MathOverflow is a question and answer site for professional mathematicians. 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

Given a Lebesgue measurable set A with strictly positive measure, can we find an open interval (a,b) such that x belongs to A for almost every x in (a,b)?

Thanks in advance for any comments!

share|cite|improve this question
the cantor set of positive measure is a counterexample. – rpotrie Sep 9 '10 at 15:12
There's also a measurable set $A\subset\mathbb{R}$ such that for all nonempty open interval $I$ one has $0 < m(A\cap I) < m(I).$ It's a well-known, old exercise. See Rudin's book. – Pietro Majer Sep 9 '10 at 15:38
See, for example, You may be interested in a weaker statement that what you ask for, see's_density_theorem – rpotrie Sep 9 '10 at 15:38
The question seems to have been answered. Voting to close as no longer relevant. – Victor Protsak Sep 9 '10 at 16:43
We don't generally close questions on MO just because they have been answered---after all, there are hundreds of non-closed questions here with excellent answers. Rather, I think what should happen is that rpotrie should answer the question with an answer about the fat Cantor set, rather than merely a comment (or someone else should if he doesn't). – Joel David Hamkins Sep 9 '10 at 20:19
up vote 3 down vote accepted

The usual Cantor set constructed by removing 1/3 at each step is nowhere dense but has measure 0. However, there exist nowhere dense sets which have positive measure. The trick is to try to remove less, for instance you remove 1/4 from each side of [0,1] during the first step then 1/16 from each pieces etc...

The resulting set is the fat Cantor set: it is nowhere dense and it has positive measure.

share|cite|improve this answer

I actually worked this problem out during my Measure Theory course a couple years ago.

Therein, you'll find a construction of a Cantor-like set having any measure strictly between 0 and 1. As with the Cantor set, you cannot find an interval contained in this Cantor-like set.

I don't suggest you accept my solution without checking it yourself, as it was written by one still coming to grips with the material. It should give you a good idea of how to construct your proof, however.

share|cite|improve this answer
ThankS Austin Mohr, both your comments and pdf file are very useful. It is really hard for me to choose an answer between you and Alephomega's and also Rpotrie's one. I just chose the one with highest mark. Thank you again. :-) – Anand Sep 15 '10 at 12:06

From the positive side, as I've mentioned on the comments, you should look at Lebesgue's density theorem. It says that you will get intervals where the measure of the set intersected with the interval is as close to full as you like, in fact, this can be done for intervals around almost every point of the set, and it holds in other contexts also.

share|cite|improve this answer
Thanks Rpotrie for your useful comments. :-) – Anand Sep 15 '10 at 12:06

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.