Solutions to More Delta Epsilon Questions

1) What is the largest tolerable error in the length of the edge of a cube if the volume of
the cube must be within ±1.2% of 8, 000 cm3 .
Solution. Let the length of the edge of the cube be x. The volume of the cube must be
1.2
within 100 8000 = 96 of 8000. Hence
√ √
7904 = 8000 − 96 ≤ x3 ≤ 8000 + 96 = 8096 ⇐⇒
3 3
7904 ≤ x ≤ 8096

To have x3 = 8000, we need x = 20. The maximum allowed error in x is the smaller of
√
3
√ √
8096 − 20 and 20 − 3 7904, which is 3 8096 − 20 ≈ 0.0797 .

2) Use the formal definition of the limit to verify that

lim x−22 =0
x→2 1+x

Solution. We must prove that

x−2
∀ ε > 0 ∃ δ > 0 such that 1+x2 < ε whenever 0 < |x − 2| < δ

Let ε > 0.
Choose δ =“to be filled in later”.
Then, whenever 0 < |x − 2| < δ, we have
x−2
1+x2 ≤ |x − 2| since 1 + x2 ≥ 1
<δ
≤ε if δ ≤ ε
x−2
We have now shown that, whenever 0 < |x − 2| < δ, we have 1+x2
< ε, provided δ
obeys δ ≤ ε. So we may choose δ = ε. That is, we may choose “to be filled in later”
above to be ε.

3) Use the formal definition of the limit to verify that

Solution. We must prove that

1
 Let ε > 0.
Choose δ =“to be filled in later”.
Then, whenever 0 < |x − 3| < δ, we have
√ 2
2x + 3 − 3 = (2x+3)−3
√
2x+3+3
= √ 2x−6
2x+3+3
2(x−3)
√
≤ 3 since 2x + 3 ≥ 0
< 32 δ
≤ε if δ ≤ 23 ε
√
We have now shown that, whenever 0 < |x − 3| < δ, we have 2x + 3 − 3 < ε, provided
3 3
δ obeys δ ≤ 2 ε. So we may choose δ = 2 ε. That is, we may choose “to be filled in later”
3
above to be 2
ε.

4) Prove that lim sin x1 does not exist. Do so by directly verifying that the formal definition
x→0
1
of lim sin x
=L cannot be satisfied for any L.
x→0
Solution. Let L ∈ IR. Assume that lim sin x1 = L. We will derive a contradiction. To
x→0
do so, we must show that the statement

∀ ε > 0 ∃ δ > 0 such that sin x1 − L < ε whenever 0 < |x| < δ

is false.
To do so, it suffices for us to exhibit one “bad” value of ε. Choose ε = 10−6 . (This is
just a guess, which happens to work out. Lots of other guesses would work too.) Now
that we have chosen a value of ε, we must show that the statement

∃ δ > 0 such that sin x1 − L < 10−6 whenever 0 < |x| < δ

is false. To do this, we must show that, for every δ > 0, the statement

sin x1 − L < 10−6 whenever 0 < |x| < δ (∗)

is false.
So let δ > 0. If the statement (∗) were true, this would imply that sin x1 takes values only
between L − 10−6 and L + 10−6 for all 0 < |x| < δ. Note that L − 10−6 ≤ y ≤ L + 10−6 is
an interval of length 2 × 10−6 . As x runs over the interval (0, δ), (so that, in particular,
0 < |x| < δ) x1 covers the set δ1 , ∞ . This contains many intervals of length 2π and

hence many periods of sin. So, as x runs over the interval (0, δ), sin x1 covers all of [−1, 1].
1
But [−1, 1] is an interval of length 2 and so cannot be contained in L − 2 ≤ y ≤ L + 21 .
So (∗) is false.

2
5) Prove that limits are unique. That is, prove that if lim f (x) = L and lim f (x) = M ,
x→a x→a
then L = M .
Solution. We are told that,

∀ ε > 0 ∃ δ1 > 0 such that f (x) − L < ε whenever 0 < |x − a| < δ1

and

∀ ε > 0 ∃ δ2 > 0 such that f (x) − M < ε whenever 0 < |x − a| < δ2

and we are to conclude that L = M . Consider any ε > 0. There is a δ1 > 0 such that
f (x) − L < ε for all 0 < |x − a| < δ1 and a δ2 > 0 such that 0 < f (x) − M < ε
for all |x − a| < δ2 . (These two δ’s are allowed to be different. That’s why I have
chosen two different names, δ1 and δ2 .) Pick any x0 obeying both 0 < |x0 − a| < δ1 and
0 < |x0 − a| < δ2 . Then

|L − M | = L − f (x0 ) + f (x0 ) − M ≤ L − f (x0 ) + f (x0 ) − M < 2ε.

We have just shown that |L − M | < 2ε for every ε > 0. This forces L = M . (If L 6= M ,
we may choose ε = |L−M4
|
and conclude that |L − M | < 2 |L−M
4
|
, which would imply that
|L−M |
2
< 0, which is impossible.)

6) Suppose that a < b < c, that f (x) < g(x) < h(x) for all a < x < c, and that lim f (x) =
x→b
lim h(x) = L. Prove that lim g(x) = L. (This is called the Squeeze Theorem.)
x→b x→b
Solution. We must prove that

∀ ε > 0 ∃ δ > 0 such that g(x) − L < ε whenever 0 < |x − b| < δ

Let ε > 0. Note that we are told that there is a δ1 > 0 such that f (x) − L < ε for all
0 < |x − b| < δ1 and a δ2 > 0 such that h(x) − L < ε for all 0 < |x − b| < δ2 .
Choose δ =“to be filled in later”.
Then, whenever 0 < |x − b| < δ, we have
(1) (2) (2) (3)
L − ε < f (x) < g(x) < h(x) < L + ε
=⇒ L − ε < g(x) < L + ε
=⇒ g(x) − L < ε
provided δ ≤ δ1 (for the inequality (1)) and δ ≤ c − b, b − a (so that a < x < c for the
inequalities (2)) and δ ≤ δ2 (for the inequality (3)). We have now shown that, whenever
0 < |x − b| < δ, we have g(x) − L < ε, provided δ obeys δ ≤ δ1 and δ ≤ δ2 and δ ≤ c − b
and δ ≤ b − a. So we may choose δ = min{δ1 , δ2 , b − a, c − b}. That is, we may choose
“to be filled in later” above to be min{δ1 , δ2 , b − a, c − b}.