Taming Genus 0 (or 1) components on variables-separated equations

Abstract

To figure properties of a curve of form $C_{f,g} = {(x,y)| f(x) - g(y)= 0}$ you must address the genus 0 and 1 components of its projective normalization $\tilde C_{f,g}$. For $f$ and $g$ polynomials with $f$ indecomposable, [Fr73a] distinguished $\tilde C_{f,g}$ with $u=1$ versus $u > 1$ components (Schinzel's problem). For $u = 1$, [Prop. 1, Fr73b] gave a direct genus formula. To complete $u > 1$ required an adhoc genus computation. [Pak22] dropped the indecomposable and polynomial restrictions but added $\tilde C_{f,g}$ is irreducible ($u = 1$). He showed - for fixed $f$ - unless the Galois closure of the cover for $f$ has genus 0 or 1, the genus grows linearly in deg($g$). Method I and Method II extend [Prop. 1, Fr73b}] using Nielsen classes to generalize Pakovich's formulation for $u > 1$. Method I plays on the covers $f$ and $g$ to the $z$-line, $P^1_z$, from which we compute the fiber product. Method II uses the projection to the $y$-line, $P^1_y$, based on explicitly computing branch cycles for this cover. Hurwitz families track the significance of these components. Expanding on [Prop. 2, Fr73a] shows how to approach Pakovich's problem. With no loss, start with ($f^*,g^*$) which have the same Galois closures, and for which their canonical representations are entangled. They, therefore, produce more than one component on the fiber product. Then, we classify the possible component types, $W$, that appear on $\tilde C_{f^*,g^*}$ using the branch cycles for $W$ that come from Method II. The result is a Nielsen class formulation telling explicitly what $g_1\,$s to avoid to assure the growth of the component genuses of $\tilde C_{f*,g*og_1}$ as deg($g_1$) increases. Of particular note: using and expanding on Nielsen classes and the solution of the genus 0 problem (classifying the monodromy groups of indecomposable rational functions).