1
$\begingroup$

I am working on a problem in which we take some entire function $F$ which has zeros at every integer square root, i.e. $F(\sqrt{n}) = 0$ for every $n \in \mathbb{Z}^+$.

I need to show that, for any radius $r>0$, $\#\{z \in \overline{\mathbb{D}}_r(0) : F(z)=0 \} \geq \frac{r^2}{2}$. That is, the number of zeros of $F$ in a closed disc of radius $r$ is greater than or equal to $r^2/2$.

I am totally stumped on how to show this - any hints would be great.

Following this, I am to show that if $F \not \equiv 0$, $F(0)\neq 0$, and $|F| \leq \exp(\alpha |z|^{\beta})$ holds for $|z| \geq 1$ then $$r^2 \leq 16\alpha r^\beta.$$

For this, I have used Jensen's Formula with the above inequality for the number of zeros, and have obtained $$\frac{r^2}{4} \leq \alpha r^{\beta} - \log|F(0)|$$

How would I deal with the $\log|F(0)|$ term? Is there a way I can justify $r$ getting larger rendering this term obsolete?

EDIT - added further question

If, by the answer below, we have $$r^2 - 1 \leq 4\alpha r ^{\beta} - 4\log|f(0)|$$ then is the following argument valid to reach the conclusion we are instructed to find:

Divide both sides by $r^\beta$ to obtain: $$r^{2-\beta} - \frac{1}{r^\beta} \leq 4A - \frac{4\log|f(0)|}{r^\beta}$$ then as $r \to \infty$ we have $r^{2-\beta} \leq 4A$ and so $r^2 \leq 4Ar^\beta \leq 16Ar^\beta$ as required.

Then, for proving that $\beta<2 \implies F\equiv 0$, we go via contradiction. Let $\beta < 2$ and then assuming $F \not \equiv 0$ and WLOG $F(0)\neq 0$, the above result tells us for large $r$ we have $$r^2 \leq 16Ar^\beta$$ $$\implies r^{2-\beta} \leq 16A.$$

Since $\beta < 2$, we have that $r^{2-\beta}$ will tend to $\infty$ as $r \to \infty$. This contradicts the fact that $r^{2-\beta} \leq 16A$ and so we must have $F \equiv 0$.

$\endgroup$
6
  • 1
    $\begingroup$ for the first part count how many given zeroes (square roots) are in the disc of radius $r$ - there are entire functions which have only such as zeroes so the bound you want must follow from them only $\endgroup$
    – Conrad
    Mar 18 at 19:27
  • $\begingroup$ Would you propose that this is $\frac{r^2}{2}$? And if so, would it be a proof by induction that would give the desired result? $\endgroup$
    – 2307
    Mar 18 at 19:41
  • 1
    $\begingroup$ Just count them - enumerate the first few roots and see what happens for small $r$; as noted in general you cannot assume you have other roots unless more is given about the function $\endgroup$
    – Conrad
    Mar 18 at 19:48
  • $\begingroup$ If it's not an integer than you get $[r^2]$ $\endgroup$
    – Conrad
    Mar 18 at 20:14
  • $\begingroup$ Note that $\frac{r^2}{4} \leq \alpha r^{\beta} - \log|F(0)|$ is sufficient to prove that $\beta \ge 2$. Perhaps that is what you really need? $\endgroup$
    – Martin R
    Mar 19 at 9:37

1 Answer 1

2
$\begingroup$

Let $n(r)$ denote the number of zeros of $F$ in the closed disk $D_r(0)$.

If $r > 0$ and $k = \lfloor r^2 \rfloor $ is the largest integer with $k^2 \le r$ then $f$ has zeros at $\sqrt 1, \sqrt 2, \cdots, \sqrt k \in D_r(0)$, so that $$ n(r) \ge k = \lfloor r^2 \rfloor \, . $$ If $r \ge 1$ then $\lfloor r^2 \rfloor \ge r^2/2$ and therefore $n(r) \ge r^2/2$.

Now we use Jensen's formula in the form $$ \frac{1}{2\pi} \int_0^{2\pi} \log |F(re^{i\theta})| \, d\theta - \log |F(0)| = \int_0^r \frac{n(t)}{t} \, dt $$ If $r \ge 1$ then the right-hand side is $$ \ge \int_1^r \frac{n(t)}{t} \; dt \ge \int_1^r \frac{t^2/2}{t} \, dt = r^2 - 1 \, . $$ And if $|F(z)| \leq \exp(\alpha |z|^{\beta})$ then the left-hand side is $$ \le \alpha r^\beta - \log |F(0)| \, . $$ It follows that $$ r^2 - 1 \le \alpha r^\beta - \log |F(0)| $$ or $$ 1 + \frac{\log |F(0)|-1}{r^2} \le \alpha r^{\beta -2} $$ for $r \ge 1$. If $\beta < 2$ then the right-hand side converges to zero for $r \to \infty$ whereas the left-hand side converges to one. That is not possible, so we necessarily have $$ \beta \ge 2 \, . $$

$\endgroup$
2
  • $\begingroup$ Thank you so much - very clear! A couple small follow-up questions: I have made a small edit to my question - is the argument I have taken valid? It feels a bit circular, especially for the contradiction as I feel like I am assuming $\beta \geq 2$ when I am first dividing through to prove the bound holds? Also, I can see that the inequality for floor function holds, but is there anywhere I could see a proof of how this holds? $\endgroup$
    – 2307
    Mar 22 at 17:47
  • 2
    $\begingroup$ @2307: That looks a bit too complicated. You must assume $F \not\equiv 0$ from the beginning, otherwise you can not apply Jensen's formula. So you show that if $F$ is not identically zero and has zeros at all those integer square roots and has a growth condition then $\beta \ge 2$. $\endgroup$
    – Martin R
    Mar 22 at 19:04

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .