# Jonality

Asher Auel and I have discussed the following definition, which is fun to think about.

**Jonality**. Let $X$ be an algebraic variety over a field $k$. The *jonality*
of $X$ is the smallest $g$ such that there exists a smooth proper geometrically connected curve $C$ of genus $g$
defined over $k$ and a *non-constant* morphism $C\rightarrow X$.

Jonality is pronounced as in “Johan” plus “gonality”.

The jonality of a curve of genus $g$ is … $g$.

The jonality of projective space $\bP^r$ is $0$ since $\bP^r$ contains lines.

The jonality of a degree $d$ hypersurface $X$ in $\bP^r$ is at most the jonality of a degree $d$ smooth plane curve, i.e., $\frac{(d-1)(d-2)}{2}$ since we can intersect $X$ with $(r-2)$ generic hyperplanes. But, the jonality is typically smaller: think of the lines in a cubic surface.

If $A$ is a simple abelian $g$-fold, its jonality is at least $g$. If additionally $A$ is not isogeneous to a Jacobian, then the jonality is at least $g+1$.

The jonality of K3 surfaces is interesting: apparently it is zero (at least over $\bC$).

The previous examples were geometric in nature. The jonality is also an interesting arithmetic invariant. Asher Auel and I are working on the open problem of whether every Severi–Brauer variety $X$ defined over a field $k$ contains a (possibly singular genus $1$ curve). In other words, we are asking whether the jonality of $X$ is at most $1$. This was a question asked, in different terms, by Pete Clark and David Saltman.

Let $D$ be a division algebra of degree $d$ and let $X$ be its Severi–Brauer variety, an étale-twisted form of $\bP^{d-1}$. If $d=2$, then $X$ is a genus $0$ curve so its jonality is $0$. If $d\geq 3$, then $X$ contains no maps from any (geometrically) rational curve, so its jonality is at least $1$.

If $d=3,4,5$, then Johan de Jong and Wei Ho showed [3] that there is a genus $1$ curve mapping to $X$; i.e., the jonality is $1$ in these cases. Auel has proved the same result for $d=6$ using cohomological techniques to combine the cases of $d=2,3$.

Above $d=6$, little is known in general. Recently, with an observation of Saltman, Auel and I proved the following theorem.

**Theorem**. If $D$ is a division algebra of degree $7$ over a *global* field
of characteristic prime to $7$,
then $D$ is split by a genus $1$ curve; i.e., the jonality of the
Severi–Brauer variety of $D$ is $1$.

This result will appear in forthcoming work which also gives new examples in
degrees $8,9,10$. (**Added 11 June 2021**: this paper has now appeared on the
arXiv.) In the meantime, I would be very interested to hear of other
arithmetically interesting cases of jonality computations.

**Added 04 March 2021**: Asher has done some more digging and has uncovered the
following additional facts. The jonality has been extensively studied for
abelian varieties.

- The jonality of abelian varieties is studied in a paper [1] of Bardelli, Ciliberto, and Verra. They write $\gamma(X)$ for the jonality of $X$ and prove some results for general abelian varieties. Their best results in all dimensions are however superseded by a result of Pirola [4], which says that if $A$ is a very general abelian $g$-fold over $\bC$, then $\gamma(A)>\frac{g(g+1)}{2}$.
- The gonality of an algebraic variety $X$ is defined in [2] by Bastianelli, De Poi, Ein, Lazarsfeld, and Ullery to be the minimum of the gonalities of all normalizations of irreducible proper curves in $X$. Since the (geometric) gonality of a genus $0$ curve is $0$, that of a genus $1$ curve is $2$, and that of a genus $2$ curve is $2$ (as they are hyperelliptic), bounds on the gonality of $X$ have implications for the jonality as well. However, plausibly $X$ might contain high-genus hyperelliptic curves and hence have large jonality but small gonality. For instance, Voisin proves [5] that the gonality of the very general abelian $g$-fold over $\bC$ is linear in $g$; specifically, if $g\geq 2k-1$, then the gonality is at least $k+1$.
- Using bend and break techniques, if $X$ has non-nef canonical bundle, then there are rational curves on $X$ (at least geometrically). So, in this case the jonality is $0$. See for example Section 8 of the lecture notes of Debarre.

# References

[1] Bardelli, Ciliberto, Verra,
*Curves of minimal genus on a general abelian variety*,
Compositio Math. **96** (1995), no. 2, 115–147.

[2] Bastianelli, De Poi, Ein, Lazarsfeld, Ullery, *Measures of irrationality for
hypersurfaces of large degree*, Compos. Math. **153** (2017), no. 11, 2368-2393.
`arXiv:1511.01359`.

[3] de Jong and Ho, *Genus one curves and Brauer–Severi varieties*, Math. Res.
Lett. **19** (2012), no. 6, 1357-1359.
`arXiv:1207.4810`.

[4] Pirola,
*Abel-Jacobi invariant and curves on generic abelian varieties*, Abelian varieties (Egloffstein, 1993), 237–249, de Gruyter, Berlin, 1995.

[5] Voisin,
*Chow ring and gonality of general abelian varieties*,
Ann. H. Lebesgue 1 (2018), 313–332.
`arXiv:1802.07153`.