Duality of real projective plane curves: Klein's equation (Q1913473)

Let \(\Gamma\) be a real projective plane curve of order \(n\), whose dual \(\widehat \Gamma\) has order \(\widehat n\). For each singular point \(p\in \Gamma\), let \(r_p (\Gamma)\), and \(m_p(\Gamma)\) be the number of real branches of \(\Gamma\), and the multiplicity of \(\Gamma\) at \(p\). The main result of the paper under review says that \[ n-\widehat n=\sum \bigl\{m_p(\Gamma) -r_p(\Gamma) \mid p\in\text{Sing} (\Gamma)\bigr\}-\sum \bigl\{m_p(\widehat \Gamma)- r_p(\widehat \Gamma)\mid p\in\text{Sing} (\widehat \Gamma) \bigr\}. \] As a particular case it follows the classical formula of Klein for real curves having only singularities of Plücker type: \(n+2\tau_2'+ i'=m+2 \delta_2'+k'\), where \(n\) denotes the order, \(\tau_2'\) the number of real bitangents with non-real points of tangency, \(i'\) the number of real flexes, \(m\) is the class, \(\delta_2'\) is the number of isolated real nodes and \(k'\) is the number of real cusps. Klein's proof (1876) of this formula was somewhat heuristic, and thirty years later Juel (1905) provided a new one which was considered correct by \textit{J.-L. Coolidge}, who included it in his book ``A treatise on algebraic plane curves'' (Oxford 1931; JFM 57.0820.06). This is not exactly so and one of the main merits of this clear, precise and beautiful paper is to produce a really correct proof, understandable by a modern reader. The author explains very carefully the topology of the real and complex flag varieties \(F=\{(X,\widehat X) \in P\times \widehat P\mid X\) and \(\widehat X\) are incident\}. As the author points out, this is the point on which the earlier ``proofs'' were silent.