Let $X$ be a del Pezzo surface (over $\mathbb{C}$), which is obtained by a blow up $\pi: X \rightarrow \mathbb{P}^{2}$ in a collection of points. Let $H$ be the hyperplane class of $\mathbb{P}^{2}$.

Question: Consider smooth, irreducible curves $\Sigma$ satisfying the inequality $-K_{X}\cdot \Sigma > \frac{1}{2}\Sigma \cdot \Sigma + \pi^{*}H \cdot \Sigma $.

Is the genus of such curves bounded above?


Here is one proof that the genus is bounded (although presumably a more elegant solution exists).

Write $\Sigma=d\pi^*H - \sum_{i=1}^nm_iE_i$ in the usual basis for $\operatorname{NS}(X)$ (i.e. $\Sigma$ is the strict transform of a curve of degree $d$ in $\mathbb{P}^2$, with multiplicity $m_i$ at each of the points which is blown up).

We can rearrange the inequality into the form $\tfrac12(\Sigma^2 + K_X\Sigma) < \tfrac12(d-\sum m_i)$. Then, from the usual genus formula for $\Sigma\subset X$, we have \begin{equation*} g_\Sigma = 1 + \tfrac12(\Sigma^2 + K_X\Sigma) < 1 + \tfrac12\left(d-\sum_{i=1}^n m_i\right) \end{equation*} so we may assume that $d > 2 + \sum m_i$, else $g_\Sigma \leq 2$.

Now we can prove the result by induction on $n$ (the number of points blown up on $\mathbb{P}^2$). If $n=0$ it is easy to see that the inequality becomes $3d > \tfrac12d^2 + d$, hence $d<4$ and $g\leq1$.

By the induction hypothesis, if $n>0$ then we may assume each $m_i\geq 1$ (else blow down the point with $m_i=0$ and reduce to the $n-1$ case). The inequality is \begin{equation*} 3d - \sum_{i=1}^nm_i > \frac12 \left(d^2 - \sum_{i=1}^n m_i^2 \right) + d \end{equation*} which can be rearranged as \begin{equation*} (d - 2)^2 < \sum_{i=1}^n (m_i-1)^2 + 4 - n \end{equation*} But, since $m_i\geq 1 \: \forall i$, we have $\sum (m_i-1)^2 \leq \left(\sum(m_i-1)\right)^2$ and since $\sum m_i < d - 2$ we have \begin{equation*} (d - 2)^2 < (d-2-n)^2 + 4 - n \end{equation*} which gives the bound $d < \frac{n^2+3n+4}{2n}$, and hence a bound for $g$.

  • $\begingroup$ Very nice, thanks Tom! I think the inequality coming after "The inequality is..." should be reversed, (although the sign is correct again by the next line, so it doesn't affect anything). $\endgroup$
    – Nick L
    Jan 31 '18 at 19:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.