Theorem 2.38 states the following: If $\{I_n\}$ is a sequence of intervals in $R^1$, such that $I_n$ contains $I_{n+1}$ $(n=1,2,3,…),$ then the infinite intersection of the sets $\{I_n\}$ is not empty.

In the proof, Rudin then defines $\{I_n\} = [a_n, b_n]$, and set $E$ as the set of all $a_n$. Letting $x =$ sup $E$ and arbitrary $m, n \in \mathbf{N}$, the following is true $$a_n \leq a_{m+n} \leq b_{m+n} \leq b_{m}.$$

He then establishes the fact that $x \leq b_m$ for any $m \in \mathbf{N}$. This is the step that I fall short on. Why is this statement true?

  • 2
    $\begingroup$ From the above $a_n\le b_m$, so $b_m$ is an upper bound of $E$, and so $\ge$ the least upper bound of $E$. $\endgroup$ Jun 22, 2017 at 6:18
  • $\begingroup$ Oh that was easy $\endgroup$
    – lithium123
    Jun 22, 2017 at 6:20

1 Answer 1


Suppose there exists $n\in\mathbb N$ such that $x > b_n$. Letting $\varepsilon = x-b_n>0$, there then exists $m\in\mathbb N$ such that $x-a_m < \varepsilon/2$. But then $a_m-b_n = a_m-x+x-b_n>\varepsilon/2$, and so $a_m > b_n$, which is impossible.


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.