Markov's theorem classifies the worst irrational numbers with respect to rational approximation and the indefinite binary quadratic forms whose values for integer arguments stay farthest away from zero. The main purpose of this paper is to present a new proof of Markov's theorem using hyperbolic geometry. The main ingredients are a dictionary to translate between hyperbolic geometry and algebra/number theory, and some very basic tools borrowed from modern geometric Teichmüller theory. Simple closed geodesics and ideal triangulations of the modular torus play an important role, and so does the problem: How far can a straight line crossing a triangle stay away from the vertices? Definite binary quadratic forms are briefly discussed in the last section.
11J06, 32G15arXiv:1702.05061v1 [math.GT] 16 Feb 2017It is fun and a recommended exercise to consider this question in elementary euclidean geometry. Here, we need to deal with ideal hyperbolic triangles, decorated with horocycles at the vertices, and "distance from the vertices" is to be understood as "signed distance from the horocycles" (Sec. 13).The subjects of this article, Diophantine approximation, quadratic forms, and the hyperbolic geometry of numbers, are connected with diverse areas of mathematics, ranging from from the phyllotaxis of plants [16] to the stability of the solar system [36], and from Gauss' Disquisitiones Arithmeticae to Mirzakhani's recent Fields Medal [51]. An adequate survey of this area, even if limited to the most important and most recent contributions, would be beyond the scope of this introduction. The books by Aigner [2] and Cassels [11] are excellent references for Markov's theorem, Bombieri [6] provides a concise proof, and more about the Markov and Lagrange spectra can be found in Malyshev's survey [46] and the book by Cusick and Flahive [20]. The following discussion focuses on a few historic sources and the most immediate context and is far from comprehensive.One can distinguish two approaches to a geometric treatment of continued fractions, Diophantine approximation, and quadratic forms. In both cases, number theory is connected to geometry by a common symmetry group, GL 2 ( ). The first approach, known as the geometry of numbers and connected with the name of Minkowski, deals with the geometry of the 2 -lattice. Klein interpreted continued fraction approximation, intuitively speaking, as "pulling a thread tight" around lattice points [40] [41]. This approach extends naturally to higher dimensions, leading to a multidimensional generalization of continued fractions that was championed by Arnold [3] [4]. Delone's comments on Markov's work [22] also belong in this category (see also [29]).In this article, we pursue the other approach involving Ford circles and the Farey tessellation of the hyperbolic plane (Fig. 6). This approach could be called the hyperbolic geometry of numbers. Before Ford's geometric proof [27] of Hurwitz's theorem [37] (Sec. 2), Speiser had apparently used the Ford circles to prove a weaker approximation theo...