Unit 1: Limits and Continuity Review

The Limit

We looked at finding the limit as the value the neighbourhood of a function approaches for arbitrarily small
neighbuorhoods around a point 𝑥 = 𝑐. Consider
1 √6 − 𝑥
−
lim 𝑥 4
𝑥→2 𝑥−2

Then the graph reveals that as we tighten up our

neighbourhood around 𝑥 = 2 that 𝑓(𝑥) looks like it
approaches a value around −0.2.

A table of values gives us a better picture and we see

that the limit is around −0.1875 (at least to four
decimal places).

The only way to do better is to use algebra and simplify

1 √6 − 𝑥
− 4 4 − 𝑥√6 − 𝑥
lim 𝑥 = lim
𝑥→2 𝑥−2 𝑥→2 4𝑥(𝑥 − 2)
16 − 𝑥 2 (6 − 𝑥)
= lim
𝑥→2 4𝑥(𝑥 − 2)(4 + 𝑥√6 − 𝑥)

(𝑥 − 2)(𝑥 2 − 4𝑥 − 8)
= lim
𝑥→2 4𝑥(𝑥 − 2)(4 + 𝑥√6 − 𝑥)

𝑥 2 − 4𝑥 − 8
= lim
𝑥→2 4𝑥(4 + 𝑥√6 − 𝑥)
3
=−
16
Which happens to be −0.1875 exact!

𝑥 1.99 1.999 1.9999 2.0001 2.001 2.01

𝑓(𝑥) −0.18880 −0.18763 −0.18751 −0.18749 −0.18737 −0.18622

We say the limit does not exist if it does not approach a real number from both sides (left and right) and in the case
1 √6−𝑥 1 √6−𝑥
− −
𝑥 4 𝑥 4
above where there is an asymptote at 𝑥 = 0 we could say that lim− 𝑥−2
= ∞ and lim+ 𝑥−2
= −∞ BUT keep in
𝑥→0 𝑥→0
mind that infinity in not a number and we can not encircle it like we can real limit points.

Functions grow in the following order from fastest to slowest

1. Exponential: 𝑦 = 𝑒 𝑥
2. Polynomial: 𝑦 = 𝑥 𝑛 , 𝑛 > 1
3. Linear: 𝑦 = 𝑥
4. Polynomial (radical): 𝑦 = 𝑥 𝑛 , 0 < 𝑛 < 1
5. Logarithmic: 𝑦 = ln 𝑥
6. Bounded (trig, constants, and exponential in the other direction – that is no growth): 𝑦 = 𝑒 −𝑥

When adding or subtracting functions only the largest growth matters. When dividing we have three cases
𝐹𝑎𝑠𝑡 𝑆𝑙𝑜𝑤 𝑠𝑎𝑚𝑒 𝑎
a. 𝑆𝑙𝑜𝑤 → ∞ b. 𝐹𝑎𝑠𝑡 → 0 c. →𝐿=
𝑠𝑎𝑚𝑒 𝑏
Where 𝑎, 𝑏 are related to the leading coefficients
Unit 1: Limits and Continuity Review

We also used Sandwich Theorem to find limits of complex functions using simpler ones. If 𝑔(𝑥) < 𝑓(𝑥) < ℎ(𝑥) for all
𝑥 ∈ (𝑎, 𝑏) expect at some point 𝑥 = 𝑐 AND lim 𝑔(𝑥) = lim ℎ(𝑥) = 𝐿 then lim 𝑓(𝑥) = 𝐿
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

sin 𝐴 1−cos 𝐴
Namely we saw that lim = 1 and lim =0
𝐴→0 𝐴 𝐴→0 𝐴

Things we need to know and understand:

• How to calculate the limit from a graph or table of values (with or without graphing calculator)
• How to calculate the limit using algebra and known limit values
• How to prove a limit using epsilon-delta
• That the limit exists when its one-sided limits approach the same value.
• Sandwich theorem is useful to find the limit of a complex function that is stuck between two simple functions.
• Limits at infinity depend on end behaviour that is determined by the order the function grows.

Review Questions:
1. What is a limit?
|𝑥|
2. Why does lim not exist?
𝑥→0 𝑥
3
−√𝑥+8
2−𝑥
3. Determine lim 𝑥−1
𝑥→1
4. Determine lim 3𝑥 ⋅ csc(𝑥 2 − 2𝑥)
𝑥→0
2
5. Prove lim 𝑥 = 1
𝑥→2
0
2𝑛𝑥 +𝑥 𝑚
6. Use Sandwich Theorem to determine the values of 𝑛 > 0 and 𝑚 so that lim = {1
𝑥→∞ 8𝑥−1
∞
0
2𝑛𝑥 +𝑥 𝑚
7. Use Sandwich Theorem to determine the values of 𝑛 > 0 and 𝑚 so that lim 8𝑥−1 = { 1
𝑥→−∞
∞
Solutions:
1. Weak: A value (real number), 𝐿, that we can make arbitrarily close to the value of the function if we pick
values of 𝑥 close enough to the limit point.
Strong. The limit of 𝑓(𝑥) as 𝑥 approaches 𝑐 is 𝐿 if and only if ∀ 𝜀 > 0, ∃ 𝛿 > 0 such that if |𝑥 − 𝑐| < 𝛿 then
|𝑓(𝑥) − 𝐿| < 𝜀.
|𝑥| 𝑥 |𝑥| −𝑥
2. Because lim+ = lim+ = 1 and lim− = lim− = − 1 which are not the same.
𝑥→0 𝑥 𝑥→0 𝑥 𝑥→0 𝑥 𝑥→0 𝑥
17
3. 6
Unit 1: Limits and Continuity Review

4. −1.5
2
5. We have that |𝑥 − 2| < 𝛿 and we want | − 1| < 𝜀 for any 𝜀 > 0. Start with the function and its limit:
𝑥
2 |2−𝑥| |𝑥−2| 𝛿 1
|𝑥− 1| = = < as long as 𝛿 < 2 (since this guarantees that 𝑥 > 0) We now need to bound
𝑥 𝑥 𝑥 𝑥
which effectively minimizes 𝑥. Since we currently have that 𝑥 ∈ (0,4) there is no good lower bound so we can
1 2 𝛿
make 𝛿 < 1 so that 𝑥 ∈ (1,3). Now < 1 and we have that | − 1| < < 𝛿 so let 𝛿 = min(𝜀, 1) ∎
𝑥 𝑥 𝑥
6. First, 𝑚 can be any real number, it won’t matter since it is next to an exponential and the numerator grows
𝒪((2𝑛 )𝑥 ). For the limit to be 0, we need 2𝑛 < 8, hence 𝑛 < 3.
𝑥
2𝑛𝑥 + 𝑥 𝑚 2𝑛𝑥 + 𝑥 𝑚 2𝑛𝑥 𝑥𝑚 2𝑛 𝑥𝑚
lim = 8 ⋅ lim = 8 ⋅ lim + 8 ⋅ lim = 8 ⋅ lim ( ) + 8 ⋅ lim
𝑥→∞ 8𝑥−1 𝑥→∞ 8𝑥 𝑥→∞ 8 𝑥 𝑥→∞ 8 𝑥 𝑥→∞ 8 𝑥→∞ 8 𝑥
And the last piece will go to 0 via Sandwich Theorem.
𝑥𝑚 2𝑥
0 < 8 ⋅ lim 𝑥 < 8 ⋅ lim 𝑥 = 0
𝑥→∞ 8 𝑥→∞ 8

Likewise, for it to blow up we need 2𝑛 > 8, so 𝑛 > 3.

𝑥
2𝑛𝑥 + 𝑥 𝑚 2𝑛𝑥 + 𝑥 𝑚 2𝑛𝑥 2𝑛
lim = 8 ⋅ lim > 8 ⋅ lim 𝑥 = 8 ⋅ lim ( ) = ∞
𝑥→∞ 8𝑥−1 𝑥→∞ 8𝑥 𝑥→∞ 8 𝑥→∞ 8

We can never make the limit 1 as if 𝑛 = 3 the limit is 8.

7. Same as above but now the denominator is going to 0 and the top grows 𝒪(𝑥 𝑚 ) so 𝑛 can be any positive
number. As long as 𝑚 ≥ 0 the numerator will grow larger (or at least be 1) while the denominator will
approach 0+ and the overall limit will go to infinity.
2𝑛𝑥 + 𝑥 𝑚 𝑥𝑚 1
lim 𝑥−1
> lim 𝑥−1
≥ lim 𝑥−1 = ∞
𝑥→−∞ 8 𝑥→−∞ 8 𝑥→−∞ 8

If 𝑚 < 0 then we can change the limit to

2𝑛𝑥 + 𝑥 𝑚 𝑥𝑚 8𝑥+1 8𝑥+1
lim > lim = lim > lim =∞
𝑥→−∞ 8𝑥−1 𝑥→−∞ 8 𝑥−1 𝑥→∞ 𝑥 |𝑚| 𝑥→∞ 2 𝑥

Continuity

We used the definition of continuity to be the following:

𝑓 is continuous at 𝑥 = 𝑐 if and only if lim 𝑓(𝑥) = 𝑓(𝑐)

𝑥→𝑐

This also requires that 𝑓(𝑐) is well defined and that lim 𝑓(𝑥) exists, and this gives us a way to describe familiar
𝑥→𝑐
discontinuities. If there is an asymptote or a jump, then the limit does not exist, and if there is a hole (removeable
discontinuity) then the function is not defined or does not equal the limit.

Once we have continuity, we have a fantastic result called Intermediate Value Theorem which states.

If 𝑓 is continuous on [𝑎, 𝑏] then 𝑓(𝑥) takes on every value between 𝑓(𝑎) and 𝑓(𝑏).

This result should be intuitive. If we have a connected path from 𝑓(𝑎) to 𝑓(𝑏) then the path must pass through each
value between the endpoints.
Unit 1: Limits and Continuity Review

Things we need to know and understand:

• How to determine continuity at a point
• Consequences of Intermediate Value Theorem
Review Questions:
𝑥−2
, 𝑥<2
√𝑥+2−2
8. Show that 𝑓(𝑥) = 4 , 𝑥 = 2 is continuous at 𝑥 = 2.
4𝑥−8
{ sin(𝑥 2 −3𝑥+2)
, 𝑥>2
9. Construct a function that is continuous on (𝑎, 𝑏) but does not satisfy the Intermediate Value Theorem.
10. Show that tan 𝑥 = ln 𝑥 2 has a solution.

Solutions:
𝑥−2 (𝑥−2)(√𝑥+2+2) 4𝑥−8
8. Both the one-sided limits and 𝑓(2) = 4. lim 𝑥+2−2
= lim 𝑥−2
= 4 and lim sin(𝑥 2 −3𝑥+2) =
𝑥→2 √ 𝑥→2 𝑥→2
4(𝑥−2) (𝑥−2)(𝑥−1)
lim = 4 ⋅ lim (𝑥−1)⋅sin((𝑥−2)(𝑥−1)) = 4
𝑥→2 sin((𝑥−2)(𝑥−1)) 𝑥→2
9. Consider 𝑓(𝑥) = 𝑓𝑙𝑜𝑜𝑟(𝑥) which is continuous on (0,1) but 𝑓(1) = 1 and 𝑓(0) = 0 clearly there is no value
𝑐 ∈ (0,1) such that 𝑓(𝑐) ∈ (0,1) since 𝑓(𝑥) = 0 for every 𝑥 ∈ (0,1). [Note: 𝑓𝑙𝑜𝑜𝑟(𝑥) rounds down to the
nearest integer.]
10. We want to show that 𝑓(𝑥) = tan 𝑥 − ln 𝑥 2 has a zero. Try some values of 𝑥 to find a positive value. Upon
checking 𝑓(1) = 1.6 and 𝑓(2) = −3.6. Since 𝑓 is continuous on [1,2] we have that there must be a zero by
IVT.

Differentiability

A function is said to be differentiable if the slope (the derivative) exists at 𝑥 = 𝑐. That means that the following limit
exists:
𝑓(𝑐 + ℎ) − 𝑓(𝑐) 𝑓(𝑥) − 𝑓(𝑐)
lim = lim
ℎ→0 ℎ 𝑥→𝑐 𝑥−𝑐

This is the definition of the slope at 𝑥 = 𝑐 and so you can intuit from the graph or data what the value of this limit
should be by imagining a tangent line and when it doesn’t exist. Instances when it doesn’t exist is when there is no well-
defined tangent line, or the slope of the tangent line is undefined. See our notes for complete list of cases where
differentiability breaks down.

To define the derivative at any point 𝑥 we need to change our label of 𝑐 = 𝑥

𝑓(𝑥 + ℎ) − 𝑓(𝑥) 𝑓(𝑐) − 𝑓(𝑥)

lim = lim
ℎ→0 ℎ 𝑐→𝑥 𝑐−𝑥
𝑑 𝑑𝑓 𝑑
And we denote this as 𝑓′(𝑥) or 𝑑𝑥 𝑓(𝑥) or 𝑑𝑥. Since this is slope, we are measuring a rate of change and the 𝑑𝑥 notation
reflects this as it means the change of some function with respect to how much 𝑥 changes. That is for a small change in
𝑥 how much greater or less will the function change. We can graph this on our calculators using nDeriv(𝑓(𝑥), 𝑥, 𝑥).

𝑑
|
If we want to know the derivative at point, we will denote it as 𝑓′(𝑐) or 𝑓(𝑥) 𝑥=𝑐 and if we are using our calculator
𝑑𝑥
can compute it using nDeriv(𝑓(𝑥), 𝑥, 𝑐). We just need to be careful when using our calculator to test at a point that may
not be differentiable. Our calculator is not actually using a limit but an approximation to the limit and will likely give
incorrect results as a result. For example, nDeriv(|𝑥|, 𝑥, 0) = 0 when it actually is not defined since the left and right-
hand limits of the derivative at 𝑥 = 0 are not the same.
Unit 1: Limits and Continuity Review

In these cases where the function may not be differentiable you can check by zooming in and observing if the function is
approaching a linear function. This is called local linearization and it is a property of differentiability.

1
𝑥 ⋅ sin 𝑥 , 𝑥 ≠ 0
For example, the function 𝑓(𝑥) = { is not locally linear at 𝑥 = 0 and therefore not differentiable.
0, 𝑥 = 0

Lastly, if 𝑓 is differentiable then it is also continuous. The proof is an example below, but you should be able to intuit
that if lim 𝑓(𝑥) ≠ 𝑓(𝑐) then either there is a jump or asymptote which cannot have a real valued tangent slope, or there
𝑥→𝑐
is a hole and we will need to use the value of 𝑓(𝑐) which is different than the neighbourhood around 𝑐.

Things we need to know and understand:

• How to find the slope at a point using limits and a calculator
• How to find the graph of 𝑓′ given the graph or data of 𝑓, and can graph the derivative on a calculator using
the original
• How to determine differentiability using limits and the shape of the function
• That differentiable functions are locally linear and continuous
Review Questions:
𝑥
11. Determine 𝑦′ if 𝑦 = 2 and state where it is differentiable. Compute nDeriv where it is not differentiable.
√𝑥 +1
𝑑
12. Determine 𝑑𝑥 |𝑥 2 − 1| and state where it is differentiable. Compute nDeriv where it is not differentiable.
13. Given the graph of 𝑓 graph 𝑓′
Unit 1: Limits and Continuity Review

14. Given the table of 𝑓 graph 𝑓′ . Assume 𝑓 is continuous

𝑥 2 4 8 11 12 15 20 21 25
𝑓(𝑥) −2 6 10 1 −5 1 6 9 11

15. Prove that differentiability implies continuity.

Solutions:
𝑐 𝑥
−
′ √𝑐2 +1 √𝑥2 +1 𝑐√𝑥 2 +1−𝑥√𝑐 2 +1 𝑐√𝑥 2 +1+𝑥√𝑐 2 +1 𝑐 2 (𝑥 2 +1)−𝑥 2 (𝑐 2 +1)
11. 𝑦 = lim = lim ⋅ = lim
𝑐→𝑥 𝑐−𝑥 𝑐→𝑥 (𝑐−𝑥)√𝑐 2 +1⋅√𝑥 2 +1 𝑐√𝑥 2 +1+𝑥√𝑐 2 +1 𝑐→𝑥 (𝑐−𝑥)√𝑐 2 +1⋅√𝑥 2 +1(𝑐√𝑥 2 +1+𝑥√𝑐 2 +1)
3
𝑐 2 −𝑥 2 𝑐+𝑥 2𝑥
′
𝑦 = lim = lim = = (𝑥 2 + 1)−2
𝑐→𝑥 (𝑐−𝑥)√𝑐 2 +1⋅√𝑥 2 +1(𝑐√𝑥 2 +1+𝑥√𝑐 2 +1) 𝑐→𝑥 √𝑐 2 +1⋅√𝑥 2 +1(𝑐√𝑥 2 +1+𝑥√𝑐 2 +1) (𝑥 2 +1)(2𝑥√𝑥 2 +1)
Since 𝑥 2 + 1 > 0 always, 𝑦′ does not have any discontinuity and therefore 𝑦′ exists ∀ 𝑥 and is differentiable
everywhere.
𝑥 2 − 1, 𝑥 ≤ −1 𝑜𝑟 𝑥 ≥ 1
12. |𝑥 2 − 1| = { so, we need to consider two cases.
1 − 𝑥2 , −1<𝑥 <1
Case 1: |𝑥| ≥ 1 then
𝑑 2 (𝑥 + ℎ)2 − 1 − (𝑥 2 − 1) 2𝑥ℎ + ℎ2
|𝑥 − 1| = lim = lim = 2𝑥
𝑑𝑥 ℎ→0 ℎ ℎ→0 ℎ
Case 2: |𝑥| < 1 then
𝑑 2 1 − (𝑥 + ℎ)2 − (1 − 𝑥 2 ) −2𝑥ℎ − ℎ2
|𝑥 − 1| = lim = lim = −2𝑥
𝑑𝑥 ℎ→0 ℎ ℎ→0 ℎ
Hence, if we wanted to evaluate the derivative at ±1 we would get ±2 depending the side of ±1 we
approach, and the limit definition would not exist. So |𝑥 2 − 1| is differentiable ∀ 𝑥 ∈ ℝ, 𝑥 ≠ ±1
nDeriv(|𝑥 2 − 1|, 𝑥, −1) = −0.001 and nDeriv(|𝑥 2 − 1|, 𝑥, 1) = 0.001
13.
Unit 1: Limits and Continuity Review

14.

The dotted horizontal lines assume piecewise linear, the solid line is connecting the midpoints of each region
with linear equations.
15. The textbook gives a proof that shows if 𝑓 is differentiable then 𝑓 is continuous. My style is to show that if 𝑓 is
not continuous then it is not differentiable (stop to think why this is an equivalent statement).

𝑓(𝑥)−𝑓(𝑐)
So, if 𝑓 is not continuous at 𝑥 = 𝑐 then maybe 𝑓(𝑐) is not defined. In that case lim 𝑥−𝑐
cannot be
𝑥→𝑐
evaluated since it requires that 𝑓(𝑐) be something. If we have that 𝑓(𝑐) exists, then either lim 𝑓(𝑥) does not
𝑥→𝑐
exist or lim 𝑓(𝑥) ≠ 𝑓(𝑐). In the both cases, we are going to break the definition of a limit.
𝑥→𝑐

𝑓(𝑥)−𝑓(𝑐)
We want that ∀ 𝜀 > 0, ∃ 𝛿 > 0 such that if |𝑥 − 𝑐| < 𝛿 then | | < 𝜀. In the first case where the limit of
𝑥−𝑐
𝑓(𝑥)−𝑓(𝑐)
𝑓 does not exist, we have no hope for | 𝑥−𝑐 | < 𝜀 since this requires us to bound 𝑓(𝑥) around 𝑥 = 𝑐.
𝑓(𝑥) − 𝑓(𝑐) |𝑓(𝑥) − 𝑓(𝑐)|
| |>
𝑥−𝑐 𝛿
And we have no way to bound 𝑓 around anything. If instead, the limit of 𝑓 exists but is not 𝑓(𝑐) then we let
Δ = |𝑓(𝑥) − 𝑓(𝑐)| > 0
Our inequality will become
𝑓(𝑥) − 𝑓(𝑐) |𝑓(𝑥) − 𝑓(𝑐)| Δ
| |> >
𝑥−𝑐 𝛿 𝛿
Which is fixed and making 𝛿 smaller only makes the lower bound greater so this will never be smaller than an
arbitrarily small 𝜀 ∎

You could use a similar argument starting with differentiability and building |𝑓(𝑥) − 𝑓(𝑐)| < 𝜀 for 𝛿 = 𝜀1 𝛿1
𝑓(𝑥)−𝑓(𝑐)
where |𝑥 − 𝑐| < 𝛿1 ⇒ | 𝑥−𝑐
| < 𝜀1 in a much more elegant argument, but I am partial to arguments that
use negations.
Unit 1: Limits and Continuity Review