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 $\displaystyle f: [0,1] \rightarrow \mathbb{R}$ given by

$$f(x) = \begin{cases} 0 & x \notin \mathbb{Q} \\ \\ 0 & x = 0 \\ \\ \frac{1}{q_x} & x = \frac{p_x}{q_x} \in \mathbb{Q} \backslash \{0\}, \ p_x \in \mathbb{Z}, \ q_x \in \mathbb{N}, \ \text{gcd}(|p_x|, q_x) = 1 \end{cases}$$

Is $\displaystyle f$ Riemann integrable?

I am trying to use the equivalent statements $\displaystyle g:[a,b] \rightarrow \mathbb{R}$ is Riemann integrable and $\displaystyle \forall \epsilon >0 \ \exists$ step functions $\displaystyle \rho, \psi$ with $\displaystyle \rho \leq g \leq \psi$ such that $\displaystyle \int_a^b (\psi - \rho) \leq \epsilon$.

I guess that means I would have to somehow show that given $\displaystyle \epsilon$, there exists only a finite amount of $\displaystyle x \in [0,1]$ with $\displaystyle f(x) \geq \epsilon$?

Is it recommended that I consider something else instead? If not, how should I do this?

share|cite|improve this question
See this. – Shai Covo Jan 12 '11 at 20:13
Your "guess" sounds like a good approach to me. Hint: given $\epsilon$, choose an integer $n$ so large that $1/n < \epsilon$. For which $x$ is $R(x) \ge 1/n$? Try it for the first few values of $n$. Then, if there are only finitely many $x$ with $R(x) \ge \epsilon$, try constructing a step function $\psi$ such that $0 \le R \le \psi$ and $\int_0^1 \psi \le 2 \epsilon$. – Nate Eldredge Jan 12 '11 at 20:16
up vote 3 down vote accepted

Using that given $\epsilon$, $R(x) > \epsilon$ for only finitely many $x_i\in [0,1]$ is a step in the right direction. Think about how to make the intervals containing those $x_i$ arbitrarily small, and $\epsilon$ arbitrarily small at the same time (hint: the size of the intervals should depend on [be bounded by] $\epsilon$).

There is a complete version of this proof in Analysis by Steven Lay on page 278.

edit: I have left out many details and I can give more info about how to go about it if you get stuck.

share|cite|improve this answer

Since you asked for recommendations: It is Riemann integrable because it is continuous almost everywhere (at the irrationals) and bounded. Of course, that is not how that Riemann integrability of this function was originally proved (this is an important example/counter-example function), but the result about continuity almost everywhere is quite important.

Since this is homework: here is a hint using step functions. See Exercise 5.24.

share|cite|improve this answer
Is there a name for this function? If so, I could try and find out more background information... – ghshtalt Jan 12 '11 at 20:11
@user: Some people call it the Modified Dirichlet Function. – Aryabhata Jan 12 '11 at 20:15
@user: Instead of trying to research this function, I suggest you think about it! It is worth thinking about and easy enough to be assigned as homework. – Aryabhata Jan 12 '11 at 20:42

For any $\epsilon > 0$, there are finitely many points in $[0,1]$ such that $R(x) \ge \epsilon / 2$ (specifically, the rational numbers $x$ with denominators $q_x \le 2 / \epsilon$). You can construct a step function $\psi$ that is equal to $1$ when sufficiently near one of these points, equal to $\epsilon / 2$ otherwise, and whose total area is $\le \epsilon$. That should let you conclude that $R$ is Riemann-integrable.

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.