Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

Here is a statement of Egoroff's Theorem:

Egoroff's Theorem. Suppose that $\mu(X) < \infty$, and $f, f_1, f_2, f_3, \ldots$ are all measurable complex-valued functions on $X$ such that $f_n \to f$ almost everywhere. Then for every $\varepsilon > 0$, there exists $E \subseteq X$ such that $\mu(E) < \varepsilon$ and $f_n \to f$ uniformly on $E^c$ (the complement of $E$ in $X$).

Is it possible to replace the "$\mu(E) < \varepsilon$" condition by "$\mu(E) = 0$" in the statement and still have the theorem be true? I suspect that the answer is no, but I am having trouble thinking of an appropriate counterexample.

share|cite|improve this question
Patrick Da Silva has already given an example, but if you want a more extended discussion about Egoroff's Theorem, see the essay I posted in this thread: (moments later) Oops, this essay is mostly on Luzin's Theorem. Nonetheless, it should still be fairly relevant. – Dave L. Renfro Feb 16 '12 at 15:00
Many times the condition $\mu(E)\lt \epsilon$ is all that we need. See here for a example. – leo Feb 16 '12 at 15:06
up vote 6 down vote accepted

Let $X = [0, \pi/2)$ and $f_n(x) = \frac 1n \tan(x)$. Clearly $f_n \to 0 = f(x)$. Suppose $\mu(E) = 0$ with $E \subseteq X$. Then since for every $\delta > 0$, $\mu( (\pi/2 - \delta, \pi/2) ) = \delta > 0$, then $\mu( (\pi/2 - \delta, \pi/2) \cap E^c ) = \delta > 0$ (assuming $\delta < \pi/2$ obviously) and thus you have a sequence of points $x_m$ such that $x_m \in E^c$ and $x_m \to \pi/2$. But $f_n(x_m) \to \infty$ as $m \to \infty$, so that $\sup_{X \cap E^c} |f_n(x) - f(x)|$ can never even be finite, so you can't possibly hope it goes to zero as $n \to \infty$. Therefore convergence is not uniform.

Hope that helps,

share|cite|improve this answer
Actually I didn't need $\tan$ : I just needed a function that went to infinity at some point and then multiply it by $1/n$ and take a neighborhood close enough to it. I could've taken $g(x) = 1/x$, $X = (0,1)$ and $f_n(x) = (1/n) g(x)$, a similar argument would've worked too. – Patrick Da Silva Feb 16 '12 at 4:38

Here is a simpler example $f_{n}:(0,1)\rightarrow \mathbb{R}$ such that $f_{n}(x)=x^n$, $f_{n}(x)\rightarrow 0$ but $f_{n}$ cannot converges uniformly outside any set of measure zero.

share|cite|improve this answer
You replace my issue at infinity by an issue of discontinuity... it's a matter of choice whether it is simpler to have an infinity-kind of discontinuity or a gap. But still you gave the other point of view, so I'm happy to +1 :) (A bit late though, but I admit I didn't read it back then I guess) – Patrick Da Silva Jun 10 '13 at 3:08

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.