# proving $\lim\limits_{n\to\infty} \int_{0}^{1} f(x^n)dx = f(0)$ when f is continous on [0,1]

$$\lim\limits_{n\to\infty} \int_{0}^{1} f(x^n)dx = f(0)$$ when f is continuous on $[0,1]$

I know it can be proved using bounded convergence theorem but, I wanna know proof using only basic properties of riemann integral and fundamental theorem of calculus and MVT for integrals ... Thank you.

• $\int_{0}^{1-\xi}f(x^n)dx+\int_{1-\xi}^1f(x^n)dx$ – Nirvanacs Nov 18 '13 at 13:35
• Actually i did it. and using MVT for integral and proved first term of equation be f(0) but then I coudln't dealt with limits and etc. So could you gimme some more hints? – quicksilver Nov 18 '13 at 13:36
• I think that's enough.Can you explain your problems in this method? – Nirvanacs Nov 18 '13 at 13:40

take any $\epsilon$, choose $\delta > 0$ such that $|f(x) - f(0)| < \epsilon$ on $[0,\delta]$. choose $n$ big enough such that $(1-\epsilon)^n < \delta$ then $$|\int_0^1 f(x^n) dx - f(0) |= |\int_0^{(1-\epsilon)} [f(x^n) - f(0)]|dx + \int_{1-\epsilon}^1 [f(x^n) - f(0) ]dx | \leq$$ $$\int_0^{(1-\epsilon)} |f(x^n) - f(0)|dx + \int_{1-\epsilon}^1 |f(x^n) - f(0) |dx$$ first factor is smaller than $\epsilon(1 - \epsilon)$ thanks to the choice of $\delta$ and $n$, second one is smaller than $\epsilon \cdot 2 \sup |f|$ because length of your interval of integration is $\epsilon$ so the result follows since $\epsilon$ was arbitrarily small

Hint:

$f(0)=\int_0^1 f(0) dx$

$x^n\to 0$ for $0\leq x<1$

and $f$ is continuous.

• Technically you need that $f$ is uniformly continuous to pull the limit inside the integral (which is itself a limit). In this case, that is guaranteed by continuity of $f$ on $[0,1]$ and compactness of $[0,1]$. A detail which might be okay to omit for a student far enough along to know the dominated convergence theorem, but I feel it is worth noting it for those who do not. – zibadawa timmy Nov 18 '13 at 21:39
• @zibadawatimmy that's what i meant:) – Haha Nov 19 '13 at 1:12

Handsome tip:

Prove it for $g\left(x\right)=f\left(x\right)-f\left(0\right)$ where $g\left(0\right)=0$ and then make use of $\int_{0}^{1}f\left(x^{n}\right)dx=\int_{0}^{1}g\left(x^{n}\right)+f\left(0\right)dx=\int_{0}^{1}g\left(x^{n}\right)dx+f\left(0\right)$.