1.

Derivatives, Rates of Change and Derivative Functions

In Chapter 1, we looked at functions, how they represented both algebraically and graphically. In Chapter 2, we looked at what happens
when a sequence of numbers (a recursion relation) and function approach a particular value (asymptotes) and towards infinity.

In this chapter, we will study a special type of limit, called a derivative. Derivatives arise when we want to find a rate of growth, a
velocity, the slope of a tangent line, or any instantaneous rate of change.

We will soon delve into the applications of derivatives in Mechanics (most common but certainly not the only one) and it is important
to see how the concept is used in often overlooked areas like in Healthcare. So far, I hope that you can see that Mathematics is really all
around us and this is largely thanks to mathematical modelling.

While mathematical modelling is about trying to fit data to functions and equations, the purpose is not just to predict but to be able to
interpret and make sense of the data through the functions. And one of the best (robust – nobody will argue with you) ways is through
Calculus which is the core of this module. The reason is because Calculus is fundamentally about modelling change.

In the following example, we revisit a concept that was discussed in Chapter 2 Section 4 – Limits of Functions at Finite Numbers. While
the focus in Chapter 2 was calculating the limit, this Chapter will focus on its broader meaning and implication. 1
 1. Derivatives, Rates of Change and Derivative Functions
Example 1: Biomedical scientists have studied the chemical and physiological changes in the body that result from alcohol consumption.
The reaction in the human body occurs in two stages: a fairly rapid process of absorption and a more gradual one of metabolism. To
predict the effect of alcohol consumption, one needs to know the rate at which alcohol is absorbed and metabolized.

Medical researchers measured the blood alcohol concentration (BAC) of eight fasting adult male subjects after rapid consumption of 15
mL of ethanol (corresponding to one alcoholic drink). The data they obtained were modeled by the concentration function
𝐶 𝑡 = 0.0225(𝑡)𝑒 −0.0467𝑡
where 𝑡 is measured in minutes after consumption and 𝐶 is measured in mg/mL. How quickly is the BAC increasing after 10 minutes?

Solution: We are asked to find the instantaneous rate of change of C with respect to 𝑡 when 𝑡 = 10. The difficulty is that we are dealing
with a single instant of time (𝑡 = 10 min) and so no time interval is involved.

However as we did in Chapter 2 Section 4, we can approximate the desired quantity by calculating the average rate of change with
respect to the time interval from 𝑡 = 10 to 𝑡 = 11 (for instance).

change in 𝐶 Δ𝐶 𝐶 11 − 𝐶(10) 0.148073 − 0.141048

average rate of change = = = ≈ = 0.007025 (mg/mL)/min
change in 𝑡 Δ𝑡 11 − 10 1 2
 1. Derivatives, Rates of Change and Derivative Functions
The following table shows the results of similar calculations of the average rates of change [in (mg/mL)/min] over successively smaller
time periods.

It appears that as we shorten the time period, the average rate of change is becoming closer and closer to a number between 0.00752
and 0.00753 (mg/mL)/min. The instantaneous rate of change at 𝑡 = 10 is defined to be the limiting value of these average rates of
change over shorter and shorter time periods that start or end at 𝑡 = 10. So we estimate that the BAC increased at a rate of about
0.0075 (mg/mL)/min. We will soon show how this was calculated precisely! So stay tuned!

We can, of course, use the limits definition to find the instantaneous rate of change but it is complicated (uses a lot of other ideas in
Maths) and it is out of the scope of this module. Instead, we can use well established formulae. So, this Chapter is dedicated to the
derivative – how it is used in various scenarios.
3
 1. Derivatives, Rates of Change and Derivative Functions
When we calculated the average rate of change, we set up a fraction comprising of a change in the dependent variable divided by a
change in the independent variable. Effectively, this is the formula to find the slope of a secant line (recall what was covered in Chapter
1 Section 1). In the figure on the bottom left, we showed how to find the slope of the secant line that joins two points 𝑃(𝑎, 𝑓(𝑎)) and
𝑄(𝑥, 𝑓(𝑥)) on the
curve. This secant
line would be the
approximation to
find the average rate
of change at 𝑥 = 𝑎.

The word tangent is derived from the Latin word tangens, which means “touching.” Thus, a tangent to a curve is a line that touches the
curve. So what is the relationship between the secant and tangent lines? The tangent line is the limit of the secant line when the second
point 𝑄 is brought closer to the reference point 𝑃 until it touches the point 𝑃! Mathematically speaking,

𝑓 𝑥 − 𝑓(𝑎) 𝑓 𝑎 + ℎ − 𝑓(𝑎)
slope of tangent line 𝑚 = lim = lim (if limit exists)
𝑥→𝑎 𝑥−𝑎 ℎ→0 ℎ 4
 1. Derivatives, Rates of Change and Derivative Functions
By definition, the derivative is the slope of the tangent line. Hence, the derivative is given by 𝑓′(𝑎) and is mathematically defined as

𝑓 𝑥 − 𝑓(𝑎) 𝑓 𝑎 + ℎ − 𝑓(𝑎)
𝑓 ′ 𝑎 = lim = lim (if limit exists)
𝑥→𝑎 𝑥−𝑎 ℎ→0 ℎ
We can also say that 𝑓′(𝑎) is the instantaneous rate of change of 𝑓 at 𝑎.

Example 2: Find an equation of the tangent line to the parabola 𝑦 = 𝑥 2 − 8𝑥 + 9 at the point (3, −6).

Solution: We are asked to find the equation of the tangent line and so must recall the formula for the equation of a line: 𝑦 = 𝑚𝑥 + 𝑐,
where 𝑚 is the slope of the line and c is the 𝑦-intercept (see Chapter 1 Section 2). Here, we assign a function 𝑓 𝑥 = 𝑥 2 − 8𝑥 + 9 and
we aim to find 𝑓′(𝑥) at 𝑥 = 3 or simply 𝑓’(3). While there will be formula we can use, let us gain (just a little) more experience in
manipulating limits.

′
𝑓 𝑥 − 𝑓(3) 𝑥 2 − 8𝑥 + 9 − (32 − 8 3 + 9) 𝑥 2 − 8𝑥 − 9 + 24
𝑓 3 = lim = lim = lim
𝑥→3 𝑥−3 𝑥→3 𝑥−3 𝑥→3 𝑥−3

𝑥 2 − 8𝑥 + 15 (𝑥 − 5)(𝑥 − 3)
= lim = lim = lim (𝑥 − 5) = −2
𝑥→3 𝑥−3 𝑥→3 𝑥−3 𝑥→3

Hence, the slope of the tangent line is −2.

5
 1. Derivatives, Rates of Change and Derivative Functions
Putting everything together now,
𝑦 = −2 𝑥 + 𝑐 → −6 = −2 3 + 𝑐 → 𝑐=0 → 𝑦 = −2𝑥 is the equation of tangent line.

Let us try the other approach. Note that either approach is acceptable if you are asked to evaluate the slope of the tangent line in this
way.
2
′
𝑓 3 + ℎ − 𝑓(3) 3+ℎ − 8 3 + ℎ + 9 − (32 − 8 3 + 9)
𝑓 3 = lim = lim
ℎ→0 ℎ ℎ→0 ℎ
2
3+ℎ − 8 3 + ℎ + 9 − (32 − 8 3 + 9) 9 + 6ℎ + ℎ2 − 24 − 8ℎ + 9 − 9 + 24 − 9
= lim = lim
ℎ→0 ℎ ℎ→0 ℎ

ℎ2 + 6ℎ − 8ℎ
= lim = lim (ℎ − 2) = −2
ℎ→0 ℎ ℎ→0

And once again, the slope of the tangent line is −2 and as we did earlier, 𝑦 = −2𝑥 is the equation of tangent line.

So revisiting Example 1, the slope of the tangent line at 𝑡 = 10 is the instantaneous rate of change of the BAC (blood alcohol
concentration) at time 𝑡 = 10.

6
 1. Derivatives, Rates of Change and Derivative Functions
A familiar example of a rate of change is velocity in Mechanics. In Example 17 of Chapter 2, we found the instantaneous velocity of a
ball dropped from the CN Tower as the limit of average velocities over shorter and shorter time periods.

More generally, if 𝑠 = 𝑓(𝑡) is the displacement (distance from the starting point) function of a particle that moves along a straight line,
then 𝑓’(𝑎) is the rate of change of of the displacement 𝑠 with respect to the time 𝑡.

In other words, 𝑓’(𝑎) is the velocity of the particle at time 𝑡 = 𝑎. The speed of the particle is the absolute value of the velocity, that is,
|𝑓’(𝑎)|.

The reason for the difference (as you will see in your Physics module) is that speed is independent of direction whereas velocity has a
direction and so can take on different signs to denote the direction. So let us revisit Example 17 of Chapter 2.

Example 3: Suppose that a ball is dropped from the upper observation deck of the CN Tower, 450 m above the ground. What is the
velocity of the ball after 5 seconds? How fast is the ball traveling when it hits the ground?

7
 1. Derivatives, Rates of Change and Derivative Functions
Solution: Let the acceleration due to gravity be 𝑎 = 9.8 m/s2
We will need to find the velocity both when t=5 and when the ball hits the ground, so it is efficient to start by finding the velocity at a general
time 𝑡. Assuming that the ball falls from rest at the top of the CN tower, the equation of motion (initial velocity 𝑢 = 0) is

𝑎 2 9.8 2
𝑠 = 𝑓 𝑡 = 𝑢𝑡 + 𝑡 = 0(𝑡) + 𝑡 = 4.9 𝑡 2
2 2

The instantaneous velocity at time t is given by

′
𝑓 𝑡 + ℎ − 𝑓(𝑡) 4.9 𝑡 + ℎ 2 − 4.9 𝑡 2
4.9(𝑡 + ℎ − 𝑡)(𝑡 + ℎ + 𝑡)
𝑣 𝑡 = 𝑓 𝑡 = lim = lim = lim
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ

4.9 ℎ 2𝑡 + ℎ
= lim = 4.9 2𝑡 + 4.9 0 = 9.8𝑡
ℎ→0 ℎ

So the velocity of the ball after 5 seconds is 9.8(5) = 49 m/s (as we saw in Example 17 in Chapter 2). To find the velocity at which the ball
hits the ground, we need to first find the time in which the ball travels the height of the tower of 450 m (to fall onto the ground). In other
words, we have to first solve for time 𝑡 in the equation 𝑠 = 4.9 𝑡 2 = 450 → 𝑡 2 = 450/4.9 → 𝑡 ≈ 9.6 s.
Then, we substitute the time it took the ball to hit the ground into the velocity function, we can find 𝑣(9.6) = 9.8(9.6) = 94 m/s.
8
 1. Derivatives, Rates of Change and Derivative Functions
Knowing how to calculate the derivative is one matter, we need to take a brief step back and ask ourselves whether do we really have a
derivative. This question leads us to discuss the topic of differentiability. Recall what we said earlier this section:

𝑓 𝑥 − 𝑓(𝑎) 𝑓 𝑎 + ℎ − 𝑓(𝑎)
𝑓 ′ 𝑎 = lim = lim (if limit exists)
𝑥→𝑎 𝑥−𝑎 ℎ→0 ℎ
The key is the word if the limit exists. A function is differentiable if the derivative exists and the derivative exists if either one of the two
limits (shown above) exists.

Example 4: Where is the function 𝑓(𝑥) = |𝑥| differentiable?

Solution: First, it is important to know how does a function generally look like? Recall from Chapter 1 that the absolute value function is a
piecewise function.

𝑥 𝑥≥0
𝑓 𝑥 = 𝑥 =ቊ
−𝑥 𝑥<0

9
 1. Derivatives, Rates of Change and Derivative Functions
Because ±𝑥 are polynomial functions and they are defined for all real numbers, the functions are thus continuous (recall that continuous

means that 𝑓 𝑎 = lim+ 𝑓(𝑥) = lim− 𝑓(𝑥)). So for completeness, we consider two regions: 𝑥 < 0 and 𝑥 > 0.
𝑥→𝑎 𝑥→𝑎

𝑓 𝑥+ℎ −𝑓 𝑥 − 𝑥 + ℎ − −𝑥 −ℎ
𝑥 < 0, 𝑓 𝑥 = −𝑥 ∶ 𝑓 ′ 𝑥 = lim = lim = lim = −1
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ

𝑓 𝑥+ℎ −𝑓 𝑥 𝑥+ℎ − 𝑥 ℎ
𝑥 > 0, 𝑓 𝑥 = 𝑥 ∶ 𝑓 ′ 𝑥 = lim = lim = lim = 1
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ

The above two lines of working imply that the function is differentiable in the two regions 𝑥 < 0 and 𝑥 > 0. All that remains is to consider
𝑓 𝑥 −𝑓 𝑎
the transition point at 𝑥 = 0. So to establish differentiability at 𝑥 = 𝑎 = 0, we see that we need to show that lim exists. Recall
𝑥→𝑎 𝑥−𝑎

from Chapter 2 (Section 4) that lim 𝑓(𝑥) = 𝐿 to exist, we must have lim 𝑓(𝑥) = 𝐿 = lim+ 𝑓(𝑥) = lim− 𝑓(𝑥). In other words, we need to
𝑥→𝑎 𝑥→𝑎 𝑥→𝑎 𝑥→𝑎
show for 𝑥 = 0,

𝑓 𝑥 − 𝑓(0) 𝑓 𝑥 − 𝑓(0)
lim+ = lim− ?
𝑥→0 𝑥−0 𝑥→0 𝑥−0
Let us consider the two sides separately.

10
 1. Derivatives, Rates of Change and Derivative Functions
𝑓 𝑥 − 𝑓(0) 𝑥−0 𝑓 𝑥 − 𝑓(0) −𝑥 − 0
lim+ = lim+ = 1, lim− = lim− = −1
𝑥→0 𝑥−0 𝑥→0 𝑥 𝑥→0 𝑥−0 𝑥→0 𝑥
We use 𝑓(𝑥) = 𝑥 here because we are looking at 𝑥 approaching 0 from the right (𝑥 > 0) and we use 𝑓(𝑥) = −𝑥 here because we are
looking at 𝑥 approaching 0 from the left (𝑥 < 0).

An alternative formulation (not necessary but just in case you have doubts) is to show

𝑓(0 + ℎ) − 𝑓(0) 𝑓 0 + ℎ − 𝑓(0)

lim+ = lim− ?
ℎ→0 ℎ ℎ→0 ℎ

𝑓(0 + ℎ) − 𝑓(0) ℎ−0 𝑓 0 + ℎ − 𝑓(0) − ℎ −0

lim+ = lim+ = 1, lim− = lim− = −1
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ
Because we are looking at ℎ approaching 0 from the right (ℎ > 0), 𝑥 = (0 + ℎ) is clearly positive and must use 𝑓 𝑥 = 𝑥 → 𝑓 0 + ℎ = ℎ.
Because we are looking at ℎ approaching 0 from the left (ℎ < 0), 𝑥 = (0 + ℎ) is clearly negative and we use 𝑓 𝑥 = −𝑥 → 𝑓 0 + ℎ = −ℎ.

Clearly, the left- and right-hand limits are not equal. While the function 𝑓(𝑥) = |𝑥| is continuous everywhere (that is, for all real numbers), it
is differentiable everywhere except at 𝑥 = 0.
11
 1. Derivatives, Rates of Change and Derivative Functions
From the figure below, we can see the derivative of the function 𝑓(𝑥) = |𝑥| is discontinuous at 𝑥 = 0.

So from the last example, we see that when there are scenarios in which the function is not differentiable. The absolute value function has a
noticeable feature: a sharp kink/corner at 𝑥 = 0. The reason why the function is not differentiable at 𝑥 = 0 is because the derivative (from
the limit definition) does not exist, as the left limit (= −1) is not equal to the right limit (= 1) for [𝑓(𝑥) − 𝑓(0)]/(𝑥 − 0).

In the figure on the right, we see examples

where the function is not differentiable at
𝑥 = 𝑎.

12
 1. Derivatives, Rates of Change and Derivative Functions
If a function has a single derivative, the derivative is given by 𝑓’(𝑥). Higher derivatives are then the second derivative 𝑓’’(𝑥), the third
derivative 𝑓 (3) (𝑥) or 𝑓’’’(𝑥), the fourth derivative 𝑓 4 (𝑥) and so on.

As we will see when we do differential equations, we can express the derivatives in another way – Leibnitz (pronounced lai-b-nits) notation.
So for 𝑦 = 𝑓(𝑥), the definition of the derivative can be written as

𝑓 𝑥 − 𝑓(𝑎) Δ𝑦 𝑑𝑦 𝑑
𝑓 ′ 𝑎 = lim = lim = = (𝑦)
𝑥→𝑎 𝑥−𝑎 Δ𝑥→0 Δ𝑥 𝑑𝑥 𝑑𝑥

𝑑𝑦/𝑑𝑥 where the “d” here refers to the differential (meaning a small difference). It makes sense since we are considering the slope of the
secant line as 𝑥 → 𝑎. When this happens, the changes in the vertical and horizontal direction approaches 0 individually, but never 0!
13
 1. Derivatives, Rates of Change and Derivative Functions
So if the first derivative for 𝑦 = 𝑓(𝑥) is written as

𝑑𝑦 𝑑
𝑓′ 𝑥 = = (𝑦)
𝑑𝑥 𝑑𝑥
Then, the higher order derivatives are:

′′
𝑑 𝑑𝑦 𝑑2 𝑦 (3) ′′′
𝑑 𝑑2 𝑦 𝑑3 𝑦 (4)
𝑑 𝑑3 𝑦 𝑑4 𝑦
𝑓 𝑥 = = 2, 𝑓 𝑥 =𝑓 𝑥 = = 3, 𝑓 𝑥 = = 4, ⋯
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 2 𝑑𝑥 𝑑𝑥 𝑑𝑥 3 𝑑𝑥

In general, we can interpret a second derivative as a rate of change of a rate of change. The most familiar example of this is in Mechanics:
acceleration, which we define as follows. If 𝑠 = 𝑠(𝑡) is the displacement function of an object that moves in a straight line, we know that its
first derivative represents the velocity 𝑣(𝑡) of the object as a function of time:

𝑑𝑠
𝑣 𝑡 =
𝑑𝑡
The instantaneous rate of change of velocity with respect to time is called the acceleration 𝑎(𝑡) of the object. Thus, the acceleration function
is the derivative of the velocity function and is therefore the second derivative of the position function: 𝑎 𝑡 = 𝑣 ′ 𝑡 = 𝑠′′(𝑡)
or in Leibnitz notation:

𝑑𝑣 𝑑 𝑑𝑠 𝑑2 𝑠
𝑎 𝑡 = = = 2
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡
14
 1. Derivatives, Rates of Change and Derivative Functions
Because 𝑓’(𝑥) represents the slope of the curve 𝑦 = 𝑓(𝑥) at the point (𝑥, 𝑓(𝑥)), it tells us the direction in which the curve proceeds at each
point.
So, it is reasonable to expect that information about 𝑓’(𝑥) will provide us with information about 𝑓(𝑥).

In particular, to see how the derivative of 𝑓 can tell us where a function is increasing or decreasing, look at the figure below.

Between A and B and between C and D, the tangent lines have positive slope and so 𝑓’(𝑥) > 0.

Between B and C, the tangent lines have negative slope and so 𝑓’(𝑥) < 0.

Thus, it appears that 𝑓 increases when 𝑓’(𝑥) is positive and decreases when 𝑓’(𝑥) is negative.

At the points B and C, we note that there is a smooth rounded bend. At those points, we
say that f’(x)=0 and so at this point, f does not change (i.e. neither increases nor decreases).

15
 1. Derivatives, Rates of Change and Derivative Functions
Sometimes when we are presented with a graph that represents the values of the derivative, it helps to be able to look at the graph of the
derivative and then make an educated guess about the original function. This practice occurs a lot in Mechanics when we are given the graph
of the velocity and have to make a sketch of the displacement function. We present a related (mathematical) example of this below.

Example 5: If it is known that the graph of the derivative 𝑓’ of a function is as shown in the figure below, what can we say about 𝑓?
If it is known that 𝑓(0) = 0, sketch a possible graph of 𝑓.

Solution: We always begin by noting that the graph is for the derivative 𝑓’.

The points where the graph of 𝑓’ is above the horizontal axis, the derivative is positive and
we expect the graph of 𝑓 to have a positive slope.

The points where the graph of 𝑓’ is on the horizontal axis, the derivative is zero and we expect the graph of 𝑓 to have a zero (flat) slope.

The points where the graph of 𝑓’ is below the horizontal axis, the derivative is negative and we expect the graph of 𝑓 to have a negative
slope. 16
 1. Derivatives, Rates of Change and Derivative Functions
We observe from the graph of the derivative function that 𝑓’(𝑥) < 0 when −1 < 𝑥 < 1, so the original function 𝑓 must be decreasing on
the interval (−1,1).

Similarly, 𝑓’(𝑥) > 0 in 𝑥 < −1 and 𝑥 > 1, so f is increasing on the intervals

(−∞, −1) and (1, ∞).

Also note that, since 𝑓’(−1) = 𝑓’(1) = 0, the graph of 𝑓 has horizontal tangents when 𝑥 = ±1.

Putting the above information together and the given information that 𝑓(0) = 0, we can
sketch of the original function as shown on the right.
All that remains is to note the asymptotes of the graph of the derivative. We note that
when 𝑥 → ±∞, the derivative goes to value of 1 (see the red-circled portions of the graph
above).
So this means that the original function must have a positive slope of 1 as 𝑥 → ±∞ (see
the green-circled portions of the graph on the left).

17
 1. Derivatives, Rates of Change and Derivative Functions
We wrap up this section with an admittedly more complicated proof example. The goal of this question is to help deepen your understanding
(within limits – no excessive, no pun intended) of the relationship between even and odd functions and how they relate to derivatives.

Example 6: Recall that a function f is called even if 𝑓(𝑥) = 𝑓(−𝑥) for all 𝑥 in its domain and odd if 𝑓(𝑥) = −𝑓(−𝑥) for all such 𝑥. Prove each
of the following.
(a) The derivative of an even function is an odd function.
(b) The derivative of an odd function is an even function.

Solution: The strategy is to always recall the definitions of the even and odd functions and how the derivative is obtained from the limit
definition which we flash below.

𝑓 𝑥 + ℎ − 𝑓(𝑥) 𝑓 𝑥 + ℎ − 𝑓(𝑥)
𝑓 ′ 𝑥 = lim = lim
ℎ→0 𝑥+ℎ −𝑥 ℎ→0 ℎ
Then, with this type of problems, we consider starting from one side of the equation and show that we will get the other side of the
equation.

So what is the equation? The equation will be the statement we seek to prove.
18
 1. Derivatives, Rates of Change and Derivative Functions
We want to first prove (a). The statement is to show that for an even function 𝑓, 𝑓 ′ 𝑥 = −𝑓′(−𝑥). The latter means that the derivative is an
odd function. Knowing what we are supposed to prove is important because it gives us a goal to work towards.

If the function 𝑓(𝑥) is even, then we have the following

𝑓 −𝑥 + ℎ − 𝑓(−𝑥) 𝑓 −(−𝑥 + ℎ) − 𝑓(𝑥) 𝑓 𝑥 − ℎ − 𝑓(𝑥)

𝑓 ′ −𝑥 = lim = lim = lim
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 𝒉
where the terms in green and purple above utilize the definition of the even function: 𝑓 𝑏 = 𝑓(−𝑏).

But when we compare the term in red to the following expression for the derivative, we try our best to match the numerators:

𝑓 𝑥 + ℎ − 𝑓(𝑥) 𝑓 𝑥 − ℎ − 𝑓(𝑥) 𝑓 𝑥 − ℎ − 𝑓(𝑥)

𝑓 ′ 𝑥 = lim → 𝑓 ′ 𝑥 = lim = lim
ℎ→0 𝑥+ℎ −𝑥 ℎ→0 𝑥−ℎ −𝑥 ℎ→0 −𝒉

I hope you can see the difference between the denominators of the purple and red terms (which is highlighted in bold). In other words,

𝑓 𝑥 − ℎ − 𝑓(𝑥) 𝑓 𝑥 − ℎ − 𝑓(𝑥) 𝑓 𝑥−ℎ −𝑓 𝑥

𝑓 ′ −𝑥 = lim = lim = − lim = −𝒇′(𝒙)
ℎ→0 𝒉 ℎ→0 −(−𝒉) ℎ→0 −𝒉
And we have 𝑓 ′ −𝑥 = −𝑓 ′ 𝑥 → 𝑓 ′ 𝑥 = −𝑓′(−𝑥) which implies that the derivative is an odd function. Q.E.D.
19
 1. Derivatives, Rates of Change and Derivative Functions
Now we want to first prove (b). The statement is to show that for an odd function 𝑓, 𝑓 ′ 𝑥 = 𝑓′(−𝑥). The latter means that the derivative is
an even function. Knowing what we are supposed to prove is important because it gives us a goal to work towards.

If the function 𝑓(𝑥) is odd, then we have the following

𝑓 −𝑥 + ℎ − 𝑓(−𝑥) −𝑓 − −𝑥 + ℎ − [−𝑓 𝑥 ] 𝑓 𝑥 − 𝑓(𝑥 − ℎ)

𝑓 ′ −𝑥 = lim = lim = lim
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ
where the terms in green and purple above utilize the definition of the odd function: 𝑓 𝑏 = −𝑓(−𝑏).

But when we compare the term in red to the following expression for the derivative, we try our best to match the numerators:

𝑓 𝑥 + ℎ − 𝑓(𝑥) 𝑓 𝑥 − 𝑓(𝑥 − ℎ) 𝑓 𝑥 − 𝑓(𝑥 − ℎ)

𝑓 ′ 𝑥 = lim → 𝑓 ′ 𝑥 = lim = lim
ℎ→0 𝑥+ℎ −𝑥 ℎ→0 𝑥 − (𝑥 − ℎ) ℎ→0 ℎ

I hope you can see there is no difference between the denominators of the purple and red terms. In other words,

𝑓 𝑥 − 𝑓(𝑥 − ℎ) 𝑓 𝑥 − 𝑓(𝑥 − ℎ)
𝑓 ′ −𝑥 = lim = lim = 𝒇′(𝒙)
ℎ→0 ℎ ℎ→0 ℎ
And we have 𝑓 ′ −𝑥 = 𝑓 ′ 𝑥 which implies that the derivative is an even function. Q.E.D.
20
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
If it were always necessary to compute derivatives directly from the definition, as we did in the preceding section, such computations would
be tedious and the evaluation of some limits would require ingenuity.

Fortunately, several rules have been developed for finding derivatives without having to use the definition directly. These formulae greatly
simplify the task of differentiation. In this section we learn how to differentiate constant functions, power functions, polynomials,
exponential functions, and the sine and cosine functions.

The proof of the formulae is not examinable as we will be learning how to use them instead.

1. Constant Functions – Such functions imply that the value does not change. So if a function does not change, then its rate of change must
be zero.

𝑑𝑦 𝑑
𝑦=𝑐 → = 𝑐 =0
𝑑𝑥 𝑑𝑥

21
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
2. Power Functions – Such functions are of the form 𝑥 𝑛 where n is a real number (except 𝑛 = 0).

𝑑𝑦 𝑑 𝑛
𝑦 = 𝑥𝑛 → = 𝑥 = 𝑛 ⋅ 𝑥 𝑛−1
𝑑𝑥 𝑑𝑥

The reason why we exclude the case when 𝑛 = 0 is because 𝑦 = 𝑥 0 = 1 which is a constant and we recall what we learned about the
derivative of constant functions.

3. Constant multiple Rule – If 𝑐 is a constant and 𝑓(𝑥) is a differentiable function, then

𝑑𝑦 𝑑 𝑑
𝑦 =𝑐⋅𝑓 𝑥 → = 𝑐⋅𝑓 𝑥 =𝑐 𝑓 𝑥
𝑑𝑥 𝑑𝑥 𝑑𝑥
The constant 𝑐 simply “comes out” as shown.

4. The Sum and Difference Rule - If 𝑓(𝑥) and 𝑔(𝑥) are differentiable functions, then

𝑑𝑦 𝑑 𝑑 𝑑
𝑦=𝑓 𝑥 ± 𝑔 𝑥 → = 𝑓 𝑥 ± 𝑔(𝑥) = 𝑓 𝑥 ± [𝑔(𝑥)]
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥

22
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
5. Exponential Functions

𝑑𝑦 𝑑 𝑥
𝑦 = 𝑒𝑥 → = 𝑒 = 𝑒𝑥
𝑑𝑥 𝑑𝑥

6. Sine and Cosine Functions

𝑑𝑦 𝑑
𝑦 = sin 𝑥 → = sin 𝑥 = cos 𝑥
𝑑𝑥 𝑑𝑥

𝑑𝑦 𝑑
𝑦 = cos 𝑥 → = cos 𝑥 = −sin(𝑥)
𝑑𝑥 𝑑𝑥

Example 7: Find the derivative of the following functions:

𝑓 𝑥 = 𝑥 + 5𝑥 + 4, 𝑔 𝑥 = cos 𝑥 − 3 𝑥 − 5𝑥,

1
ℎ 𝑥 = 2𝑒 𝑥 − 5 sin 𝑥 + 4𝑥 3 , 𝑝 𝑥 = − sin(𝑥)
2 𝑥
23
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Solution:

1
′
1 1
−1 1 −1 1
𝑓 𝑥 = 𝑥2 + 5𝑥 + 4 → 𝑓 𝑥 = ⋅ 𝑥 2 +5⋅1+0= ⋅𝑥 2+5+0= +5
2 2 2 𝑥

1
𝑔 𝑥 = cos 𝑥 − 3 𝑥 − 5𝑥 = cos 𝑥 − 𝑥 3 − 5𝑥

′
1 1
−1 1 −2 1 1
→ 𝑔 𝑥 = − sin 𝑥 − ⋅ 𝑥 3 − 5 = − sin 𝑥 − 𝑥 3 − 5 = − sin 𝑥 − 2 − 5 = − sin 𝑥 − 3 −5
3 3 3 𝑥 2
3𝑥 3

ℎ 𝑥 = 2𝑒 𝑥 − 5 sin 𝑥 + 4𝑥 3 → ℎ′ 𝑥 = 2𝑒 𝑥 − 5 ⋅ cos 𝑥 + 4 ⋅ 3 ⋅ 𝑥 3−1 = 2𝑒 𝑥 − 5 cos 𝑥 + 12𝑥 2

1 1 −1 1
′
1 1
− −1 1 −3 1
𝑝 𝑥 = − sin 𝑥 = 𝑥 2 − sin 𝑥 → 𝑝 𝑥 = ⋅ − 𝑥 2 − cos 𝑥 = − 𝑥 2 − cos 𝑥 = − − cos(𝑥)
2 𝑥 2 2 2 4 4 𝑥3

24
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
You will not need to know how to prove the formula for derivatives of sine and cosine. However, we will go through them here because they
are a very good exercise in the use of the two most common trigonometric limits shown below. These were proven in Chapter 2 Section 4.

sin 𝜃 cos 𝜃 − 1
lim = 1, lim =0
𝜃→0 𝜃 𝜃→0 𝜃
Useful identities that we learned in Chapter 1: sin 𝑎 + 𝑏 = sin 𝑎 cos 𝑏 + cos 𝑎 sin 𝑏 , cos 𝑎 + 𝑏 = cos 𝑎 cos 𝑏 − sin 𝑎 sin(𝑏)

If 𝑓 𝑥 = sin(𝑥), then

𝑓 𝑥 + ℎ − 𝑓(𝑥) sin 𝑥 + ℎ − sin(𝑥) sin 𝑥 cos ℎ + sin ℎ cos 𝑥 − sin(𝑥)

𝑓 ′ 𝑥 = lim = lim = lim
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ

cos ℎ − 1 sin ℎ
= lim sin 𝑥 + cos 𝑥 = sin 𝑥 ⋅ 0 + cos 𝑥 ⋅ 1 = cos(𝑥)
ℎ→0 ℎ ℎ

If 𝑓 𝑥 = cos(𝑥), then

cos 𝑥 + ℎ − cos(𝑥) cos 𝑥 cos ℎ − sin 𝑥 sin ℎ − cos(𝑥) cos ℎ − 1 sin ℎ

𝑓 ′ 𝑥 = lim = lim = lim cos 𝑥 − sin 𝑥
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ ℎ
= cos 𝑥 ⋅ 0 − sin 𝑥 ⋅ 1 = −sin(𝑥)
25
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Example 8: When we consider the flow of blood through a blood vessel, such as a vein or artery, we can model the shape of the blood vessel
by a cylindrical tube with radius 𝑅 and length 𝑙 as illustrated in the figure
on the right.

Because of friction at the walls of the tube, the velocity 𝑣 of the blood is greatest along the central axis of the tube and decreases as the
distance 𝑟 from the axis increases until 𝑣 becomes 0 at the wall. The relationship between 𝑣 and 𝑟 is given by the law of laminar flow
discovered by the French physician Jean Poiseuille (pronounced John Pwa-zwei) in 1840. This law states that

𝑃
𝑣= (𝑅 2 − 𝑟 2 )
4𝜂𝑙
where 𝜂 (pronounced as eta) is the viscosity of the blood (a number that measures how a fluid’s resistance to flow – e.g. honey has higher
viscosity than water) and 𝑃 is the pressure (physical force per unit area acting on an object) difference between the ends of the tube. If 𝑃
and 𝑙 are constant, then 𝑣 is a function of 𝑟 with domain [0, 𝑅].

The velocity gradient is the instantaneous rate of change of velocity with respect to 𝑟. In general, the velocity gradient is a measure of the
rate of change of velocity with respect to distance. Simply put, they measure how the velocity of a fluid changes within different spatial
points of a fluid. The concept is also widely used to understand how solid material deform.
26
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
𝑃 𝑑𝑣 𝑃 𝑃𝑟
𝑣= 𝑅2 − 𝑟 2 → = −2𝑟 = − 0≤𝑟≤𝑅
4𝜂𝑙 𝑑𝑟 4𝜂𝑙 2𝜂𝑙

For one of the smaller human arteries we can take 𝜂 = 0.027, 𝑅 = 0.008 cm, 𝑙 = 2 cm, and 𝑃 = 4000 dynes/cm2 (0.4 Pascals or 0.4 Pa) ,
which gives

4000
𝑣= 0.0082 − 𝑟 2 ≈ 1.85 × 10−4 (6.4 × 10−5 − 𝑟 2 )
4(0.027)(2)
At 𝑟 = 0.002 cm, the blood is flowing at a speed of
𝑣 ≈ 1.85 × 10−4 6.4 × 10−5 − 4 × 10−6 = 1.11 cm/s
and the velocity gradient at that point is
You might be tempted to simplify but
𝑑𝑣 𝑃𝑟 4000 0.002 you must resist this temptation because
=− =− ≈ −74 (cm/s)/cm the simplification would lose its meaning!
𝑑𝑟 2𝜂𝑙 2 0.027 2
To get a feeling for what this statement means, let’s change our units from centimeters to micrometers (1 cm = 10,000 𝜇m). Then the radius
of the artery is 80 𝜇m. The velocity at the central axis is 11,850 𝜇m/s, which decreases to 11,110 𝜇m/s at a distance of 𝑟 = 20 𝜇m. The fact
that dv/dr=-74 (𝜇m/s)/𝜇m means that, when 𝑟 = 20 𝜇m, the velocity is decreasing at a rate of about 74 (𝜇m/s)/𝜇m for each micrometer
that we proceed away from the center of the artery.
27
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Let us explore the relationship between displacement, velocity and acceleration a little more with the following example. Before we begin
the example, let us ask ourselves these questions. They may appear to be a kinematics question but careful mathematical analysis can
actually address these questions.

Example 9: Suppose the displacement 𝑠 (in meters) of an object moving horizontally at time 𝑡 (in seconds) is given by 𝑠 𝑡 = 𝑡 2 − 5𝑡,
0 ≤ 𝑡 ≤ 5. Assume that positive values of 𝑠 correspond to positions to the right of 𝑠 = 0. This is an important point to note, it will appear
again in your Physics class!
• Sketch the velocity function on the interval 0 ≤ 𝑡 ≤ 5 and determine when the object is stationary, moving to the left, and moving to the
right.
• Describe the motion of the object.
(The above question about describing the motion is added in because mathematical modelling usually requires one to translate
mathematical results so let us gain some experience here)

Solution: Whenever you see problems, it is important to always first recall what you have learned. Here, you should begin to recall what is
the relationship between displacement, velocity and acceleration. That is, 𝑎 𝑡 = 𝑣 ′ 𝑡 = 𝑠 ′′ 𝑡 , 𝑣 𝑡 = 𝑠′(𝑡)

28
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
A sketch of the displacement function is given on the right. Recall from Chapter 1 that this is a quadratic function. The fact that the
displacement is negative shows that the object’s motion is to the left of the start point.

We need to differentiate the displacement function to get the velocity. So, 𝑣 𝑡 = 2𝑡 − 5 and
recalling what we learned in Chapter 1, the velocity is a straight line as shown in the bottom right.

Solving 𝑣 𝑡 = 2𝑡 − 5 > 0 and we see that the velocity is positive (i.e. moves to the right) for
x
5 5
< 𝑡 ≤ 5. Similarly, the velocity is negative (moves to the left) for 0 ≤ 𝑡 < . Though the velocity
2 2

5
of the object is increasing at all times, its speed is decreasing for 0 ≤ 𝑡 < , and then increasing
2

5 5
for < 𝑡 ≤ 5. Observe that 𝑡 = , the object is at its furthest position from the start point at
2 2

25 5
𝑠=− . What all this means is that the object first moved to the left from 𝑡 = 0 to at a speed x
4 2

5
of 5 m/s. Then at 𝑡 = , the object stops (see point x on displacement and velocity graphs) and
2

changes direction to move to the right. It then increases its speed till it returns to the start point
at 𝑠 = 0.
29
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Example 10: A company makes computer chips from square wafers of silicon. It wants to keep the side length of a wafer very close to 15 mm
and it wants to know how the area 𝐴(𝑥) of a wafer changes when the side length 𝑥 changes.
• Find 𝐴’(15) and explain its meaning in this situation.
• Show that the rate of change of the area of a square with respect to its side length is half its perimeter. Try to explain geometrically why
this is true by drawing a square whose side length 𝑥 is increased by an amount Δ𝑥. How can you approximate the resulting change in
area Δ𝐴 if Δ𝑥 is small?

Solution: The first thing is to find the area function. Since it is a square of side 𝑥, the area is
𝐴(𝑥) = 𝑥 2 . So 𝐴’(𝑥) = 2𝑥 and 𝐴′ 15 = 2 15 = 30 (mm2 )/mm is the rate at which the
area is increasing with respect to the side length as the side length 𝑥 reaches 15 mm.

The perimeter of the square is 𝑃 𝑥 = 4𝑥 → 𝑃 𝑥 = 2𝐴′ 𝑥 → 𝐴′ 𝑥 = 𝑃(𝑥)/2.

From the figure on the right, we see that if each side of the square is increased by Δ𝑥 then
the new area must be the sum of both of the original square plus the area of the shaded regions
2
shown. So, Δ𝐴 = new area − old area = area of shaded regions = 𝑥 Δ𝑥 + 𝑥 Δ𝑥 + Δ𝑥
= 2𝑥Δ𝑥 + Δ𝑥 2 . So if Δ𝑥 is small, then Δ𝐴 ≈ 2𝑥Δ𝑥.
30
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
The product rule involves the derivative of the product of two functions 𝑓(𝑥) and 𝑔(𝑥):

𝑑 𝑑 𝑑
𝑓 𝑥 ⋅𝑔 𝑥 = 𝑓 𝑥 ⋅𝑔 𝑥 + 𝑔 𝑥 ⋅ 𝑓 𝑥 = 𝑓 ′ 𝑥 ⋅ 𝑔 𝑥 + 𝑓 𝑥 ⋅ 𝑔′(𝑥)
𝑑𝑥 𝑑𝑥 𝑑𝑥

The quotient rule involves the derivative of the quotient of two functions 𝑓(𝑥) and 𝑔(𝑥), provided that 𝑔 𝑥 ≠ 0:

𝑑 𝑑
𝑑 𝑓 𝑥 𝑓 𝑥 ⋅𝑔 𝑥 − 𝑔 𝑥 ⋅𝑓 𝑥 𝑓 ′ 𝑥 ⋅ 𝑔 𝑥 − 𝑓 𝑥 ⋅ 𝑔′(𝑥)
= 𝑑𝑥 𝑑𝑥 =
𝑑𝑥 𝑔 𝑥 𝑔 𝑥 2 𝑔 𝑥 2

When using these rules, it is important to recognize which expression is 𝑓(𝑥) and 𝑔(𝑥) first!

Example 11: Differentiate the following functions:

𝑓 𝑥 = sin 𝑥 𝑒 𝑥 − 4𝑥 2 cos 𝑥 , 𝑔 𝑡 = 𝑎 𝑡 1 + 𝑡3 ,

𝑥 2 − 5𝑥 + 4
ℎ 𝑥 = , 𝑝 𝑥 = tan(𝑥)
cos 𝑥
31
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Solution: The complexity in the expressions you see arise in mathematical modelling so let us go through them step by step.

The product rule has been applied to the terms in red and green below and expanded accordingly.

𝑓 𝑥 = sin 𝑥 𝑒 𝑥 − 4𝑥 2 cos 𝑥
→ 𝑓 ′ 𝑥 = cos 𝑥 ⋅ 𝑒 𝑥 + sin 𝑥 ⋅ 𝑒 𝑥 − 8𝑥 ⋅ cos 𝑥 + 4𝑥 2 ⋅ − sin 𝑥
→ 𝑓 ′ 𝑥 = 𝑒 𝑥 cos 𝑥 + sin 𝑥 − 4𝑥(2 cos 𝑥 − 𝑥 sin(𝑥))

In the working below, we present two ways of solving the question: either by directly multiplying the terms in t or using the product rule.
Either way is acceptable.
1 1 1 1 7
3 3 3+
𝑔 𝑡 =𝑎 𝑡 1+𝑡 = 𝑎𝑡 2 1+𝑡 = 𝑎𝑡 2 + 𝑎𝑡 2 = 𝑎𝑡 2 + 𝑎𝑡 2

′
1 1−1 3
1
2
𝑎 −1 3
5 𝑎 −1 1 5 𝑎 −1 7 5
→ 𝑔 𝑡 = 𝑎 ⋅ 𝑡 2 1 + 𝑡 + 𝑎𝑡 2 3𝑡 = 𝑡 2 1 + 𝑡 + 3𝑎𝑡 2 = 𝑡 2 + + 3 𝑎𝑡 2 = 𝑡 2 + 𝑎𝑡 2
2 2 2 2 2 2

′
1 1−1 7 7−1 𝑎 −1 7 5
→ 𝑔 𝑡 = 𝑎 ⋅ 𝑡 2 + 𝑎 ⋅ 𝑡 2 = 𝑡 2 + 𝑎𝑡 2
2 2 2 2
32
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
The expression below is a straightforward application of the quotient rule.

𝑥 2 − 5𝑥 + 4
ℎ 𝑥 =
cos 𝑥

′
2𝑥 − 5 cos 𝑥 − (𝑥 2 − 5𝑥 + 4)(− sin 𝑥 ) 2𝑥 − 5 2
tan 𝑥
→ ℎ 𝑥 = = + 𝑥 − 5𝑥 + 4 ⋅
cos 𝑥 2 cos(𝑥) cos 𝑥
There is no need to oversimplify the expression so something like the above (in green) is acceptable.
For the last example, we can find the derivative of such functions using the quotient rule.

2
sin 𝑥 ′
cos 𝑥 ⋅ cos 𝑥 − sin 𝑥 ⋅ (− sin 𝑥 ) cos 2 𝑥 + sin2 (𝑥) 1 1
𝑝 𝑥 = tan 𝑥 = → 𝑝 𝑥 = = = = = sec 2 (𝑥)
cos 𝑥 cos 𝑥 2 cos(𝑥) 2 cos2 (𝑥) cos 𝑥

We present the derivatives of trigonometric functions below for your reference. They are a good exercise in using the quotient rule.
1
cot 𝑥 = ,
tan(𝑥)
1
sec 𝑥 = ,
cos 𝑥
1
csc 𝑥 =
sin 𝑥
33
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Example 12: Differentiate the following function and determine where are its horizontal tangents.

sec(𝑥)
𝑓 𝑥 =
1 + tan(𝑥)
Solution: Recall that the tangent to a function is the instantaneous slope. So a horizontal tangent refers to the horizontal slope and we are
looking for points where the function has zero derivative.

′
′
sec 𝑥 ⋅ 1 + tan 𝑥 − sec 𝑥 ⋅ [1 + tan(𝑥)]′ sec 𝑥 tan 𝑥 1 + tan 𝑥 − sec 𝑥 ⋅ [sec 2 (𝑥)]
𝑓 𝑥 = =
1 + tan 𝑥 2 1 + tan 𝑥 2

Recall that sin2 x + cos 2 𝑥 = 1 sec 𝑥 [tan 𝑥 + tan2 𝑥 − sec 2 (𝑥)]

=
1 + tan 𝑥 2
sin2 (𝑥) 1
→ + 1 =
cos 2 (𝑥) cos 2 (𝑥) sec 𝑥 [tan 𝑥 − 1]
=
1 + tan 𝑥 2
→ tan2 𝑥 + 1 = sec 2 (𝑥)

Remember that the zero derivative implies that the numerator expression is zero. So this means that
1
we need to look for values of 𝑥 where sec(𝑥) = 0 and tan(𝑥) = 1. Since sec(𝑥) = so it will never be zero because the numerator is 1.
cos 𝑥

𝜋
On the other hand, tan(𝑥) = 1 when 𝑥 = 𝑛𝜋 + . The plot of the function is given in above.
4 34
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
So far, we have looked at the derivative of basic functions. However, the reality is that most functions in mathematical modelling are
complex and are at least composite functions. Do review Chapter 1 Section 1 if it is a little hazy. To do this section, you will need to apply
what was learned in the previous section.

Definition: For the composite function 𝑓(𝑔(𝑥)) that is differentiable, the derivative of the composite function follows the chain rule:

𝑑 𝑑𝑓 𝑑𝑔
𝑓 𝑔 𝑥 = 𝑓′ 𝑔 𝑥 ⋅ 𝑔′ 𝑥 = ⋅
𝑑𝑥 𝑑𝑔 𝑑𝑥
We see that it is basically differentiating the outer function 𝑓 (in terms of 𝑔) and multiplying it with the derivative of the inner function 𝑔.

Example 13: Find the derivative of the following functions.

1 2
𝑎 𝑝 𝑥 = 𝑥 2 − 3𝑥, 𝑏 𝑝 𝑥 = sin 𝑥 2 , 𝑐 𝑝 𝑥 = sin2 (𝑥) , 𝑑 𝑝 𝑥 = sin ⋅ 𝑒 𝑥 −4𝑥
𝑥

Solution:
(a) The inner function is 𝑔 = 𝑥 2 − 3𝑥 and the outer function is 𝑔. Then,

′
1 −1 1 2𝑥 − 3
𝑝 𝑥 = g 2 ⋅ 2𝑥 − 3 = 2𝑥 − 3 =
2 2 𝑥 2 − 3𝑥 2 𝑥 2 − 3𝑥 35
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
(b) The inner function is 𝑔 = 𝑥 2 and the outer function is sin(𝑔). Then, 𝑝′ 𝑥 = cos 𝑔 ⋅ 2𝑥 = cos 𝑥 2 ⋅ 2𝑥

(c) The inner function is 𝑔 = sin(𝑥) and the outer function is 𝑔2 . Then, 𝑝′ 𝑥 = 2𝑔 ⋅ cos(𝑥) = 2 sin 𝑥 cos(𝑥)

(d) This question is more complicated. We first see that we have the product of composite functions. The following solution method is
proposed. We begin by looking at the derivative of both terms in the product.

1 1 1
𝑎 𝑥 = sin = sin 𝑥 −1 → 𝑎′ 𝑥 = cos 𝑥 −1 ⋅ −1 𝑥 −1−1 = cos −
𝑥 𝑥 𝑥2
2 −4𝑥 2 −4𝑥
𝑏 𝑥 = 𝑒𝑥 → 𝑏′ 𝑥 = 𝑒 𝑥 ⋅ 2𝑥 − 4
Then we have

𝑑 1 1 𝑥 2 −4𝑥
1 𝑥 2 −4𝑥
𝑎 𝑥 ⋅𝑏 𝑥 = 𝑎′ 𝑥 ⋅ 𝑏 𝑥 + 𝑎 𝑥 ⋅ 𝑏 ′ 𝑥 = cos − ⋅ 𝑒 + sin ⋅ 𝑒 ⋅ 2𝑥 − 4
𝑑𝑥 𝑥 𝑥2 𝑥

2 −4𝑥 1 1 1
= 𝑒𝑥 − cos ⋅ 2 + 2𝑥 − 4 sin
𝑥 𝑥 𝑥

36
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Example 14: When the diameter of a spherical tumor is 16 mm it is growing at a rate of 0.4 mm a day. How fast is the volume of the tumor
changing at that time?

Solution: We are told that the tumour is spherical and so we can use the volume of the sphere. In addition, the question gives us important
information: radius is 16/2 mm and the radius’ growth rate is given as 0.4 /2 mm/day. What the question wants us to find is the rate of
change of volume when the radius 𝑟 = 8 mm.

4 3 𝑑𝑟 0.4 𝑑𝑉
𝑉= 𝜋𝑟 , = = 0.2, at 𝑟 = 8?
3 𝑑𝑡 2 𝑑𝑡
The Chain rule in this case has the following form:

𝑑𝑉 𝑑𝑉 𝑑𝑟 4 𝑑𝑟 𝑑𝑟
= ⋅ = 𝜋 ⋅ 3𝑟 2 ⋅ = 4𝜋𝑟 2 ⋅
𝑑𝑡 𝑑𝑟 𝑑𝑡 3 𝑑𝑡 𝑑𝑡

So when 𝑟 = 8 and 𝑑𝑟/𝑑𝑡 = 0.2,

𝑑𝑉
= 4𝜋 8 2 (0.2) ≈ 161 mm3 /day
𝑑𝑡
Note that the above value is a specific volume growth rate at a specific radius and its corresponding growth rate.

37
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
The Chain Rule can also be used to find the derivative of exponential functions with arbitrary bases. Recall what was covered in Chapter 1.
For a constant 𝑏, we have

𝑥) 𝑑 𝑥 𝑑 𝑥 ln 𝑑
𝑏 𝑥 = 𝑒 ln(𝑏 = 𝑒 𝑥 ln(𝑏) → 𝑏 = 𝑒 𝑏
= 𝑒 𝑥 ln(𝑏) ⋅ 𝑥 ⋅ ln 𝑏 = ln 𝑏 𝑒 𝑥 ln 𝑏
𝑑𝑥 𝑑𝑥 𝑑𝑥
𝑑 𝑥
→ [𝑏 ] = 𝑏 𝑥 ln(𝑏)
𝑑𝑥
The Chain rule can also be applied to composite functions that comprise more than 2 basic functions.

Example 15: Find the derivative of the function 𝑝 𝑥 = sin(cos tan 𝑥 ).

Solution: The first thing to do is to determine the inner, middle and outer functions: 𝑓, 𝑔 and ℎ respectively.
𝑝 𝑥 = sin cos tan 𝑥 → 𝑓 = tan 𝑥 , 𝑔 = cos 𝑓 , 𝑝 = sin(𝑔)
Then, the extended Chain rule will look like

𝑑𝑝 𝑑𝑝 𝑑𝑔 𝑑𝑓
= ⋅ ⋅ = cos 𝑔 ⋅ − sin 𝑓 ⋅ sec 2 𝑥
𝑑𝑥 𝑑𝑔 𝑑𝑓 𝑑𝑥

𝑑𝑝
→ = cos cos tan 𝑥 ⋅ − sin tan 𝑥 ⋅ sec 2 (𝑥)
𝑑𝑥 38
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
As we mentioned in Chapter 1, a function is a function when it is one-to-one, i.e. an element of the domain maps to one element in the
range. Mathematically, we can write a function of the form 𝑦 = 𝑓(𝑥) for some function 𝑓.

However, there are curves where we cannot express in an explicit form: 𝑦 = 𝑓(𝑥). And with the tools we have so far, it is not possible to find
the derivative (or slope) of a point on the curve. If we cannot write 𝑦 = 𝑓(𝑥) in the first place, then how do we find 𝑑𝑦/𝑑𝑥?

This is where implicit differentiation comes in. Let us explore this through an example.

Example 16: Obtain the slope of any point on the red curve shown on the right. This red curve is called the folium
of Descartes (pronounced as day-car). Find the equation of the tangent line at the point (3,3) on the curve.

Solution: We assume that 𝑦 = 𝑓(𝑥) is defined implicitly. Then, we differentiate the given equation with respect to
𝑥. The chain rule will have to employed (see green terms) and then solve for 𝑑𝑦/𝑑𝑥.

𝑑 3 𝑑
𝑥 + 𝑦3 = 6𝑥𝑦
𝑑𝑥 𝑑𝑥
𝑑 3 𝑑 𝑑
→ 3𝑥 2 + 𝑦 = 6𝑥 ⋅ 𝑦 + 6𝑥 ⋅ 𝑦
𝑑𝑥 𝑑𝑥 𝑑𝑥 39
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
𝑑𝑦 𝑑𝑦
→ 3𝑥 2 + 3𝑦 2 ⋅ = 6𝑦 + 6𝑥 ⋅
𝑑𝑥 𝑑𝑥
𝑑𝑦
→ 3𝑥 2 − 6𝑦 = ⋅ (6𝑥 − 3𝑦 2 )
𝑑𝑥

𝑑𝑦 (3𝑥 2 − 6𝑦) (𝑥 2 − 2𝑦)

→ = =
𝑑𝑥 (6𝑥 − 3𝑦 2 ) (2𝑥 − 𝑦 2 )
The above expression gives us the value of the tangent (or slope or derivative) at some point (𝑥, 𝑦) on the curve.

The equation of the tangent line to a curve is an important concept that is often used to find an approximation to a function. More details
will be made later in this Chapter so stay tuned! But for now, let us obtain the equation of the tangent line. At the point (3,3), the slope of
the tangent line is given by

𝑑𝑦 (𝑥 2 − 2𝑦) 32 − 2(3) 9 − 6
= = = = −1
𝑑𝑥 (2𝑥 − 𝑦 2 ) 2 3 − 32 6 − 9
Recall the equation of a line is 𝑦 = 𝑚𝑥 + 𝑐 and so with 𝑚 = −1 (from above),
𝑦 = −𝑥 + 𝑐 → 3 = −3 + 𝑐 → 𝑐 = 6
∴ 𝑦 = −𝑥 + 6 is the equation of the tangent line to the curve at 3,3 .
40
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
We wrap up this section with a discussion of the derivatives of logarithmic and inverse trigonometric functions.

1. Logarithmic Functions of form 𝑦 = log 𝑏 𝑥

𝑦 = log 𝑏 𝑥
→ 𝑏𝑦 = 𝑥

We differentiate the above equation implicitly with respect to 𝑥 on both sides:

𝑑 𝑦 𝑑
𝑏 = 𝑥
𝑑𝑥 𝑑𝑥
𝑑𝑦
→ 𝑏 𝑦 ln 𝑏 ⋅ =1
𝑑𝑥

𝑑𝑦 1 1
→ = =
𝑑𝑥 𝑏 𝑦 ln(𝑏) 𝑥 ln(𝑏)
When the base 𝑏 = 𝑒, then

𝑑𝑦 1
𝑦 = log 𝑒 𝑥 = ln 𝑥 → =
𝑑𝑥 𝑥
41
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
Example 17: Differentiate the following functions:

𝑥+1
𝑎 𝑝 𝑥 = ln 𝑥 3 + 1 , 𝑏 𝑝 𝑥 = ln sin 𝑥 , 𝑐 𝑝 𝑥 = ln( 𝑥), 𝑑 𝑝 𝑥 = ln
𝑥−2

Solution:

3 ′
1 2
3𝑥 2
′ ′
𝑎 𝑓 = ln 𝑔 , 𝑔 =𝑥 +1 → 𝑝 𝑥 = 𝑓 𝑔 ⋅𝑔 𝑥 = 3 ⋅ 3𝑥 = 3
𝑥 +1 𝑥 +1

1
𝑏 𝑓 = ln 𝑔 , 𝑔 = sin(𝑥) → 𝑝′ 𝑥 = 𝑓 ′ 𝑔 ⋅ 𝑔′ 𝑥 = ⋅ cos(𝑥) = cot(𝑥)
sin 𝑥

′
1 1 1−1 1 1 1
′
1 ′
𝑐 𝑓 = ln 𝑔 , 𝑔 = 𝑥 → 𝑝 𝑥 =𝑓 𝑔 ⋅𝑔 𝑥 = ⋅ 𝑥2 = ⋅ ⋅ =
𝑥 2 2 𝑥 𝑥 2𝑥
We could also do the following as well.
1 1 1 1 1
𝑝 𝑥 = ln 𝑥 = ln(𝑥 2 ) = ln 𝑥 → 𝑝′ 𝑥 = ⋅ =
2 2 𝑥 2𝑥
42
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
𝑥+1
𝑑 𝑓 = ln 𝑔 , 𝑔=
𝑥−2

1 1
−
1 1⋅ 𝑥−2 − 𝑥+1 ⋅ 𝑥−2 2 ⋅1
′ ′
→ 𝑝 𝑥 =𝑓 𝑔 ⋅𝑔 𝑥 = ′
⋅ 2
2
𝑥+1 𝑥−2
𝑥−2

1 𝑥+1 1 𝑥+1 2 𝑥−2 −𝑥−1 𝑥−5

→ 𝑝′ 𝑥 = ⋅ 𝑥−2⋅ 𝑥−2 − = 𝑥−2 − = =
𝑥+1 𝑥−2 2 𝑥−2 𝑥+1 𝑥−2 2 2 𝑥+1 𝑥−2 2(𝑥 + 1)(𝑥 − 2)

We could also do the following as well, which involves the laws of logarithms:

𝑥+1 1 1
𝑝 𝑥 = ln = ln 𝑥 + 1 − ln 𝑥 − 2 = ln 𝑥 + 1 − ln 𝑥 − 2 2 = ln 𝑥 + 1 − ln 𝑥 − 2
𝑥−2 2

1 1 1 𝑥−5
→ 𝑝′ 𝑥 = − = 2 𝑥−2 − 𝑥+1 =
𝑥+1 2 𝑥−2 2 𝑥+1 𝑥−2 2(𝑥 + 1)(𝑥 − 2)

43
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
We do not have to be restricted to finding a logarithm function in order to use it in differentiation. Whenever we have a rational function
that comprises of products (such as the following example), we can employ the technique of logarithmic differentiation.

Example 18: Differentiate the following function.

3
𝑥4𝑥2 + 1
𝑦=
3𝑥 + 2 5
Solution: Apply the logarithm on both sides of the equation and apply the laws of logarithms:

3
𝑥4 𝑥2 + 1 3
5
3 1
ln 𝑦 = ln = ln 𝑥 4 + ln 𝑥 2 + 1 − ln 3𝑥 + 2 = ln 𝑥 + ln 𝑥 2 + 1 − 5 ln(3𝑥 + 2)
3𝑥 + 2 5 4 2
Now we differentiate the equation implicitly with respect to x:

1 𝑑𝑦 3 1 5
⋅ = + ⋅ 2𝑥 − ⋅3
𝑦 𝑑𝑥 4𝑥 2 𝑥 2 + 1 3𝑥 + 2

3
𝑑𝑦 3 𝑥 15 𝑥2 + 1 3 𝑥4
𝑥 15
→ =𝑦⋅ + 2 − = ⋅ + −
𝑑𝑥 4𝑥 𝑥 + 1 3𝑥 + 2 3𝑥 + 2 5 4𝑥 𝑥 2 + 1 3𝑥 + 2
44
 2. Basic Differentiation Formula, Product, Quotient and Chain Rules
The final formula for this section involves differentiating the inverse tangent function. Again, we employ the idea of implicit differentiation.
𝑦 = tan−1 (𝑥) → tan 𝑦 = 𝑥

𝑑 𝑑 𝑑𝑦 𝑑𝑦 1
→ tan 𝑦 = 𝑥 =1 → sec 2 𝑦 ⋅ =1 → =
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 sec 2 𝑦

But recall the relationship: tan2 (𝜃) + 1 = sec 2 (𝜃). So,

𝑑𝑦 1 1
= =
𝑑𝑥 sec 2 𝑦 𝑥2 + 1
Example 19: Differentiate the following function:
𝑦 = tan−1 𝑥 = arctan( 𝑥) → tan 𝑦 = 𝑥
Now we implicitly differentiate with respect to x:

𝑑 𝑑 𝑑𝑦 1
tan 𝑦 = 𝑥 → sec 2 (𝑦) ⋅ =
𝑑𝑥 𝑑𝑥 𝑑𝑥 2 𝑥
However unlike what we saw earlier, tan2 (𝜃) + 1 = sec 2 (𝜃) → tan2 𝑦 + 1 = sec 2 (𝑦) → 𝑥 2
+ 1 = sec 2 (𝑦)

𝑑𝑦 1 1 1 1 1
= ⋅ = ⋅ =
𝑑𝑥 sec 2 𝑦 2 𝑥 𝑥 + 1 2 𝑥 2 𝑥 + 1 𝑥
45
 3. Related Rates
You might have noticed that we went a little light on the applications in the last two sections. This is deliberate as we are now going to delve
into how to apply the derivative formula and rules in the mathematical modelling context. There is nothing new mathematically except the
general approaches taken when solving practical problems.

This section will focus on how we translating a word problem into a mathematical one and the emphasis is on derivatives. While this is what
we have been doing since Chapter 1, we are going to slowly turn the complexity up a little notch. The translation of a word problem into a
mathematical one is the most critical part of the mathematical modelling journey because if the problem to be solved is wrongly set up,
then you will not be solving what is required.

Below are some steps which may be helpful to you as you attempt to solve a word problem involving derivatives.
1. Read the problem carefully, making a sketch to organize the given information (only if you have difficulty visualizing the problem).
Identify the rates that are given and the rate that is to be determined.
2. Write one or more equations that express the basic relationships among the variables.
3. Introduce rates of change by differentiating the appropriate equations with respect to time t.
4. Substitute the known values and solve the desired quantity. Check that units are consistent and the answer is reasonable.
(e.g. correct sign or if you have some background knowledge, then does the answer make contextual sense)
46
 3. Related Rates
Example 20. An oil rig springs a leak in calm seas, and the oil spreads in a circular patch around the rig. If the radius of the oil patch increases
at a rate of 30 m/hr, how fast is the area of the patch increasing when the patch has a radius of 100 meters?

Background: The increase in the radius of the oil patch the speed of the prevailing and forecasted sea states (current, wind). When such
incidents occur, the local Meteorological Bureau will plot trajectory maps of an oil spill using ocean data like speed of currents. A
combination of satellite imaging and sea-based buoys are often used to obtain the sea states. The trajectory maps are usually used to warn
affected communities and to forecast the size of the affected zone. The latter is necessary to know where to position the containment boom
which will “scoop up” the oil spill. While the example is idealistic (calm seas and circular patch), the model presented in the solution is the
basis of more complex models.

47
 3. Related Rates
You should notice that there are two variables that are changing simultaneously: the radius of the circle and its area.

The key relationship between the radius and area is 𝐴 = 𝜋𝑟 2 . It helps to rewrite the basic relationship showing explicitly which quantities
vary in time. In this case, we rewrite 𝐴 and 𝑟 as 𝐴(𝑡) and 𝑟(𝑡) to emphasize that they change with respect to 𝑡 (time). Then, the general
expression relating the radius and area at any time t is then 𝐴 𝑡 = 𝜋 𝑟(𝑡) 2 .

The goal is to determine how fast is the area of the patch increasing. Mathematically, it is to find 𝑑𝐴/𝑑𝑡 and we can use the chain rule.

2
𝑑𝐴 𝑑𝑟 𝑑𝑟
𝐴 𝑡 = 𝜋 𝑟(𝑡) → =𝜋⋅2 𝑟 𝑡 ⋅ = 2𝜋 𝑟 𝑡 ⋅
𝑑𝑡 𝑑𝑡 𝑑𝑡

𝑑𝑟 𝑑𝐴 m2 m2
𝑟 = 100, = 30 → = 2𝜋 100 ⋅ 30 = 6000𝜋 ≈ 18,850
𝑑𝑡 𝑑𝑡 h h

It is important to remember that substitution of specific values of the variables occurs after differentiating.
Notice that the rate of change of the area depends on the radius of the spill.
As the radius increases, the rate of change of the area also increases.
48
 3. Related Rates
Example 21: Two small planes approach an airport, one flying due west at 120 miles per hour and the other flying due north at 150 miles per
hour. Assuming they fly at the same constant elevation, how fast is the distance between the planes changing when the westbound plane is
180 miles from the airport and the northbound plane is 225 miles from the airport?

Solution: You will need a little background knowledge here: velocity is the rate of change of
displacement with respect to time. Note that the direction will influence the sign of the velocity.

Let 𝑥(𝑡) and 𝑦(𝑡) denote the distance from the airport to the westbound and northbound
planes, respectively. In such problems, it is important to choose a frame of reference (shown in
red on the right) and it makes sense to choose the airport because we are not expecting the
airport to be mobile. This idea will prove helpful to you when studying Mechanics.

At the reference point, 𝑥 = 𝑦 = 0 and we insist that 𝑥 to increases eastwards and 𝑦 increases
northwards. This means that the positions of the two small planes are 𝑥 = 180 and 𝑦 = −225. Because the westward plane is travelling to
𝑑𝑥 dx
the left (opposite the direction of increasing 𝑥), <0 → = −120. Because the northward plane is travelling up (in the direction of
𝑑𝑡 dt

𝑑𝑦 𝑑𝑦
increasing 𝑦), >0 → = 150. 49
𝑑𝑡 𝑑𝑡
 3. Related Rates
The paths of the two planes form the legs of a right triangle, and the distance between them, denoted 𝑧(𝑡), is the hypotenuse. By the
Pythagoras’ theorem (we can use this because the two planes are flying at the same elevation): [𝑧(𝑡)]2 = [𝑥(𝑡)]2 +[𝑦(𝑡)]2 .

Note that while the lengths of the sides of the right triangle will change with time, the theorem will still be relevant because the two small
planes are maintaining both their altitude (height from the ground) and heading (direction of flight).

The goal is to find how fast is the distance between the planes changing. Mathematically, this means that we want to find 𝑑𝑧/𝑑𝑡.

𝑑 𝑑 𝑑
[𝑧(𝑡)]2 = [𝑥(𝑡)]2 +[𝑦(𝑡)]2 → 𝑧 𝑡 2
= 𝑥 𝑡 2
+ 𝑦 𝑡 2
𝑑𝑡 𝑑𝑡 𝑑𝑡
𝑑𝑧 𝑑𝑥 𝑑𝑦
→ 2𝑧 ⋅ = 2𝑥 ⋅ + 2𝑦 ⋅
𝑑𝑡 𝑑𝑡 𝑑𝑡
𝑑𝑥 𝑑𝑦
We are told that the two planes are travelling at a constant speed: = −120, = 150. At the time of interest, we are given 𝑥 = 180 and
𝑑𝑡 𝑑𝑡

𝑦 = −225. The corresponding value of z is 𝑧 = 1802 + 2252 ≈ 288. So,

𝑑𝑧 1 𝑑𝑥 𝑑𝑦 1
= 2𝑥 ⋅ + 2𝑦 ⋅ = 180 ⋅ −120 + (−225)(150)] ≈ −192 mi/h < 0
𝑑𝑡 2𝑧 𝑑𝑡 𝑑𝑡 288

The negative value for 𝑑𝑧/𝑑𝑡 makes sense because the planes are approaching the airport and 𝑧 is decreasing. 50
 3. Related Rates
Example 22: A 12-foot ladder is leaning against a vertical wall when Jack begins pulling the foot of the ladder away from the wall at a rate of
0.2 ft/s. What is the configuration of the ladder at the instant that the vertical speed of the top of the ladder equals the horizontal speed of
the foot of the ladder?

Solution: In this question, it is important to remember that the ladder is not going to change its
length. We are assuming that the wall is vertical and the floor is horizontal. The reference point
here is the base of the vertical wall, where ℎ increases upwards and 𝑥 increases leftwards.

We are told that Jack is pulling the foot of the ladder away from the wall at a rate of 0.2 ft/s. That is, 𝑑𝑥/𝑑𝑡 = 0.2. Because the friction of
the floor and the wall are not usually the same (the floor is not always made of brick and the wall is not always made of concrete), we
cannot assume that the magnitude of dh/dt is the same as 𝑑𝑥/𝑑𝑡. However in this instance, we are fortunate that the question wants us to
consider the case where |𝑑𝑥/𝑑𝑡| = |𝑑ℎ/𝑑𝑡|. And because h is decreasing as time passes, 𝑑ℎ/𝑑𝑡 = −0.2.

The goal here is to find the configuration of the ladder and given the question, the only thing this means what is the value of x and h in
which the vertical speed of the top of the ladder equals the horizontal speed of the foot of the ladder.

51
 3. Related Rates
Pythagoras’ theorem applies here as we have a right triangle and we have
𝑥 2 + ℎ2 = 12 2
= 144

𝑑 2 𝑑
→ 𝑥 + ℎ2 = 144
𝑑𝑡 𝑑𝑡
𝑑𝑥 𝑑ℎ
→ 2𝑥 ⋅ + 2ℎ ⋅ =0
𝑑𝑡 𝑑𝑡
→ 2𝑥 ⋅ 0.2 + 2ℎ ⋅ −0.2 = 0

12 12 2
→ 𝑥−ℎ=0 → 𝑥 =ℎ → 𝑥 2 + ℎ2 = 12 → 2𝑥 2 = 12 → 𝑥 = ℎ = = ⋅ =6 2
2 2 2

Example 23: A 5-foot-tall woman walks at 8 ft/s toward a street lamppost that is 20 ft above the ground. What is the rate of change of the
length of her shadow when she is 15 ft from the street lamppost? At what rate is the tip of her shadow moving?

Solution: There are two objects of interest here: a 5-foot-tall woman and a street lamppost. We are assuming that the lamppost is not
moving (they generally don’t).

52
 3. Related Rates
If we set the reference point to be at the base of the lamppost, then
• the distance 𝑥 be the horizontal distance between the lamppost and the woman. So, the
woman’s speed is given by 𝑑𝑥/𝑑𝑡.
• let 𝑦 be the horizontal distance between the feet of the woman and the tip of her shadow
on the ground.
𝑑𝑥
• 𝑥 and 𝑦 increase in a leftward direction. As the woman is walking towards the lamppost, = −8.
𝑑𝑡

The goal of this question is to determine the rate in which the tip of her shadow is moving. Mathematically, this means we want

𝑑
(𝑥 + 𝑦)
𝑑𝑡
The above expression looks strange but it is an important application of relative speeds (seen a lot in Mechanics). The reason is that given
how we have defined 𝑦, 𝑑𝑦/𝑑𝑡 is only referring to the speed of the tip of the woman’s shadow relative to her position.

So unless this is specifically mentioned in the question, we always assume that the required rate is with respect to a fixed point.

53
 3. Related Rates
Using the law of similar triangles, we have

𝑥+𝑦 𝑦
=
20 5
→ 𝑥 + 𝑦 = 4𝑦 → 𝑥 = 3𝑦

The velocity of the woman and the shadow are related by

𝑑𝑥 𝑑𝑦 𝑑𝑦 1 𝑑𝑥 1 8
=3 → = = −8 = −
𝑑𝑡 𝑑𝑡 𝑑𝑡 3 𝑑𝑡 3 3

Then, the required rate is then

20𝑦 𝑑 𝑑
→ 𝑥+𝑦 = → 𝑥+𝑦 =4⋅ 𝑦
5 𝑑𝑡 𝑑𝑡

𝑑 𝑑𝑦 8 32 ft
→ 𝑥+𝑦 =4 =4 − =−
𝑑𝑡 𝑑𝑡 3 3 s

Again, this result makes sense since we expect that as the woman walks towards the lamppost, the tip of the shadow will also be moving
towards the lamppost.
54
 3. Related Rates
Example 24: A ship leaves port traveling southwest at a rate of 12 miles per hour. At noon, the ship reaches its closest approach to a radar
station, which is on the shore 1.5 miles from the port. If the ship maintains its speed and course, what is the rate of change of the tracking
angle 𝜃 between the radar station and the ship at 1:30 p.m. (see coloured figure below)? Hint: Use Law of Sines.

Solution: Let 𝑠 be the distance travelled by the ship from the port in a
southwest direction. The relevant triangle is given below, bearing in mind
that the distance between the radar station and port (1.5 mi) and the heading 𝑥

of the ship will not change. 1.5

𝜃 𝜋
4
We are given 𝑑𝑠/𝑑𝑡 = 12
Given the triangle on the left, law of sines will allow us to obtain the equation:
and we want to find 3𝜋 𝑠 = 𝑥 + 12 1.5
−𝜃 = 𝑥 + 18
4 3𝜋
𝑑𝜃/𝑑𝑡. At noon, the shortest sin −𝜃
sin 𝜃 4 3𝜋
distance between the = → 1.5 sin 𝜃 = 𝑠 ⋅ sin −𝜃
𝑠 1.5 4
ship and the radar station is ⊥ to the
3𝜋 3𝜋 𝑠 2
ship’s trajectory (ref. the coloured figure). → 1.5 sin 𝜃 = 𝑠 ⋅ sin cos 𝜃 − cos sin 𝜃 = cos 𝜃 + sin 𝜃
4 4 2
𝑥 𝜋 3 2 3 2
So, = cos → 𝑥= =
1.5 4 2 2 4 55
 3. Related Rates
Now we want to simplify the expression in order to have 𝜃(𝑠). To do this, we recall what we learned in Chapter 1. We are trying to get a
single function so that the we can work out the inverse function.

𝑠 2 3 𝑠 2
1.5 sin 𝜃 = cos 𝜃 + sin 𝜃 → tan 𝜃 = 1 + tan 𝜃 → tan 𝜃 3 − 𝑠 2 = 𝑠 2
2 2 2

𝑠 2
→ tan 𝜃 =
3−𝑠 2
But there is no need to find 𝜃(𝑠) because we can now differentiate the above equation with respect to time, using 𝑑𝑠/𝑑𝑡 = 12 to find
𝑑𝜃/𝑑𝑡.

d 𝑑 𝑠 2 𝑑𝜃 2 3−𝑠 2 −𝑠 2 − 2 𝑑𝑠
[tan 𝜃 ] = → sec 2 𝜃 ⋅ = 2 ⋅
dt 𝑑𝑡 3 − 𝑠 2 𝑑𝑡 3−𝑠 2 𝑑𝑡

𝑑𝜃 1 2 3−𝑠 2 −𝑠 2 − 2 𝑑𝑠 1 3 2 − 2𝑠 + 2𝑠 1 36 2
→ = ⋅ 2 ⋅ = ⋅ 2 ⋅ 12 = ⋅ 2
𝑑𝑡 sec 2 𝜃 3−𝑠 2 𝑑𝑡 sec 2 𝜃 3−𝑠 2 sec 2 𝜃
3−𝑠 2

56
 3. Related Rates
𝑠 2
tan 𝜃 = & 1 + tan2 𝜃 = sec 2 𝜃
3−𝑠 2

2 2 2
2
𝑠 2 3−𝑠 2 + 2𝑠 2 3−𝑠 2 + 2𝑠 2 9 − 6𝑠 2 + 2𝑠 2 + 2𝑠 2 9 − 6𝑠 2 + 4𝑠 2
→ sec 𝜃) = 1 + = 2 = 2 = 2 = 2
3−𝑠 2 3−𝑠 2 3−𝑠 2 3−𝑠 2 3−𝑠 2

2
𝑑𝜃 1 36 2 3−𝑠 2 36 2 36 2
→ = ⋅ 2 = ⋅ 2 =
𝑑𝑡 sec 2 𝜃 3−𝑠 2 9 − 6𝑠 2 + 4𝑠 2 3−𝑠 2 9 − 6𝑠 2 + 4𝑠 2

The distance travelled by the ship is the product of speed and time and we have to add the distance from the time of departure to 12 pm
(when radar station is closest to ship) and the distance from 12 pm to 1:30 pm (distance covered in 1.5 hours), i.e.

3 2 𝑑𝜃 36 2 36 2
𝑠 = 𝑥 + 12 1.5 = + 18 ≈ 19.06 → = = 2
≈ 0.04
4 𝑑𝑡 9 − 6 𝑠 2 + 4 𝑠 2 9 − 6 19.06 + 4 19.06

The rate of change of the tracking angle with respect to time is 0.4 radians per hour.
57
 3. Related Rates
Example 25: Two boats leave a port at the same time; one travels west at 20 mi/h and the other travels south at 15 mi/h. At what rate is the
distance between them changing 30 minutes after they leave the port?

Solution: Let 𝑥(𝑡) be the distance that the westbound boat has travelled at time 𝑡 and 𝑦(𝑡) be the distance travelled by the southbound
boat at time 𝑡. Let the reference point be situated at the port where 𝑥(𝑡) increases westwards and 𝑦(𝑡) increases southwards. As we did
earlier, let the distance between the two boats be 𝑧(𝑡). By Pythagoras’ theorem, we have

2 2 2
𝑑𝑧 𝑑𝑥 𝑑𝑡 𝑑𝑧 1 𝑑𝑥 𝑑𝑦
𝑧 𝑡 =𝑥 𝑡 +𝑦 𝑡 → 2𝑧 = 2𝑥 + 2𝑦 → = 𝑥 +𝑦
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑧 𝑑𝑡 𝑑𝑡

2
𝑑𝑥 𝑑𝑦 2 2 15
= 20, = 15, 𝑧 0.5 = 𝑥 0.5 + 𝑦 0.5 = 102 + = 12.5
𝑑𝑡 𝑑𝑡 2

𝑑𝑧 1 15
∴ = 10 ⋅ 20 + 15 ⋅ = 25 mi/h
𝑑𝑡 12.5 2

58
 3. Related Rates
Example 26: A camera is set up at the starting line of a drag race 50 ft from a dragster at the starting line (camera 1 in the figure). Two
seconds after the start of the race, the dragster has traveled 100 ft and the camera is turning at 0.75 rad/s while filming the dragster.
a. What is the speed of the dragster at this point (after 2 seconds)?
b. A second camera (camera 2 in the figure) filming the dragster
is located on the starting line 100 ft away from the dragster at the
start of the race. How fast is camera 2 turning 2 seconds after
the start of the race?

Solution: Let A be the point where the dragster started, let B be the
point where camera 1 is located and let C denote the position of the dragster at time 𝑡 and 𝑦(𝑡) be the distance between A and C. Let 𝜃 be
angle ABC, as shown on the right. We C

𝑦 𝑑𝜃 1 𝑑𝑦 𝑦(𝑡)
tan 𝜃 = → sec 2 (𝜃) ⋅ = ⋅
50 𝑑𝑡 50 𝑑𝑡
𝑦2
𝜃
2 2 2 2 2
Recall 1 + tan 𝜃 = sec 𝜃 → 1+ = sec (𝜃). At 𝑡 = 2, 𝑦 = 100 → sec 𝜃 = 1 + 2 = 5 B 50 A
50 2

𝑑𝑦 𝑑𝜃
= 50 ⋅ sec 2 𝜃 ⋅ = 50 ⋅ 5 ⋅ 0.75 = 187.5 ft/s
𝑑𝑡 𝑑𝑡 59
 3. Related Rates
Example 27: A camera is set up at the starting line of a drag race 50 ft from a dragster at the starting line (camera 1 in the figure). Two
seconds after the start of the race, the dragster has traveled 100 ft and the camera is turning at 0.75 rad/s while filming the dragster.
a. What is the speed of the dragster at this point (after 2 seconds)?
b. A second camera (camera 2 in the figure) filming the dragster
is located on the starting line 100 ft away from the dragster at the
start of the race. How fast is camera 2 turning 2 seconds after
the start of the race?

Solution: Let D be the point where camera 2 is located.

𝑦 𝑑𝜙 1 𝑑𝑦
tan 𝜙 = → sec 2 𝜙 ⋅ = ⋅ C
100 𝑑𝑡 100 𝑑𝑡
𝑑𝑦
Since camera 2 is tracking the same dragster, = 187.5 and we want to find 𝑑𝜙/𝑑𝑡. At the same 𝑡 = 2, 𝑦(𝑡)
𝑑𝑡

2 𝜙
2 2
𝑦2 2
100 2
1 + tan 𝜙 = sec 𝜙 → 1+ = sec (𝜙) → 1+ = 2 = sec (𝜙) D
100 2 100 100 A

𝑑𝜙 1 rad This result makes sense since camera 2 is further away

→ = ⋅ 187.5 = 0.9375
𝑑𝑡 200 s from the car and has to rotate faster than camera 1. 60
 3. Related Rates
Example 28: A water heater that has the shape of a right cylindrical tank with a radius of 1 ft and a height of 4 ft is being drained. How fast is
water draining out of the tank (in ft 3 /min) if the water level is dropping at 6 in/min? Note: 1 feet (ft) = 12 inches (in).

Solution: Let ℎ(𝑡) be the height of the water in the tank at time 𝑡. Then, the volume of the water in the tank at time 𝑡 is given by the volume
of the cylinder 𝑉 = 𝜋𝑟 2 ℎ = 𝜋ℎ. Our goal is to find 𝑑𝑉/𝑑𝑡 when 𝑑ℎ/𝑑𝑡 = −0.5 ft/min.

𝑑𝑉 𝑑𝑉 𝑑ℎ 1
= ⋅ = 𝜋 ⋅ −
𝑑𝑡 𝑑ℎ 𝑑𝑡 2
𝜋 𝜋
So the volume of water in the tank is decreasing at ft 3 /min, the water is draining out at cubic feet per minute as well. It is interesting to
2 2

note that the water drainage rate is independent of the tank’s height. This makes sense because the volume of a cylindrical tank is linearly
related to the height. So, 𝑑𝑉/𝑑𝑡 is proportional to 𝑑ℎ/𝑑𝑡. So if we are given one quantity (e.g. 𝑑ℎ/𝑑𝑡), we can find the other (𝑑𝑉/𝑑𝑡).

Motivation time: It is hoped that the few examples we looked at here will give you a flavour of the potential applications that can arise in
real life. A really robust way to model real world scenarios is to try to strike a balance between a sufficiently simple model coupled with
sufficiently realistic conditions. So, the more Maths one learns, the more sophisticated one’s models can be. So keep learning new
techniques and keep persevering! 61
 4. Linear Approximations, Taylor Polynomials
We have seen that a curve lies very close to its tangent line near the point of tangency. In fact, by zooming in toward a point on the graph of
a differentiable function, we noticed that the graph looks more and more like its tangent line. See the figure below. This observation is the
basis for finding the approximate values of functions.

We can use the tangent line at (𝑎, 𝑓(𝑎)) as an approximation to the curve 𝑦 = 𝑓(𝑥) when 𝑥 is near 𝑎.

So how do we find the equation of the tangent line to a curve 𝑦 = 𝑓(𝑥) at some point 𝑥 = 𝑎?

We start with two points: (𝑎, 𝑓(𝑎)) on the curve 𝑦 = 𝑓(𝑥) and some arbitrary point on the tangent
line: (𝑥, 𝑦).

Then, the slope of the tangent line is given by 𝑓’(𝑎). So,

𝑦 − 𝑓(𝑎)
𝑓′ 𝑎 = → 𝑦 = 𝑓 𝑎 + 𝑓 ′ 𝑎 ⋅ 𝑥 − 𝑎 = 𝐿(𝑥)
𝑥−𝑎
This is the equation of the tangent line to the curve 𝑦 = 𝑓(𝑥) at the point (𝑎, 𝑓(𝑎)) and is also the linear (or tangent line) approximation to
𝑦 = 𝑓(𝑥) when 𝑥 is near 𝑎. 𝐿(𝑥) is called the linearization of the function 𝑓 at 𝑎.
62
 4. Linear Approximations, Taylor Polynomials
Example 29: Find the linearization of the function 𝑓 𝑥 = 𝑥 + 3 at 𝑎 = 1 and use it to approximate the numbers 3.98 and 4.05. Are
these approximations overestimates or underestimates?

Solution: Overestimate means that the approximate value is more than what it actually is and similarly, an underestimate means the
approximate value is less than what is actually is.

1
′
1 1
−1 𝑑 1
𝑓 𝑥 = 𝑥+3= 𝑥+3 2 → 𝑓 𝑥 = 𝑥+3 2 ⋅ 𝑥+3 =
2 𝑑𝑥 2 𝑥+3
So, the linearization of 𝑓 at 𝑥 = 𝑎 = 1 is given by
𝐿 𝑥 = 𝑓 𝑎 + 𝑓′ 𝑎 ⋅ 𝑥 − 𝑎 = 𝑓 1 + 𝑓′ 1 ⋅ 𝑥 − 1

1 1 𝑥+7
→ 𝐿(𝑥) = 1 + 3 + ⋅ 𝑥−1 =2+ 𝑥−1 =
2 1+3 4 4

So when we try to approximate 3.98 = 0.98 + 3 = 𝑓(0.98) (so we see that 0.98 is near 𝑥 = 𝑎 = 1) can be approximated by

0.98 + 7 7.98
3.98 = 𝑓 0.98 ≈ = = 1.995
4 4
(actual value of 3.98 = 1.9949; very close!; overestimate)

63
 4. Linear Approximations, Taylor Polynomials
Likewise to approximate 4.05 = 1.05 + 3 = 𝑓 1.05 , we have

1.05 + 7 8.05
4.05 = 𝑓 1.05 ≈ = = 2.0125
4 4
(actual value is 2.01246; very close! ; overestimate)

Example 30: Find the linear approximation to sin(𝑥) at 𝑥 = 0.

Solution: The linearization to the function 𝑓(𝑥) = sin(𝑥) at 𝑥 = 𝑎 = 0 is given by

𝐿 𝑥 = 𝑓 𝑎 + 𝑓′ 𝑎 ⋅ 𝑥 − 𝑎
→ 𝐿 𝑥 = 𝑓 0 + 𝑓 ′ 0 ⋅ 𝑥 − 0 = sin 0 + cos 0 ⋅ 𝑥 = 𝑥
So we see that sin 𝑥 ≈ 𝑥, 𝑥 is near 0.

This is also why

sin 𝑥
lim =1
𝑥→0 𝑥
as the behaviour of sin(𝑥) around 𝑥 = 0 is similar to the function 𝑦 = 𝑥.
64
 4. Linear Approximations, Taylor Polynomials
The tangent line approximation 𝐿(𝑥) is the best first-degree (linear) approximation to 𝑓(𝑥) near 𝑥 = 𝑎 because 𝑓(𝑥) and 𝐿(𝑥) have the
same rate of change (derivative) at 𝑥 = 𝑎.

For a better approximation than a linear one, let’s try a second-degree (quadratic) approximation 𝑃(𝑥).

In other words, we approximate a curve by a parabola instead of by a straight line. To make sure that the approximation is a good one, we
stipulate the following:
• 𝑃(𝑎) = 𝑓(𝑎)
• 𝑃’(𝑎) = 𝑓’(𝑎)
• 𝑃’’(𝑎) = 𝑓’’(𝑎)

The second order polynomial will then have the form: 𝑃(𝑥) = 𝐴 + 𝐵(𝑥 − 𝑎) + 𝐶 𝑥 − 𝑎 2 . Applying the three conditions above, we have
• 𝑃 𝑎 =𝑓 𝑎 → 𝐴=𝑓 𝑎
• 𝑃’ 𝑎 = 𝑓’ 𝑎 → 𝐵 = 𝑓′(𝑎)

′′ ′′ ′′ 𝑓′′ 𝑎
• 𝑃 𝑎 =𝑓 𝑎 → 2𝐶 = 𝑓 𝑎 → 𝐶=
2

65
 4. Linear Approximations, Taylor Polynomials
So, the required quadratic approximation to 𝑓(𝑥) [also called second-degree Taylor polynomial of 𝑓 centered at 𝑎] is given by

𝑓′′(𝑎) 2
𝑃 𝑥 = 𝑓 𝑎 + 𝑓’ 𝑎 ⋅ 𝑥 − 𝑎 + ⋅ 𝑥−𝑎
2

Then following the approach we have taken so far, the 𝑛-th degree Taylor polynomial of 𝑓 centered at 𝑎 is given by

′
𝑓 ′′ 𝑎 2
𝑓 (3) 𝑎 3
𝑓 (𝑛) 𝑎 𝑛
𝑇𝑛 𝑥 = 𝑓 𝑎 + 𝑓 𝑎 ⋅ 𝑥 − 𝑎 + ⋅ 𝑥−𝑎 + ⋅ 𝑥−𝑎 +⋯ ⋅ 𝑥−𝑎
2 3! 𝑛!

where 𝑛 is a positive integer and 𝑛! = 𝑛(𝑛 − 1)(𝑛 − 2) … (3)(2)(1).

Example 31: Find the first three Taylor polynomials for the function 𝑓(𝑥) = ln(𝑥) centered at 𝑎 = 1.

Solution: We begin by finding the appropriate derivatives of 𝑓(𝑥).

1 1 2
𝑓 𝑥 = ln 𝑥 → 𝑓′ 𝑥 = → 𝑓 ′′ 𝑥 = −1 𝑥 −1−1 = − → 𝑓 ′′′ 𝑥 = −1 −2 𝑥 −2−1 =
𝑥 𝑥2 𝑥3

66
 4. Linear Approximations, Taylor Polynomials
1 1 2
𝑓 1 = ln 1 = 0, 𝑓′ 1 = = 1, 𝑓 ′′ 1 = − = −1, 𝑓 ′′′ 1
= =2
1 12 13

Then, the required Taylor polynomials that are centered at 𝑎 = 1 are given by

1
𝑇1 𝑥 = ln 1 + ⋅ 𝑥−1 =𝑥−1
(1)

1 −1 2
𝑇2 𝑥 = ln 1 + ⋅ 𝑥−1 + ⋅ 𝑥−1
1 1 22

1 2
→ 𝑇2 (𝑥) = 𝑥 − 1 − 𝑥−1
2

1 −1 2
2 3
𝑇3 𝑥 = ln 1 + ⋅ 𝑥−1 + ⋅ 𝑥−1 + ⋅ 𝑥−1
1 1 22 1 3 3!

1 2
1 3
→ 𝑇3 (𝑥) = 𝑥 − 1 − 𝑥−1 + 𝑥−1
2 3
67
 4. Linear Approximations, Taylor Polynomials

Notice that these polynomial approximations are better when 𝑥 is close

to 1 and that each successive approximation is better than the
preceding ones.

This suggests that if we want the approximation to be very accurate, we

require 𝑇𝑛 𝑥 , 𝑛 → ∞.

It is this idea that leads us to the Taylor series for a function centered at 𝑎.

′
𝑓 ′′ 𝑎 2
𝑓 𝑛
𝑎 𝑛
𝑓 𝑥 =𝑓 𝑎 +𝑓 𝑎 ⋅ 𝑥−𝑎 + ⋅ 𝑥−𝑎 + ⋯+ ⋅ 𝑥−𝑎 +⋯
2 𝑛!

∞ 𝑛
𝑓 𝑎 𝑛
→ 𝑓(𝑥) = ෍ ⋅ 𝑥−𝑎 0! = 1
𝑛!
𝑛=0
68
 4. Linear Approximations, Taylor Polynomials
The Maclaurin Series is a special case of the Taylor series that is centered at 𝑎 = 0.

∞
′
𝑓 ′′ 0 2
𝑓 𝑛
0 𝑛
𝑓 𝑛
0 𝑛
𝑓 𝑥 =𝑓 0 +𝑓 0 ⋅ 𝑥−0 + ⋅ 𝑥−0 +⋯+ ⋅ 𝑥−0 +⋯= ෍ ⋅ 𝑥 0! = 1
2 𝑛! 𝑛!
𝑛=0

We saw in the previous slide that the Taylor series for a function centered at 𝑎 works by summing an infinite number of Taylor polynomials
at a value of 𝑥 that is close to 𝑎. In practical reality, we will always assume that we will pick 𝑥 close to a.

69