Conjecture Let be the disjoint union of the graphs and and let be a surface. Is it true that every optimal drawing of on has the property that and are disjoint?
We insist on the usual restrictions for drawings (as in The Crossing Number of the Complete Graph).
Although both crossing numbers and embeddings of graphs on general surfaces are rich and well-studied subjects, their common generalization - drawing graphs on general surfaces has received very little attention. The question highlighted here appears to be quite basic in nature, but due to the combined difficulties of crossings and general surfaces, it may be quite difficult to resolve.
This conjecture is trivially true when is the plane, and DeVos, Mohar, and Samal have proved that it also holds when is the projective plane. It is open for all other surfaces to the best of my (M. DeVos) knowledge.
Bibliography
* indicates original appearance(s) of problem.