# Does $\int_{\mathbb R} f(x)x^n dx = 0$ for $n=0,1,2,\ldots$ imply $f=0$ a.e.?

Let $f(x)$ be a real-valued function on $\mathbb{R}$ such that $x^nf(x), n=0,1,2,\ldots$ are Lebesgue integrable.

Suppose $$\int_{-\infty}^\infty x^n f(x) dx=0$$ for all $n=0,1,2,\ldots.$

Does it follow that $f(x)=0$ almost everywhere?

-
This is a duplicate question. – PEV Jan 29 '11 at 23:49
– PEV Jan 29 '11 at 23:51
@TCL: Note also that the indefinite integral cannot equal $0$... – Arturo Magidin Jan 29 '11 at 23:59
This is more apt I think: math.stackexchange.com/questions/17026/… – Aryabhata Jan 29 '11 at 23:59
Neither of the questions linked in the comments above is a duplicate, but Hans Lundmark's answer on the first question shows that the answer to this one is no: math.stackexchange.com/questions/16831/… – Jonas Meyer Jan 30 '11 at 0:08

Let $f$ be such that for the Fourier transform $\hat f$ it holds that for all $n$ \begin{equation*} \hat f^{(n)}(0) = 0. \end{equation*} Note that this does not imply that $f$ itself is zero since there are infinitely flat functions. By using the rules for Fourier transforms and derivatives you see that $\int x^n f(x) dx = 0$ for all $n$.