Topological ovals in 4-dimensional translation planes (Q1337782)

Let \(f : \mathbb{R}^ 2 \to \mathbb{R}^ 2\) be a smooth map, and let \(F \subseteq \mathbb{R}^ 2 \times \mathbb{R}^ 2\) denote the graph of \(f\). The map \(f\) is called a planar partition function, if (1) for each \(d \in \mathbb{R}^ 2 - \{0\}\), the map \(x \mapsto f(x + d) - f(x)\) is bijective, and if (2) the collection of all tangent spaces \(T_{(x,f(x))} F\), together with \(\{0\} \times \mathbb{R}^ 2\), forms a partition of \(\mathbb{R}^ 2 \times \mathbb{R}^ 2\). The classical example is the map \(\mathbb{C} \to \mathbb{C}\), \(z \mapsto z^ 2\). Associated with a planar partition function are two 4-dimensional locally compact affine planes, the shift plane \(I(f)\), whose lines are the translates of the graph \(F\), as well as the subsets \(\{x\} \times \mathbb{R}^ 2\), \(x \in \mathbb{R}^ 2\), and the derived affine translation plane \(T_ f\) determined by the partition given in (2). Every continuous collineation of the shift plane \(I(f)\) induces a continuous collineation of the translation plane \(T_ f\) (Prop. 1). Now let \(\overline{T}_ f\) denote the projective closure of \(T_ f\), and let \(p_ \infty\) denote the point at infinity on the vertical line through the origin of \(\mathbb{R}^ 2 \times \mathbb{R}^ 2\). The maps \((x,y) \mapsto (xy - {1\over 3} x^ 3, {1\over 2} y^ 2 - {1\over 12} x^ 4)\) and \((x,y) \mapsto (xy - {1\over 3}x^ 3 - x\), \({1\over 2}(y^ 2 - x^ 2) - {1\over 12} x^ 4)\) are the first known examples of nonclassical planar partition functions with the additional property that \(F \cup \{p_ \infty\}\) is a topological (i.e. closed) oval in \(\overline{T}_ f\) (Prop. 3).