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

I am trying to prove that if $\mathcal U$ is a non-convergent ultrafilter on $[0,\infty)$ with its usual topology, then $\mathcal U$ must contain intervals of the form $[a,\infty)$ for every $a \ge 0$.

I am trying to argue by contradiction: Suppose that for some $a \ge 0$, $\mathcal U$ does not contain $[a,\infty)$. Since $\mathcal U$ is an ultrafilter, it must therefore contain $[0,a)$. I want to argue somehow that $\mathcal U \to a$, contradicting that $\mathcal U$ is non-convergent. Thus, I need to show that $\mathcal U$ contains every interval of the form $(c,d)$ with $a \in (c,d)$. Given such a $(c,d)$, I know that $[0,d) \in \mathcal U$ since $[0,a) \subseteq [0,d)$. Somehow I need to show that $\mathcal U$ contains $(c,\infty)$ or something comparable to it, in which case $\mathcal U$ will contain $(c,\infty) \cap [0,d) = (c,d)$. I have tried arguing by contradiction again, but I seem to be getting nowhere. Any tips?

share|cite|improve this question
I think that you don't want to use some $a$. You want to take the set $\{x\in[0,\infty)\mid [x,\infty)\notin\cal U\}$, and argue that since it isn't empty, it must have an infimum, and take that to be your $a$. I mean, note that if $[1,\infty)\notin\cal U$ then certainly $[2,\infty)\notin\cal U$. But it is impossible that $\cal U$ converges to both points, since the space is Hausdorff. – Asaf Karagila Jun 5 '13 at 13:19
up vote 1 down vote accepted

I think your start is OK: we have some $a > 0$ such that $[0,a) \in \mathcal{U}$, and so also $[0,a] \in \mathcal{U}$, as this is an enlargement.

Now $\mathcal{U}' = \{ A \cap [0,a]: A \in \mathcal{U} \}$ is an ultrafilter on the compact set $[0,a]$, so it converges to some unique $p \in [0,a]$. But then $\mathcal{U}$ also converges to $p$ and this contradicts that $\mathcal{U}$ is not convergent.

share|cite|improve this answer
Well that was easy. Why do I complicate my life all the time... – echoone Jun 5 '13 at 17:25
Most things are easy once you've seen them. It's a logical idea, as the space is "almost" compact: all closed initial segments are. So the "escape the compactness" (because we have a non-convergent ultrafilter) it must "live" on the non-compact part: the tail. Glad to be able to help. – Henno Brandsma Jun 5 '13 at 17:33

You may want to take the infimum of all $a$'s so that $[0,a)$ is included in $\mathcal U$.

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.