Article Contents
Article Contents

# Slope gap distribution of saddle connections on the $2n$-gon

• *Corresponding author: Caglar Uyanik

This work started as a project during 2020 SUMRY and the authors are grateful for the support from Yale University. The second named author is grateful for support from the NSF Postdoctoral Fellowship DMS-1903099

• We explicitly compute the limiting slope gap distribution for saddle connections on any $2n$-gon for $n\geq 3$. Our calculations show that the slope gap distribution for a translation surface is not always unimodal, answering a question of Athreya. We also give linear lower and upper bounds for number of non-differentiability points as $n$ grows. The latter result exhibits the first example of a non-trivial bound on an infinite family of translation surfaces and answers a question by Kumanduri-Sanchez-Wang.

Mathematics Subject Classification: Primary: 37D40; Secondary: 37E35.

 Citation:

• Figure 1.  The regular decagon, $O_{10}$. Black vertices are identified together and red vertices are identified together

Figure 2.  A saddle connection $\gamma$ on $X_{10}$

Figure 3.  An example of $\mathcal{S}$ for $n = 7$

Figure 4.  Cylinder decomposition of $O_{2n}$ in the $\theta = \pi/2n$ direction

Figure 5.  A staircase representative of $O_{12}$

Figure 6.  The saddle connections of interest for $\Omega_1$ (in magenta) on $\mathcal{S}$ for $n = 7$

Figure 7.  An illustration of $\Omega_2$ with coordinates in $\mathbb{R}^2$, for any $n$

Figure 8.  Color-coded relevant saddle connections for $n = 7$

Figure 9.  The corresponding division of $\Omega_1$ for $n = 7$ using coordinates in $\mathbb{R}^2$. Each saddle connection is color coded to match the region it wins in. The purple region is $P_1$, the blue region is $P_2$, the green region is $P_3$, the yellow region is $P_4$, the gray region is $P_5$, the orange region is $P_6$, and the red region is $P_7$

Figure 10.  Numerical computation of the volume of ${\rm SL}_2(\mathbb{R})/\Gamma_{O_{2n}}$ using our Poincaré section and return time function compared to the exact volume. Within the limits of our computational software, there was essentially no error in the computed volume

Figure 11.  Gap distributions for the $2n$-gon for several values of $n$. Note the changed scaling on the horizontal axis in the last two graphs

Figure 12.  An illustration of $P_1$ with coordinates in $\mathbb{R}^2$, for any $n$, along with the hyperbola $R(x,y) = t$ (in black; here $t = 3$)

Figure 13.  In this example, $n = 7$. The black line is the hyperbola $R(x,y) = 3.6$

Figure 14.  In this example, $n = 7$ and $i = 3$. The black curve is the hyperbola $R(x,y) = 3.6$

Figure 15.  This is the example of $n = 7$. The black curve is the equation $R(x,y) = 2$

Figure 16.  The region $P_n$ with $n = 7$ with the graph of the equation $R(x,y) = 3$

Figure 17.  Time stamps for boundary crossings

Figure 18.  Plot of the upper bound from Theorem 5.1 (olive dashed line) and the actual number of non-differentiable points (pink squares) against $n$

Figure 19.  Table of the upper bound from Theorem 5.1 and number of non-differentiable points for small $n$

Figure 20.  Labeling of $\mathcal{S}'$ using $\sigma$'s and $\nu$'s for $n = 7$

Figure 21.  Labeling of $\mathcal{S}'$ with $\lambda$'s for $n = 7$

Figure 22.  An example of vertex labeling on $\mathcal{S'}$ for $n = 7$

Figure 23.  If fix $i = 2$ in this case, we see that any vector linking $L_j$ and $L_{k}$ (black vectors) with $k \leq i+1$ while staying in the staircase is either steeper than $\nu_1$ (green vector) or $\nu_2$ (yellow vector)

Figure 24.  If fix $i = 2$ in this case, we see that any vector linking $L_j$ and $R_{k}$ with $k \leq i+1$ while staying in the staircase (black vectors) is either steeper than $\nu_1$ (green vector) or $\nu_2$ (yellow vector)

Figure 25.  For the example of $k = 2$, we see that any vector connecting some $R_j$ or $L_j$ (in this case $L_0$) and $R_2$ will pass above $R_1$ and hence be longer and shallower than $\sigma_2$

Figure 26.  For the example of $k = 1$, we see that any vector passing through $v_1$ is longer and shallower than $\sigma_2$

Figure 27.  If fix $k = 1$ and $i = 2$ in this case, we see that any vector passing through the side of length $h_1$ is steeper than $\nu_1$ (green vector), which in turn is steeper than $\nu_2$

Figure 28.  Vectors ending at $L_k$ for any $k$ (shown in black here) are necessarily steeper than the $\nu$ vector that ends at $L_k$ (shown in green or yellow), which is in turn steeper than every $\sigma$ vector

Figure 29.  Let $n = 7$ and $i = 3$ (so the vector we are trying to show victory for is $\sigma_3$). Vectors ending at $R_k$ for any $k < i$ (shown in black here) are necessarily shallower and longer than the $\sigma$ vector that ends at $R_k$ (in orange), which in turn violates Condition 3

Figure 30.  Let $n = 7$ and $i = 2$ (so the vector we are trying to show victory for is $\sigma_2$). Vectors starting at $R_j$ for any $j \leq i$ (shown in black here) are necessarily steeper than the $\sigma$ vector that starts at $R_j$ (in red), which in turn violates Condition 2

Figure 31.  Let $n = 7$ and $i = 3$ (so the vector we are trying to show victory for is $\sigma_3$). Vectors starting at some $L_j$ and ending at $R_k$ for any $k < i$ (shown in black here) are necessarily shallower and longer than the $\sigma$ vector that ends at $R_k$ (in orange), which in turn violates Condition 3

Figure 32.  Let $n = 7$. Vectors starting at $L_j$ and ending at $R_{j+1}$ are precisely the diagonals of the large rectangles, which have aspect ratio equal to $2 + 2\cos(\pi/n)$

Figure 33.  For instance, any vector passing through the side of length $h_1$ is steeper than $\nu_1$ (green vector), which in turn is steeper than every $\sigma$ vector

Figure 34.  Fix $i = 3$. For instance, we see that any vector passing through $v_1$ is longer and shallower than $\sigma_2$, which fails Condition 3

•  [1] J. S. Athreya, Gap distributions and homogeneous dynamics, Geometry, Topology, and Dynamics in Negative Curvature, London Math. Soc. Lecture Note Ser., Cambridge Univ. Press, Cambridge, 425 (2016), 1-31. [2] J. S. Athreya and J. Chaika, The distribution of gaps for saddle connection directions, Geom. Funct. Anal., 22 (2012), 1491-1516.  doi: 10.1007/s00039-012-0164-9. [3] J. S. Athreya, J. Chaika and S. Lelièvre, The gap distribution of slopes on the golden L, Recent Trends in Ergodic Theory and Dynamical Systems, Contemp. Math. Amer. Math. Soc., Providence, RI, 631 (2015), 47-62.  doi: 10.1090/conm/631/12595. [4] J. S. Athreya and Y. Cheung, A Poincaré section for the horocycle flow on the space of lattices, Int. Math. Res. Not. IMRN, (2014), 2643-2690. doi: 10.1093/imrn/rnt003. [5] S. G. Dani and J. Smillie, Uniform distribution of horocycle orbits for Fuchsian groups, Duke Mathematical Journal, 51 (1984), 184-194.  doi: 10.1215/S0012-7094-84-05110-X. [6] A. Eskin and H. Masur, Asymptotic formulas on flat surfaces, Ergodic Theory and Dynamical Systems, 21 (2001), 443-478.  doi: 10.1017/S0143385701001225. [7] B. Heersink, Poincaré sections for the horocycle flow in covers of SL$(2, \Bbb R)/{\rm SL}(2, \Bbb Z)$ and applications to Farey fraction statistics, Monatsh. Math., 179 (2016), 389-420.  doi: 10.1007/s00605-015-0873-x. [8] P. Hubert and T. A. Schmidt, Invariants of translation surfaces, Ann. Inst. Fourier (Grenoble), 51 (2001), 461-495.  doi: 10.5802/aif.1829. [9] P. Hubert and T. A. Schmidt, An introduction to Veech surfaces, Handbook of Dynamical Systems, Elsevier B. V., Amsterdam, 1B (2006), 501-526.  doi: 10.1016/S1874-575X(06)80031-7. [10] L. Kumanduri, A. Sanchez and J. Wang, Slope gap distributions of Veech surfaces, Preprint, http://arXiv.org/abs/2102.10069, arXiv: 2102.10069. [11] J. Marklof and A. Strömbergsson, The distribution of free path lengths in the periodic lorentz gas and related lattice point problems, Annals of Mathematics, 172 (2010), 1949-2033.  doi: 10.4007/annals.2010.172.1949. [12] H. Masur, Ergodic theory of translation surfaces, Handbook of dynamical systems. Vol. 1B. Elsevier B. V., Amsterdam, 1B (2006), 527-547. doi: 10.1016/S1874-575X(06)80032-9. [13] A. Sanchez, Gaps of saddle connection directions for some branched covers of tori, Ergodic Theory and Dynamical Systems, 42 (2022), 3191-3245. [14] P. Sarnak, Asymptotic behavior of periodic orbits of the horocycle flow and eisenstein series, Comm. Pure Appl. Math., 34 (1981), 719-739.  doi: 10.1002/cpa.3160340602. [15] J. Smillie and C. Ulcigrai, Beyond Sturmian sequences: Coding linear trajectories in the regular octagon, Proc. Lond. Math. Soc. (3), 102 (2011), 291-340.  doi: 10.1112/plms/pdq018. [16] D. Taha, The Boca-Cobeli-Zaharescu map analogue for the Hecke triangle groups $G_q$, Preprint, http://arXiv.org/abs/1810.10668, arXiv: 1810.10668. [17] D. Taha, On cross sections to the geodesic and horocycle flows on quotients of SL$(2, \mathbb{R})$ by hecke triangle groups $G_q$, Preprint, http://arXiv.org/abs/1906.07250, arXiv: 1906.07250. [18] C. Uyanik and G. Work, The distribution of gaps for saddle connections on the octagon, Int. Math. Res. Not. IMRN, 18 (2016), 5569-5602.  doi: 10.1093/imrn/rnv317. [19] W. A. Veech, Teichmüller curves in moduli space, Eisentein series and an application to triangular billiards, Inventiones Mathematicae, 97 (1989), 553-583.  doi: 10.1007/BF01388890. [20] W. A. Veech, Geometric realizations of hyperelliptic curves, Algorithms, Fractals, and Dynamics, (Okayama/Kyoto, 1992), Plenum, New York, (1995), 217-226. [21] W. A. Veech, Siegel measures, Annals of Mathematics, 148 (1998), 865-944.  doi: 10.2307/121033. [22] Y. B. Vorobets, Plane structures and billiards in rational polygons: The Veech alternative, Uspekhi Mat. Nauk, 51 (1996), 3-42.  doi: 10.1070/RM1996v051n05ABEH002993. [23] G. Work, A transversal for horocycle flow on $\mathcal{H}(\alpha)$, Geom. Dedicata, 205 (2020), 21-49.  doi: 10.1007/s10711-019-00464-y. [24] A. Zorich, Flat surfaces, Frontiers in Number Theory, Physics, and Geometry. I, Springer, Berlin, (2006), 439-585. doi: 10.1007/978-3-540-31347-2_13.

Figures(34)