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

Let $(X,d)$ be a metric space and $S\subseteq X$. Let $\tau$ be the topology on $X$ induced by $d$ and $\tau_S$ be the subspace topology on $S$: $$ \tau_S = \{S\cap V:V\in \tau\}. $$ Denote $B_r(x)= \{y\in X:d(x,y)<r\}$ and $B'_r(x)= \{y\in S:d(x,y)<r\}$.

I should prove that for $U\subseteq S$ it holds $U\in \tau_S$ iff for any $x\in U$ there is $r_x$ s.t. $B'_{r_x}(x)\subseteq U$.

Proof: Clearly, $B'_r(x) = S\cap B_r(x)$.

  1. If $U\in \tau_S$ then there is $V\in \tau$ s.t. $U = V\cap S$. So, for any $x\in U\subseteq V$ there is $r_x$ s.t. $B_{r_x}(x)\subseteq V$ and hence $B'_{r_x}(x) = B_{r_x}(x)\cap S\subseteq V\cap S = U$.

  2. Conversely, if for any $x\in U$ there is $r_x$ s.t. $B'_{r_x}(x)\subseteq U$, then we can put $$ V = \bigcup\limits_{x\in U}B_{r_x}(x) $$ so $V\cap S =\bigcup\limits_{x\in U}B_{r_x}(x)\cap S = \bigcup\limits_{x\in U}B'_{r_x}(x) =U$.

If you write this proof formally that is quite long. I wonder if there are more direct and nice proofs.

share|cite|improve this question
Isn't this already written formally? Looks like a perfectly good proof to me. Also, you can't do much better than two lines for each direction, no? – Sam Oct 20 '11 at 9:22
@Sam: I wonder if there is an argument simpler then constructing $V$ as an infinite union of balls. – Ilya Oct 20 '11 at 9:30
up vote 3 down vote accepted

The indefinite articles in the second sentence are confusing. From the title and the rest of the question, it seems that what you actually mean is:

Let $\tau$ be the topology on $X$ (namely the one induced by $d$), and let $\tau_S$ be the subspace topology on $S$ (namely the one induced by $\tau$).

Another confusing aspect is that in part 2 you write "Conversely, if $U$ is open w.r.t. to the induced metric on $S$". This is of course equivalent to the actual premise ("for any $x\in U$ there is $r_x$ s.t. $B'_{r_x}(x)\subseteq U$"), but this equivalence is a significant part of what's to be proved. If you're already assuming this equivalence as known, then what you're trying to prove is what the title seems to suggest, namely the equivalence of the subspace topology $\tau_S$ on $S$ induced by $\tau$ and the metric topology $\tau'_S$ on $S$ induced by the metric on $S$ induced by $d$. In that case you can simplify the proof by not talking about the individual points of $U$ at all:

Equivalent claim: For $U\subseteq S$ we have $U\in \tau_S$ if and only if $U\in \tau'_S$.

Proof: Since the $B_r(x)$ form a base of $\tau$, the $B'_r(x)$ form a base of $\tau_S$: For $U\in\tau_S$, there is $V\in\tau$ with $U=S\cap V$ and $V=\bigcup B_{r_i}(x_i)$, so $U=S\cap\bigcup B_{r_i}(x_i)=\bigcup\left(S\cap B_{r_i}(x_i)\right)=\bigcup B'_{r_i}(x_i)$. By definition, the topology with base $B'_r(x)$ is $\tau'_S$.

[Edit regarding the clarified question:]

It's clear from the clarified question that you do want to start from the premise "for any $x\in U$ there is $r_x$ s.t. $B'_{r_x}(x)\subseteq U$" in part 2. In that case, you're not just proving what the title says, but the combination of that and "a set is open if and only if it is a neighbourhood for each of its points". The union of balls with one ball for each point is only required for that second claim, which has nothing to do with the equivalence of the subspace topology and the metric topology. Any proof of that second claim must necessarily use one ball for each point, since that's all that's given and the set wouldn't be open if there weren't such a ball for one of the points.

share|cite|improve this answer
Thank you for pointing out confusing points - is the question correctly written now? – Ilya Oct 20 '11 at 11:29
I think so, except for an extra $U$ after "Conversely, if". – joriki Oct 20 '11 at 11:33
Ok, I fixed it, thanks. – Ilya Oct 20 '11 at 11:38
@Gortaur: I've expanded my answer in response to your clarification of the question. – joriki Oct 20 '11 at 11:49
I'm a bit confused with your answer and have some questions. May I ask you to join the chat room I've created? – Ilya Oct 20 '11 at 12:13

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.