Crossing sequences
Then there exists a graph that be drawn on a surface with orientable (nonorientable, resp.) genus with crossings, but not with less crossings.
This actually are two conjectures, one for the orientable case and another for nonorientable one. For sequences the nonorientable case was resolved in [ABS] and the orientable one in [DMS].
The conclusion also holds (for the orientable case) whenever the sequence is convex [S], that is whenever is nonincreasing. It might seem that this condition is also necessary: For the most extreme sequence (suggested by Salazar) one needs to construct a graph for which adding one handle saves just one crossing, while adding another saves many -- but then why not add the second handle first? Somewhat surprisingly, graphs with this counterintuitive property exist, at least for sequences .
An interesting open case is to consider sequences for which for some and small .
Bibliography
*[ABS] Dan Archdeacon, C. Paul Bonnington, and Jozef Siran, Trading crossings for handles and crosscaps, J.Graph Theory 38 (2001), 230--243.
[DMS] Matt DeVos, Bojan Mohar, Robert Samal, Unexpected behaviour of crossing sequences, in preparation
[S] Jozef Siran, The crossing function of a graph, Abh. Math. Sem. Univ. Hamburg 53 (1983), 131--133.
* indicates original appearance(s) of problem.