Real plane separating (M-2)-curves of degree d and totally real pencils of degree d-3
Matilde Manzaroli
TL;DR
The paper generalizes the classical quintic result that a real plane curve is separating if and only if its ovals are in non-convex position to all real plane separating $(M-2)$-curves of degree $d\ge 4$. It proves that such curves admit infinitely many totally real pencils of degree $d-3$, with the number of base points determined by the separating gonality: $g-1$ base points when sepgon$(C)=g-1$ and $g-2$ base points when sepgon$(C)=g$, where $g=\frac{(d-1)(d-2)}{2}$ is the genus. The approach combines Gabard’s bound on separating gonality, constructions via separating morphisms, and Bezout-type intersection arguments, with supplementary analysis of complex orientations and the separating semigroup Sep$(C)$. A quintic example ($C_5$) is used to illustrate constraints on separating gonality and to discuss implications for the semigroup and orientation data, highlighting the broader applicability to real plane curves. Overall, the work provides a constructive path to totally real pencils on separating $(M-2)$-curves and deepens connections between real algebraic geometry, morphism theory, and complex orientations.
Abstract
It is well known that a non-singular real plane projective curve of degree five with five connected components is separating if and only if its ovals are in non-convex position. In this article, this property is set into a different context and generalised to all real plane separating (M-2)-curves.
