.: x = 1; max.: x = −1;
(f) min.: x = ln(2); infl. pt.: x = ln(4).
• Problem 10.15:
C = 4, k = −0.5
• Problem 10.16:
(a) decreasing; (b) increasing; y1 (0) = y2 (0) = 10; y1 half-life = 10 ln(2);
y2 doubling-time = 10 ln(2)
• Problem 10.17:
41.45 months.
• Problem 10.18:
(c) r = 0.0101 per year, C = 0.7145 billions.
• Problem 10.19:
Not provided
• Problem 10.20:
r1 ≈ −14.5, r2 ≈ −0.9 per unit time, C1 = 55,C2 = 45.
• Problem 10.21- 10.23:

Not Provided
• Problem 10.24:
crit.pts.: x = 0, x ≈ ±1.64; f (0) = 1; f (±1.64) ≈ −0.272
• Problem 10.25:
(a) x = 1/β , (b) x = ln(α )/β .
• Problem 10.26:

ar R
(a) x = r; (c) x = ln ; (d) decrease; (e) decrease.
a−r A
• Problem 10.27:
q
x = b ln((a2 + b2 )/b2 )
• Problem 10.28:
Not Provided

• Problem 11.1:
Not Provided
• Problem 11.2:
(a) C any value, k = −5; (b) C any value, k = 3.
• Problem 11.3:
Not Provided
• Problem 11.4:
(a) y(t ) = Ce−t ; (b) c(x) = 20e−0.1x ; (c) z(t ) = 5e3t .
• Problem 11.5:
e2.08t .
• Problem 11.6:
Not provided
• Problem 11.7:
(a) P(5) ≈ 1419; (b) t ≈ 9.9 years.
• Problem 11.8:
dN
= 0.05N; N (0) = 250; N (t ) = 250e0.05t ; 2.1 × 1010 rodents
dt
• Problem 11.9:
(a) dy/dt = 2.57y; (b) dy/dt = −6.93y.
• Problem 11.10:
(a) 12990; (b) 30792 bacteria.
• Problem 11.11:
1.39 hours; 9.2 hours
• Problem 11.12:
(a) y1 growing, y2 decreasing; (b) 3.5, 2.3; (c) y1 (t ) = 100e0.2t , y2 (t ) =
10000e−0.3t ; (d) t ≈ 9.2 years.
• Problem 11.13:
12265 people/km2
• Problem 11.14:
6.93 years
• Problem 11.15:
(a) 1 hour; (b) r = ln(2); (c) 0.25 M; (d) t = 3.322 hours.
• Problem 11.16:
20 min; 66.44 min
• Problem 11.17:
ln(2)
τ=
ln(10)
• Problem 11.18:
(a) 57300 years; (b) 22920 years
• Problem 11.19:
(a) 29 years; (b) 58 years; (c) 279.7 years.
• Problem 11.20:
(a) 80.7%; (b) 12.3 years.
• Problem 11.21:
1.7043 kg
• Problem 11.22:
y ≈ 707.8 torr

• Problem 12.1:
√
2 h0
Until t = .
k
• Problem 12.2:
• Problem 12.3-12.7:
Not Provided
• Problem 12.8:
(a) C = −12; (b) C1 = 1, C2 = −5; (c) C1 = −1, C2 = 0.
• Problem 12.9:
g g g
(a) v(t ) = − e−kt + ; (b) v = .
k k k
• Problem 12.10:
k k
c(t ) = − e−st +
s s
• Problem 12.11-12.12::
Not provided
• Problem 12.13:
(b) 46 minutes before discovery.
• Problem 12.14:
10.6 min
• Problem 12.15:
(a) Y = Y0 − kt, (d) k = 0.0333 per min.
• Problem 12.16:
(b) k = 3/2.
• Problem 12.17:
(a) Input rate I, αF fish caught per day. Birth and mortality neglected. (b)
Steady state level F = I/αN. (c) 2 ln(2)/αN days. (d) t = Flow /I days.
• Problem 12.18:
Not provided
• Problem 12.19:
64.795 gm, 250 gm
• Problem 12.20:
Q r
(a) Q0 (t ) = kr − r = − [Q − kV ]; (b) Q = kV ; (c) T = V ln 2/r.
V V
• Problem 12.21:
dQ −2
(a) = kQ; Q(t ) = 100e(−8.9×10 )t ; (b) 7.77 hr.
dt
• Problem 12.22:
√
2A y0 √
(b) y0 ; (c) t = ; (d) −k y0 .
k
• Problem 12.23:
a = 0, b = −1
• Problem 12.24:
95.12, 90.48.
• Problem 12.25:
(a) y5 = 1.61051; y(0.5) = 1.6487213; error = 0.03821; (b) y5 = 0.59049;
y(0.5) = 0.60653; error = 0.01604.
• Problem 12.26-12.27:
Not Provided

• Problem 13.1: Not provided
• Problem 13.2:
(a) Steady state: y = 0. If y(0) = 1, y → 0.

(b) Steady states: y = 0, 2. If y(0) = 1, y → 2.
(c) Steady states: y = 0, 2, 3. If y(0) = 1, y → 2.
• Problem 13.3:
(a) At y = − 23 , TL slope is 0. If y < − 23 , TL slope negative. If y > 32 , TL

slope positive.
(b) Slope = 0 at y = 0 and y = 2. If y < 0, slope positive. If 0 < y < 2,
slopes negative. If y > 2, slopes are positive.
(c) Slope = 0 at y = 1 and y = 2. If y < 1, the slopes are positive. If
1 < y < 2, the slopes are negative. If y > 2, the slopes are positive.
(d) Slope = 0 at y = 3. If y < 3, slopes are negative. If y > 3, slopes are also
negative.
(e) y(t ) increases for all real t.
(f) The slopes are zero at y = −1, 0, 1. If y < −1, the slopes are negative. If
−1 < y < 0, the slopes are positive. If 0 < y < 1, the slopes are negative.
If y > 1, the slopes are positive.
(g) If 0 < y < 2, the slopes are negative. If 2 < y < 3, the slopes positive. If
y > 3, the slopes are positive.
• Problem 13.4:
Linear differential equations: only (a)
• Problem 13.5:
Not Provided
• Problem 13.6:
(B)
• Problem 13.7:
(B)
• Problem 13.8:
h(t ) → (I/K )2 .
• Problem 13.9:
dx k 1
(a) = (V0 − x3 ); (d) V = V0 .
dt 3 2
• Problem 13.10:
Not provided
• Problem 13.11:
Kmax
(a) Kmax , c = k; (b) ln(2)/r; (c) c = 0, c = − k.
r
• Problem 13.12-13.13:
Not provided
• Problem 13.14:
dy/ds = y(1 − y)
• Problem 13.15:
0.55, 0.5995, 0.6475, 0.6932, 0.7357
• Problem 13.16:
Not provided
• Problem 13.17:
φ = 0.857 = 86%.
• Problem 13.18:
(c) Steady states at at y2 = 0 and y2 = P − a/b. (d) Social media persists if
Pb/a > 1.
• Problem 13.19:
β p
(b) Stable steady state at a = −1 + 1 + 4µM/β
2µ

• Problem 14.1:
(a) 180o (b) 300o (c) 164.35o (d) 4320o
(e) 5π/9 (f) 2π/45 (g) 5π/2 (h) π/2
√ √
(i) 1/2 (j) 2/2 (k) 3/3
• Problem 14.2-refprob:chap10-6:
Not Provided
• Problem 14.5:
(a) T (t ) = 37.1 + 0.4 cos[π (t − 8)/12]; (b) W (t ) = 0.5 + 0.5 cos[π (t −
8)/6].
• Problem 14.6:
r
g 2π π
(a) S = 3 cos t ; (b) y = 2 sin t+ + 10.
l 3 6
• Problem 14.7:
(E)
• Problem 14.8:
√ √
(a) x; (b) x/ 1 − x2 ; (c) 1 − x2 .
• Problem 14.9:
(D)
• Problem 14.10:
(B)
• Problem 14.11:
Not provided

• Problem 15.1:
dy dy dy 2 2 √ √
(a) = 2x cos x2 ; (b) = sin 2x; (c) = − x− 3 (cot 3 x)(csc2 3 x);
dx dx dx 3
dy dy
(d) = (1 − 6x) sec(x − 3x2 ) tan(x − 3x2 ); (e) = 6x2 tan x + 2x3 sec2 x;
dx dx
dy cos x + x sin x dy dy sin 2x
(f) = 2
; (g) = cos x − x sin x; (h) = 2 1
; (i)
dx cos x dx dx x2 esin x
dy dy
= 6(2 tan 3x + 3 cos x)(2 sec2 3x − sin x); (j) = − sin(sin x) · cos x +
dx dx
cos 2x.
• Problem 15.2:
p
0 −(4x3 + 10x) sin(ln(x4 + 5x2 + 3)) 0 (3x2 − 2 cos(x) sin(x)) cos( cos2 (x) + x3 )
(a) f (x) = ; (b) f (x) = ;
(x4 + 5x2 + 3)
p
(2 cos2 (x) + x3 )
1
(c) f 0 (x) = 6x2 + ; (d) f 0 (x) = 4(x2 ex + tan(3x))3 (2xex + x2 ex +
x ln(3)
3x2 (sin2 (x) cos(x) − cos2 (x) sin(x))
q
3 sec2 (3x)); (e) f 0 (x) = 2x sin3 (x) + cos3 (x) + q .
2 sin3 (x) + cos3 (x)
• Problem 15.3:
√
− 3/20, 1/20
• Problem 15.4:
(a) [0, π/4], [5π/4, 2π ]; (b) [3π/4, 7π/4]; (c) x = 3π/4, 7π/4.
• Problem 15.5:
±( π8 , 1)
• Problem 15.6:
−0.021 rad/min
• Problem 15.7:
0.125 radians per minute
• Problem 15.9:
dy Cx dθ C
(a) = −√ ; (b) = .
dt L2 − x2 dt y
• Problem 15.8:
1 2
R= 32 v0
• Problem 15.10:
8π m/s; 0 m/s
• Problem 15.11:
Not Provided
• Problem 15.12:
S = 2v/Kcrit
• Problem 15.13:
v = Kcrit S/4
• Problem 15.14:
dy 1 dy 1 dθ
(a) = q ; (b) = 2√ ; (c) =
dx 2 2 dx 3(arcsin x) 3 1 − x2 dr
3x 3 1 − x 3
1 dy dy −2x2 + a2 − a
2
; (d) = arcsec 1x − √ x 2 ; (e) = √ ;
2r + 2r + 1 dx 1−x dx a a2 − x2
1 + t 2 2(1 − t 2 )

dy
(f) = − · .
dt 1 − t 2 (1 + t 2 )2
• Problem 15.15:
30π cm/s; to the right
• Problem 15.16:
√ √
(a) h2 + 2hR; (b) −v h2 + 2hR/R.
• Problem 15.17:
dy 4 sec2 (2x + y) dy 2 sin x dy y cos x + sin y
(a) = ; (b) = ; (c) =− .
dx 1 − 2 sec2 (2x + y) dx cos y dx x cos y + sin x
• Problem 15.18:
y = −x + 2
• Problem 15.19:
√
y0 = a/ 1 − a2 x2
• Problem 15.20:
0.4 m
• Problem 15.21:
5
26 rad/s
• Problem 15.22:
Not Provided
• Problem 15.23:
p
(c) y(t ) = A cos( g/Lt ).
• Problem 15.24:
(a) π/8; (b) 5π/8.
• Problem 15.25:
(b) t = π/4 + nπ.
H.0.1 Answers to Chapter 16 Problems

• Problem 16.1:
1(d), 2(b), 3(a), 4(c) , 5(a), 6(a), 7(d), 8(b), 9(e), 10(a).
• Problem 16.1:
(E)
• Problem 16.1:
(E)
• Problem 16.1:
(D)
• Problem 16.2:
3/4 cm/sec.
• Problem 16.3:
(a) v(t ) = 3(t + 2)2 + λ , a(t ) = dv/dt = 6(t + 2). (b) λ .
• Problem 16.4:
(a) t = 1 and t = 4, (b) 9 m/s (c) 12 m.
• Problem 16.5:
Local max at t = v0 /g.
• Problem 16.10:
(a) Inflection point at x = 1. (b) Local maximum.
• Problem 16.17:
(a) Critical points at x = 0, 3, inflection at x = 9/4. (b) Global min x = −1,
global max x = 0.
• Problem 16.18:
Local min x = −9/5, inflection point x = 0.
• Problem 16.6:
4.0004
• Problem 16.7:
ln(350)
t= 0.05 days.
• Problem 16.8:
(a)x = 0, ln(A)/b (b)x = 1/b.
• Problem 16.9:
p(x) = −x3 + 3x.
• Problem 16.11:
Local min at x = (2a/b)1/6 .
• Problem 16.12:
√
Square of side length r/ 2.
• Problem 16.13:
x = 41 P.
• Problem 16.14:
dAlea f
dt = 2y(t ) + x(t ) + πy2(t )
• Problem 16.15:
q
(a) r = 21 πS and h = 0, (b) a sphere, (c) No.
• Problem 16.16:
dV
= 2π [r (t )h(t ) + r (t )2 ] + 4πr (t )2 .

dt
• Problem 16.19:
• Problem 16.20:
k ≈ 1/27
• Problem 16.21:
(a) k = ln(2)/3 per hr (b) 6 more hrs.
• Problem 16.22:
(a) dP/dt = C − γP (b) P = C/γ (d) t = (ln(1/0.2)/γ.
• Problem 16.23:
(a) t = 3h (b) 23 h.
• Problem 16.24:
(a) y = 4S (b) T = 2τ1/2 = 2 ln(2)/ f (c) y → 50
• Problem 16.25:
√
y = D/ 8
• Problem 16.29:
0.125m/s
• Problem 16.28:
696 ys.
• Problem 16.30:
200
√ 1/2
(a) π/150 radians/s (b) π/6 radians (c) D = π (2 − 3)
dD 2
(d) = √
dt 3(2 − 3)1/2
• Problem 16.31:
(a) φ = arcsin(A/B), (b) −1 < arcsin(A/B) < 1, (c) φ → π/6
• Problem 16.27:
(b) y(0) = 0.5 (c) at y = 0.5 (d) y → 1.
• Problem 16.32:
7
(a) r10 (t ) = cm/year, (b) S0 (t ) = 14.
4π
• Problem 16.33:
√
(a) BA length L − x, BP length x2 + d 2 .
p r4 d
(b) Resistance of BA is (L − x)/R4 , BP is x2 + d 2 /r4 . (c) √ .
R8 − r8
• Problem 16.26:
• Problem 16.34:
1
(a) , (b) 0, (c) 1.
tan(x) tan(y)
Bibliography
Anne Bernheim-Groswasser, Sebastian Wiesner, Roy M Golsteyn, Marie-

France Carlier, and Cécile Sykes. The dynamics of actin-based motility
depend on surface parameters. Nature, 417(6886):308–311, 2002.
Benjamin P Bouchet, Ivar Noordstra, Miranda van Amersfoort, Eugene A

Katrukha, York-Christoph Ammon, Natalie D Ter Hoeve, Louis Hodg-
son, Marileen Dogterom, Patrick WB Derksen, and Anna Akhmanova.
Mesenchymal cell invasion requires cooperative regulation of persistent mi-
crotubule growth by slain2 and clasp1. Developmental cell, 39(6):708–723,
2016.
C.M. Breder. Structure of a fish school. Bull. Amer. Mus. Nat. Hist., 98:1–27,
1951.
C.M. Breder. Equations descriptive of fish schools and other animal aggrega-
tions. Ecology, pages 361–370, 1954.
Eric L Charnov. Optimal foraging, the marginal value theorem. Theoretical

population biology, 9(2):129–136, 1976.
Michael Crichton. The Andromeda Strain. Knopf, Dell, NY, 1969.
Lawrence M Dill. The escape response of the zebra danio (Brachydanio rerio)
I. The stimulus for escape. Animal Behaviour, 22(3):711–722, 1974a.
Lawrence M Dill. The escape response of the zebra danio (Brachydanio rerio)
II. The effect of experience. Animal Behaviour, 22(3):723–730, 1974b.
Reuven Dukas and Stephen Ellner. Information processing and prey detection.
Ecology, 74(5):1337–1346, 1993.
Reuven Dukas and Alan C Kamil. The cost of limited attention in blue jays.
Behavioral Ecology, 11(5):502–506, 2000.
Reuven Dukas and Alan C Kamil. Limited attention: the constraint underlying
search image. Behavioral Ecology, 12(2):192–199, 2001.
John T Emlen Jr. Flocking behavior in birds. The Auk, 1952.

Leonhard Euler. Introductio in analysin infinitorum, auctore Leonhardo

Eulero... apud Marcum-Michaelem Bousquet, 1748.
MP Hassell, JH Lawton, and JR Beddington. Sigmoid functional responses

by invertebrate predators and parasitoids. The Journal of Animal Ecology,
46(1):249–262, 1977.
Brian Hopkins. The species-area relations of plant communities. The Journal

of Ecology, pages 409–426, 1955.
Conder P. J. Individual distance. Ibis, 91:649–655, 1949.
Corinne Le Quéré, Robbie M Andrew, Josep G Canadell, Stephen Sitch,

Jan Ivar Korsbakken, Glen P Peters, Andrew C Manning, Thomas A
Boden, Pieter P Tans, Richard A Houghton, et al. Global carbon budget
2016. Earth System Science Data, 8(2):605, 2016.
ME Lutcavage, RW Brill, GB Skomal, BC Chase, JL Goldstein, and J Tutein.

Tracking adult north atlantic bluefin tuna (thunnus thynnus) in the north-
western atlantic using ultrasonic telemetry. Marine Biology, 137(2):347–358,
2000.
Rajat Rohatgi, Peter Nollau, Hsin-Yi Henry Ho, Marc W Kirschner, and
Bruce J Mayer. Nck and phosphatidylinositol 4, 5-bisphosphate synergisti-
cally activate actin polymerization through the N-WASP-Arp2/3 pathway.
Journal of Biological Chemistry, 276(28):26448–26452, 2001.
Miller R.S. and Stephen W. J.D. Spatial relationships in flocks of sandhill

cranes. Ecology, 47(2):323–327, 1966.
D. W. Stephens and J. R. Krebs. Foraging theory. Princeton University Press,

Princeton NJ, 1986.
Lewis G Tilney and Daniel A Portnoy. Actin filaments and the growth,
movement, and spread of the intracellular bacterial parasite, listeria mono-
cytogenes. The Journal of cell biology, 109(4):1597–1608, 1989.
Index
amplitude, 308, 309, 312, 345 arctangent, 318

analytic, 126, 251 arguement
confirmation, 193 geometric, 280
derivative, 75 argument
methods, 277 algebraic, 83
solution, 249, 260, 263 geometric, 30, 73, 157
analytic solution, 260 qualitative, 284, 287
angle astroid, 192
changing, 328 average
converting, 321 rate of change, 56, 57, 78
degrees, 304 rate of energy gain, 156
radians, 114, 304, 305, 309, value, 306, 312
316 velocity, 56, 57, 398
visual, 332
antiderivative, 92, 93, 98 balance equation, 231, 232, 258,
polynomial, 93 291
sketch, 101 base, 206
antidifferentiation, 92, 98, 101 base point, 113
aphids, 39 Beverton-Holt model, 38
approximation, 30, 114, 207, 219,
260 chain rule, 95, 167, 168, 183
2 ≈ 103 , 204
n
circle, 189
decimal, 27 circumference of, 304
error, 272 coefficient
linear, 113, 114, 350 binomial, 90
Newton’s method, 35, 117 Hill, 38
numerical, 80 power function, 26
with tangent line, 67, 124 composite
approximaton functions, 95
Euler’s method, 261 concave
arc length, 305 down, 128
arccosine, 317 up, 127, 128
arcsine, 316 concavity, 114, 128, 175, 190
cone, 186 ellipse

constraint, 27, 149, 151 rotated, 192
converge, 118 endpoints
convergence, 117 maxima at, 153
cosine, 306 error
derivative, 326 in approximation, 116
cosines numerical, 265
law of, 329 Euler, 225
coupled Euler’s method, 260, 261, 289,
ODEs, 292 300
Crichton even
Michael, 203 function, 32, 42, 307
critical point, 35, 132 exponential function, 227
critical points base 10, 208
classifying, 136 base 2, 207
cubic, 33, 78 base e, 208
curvature, 128 exponential growth, 262
cycle, 305 extrema, 134
peridic, 309
cylinder
family
surface area, 149
of solutions, 101, 229
volume, 148
finite difference, 82
equation, 261
decreasing
first derivative
function, 127, 128
test, 133, 157
degree
Folium of Descartes, 201
of polynomial, 91
frequency, 309
derivative, 51, 61, 70
function
definition, 61
composition, 167
second, 127
continuous, 75
differential equation, 95, 211,
cubic, 26
227–229, 249, 405
differentiation discrete, 75
implicit, 189 even, 32
Dill Hill, 38
Larry, 333 odd, 32
dimensionless, 305 one-to-one, 32
direction field, 280 polynomial, 33
discontinuity power, 25, 26, 29
jump, 76 rational, 33, 35, 36
removable, 76
doubling, 203 Galileo, 53
Dukas geometric argument, 30, 403
Reuven, 175 geometric relationships, 183
INDEX 443
harmonic oscillator, 405 of limits, 378

Hill Lineweaver-Burk, 49, 396
function, 49, 87 local
behaviour, 69
identity maximum, 33, 132
trigonometric, 307 minimum, 32, 132
implicit log-log plot, 216
differentiation, 97, 189 logarithm
function, 188 natural, 212
increasing logistic equation, 276, 287, 300
function, 127
independent variable, 32, 34, 146 maximum, 145
inflection absolute, 137
point, 128 global, 137
initial local, 33
condition, 230 Michaelis-Menten
value, 230 kinetics, 37, 104
initial value minimum, 145
problem, 238, 260 absolute, 137
instantaneous global, 137
rate of change, 63 local, 32
intercept, 54 model
inverse function, 191, 212, 315 mathematical, 27
iterated
method, 117 natural base, 208
iteration, 117, 261 Newton
Isaac, 254
Kepler, 150, 162 Newton’s method, 27, 35,
kymograph, 51 109–112, 116
nonlinear
ladybugs, 39 differential equation, 275, 276
law of cosines, 329, 392 numerical
level curve, 358 solution, 260
license, 4 numerical solution, 249
limit
DNE, 76 odd
exists, 76 function, 32, 42, 307
right and left, 76 one-to-one, 32, 315
linear optimization, 145
approximation, 109
operation, 91 parameter, 28, 232
relationship, 54 perimeter
linearity, 91 of circle, 305
of derivative, 91 period, 307–309
periodic function, 307, 309 second derivative, 92, 97, 133

phase, 308 test, 134, 146, 147, 174
phase line, 284 second order
phase shift, 310 DE, 405
Pi (π), 304 sign change, 130
point sine, 306
of intersection, 26, 27 derivative, 326
of tangency, 109 sketching, 127
polynomial, 33, 91 antiderivative, 101
derivative of, 91 cubic, 33
polynomials, 25 graph, 26, 31
population polynomial, 33, 35
fish, 38, 210 rational, 35
population growth, 105 the derivative, 72, 100
power slope
dominant, 26 of a straight line, 33
function, 25, 26, 29, 31, 35 of straight line, 54
power rule, 89, 95, 191 of tangent line, 70
proportional, 28 slope field, 280
proportionality constant, 28 solution
Pythagoras theorem, 316, 317, analytic, 260
401 numerical, 260
Pythagorian to differential equation, 229
triangle, 304 solution curve, 242
spreadsheet, 118, 263
radian, 114, 304, 305 stability, 287
rate stable
constant, 234 steady state, 287
rate of change state space, 280, 284
average, 51, 54, 56, 57, 66, 69 steady state, 252, 256, 257, 284,
averate, 67 286
instantaneous, 51, 58, 69 step function, 84
rational function, 25, 33, 35–37 step size, 260
recurrence relation, 261 straight line, 54
related rates, 183, 328, 332 system
rescaling, 279 of equations, 292
restricting the domain, 315
Ricker equation, 210 tangency
root, 41, 44 point of, 109
of equation, 116 tangent
rooted tangent, 403 rooted, 403
tangent line, 62, 67, 70, 109–111
secant transformation, 279, 309
line, 51, 55, 67 trigonometric
INDEX 445
functions, 303, 304 velocity, 68

identities, 304 vertical line property, 188
trigonometric identities, 392
trigonometry, 67
zero, 33, 44, 72, 109, 116, 131
unbounded, 32 zeros
unit circle, 305 of a function, 136
unstable, 287 zoom, 70
Index of Applications
acceleration, 68, 85, 97, 160 nutrition, 182

uniform, 98, 99 population, 354
ActA, 397 birth rate
actin, 95, 397 human, 234
age distribution, 233 blood alcohol level, 268, 354
uniform, 233 blood vessel branching, 358
albedo, 30, 31, 47, 171, 181 box
alcohol, 268, 354 rectangular, 400
allometric constants, 216 Boyle’s Law, 197
allometry
metabolic rate, 47 carbon 14, 247
pulse rate, 47 carnivores, 170
Andromeda strain, 203, 215 carrying capacity, 146, 276
ant trails, 173, 182 cell
antibiotics, 246 cylindrical, 148
atmospheric pressure, 248 epidermal, 200
attention, 175, 182 epithelia, 87
attraction, 396 growth, 196
Avogadro’s length, 150
number, 238 nutrition, 160
shape, 27, 28, 46, 352
bacteria spherical, 147
Andromeda strain, 203, 215, cesium-137, 239
244 chemical
colony, 246 decay, 258
colony shape, 105 production, 258
E. coli, 204, 352 Chernoby, 239
population, 246, 348 chess
bird invention, 219
duck, 356 climate change, 47
flock, 356, 396 clock hands, 330, 331
formation, 341 coffee budget, 171
immigration, 354 comet tail, 397
conservation, 165, 301 enzyme, 36, 48

convergent extension, 184 equinox, 313
cooling, 52, 286 escape
cooperativity, 38 response, 333
Euler’s method, 272
data exponential peeling, 223
refined, 58, 65 extinction, 245
temperature, 269
daylight cycle, 313 falling object, 53, 65, 97, 107,
decay 258, 350
equation, 238, 242 fertility, 233, 245
exponential, 239 fireworks, 344
density dependent fish, 170
growth, 145 population, 299, 351
growth rate, 276 population density, 165
disease, 291 populations, 224, 246
displacement, 57, 322 salmon, 224, 351
distance, 160 school, 396
diversity, 198 school of, 50, 224
DNA stocks, 270
amplification, 219 zebra danio, 333, 342
doubling fluorescence, 95
time, 235, 236 food
drug cryptic, 175
doses, 163 patch, 154
sensitivity, 163 type, 175
dynein, 222 foraging, 105, 166
optimal, 154
E. coli, 204 time, 154
Earth friction, 268
emissivity of, 30 fungal colony
temperature of, 30, 47, 90, 96, growth, 196
171, 181, 190
ECG, 303 gene expression, 223
ecosystem, 105 genomics, 223
electrocardiogram, 303 gravity, 97
emissivity, 181 greenhouse gas, 30, 47, 172, 181
endemic growth
disease, 294 cell, 196
energy crystal, 298
Arrhenius activation, 210 density-dependent, 276
balance, 30, 47, 90, 163 fungal colony, 196
gain, 105, 154, 166, 182 limb, 196
loss, 163 tumour, 357
I N D E X O F A P P L I C AT I O N S 449
vine, 353 Lennard-Jones potential, 351

growth rate lens
intrinsic, 146 convex, 197
limb development, 196
half life, 239 limnology, 163
half-life, 246–248, 271 Lineweaver-Burk, 49, 396
heartbeat, 303 logistic growth, 104, 145, 165
heat harvesting, 299
loss, 355 Lysteria
heating, 52, 65 monocytogenes, 397
Hill
coefficient, 38 Michaelis-Menten
function, 38, 49, 87 kinetics, 49, 80, 104, 142
HIV, 290, 350 microtubule, 73, 86, 222
hormone cycle, 313 model
mathematical, 27
ideal gas, 197 molecular collision, 210
infant weight gain, 270 molecular motor, 222
infection moon phase, 314
spread, 300 mortality, 234, 245
influenza, 290 motion
initial of a ball, 141
population, 246 of a cannon-ball, 341
velocity, 98 of a cannonball, 160
insect, 181 of a cell, 200
ant, 173, 181 of a particle, 350
bee, 105, 163 uniform, 53
intrinsic growth rate, 276 motor
invasive species, 106, 221 molecular, 73, 86
muskrats, 197 moving bead, 397
iodine-131, 239 murder mystery, 256
isotherms, 201 muscle
shortening, 142
Kepler
wedding, 150, 162 net growth rate, 232
kinesin, 73 Newton’s
kinetics, 247 law of cooling, 254, 263, 269,
353
Lactobacillus, 52 nM
Law of Mass Action, 277, 299 nano Molar, 37
leaf nuclear power plant, 239
shape, 351 nutrient
learning, 354 absorption, 29, 147
least cost, 355 balance, 27
consumption, 29, 147 radioactivity, 248

increase, 160 rainfall, 170
nutrition reaction, 47
gain, 182 bimolecular, 301
biochemical, 36, 299
optics, 197 distance, 336
optimal enzyme-catalyzed, 48
foraging, 154 rate, 247
strategy, 182 reversible, 301
speed, 36, 49, 80, 142
particle motion, 106 reproduction, 164
per capita number, 295
birth rate, 232 repulsion, 396
mortality rate, 232 residence time, 154
per capita birth rate, 245 Ricker
percent equation, 224
growth, 234
perimeter SARS, 290
maximal, 153 saturation, 37
pheromone, 173 seed
pollution, 170, 271 distribution, 225
polymerase Shannon Entropy, 223
chain reaction, 219 shortest path, 173, 181
polymerization, 48 skydiver
population equation, 258
density, 146, 165 sleep-wake cycles, 322
growth, 242 social media, 300
uncontrolled growth, 231 spacing distance, 396
population density, 245 steam power, 355
population doubling, 245 stroboscope, 59
population growth, 104, 145, 198, strontium-90, 248
244–246 substrate, 36
human, 66, 222, 246, 356 sunspot, 322
muskrats, 197 surface area
rodent, 245 minimal, 148
predator, 342 sustainability, 30, 47, 66, 90, 96,
Predator Response, 39 145, 165, 171, 181, 215, 221,
predators, 164 222, 224, 231, 235, 246, 270,
prey, 342 271, 276, 291, 299, 300, 355
probability
of decay, 237 temperature
ambient, 254, 269
race track, 329, 351, 356, 357 bee swarm, 163
radioactive decay, 237 cofee, 269
I N D E X O F A P P L I C AT I O N S 451
daily variation, 322 initial, 98

milk, 52, 65 instantaneous, 60
thermodynamics, 201 terminal, 258, 268
time of death, 256 vesicle, 86
tissue, 87, 184, 271 visual angle, 332, 338
tumor growth, 183
tuna, 53, 66 water
density, 163
vaccination, 300 ice melting, 197
Van der Waal’s equation, 201 wine barrel, 150, 162
velocity, 61, 85, 97, 160
average, 56, 57, 66, 68, 398 yoghurt, 52

Ubc Diffcal PDF

Uploaded by

Copyright:

Available Formats

Ubc Diffcal PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ubc Diffcal PDF

Uploaded by

Copyright:

Available Formats

“ T H E R E I S N O K N O W L E D G E T H AT I S N O T P O W E R .

Printing as of January 2020

1 Power functions as building blocks 25

2 Average rates of change, average velocity and the secant line 51

3 Three faces of the derivative: geometric, analytic, and computational 69

3.2 The analytic view: calculating the derivative 75

4 Differentiation rules, simple antiderivatives and applications 89

5 Tangent lines, linear approximation, and Newton’s method 109

6 Sketching the graph of a function using calculus tools 127

8 Introducing the chain rule 167

9 Chain rule applied to related rates and implicit differentiation 183

10 Exponential functions 203

11 Differential equations for exponential growth and decay 227

12 Solving differential equations 249

13 Qualitative methods for differential equations 275

14 Periodic and trigonometric functions 303

15 Cycles, periods, and rates of change 325

16 Additional exercises 347

A A review of Straight Lines 363

B A precalculus review 365

C A Review of Simple Functions 367

D.3 Limits of rational functions 379

F Trigonometry review 391

G For further study 395

H Short Answers to Problems 407

1.1 Graphs of power functions 26

2.1 Kymograph used to track motion of a labeled protein 51

3.1 Zooming in on the graph of a function 70

3.9 Function derivative sketching 73

4.1 Position, velocity, and acceleration for falling object 99

5.1 A parabola and two tangent lines 110

6.1 Increasing and decreasing derivatives and concavity 128

6.6 B(x) = C (x2 − x3 ): which power dominates? 135

7.1 Growth rate for Logistic growth 146

8.1 Function composition 167

9.1 Growth of a spherical tumor 183

10.1 Powers of 2 204

10.4 Tangent line to ex 209

11.1 Exponential growth and decay 229

12.1 Steady state 252

13.1 Direction field and solution curves for dydt = 2y 282

14.1 Electrical activity of the heart: a repeating pattern 303

14.9 Periodic moon phases 314

15.1 Law of cosines 329

16.1 A function and its tangent line 349

A.1 Slope of a straight line 363

C.1 An ellipse cannot be described by a single function 367

C.7 Graphs of inverse functions 374

D.1 A function with a “hole” 381

F.1 Ratios of side lengths in right triangle 391

G.1 Lysteria, a parasite that lives inside a cell 397

2.1 Heating milk temperature data. 52

3.1 Some computed derivatives. 83

4.1 The Power Rule 90

6.1 Function behaviour of a polynomial 137

8.1 Data for Newton’s method 178

9.1 Geometric relationships for related rates 183

10.1 Constant in derivative of 2x 207

12.1 Checking if a function satisfies a differential equation 250