MathOverflow is a question and answer site for professional mathematicians. 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 $a$ be a closed point in $\mathbf{P}^1_{\overline{\mathbf{Q}}}$.

Let $Y \cong \mathbf{P}^1_{\overline{\mathbf{Q}}} $ and let $\pi:Y\to \mathbf{P}^1_{\overline{\mathbf{Q}}}$ be a finite morphism which is unramified above $a$. Let $b$ be a point in $\pi^{-1}(a)$.

Can one effectively bound the height of $b$ in terms of the degree of $\pi$, some data depending on the branch points of $\pi$ and the height of $a$?

Ineffectively this should be possible.

Maybe it's more natural to ask if one can bound the height of $\pi^{-1}(a)$ in terms of the degree of $\pi$, some data depending on the branch points of $\pi$ and the height of $a$.

share|cite|improve this question
What do you mean by the "naive height of b" ? (a Weil height depends on a line bundle). ALso, I think $\pi=f$ in your post. – Damian Rössler Oct 3 '11 at 9:10
Since you assume $Y\simeq \mathbf P^1$, it is more natural to give yourself $\pi$ as a rational function of some degree $d$. Then, there are effective constants $c$ and $c'$ such that $ d h(y) -c \leq h(\pi(y)) \leq d h(y)+c'$ for all $y\in \mathbf P^1(\overline{\mathbf Q})$. The constant $c'$ is easy to obtain. The effectivity of $c$ is more subtle (one can get one through resultants, but in a more general setting, some form of effective Nullstellensatz is necessary). – ACL Oct 3 '11 at 21:09
up vote 2 down vote accepted

Let me reformulate the problem in less fancy terms. Let $\pi(x)\in\overline{\mathbb{Q}}$ be a rational function of degree $d\ge1$. As ACL noted, there is a standard estimate $$ dh(y) - c_1(\pi) \le h(\pi(y)) \le dh(y) + c_2(\pi), \quad (*) $$ where it is relatively easy to give explicit formulas for $c_1$ and $c_2$ in terms of $d$ and the coefficients of the polynomials defining $\pi$. However, what you seem to be asking is whether we can give formulas for $c_1$ and $c_2$ that depend only on $d$ and the branch points of $\pi$.

I suspect that the answer is no, for the following reason. For simplicity, lets assume that $\pi(\infty)\ne0$, so I can normalize $\pi$ so that its numerator is monic of degree $d$. A rational function of degree $d$ has $2d-2$ branch points, but with this normalization, it has $2d+1$ free coefficients. So fixing the branch points should give a 3 dimensional family of maps with those branch points. By varying $\pi$ within such a family, the values of $c_1(\pi)$ and $c_2(\pi)$ that come from the usual proof will not be bounded. This makes me suspect that by varying $\pi$ within the family, the inequality $(*)$ will fail if we require $c_1$ and $c_2$ to be fixed.

This isn't a proof, of course, but I hope it suggests where one might look for a negative answer to your second question.

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.