Inflection points and double tangents on anti-convex curves in the real projective plane (Q940726)

A parametrized \(C^1\)-curve in the real projective plane \(\mathbb{P}^2\) is said to be anti-convex (resp. satisfying the BARNER-condition) if for each point \(p\) on the curve, there exists a line which is transveral to the curve and meets the curve only at \(p\). In the excellently written paper the authors prove three fundamental theorems: A: Let \(\gamma\) be a \(C^1\)-regular anti-convex curve in \(\mathbb{P}^2\) which is not a line. If the number \(i(\gamma)\) of independent true inflection points on \(\gamma\) is finite, then so is the number \(\delta(\gamma)\) of elements in a maximal set of independent double tangents, and \(i(\gamma)-2\delta(\gamma)= 3\) holds. In particular, the number \(\delta(y)\) does not depend on the choice of a maximal set of independent double tangents if \(i(\gamma)\) is finite. B: Let \(\gamma\) be a \(C^2\)-regular anti-convex curve in \(\mathbb{P}^2\) which is not a line. Then \(\gamma\) has at least three inflection points with the property that the tangent lines at these inflection points cross \(\gamma\) only once. C: Let \(\gamma\) be a \(C^3\)-regular strictly convex curve of constant width \(d\), which is not a circle. Then there exist at least three osculating \(d\)-circles which cross \(\gamma\) exactly twice, both times tangentially. Moreover, these circles coincide with the osculating circles (in the usual sense) at each of their crossing points on \(\gamma\). The paper is completed by very instructive figures.