Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Can you tell me if the following proof is correct?


If $f$ is a continuous and compactly supported function from a metric space $X$ into $\mathbb{R}$ then $f$ is uniformly continuous.


The proof is in two parts. First we want to show that $f$ is uniformly continuous on $K := \operatorname{supp}{f}$:

Let $\varepsilon > 0$. Because $f$ is continuous we have that for each $x$ in $K$ there is a $\delta_x$ such that for all $y$ with $d(x,y) < 2 \delta_x$ we have $|f(x) - f(y)| < \varepsilon$ and because $\{ B(x, \frac{\delta_x}{2}) \}_{x \in K}$ is an open cover of $K$ there is a finite subcover which we denote $\{ B(x_i, \frac{\delta_i}{2}) \}_{i=1}^n$. Define $\delta := \min_i \frac{\delta_i}{2}$ and let $x$ and $y$ be any two points in $K$ with $d(x,y) < \delta$. $\{ B(x_i, \frac{\delta_i}{2}) \}_{i=1}^n$ is a cover so there exists an $i$ such that $x$ is in $B(x_i, \frac{\delta_i}{2})$ which means that $d(x,x_i) < \frac{\delta_i}{2}$. Then $d(x_i ,y) \leq d(x_i ,x) + d(x,y) < \frac{\delta_i}{2} + \delta \leq \delta_i$ hence $y$ is also in $B(x_i, \delta_i)$.

Since $d(x_i,y) < \delta_i$ and $d(x, x_i) < \delta_i$ we have $|f(x) - f(y)| \leq |f(x) - f(x_i)| + |f(x_i) - f(y)| < 2 \varepsilon$.

Next we want to show that if $f$ is uniformly continuous on $K$ then it is uniformly continuous on all of $X$:

Let $\varepsilon > 0$. For any two points $x$ and $y$ we're done if either both are in $K$ or both are outside $K$ so let $x \in X \setminus K$ and $y \in K$ with $d(x,y) < \delta$. Then there is an $i$ such that $y$ is in $B(x_i, \frac{\delta_i}{2})$. Then $d(x,x_i) \leq d(x,y) + d(y,x_i) < \delta_i$ and hence $|f(x) - f(y)| \leq |f(x) - f(x_i)| + |f(x_i) - f(y)| < 2 \varepsilon$.

Is it necessary to prove this in two parts or is the second part "obvious" and should be left away?

Thanks for your help.

share|cite|improve this question
I'm more wondering whether the first part is unnecessarily detailed. Don't you have a canned theorem that a continuous function with compact domain is uniformly continuous which you can just apply to $f|_K$? – Henning Makholm Jan 2 '12 at 0:19
In the sentence "So $x,y \in B(x_i, \delta_i)$ which means that $d(x,y) < 2 \delta_i$ and therefore $|f(x) - f(y)| < \varepsilon$." where does this come from? You know that $|f(x) - f(y)| \lt |f(x)-f(x_i)| + |f(x_i)-f(y)| \lt 2\varepsilon$, which is good enough, but I fail to see how you get $\lt \varepsilon$. Is this a typo, or am I missing something? – t.b. Jan 2 '12 at 0:24
@t.b. That was a "thinko", not a typo. Yes, of course it's less than $2 \varepsilon$. Thank you! – Rudy the Reindeer Jan 2 '12 at 9:29
@Matt You have not quite corrected the "thinko" t.b. pointed out. You say that since $d(x,y) < \delta_i$ you know that $|f(x)-f(y)|$ is small by assumption. In fact you have not made any such assumption. The assumption you made is that $|f(x)-f(y)|$ is small if $d(x,y)<\delta_x$. – Johan Jan 2 '12 at 10:24
@Johan For a continuous function $f$ I have that for an $\varepsilon$ I can find a $\delta$ such that $f(B(x, \delta)) \subset B(f(x), \varepsilon)$. Above I have chosen $\delta_i$ such that $f(B(x_i, \delta_i)) \subset B(f(x_i), \varepsilon)$. I also have $x,y \in B(x_i,\delta_i)$ so $f(x), f(y) \in f(B(x_i,\delta_i)) \subset B(f(x_i), \varepsilon)$ and therefore $|f(x) - f(y)| < 2 \varepsilon$. Can you point out my thinko more explicitly please? I still don't see it, thank you. – Rudy the Reindeer Jan 2 '12 at 14:04
up vote 3 down vote accepted

That looks good except for the correction that t.b. pointed out. In the spirit of Henning Makholm's comment, here is a "canned theorem" approach.

A continuous function on a compact metric space is uniformly continuous, so $f|_K$ is uniformly continuous. Let $\varepsilon>0$ be given. Then $K_\varepsilon:=\{x:|f(x)|\geq \varepsilon\}$ is a closed subset of $K$, hence compact, and $\{x:f(x)=0\}$ is a closed set disjoint from $K_\varepsilon$, so there is a positive distance $\delta_1$ between $K_\varepsilon$ and $\{x:f(x)=0\}$. Let $\delta_2$ be such that if $x$ and $y$ are in $K$ and $d(x,y)<\delta_2$, then $|f(x)-f(y)|<\varepsilon$. Let $\delta=\min\{\delta_1,\delta_2\}$.

If $x$ and $y$ are in $X$ and $d(x,y)<\delta$, then:

  • $x$ and $y$ are both in $K$, and since $d(x,y)<\delta_2$, we have $|f(x)-f(y)|<\varepsilon$; or
  • one of $x$ or $y$ is not in $K$. WLOG suppose $x$ is not in $K$. Then $f(x)=0$, and since $d(x,y)<\delta_1$, $y$ is not in $K_\varepsilon$, meaning $|f(x)-f(y)|=|f(y)|<\varepsilon$. $\square$
share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.