Ma · Mathematics

How to Use an Epsilon–Delta Proof to Prove Continuity

Published By the 9to5Science editorial team

At a glance

  1. 1. State the definition

    Write the epsilon–delta condition for the function at the chosen point.

  2. 2. Estimate the difference

    Factor, simplify, or bound the expression involving f(x) minus f(a).

  3. 3. Control x minus a

    Choose a delta that makes the bound smaller than epsilon.

  4. 4. Check the choice

    Verify the delta is positive and works for every allowed x.

Continuity at a point means that changing the input by a sufficiently small amount changes the output by a sufficiently small amount. The epsilon–delta proof makes the phrases “sufficiently small” and “sufficiently close” precise.

The usual strategy is to fix an arbitrary positive epsilon, find a positive delta depending on epsilon, and prove the required implication. The choice of delta may depend on epsilon and on the point a, but it must not depend on x.

The epsilon–delta definition

A function ff is continuous at aa if, for every ε>0\varepsilon > 0, there exists a δ>0\delta > 0 such that, for every xx in the domain of ff,

∣x−a∣<δ⟹∣f(x)−f(a)∣<ε.|x-a|<\delta \quad\Longrightarrow\quad |f(x)-f(a)|<\varepsilon.

This definition assumes that aa is in the domain of ff. A function is continuous on a set if it is continuous at every point of that set.

The expression ∣x−a∣<δ|x-a|<\delta says that xx is within δ\delta of aa. The conclusion ∣f(x)−f(a)∣<ε|f(x)-f(a)|<\varepsilon says that the corresponding function values are within ε\varepsilon of f(a)f(a).

A general proof template

To prove continuity at a particular point aa, use the following structure:

  1. Let ε>0\varepsilon>0 be arbitrary.
  2. Compute or estimate ∣f(x)−f(a)∣|f(x)-f(a)|.
  3. Find a bound involving ∣x−a∣|x-a|.
  4. Choose δ\delta so that the bound is less than ε\varepsilon.
  5. Assume ∣x−a∣<δ|x-a|<\delta and verify that ∣f(x)−f(a)∣<ε|f(x)-f(a)|<\varepsilon.
  6. State that this proves continuity at aa.

The central calculation is usually step 2. You are working backward: determine what bound on ∣x−a∣|x-a| would make the output difference smaller than ε\varepsilon.

Worked example: proving that f(x)=x2f(x)=x^2 is continuous at 3

Let f(x)=x2f(x)=x^2. We want to prove continuity at a=3a=3.

Let ε>0\varepsilon>0 be given. We need to control

∣f(x)−f(3)∣=∣x2−9∣.|f(x)-f(3)|=|x^2-9|.

Factor the difference of squares:

∣x2−9∣=∣x−3∣∣x+3∣.|x^2-9|=|x-3||x+3|.

The first factor is exactly the input distance ∣x−3∣|x-3|. The second factor, ∣x+3∣|x+3|, is not automatically bounded by a fixed number just from knowing that xx is near 3. We create a convenient local bound by requiring ∣x−3∣<1|x-3|<1.

If ∣x−3∣<1|x-3|<1, then

∣x∣=∣(x−3)+3∣≤∣x−3∣+3<4,|x|=|(x-3)+3|\leq |x-3|+3<4,

so

∣x+3∣≤∣x∣+3<7.|x+3|\leq |x|+3<7.

Therefore, whenever ∣x−3∣<1|x-3|<1,

∣x2−9∣=∣x−3∣∣x+3∣≤7∣x−3∣.|x^2-9|=|x-3||x+3|\leq 7|x-3|.

To make this less than ε\varepsilon, it is enough to require

7∣x−3∣<ε,7|x-3|<\varepsilon,

which is guaranteed by ∣x−3∣<ε/7|x-3|<\varepsilon/7.

Choose

δ=min⁡(1,ε7).\delta=\min\left(1,\frac{\varepsilon}{7}\right).

This is positive because ε>0\varepsilon>0. Now suppose that ∣x−3∣<δ|x-3|<\delta. Since δ≤1\delta\leq 1, we have ∣x−3∣<1|x-3|<1, and hence ∣x+3∣<7|x+3|<7. Also, δ≤ε/7\delta\leq \varepsilon/7, so ∣x−3∣<ε/7|x-3|<\varepsilon/7. Thus

∣f(x)−f(3)∣=∣x2−9∣=∣x−3∣∣x+3∣≤7∣x−3∣<7(ε7)=ε.|f(x)-f(3)|=|x^2-9| =|x-3||x+3| \leq 7|x-3| <7\left(\frac{\varepsilon}{7}\right) =\varepsilon.

Therefore, for every ε>0\varepsilon>0, the choice δ=min⁡(1,ε/7)\delta=\min(1,\varepsilon/7) works. Hence f(x)=x2f(x)=x^2 is continuous at 33.

Why the minimum appears

The number ε/7\varepsilon/7 controls the final estimate, while 11 supplies the local bound on ∣x+3∣|x+3|. Both conditions are needed:

∣x−3∣<1and∣x−3∣<ε7.|x-3|<1 \quad\text{and}\quad |x-3|<\frac{\varepsilon}{7}.

A single choice satisfying both is

∣x−3∣<min⁡(1,ε7).|x-3|<\min\left(1,\frac{\varepsilon}{7}\right).

More generally, if a proof needs ∣x−a∣<r|x-a|<r for one part and ∣x−a∣<s|x-a|<s for another, choose δ=min⁡(r,s)\delta=\min(r,s).

A simpler example: an affine function

For f(x)=5x−2f(x)=5x-2, continuity at any aa follows directly:

∣f(x)−f(a)∣=∣(5x−2)−(5a−2)∣=5∣x−a∣.|f(x)-f(a)| =|(5x-2)-(5a-2)| =5|x-a|.

Given ε>0\varepsilon>0, choose δ=ε/5\delta=\varepsilon/5. Then

∣x−a∣<δ⟹∣f(x)−f(a)∣=5∣x−a∣<5ε5=ε.|x-a|<\delta \quad\Longrightarrow\quad |f(x)-f(a)|=5|x-a|<5\frac{\varepsilon}{5}=\varepsilon.

Thus every affine function is continuous everywhere.

Common mistakes

Choosing delta without checking it

Writing “take δ=ε/7\delta=\varepsilon/7” is not enough when the proof also requires a bound such as ∣x−3∣<1|x-3|<1. Use a minimum or otherwise impose every needed restriction.

Letting delta depend on x

The number δ\delta may depend on ε\varepsilon and aa, but it must be selected before xx is chosen. A statement such as “choose δ=∣x−a∣\delta=|x-a|” does not prove continuity.

Proving only a limit statement

Showing

lim⁡x→af(x)=f(a)\lim_{x\to a}f(x)=f(a)

is equivalent to continuity at aa when aa is a limit point of the domain, but an epsilon–delta proof must explicitly establish the required implication. Do not merely say that the function is familiar or that its graph looks smooth.

Forgetting the function’s domain

For a function such as f(x)=1/xf(x)=1/x, continuity can be proved at every a≠0a\ne0, but there is no value f(0)f(0) in the usual real-valued definition. Therefore continuity at 00 is not defined for this function, whose domain is R∖{0}\mathbb{R}\setminus\{0\}.

Using an unnecessarily exact delta

Many different choices work. A smaller positive delta is also valid. The goal is not to find the largest possible delta, but one that makes the implication true.

A useful shortcut after the definition

Once the epsilon–delta proofs have established the basic continuous functions, continuity rules can reduce later work. Sums, differences, products, and compositions of continuous functions are continuous wherever they are defined. Quotients are continuous where the denominator is nonzero. These rules explain why polynomials are continuous everywhere and rational functions are continuous at every point where their denominators do not vanish.

However, when a problem specifically asks for an epsilon–delta proof, write the estimate explicitly rather than citing only a continuity rule.

References

  1. Understanding Analysis, Stephen Abbott
  2. Principles of Mathematical Analysis, Walter Rudin

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.