Diophantine approximation on hyperbolic orbifolds (Q1107565)

In recent years a connection has been discovered between simple geodesics on certain Riemann surfaces (perhaps branched) and analogues of the classical Markov spectrum of diophantine approximation. On page 74 of \textit{Harvey Cohn}'s paper [Ann. Math. Stud. 97, 73--85 (1981; Zbl 0455.30030)] a private communication from Werner Fenchel is described that attributes a kind of ``non-intersection'' property to the geodesics associated with certain generators of \(\Gamma'\), the commutator subgroup of \(\Gamma(1)\). (The quotation marks are in Cohn's paper.) Cohn's generators have traces coincident with the Markov spectrum and the surface in question has signature \((1;\infty)\). About 5 years later independent and simultaneous progress was made by the author [[H] Acta Math. 156, 33--82 (1986; Zbl 0593.10028)] and \textit{J. Lehner} and the reviewer [[L-S] Bull. Am. Math. Soc., New Ser. 11, 359--362 (1984; Zbl 0556.10015)]. In Haas' case, he characterized simple geodesics on a Riemann surface of type \((1;\infty)\) in terms of the deepest penetration into the horocyclic neighborhood of \(\infty\). If this penetration was too deep, the geodesic cannot be simple whereas if it is below a certain bound depending on the Teichmüller class of the surface, must be simple. In this case the penetration values turn out to be a discrete set with a single limit point (the bound of the previous sentence). All simple geodesics are limits of the simple closed ones. In the case of \(\Gamma'\), the set of ``simple'' penetration constants is the discrete portion of the Markov spectrum. This Haas paper contains other interesting results concerning the behavior of infinite simple geodesics related to ideas of Thurston. In [L-S], a proof was outlined that the simple closed geodesics on \(H^+/\Gamma (3)\) are exactly the projections of fixed lines of hyperbolic matrices whose traces comprise the Markov spectrum. This surface is a single point in the Teichmüller class with signature \((0;\infty,\infty,\infty,\infty).\) A nice discussion of all these results was given by \textit{C. Series} [Math. Intell. 7, No. 3, 20--29 (1985; Zbl 0566.10024)]. In [Sh] Duke Math. J. 52, 535--545 (1985; Zbl 0578.10027), the reviewer showed that the Markov spectrum (in exactly the same way) characterizes the simple closed geodesics on \(H^+/\Gamma^ 3\). (This last is the subgroup of \(\Gamma(1)\) generated by the cubes of all its elements.) The surface has signature \((0;2,2,2,\infty)\). Because, as this paper acknowledged and as I reiterate here, I had learned from Haas how to apply the work of \textit{A. L. Schmidt} [J. Reine Angew. Math. 286--287, 341--368 (1976; Zbl 0332.10015)], this time the results were for the Teichmüller class, not just the congruence subgroup as in [L-S]. In addition this paper showed that the fact that \(\Gamma^ 3\supseteq \Gamma (3)\), \(\Gamma'\) can be used to show that the previous results of [H] and [L-S] on simple closed geodesics follow from the analogous result on this surface. Again, this works for the entire Teichmüller class. The paper under review extends the results of [H], concerning all simple geodesics, not just the closed ones, to the surfaces of signature \((0;2,2,2,q)\). Here \(q>2\) or \(q=\infty\). As in [Sh], this implies similar results for surfaces of signature \((0;q,q,q,q)\) and \((1;q)\). Also, for the above groups punctures may be replaced by holes; that is the groups may be of the second kind. Taken together with [H], these works of Haas subsume all previous discussions of this connection between number theory and the geometry of simple geodesics on these Riemann surfaces.