I need to show that there are distinct $\sigma$-finite measures in $B(\mathbb{R})$ that are the same in the open sets. Im not really seeing a good example cause every open set is a countable union of compact sets, so the sum has to be the same , but they have to interchange themselves. Any help is appreciated.


For a real number $x\in\mathbb R$ take $\mu_x(A)=\int_{A}\frac{\mathrm dt}{t^2}+x\delta_0(A)$

Then, $\mathbb R=\{0\}\cup \bigcup_n\left(]-\infty,-\frac1n[\cup]\frac1n,+\infty[\right)$ and each subset has finite measure so $\mu_x$ is $\sigma$-finite. Also it is clear that every $\mu_x$ has the same value on an open set which does not contain $0$, and is infinite as soon as $0$ is in the open set, so that all $\mu_x$ coincide on all open subsets but are not equal.

Thank you Kavi Rama Murthy for noticing me that my former answer was wrong.


I think i found an example. If we consider $\mu(A) = \# (\mathbb{Q} \cap A)$ and $\nu(A) = \# (A \cap(\mathbb{Q}\cup \sqrt{2}))$.

  • $\begingroup$ Are these measures $\sigma$-finite? $\endgroup$ – EpsilonDelta May 13 at 12:41
  • $\begingroup$ Yeah if we consider R to the union of the $\q_{ns}$ and the irrationals. $\endgroup$ – Pedro Santos May 13 at 12:42
  • $\begingroup$ No they have 0 and 1 i think $\endgroup$ – Pedro Santos May 13 at 12:44
  • $\begingroup$ Surely we have $\mu(\mathbb{Q}) = |\mathbb{Q}| = \infty$? $\endgroup$ – EpsilonDelta May 13 at 12:45
  • $\begingroup$ Yes we do , but we consider X to be the countable union of $q_{n}$ and each one of them has measure 1. $\endgroup$ – Pedro Santos May 13 at 18:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.