1: \documentclass[a4paper,11pt]{article}
2: %\linespread{1.1}
3: \usepackage{latexsym}
4: \usepackage[namelimits,intlimits,sumlimits]{amsmath}
5: \usepackage{amsfonts}
6: \usepackage{amssymb}
7: \usepackage{epsfig}
8:
9: \setlength{\oddsidemargin}{-0.04cm}
10: \setlength{\evensidemargin}{-0.04cm}
11: \setlength{\textwidth}{15.5cm}
12: \setlength{\topmargin}{-1.04cm}
13: \setlength{\headheight}{1.0cm}
14: \setlength{\headsep}{0.5cm}
15: \setlength{\textheight}{23.7cm}
16: \setlength{\footskip}{1cm}
17: \setlength{\parskip}{1ex plus0.5ex minus0.2ex}
18: \setlength{\parindent}{0cm}
19:
20: \newtheorem{satz}{Theorem}
21: \newtheorem{lemma}{Lemma}
22: \newtheorem{rem}{Remark}
23: \newtheorem{kor}{Corollary}
24: \newtheorem{prop}{Proposition}
25: \newtheorem{defin}{Definition}
26:
27:
28: \newcommand{\slr}{ $SL(2,\R)$ }
29: \newcommand{\abs}[2]{\parbox{2cm}{#1} \parbox[t]{13cm}{#2} \newline[2mm]}
30:
31: \newcommand{\R}{\mathbb{R}}
32: \newcommand{\C}{\mathbb{C}}
33: \newcommand{\Q}{\mathbb{Q}}
34: \newcommand{\Z}{\mathbb{Z}}
35: \newcommand{\N}{\mathbb{N}}
36: \newcommand{\T}{\mathbb{T}}
37: \renewcommand{\H}{\mathbb{H}}
38: \newcommand{\V}{\mathbb{V}}
39: \newcommand{\CC}{\mathcal{C}}
40: \newcommand{\M}{\mathcal{M}}
41: \newcommand{\F}{\mathcal{F}}
42: \renewcommand{\d}{d}
43:
44: \DeclareMathOperator{\ord}{ord}
45: \DeclareMathOperator{\q}{q}
46: \DeclareMathOperator{\K}{K}
47: \DeclareMathOperator{\h}{h}
48: \DeclareMathOperator{\rpt}{rpt}
49: \DeclareMathOperator{\D}{D}
50: \DeclareMathOperator{\lcm}{lcm}
51:
52:
53: \newenvironment{proof}{{\it Proof}: }{ $\mbox{ }$ \\ $\mbox{ }$
54: \hfill $\square$ \newline}
55: \newenvironment{beweis}{{\it Proof}: }{ $\mbox{ }$ \\ $\mbox{ }$
56: \hfill $\square$ \newline}
57:
58: %\includeonly{corect.tex,corect2.tex}
59: \title{ On the asymptotic quadratic growth rate of saddle connections and
60: periodic orbits on marked flat tori}
61:
62: \author{Martin Schmoll}
63:
64:
65:
66:
67:
68:
69: \begin{document}
70:
71: \maketitle
72:
73: \begin{abstract}
74: Asymptotic quadratic growth rates of saddle connections and families
75: of periodic cylinders on translation tori with $n$ marked points are studied.
76: For any marking the existence of limits of the quadratic growth rate is shown
77: using elementary methods (not Ratners theorem).
78: We study the growth rate limit as function of the marking.
79: Precise formulas for this function in the case of two marked points are given and
80: the sets where the growth rate function is maximal and continuous are described.
81: For rational two markings the index of the Veech group in $SL_2(\Z)$ is
82: calculated in two different ways.
83: \end{abstract}
84: \footnotetext{1991 Mathematics Subject Classification: 37F99, 40E10, 11P21, 51F99}
85: \tableofcontents
86:
87:
88:
89:
90:
91: \section{Introduction}
92: The motivation of this paper is the main Theorem in
93: ``Pointwise asymptotic formulas on flat surfaces" of A. Eskin and H. Masur
94: \cite{esmr98}. They prove that on strata of quadratic differentials $QD(g,P)$,
95: parametrizing half translation surfaces, there exists almost everywhere a limit
96: of the quadratic growth rate and it is constant
97: (with respect to a natural $SL_2(\R)$ invariant Liouville measure $\mu$ on $QD(g,P)$
98: see Veech \cite{ve86}).
99: The following questions arise:\\
100: 1. Does the quadratic growth rate associated to these constants majorize all other
101: quadratic growth rates?\\
102: 2. How can one compute the constants?\\
103: 3a. Are there (sub) spaces of quadratic differentials, where one can prove not only
104: almost everywhere results but pointwise results?\\
105: 3b. And if one has pointwise results, on which sets do the constants take ``different"
106: values?\\
107:
108: Our example the simplest one can find, but it has the advantage that one
109: can compute things elementary. The above stated questions could all be answered
110: in a precise way.\\
111: The paper contains results about the growth rates of cylinders of periodic
112: trajectories or saddle connections on the two dimensional torus $\T^2:=(\R^2/\Z^2,dz)$, $dz=dx+idy$
113: with $n$ marked points. If the number of cylinders of periodic orbits (po)
114: or saddle connections (sc) of length less than $T$ on $\T^2_{[x_0,...,x_{n-1}]}$
115: is denoted by $N_{po/sc}(\T^2_{[x_0,...,x_{n-1}]},T)$ then the existence
116: of the limits
117: \[\lim_{T\rightarrow \infty}\frac{N_{po/sc}(\T^2_{[x_0,...,x_{n-1}]},T)}{T^2}\]
118: for every constellation of marked points i.e. for all $[x_0,...,x_{n-1}]\in {\T^{2n}}$ is proved.
119: \footnote{For all arguments the Parameterspace could be restricted to the case $x_0=0$, because this
120: differs from the general marking only by a translation. All our counting results are invarinat under
121: translations of the marking.}
122:
123: Remark: In the mathematical sense this is not the true parameterspace, because we would like to have
124: all marked points different, but since we make either pointwise statements with respect to one marking
125: or only measure theoretic statements with respect to the Lebesgue measure on this space
126: the difference to a well choosen parameter space plays no role. Moreover all statements are
127: true for any conformal class the tori might have.
128:
129: The existence of the limits is established by counting lattice
130: points of certain lattices in the universal cover $\R^2$ of $\T^2$. As a biproduct
131: of this method we conclude that there is a subset $\T^{2n}_{gp(sc/po)} \subset \T^{2n}$
132: (which depends on what one counts: saddle connections or periodic families)
133: of measure one in the parameter space of all n markings
134: (contained in the set of all non rational n markings)
135: where the above limit takes a maximal value $c^{max}_{sc/po}$. Moreover it is easy
136: to see that if one approaches a point of the set $\T^{2n}_{gp(sc/po)}$ the asymptotic constants
137: of the sequence converge to $c^{max}_{sc/po}$. In other words the map
138: \[[x_0,...,x_{n-1}] \longmapsto \lim_{T\rightarrow \infty}
139: \frac{N_{po/sc}(\T^2_{[x_0,...,x_{n-1}]},T)}{T^2}\]
140: is continous at the points of the set $\T^{2n}_{gp(sc/po)} $. In the case of two marked points
141: and only in this case the set $\T^{2}_{gp(sc/po)} $ does not depend on what is counted (po's or sc's).
142: Without restrictions one can assume that the first marking is $[0]$, then
143: the set of points in general position consists exactly of all the non rational points
144: $\R^2\backslash \Q^2 \bmod \Z^2$.
145: In this case the quadratic growth rate function for saddle connections, assuming $\gcd(p_1,p_2,n)=1$, is given by:
146: \begin{eqnarray}
147: \left[\frac{p_1}{n}, \frac{p_2}{n} \right] &\longmapsto& \frac{6}{\pi}\left
148: (1 + \prod_{p| n\ prime}\left(1-\frac{1}{p^2}\right)^{-1}
149: \sum_{\gcd(i,n)=1}\left(\frac{1}{i^2}-\frac{1}{n^2}
150: \right)\right) \quad \mbox{on } \Q^2 \nonumber\\
151: \left[x,y \right] &\longmapsto& \frac{6}{\pi}+\pi \quad \mbox{for } (x,y) \notin \Q^2
152: \end{eqnarray}
153: The continuity at non rational points implies that the above formula approaches
154: $\frac{6}{\pi}+\pi$ as $(\frac{p_1}{n}, \frac{p_2}{n})$ approximates
155: a non rational number. In contrast to the general for two marked points
156: this continuity is calculated directly by an estimate.
157: Furthermore for the two marked torus we write down the Veech and affine groups for rational markings.
158: An argument shows that every Veech surface coming from a rational marking is
159: a $SL_2(\Z)$ deformation of one of the family $F=\{\T^2_{[\frac{1}{n},0]}: \ n \in \N\}$.
160: Therefore all the Veech groups of rationally marked tori are isomorphic
161: to one of the Veech groups $V_{[\frac{1}{n},0]}$ associated to the family of surfaces $F$.
162: Finally the asymptotic formula of Veech and the asymptotic counting of lattice points
163: in the plane is used to compute the index $[SL_2(\Z):V_{[\frac{1}{n},0]}]$.
164: With the help of the description of the Veech groups for rational 2 markings we
165: rediscover the index of some classical congruence subgroups of $SL_2(\Z)$.
166:
167: It is now well known \cite{ems} that one can also use Ratners Theorem to obtain pointwise results
168: in modulispaces of torus coverings branched over n marked points.
169: Of course from the point of pure existence of the growth rates
170: this is more general and in fact covers the examples described in the paper.
171: On the other hand the results in this paper are explicit and the extremality, or more restrictive
172: the maximality of the almost everywhere constants are proved for the above cases.
173: In this direction one of the main objects of the paper was to see if there is
174: something in between purely rational markings and markings in general position which has not
175: an extremal growth rate. This question also could be answered with ``yes" (see Theorem
176: \ref{tlimes} and the preparatory Lemma \ref{schnigi}).
177: Unfortunately our elementary methods can not directly taken over to prove
178: something on families of branched coverings of tori, but nevertheless with the
179: more sophisticated approach in \cite{ems} one can compute all the Siegel Veech constants
180: for two marked torus coverings which are not Veech. The results in \cite{ems} show,
181: that in this cases the Siegel Veech constants, as functions of the degree
182: of the covering, behave monotonically, too.
183: In the light of Eskins and this results it would be of interest to compute and prove
184: related statements for marked Veech surfaces and their branched coverings.
185:
186: Acknowledgements:\\
187: I have to thank for the hospitality of the CPT in Marseille Luminy in person
188: of Ricardo Lima and his group during the jear '99
189: as well as Alex Eskin and Pascal Hubert for fruitful discussions.
190: I owe thanks to Eugene Gutkin for pointing out an unprecise statement to me
191: and the referee for finding a mistake.
192: This paper is part of my thesis and I like to give many thanks to
193: Tyll Kr\"uger and Serge Troubetzkoy for their continuous encouragement
194: making this work possible.
195: %----------------------------------------------------------------
196:
197:
198:
199:
200:
201: \section{Notation and preparation}
202:
203:
204:
205:
206:
207: %---------------------------------------------------------------
208:
209: For $[x_0],...,[x_n]\in \R^2/ \Z^2$ let $\T^2_{([x_0,...,x_{n-1}])} $ be the translation torus
210: $\T^2=(\R^2/\Z^2, dz=dx+idy)$ marked at the points $[x_0],[x_1],...,[x_{n-1}]\in \T^2$ .
211: If we do not write down the zeroth entry $[x_0]$ in the index vector we assume $[x_0]=0+\Z^2$.
212: Our calculations are restricted to the torus which is isometric to
213: $(\R^2/\Z^2,g:=d\!x^2+d\!y^2)$ with the induced Euclidean metric.
214: With respect to this choice the periodic directions (of the geodesic flow)
215: are exactly the ones with rational slope.
216: The results are not restricted to these special translation tori, they hold for all marked translation tori.
217: In fact all the arguments are invariant under $SL_2(\R)$ deformations of the marked tori and any
218: marked translation torus is an $SL_2(\R)$ deformation of the studied ones.
219: The vector $v_s$ {\bf associated} to a saddle connection $s$
220: (or a prime periodic orbit) is the unique vector $v_s$ in $(\R^2, d\!z)$ that
221: maps isometrically to $s$ in $\R^2 \rightarrow \T^2$ under the natural projection map.
222: Recall that in the case of a periodic orbit $s$ the associated vector $v_s$ characterizes
223: a whole cylinder of closed geodesics.
224: A rational two vector like $(\frac{p}{n},\frac{q}{n})$ is always assumed to fulfill the condition
225: $\gcd(p,q,n)=1$.
226: \begin{defin} \nopagebreak $\mbox{}$ \linebreak
227: \begin{itemize}
228: \item[1.] Two points $[x],[y]\in \T^2$ are said to be {\bf relatively rational}, if
229: $x-y \in \Q^2 $.
230: \item[2.]$SC_{(x,y)}$ denotes the set of saddle connections between the points
231: $[x]\in\T^2$ and $[y]\in\T^2$.
232: \item[3.] $l(s)$ is the length of a $s\in SC_{(x,y)}$ and finally if
233: $[w],[z]\in s \in SC_{(x,y)}$ let $l_{s}(w,z)$ be the distance between $[w]$
234: and $[z]$ \bf{measured along $s$}.
235: \end{itemize}
236: \end{defin}
237: Relatively rational is obviously an equivalence relation. Every marking splits up in classes
238: of relatively rational points. Each class defines a marking, eventually of lower order.
239: By the results of Veech \cite{ve89}
240: or Gutkin and Judge \cite{gj97} $\T^2_{[x_0,...,x_{n-1}]}$
241: is a covering of the one marked torus $\T^2_{[x_0]}$ if and only if all marked points
242: $x_i$ are pairwise relatively rational.
243: Thus $\T^2_{[x_0,...,x_{n-1}]}$ is a Veech surface if and only if
244: for all markings are in one and the same marking class
245: and from \cite{gj97} it follows:
246: $V\left(\T^2_{[x_0,...,x_{n-1}]}\right)\subset SL_2(\Q)$, where
247: $V\left(\T^2_{[x_0,...,x_{n-1}]}\right)$ denotes the Veech group of
248: $\T^2_{[x_0,...,x_{n-1}]}$.
249: Any affine diffeomorphism $\phi$ of the unmarked torus $(\T^2,dz)$ is
250: given by an element $(A,[v])\in SL_2(\Z)\oplus \T^2$: $\phi([z])=[Az+v]$.
251: Since every affine diffeomorphism of the marked torus defines one of the unmarked torus
252: by forgetting marked points one obtains:
253: \begin{equation}\label{sl2}
254: V\left(\T^2_{[x_0,...,x_{n-1}]}\right)\subset SL_2(\Z).
255: \end{equation}
256:
257: %\begin{prop}
258: %Let $[x],[y]\in \T^2$ and $s_1,s_2\in SC_{(x,y)}$. Then for all
259: %$p \in s_1 \cap s_2 $:
260: %\[l^{x}_{s_i}(p)=\frac{q_p}{n}l(s_i)\quad \mbox{where}\ q_p \in \Z \
261: %\mbox{with}\ 1\leq q_p \leq n
262: %=\left|\{p: p \in s_1\cap s_2\}\right| \] holds.
263: %The ratio $q_p/n$ does not depend on the choice of the saddle connection $s_1$ or $s_2$,
264: %we call it the ``d relation" (as a shorthand for division relation) of $p$
265: %[w.r.t. $SC_{(x,y)}$].
266: %The distance $d_{s_i}(p_1,p_2)$ between two consecutive intersection points $p_1,p_2$ is
267: %\[d_{s_i}(p_1,p_2)=\frac{1}{n} (l^{x}_{s_i}(p_2)-l^{x}_{s_i}(p_1)) \ \mbox{for } i=1 \mbox{ or }
268: %i=2\]
269: %measured always along $s_1$ or $s_2$.
270: %\end{prop}
271: %\begin{proof}
272: %We take representatives in $\R^2$, $[x]$ is supposed to be represented by $(0,0)$.
273: %The saddle connections $s_i$ are represented by the lines $\tilde{s}_i$,
274: %starting at the origin. Since we have to identify points modulo $\Z^2$,
275: %the statement follows already from the Strahlensatz (see Figure 1) of elementary geometry.
276: %\end{proof}
277: %\begin{center}
278: %\includegraphics[width=9cm]{strahl.eps}
279: %\end{center}
280:
281:
282:
283:
284:
285: \begin{defin}
286: A {\bf lattice} $G$ is a subset of $\R^2$ of the form
287: \[G=(p_1+q_1\Z,p_2+q_2\Z)\quad p_i,q_i \in \R^+\]
288: A {\bf point distribution} is a finite union of lattices.
289: The set of {\bf visible points} $G^V$ of of a discrete subset $G \subset \R^2$ consists
290: of all points $p\in G$ such there is no $0<|\lambda| <1$ and a point
291: $q\in G$ with $q=\lambda p$.
292: \end{defin}
293: In other words: There is no other point of $G$
294: on the line between the origin and the points $\pm p$.
295: For the correct counting of periodic orbits and saddle connections
296: we need further:
297: \begin{defin}
298: If a point distribution $G$ is contained in another point distribution
299: $H$, then the visible points $H^{V_G}$ of $H$ with respect to $G$
300: are those $p\in H^V$ for which there is a $\lambda \neq 0$ such that
301: $\lambda p \in G$.
302: $G \subset H$ is called ${\mathbf H}$ {\bf complete} if $H^{V_G}=G^V$.
303: \\
304: $G$ is called {\bf closed}, if $mG \subset G$ for any integer $m \in \Z \backslash \{0\}$
305: such that there exist two points $x,y \in G$ with $x=my$.
306: \end{defin}
307:
308:
309:
310:
311:
312: Let $[x],[y]\in \T^2$, again we put $x=0$. Then the set of saddle connections
313: $SC_{(x,y)}$ is described by the visible points of the lattice of associated vectors
314: $G_{(x,y)}=y+\Z^2$.\\
315: For a discrete subset $G \subset \R^2$ (in general point distributions) we define for $T>0$
316: \[N(G^{(V)},T):=card(G^{(V)}\cap B(0,T)).\]
317: Here $B(0,T)$ is the (closed) disc of radius $T$ centered in the origin of $\R^2$.
318: %\begin{lemma}\label{schnigi}
319: % Let $[x],[y],[z]\in \T^2$ be different points
320: %\begin{description}
321: %\item $[z]\in s \in SC_{(x,y)}$ is an intersection point of two saddle connections,
322: %if and only if $l^{x}_{s}(z)\in \frac{p}{q} l(s) \ 1\leq p\leq q\quad p,q\in \N $
323: %\item If $[z]$ is the intersection point of two saddle connections in $SC_{(x,y)}$
324: %then the following relation for the lattices of vectors holds
325: %\[\frac{q}{p}G_{(x,z)} \cap G_{(x,y)}= qG_{(x,y)}\]
326: %where $p/q$ is the d relation of $z$ and $qG_{(x,y)}:=y+q\Z^2$.
327: %\end{description}
328: %\end{lemma}
329: %\begin{proof}
330: %Because of Proposition (1) it is enough to show: if
331: %\[ [z]\in s \in SC_{(x,y)}\ \mbox{ and }
332: %\ l^{x}_{s}(z)\in \frac{p}{q} l(s) \quad 1\leq p\leq q\quad p,q\in \N \]
333: %then there exists some other saddle connection
334: %$s \in SC_{(x,y)}$, passing through $[z]$. This follows in fact from the second statement,
335: %which we prove now.
336: %By assuming $x=0$ we have $z=\frac{p}{q}y$ and therefore the condition that
337: %$z+(k,l)$ with $(k,l)\in \Z^2$ lies on a saddle connection represented by
338: %$y+(m,n) $ with $\ (m,n)\in \Z^2$ is:
339: %\[z+(k,l)=\frac{p}{q}y+(k,l)=\lambda(y+(m,n))\quad \mbox{with}\ 0<\lambda<1.\]
340: %Solutions of this equation which are in $G_{(x,y)}=y+\Z^2$ are exactly
341: %given by $\lambda =\frac{p}{q}$
342: %together with pairs of $(m,n)$ which are multiples of $q$, so we have:
343: %\[G_{(x,z)}\cap\frac{q}{p}G_{(x,y)}= \frac{p}{q}y+p\Z^2=pG_{(x,z)}.\]
344: %The statement follows by multiplying with $\frac{q}{p}$.
345: %\end{proof}
346:
347: Given a nonrational (marked) point $[y]$ on the marked torus $\T_{[0]}^2$.
348: Take a saddle connection $s_{y}\in S_{(y)}:=S_{(0,y)}$ from the origin to $[y]$.
349: It might be represented in $\R^2$ by $s_{y}=y +(m_y,n_y)$ with two fixed integers
350: $m_y,n_y$ and an $y$ choosen from the set $[0,1)^2$.
351: Assume we take a second point $[x]\in \T_{[0]}^2$ $x \in [0,1)^2$, different from $[y]$
352: so, that there is a saddle connection $s_{x}\in S_{(x)}$,
353: which fulfills an equation
354: \begin{equation}\label{relation}
355: s_{x}=x +(m_{x},n_{x})=r s_{y},
356: \ \mbox{with }\ r \in (0,1) \mbox{ and } n_x, m_x \in \Z.
357: \end{equation}
358: This means after mark the point $[x]$ the saddle connection $s_{x}$ destroys
359: $s_{y}$ and generates two new saddle connections, $s_{x}$ and one in the set $S_{(x,y)}$.
360: It is clear by equation (\ref{relation}), that
361: if the factor $r$ is a rational number, say $r=p/q$, all saddle
362: connections represented by $y +(m_y,n_y) + q(k,l)$, $k,l \in \Z$ are divided by the marking
363: $x$ to give saddle connections in $S_{(x,y)}$ and $S_{(x)}$.
364: Now assume $[x]$ rationally divides two saddle connections from $S_{(y)}$ with the equations:
365: \begin{eqnarray}
366: x +(m_1,n_1)=\frac{p_1}{q_1}(y+(k_1,l_1))\\
367: x +(m_2,n_2)=\frac{p_2}{q_2}(y+(k_2,l_2))
368: \end{eqnarray}
369: with integers $k_i,l_i,m_i,n_i,p_i,q_i$, $i=1,2$ and $x,y$ as above.
370: Subtract the two equations to obtain:
371: \[(m_1,n_1)-(m_2,n_2)= \left( \frac{p_1}{q_1}-\frac{p_2}{q_2} \right)y +\frac{p_1}{q_1}(k_1,l_1)-\frac{p_2}{q_2}(k_2,l_2)\]
372: Since $y$ is non rational $ \frac{p_1}{q_1}=\frac{p_2}{q_2}$.
373: Assuming this we have:
374: \[(m_1,n_1)-(m_2,n_2)= \frac{p_1}{q_1}((k_1,l_1)-(k_2,l_2)).\]
375: This equation has solutions if $(k_1,l_1)-(k_2,l_2)$ is (componentwise) divisible by $q_1$.
376: Thus the saddle connections in the set $S(x)$ dividing the ones
377: from $S(y)$ rationally, all have a common stretch factor $p_1/q_1$. Hence, if there is
378: more than one rationally dividing saddle connection, then they are all
379: given by the (lattice) set of vectors
380: \[\left\{\frac{p_1}{q_1}(y+(k,l)):\ (k,l)\equiv (k_1,l_1)\mod{q_1} \right\} \]
381:
382: If $[y]$, $p_1,q_1$ are fixed the different sets of rationally dividing saddle connections
383: for the stretch factor $p_1/q_1$ are parameterized by $(k_1,l_1) \in \{0,...,q_1 -1\}^2$.
384: It is now clear that $S_{(x,y)}$ contains a lattice.
385:
386: \begin{figure}
387:
388: \epsfxsize=7truecm
389:
390: \centerline{\epsfbox{strahl2.eps}}
391: \caption{shows the `rational' intersection points of two saddle connections from $p$ to $y$ with the same horizontal
392: coordinates.}
393: \end{figure}
394: To make shure there are not too much
395: parallel saddle connections in $S_{(x)}$ and $S_{(y)}$ having non rational length ratio, consider the set
396: $S^{nr}_{(x,y)} \subset S_{(x,y)}$ of saddle connections that are represented by $r\alpha$, where $r \in (0,1)$ is non rational.
397:
398: \begin{prop}\label{countnrat}
399: $N(S^{nr}_{(x,y)},T)= O(T)$.
400: \end{prop}
401: \begin{proof}
402: Assume that the growth rate in question is of bigger than linear order, then there must be an
403: integer $i \in \Z$ where one can find two saddle connections
404: $s_{1},s_{2} \in S^{nr}_{(x,y)}$ represented by the points $s_1=r_1(y+(i,j_1))=r_1s_{1,y}$ and $s_2=r_2(y+(i,j_2))=r_2s_{2,y}$,
405: $s_{1,y},s_{2,y}\in S_{(y)}$, $j_1,j_2 \in \Z$ (we assume $j_1<j_2$).
406: If that would be not the case, for each $i \in \Z$ there could be at most one saddle connection of the
407: form $r(y+ (i,l))$ in $S^{nr}_{(x,y)}$, therefore $N(S^{nr}_{(x,y)},T)\leq 2T$ contradicting our assumption.
408: Now by the Strahlensatz of elementary geometry (see figure 1) each intersection point of the saddle connections
409: $s_{1,y}$ and $s_{2,y}$ as above, must divide the length in rational amounts contradicting the assumption.
410: \end{proof}
411:
412:
413: \begin{defin} Let $SC_{(x,y)}$ and $SC_{(x,z)}$ be two sets of saddle connections
414: on the marked tori $\T^2_{[x,y]}$, $\T^2_{[x,z]}$ and suppose $[z]$ is
415: rationally divided by $[y]$.
416: Then $G^{(x,z)}_{(x,y)}\subset G_{(x,y)}$ is the set of vectors associated to
417: saddle connections $S_{(x,y)}$ on $\T^2_{[x,y,z]}$ which are parallel to vectors associated to
418: saddle connections in $S_{(x,z)}$ and shorter in length.\\
419: %Analogously let $G^{\supset(x,y)}_{(x,z)}\subset G_{(x,z)}$ be the lattice of vectors
420: %which are parallel to vectors from $G_{(x,y)}$.
421: \end{defin}
422:
423: To summarize:
424: \begin{lemma}\label{schnigi}
425: Let $[x],[y],[z]\in \T^2$ be different points and $z-x \in [0,1)^2 \notin \Q^2$.
426: If $[y]\in s \in SC_{(x,z)}$, then $G^{(x,z)}_{(x,y)}$ contains a lattice,
427: if and only if there is a rational number $p/q \in \Q$ such that
428: $l_{s}(x,y)= \frac{p}{q} l_{s}(x,z)=\frac{p}{q} l(s)$, moreover the number $p/q$ is unique in the sense that
429: this equation holds for any saddle connection represented by elements of $G^{(x,z)}_{(x,y)}$.
430: \end{lemma}
431: \begin{proof}
432: These are just the statements above translated to the situaion $[x]\neq [0]$.
433: \end{proof}
434:
435:
436: The Lemma is viewed as the generalisation of the following observation on the torus
437: $\T^2=\R^2/\Z^2$ (with the origin as a marked point): a point $p$ is rational if and only if
438: it is in the intersection of saddle connections. If this is the case the
439: saddle connections through $p$ give rise to a lattice (by taking representatives in $\R^2$).
440:
441: The moral of the Lemma is: when counting saddle connections for general markings,
442: all markings which are lying on some saddle connection and dividing the length
443: of it rationally will cause a lattice of intersection points.
444: This is important as we will see later, this decreases the asymptotic constant of that marking.
445: Because of this observation we also have to take care about the definition of the set of
446: markings in $\T^{2n}$, where we assume the growth rates to be extremal.
447: Thus for saddle connections we define the set of markings `in general position' (shortly: `gp')
448: in $\T^{2n}$ inductively:
449: \begin{equation}
450: \begin{array}{l}
451: \T^{4}_{gp(sc)}=\{[x_0,x_1]\in \T^{4}: \ x_1-x_0 \notin \Q^2 \}\\
452: \mbox{ } \\
453: \T^{2(n+1)}_{gp(sc)}=\left\{
454: \begin{array}{l}
455: [x_0,...,x_n]\in \T^{2(n+1)}:[x_0,...,x_{n-1}]\in \T^{2n}_{gp(sc)},[x_n] \in \T^2 \ \mbox{and} \\
456: \forall \ i \in \{ 0,...,n \} \ x_i \mbox{ does not divide any saddle connection}\\
457: \mbox{defined by the marking }[x_0,...,x_{i-1},x_{i+1},...,x_{n}]
458: \mbox{ rationally}
459: \end{array}
460: \right\}.
461: \end{array}
462: \end{equation}
463:
464: Thus the set of non extremal n markings is a set of full Lebesque measure
465: on $\T^{2n}$ (see Theorem \ref{tlimes}).
466: For periodic orbits the set of markings in general position is defined in the following way:
467: \begin{equation}
468: \T^{2n}_{gp(po)}=\{[x_1,...,x_n]:\ x_j-x_i \notin \Q \ \forall \ i,j \in \{0,...,n\} \mbox{ and } j \neq i \}.
469: \end{equation}
470: Observe that for all $n \in \N$ $\T^{2n}_{gp(sc)}\subseteq \T^{2n}_{gp(po)}$. The sets are equal
471: in the cases $n=1$ (trivial) and $n=2$, for more marked points the inclusion is strict.
472: It will turned out the markings in general position are exactely the sets of markings on which the growth rates are maximal.
473: Note that in the case of two marked points the set of markings in general position is the same as the
474: set of non rational markings, always assuming one of the markings to be $[\vec{0}]$.
475: It should be noted here, that the pure existence of the growth rate functions
476: for any choice of marking could also be seen using the Siegel Veech Formula together with
477: Ratners Theorem for ergodic measures on homogeneous spaces (see \cite{ems}).
478: The more elementary argumentation here and in the chapter on general markings shows,
479: that even for complicated markings one has to count only simple lattice
480: point distributions to obtain the quadratic growth constant.
481: %-----------------------------------------------------------------------------
482:
483:
484:
485:
486:
487: \section{Quadratic growth rates of lattices and point distributions in $\R^2$}
488:
489:
490:
491:
492:
493: %-----------------------------------------------------------------------------
494: To compute quadratic growth rates we recall some preparatory Lemmas.
495: The first is quite well known and uses an argument going back to Gauss.
496: %If $G \subset \R^2$ (in general a lattice) we define for $T>0$
497: %\[N(G^{(V)},T):=|G^{(V)}\cap B(0,T)|.\]
498: \begin{lemma}\label{gauss}
499: Let $G=\{ (x,y)\in \R^2: (x,y)=(p_1 +q_1 \Z,p_2 +q_2\Z) \}$.
500: Then \[N(G,T)= card(G\cap B(T,0)) = \frac{\pi}{q_1q_2}T^2 + o(T^2).\]
501: \end{lemma}
502: \begin{proof}
503: To begin we treat the case $p_1=p_2=0$. By drawing horizontal and vertical lines
504: through the lattice points, we got a decomposition of the plane in rectangles with
505: sidelengthes $(q_1,q_2)$. Every lattice point of $G$ can be viewed
506: as the upper right vertice of one rectangle. So the number of points of $G$ in the
507: concentric circle $B(0,T)$ of radius $T$ is determined by the number of rectangles
508: in that circle up to a certain correction term coming from the fact that a rectangle might
509: be only partly inside the ball. But after increasing the radius of the ball
510: by the length of the diagonal of the rectangles $\sqrt{q^2_1+q^2_2}$ the estimate
511: \[N(G,T)\leq \frac{Vol\left(B\left(0,T+\sqrt{q^2_1+q^2_2}\right)\right)}{q_1q_2}\]
512: holds. And by decreasing the radius with the same amount we have:
513: \[\frac{Vol\left(B\left(0,T-\sqrt{q^2_1+q^2_2}\right)\right)}{q_1q_2})\leq N(G,T)\]
514: Putting together these two estimates leds to:
515: \[ \left|N(G,T)-\frac{\pi}{q_1q_2}T^2\right|\leq \frac{\pi}{q_1q_2}
516: \left(2T\sqrt{q^2_1+q^2_2} +q^2_1+q^2_2\right)\]
517: In the general case we have the estimate:
518: \begin{eqnarray}
519: \left|(G-(p_1,p_2))\cap B\left(0,T-\sqrt{p^2_1+p^2_2}\right)
520: \right| \leq \left|G \cap
521: B(0,T)\right| \leq \nonumber \\
522: \leq \left|\left(G-(p_1,p_2)\right)\cap B\left(0,T+\sqrt{p^2_1+p^2_2}\right)\right|
523: \end{eqnarray}
524: the statement follows easily.
525: \end{proof}
526: Before stating the next proposition let us make a general remark on our strategy of counting.
527: If our interest is for example counting prime closed geodesics, we associate a certain point
528: distribution in $G \subset \Z^2$ to all geodesics in question.
529: This point distribution $G$ does not reflect
530: a priory the prime geodesics, but it will be choosen in a way that its quadratic asymptotics
531: as in the last lemma can be calculated easily. The points in $G$ which are associated to
532: prime geodesics will be the visible points of $G$ with respect to $\Z^2$.
533: The price we have to pay is that we cannot relate the distribution of visible points and
534: all points of (the completed and closed) $G$ just by the factor $\frac{6}{\pi^2}$.
535: We have to choose our multiplicities with respect to $G$:
536: \begin{prop}\label{vielf}
537: Let $G\subset\Z^2$ be a $\Z^2$ complete and closed lattice.
538: Let $K\subset \Z \backslash \{0\}$ be the set of numbers $l$ such that
539: $lG^{V}\subset G$.
540: %Further assume that the limit
541: %\[\lim_{T \rightarrow \infty} \frac{N(G^{V},T)}{T^2} \]
542: Then the relation
543: \begin{equation}
544: N(G^{V},T)=N(G,T)\left(\sum_{l \in K}\frac{1}{l^2}\right)^{-1} + o(T^2)
545: \end{equation}
546: holds between the numbers of points of $G$ and the number of visible points $G^V$ of $G$.
547: For the lattice $G=\Z^2$ this implies:
548: \begin{equation}
549: N(\Z^2,T)=N({\Z^2}^{V},T)\frac{3}{\pi^2} + o(T^2).
550: \end{equation}
551: \end{prop}
552: \begin{proof}
553: First let us check that there is no $ x \in G$ that is not an integer multiple of some
554: $y \in G^V$. Assume there is, then there exists a $z \in G^V$ and a rational number
555: $\frac{p}{q}$ with $x=\frac{p}{q}z$. So $q$ divides $z$, a contradiction to the
556: $\Z^2$ completeness of $G$. Since $G$ is closed and every element in $G$ is
557: some $K$-multiple of an element in $G^V$, thus:
558: \begin{eqnarray} \label{factor}
559: N(G,T)= \sum_{l\in K}N\left(G^{V},\frac{T}{l}\right) +o(T^2) =
560: \left(\sum_{l\in K}\frac{1}{l^2}\right)N(G^{V},T)+o(T^2).
561: \end{eqnarray}
562: The lattice $G=\Z^2$ is obviously $\Z^2$ complete, closed and $K=\Z\backslash \{0 \}$, thus
563: \begin{eqnarray}
564: N(\Z^2,T)&=&\left(\sum^{\infty}_{\substack{l=-\infty \\ l\neq 0}}
565: \frac{1}{l^2}\right)N({\Z^2}^{V},T)+o(T^2) = \nonumber \\
566: &=&2\left(\sum^{\infty}_{l=1}\frac{1}{l^2}\right)N({\Z^2}^{V},T)+o(T^2)=
567: \frac{\pi^2}{3}N({\Z^2}^{V},T)+o(T^2). \nonumber
568: \end{eqnarray}
569: \end{proof}
570: Counting only a certain percentage $p$ of the length of each line
571: in a point distribution $G$ is equivalent to counting the length of
572: vectors in the point distribution \\ $pG:=\{pv: v \in G \}$. Thus there is a relation:
573: \begin{equation}\label{scaling}
574: N(pG,T) = N\left(G,\frac{T}{p} \right)= \frac{1}{p^2}N(G,T)+o(T^2).
575: \end{equation}
576: Remark: As is seen above the relating factor between cylinders of geodesics
577: and prime geode\-sics will be $\frac{3}{\pi^2}$ instead of $\frac{6}{\pi^2}$ because we don't
578: distinguish between the two directions a prime geodesic might have. Our decision to do that
579: is based on the point of view : count only (the length of) the geometric object.
580: \begin{defin} Define the set of lattices parametrized by $ p_1,q_1,p_2,q_2 \in \Z $, with
581: $\gcd(p_i,q_i)=1$ as:
582: \[G_{(q_1,p_1,q_2,p_2)}:=\left(q_1 \Z+ p_1,q_2 \Z+ p_2\right) \subset \Z^2.\]
583: Further for $n \in \N$
584: and a divisor $m$ of $n$ define $I^n_m$ to be the set of numbers which are
585: relatively prime to $\frac{n}{m}$
586: \[I^n_m= \left\{j\in \left\{1,2,...,\frac{n}{m}-1\right\}: \gcd \left(j,\frac{n}{m}\right)=1\right\}\]
587: (if $m=1$ we write shortly $I^n$ instead of $I^n_m$) and the point distributions
588: \[G_{I^n_m}:=\bigcup_{k\in I^n_m}G_{(n,mk,n,0)}\subset \Z^2.\]
589: \end{defin}
590: With this conventions
591: \begin{lemma}\label{pvert}
592: The point distributions $G_{I^n}$ are for all $n \in \N$ complete with respect to $\Z^2$
593: and closed. We have:
594: \begin{equation}\label{vf}
595: N(G_{I^n},T)=2N(G^{V}_{I^n},T)\left[\sum_{k\in I^n}
596: \sum^{\infty}_{l=1 }\frac{1}{(nl+k)^2}\right].
597: \end{equation}
598: As a consequence, for all prime numbers $p$
599: \begin{equation} \label{primvf}
600: N(G_{I^p},T)=N(G^{V}_{I^p},T)\frac{\pi^2}{3}\left(1-\frac{1}{p^2}\right).
601: \end{equation}
602: \end{lemma}
603: \begin{proof}
604: First one has to find all integers which map $G_{I^n}$ under elementwise
605: multiplication to $G_{I^n}$ itself.
606: For $l \in \Z$ and $p\in G_{I^n}$ we have $lp\in G_{I^n_m}$, if and only if
607: \[l \equiv r \mod{n} \quad , \mbox{with } r \in I^n.\]
608: Therefore $l$ has to be in $ n\Z+r \quad \mbox{with }\ r \in I^n$.
609: Obviously the multiplication with a number from this set maps points of $G_{I^n}$ again to
610: $G_{I^n}$.
611: So if the point distributions are complete with respect to $\Z^2$ they are also
612: closed. To see the last statement take $l\in \Z \backslash \{0\}$
613: and choose $(p_1,p_2)\in \Z^2$ with $l(p_1,p_2) \in G_{I^n}$.
614: This means $l(p_1,p_2) = (kn+r,mn)$ with $r \in I^n$ and $k,m \in \Z$.
615: The equation in the first coordinate says that $l$ and $p_1$ are relatively prime
616: to $n$ and from this it follows that $p_2$ has to be a multiple of $n$. Thus $(p_1,p_2)$
617: is already contained in $G_{I^n}$ and the completeness follows.
618: Using equation (\ref{factor}) from Proposition \ref{vielf}
619: gives the first part of the Lemma.
620: For the second part we simply observe that because $p$ is prime $I^p=\{1,2,...,p-1\}$:
621: \[\sum_{k\in I^p} \sum^{\infty}_{l=1 }\frac{1}{(pl+k)^2}=
622: \sum^{\infty}_{l=1 }\frac{1}{l^2}-\sum^{\infty}_{l=1 }\frac{1}{(pl)^2}\]
623: \end{proof}
624: Remark: Obviously the lattices $I^{n}_m$ are for $m\neq 1$ never $\Z^2$ complete.
625: %Rigorously viewed the next Lemma is not necessary to develop the theory
626: %(and of course it is well known):
627: %\begin{lemma}
628: %\begin{equation}\label{sin}
629: %\frac{\pi^2}{n^2 \sin^2(\frac{\pi}{n})}=\sum^{\infty}_{l=1 }\frac{1}{(nl+1)^2}
630: %\end{equation}
631: %\end{lemma}
632: %\begin{proof}
633: %We assume the product development
634: %\[\sin(\pi z)=\pi z \prod^{\infty}_{l=1}\left(1-\frac{z}{l}\right)
635: %\left(1+\frac{z}{l}\right)\]
636: %as well known. With a logarithmic derivative we get
637: %\[\pi \frac{\cos(\pi z)}{\sin(\pi z)}=\frac{1}{z}+\sum^{\infty}_{l=1}
638: %\left[\frac{1}{z-l}+\frac{1}{z+l}\right]\]
639: %Now we take the derivative with respect to the variable $z$ and
640: %evaluate the result at the point $z=\frac{1}{n}$. The statement follows.
641: %\end{proof}
642: %A simple consequence of this identity is:
643: %\begin{eqnarray}\label{sinaf}
644: %\sum^{n-1}_{k=1} \frac{1}{\sin^2(\frac{k\pi}{n})}&=&\frac{1}{\pi^2}\sum^{n-1}_{k=1}
645: %\sum^{\infty}_{l=1}\frac{1}{\left(l+\frac{k}{n}\right)^2}
646: %=\frac{1}{\pi^2}\left[\sum^{\infty}_{l=1 }\frac{n^2}{l^2}-
647: %\sum^{\infty}_{l=1}\frac{1}{l^2}\right]=\nonumber\\ &=&\frac{1}{3}(n^2-1)
648: %\end{eqnarray}
649: Before we compute the relating factor of equation (\ref{vf}) in general, an
650: observation is helpful: Let ${\chi }_{0}: \Z \rightarrow \Z/2\Z$ be the principal
651: character modulo $n$,
652: that is ${\chi}_0(l)=1$ if $\gcd(l,n)=1$ and is equal to $0$ if not, then obviously:
653: \[\sum_{k\in I^n} \sum^{\infty}_{l=1 }\frac{1}{(nl+k)^2}=
654: \sum^{\infty}_{l=1 }\frac{{\chi}_{0}(l)}{l^2}= L({\chi}_0,2)\]
655: where $L({\chi}_0,2)$ is the $L$ series with respect to ${\chi}_0$ evaluated at $2$.
656: The following is well known
657: (see for example exercise (3.d) page 166 in \cite{ten} or Serre \cite{se})
658: \begin{prop} \label{sumform}
659: If $s$ is a complex number with real part bigger than $1$
660: then:
661: \begin{equation}
662: L({\chi}_0,s)=\sum^{\infty}_{l=1 }\frac{\chi_{0}(l)}{l^s}=\zeta(s)\prod_{p|n\ prime}
663: \left(1-\frac{1}{p^s}\right)
664: \end{equation}
665: \end{prop}
666: \begin{proof}
667: The condition on the real part of $s$ guaranties absolute convergence
668: of the series so that the following manipulations are justified.
669: We start with the right hand side:
670: \begin{eqnarray}
671: \zeta(s)\prod_{p|n\ prime}\left(1-\frac{1}{p^s}\right)=
672: \sum^{\infty}_{l=1}\frac{1}{l^s}\left(\sum_{m|n}\frac{\mu(m)}{m^s}\right)= \nonumber \\
673: =\sum^{\infty}_{k=1}\sum_{m|\gcd(k,n)}\frac{\mu(m)}{k^s} =
674: \sum^{\infty}_{k=1}\frac{1}{k^s}\left(\sum_{m|\gcd(k,n)}\mu(m)\right)=
675: \sum^{\infty}_{k=1 }\frac{{\chi}_{0}(k)}{k^s}
676: \end{eqnarray}
677: We have used the well known M\"obius $\mu$ function, the formula
678: \[\sum_{d|n}\mu(d)=\left\{ \begin{array}{l}
679: 1 \mbox{ if } n=1\\
680: 0 \textrm{ if } n>1
681: \end{array} \right. \]
682: and summation over $k=lm$ to reorder the sum in a useful way.
683: \end{proof}
684: Evaluating the above $L$ series for $s=2$ we find
685: \begin{equation}
686: N(G_{I^n},T)=N(G^{V}_{I^n},T)\frac{\pi^2}{3}\prod_{p|n \ prime}
687: \left(1-\frac{1}{p^2}\right)+o(T^2).
688: \end{equation}
689: The left hand side of this equation is easy to compute with the help of
690: Lemma \ref{gauss} and the definition of $G_{I^n}$
691: \begin{eqnarray}
692: N(G_{I^n},T)=\frac{\pi}{n^2}|I^n|T^2+o(T^2).
693: \end{eqnarray}
694: Use Eulers $\varphi$ function to identify
695: \begin{equation}\label{in}
696: |I^n|=\varphi(n)=n\prod_{p|n \ prime}\left(1-\frac{1}{p}\right)
697: \end{equation}
698: and finally write
699: \begin{eqnarray}\label{vffinal}
700: \lim_{T \rightarrow \infty}
701: \frac{N(G^{V}_{I^n},T)}{T^2}&=&\frac{3}{\pi}\frac{\varphi(n)}{n^2}\prod_{p|n \ prime}
702: \left(1-\frac{1}{p^2}\right)^{-1}=
703: \nonumber\\
704: &=&\frac{3}{n\pi}\prod_{p|n \ prime}\left(1+\frac{1}{p}\right)^{-1}.
705: \end{eqnarray}
706: For later use we show
707: \begin{prop}\label{completion}\hfill \\
708: $G_{I^n}$ containes the $\Z^2$ completion of the lattice
709: $G_{(n,r,n,0)}=$$\{(n\Z+r,n\Z):\ r \in I^n \} $.
710: \end{prop}
711: \begin{proof}
712: It is enough to show that a multiple of each point $(kn+l,mn) \in G_{I^n}$
713: is in $G_{(n,r,n,0)}$. That is we have to solve $rx \equiv l \pmod{n}$.
714: But this is always possible since $r$ and $l$ are relatively prime with respect
715: to $n$.
716: \end{proof}
717: %----------------------------------------------------------------------
718:
719:
720:
721:
722:
723: \section{The Veech group of $\mathbf{\T^2_{\left[\frac{p_1}{q_1},\frac{p_2}{q_2}\right]}}$ and its
724: index in $\mathbf{SL_2(\Z)}$}
725:
726:
727:
728: %----------------------------------------------------------------------
729: We now describe the affine group $A\mathit{ff}(\T^2_{x})$ for a rational point $x\in \Q^2$.
730: \begin{prop} \label{afftor}
731: Assume $(\frac{1}{2}, \frac{1}{2}),(0, \frac{1}{2}), (\frac{1}{2}, 0)\neq x\in \Q^2$.
732: Then
733: \[A\mathit{ff}(\T^2_{x})=\left\{z \mapsto Az+v:\! A\in SL_2(\Z) \mbox{ with \,}\left\{\! \begin{array}{ll}
734: Ax \equiv -x \pmod{\Z^2} & \textrm{ if } v=x\\
735: Ax \equiv x \pmod{\Z^2} & \textrm{ if } v=0
736: \end{array} \right. \!\right\} \]
737: If $y=(\frac{1}{2}, \frac{1}{2}),(0, \frac{1}{2}), (\frac{1}{2}, 0)$ we have
738: \[ A\mathit{ff}(\T^2_{y})=\left\{z \mapsto Az+v: A\in SL_2(\Z) \mbox{ with }
739: \left\{\begin{array}{l}
740: Ay \equiv y \pmod{\Z^2} \\
741: v=y
742: \end{array} \right.\right\} \]
743: \end{prop}
744: \begin{proof}
745: $A\in SL_2(\Z)$ is a consequence of the discussion around equation (\ref{sl2}) in the preparation section.
746: The only thing which is left to prove is the statement on the translation vector $v$.
747: It is defined only modulo $\Z^2$, thus one can assume $v\in[0,1)^2$.
748: The subgroup $N(\T^2_x)$ of pure translations is trivial if
749: $x\neq (\frac{1}{2}, \frac{1}{2}),(0, \frac{1}{2}), (\frac{1}{2}, 0)$:\\
750: Assume the lattice $\Z^2_x=x+\Z^2$ generated is by use of $v$ translated to $\Z^2$ and
751: by the same translation $\Z^2$ is moved to $\Z^2_x$, then $v\equiv x \equiv -x \pmod{\Z^2}$ must hold.
752: Thus either $v$ is trivial, or $x$ is one of the exceptions.
753: In general the same problem comes up if the map should interchange the lattices
754: $\Z^2$ and $\Z^2_x$. The only possibility is: $A$ has to map $\Z^2_x$ to
755: $\Z^2_{-x}$ and $v$ translates this into $\Z^2$. Thus we obtain the conditions:
756: \[Ax \equiv -x \pmod{\Z^2}\ \mbox{ and } v=x \quad \mbox{or}\quad Ax\equiv x
757: \pmod{\Z^2}\ \mbox{ and }v=0\]
758: For the three exceptions there is in fact only one condition,
759: because in these cases $-x \equiv x \pmod{\Z^2}$, thus $-id$ is automatically contained
760: in these affine groups.
761: Moreover the translation $v=x$ is contained in the groups.
762: \end{proof}
763: \begin{satz} \label{kong}The Veech group $V(\T^2_x)$ where
764: $x=\left[\frac{p_1}{q_1},\frac{p_2}{q_2}\right]$
765: \begin{itemize}
766: \item for $ 1\leq \gcd(q_1,q_2)< \min(q_1,q_2) $ is
767: \begin{displaymath}
768: V(\T^2_x)=
769: \left\{\begin{array}{cc} \pm \left(
770: \begin{array}{cc} a & b \\ c & d
771: \end{array} \right) \in SL_2(\Z):&
772: \begin{array}{cc}
773: a\equiv 1 \quad \pmod{q_1} & b \equiv 0 \pmod{q_2} \\
774: c\equiv 0 \quad \pmod{q_1} & d \equiv 1 \pmod{q_2}
775: \end{array}
776: \end{array}
777: \right\}
778: \end{displaymath}
779: \item for $ q_1|q_2 $ (or $ q_2|q_1 $ analogously) is
780: \begin{displaymath}
781: V(\T^2_x)=
782: \left\{ \begin{array}{cc} \pm \left(
783: \begin{array}{cc}
784: a & b \\
785: c & d
786: \end{array} \right) \in SL_2(\Z): &
787: \begin{array}{c}
788: a+\frac{q_1p_2}{q_2}b \equiv 1 \pmod{q_1} \\
789: b \equiv 0 \pmod{\frac{q_2}{q_1}} \\
790: c\equiv 0 \pmod{q_1} \\
791: d \equiv 1 \pmod{q_2}
792: \end{array}
793: \end{array}
794: \right\}
795: \end{displaymath}
796: \item for $ q_1=q_2 $ is
797: \begin{displaymath}
798: V(\T^2_x)=
799: \left\{\!\begin{array}{ll} \pm \left(
800: \begin{array}{cc}
801: a & b \\
802: c & d
803: \end{array} \right) \!\in SL_2(\Z):\!\! &
804: \begin{array}{c}
805: \left(
806: \begin{array}{cc}
807: a & b \\
808: c & d
809: \end{array} \right)
810: \left(\begin{array}{c} p_1 \\ p_2 \end{array}\right)
811: \equiv \left(\begin{array}{cc}
812: 1& 0 \\
813: 0 & 1
814: \end{array} \right)
815: \pmod{q_1}
816: \end{array}
817: \end{array}
818: \! \right\}
819: \end{displaymath}
820: In the case $q_1=2$ that is $(\frac{p_1}{q_1},\frac{p_2}{q_2})=(\frac{1}{2},\frac{1}{2})$ we
821: have to add the rotation:
822: \begin{equation}\nonumber
823: r_{\frac{\pi}{2}}=\begin{pmatrix}
824: _{0} & _{1} \\
825: _{-1} & _{0}
826: \end{pmatrix}
827: \end{equation}
828: \item for $ q_2=0 $ (or equivalently $q_1=0$) is
829: \begin{displaymath}
830: V(\T^2_x)=
831: \left\{\begin{array}{ll} \pm \left(
832: \begin{array}{cc}
833: a & b \\
834: c & d
835: \end{array} \right) \in SL_2(\Z):&
836: \begin{array}{cc}
837: \left(
838: \begin{array}{cc}
839: a & b \\
840: c & d
841: \end{array} \right)& \equiv \left(
842: \begin{array}{cc}
843: 1& b \\
844: 0 & 1
845: \end{array} \right) \pmod{q_1}
846: \end{array}
847: \end{array}
848: \right\}
849: \end{displaymath}
850: \end{itemize}
851: \end{satz}
852: \begin{proof}
853: By the last proposition we have to evaluate the condition
854: \[ \begin{array}{cc}
855: \left(
856: \begin{array}{cc}
857: a & b \\
858: c & d
859: \end{array} \right) \left(
860: \begin{array}{c}
861: \frac{p_1}{q_1}\\
862: \frac{p_2}{q_2}
863: \end{array} \right) \equiv
864: \pm\left(\begin{array}{c}
865: \frac{p_1}{q_1}\\
866: \frac{p_2}{q_2}
867: \end{array} \right) \pmod{\Z^2}
868: \end{array}
869: \] on the coefficents $a,b,c,d \in \Z$ . If for example $p_2=0$
870: then we have
871: \[ \begin{array}{l}
872: \left(
873: \begin{array}{c}
874: \frac{p_1}{q_1}a \\
875: \frac{p_1}{q_1}c
876: \end{array} \right) \equiv \pm \left(
877: \begin{array}{c}
878: \frac{p_1}{q_1}\\
879: 0
880: \end{array} \right) \pmod{\Z^2}
881: \end{array}
882: \]
883: It follows directly that $a\equiv \pm 1 \pmod{q_1}$, $c \equiv 0 \pmod{q_1}$ and
884: $b\in \Z$ is arbitrary. Since the determinant of the matrix is always $1$ we have the condition
885: $ad \equiv 1 \pmod{q_1}$. Therefore $d \equiv \pm 1 \pmod{q_1}$, if $a \equiv \pm 1 \pmod{q_1}$.
886: The other three cases are equally simple, we omit them.
887: \end{proof}
888: We prove now that all 2 marked tori are affine equivalent to one from a
889: special family, if the marking is rational.
890: \begin{prop}\label{reduction} If $x=\left[\frac{p}{n},\frac{q}{n}\right]$ ($\gcd(p_1,p_2,n)=1$),
891: $\T^2_x$ is affine isomorphic to $\T^2_{\left[\frac{1}{n},0\right]}$.
892: More precisely: there exists an affine diffeomorphism of $\T^2$ represented by an element of $SL_2(\Z)$
893: which maps $\left[\frac{p}{n},\frac{q}{n}\right]$ to $\left[\frac{1}{n},0\right]$
894: (and preserves $[0]$).
895: \end{prop}
896: \begin{proof}
897: The Veech group $V(\T^2)$ of the one marked torus is represented by $SL_2(\Z)$.
898: Thus it is enough to show: given a point $[p/n,q/n]\in\T^2$
899: with $\gcd(p,q,n)=1$, there is a linear map in $SL_2(\Z)$ that maps it to the point
900: $[1/n,0]\in\T^2$.
901: Let $c:=\gcd(p,q)$ then the equation $\frac{q}{c}p-\frac{p}{c}q=0$ holds. We want
902: the vector $(-\frac{q}{c},\frac{p}{c})$ to become the second row of a matrix $A\in SL_2(\R)$, thus
903: the vector $(a,b)$ representing the first row has to fulfill $1=\det(A)=a\frac{p}{c}+b\frac{q}{c}$.
904: This equation has solutions $a,b \in \Z$ because $\frac{p}{c}$ and $\frac{q}{c}$ are relatively prime.
905: Taking solutions $a,b$ defines an $A\in SL_2(\R)$ with $A(p,q)=(c,0)$ with $c$ relatively prime to $n$.
906: Hence the congruence $kc\equiv 1 \pmod{n}$ has a solution $k \in \Z$ and one can use the linear Dehn twists
907: \[\begin{array}{cc}
908: \left(
909: \begin{array}{cc}
910: 1 & 0 \\
911: k & 1
912: \end{array} \right) \left(
913: \begin{array}{c}
914: c\\
915: 0
916: \end{array} \right) \equiv
917: \left(\begin{array}{c}
918: c\\
919: 1
920: \end{array} \right)
921: \end{array} \pmod{n}\]
922: and
923: \[\begin{array}{cc}
924: \left(
925: \begin{array}{cc}
926: 1 & -c \\
927: 0 & 1
928: \end{array} \right) \left(
929: \begin{array}{c}
930: c\\
931: 1
932: \end{array} \right) \equiv
933: \left(\begin{array}{c}
934: 0\\
935: 1
936: \end{array} \right)
937: \end{array}\pmod{n} \]
938: to obtain the result up to a rotation of 90 degrees.
939: \end{proof}
940: \begin{kor}\label{isom}
941: The Veech groups $V(\T^2_x) \subset SL_2(\Z)$ for points
942: $x=[\frac{p}{n},\frac{q}{n}]$ with $\gcd(p,q,n)=1$
943: are all isomorphic. The isomorphism is given by conjugation with an element of
944: $SL_2(\Z)$. Moreover the asymptiotic growth rates are the same, if the surfaces are
945: affine isomorphic.
946: \end{kor}
947: \begin{proof}
948: The first two statements follow from the above proposition. Veechs asymptotic formula depends
949: only on the conjugacy class of the Veech group in $SL_2(\R)$ and on the areas of the
950: cylinders of closed geodesics (see for example Vorobetz paper \cite{va}),
951: which are invariant under affine isomorphisms. In other words the asymptotic constants
952: depend only on the $SL_2(\R)$ orbit of the given Veech surface in the moduli space.
953: But Proposition (\ref{reduction}) shows the marked surfaces in question are all on one $SL_2(\R)$ orbit.
954: \end{proof}
955: \begin{rem} Since every rational number $x$ has a unique representation as above,
956: one can restrict the considerations to families of markings defined by
957: \[x_{n}:=\left[\frac{1}{n},0\right]\quad \mbox{or}
958: \quad x_{(n,n)}:=\left[\frac{1}{n},\frac{1}{n}\right]\quad \mbox{where }\ 1\neq n \in \N \]
959: and to the corresponding tori. We choose the family $\T^2_{x_n}$.
960: \end{rem}
961: To compute the index $[SL_2(\Z):V(\T^2_{[\frac{p}{n},\frac{q}{n}]})]$, the first observation is that
962: the only affine map $\phi \in A\mathit{ff}(\T^2_{[\frac{p}{n},\frac{q}{n}]})$ ($\gcd(p,q,n)=1$)
963: which interchanges $[0]$ and $[\frac{p}{n},\frac{q}{n}]$ is given by rotation of 180 degrees
964: and a translation by the vector $(\frac{p}{n},\frac{q}{n})$ (for $n>2$). Since $\phi$ maps
965: $[\frac{p}{n},\frac{q}{n}]$ to $0$ it is of no interest for the following discussion.
966: Up to the subgroup of generated by the idempotent element $\phi$,
967: $A\mathit{ff}(\T^2_{[\frac{p}{n},\frac{q}{n}]})$ can be viewed as the isotropy subgroup
968: $A\mathit{ff}_{[\frac{p}{n},\frac{q}{n}]}(\T^2)\subset A\mathit{ff}(\T^2) $. For a translation structure
969: $u:=(S_{g},\omega)$ and $x\in S_{g}$ one defines the isotropy group $A\mathit{ff}_x(u)$ of $x$
970: and its associated Veech group in the usual way:
971: \begin{eqnarray} &A\mathit{ff}_x(u) := \{g \in A\mathit{ff}(u): gx=x\} \quad \mbox{and}\nonumber \\
972: &V_x(u):=d\circ A\mathit{ff}_x(u)\subset V(u).\nonumber
973: \end{eqnarray}
974: These are subgroups of $A\mathit{ff}(u)$, $V(u)$ respectively.
975: As a direct consequence from the definitions and elementary group theory one obtains:
976: \begin{prop}\label{shindex}
977: Let $u$ be a translation structure and $u_x$ the structure $u$ where the point $x$ is marked.
978: Then
979: \begin{equation}
980: [A\mathit{ff}(u):A\mathit{ff}_x(u)]= \ord(A\mathit{ff}(u)/ A\mathit{ff}_x(u))=
981: |O_{A\mathit{ff}(u)}(x)|
982: \end{equation}
983: and
984: \begin{equation}\label{vindex}
985: [V(u):V_x(u)]=\ord (V(u)/V_x(u))
986: \end{equation}
987: where $O_{A\mathit{ff}(u)}(x):=\{y\in S_{g}: y \in A\mathit{ff}(u).x\}$ is the orbit of $x$
988: under $A\mathit{ff}(u)$ and $G/H$ denotes the set of right cosets $xH$ of the subgroup $H\subset G$.
989: \end{prop}
990: \begin{beweis}
991: See for example Chapter I \S 5 in S. Langs book on algebra \cite{langa}.
992: \end{beweis}
993: Lemma \ref{reduction} above implies together with Proposition
994: \ref{shindex} for $n>2$
995: \begin{eqnarray}[A\mathit{ff}_{[\frac{p}{n},\frac{q}{n}]}(\T^2):A\mathit{ff}(\T^2)]
996: &=&|\{(a,b): 0< a,b \leq n \mbox{ and } \gcd (a,b,n)=1 \}|\nonumber \\
997: &=& n^2\prod_{p|n \ prime}\left(1-\frac{1}{p^2}\right)
998: \end{eqnarray}
999: The last equation on the cardinality of the set of pairs $(a,b)$ is well known.
1000: $d\phi=-id$ is normal in $SL_2(\Z)$ and therefore also in $V(\T^2_{[\frac{p}{n},\frac{q}{n}]})$, thus
1001: \[V_{[\frac{p}{n},\frac{q}{n}]}(\T^2)\cong V(\T^2_{[\frac{p}{n},\frac{q}{n}]})/(-id).\]
1002: Since $-id$ generates a subgroup of order $2$ and the subgroup $N(\T^2_{[\frac{p}{n},\frac{q}{n}]}) \subset
1003: A\mathit{ff}(\T^2_{[\frac{p}{n},\frac{q}{n}]})$ of pure translations is trivial if $n\ne 2$
1004: we have
1005: \[[SL_2(\Z):V(\T^2_{[\frac{p}{n},\frac{q}{n}]})]=\frac{n^2}{2}\prod_{p|n \ prime}
1006: \left(1-\frac{1}{p^2}\right)\]
1007: For $n=2$ the index is $3$ because $-id$ is already an element in
1008: $V_{[\frac{p}{n},\frac{q}{n}]}(\T^2)$.
1009:
1010:
1011: In contrast to this calculation there is a second way to compute the index of the Veech groups
1012: using the asymptotic formula of Veech and the results from the last section.
1013: To begin, the Veech groups are presented as subgroups of $SL_2(\Z)$, but since we do not want
1014: to count orbits in both directions we take the quotient modulo $\pm id$: $PSL_2(\Z)$.
1015: Since both modulo $-id$ equivalent elements are in $V(\T^2_{x_n})$ we can do the same with
1016: the Veech groups and loose no information when taking the preimage in $SL_2(\R)$.
1017: Thus, denote by $V_{p}(\T^2_{x_n})$ the image of $V(\T^2_{x_n})$ in $PSL_2(\R)$.
1018: The orbits $V_{p}(\T^2_{x_{n}})(\pm v_{\CC_i})\in \R^2 ,\ i=1,...,k(x_{n})$
1019: ($k(x_{n})$ is the number of different orbits) of the vectors $\pm v_{\CC_i}$
1020: associated to a maximal cylinder $\CC_i$ of closed geodesics which fills the torus under
1021: $V_{p}(\T^2_{x_{n}})$ give always the same asymptotic constant
1022: (for the formula see Proposition 6.3 in \cite{gj97}:
1023: \[\lim_{T \rightarrow \infty}\frac{N(V_{p}(\T^2_{x_{n}})(\pm v_{\CC_i}),T)}
1024: {T^2}=vol\left(\H/V_{p}(\T^2_{x_{n}})\right)^{-1}
1025: \frac{[A\mathit{ff}(\CC_i):A\mathit{ff}_0(\CC_i)]}{area(\CC_i)}
1026: %=\frac{3}{\pi}[SL_2(\Z):V(\T^2_{x_{n}})]^{-1}.
1027: \]
1028: Since each cylinder $\CC_i$ fills the torus we have $area(\CC_i)=1$ and
1029: $[A\mathit{ff}(\CC_i):A\mathit{ff}_0(\CC_i)]=1$.
1030: Here $A\mathit{ff}_0(\CC_i)$ is the group generated by linear Dehn twists around $\CC_i$.
1031: %Moreover since $x_n$ is rational $\T^2_{x_{n}}$ defines a covering of degree $n$ of a 1-marked torus.
1032: %To see this just tile the torus $\T^2_{x_{n}}$ by rectangles of height $1$ and width $1/n$,
1033: %or more formally use the map $f_n:[x,y]\longrightarrow [nx,y] \mod{1}$.
1034: Thus we can write:
1035: \begin{eqnarray}
1036: vol\left(\H/V_{p}(\T^2_{x_{n}})\right)&=&[PSL_2(\Z):V_{p}(\T^2_{x_{n}})] \
1037: vol\left(\H/PSL_2(\Z)\right)= \nonumber \\
1038: &=&\frac{\pi}{3}[SL_2(\Z):V(\T^2_{x_{n}})].
1039: \end{eqnarray}
1040: The other side of Veech's asymptotic formula is the quadratic growth rate of the lengths of
1041: the maximal cylinders that completely fill the torus. Their directions are exactly the directions of lines
1042: in $\R^2$ which begin at the origin and cross some point of the lattice $x_{n}+\Z^2$.
1043: To get a set of integer points which represent (eventually multiples of the length of)
1044: the closed cylinders in question we simply multiply with $n$ and get the
1045: lattice $G_{(n,1,n,0)}=n\Z^2+(1,0) \subset \Z^2 $ studied earlier. Since we are on the torus $\R^2/\Z^2$
1046: to count the length spectrum of the cylinders correctly one has to look for the
1047: visible points of the completion of $G_{(n,1,n,0)}$ with respect to $\Z^2$.
1048: This completion, by Proposition \ref{completion}, is the point distribution $G_{I^n}$,
1049: thus finally:
1050: \begin{equation}\label{kxn}
1051: \lim_{T\rightarrow \infty}\frac{N(G^{V}_{I^n}, T)}{T^2}=
1052: \frac{3k(x_{n})}{\pi}[SL_2(\Z):V(\T^2_{x_{n}})]^{-1}
1053: \end{equation}
1054: What is left to calculate is $k(x_{n})$ :
1055: \begin{prop}\label{invariante}
1056: \begin{equation}
1057: k(x_{n})=\left\{ \begin{array}{lll}
1058: \frac{1}{2}\varphi(n)=\frac{n}{2}\prod_{p|n \ prime}\left(1-\frac{1}{p}\right)
1059: &\mbox{ if}& n>2 \\
1060: 1 & \mbox{ if} & n=2
1061: \end{array}\right.
1062: \end{equation}
1063: \end{prop}
1064: \begin{proof}
1065: Observe that the ratio of the length of the two saddle connections connecting
1066: $[0]$ and $[x_n]$ and bounding the same cylinder of periodic orbits is an invariant
1067: under the operation of the Veech group. Furthermore modulo $V(\T^2_{x_{n}})$ we
1068: can represent every periodic orbit or saddle connection by a point in
1069: $Q_n=\{(k,l) \in \Z^2 :\ 0< k,l\leq n\} $. Thus the directions with only one
1070: periodic family are represented by:
1071: \[\{(i,n)\in Q_n: \gcd(i,n)=1\}\]
1072: The line from $(0,0)$ to $(i,n)$ intersects exactly one point of the set $x_{n}+\Z^2$
1073: characterized by
1074: \[\left(x\frac{i}{n},x \right)\ \mbox{ with } xi\equiv 1 \pmod{n}.\]
1075: If $i\in \{1,2,...,n-1\}$ runs through the numbers relatively prime to $n$, then
1076: $x$ does this as well. Thus the above regarded length ratio is in any case one of the numbers
1077: \[\frac{\min(x,n-x)}{\max(x,n-x)} \quad \mbox{ with }x\in I^n .\]
1078: We see, exactly two relatively prime numbers modulo $n$, namely $x$
1079: and $n-x$ having the same invariant. If this invariant is complete, which will
1080: be shown in the next Lemma, then the number of different one cylinder orbits under the
1081: operation of the Veech group is given by
1082: \[\frac{1}{2}\varphi(n)=\frac{1}{2}\left|\{x\in\{1,2,...,n-1\}:
1083: \ \gcd(x,n)=1\}\right|=\frac{n}{2}\prod_{p|n\ prime}\left(1-\frac{1}{p}\right).\]
1084: The exceptional case $n=2$ is trivial, because there exists only one direction of one cylinder
1085: in $Q_2$.
1086: \end{proof}
1087: \begin{lemma} Let $\T^2_x$ be a rational marked torus. Two directions $v_1,v_2$ with only one
1088: periodic cylinder are in the same orbit under the operation of the Veech group, if they
1089: have the same ratio
1090: \[I(v_i)=\frac{\min(s^{v_i}_1,s^{v_i}_2)}{\max(s^{v_i}_1,s^{v_i}_2)} \quad i=1,2\]
1091: Here $s^{v_i}_j\ j=1,2$ are the two saddle connections which are on the boundary of the
1092: cylinder in direction $v_i$.
1093: \end{lemma}
1094: \begin{proof}
1095: Obviously the condition is necessary. All that is left is to show that it is sufficent.
1096: That is we have to find an element of the Veech group that is a map between
1097: the two directions with the same invariant. Forgetting for a moment the marked
1098: point $x$ we can map any periodic direction to any other by elements of $SL_2(\Z)$.
1099: Especially the saddle connections in question are mapped on one another. Because
1100: the invariant is the same for both possible orientations of the saddle connections
1101: and $-id\in SL_2(\Z)$ we have found a map, which also transports one marked point
1102: to the other.
1103: \end{proof}
1104: Now we can compute the index with the help of the counting formula again:
1105: \begin{kor} Let $x=[\frac{p'}{n},\frac{q}{n}]$, $n>2$, $gcd(p',q,n)=1$ then
1106: \begin{equation}
1107: [ SL_2(\Z):V(\T^2_{x})]=\frac{n^2}{2}\prod_{p|n \ prime}\left(1-\frac{1}{p^2}\right)
1108: \end{equation}
1109: holds. In the case $n=2$ the index is $3$.
1110: \end{kor}
1111: \begin{proof}
1112: From equality (\ref{kxn}) together with $k(x_n)=\frac{1}{2}\varphi(n)$
1113: for $n>2$ it follows
1114: \begin{equation}\label{index}
1115: [SL_2(\Z):V(\T^2_x)]=\frac{3}{2\pi} \varphi(n)\left[\lim_{T\rightarrow \infty}
1116: \frac{N(G^{V}_{I^n}, T)}{T^2}
1117: \right]^{-1}.
1118: \end{equation}
1119: Then put in the right hand side of equation (\ref{vffinal}) to get the result.
1120: If $n=2$ then $k(x_2)=1$ and we obtain $3$.
1121: \end{proof}
1122: %Especially for $x_{(1/n,1/n)}=(\frac{1}{n},\frac{1}{n})$ and $ x_{n}=(\frac{1}{n},0)$ we have:
1123: %\begin{equation}
1124: % [ SL_2(\Z):V(\T^2_{x_{(1/n,1/n)}})]=[ SL_2(\Z):V(\T^2_{x_{n}})]=\frac{n^2}{2}
1125: % \prod_{p|n \ prime}\left(1-\frac{1}{p^2}\right)
1126: %\end{equation}
1127: \begin{rem}
1128: The two ways of computing the index of the Veech groups of two marked tori might be used
1129: to obtain the asymptotic formula (\ref{vffinal}), counting the lattice points, without using the
1130: second section. One can take the first index computation and the Veech formula together
1131: with the knowledge about the number of cusps $k(x_n)$ (see Proposition \ref{invariante})
1132: to compute $ \lim_{T\rightarrow \infty}N(G^{V}_{I^n}, T)/T^2$ backwards.
1133: \end{rem}
1134: Nevertheless the first given way to compute the index of the Veech groups seems to be a
1135: simple method to calculate for the index of the well known groups
1136: (for example compare S.Lang \cite{lang}
1137: \begin{displaymath}
1138: \Gamma_1(n)=\left\{\begin{array}{ll} \left(
1139: \begin{array}{cc}
1140: a & b \\
1141: c & d
1142: \end{array} \right) \in SL_2(\Z):&
1143: \begin{array}{cc}
1144: \left(
1145: \begin{array}{cc}
1146: a & b \\
1147: c & d
1148: \end{array} \right)& \equiv \left(
1149: \begin{array}{cc}
1150: 1& b \\
1151: 0 & 1
1152: \end{array} \right)\pmod{n}
1153: \end{array}
1154: \end{array}
1155: \right\}
1156: \end{displaymath}
1157: in $SL_2(\Z)$. Either the $\Gamma_1(n)$ are subgroups of index two in $V(\T^2_{x_n})$ if $n>2$, or
1158: in the case $n=2$ $-id \in \Gamma_1(2)$ and $V(\T^2_{x_2})$ is isomorphic to $\Gamma_1(2)$. Thus:
1159: \begin{equation}
1160: [ SL_2(\Z):\Gamma_1(n)]=
1161: \left\{ \begin{array}{cl}
1162: n^2\prod_{p|n \ prime}\left(1-\frac{1}{p^2}\right) & \textrm{ if }\ n>2 \\
1163: 3 & \textrm{ if } \ n=2
1164: \end{array} \right.
1165: \end{equation}
1166: %----------------------------------------------------------------------------
1167:
1168:
1169:
1170:
1171:
1172: \section{Quadratic growth rates and constants on marked tori}
1173:
1174:
1175:
1176:
1177:
1178:
1179:
1180: %----------------------------------------------------------------------------
1181:
1182:
1183:
1184: \subsection{The 2 marked torus $\T^2_x$}
1185:
1186:
1187:
1188:
1189:
1190: %----------------------------------------------------------------------------
1191: We begin with a general observation concerning the directions of periodic
1192: families on the torus. These directions are exactly all the ``rational" directions and
1193: this fact will not change how many points we will mark and wherever
1194: we will place them on the torus. The only thing that might change
1195: is the number of periodic families in a given rational direction, but in this case
1196: they are all of equal length.
1197: For the two marked torus $\T^2_x$ there are always two families in any
1198: rational direction if $x$ is non rational. There are saddle connections
1199: connecting $[0]$ und $[x]$ but which bound no periodic family because they
1200: are not in a rational direction. For rational markings $x$ this can never happen.
1201:
1202: Collecting things for rational $x$: either there is a rational direction with two
1203: periodic families and the bounding saddle connections connecting the same marked
1204: point, or there is only one family but then the two bounding saddle connections
1205: will connect the marked points $[0]$ and $[x]$.
1206: Therefore counting the length $ N_{po}(\T^2_x,T)$ of cylinders of periodic trajectories
1207: for extremal markings $x$ on $\T^2_x$ is easy:
1208: \begin{equation}\label{ponr}
1209: \lim_{T \rightarrow \infty} \frac{N_{po}(\T^2_x,T)}{T^2}=
1210: 2\lim_{T \rightarrow \infty} \frac{N_{po}(\T^2,T)}{T^2}= \frac{6}{\pi}
1211: \quad \forall x \notin \T^2_{rat}
1212: \end{equation}
1213: To treat the rational case we have to subtract the directions where
1214: only one family is from the doubled number of periodic orbits of the
1215: one marked torus. Let us denote the number of single cylinders on $\T^2_x$
1216: with length smaller than $T$ by $N_{po,[x]}(\T^2_x,T)$.
1217: Thus we write:
1218: \[N_{po}(\T^2_x,T)=2N_{po}(\T^2,T)-N_{po,[x]}(\T^2_x,T)+o(T^2)\]
1219: With the help of the discussion at the end of the last paragraph
1220: one computes for a rational point $x=(\frac{p_1}{n},\frac{p_2}{n})$:
1221: \begin{eqnarray}\label{po}
1222: \lim_{T \rightarrow \infty} \frac{N_{po}(\T^2_x,T)}{T^2}&=&
1223: \lim_{T \rightarrow \infty} \frac{1}{T^2}\left(2N({\Z^2}^{V},T)
1224: -N(G^{V}_{I^n},T)\right)=\nonumber\\
1225: &=&\frac{6}{\pi}\left( 1-\frac{1}{2n}\prod_{p|n \ prime}
1226: \left(1+\frac{1}{p}\right)^{-1}\right)
1227: \end{eqnarray}
1228: To count saddle connections is less easy, but again using the last paragraph
1229: (and the discussion before equation (\ref{kxn}) we can write down the associated
1230: growth rates as follows:
1231: \begin{eqnarray*}
1232: & &\lim_{T \rightarrow \infty}\frac{N_{sc}(\T^2_x,T)}{T^2}=\\
1233: & &\lim_{T \rightarrow\infty}\frac{1}{T^2}\left(2N({\Z^{2}}^{V},T)
1234: -2N(G^{V}_{I^n},T)\right)+\\
1235: &+&\lim_{T \rightarrow\infty}\frac{1}{T^2}\sum_{i\in I^n}\left(\frac{1}{i^2}+
1236: \frac{1}{(n-i)^2}\right)
1237: \frac{n^2}{|I^n|}N(G^{V}_{I^n},T) = \cdots
1238: \end{eqnarray*}
1239: Here (\ref{vffinal}) and Proposition \ref{invariante} is used. The
1240: last term needs some explanation. We have seen that $N(G^{V}_{I^n},T)$ already counts the length of the
1241: simple periodic cylinders. It counts as well the length of their boundaries
1242: that is the sum of the length of the two bounding saddle connections.
1243: By Proposition \ref{invariante} there are $\frac{1}{2}\varphi(n)$
1244: different families of orbits of saddle connections labeled by their length ratios.
1245: Equation (\ref{kxn}) shows that these orbits under the Veech group all have the
1246: same growth rate, namely $\frac{2}{\varphi(n)} N(G^{V}_{I^n},T)$.
1247: The relation of the length of a saddle connection bounding a family to the
1248: length of the family is more or less the invariant $I$. Anyway it is one
1249: of the numbers $\frac{i}{n}$ or $\frac{n-i}{n}$ where $i \in I^n$.
1250: With the help of equation (\ref{scaling}) one finds the above expression.
1251: Continuing the evaluation:
1252: \begin{eqnarray}\label{rat}
1253: \cdots &=& \frac{6}{\pi}-\frac{6}{n\pi}\prod_{p|n \ prime}
1254: \left(1+\frac{1}{p}\right)^{-1}+
1255: \frac{6}{\pi}\prod_{p|n \ prime}
1256: \left(1-\frac{1}{p^2}\right)^{-1}\left(\sum_{i\in I^n}
1257: \frac{1}{i^2}\right)=\nonumber \\
1258: &=&\frac{6}{\pi}\left(1 + \prod_{p|n \ prime}\left(1+\frac{1}{p}\right)^{-1}
1259: \left(\prod_{p|n \ prime}\left(1-\frac{1}{p}\right)^{-1}\sum_{i\in I^n}
1260: \frac{1}{i^2}-\frac{1}{n}\right)\right)=\nonumber\\
1261: &=&\frac{6}{\pi}\left(1 + \prod_{p|n \ prime}\left(1-\frac{1}{p^2}\right)^{-1}
1262: \sum_{i\in I^n}\left(\frac{1}{i^2}-\frac{1}{n^2}\right)\right)
1263: \end{eqnarray} The last equality follows with the help of equation (\ref{in}).
1264: For non rational markings one obtains:
1265: \begin{eqnarray} \label{snrat}
1266: N_{sc}(\T^2_x,T)&=&2N({\Z^{2}}^{V},T) + N(x\Z^2,T)+o(T^2)=\nonumber \\
1267: &=& \frac{6}{\pi}T^2+{\pi}T^2 +o(T^2).
1268: \end{eqnarray}
1269: Summarizing:
1270: \begin{satz}The limits
1271: \[ \lim_{T \rightarrow \infty} \frac{N_{po/sc}(\T^2_x,T)}{T^2}\]
1272: exist for all $x$, moreover there are estimates
1273: \[\lim_{T \rightarrow \infty}
1274: \frac{N_{sc}(\T^2_x,T)}{T^2}\leq \lim_{T \rightarrow \infty}
1275: \frac{N_{sc}(\T^2_y,T)}{T^2}=\frac{6}{\pi}+\pi\]
1276: and
1277: \[\lim_{T \rightarrow \infty}
1278: \frac{N_{po}(\T^2_x,T)}{T^2}\leq \lim_{T \rightarrow \infty}
1279: \frac{N_{po}(\T^2_y,T)}{T^2}=\frac{6}{\pi}\]
1280: for $y$ non rational. The inequalities are strict if $x$ is rational.
1281: Moreover for each sequence of rational points $\{x_n\}^{\infty}_{n=0}$ converging to a
1282: non rational point $x$:
1283: \[\lim_{n\rightarrow \infty}\left(\lim_{T \rightarrow \infty}
1284: \frac{N_{sc/po}(\T^2_{x_n},T)}{T^2}\right)= \lim_{T \rightarrow \infty}
1285: \frac{N_{sc/po}(\T^2_x,T)}{T^2}\]
1286: This is the continuity of the quadratic constants at non rational points.
1287: \end{satz}
1288: \begin{proof}
1289: The existence of the limits is proved in the equations (\ref{ponr}), (\ref{po}), (\ref{rat})
1290: and (\ref{snrat}), even more the computations give explicitly the values of the limits.
1291: In the case of saddle connections it
1292: is indeed not easy to see the continuity of the growth rate function
1293: from the expression (\ref{rat}) directly (in fact before the author found the esimate
1294: below he does not believe in the stated formula).
1295: By comparing equation (\ref{snrat}) (the limiting
1296: constant) to the first line of (\ref{rat}) one has to show:
1297: \[\lim_{n \rightarrow \infty}\prod_{p|n \ prime}
1298: \left(1-\frac{1}{p^2}\right)^{-1}\left(\sum_{i\in I^n}
1299: \frac{1}{i^2}\right)= \frac{\pi^2}{6}\]
1300: By using the equation (\ref{sumform}), which states
1301: \[\sum_{i \in I^n}\sum^{\infty}_{l=0}\frac{1}{(nl+i)^2}=\frac{\pi^2}{6}\prod_{p|n \ prime}
1302: \left(1-\frac{1}{p^2}\right)\]
1303: and the estimates
1304: \[\sum_{i \in I^n}\left(\frac{1}{i^2}+\frac{\pi^2}{6}\frac{1}{4n^2}\right)<
1305: \sum_{i \in I^n}\sum^{\infty}_{l=0}\frac{1}{(nl+i)^2}<
1306: \sum_{i \in I^n}\left(\frac{1}{i^2}+\frac{\pi^2}{6}\frac{1}{n^2}\right)\]
1307: we have:
1308: \begin{eqnarray}\frac{\pi^2}{6}\left(1-\frac{1}{n}\prod_{p|n \ prime}
1309: \left(1+\frac{1}{p}\right)^{-1}\right)<\sum_{i \in I^n}\frac{1}{i^2}\prod_{p|n \ prime}
1310: \left(1-\frac{1}{p^2}\right)^{-1}< \nonumber \\
1311: <\frac{\pi^2}{6}\left(1-\frac{1}{4n}\prod_{p|n \ prime}
1312: \left(1+\frac{1}{p}\right)^{-1}\right).
1313: \end{eqnarray}
1314: Clearly if $n\rightarrow \infty$ the right and left hand side of the inequalities converge to
1315: $\frac{\pi^2}{6}$.
1316: The continuity in the case of periodic orbits follows immediately from
1317: equation (\ref{po}).
1318: \end{proof}
1319: %Example:
1320: %
1321: %The billiard in an equal lateral triangle with sidelengthes $l$ has the regular 6-gon,
1322: %where the parallel sides are identified, as invariant phase surface. The
1323: %----------------------------------------------------------------------------
1324:
1325:
1326:
1327:
1328:
1329:
1330:
1331: \subsection{The general case}
1332:
1333:
1334:
1335:
1336:
1337:
1338:
1339: %----------------------------------------------------------------------------
1340:
1341: We generalize the results of the last section to arbitrary many markings,
1342: with the exception that we do not try to find explicit formulas for markings
1343: which are not in general position.
1344: \begin{satz}\label{tlimes}
1345: \begin{itemize}
1346: \item[1.] [continuity] Let $\T^2_{[x_0,...,x_{n-1}]}$ be a n marked translation torus.
1347: Then the limits
1348: \[\lim_{T \rightarrow \infty}\frac{N_{sc/po}(\T^2_{[x_0,...,x_{n-1}]},T)}{T^2}\]
1349: exist.
1350: \item[2.] [ $c^{max}_{sc/po}$ on markings in general position]
1351: The set $\T^{2n}_{gp(sc/po)}$ has full Lebesque measure on $(\T^2)^{n}$
1352: and the inequalities
1353: \[\lim_{T \rightarrow \infty}\frac{N_{sc}(\T^2_{[x_0,...,x_{n-1}]},T)}{T^2}
1354: \leq \lim_{T \rightarrow \infty}\frac{N_{sc}(\T^2_{[y_0,...,y_{n-1}]},T)}{T^2}=
1355: \frac{n(n+1)}{2}\pi+\frac{3n}{\pi}\] and
1356: \[\lim_{T \rightarrow \infty}\frac{N_{po}(\T^2_{[x_0,...,x_{n-1}]},T)}{T^2}
1357: \leq \lim_{T \rightarrow \infty}\frac{N_{po}(\T^2_{[y_0,...,y_{n-1}]},T)}{T^2}=
1358: \frac{3n}{\pi}\] hold for all $(y_0,...,y_{n-1})\in \T^{2n}_{gp(sc/po)}$.
1359: In both cases ($sc$'s and $po$'s) the inequality is strict, if
1360: $(x_0,...,x_{n})\in \T^{2(n+1)}_{rat(sc/po)}$
1361: \item[3.] [continuity at points in general position] \\
1362: If $\ \{(x^i_0,...,x^i_{n})\}^{\infty}_{i=0} \quad (x^i_0,...,x^i_{n})
1363: \in \T^{2n}$ is a sequence with \\
1364: $\lim_{i\rightarrow \infty}(x^i_0,...,x^i_{n})$ $=(y_0,...,y_{n-1})\in
1365: \T^{2n}_{gp(sc/po)}$
1366: then
1367: \begin{equation}
1368: \lim_{i\rightarrow \infty}\left(\lim_{T \rightarrow \infty}
1369: \frac{N_{sc/po}(\T^2_{[x^i_0,...,x^i_{n}]},T)}{T^2}\right)=
1370: \lim_{T \rightarrow \infty}\frac{N_{sc/po}(\T^2_{[y_0,...,y_{n}]},T)}{T^2}
1371: \end{equation}
1372: \end{itemize}
1373: \end{satz}
1374: \begin{proof}
1375: We prove the statements by induction over the number $n$ of marked points.
1376: The idea is to write down, modulo lower order terms, the numbers of saddle connections (or po's) for a given marking
1377: as sums over the numbers for an marking with one marked point less and
1378: the terms occuring when one marks the forgotten point. The results from the first
1379: section will imply the statements.
1380: To start for one point there is nothing to show, moreover we have already seen
1381: the proof for $n=2$. So we assume the statements are true for all n markings
1382: $(x_0,...,x_{n-1})$ of the torus. We add another marking $x_n$.
1383: If the resulting marking is in general position, $(x_0,...,x_n) \in \T^{2(n+1)}_{gp(sc)}$
1384: (for saddle connections) we obtain for saddle connections
1385: \[N_{sc}(\T^2_{[x_0,...,x_{n}]},T)=N_{sc}(\T^2_{[x_0,...,x_{n-1}]},T)+
1386: \sum^{n}_{i=0}N(SC_{(x_i,x_n)},T)\]
1387: furthermore by the definition of general position markings
1388: \[\lim_{T \rightarrow \infty} \frac{N(SC_{(x_i,x_n)},T)}{T^2}=\pi\]
1389: for all $i \in \{0,1,...,n-1\}$. For the case $i=n$: saddle connections
1390: starting and ending in $x_n$ are bounding cylinders of periodic trajectories.
1391: We have to take care of multiplicities when counting these
1392: \[\lim_{T \rightarrow \infty} \frac{N(SC_{(x_n,x_n)},T)}{T^2}=\frac{3}{\pi}.\]
1393: Thus
1394: \begin{eqnarray}
1395: \lim_{T \rightarrow \infty}\frac{N_{sc}(\T^2_{[x_0,...,x_{n}]},T)}{T^2}&=&
1396: \lim_{T \rightarrow \infty}\frac{N_{sc}(\T^2_{[x_0,...,x_{n-1}]},T)}{T^2}+
1397: n\pi+\frac{3}{\pi}=\nonumber \\
1398: &=& \frac{n(n+1)}{2}\pi+\frac{3n}{\pi}. \nonumber
1399: \end{eqnarray}
1400: We assume now that the marking is in general position with respect to periodic orbits.
1401: Since a saddle connection starting and ending in $x_n$ bounds a cylinder of periodic
1402: trajectories, one has to add in each rational direction
1403: a new periodic family. Thus the periodic orbit growth rates for markings in general position are:
1404: \[\lim_{T \rightarrow \infty}\frac{N_{po}(\T^2_{[x_0,...,x_{n}]},T)}{T^2}=
1405: \lim_{T \rightarrow \infty}\frac{N_{po}(\T^2_{[x_0,...,x_{n-1}]},T)}{T^2}+
1406: \frac{3}{\pi}=\frac{3n}{\pi}.\]
1407: If the marking $(x_0,..,x_n)$ is not in general position, it could be divided in
1408: classes of relatively rational points (this is of course meaningful in any case, po's or sc's).
1409: The set of these classes is denoted by $\mathcal{R}_{[x_0,...,x_n]}$.
1410: For each class the Veech theory holds
1411: (see \cite{ve89,ve92, ve98}) and guaranties that all the limits exist.
1412: For the inequalities between the limits one observes: if the marked points
1413: are rational there exists always a direction in which there are strictely less than
1414: the maximal number of $n$ cylinders of closed trajectories.
1415: The orbit of this direction under the operation of the
1416: Veech group leads to a positive asymptotic growth constant (Proposition 6.1 in \cite{gj97}).
1417: But a lower (than maximal possible) number of closed cylinders in a given direction
1418: causes ``multiplicities" (on has to count only visible points).
1419: Hence the limit growth rates are in both cases (po's and sc's) smaller than the ones for markings in general position.
1420: This shows: if ther is one class in $\mathcal{R}_{[x_0,...,x_n]}$ containing more than one point,
1421: then the growth rates have to be smaller than for markings in general position.
1422: To compute the growth rates in the cases of periodic orbits, we observe that
1423: each class $m\in \mathcal{R}_{[x_0,...,x_n]}$ defines a marked Veech torus $\T^2_m$,
1424: by marking $\T^2$ only with the points of that class. Thus:
1425: \begin{equation}
1426: N_{po}(\T^2_{[x_1,...,x_{n}]},T))=\sum_{m \in \mathcal{R}_{[x_0,...,x_n]} }
1427: N_{po}(\T^2_{m},T) +o(T^2).
1428: \end{equation}
1429: Here is used that by definition between points of different classes there are never
1430: saddle connections with rational slope (or equivalently: with bound a closed trajectory).
1431:
1432: All remains is to count the saddle connections between points of different classes.
1433: To do this let $x_i$ be a marked point not rationally equivalent to $x_n$ and
1434: $SC^{(x_0,...,x_{n-1})}_{(x_i,x_j)}$ the set of saddle connections connecting
1435: $x_i$ and $x_j$ on the n marked torus $\T^2_{[x_1,...,x_{n-1}]}$.
1436: What we have to count are sets of saddle connections from a given set of saddle connections
1437: between two marked points (for example the set $SC^{(x_0,...,x_{n-1})}_{(x_i,x_j)}$)
1438: which are parallel and shorter to another set of saddle connections (for example $SC_{(x_i,x_n)}$).
1439: However by Lemma (\ref{schnigi}) and Proposition (\ref{countnrat})
1440: sets of this kind are always lattices modulo a mistake of order $O(T)$, which does not count
1441: in the quadratic limit.\\
1442: There are two different cases: First, there is a nonempty subset in
1443: $SC_{(x_i,x_n)}$ consisting of saddle connections
1444: parallel to saddle connections in the set $SC^{(x_0,...,x_{n-1})}_{(x_i,x_j)}$ and shorter in length.
1445: The set of all indices $j$ of such $x_j$ are denoted with $I$.
1446: Second, a nonempty subset $SC_{(x_i,x_n)}$ is parallel to one
1447: from $SC^{(x_0,...,x_{n-1})}_{(x_i,x_j)}$, but the first are longer.
1448: As above denote the set of all indices of such $x_j$ with $K$.
1449: Moreover the sets of saddle are (modulo sets of lower order growth rates)
1450: given by some lattices $G^{(x_i,x_n)}_{(x_i,x_j)}$ or $G^{(x_i,x_j)}_{(x_i,x_n)}$ respectively.
1451: We can write:
1452: \begin{equation}
1453: N(SC^{(x_0,...,x_n)}_{(x_i,x_n)},T)=N(G_{(x_i,x_n)},T)-\sum_{x_j\in I}
1454: N(G^{(x_i,x_j)}_{(x_i,x_n)},T) +o(T^2).
1455: \end{equation}
1456: Here $G_{(x_i,x_n)}$ denotes the lattices of saddle connections between $x_i$ and $x_n$ on the
1457: two marked tori $\T^2_{[x_i,x_n]}$. For all $j\in K$
1458: \begin{equation}
1459: N(SC^{(x_0,...,x_n)}_{(x_i,x_j)},T)=N(SC^{(x_0,...,x_{n-1})}_{(x_i,x_j)},T)
1460: -N(G^{(x_i,x_n)}_{(x_i,x_j)},T)+o(T^2).
1461: \end{equation}
1462: Modulo terms of strictly lower order size the above expressions are all lattices,
1463: hence they have quadratic growth rate limits.
1464: After these steps for all markings $[x_i]$ the induction is complete.
1465: From the two equalities it follows immediately:\\
1466: if intersections of the above kind have a nonzero quadratic growth rate
1467: (so the marking is rational in the sense of our definition) than the associated growth rates are
1468: smaller than for a non rational marking. The continuity follows from the fact that by approximating
1469: a marking in general position all the denominators of all the length ratios are growing over all bounds.
1470: \end{proof}
1471:
1472: %---------------------------------------------------------------------------
1473:
1474:
1475:
1476: \subsection{Branched coverings of marked tori}
1477:
1478:
1479:
1480:
1481:
1482: %---------------------------------------------------------------------------
1483: %To start let $S ` ``-{^\pi}!!!>" \T^2$ be a branched covering
1484: %where $S$ is a compact orientable surface and $\pi$ is a homeomorphism. We regard the images
1485: %of the branch points $(x_1,...,x_n)$ under $\pi$ to be the marked points of $\T^2$.
1486: %By varying $(x_1,...,x_n)$ in $\T^{2n}$ we get a family of translation structures
1487: %on $S$ in the following way. Take a map $\varphi \in H_{0}(\T^2)$ that is
1488: %a map of the torus which is homeomorphic to the identity and define the
1489: %translation structure $(S,\omega_{\varphi})$ on $S$ where
1490: %\[\omega_{\varphi}= \]
1491: %Like
1492: Because of the simplicity of our methods we are not able to conclude anything about (families of)
1493: branched coverings of tori, without using the existence of the growth rate limits by the
1494: method explained in \cite{ems}. Even the sets where the growth rate
1495: is the biggest is not seen as just the preimage of the non rational markings on
1496: the torus. For closed cylinders of geodesics at least it has to be contained in this set.
1497: This is because the geometry and combinatorics of the covering might cause new sets
1498: where the growth rates are different.
1499: Geometrically the reason for this is that the length spectrum of the cylinders of periodic
1500: trajectories is connected in a way to the saddle connections bounding them, which depends on the
1501: covering and where the singular points are.
1502: If one makes the simplifying assumption that all inverse images of the marked points
1503: are itself marked (such coverings are called `balanced coverings' by some people), then
1504: from Theorem \ref{tlimes} one can conclude:
1505: \begin{kor}
1506: Let $\V \stackrel{\pi}{\rightarrow} \T^2_{[x_0,...,x_n]}$ a covering
1507: branched only over $(x_0,...,x_n)$ with the induced translation structure
1508: $\pi^{\ast}dz$ . Further denote by ${\V}_C$ the translation structure
1509: where each inverse image of a marked point is itself marked .
1510: Then, after rescaling the volume of $ \V$ (with respect to the induced flat
1511: metric) to one:
1512: \begin{equation}
1513: \lim_{T \rightarrow \infty}\frac{N_{sc}(\V_C,T)}{T^2}=
1514: deg(\pi)^2 \lim_{T \rightarrow \infty}\frac{N_{sc}(\T^2_{[x_0,...,x_{n}]},T)}{T^2}
1515: \end{equation}
1516: ($deg(\pi)$ is the degree of $\pi$). Moreover, if some of the markings are artificial
1517: (i.e. not singular points of the induced metric) then the growth rates after ``unmark"
1518: any subset of these points will be smaller than the above one. Thus
1519: \begin{equation}\label{verh}
1520: \limsup_{T \rightarrow \infty}\frac{N_{sc}(\V,T)}{T^2}\leq
1521: deg(\pi)^2 \left(\frac{n(n+1)}{2}\pi+\frac{3n}{\pi} \right).
1522: \end{equation}
1523: \end{kor}
1524: Remark: Similarly one can write down the corresponding estimate for maximal periodic cylinders. \\
1525: If one includes Ratners Theorem as done in \cite{ems} then
1526: the $\limsup$ in equation (\ref{verh}) can be replaced by a limit.\\
1527: \begin{proof}
1528: The statement is a direct consequence of the fact that each saddle connection
1529: on the torus has exactly $deg(\pi)$ preimages on $\V_C$ and Theorem \ref{tlimes}.
1530: The second part is clear, because by removing nonsingular markings,
1531: the directions in which there are saddle connections are not changed.
1532: But in any such direction the number of saddle connections could decrease and their length increase.
1533: \end{proof}
1534: Remark: With the results of Veech \cite{ve90} every translation surface can
1535: be approximated by torus coverings in spaces of abelian differentials ${\cal A}(g,P)$.
1536: Moreover torus coverings are equally distributed with respect to the
1537: Liouville measure $\mu$ mentioned in the introduction.
1538: Since periodic cylinders and saddle connections are locally stable with
1539: respect to deformations in this spaces we have some sort of control
1540: on their numbers by the Corollary. But because of the increasing
1541: number of artifical markings while approximating a general translation surface
1542: or equivalently the increasing degree of the covering, the estimate above
1543: is to weak to predict for example quadratic bounds of the growth rates of
1544: po's or sc's in general. On the other hand the equidistribution of torus coverings in ${\cal A}(g,P)$
1545: is used to compute the $\mu$ volume of these spaces of abelian differentials see \cite{eo}.
1546: The knowledge of these volumes in turn is the main step to evaluate Siegel Veech formulas, compare
1547: \cite{ems, emz}.
1548:
1549: \begin{thebibliography}{99}
1550:
1551: \bibitem[E]{esk98} Eskin, Alex, Counting problems and semisimple groups,
1552: Doc. Math. J. DMV , Extra Volume ICM II (1998) 539-552.
1553: \bibitem[EM]{esmr98} Eskin, Alex ; Masur, Howard. Pointwise asymptotic
1554: formulas on flat surfaces, Ergodic theory and dynamical systems (2001) 37 pages.
1555: \bibitem[EMS]{ems} Eskin, Alex; Masur, Howard; Schmoll, Martin.
1556: Billiards with barriers, preprint (2001) 37 pages.
1557: \bibitem[EMZ]{emz} Eskin, Alex; Masur, Howard; Zorich, Anton.
1558: The Siegel Veech constants, In preparation.
1559: \bibitem[EO]{eo} Eskin, Alex; Okounkov, Andrei.
1560: Asymptotics of numbers of branched coverings of a torus and
1561: volumes of moduli spaces of holomorphic differentials. preprint (2000) 42 pages.
1562: To appear in Inventiones Math. .
1563: \bibitem[GJ96]{gj96}Gutkin, E.; Judge, C.; The geometry and
1564: arithmetic of translation surfaces with applications to polygonal
1565: billiards, Math. Res. Lett.3 (1996), 391-403.
1566: \bibitem[GJ97]{gj97}Gutkin, E., Judge, C.; Affine mappings of
1567: translation surfaces, Geometry and Arithmetic, Duke Journal (1999).
1568: \bibitem[Lang]{langa}Lang, Serge; Algebra. Second edition. Addison-Wesley Publishing Co.,
1569: Reading, Mass., 1984. xv+714 pp. ISBN: 0-201-05487-6
1570: \bibitem[Lang]{lang}Lang, Serge; Introduction to modular forms,
1571: Grundlehren der mathematischen Wissenschaften 222, Springer Verlag 1976.
1572: \bibitem[M86]{m86}Masur, H.; Closed trajectories for
1573: quadratic Differentials with an application to billiards , Duke
1574: Math. J. 53 (1986) 307-314.
1575: \bibitem[M88]{m88}Masur, H.; Lower bounds for the number of saddle
1576: connections and closed trajectories of a quadratic differential,
1577: in Holomorphic Functions and Moduli 1, 215-228, MSRI Publ. 10,
1578: Springer Verlag 1988.
1579: \bibitem[M90]{m90}Masur, H.; The growth rate of trajectories of a
1580: quadratic differential, Ergodic Theory and Dynamical Systems 10
1581: (1990), 151-176.
1582: \bibitem[S]{se}Serre, J.P.; Cours d'arithm\'{e}tique, Le Math\'{e}maticien,
1583: Presses Universitaires de France,(1970).
1584: \bibitem[T]{ten}Tenenbaum, G.; Introduction \`{a} la th\'{e}orie analytique
1585: et probabiliste des nombres, Cours Sp\'{e}cialis\'{e}s, Soci\'{e}t\'{e} Math\'{e}matique
1586: de France,(1995).
1587: \bibitem[V86]{ve86} Veech, W.; The Teichm\"{u}ller geodesic flow,
1588: Annals of mathematics, 64 (1986),441-530.
1589: \bibitem[V89]{ve89} Veech, W.; Teichm\"{u}ller curves in Moduli space,
1590: Eisenstein series, and an application to triangular billiards,
1591: Invent. Math. 97 (1989), 553-583.
1592: \bibitem[V90]{ve90} Veech, W.; Moduli spaces of quadratic
1593: Differentials, Journal D'Analyse Math\'{e}matique Vol.55 (1990),341-379.
1594: \bibitem[V92]{ve92} Veech, W.; The billiard in the regular polygon,
1595: Geom. and Funct. Analysis 2 (1992),341-379.
1596: \bibitem[V98]{ve98} Veech, W.; Siegel measures. Ann. of Math. (2) 148 (1998), no. 3, 895-944.
1597: \bibitem[Vrb96a]{va} Vorobets, Y.; Planar structures and billiards
1598: in rational polygons, Russ. Math. Surveys 51 (1996), 177-178.
1599: \bibitem[Vrb96b]{vab} Vorobets, Y.; Planar structures and billiards
1600: in rational polygons: The Veech alternative, Russ. Math. Surveys 51
1601: (1996), 779-817.
1602: \bibitem[ZK]{zk} Zemlyakov, A.N., Katok, A.B.; Topological transitivity of Billiards
1603: in polygons. Matem. Zametki 18(2)(1975) pp. 291-300. english
1604: translation in Math. notes 18(2) (1976) pp 760-764.
1605: \end{thebibliography}
1606: \vspace*{2cm}
1607: \begin{center}
1608: Authors present adress:
1609: \vspace*{5mm}\\
1610: Technische Universit\"at Berlin \\ Strasse des 17. Juni 136 \\
1611: Sekr. MA 7-1/ MA 7-2 \\ D-10623 Berlin \\
1612: \normalsize{e-mail: \texttt{schmoll@math.tu-berlin.de} }
1613:
1614: \end{center}
1615:
1616: \end{document}
1617:
1618:
1619:
1620:
1621:
1622:
1623:
1624:
1625:
1626:
1627:
1628:
1629:
1630:
1631:
1632:
1633: