Positive neighborhoods of rational curves (Q2400474)

The problem of describing neighborhoods of compact curves on complex surfaces goes back to [Math. Ann. 146, 331--368 (1962; Zbl 0173.33004)] \textit{H. Grauert}'s theorem for negatively embedded curves, Arnold's linearisation theorem for elliptic curves and its relation to small denominators, Ueda's theory for embedded curves of zero self-intersection number, inspired by the previous work of \textit{T. Ueda} [J. Math. Kyoto Univ. 22, 583--607 (1983; Zbl 0519.32019)] and Savel'ev's linearisation theorem for the case of rational curves of zero self-intersection number [\textit{V. I. Savel'ev}, Vestn. Mosk. Univ., Ser. I 1982, No. 4, 28--32 (1982; Zbl 0511.32017)]. In the present paper the authors consider holomorphic embeddings of the complex projective line into a complex surface. The first invariant we can attach to this embedding is the self-intersection number of the image. In the paper [the first author and \textit{F. Loray}, C. R., Math., Acad. Sci. Paris 354, No. 5, 470--474 (2016; Zbl 1387.32013)] the authors give an extensive description of these embeddings when the self-intersection number is equal to \(1\). In particular they describe completely those embeddings that are equivalent to the one of the complex projective line as a line of the projective plane. In the paper under review the interest is to present a geometric approach to this case using holomorphic foliations, which perhaps may be useful to study other situations. They take a smooth rational curve \(C\) embedded in a complex surface \(S\) with self-intersection number 1. Let \(H\) be a line in the complex projective plane \(\mathbb CP(2)\). One of the results of [loc. cit.] is the following: suppose that there exist three different holomorphic fibrations over \(C\) in \(S\). Then a neighborhood of \(C\) is holomorphically equivalent to a neighborhood of \(H\) in \(\mathbb CP(2)\). In fact, the main idea to prove the previous theorem is to to analyse the curve of tangencies between two fibrations over \(C\). When this curve is a common fiber, they construct a holomorphic diffeomorphism from a neighborhood of \(C\) to a neighborhood of \(H\). The other possibilities are: (1) the curve of tangencies is \(C\) itself; (2) the curve of tangencies is transverse to \(C\) but it is not a common fiber. In the presence of three different fibrations, necessarily two of them have a common fiber. The paper is very well written.