2

Suppose that $X$ is a locally compact Hausdorff space, and let $\mu$ be a positive measure defined on the $σ$-algebra of Borel subsets of $X$, with $μ(X)<∞$.

Definition 1. One says that $μ$ is regular if, for every Borel set $E⊆X$, one has that $$ μ(E) = \sup\big\{μ(K): K ⊆ E, \text { and $K$ is compact} \big\} = $$ $$ =\inf\big\{μ(U): U\supseteq E, \text { and $U$ is open} \big\}. $$

Definition 2. An atom for $\mu$ is any Borel subset $A\subseteq X$, such that $\mu(A)>0$, and such that, for every Borel subset $B\subseteq A$, one has that $\mu(B)=0$, or $\mu(A\setminus B)=0$.

Every singleton $A=\{x_0\}$ is an atom, provided its measure is nonzero, and so is $\{x_0\}\cup N$, when $N$ has measure zero.

I can prove that, under the hypothesis that $μ$ is regular and $X$ is second-countable, these are the only atoms, but I could not find this result anywhere in the literature. Can anyone help me locate a reference?


PS 1. This is related to Must atoms of a Borel measure space be singletons?, where the result is proved, but no reference is given.

PS 2. It is known that every finite Borel measure on a second-countable, locally compact Hausdorff space, is regular.

Ruy
  • 20,073
  • Could you add the definition of a regular measure, please? And is the measure assumed to be finite or $\sigma$-finite? Positive on non-empty open sets? This might affect the answer. – Henno Brandsma Feb 11 '22 at 07:13
  • Thanks for the comment and sorry for the lack of precision in my question. I hope I have now added the info you requested. I am not assuming that the measure has full support, so it is allowed to vanish on open sets. On the other hand, you may pass to the support of the measure and hence assume it does have full support if that helps. – Ruy Feb 11 '22 at 17:59

0 Answers0