Skip to main content
Log in

A Universal Triangulation for Flat Tori

  • Published:
Discrete & Computational Geometry Aims and scope Submit manuscript

Abstract

A construction due to Burago and Zalgaller (Vestnik Leningrad Univ 15:66–80, 1960; St Petersburg Math J 7(3):369–385, 1995) shows that every orientable polyhedral surface, one that is obtained by gluing Euclidean polygons, has an isometric piecewise linear embedding into Euclidean space \(\mathbb {E}^3\). A flat torus, resulting from the identification of the opposite sides of a Euclidean parallelogram, is a simple example of polyhedral surface. The embeddings constructed according to Burago and Zalgaller may have a huge number of vertices, moreover distinct for every flat torus. Based on another construction of Zalgaller (J Math Sci 100(3):2228–2238, 2000. https://doi.org/10.1007/s10958-000-0007-3) and on recent works by Arnoux et al. (2021, in preparation), we exhibit a universal triangulation with 2434 triangles which can be embedded linearly on each triangle in order to realize the metric of any flat torus.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14
Fig. 15
Fig. 16
Fig. 17
Fig. 18
Fig. 19

Similar content being viewed by others

Notes

  1. As often the case in the literature, we use tha adjective “linear” to qualify a map that preserves a specific triangulation, and reserve the terminology “piecewise linear” for a map defined on a triangulable space without a prescribed triangulation, which is linear for some triangulation.

  2. This expression gives the same value as the formula given in [16], but is somewhat simpler.

  3. Our computations do not lead to the same formulas as in [16].

  4. While the symmetry of the cross section would allow to turn by an angle with amplitude at most \(\pi /3\), it is in fact necessary to allow for all the possible angles in order to obtain a unique isomorphism class of triangulations after the final gluing.

  5. The orbit of \({{\mathcal {M}}}^+\) under the action of \(\text {SL}_2(\mathbb {Z})\) only includes the positively oriented triangles of the Dedekind tessellation. In particular, the orbit of \({{\mathcal {M}}}^+\) by \(\varDelta \) includes one out of every two triangles in the fan.

  6. We thank the anonymous reviewers for suggesting this improvement.

  7. Here and in the sequel, we write \(x\approx y\), where \(y=\sum _{i=k}^{\ell }d_i 10^i\) is a decimal with \(d_k\ne 0\), to mean that \(|x-y| < 10^{k}\).

References

  1. Arnoux, P., Lelièvre, S., Málaga, A.: Diplotori: a family of polyhedral flat tori. In preparation (2021). https://www.imaginary.org/hands-on/diplotori-flat-polyhedral-tori

  2. Banchoff, T.F.: Geometry of the Hopf mapping and Pinkall’s tori of given conformal type. In: Tangora, M.C. (ed.) Computers in Algebra. Lecture Notes in Pure and Applied Mathematics, vol. 111, pp. 57–62. M. Dekker, New York (1988)

    Google Scholar 

  3. Borrelli, V., Jabrane, S., Lazarus, F., Thibert, B.: Flat tori in three dimensional space and convex integration. Proc. Natl Acad. Sci. U.S.A. (PNAS) 109(19), 7218–7223 (2012)

    Article  MathSciNet  Google Scholar 

  4. Burago, Y.D., Zalgaller, V.A.: Isometric piecewise-linear imbeddings of two-dimensional manifolds with a polyhedral metric into \({\mathbb{R}}^3\). Algebra Anal. 7(3), 76–95. Transl. St Petersburg Math. J. 7(3), 369–385 (1995)

  5. Burago, Y.D., Zalgaller, V.A.: Polyhedral realizations of developments. Vestnik Leningrad. Univ. 15, 66–80 (1960). (in Russian)

    MathSciNet  Google Scholar 

  6. Farb, B., Margalit, D.: A Primer on Mapping Class Groups. Princeton University Press, Princeton (2012)

    Google Scholar 

  7. Kuiper, N.: On \({C}^1\)-isometric imbeddings. Indagationes Mathematicae 17, 545–555 (1955)

    Article  Google Scholar 

  8. Lazarus, F., Tallerie, F.: A universal triangulation for flat tori. In: 38th International Symposium on Computational Geometry (SoCG 2022). Schloss Dagstuhl-Leibniz-Zentrum für Informatik (2022). https://drops.dagstuhl.de/opus/volltexte/2022/16061/pdf/LIPIcs-SoCG-2022-53.pdf

  9. Lazarus, F., Tallerie, F.: Polyhedral Realizations of Flat Tori (2022). https://www.imaginary.org/gallery/polyhedral-realizations-of-flat-tori

  10. Lévy-Leblond, J.M.: A symmetric 3D proof of Heron’s formula. Math. Intell. 43(2), 37–39 (2021). https://doi.org/10.1007/s00283-020-09996-8

    Article  MathSciNet  Google Scholar 

  11. Nash, J.F.: \({C}^1\)-isometric imbeddings. Ann. Math. 60(3), 383–396 (1954). https://doi.org/10.2307/1969840

    Article  MathSciNet  Google Scholar 

  12. Pinkall, U.: Hopf tori in \(\text{ S}^3\). Invent. Math. 81(2), 379–386 (1985). https://doi.org/10.1007/BF01389060

    Article  MathSciNet  Google Scholar 

  13. Quintanar, T.: An explicit PL-embedding of the square flat torus into \({\mathbb{E} }^3\). J. Comput. Geom. 11(1), 615–628 (2020). https://doi.org/10.20382/jocg.v11i1a24

    Article  MathSciNet  Google Scholar 

  14. Segerman, H.: Visualizing Mathematics with 3D Printing. John Hopkins University Press, Baltimore (2016)

    Book  Google Scholar 

  15. Tsuboi, T.: On origami embeddings of flat tori. arXiv Preprint (2020). arXiv:2007.03434

  16. Zalgaller, V.A.: Some bendings of a long cylinder. J. Math. Sci. 100(3), 2228–2238 (2000). https://doi.org/10.1007/s10958-000-0007-3

    Article  MathSciNet  Google Scholar 

Download references

Acknowledgements

We warmly thank Alba Málaga, Pierre Arnoux and Samuel Lelièvre for sharing with us their constructions of flat tori and showing us how to cover their moduli space with these constructions. Figures 14 and 16 were computed thanks to their Sage program. We also thank the anonymous reviewers for their many suggestions, including the improvement for the size of the overlay of the universal triangulations of short and long tori.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Francis Lazarus.

Additional information

Editor in Charge: Kenneth Clarkson

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

Here, we provide the details for the proof of Lemma 3.9. From Lemma 3.7, simple computations show that the region \({{\mathcal {M}}}_{19,2}\) is bounded by the following parametrized curves:

  • \(\lambda _2(t) = z_2 + it\) with \(t \in [0, +\infty [\) and \(z_2 = \frac{2 - \sin \frac{2\pi }{19} \cot \frac{\pi }{19} + i \sin \frac{2 \pi }{19}}{19}\),

  • \(\beta _{2,1}(t) = \frac{2}{19} - \frac{\sin \frac{2\pi }{19}}{19 \sin \frac{\pi }{19}} e^{-it}\) with \(t \in [\frac{\pi }{19}, \frac{3 \pi }{19}]\),

  • \(\beta _{2,2}(t) = \frac{2 + i\cot \frac{\pi }{19}}{19} - \frac{e^{i\frac{15\pi }{38}}}{19 \sin \frac{\pi }{19}} t\) with \(t \in [\cos \frac{16\pi }{19}, \cos \frac{3\pi }{19}]\),

  • \(\rho _2(t) = w_2 + it\) with \(t \in [0, +\infty [\) and \(w_2 = \beta _{2,2}\left( \cos \frac{16 \pi }{19}\right) \).

While \(\mathcal {M}_{19,7}\) is bounded by:

  • \(\lambda _7(t) = z_7 + it\) with \(t \in [0, +\infty [\) and \(z_7 = \frac{7 - \sin \frac{7\pi }{19} \cot \frac{\pi }{19} + i \cot \frac{\pi }{19} \left( 1 - \cos \frac{7\pi }{19}\right) }{19}\),

  • \(\beta _{7,1}(t) = \frac{7 + i \cot \frac{\pi }{19}}{19} - \frac{e^{i \frac{5\pi }{38}}}{19 \sin \frac{\pi }{19}} t\) with \(t \in [\cos \frac{6\pi }{19}, \cos \frac{\pi }{19}]\),

  • \(\beta _{7,2}(t) = \frac{7}{19} - \frac{\sin \frac{7\pi }{19}}{19 \sin \frac{\pi }{19}} e^{-it}\) with \(t \in [\frac{6\pi }{19}, \frac{8\pi }{19}]\),

  • \(\beta _{7,3}(t) = \frac{7 + i\cot \frac{\pi }{19}}{19} - \frac{e^{i \frac{5\pi }{38}}}{19 \sin \frac{\pi }{19}} t\) with \(t \in [\cos \frac{11\pi }{19}, \cos \frac{8\pi }{19}]\),

  • \(\rho _7(t) = w_7 + it\) with \(t \in [0, +\infty [\) and \(w_7 = \beta _{7,3}\left( \cos \frac{11\pi }{9}\right) \).

And \(\mathcal {M}_{19,13}\) is bounded by:

  • \(\lambda _{13}(t) = z_{13} + it\) with \(t \in [0, +\infty [\) and \(z_{13} = \frac{13 - \sin \frac{13\pi }{19} \cot \frac{\pi }{19} + i \cot \frac{\pi }{19} \left( 1 - \cos \frac{13\pi }{19}\right) }{19}\),

  • \(\beta _{13,1}(t) = \frac{13 + i \cot \frac{\pi }{19}}{19} - \frac{e^{-i\frac{7\pi }{38}}}{19 \sin \frac{\pi }{19}} t\) with \(t \in [\cos \frac{5 \pi }{19}, \cos \frac{\pi }{19}]\),

  • \(\rho _{13}(t) = w_{13} + it\) with \(t \in [0, +\infty [\) and \(w_{13} = \beta _{13,1}\left( \cos \frac{5\pi }{19}\right) \).

In the sequel we denote by \(\mathcal {B}(z, r)\) the closed disk of radius r centered at z and by \(\mathcal {C}(z, r)\) its boundary circle. We also denote by \(\Re (z)\) and \(\Im (z)\), the real part and the imaginary part of z, respectively. Recall that each \(g_{\delta }: z \mapsto \frac{1}{-z+\delta }\) sends the horizontal line \(\{\Im (z) = h\}\) onto the circle \(\mathcal {C}\left( \frac{i}{2\,h},\frac{1}{2\,h}\right) \). Furthermore, for \(\delta \ge 1\), \(g_{\delta }\) sends the imaginary axis onto the circle \(\mathcal {C}\left( \frac{1}{2\delta }, \frac{1}{2\delta }\right) \), and the line \(\{\Re (z) = \frac{1}{2}\}\) onto the circle \(\mathcal {C}\left( \frac{1}{2\delta - 1}, \frac{1}{2\delta - 1}\right) \). We denote the red, blue and yellow slices of \({{\mathcal {M}}}^+_\text {short}\) on Fig. 15 by respectively \({{\mathcal {S}}}_r\), \({{\mathcal {S}}}_b\) and \({{\mathcal {S}}}_y\).

From these facts, we deduce that \(g_5({{\mathcal {S}}}_r)\), the image of the red slice by \(g_5\) (see Fig. 16), is bounded by four arcs of circles; one from respectively \(\mathcal {C}\left( \frac{1}{9}, \frac{1}{9}\right) \), \(\mathcal {C}\left( \frac{i}{66},\frac{1}{66}\right) , \mathcal {C}\left( \frac{1}{10}, \frac{1}{10}\right) \) and \(\mathcal {C}\left( \frac{i}{50}, \frac{1}{50}\right) \). Similarly, the image \(g_3({{\mathcal {S}}}_b)\) of the blue slice is bounded by arcs from the circles \(\mathcal {C}\left( \frac{1}{5}, \frac{1}{5}\right) , \mathcal {C}\left( \frac{i}{50}, \frac{1}{50}\right) , \mathcal {C}\left( \frac{1}{6},\frac{1}{6}\right) \) and \(\mathcal {C}\left( \frac{i}{24}, \frac{1}{24}\right) \). Finally, the image \(g_1({{\mathcal {S}}}_y)\) of the yellow slice is bounded by arcs of the circles \(\mathcal {C}(1, 1), \mathcal {C}\left( \frac{i}{24}, \frac{1}{24}\right) , \mathcal {C}\left( \frac{1}{2},\frac{1}{2}\right) \) and a segment of the line \(\{ \Re (z) = \frac{1}{2} \}\). For this last slice, we note that \(g_1\) sends the arc of circle \(\{ e^{it} \mid \frac{\pi }{3} \le t \le \frac{\pi }{2}\}\) to the vertical line segment \(\{ \frac{1}{2} + i t \mid t \in [\frac{1}{2}, \frac{\sqrt{3}}{2}] \}\).

We now proceed to prove that the three curvilinear quadrilaterals \(g_5({{\mathcal {S}}}_r)\), \(g_3({{\mathcal {S}}}_b)\) and \(g_1({{\mathcal {S}}}_y)\) shown in Fig. 16 lie above the lower boundary of \(\mathcal {M}_{19}:= \mathcal {M}_{19,2} \cup \mathcal {M}_{19, 7} \cup \mathcal {M}_{19, 13}\). (A point z lies above another point w with same real part if \(\Im (z) \ge \Im (w)\).) Let us remark that for showing that such a quadrilateral lies above some boundary, it suffices to show that the bottom side and the right most side of the quadrilateral lie above this boundary as the quadrilateral is completely included in the region of the plane above these two sides.

The Red quadrilateral \(g_5({{\mathcal {S}}}_r)\). Denote the vertices of this quadrilateral as in Fig. 20. From the above description, one computes \(A_1 = \frac{6 + 44 i}{1479}, A_2 = \frac{5 + 33 i}{1114}, A_3 = \frac{1 + 5 i}{130}, A_4 = \frac{18 + 100 i}{2581}\). By the previous remark it is enough to show that the curvilinear sides \(\overset{\frown }{A_1A_2}\) and \(\overset{\frown }{A_2A_3}\) lie above the lower boundary of \(\mathcal {M}_{19}\).

Fig. 20
figure 20

The red quadrilateral \(A_1A_2A_3A_4\)

  • Arc \(\overset{\frown }{A_1A_2}\). The vertical line through \(\beta _{2,1}(t)\) cuts \(\mathcal {C}\left( \frac{i}{66}, \frac{1}{66}\right) \) (the circle containing \(\overset{\frown }{A_1A_2}\)) in two points. We denote by \(P_1(t)\) the highest of these two points. Remark that \(\overset{\frown }{A_1A_2}\) is indeed in the upper half part of \(\mathcal {C}\left( \frac{i}{66}, \frac{1}{66}\right) \). We compute \(P_1(t) = \frac{i + e^{i \tau }}{66}\) with \(0 \le \tau \le \pi \), where \(\frac{\cos \tau }{66}= \frac{2}{19} - \frac{\sin \frac{2\pi }{19}}{19 \sin \frac{\pi }{19}} \cos t\). Thus:

    $$\begin{aligned} \begin{aligned} \Im (P_1(t))&= \frac{1 + \sin \tau }{66} \ge \Im (\beta _{2,1}(t)) = \frac{\sin \frac{2\pi }{19}}{19\sin \frac{\pi }{19}} \sin t \\&\iff \frac{\sin ^2\tau }{66^2} \ge \left( -\frac{1}{66} + \frac{\sin \frac{2\pi }{19}}{19 \sin \frac{\pi }{19}}\sin t \right) ^2 \\&\iff f_1(\cos t)\ge 0, \end{aligned} \end{aligned}$$

    where

    $$\begin{aligned} f_1(x):= \frac{8x}{361}\cos \frac{\pi }{19} + \frac{2\sqrt{1 - x^2}}{627} \cos \frac{\pi }{19} - \frac{6 + 2 \cos \frac{2\pi }{19}}{361} \end{aligned}$$

    A study of \(f_1\) shows that it is non negative on [0.96, 1], so that \(f_1(\cos t)\) is non negative for \(t \in [0, t_{red}^1:= 0.283]\). Note that \(0<\pi /19<t_{red}^1<3\pi /19\). Moreover,Footnote 7\(\Re \left( z_2=\beta _{2,1}\left( \frac{\pi }{19}\right) \right) \approx 0.002 < \Re (A_1) = \frac{6}{1479}\) and \(\Re (\beta _{2,1}(t_{red}^1)) \approx 0.0055 > \Re (A_2) = \frac{5}{1114}\). This shows that the arc \(\overset{\frown }{A_1A_2}\) entirely lies above \(\beta _{2,1}\).

  • Arc \(\overset{\frown }{A_2A_3}\). The vertical line through \(\beta _{2,1}(t)\) cuts \(\mathcal {C}\left( \frac{1}{10}, \frac{1}{10}\right) \) (the circle containing \(\overset{\frown }{A_2A_3}\)) in two points. Denoting by \(P_2(t)\) the highest of these two points, similar computations as above lead to:

    $$\begin{aligned}{} & {} \Im (P_2(t)) \ge \Im (\beta _{2,1}(t))\\ {}{} & {} \quad \iff f_2(\cos t) := \frac{1}{1805}\frac{\sin \frac{2\pi }{19}}{\sin \frac{\pi }{19}} \cos t + \frac{18}{1805} - \frac{\sin ^2\frac{2\pi }{19}}{361 \sin ^2\frac{\pi }{19}} \ge 0. \end{aligned}$$

    Since \(f_2\) is non negative on [0.55, 1], we have that \(f_2(\cos t)\) is non negative for \(t \in [0, 0.9]\). Note that this interval contains the interval of definition \([\frac{\pi }{19}, \frac{3 \pi }{19}]\) of \(\beta _{2,1}\). Since \(\Re (z_2) < \Re (A_2)\) and \(\Re \left( \beta _{2,1}\left( \frac{3\pi }{19}\right) \right) \approx 0.014 > \Re (A_3) = \frac{1}{130}\), the arc \(\overset{\frown }{A_2A_3}\) is included in the region above \(\beta _{2,1}\). We thus conclude that \(A_1A_2A_3A_4\) lies entirely above \(\beta _{2,1}\).

The blue quadrilateral \(g_3({{\mathcal {S}}}_b)\). Denote the vertices of this quadrilateral \(B_1, B_2, B_3, B_4\) in analogy with what precedes (\(B_1\) is the bottom left vertex of the quadrilateral, \(B_2\) the bottom right one, \(B_3\) the top right one and \(B_4\) the top left one). We compute \(B_1 = \frac{2 + 20 i}{505}, B_2 = \frac{3 + 25 i}{634}, B_3 = \frac{1 + 4 i}{51}, B_4 = \frac{10 + 48 i}{601}\). Again, it is enough to show that the curvilinear sides \(\overset{\frown }{B_1B_2}\) and \(\overset{\frown }{B_2B_3}\) lie above the lower boundary of \(\mathcal {M}_{19}\).

  • Arc \(\overset{\frown }{B_1B_2}\). The vertical line passing by \(\beta _{2,1}(t)\) cuts \(\mathcal {C}\left( \frac{i}{50}, \frac{1}{50}\right) \) (the circle containing \(\overset{\frown }{B_1B_2}\)) in two points. Let \(Q_1(t)\) denotes the highest of these two points. We have

    $$\begin{aligned}&\Im (Q_1(t)) \ge \Im (\beta _{2,1}(t)) \iff \\&f_3(\cos t) := \frac{4 \sin \frac{2\pi }{19}}{361 \sin \frac{\pi }{19}} \cos t + \frac{\sin \frac{2\pi }{19}}{475 \sin \frac{\pi }{19}} \sqrt{1 - \cos ^2 t} -\frac{4}{361} - \frac{\sin ^2\frac{2\pi }{19}}{361 \sin ^2\frac{\pi }{19}} \ge 0. \end{aligned}$$

    Furthermore, \(f_3\) is non negative on [0.94, 1], so that \(f_3(\cos t)\) is non negative on \([0, t_{blue}^1:= 0.3]\). Note that \(0< \pi /19<0.3 < 3\pi /19\). Moreover, \(\Re (z_2) \approx 0.002 < \Re (B_1) = \frac{2}{505}\) and \(\Re (\beta _{2,1}(t_{blue}^1)) \approx 0.008 > \Re (B_2) = \frac{3}{634}\), which implies that \(\overset{\frown }{B_1B_2}\) lies entirely above \(\beta _{2,1}\).

  • Arc \(\overset{\frown }{B_2B_3}\). Denote by \(Q_2(t)\) the highest intersection point of the vertical line passing through \(\beta _{2,1}(t)\) with \(\mathcal {C}\left( \frac{1}{6},\frac{1}{6}\right) \) (the circle containing \(\overset{\frown }{B_2B_3}\)). We have

    $$\begin{aligned}{} & {} \Im (Q_2(t)) \ge \Im (\beta _{2,1}(t)) \\ {}{} & {} \quad \iff f_4(\cos t) := - \frac{7 \sin \frac{2\pi }{19}}{1083 \sin \frac{\pi }{19}} \cos t + \frac{26}{1083} - \frac{\sin ^2\frac{2\pi }{19}}{361\sin ^2\frac{\pi }{19}} \ge 0. \end{aligned}$$

    Since \(f_4\) is non negative on \([-1, 1]\) and \(\Re (z_2) < \Re (B_2)\), it follows that \(\overset{\frown }{B_2B_3}\) is above \(\beta _{2,1}\) over the interval \([\Re (B_2), \Re \left( \beta _{2,1}\left( \frac{3\pi }{19}\right) \right) ]\). Let \(\overline{\beta }_{2,2}\) be the supporting line of \(\beta _{2,2}\). The point of \(\overline{\beta }_{2,2}\) on the same vertical as \(B_2\) is \(Q_3:= \beta _{2,2}\left( \frac{1211 \sin \frac{\pi }{19}}{634 \cos \frac{15\pi }{38}} \right) \), while the point of \(\beta _{2,2}\) on the same vertical line as \(B_3\) is \(Q_4:= \beta _{2,2}\left( t_{blue}^2:= \frac{83 \sin \frac{\pi }{19}}{51 \cos \frac{15\pi }{38}} \right) \). Observe that \(t_{blue}^2 \in [\cos \frac{16\pi }{19}, \cos \frac{3\pi }{19}]\). We compute \(\Im (Q_3) \approx 0.02 < \Im (B_2) = \frac{25}{634}\) and \(\Im (Q_4) \approx 0.06 < \Im (B_3) = \frac{4}{51}\). By concavity of \(\overset{\frown }{B_2B_3}\), we deduce that \(\overset{\frown }{B_2B_3}\) lies above \(\beta _{2,2}\) over \([\Re (\beta _{2,2}(\cos \frac{3\pi }{19})),\Re (B_3)]\). We conclude that \(\overset{\frown }{B_2B_3}\) lies above \(\beta _{2,1}\cup \beta _{2,2}\).

The yellow quadrilateral \(g_1({{\mathcal {S}}}_y)\). Denote the vertices of this quadrilateral \(C_1, C_2, C_3, C_4\) analogously to what precedes. One computes

$$\begin{aligned} C_1 = \frac{2 + 48 i}{577},\quad C_2 = \frac{1 + 12 i}{145}, \quad C_3 = \frac{1 + i}{2},\quad C_4 = \frac{1 + \sqrt{3} i}{2} = e^{i \frac{\pi }{3}}. \end{aligned}$$

We show that the curvilinear sides \(\overset{\frown }{C_1C_2}\) and \(\overset{\frown }{C_2C_3}\) lie above the lower boundary of \(\mathcal {M}_{19}\).

  • Arc \(\overset{\frown }{C_1C_2}\). Let \(R_1(t)\) be the highest intersection point of the vertical through \(\beta _{2,1}(t)\) with \(\mathcal {C}\left( \frac{i}{24}, \frac{1}{24}\right) \) (the circle containing \(\overset{\frown }{C_1C_2}\)). We have

    $$\begin{aligned}&\Im (R_1(t)) \ge \Im (\beta _{2,1}(t)) \\ {}&\quad \iff f_5(\cos t) := \frac{8 \cos \frac{\pi }{19}}{361} \cos t + \frac{\cos \frac{\pi }{19}}{114} \sqrt{1 - \cos ^2 t} - \frac{6 + 2 \cos \frac{2\pi }{19}}{361} \ge 0. \end{aligned}$$

    Since \(f_5\) is non negative on [0.8, 1], it ensues that \(f_5(\cos t)\) is non negative for t in [0, 0.6]. This interval contains \([\frac{\pi }{19}, \frac{3 \pi }{19}]\). As \(\Re (z_2) \approx 0.002 < \Re (C_1) = \frac{2}{577}\) and \(\Re \left( \beta _{2,1}\left( \frac{3\pi }{19}\right) \right) \approx 0.02 > \Re (C_2) = \frac{1}{145}\), we conclude that \(\overset{\frown }{C_1C_2}\) lies entirely above \(\beta _{2,1}\).

  • Arc \(\overset{\frown }{C_2C_3}\). First we note that if a point z with \(\Re (C_3) \le \Re (z) \le \frac{1}{2}\) belongs to \(\mathcal {B}\left( \frac{1}{2}, \frac{1}{2}\right) \) then it lies below the arc \(\overset{\frown }{C_2C_3}\). Thus to show that \(\beta _{i, j}(t)\) is below \(\overset{\frown }{C_2C_3}\) it is sufficient to show that \(|\beta _{i, j}(t) - \frac{1}{2}| \le \frac{1}{2}\). We have

    $$\begin{aligned} |\beta _{2,1}(t) - \frac{1}{2}| \le \frac{1}{2} \iff f_6(\cos t):= \frac{15 \sin \frac{2\pi }{19}}{361 \sin \frac{\pi }{19}} \cos t + \frac{\sin ^2\frac{2\pi }{19}}{361 \sin ^2\frac{\pi }{19}} - \frac{34}{361} \le 0. \end{aligned}$$

    Since \(f_6\) is non positive on \([-1, 1]\), it follows that \(f_6(\cos t)\) is always non positive. Hence \(\overset{\frown }{C_2C_3}\) is above \(\beta _{2,1}\) over \(\left[ \Re (C_2), \Re \left( \beta _{2,1}\left( \frac{3\pi }{19}\right) \right) \right] \). Next, we show that \(\overset{\frown }{C_2C_3}\) lies above \(\beta _{2,2}\) in the strip \(\{z\mid \Re (z)\in {[}\Re \left( \beta _{2,1}\left( \frac{3\pi }{19}\right) \right) , \Re (z_7)]\}\). Let \(R_{2,7}\) be the point on \(\beta _{2,2}\) with real part \(\Re (z_7)\). We have \(R_{2,7} = \beta _{2,2}\left( \tau _{2, 7}:= \frac{(-5 + \cos \frac{5\pi }{38} \cot \frac{\pi }{19})\sin \frac{\pi }{19}}{\sin \frac{2\pi }{19}}\right) \) and we verify that \(\tau _{2, 7} \in \Big [\cos \frac{16 \pi }{19},\cos \frac{3 \pi }{19}\Big ]\) and \(\Im R_{2,7} \approx 0.24 > \Im z_7 \approx 0.18\). Then, to show that \(\overset{\frown }{C_2C_3}\) lies above \(\beta _{2,2}\) in the above strip, it suffices by concavity of \(\mathcal {B}\left( \frac{1}{2}, \frac{1}{2}\right) \) (as \(\beta _{2,2}\) is a line segment) to show that \(\beta _{2,2}(\cos \frac{3\pi }{19}), R_{2,7} \in \mathcal {B}\left( \frac{1}{2}, \frac{1}{2}\right) \). We indeed compute: \(|\beta _{2,2}(\cos \frac{3\pi }{19}) - \frac{1}{2}|^2 \approx 0.23 < \frac{1}{4}\), \(|R_{2,7} - \frac{1}{2}|^2 \approx 0.23 < \frac{1}{4}\). It remains to show that the lower boundaries of \(M_{19, 7}\) and \(M_{19, 13}\) lie below \(\overset{\frown }{C_2C_3}\) in the strip \(\{z\mid \Re (z)\in [\Re (z_7), \Re (C_3)=1/2]\}\). We have

    $$\begin{aligned}&\Big |\beta _{7,1}(t) - \frac{1}{2}\Big |^2 = |\beta _{7,3}(t) - \frac{1}{2}|^2 \le \frac{1}{4}\\ {}&\quad \iff f_7(t) := \frac{\cos ^2\frac{5\pi }{38}}{361 \sin ^2\frac{\pi }{19}} t^2 + \frac{5 \cos \frac{5\pi }{38}}{361 \sin \frac{\pi }{19}} t - \frac{84}{361} \\ {}&\qquad + \frac{\cos ^2\frac{\pi }{19} \cos \frac{\pi }{19} \sin \frac{5\pi }{38} + \sin ^2\frac{5 \pi }{38}}{361 \sin ^2\frac{\pi }{19}} \le 0. \end{aligned}$$

    As \(f_7\) is non positive on the interval \([-1, 1]\), which contains the domains of \(\beta _{7,1}\) and \(\beta _{7,3}\), we deduce that these two curves lie entirely below \(\overset{\frown }{C_2C_3}\). We then have

    $$\begin{aligned} |\beta _{7,2}(t) - \frac{1}{2}|^2 \le \frac{1}{4} \iff f_8(\cos t):= \frac{5 \cos \frac{5\pi }{38}}{361 \sin \frac{\pi }{19}}\cos t - \frac{84}{361} + \frac{\cos ^2\frac{5\pi }{38}}{361 \sin ^2\frac{\pi }{19}} \le 0. \end{aligned}$$

    Since \(f_8\) is non positive on \([-1, 1]\), it follows that \(f_8(\cos t)\) is non positive for all t, which shows that \(\beta _{7,2}\) lies below \(\overset{\frown }{C_2C_3}\). Finally, as previously noticed, since \(\beta _{13,1}\) is a line segment, it suffices to show that its extremities lies below \(\overset{\frown }{C_2C_3}\) by concavity. We compute: \(|z_{13} - \frac{1}{2}|^2 \approx 0.244 < \frac{1}{4}\) and \(|w_{13} - \frac{1}{2}| \approx 0.1 < \frac{1}{4}\). Thus \(\overset{\frown }{C_2C_3}\) lies above \(\beta _{13,1}\), and as\(\Re (w_{13}) \approx 0.502 > \frac{1}{2} = \Re (C_3)\), we deduce that \(\overset{\frown }{C_2C_3}\) is included in \(M_{19}\).

This ends the proof of Lemma 3.9.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lazarus, F., Tallerie, F. A Universal Triangulation for Flat Tori. Discrete Comput Geom 71, 278–307 (2024). https://doi.org/10.1007/s00454-023-00598-7

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00454-023-00598-7

Keywords

Mathematics Subject Classification

Navigation