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

Lusin Theorem (as stated by Rudin):

Let $X$ be a locally compact Hausdorff space and let $μ$ be a regular Borel measure on $X$ such that $μ(K)<∞$ for every compact $K⊆X$. Suppose $f$ is a complex measurable function on $X$, $μ(A)<∞$, $f(x)=0$ if $x∈X \setminus A$, and $ϵ>0$. Then there exists a continuous complex function $g$ on $X$ with compact support such that


But I can't seem to find in the proof anywhere a use of the fact that the measure is finite for compact sets. Is the condition nessecary? Is there a counter-example?

share|cite|improve this question

It can be shown that any $\sigma$-finite regular measure on an LCH space must have $\mu(K) < \infty$ for $K$ compact. (Exercise: Prove it.) So we will have to use a measure which is not $\sigma$-finite. Modifying dissonance's example, let $X = \mathbb{R}$ and $\mu(A) = \infty$ for all nonempty $A$, which I think is a regular measure (rather trivially so), and take $f$ to be any discontinuous function.

Rudin's proof of Lusin contains the following line:

Fix an open set $V$ such that $A \subset V$ and $\bar{V}$ is compact. There are compact sets $K_n$ and open sets $V_n$ such that $K_n \subset T_n \subset V_n \subset V$ and $\mu(V_n - K_n) < 2^{-n} \epsilon$.

This invokes Theorem 2.17 (a) (paraphrased):

For any measurable set $E$ and $\epsilon > 0$, there exists $F$ closed and $V$ open with $F \subset E \subset V$ and $\mu(V-F) < \epsilon$.

The proof of 2.17 uses the assumption that $\mu$ is finite on compact sets in the second line, when it asserts that $\mu(K_n \cap E) < \infty$.

Also, it's easy to see that all of Rudin's quoted claims fail using the example I gave.

share|cite|improve this answer
But the conditions of the theorem state that the function has a finite-measure support, which your example does not have. Also, it is easy to prove the quoted line for finite measure sets - which is the case here. – yaakov Mar 30 '11 at 21:42
:-) Thank you for the credit! I don't think I quite deserved that, though. (lol) – Giuseppe Negro Mar 30 '11 at 21:43
@yaakov: Oh, good point. So the condition is redundant. – Nate Eldredge Mar 30 '11 at 23:36

proof of Lusin's Theorem: Let f be a measurable real-valued function on [a; b] and let  > 0 be given. For each n, there is a continuous function hn on [a; b] such that mfx : jhn(x) 􀀀 f(x)j  =2n+2g < =2n+2. Let En = fx : jhn(x)􀀀f(x)j  =2n+2g. Then jhn(x)􀀀f(x)j < =2n+2 for x 2 [a; b]nEn. Let E = S En. Then mE < =4 and hhni is a sequence of continuous, thus measurable, functions that converges to f on [a; b] n E. By Egoro 's Theorem, there is a subset A [a; b] n E such that mA < =4 and hn converges uniformly to f on [a; b]n(E[A). Thus f is continuous on [a; b]n(E[A) with m(E[A) < =2. Now there is an open set O such that O  (E [ A) and m(O n (E [ A)) < =2. Then f is continuous on [a; b] n O, which is closed, and mO < . By Q2.40, there is a function ' that is continuous on (􀀀1;1) such that f = ' on [a; b] n O. In particular, ' is continuous on [a; b] and mfx 2 [a; b] : f(x) 6= '(x)g = mO < . If f is de ned on (􀀀1;1), let 0 = min(=2n+3; 1=2). Then for each n, there is a continuous function 'n on [n + 0; n + 1 􀀀 0] such that mfx 2 [n + 0; n + 1 􀀀 0] : f(x) 6= 'n(x)g < =2n+2. Similarly for [􀀀n􀀀1+0;􀀀n􀀀0]. Linearise in each interval [n􀀀0; n+0]. Similarly for intervals [􀀀n􀀀0;􀀀n+0]. Then we have a continuous function ' de ned on (􀀀1;1) with mfx : f(x) 6= '(x)g < 4 P =2n+2 = .

share|cite|improve this answer
formatting of your answer is awful – Norbert Jul 19 '13 at 8:31
Please improve the formatting of your answer (see the editing help). – Mårten W Jul 19 '13 at 8:34
Flagged this post since it is unreadable. – AD. Jul 19 '13 at 11:21
Please check your encoding and learn some LaTeX. – AD. Jul 19 '13 at 11:22

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.